Protein-based Nanoparticles for Imaging and Drug Delivery Applications

by

Ioana Laura Aanei

A dissertation submitted in partial satisfaction of the

requirements for the degree of

Doctor of Philosophy

in

Chemistry

in the

Graduate Division

of the

University of California, Berkeley

Committee in charge:

Professor Matthew B. Francis, Chair

Professor Michelle C. Chang

Professor David F. Savage

Professor Andreas Martin

Fall 2016

Protein-based Nanoparticles for Imaging and Drug Delivery Applications

Copyright © 2016

By: Ioana Laura Aanei

1

Abstract

Protein-based Nanoparticles for Imaging and Drug Delivery Applications

by

Ioana Laura Aanei

Doctor of Philosophy in Chemistry

University of California, Berkeley

Professor Matthew B. Francis, Chair

Protein cage architectures such as viral capsids, heat shock proteins, and ferritins are naturally occurring nanoscale structures with exquisite homogeneity and stability. This dissertation work describes endeavors toward the use of three different protein cage ar-chitectures: the MS2 bacteriophage, a mutant of the Tobacco Mosaic Virus (TMV), and the nanophage scaffold (derived from fd phage) as therapeutic and imaging agent deliv-ery systems. The cages were reacted with several types of organic molecules, such as drug molecules, imaging tracers, and targeting ligands. Towards the development of the platforms for therapeutic delivery, the antitumor agent doxorubicin was covalently bound on the protein scaffolds via a pH sensitive linker. The effects of these constructs on cul-tured mammalian cells, their biocompatibility, in vivo tumor homing and biodistribution were investigated. Tumor growth inhibition studies were also performed with these struc-tures. Together, the results presented herein demonstrate the utility of the protein cages as robust nanoscale platforms for the delivery of cargo.

i

Dedicated to my family and friends, who have been with me through thick and thin.

ii

Table of Contents

Chapter 1: Protein-based nanoparticles 1

1.1 Nanoparticles in medicine 21.2 Viral capsids as protein-based nanoparticles 41.3 Chemical modifications of protein materials 51.4 Conclusions 91.5 References 9

Chapter 2: MS2 bacteriophage as an imaging agent 16

2.1 Current imaging techniques 172.2 The MS2 bacteriophage 182.3 Evaluating the stability of MS2 capsids 192.4 Positron Emission Tomography and biodistribution of MS2 capsids 202.5 Conclusions 232.6 Materials and methods 232.7 References 27

Chapter 3: MS2-antibody conjugates for cancer imaging applications 30

3.1 The tumor microenvironment 313.2 Current examples of targeted virus-like particles 323.3 Synthesis of MS2-antibody conjugates 333.4 Characterization of MS2-antibody conjugates 363.5 Charge Detection Mass Spectrometry for determining the number of antibodies attached to the MS2 capsids 373.6 In vitro interactions with cancer cell models 413.7 In vitro stability of MS2-antibody conjugates 443.8 In vivo distribution of MS2-antibody 453.9 In vivo assesement of immunogenicity of MS2-antibody constructs 483.10 Efforts to remove endotoxins 483.11 Conclusions 493.12 Materials and methods 493.13 References 57

iii

Chapter 4: MS2 as a cardiovascular imaging agent 62

4.1 Imaging techniques for cardiovascular diseases 634.2 Using the MS2 capsid to target atherosclerotic plaques 644.3 Synthesis and characterization of VCAM-targeted agents 654.4 Pathology of ApoE knockout mice 674.5 Biodistribution and plaque-targeting ability of VCAM-targeted agents 674.6 Conclusions 704.7 Materials and methods 704.8 References 73

Chapter 5: Viral capsids for glioblastoma detection and treatment 75

5.1 The effect of nanoparticle properties on tumor uptake 765.2 Protein scaffolds compared in this study: MS2, TMV, NP and IDP 775.3 In silico studies of nanoparticle extravasation 795.4 Synthesis and characterization of agents 805.5 Positron Emission Tomography in glioblastoma models 805.6 Optical imaging in glioblastoma models 835.7 Drug delivery studies in glioblastoma models 845.8 Conclusions 885.9 Materials and methods 895.10 References 96

Appendix 101

iv

Acknowledgements

My journey to today was made possible and enjoyable by a series of extraordinary people I met along the way. I am eternally grateful for their help and their presence in my life.

My love for science was discovered and encouraged by my 7th grade chemistry teach-er, Mihai Draghici. I will never forget my first experiments and his words of encourage-ment. Mariana Doicescu is a wonderful teacher with a heart of gold and a patience out of this world. She prepared me for the International Chemistry Olympiad and maintains a special place in my heart. Ion Bararu is the big dreamer, the go-getter, the Science enthu-siast. He was my physics teacher in high school and the mentor of our team of 12 girls in the final for the NASA Space Settlement International Competition in Houston. I learned so much from him, varying from survival skills, to advanced physical concepts, to not being ashamed of your inner geek. I would also like to acknowledge my awesome team-mates for the NASA competition – a group of powerful, intelligent, driven women who will change the world. I am grateful to Daniel Tofan, who was my rock for many years, helped me with college applications and to develop my love for chemistry.

I am extremely grateful for the opportunity I had to attend Caltech. The people I met there were absolutely amazing and the education I got (especially the extremely difficult exams, the wonderful social sciences classes and the all-nighters) prepared me for any-thing that might come my way. I was fortunate to get to work in the laboratory of Prof. David Tirrell under the guidance of James Van Deventer in my freshman year at Caltech. Prof. Tirrell was very supportive and encouraging and Jim was patient as I made my debut in a chemistry lab. During my summer at the University of Washington in Seattle, Prof. Suzie Hwang-Pun was very welcoming and helped me learn about patterning and non-viral delivery methods. Chung H.K. Wang (Kat) was a kind and inspiring mentor and made my stay very enjoyable. During my last year at Caltech, I worked in the Mark Davis lab with Chung H.J. Choi (Jonathan). There, I learned about breast cancer targeting with gold nanoparticles and made many good friends.

U.C. Berkeley has been my home for the last 5.5 years. Looking back, I know I made the right choice to come here for graduate school. The department has some of the most talented, passionate, intelligent and creative people in the world and I am extremely hon-ored to have been their colleague for a few years. As I was a part of the Chemical Biology program, I got the opportunity to rotate in 3 labs. In the Christopher Chang lab, I worked with Sheel Dodani and learned a lot about organic synthesis and fluorescent probes. In the Michelle Chang lab, I worked with Mark Walker and Ben Thuronyi, two wonderful and patient mentors who taught me about protein expression and synthetic biology. I am very grateful to both of these labs for their warm welcoming and the knowledge they shared with me.

After my rotations, I joined the Francis lab, where I met some of the nicest, most tal-ented people I know. I have greatly enjoyed my time in the lab and I am forever grateful for all the support, from showing me where to find supplies to helping me with the Gradu-

v

ate Researcher Seminar and qualifying exam. The postdoctoral students in our lab have been outstanding and we benefited greatly from their experience. Michelle Farkas was the first postdoc I worked with and she graciously showed me the ropes of in vivo exper-iments. I will never forget the late nights spent running experiments with radioactivity, or the time when we fell asleep in the lab, running a gel at 3 am. Adel Elsohly is one of the kindest, funniest, most reliable, intelligent, creative and talkative people I know! We have worked closely together and share many memorable moments, including our drives to Stanford, our D’yar lunches and boba runs. I am forever grateful for everything he taught me and I hope that our friendship will only grow stronger with time. Meera Rao is kind and sweet, but a beast when it comes to working out! She is a strong, intelligent woman with a lot to offer and I am so happy that I got a chance to befriend her. Henrik Munch brought a European flavor to the lab, reminding us of the importance of soccer and beer. Ayodele Ogunkoya is a laid back, happy person, who refuses to play by other people’s rules. I hope that he will always be this happy and excited about science. Rafi Mohammad always had a smile on his face and a good question to ask. Alex Hoepker is a passionate scientist with big ambitions and the resources to achieve his goals. I wish him best of luck in all his future endeavors. Christian Rosen is a fast learner and a driven scientist. I am sure he will do well in whatever he sets his mind to. Carson Bruns has brought a completely different flavor in our lab and opened our eyes to the beauty of the mechani-cal bond. Ariel Furst also brought her experience in electrochemistry and her passion for baking, for which we are grateful.

The graduate students in the Francis lab are exquisite scientists and wonderful peo-ple. I will greatly miss our holiday parties, ski trips, group meetings and box-soccer games on the hallways. Christopher Behrens was my mentor during my rotation and I inherited his desk after he left. He graciously taught me about MS2 expression and bioconjugation techniques. Gary Tong was one of the senior members in the lab when I joined and kindly shared with us his music taste. Wesley Wu was a model for work ethics, while Kanna Palaniappan always had a good suggestion to share during subgroup meeting. Michel Dedeo was very talented at fixing everything in the lab and was a very creative scientist. Zac Carrico and Kristen Seim were models for hard work and pioneered very interesting projects in our lab. Mike Coyle has the ability to pick up new concepts very fast and come up with excellent questions. Allie Obermeyer is the whole package: her excellent commu-nication skills, creative solutions for troubleshooting, beautiful graphics and impressive work ethics have landed her a dream job at Columbia University. I wish her best of luck with her lab! Katherine MacKenzie was always encouraging and friendly and had excel-lent advice for troubleshooting experiments. Stacy Capehart worked hard and ran even harder. She always had something nice to say and good suggestion during meetings. Jeff Glasgow is one of the nicest people I know and his calm demeanor and excellent scientif-ic insight made it a pleasure to work with him. Dan Finley’s humor and friendly personality made him one of the main characters at all our gatherings. Leah Witus was one of the most helpful, kind, intelligent people I met in grad school. Amy Twite is a great scientist and a dear friend. Her projects were some of the most challenging, interesting and revo-lutionary undertakings in our lab. Troy Moore is a walking encyclopedia of knowledge, be it regarding to Chemistry or where to find the best sushi in town. Kareem El Muslemany

vi

dared to dream big and worked very hard to get to where he wanted to. Abby Knight al-ways brought a note of positivity to the lab. Chawita (Jelly) Netirojjanakul is an extremely talented scientist with an impressive drive. Jess Lee is nice, considerate and friendly and just a pleasure to be around. Nina Hentzen was a temporary addition to our lab, but it was great to work with her and brainstorm how to troubleshoot experiments together.

Jim MacDonald is one of the most talented chemists I know and I will miss walking into 748 and hearing his music blasting. Richard Kwant has one of the most beautiful minds out there and I am sure he will be able to tackle any challenge is thrown at him. It was a pleasure to get to know Jake Jaffe (and his better half, Danielle). He has great scientific insight and excellent work ethics. Jenna Bernard is one of the two people in this world who can convince me to drink more than I should. Over the years, she (and her husband Matt) became one of my closest friends and we share many great memories, from writing grants together to sweating in the gym and doing a Tough Mudder during a sand storm. Kanwal Palla sat at the desk next to me for 5 years. Believe it or not, we still didn’t get enough of each other, so we continued to hang out outside of the lab, either grading on her wonderful couch or going on our car adventures. She was my rock during many chal-lenges and I am forever grateful for her friendship and advice.

Joel Finbloom is extremely talented and it was a pleasure to work with him on several occasions. I think his calm demeanor, excellent memory and work ethics will take him very far. Rapeepat (Am) Sangsuwan is one of the nicest people I met and she always has a smile on her face. Matt Smith is a creative scientist with an eccentric side. He makes science fun and fashion seem ridiculous. I will miss Adam Childs’ humor and Rachel Bisiewicz’s baking (those brownies!). Emily Hartman is a talented and passionate scien-tist and has the drive to tackle difficult projects. I have been impressed by her kindness and strength and wish her best of luck in all her future endeavors. Kristin Wucherer is sweet, kind, sharp, hard working and thoughtful. We didn’t get to be gym buddies for a long time, but I will always have fond memories of our interactions. Tyler Hurburt kindly accepted me calling him “Kanwal” for half a year while I was missing her. He even threw paperclips at me (as she used to do), in her name. He is funny and dedicated and a plea-sure to be around. Marco Lobba is always happy and excited about science. I think he is a great addition to our lab and I wish him best of luck with his challenging project. Sarah Klass has the drive, curiosity, intelligence and work ethics to achieve great things. Jing Dai is a gifted scientist with a quiet confidence that comes out just at the right time. I wish her best of luck with her projects. Melanie Gut has quickly integrated in our lab and hit the ground running on her project. She is kind and friendly and it was great to get to know her. Nick Dolan is a new addition to our lab, but I already had a chance to see his dedication and curiosity shine through.

Our undergraduate students have been a wonderful presence in the lab. Effie Zhou and Susanna Elledge are two extremely talented scientists, on their way to big things. Hyun Shin (Ken) Park worked with me for almost two years. I saw him develop into an independent scientist and get into all the graduate schools he applied to. I am very proud of him and wish him best of luck with his future career. Nielson Weng was a very talented undergraduate and his hard work landed him a position in the Medical Scientist Training

vii

Program at Stanford. Jane Honda and Zoe Adams have added their positive attitude to the pleasant work environment in the lab. Kenneth Han is a very capable student and it has been a pleasure to see him grow. Sorry I’ve been terrorizing you with the WD40!

I would like to thank my friends from other groups in the Chemistry Department. Ma-son Appel has been a constant source of joy and random trivia. Frances Rodriguez-Ri-vera is a dear friend; I will miss our long talks and coffee dates. I am sure that her talent will be appreciated at Merck and she will achieve great things. David Spiciarich has taken up a challenging and clinically relevant project and his creativity and hard work have helped him make great progress. I will miss our lunches and enjoying chocolate together. I will also miss going to the 8th floor to bother Ioannis Mountziaris while he was running a column. His infectious laughter and unique sense of humor will be greatly missed. Cheri Ackerman is one of the most driven, talented, no-nonsense scientists I know. Her ambi-tion and acuity will take her far. Eva Nichols is funny and humble and an extraordinary scientist. She was also my guide in the world of climbing, for which I am truly grateful. Karla Ramos Torres brought her insight and latin flavor into the Chris Chang lab, has been a good friend and a great scientist. Ala Bunescu (and her husband Pierre-Yves) is a close friend and always took the time to listen to my woes, either science- or life-related. Her cooking is exquisite and her grasp of Chemistry is admirable. Aditya Gottumukkala introduced me to BPEP and encouraged me to follow my dreams. I will miss our conver-sations and our long walks. I am also thankful to the other Chemical Biology students, Carisa Orwig and the class of 2011.

My work would have not been possible without the help from the great people at several core facilities. I would like to acknowledge Dr. Byron Hann and his team at the UCSF Preclinical Core and Jennifer Bolen at the UCSF Mouse Pathology Core. Holly Aaron at the Molecular Imaging Center and Hector Nolla at the Flow Cytometry Facility were always available to help and did it with efficiency and kindness. The Tissue Culture Core staff was helpful and provided cell lines and advice. I would like to thank Nick Ellis in the Miller lab for his help with the microtome and the Bertozzi lab and Chris Chang lab for kindly allowing me to use some of their instruments. I want to acknowledge CARA for instrumentation and funding and Dr. Thorsten Staudt for scientific insight.

During my graduate studies, I had the honor to collaborate with many wonderful sci-entists. Dr. James O’Neil taught me everything I know about radioactivity and was a great mentor. I am truly grateful for his friendship and support. Mustafa Janabi, Nick Vandehey and Steve Hanrahan helped me learn about being creative and resourceful and making an experimental setup work. The cyclotron gang has been my second “lab home” and I have many great memories from my time spent at LBL. Dr. Youngho Seo in the Department of Radiology and Biomedical Imaging at UCSF is a model for work ethics and scientific curiosity. He has been a great source of support for all our in vivo experiments and it was a pleasure to work with him. Stephanie Taylor, Melanie Regan, Sergio Wong and Tony Huynh were all delightful collaborators and very nice people. We spent many late nights working on complicated experiments and I will be forever grateful for their dedication and hard work. Dr. Henry VanBrobrocklin and Joe Blecha were instrumental in performing experiments with radioactivity at UCSF and I am very thankful for their help. Prof. Sean

viii

Bendall, Prof. Gary Nolan and Astraea Jager from the Department of Microbiology and Immunology at Stanford University graciously taught us about mass cytometry and the impressive potential it has for studying cellular processes. Kevin Bond and Prof. Martin Jarrold from Indiana University showed us the power of Charge Detection Mass Spec-trometry and made the collaboration pleasant and interactive. Tiras Lin, Preyas Shah and Prof. Eric Shaqfeh from Stanford are courageously taking on the challenging task of modeling the flux of nanoparticles through tumor pores. We have learned a lot from this collaboration and I am grateful for the insight into the world of computational chemistry. We also had the opportunity to collaborate with a wonderful group of scientists from UCSF: Dr. Tomoko Ozawa, Raquel Santos, Edgar Lopez-Lepe, Theodore Nicolaides, MD and Mitchel Berger, MD. We have learned about glioblastoma and the latest techniques to combat it. I am very grateful for their dedication, professionalism and affability.

I want to take this chance to thank my thesis committee (Prof. David Savage, Prof. Michelle Chang and Prof. Andreas Martin) for their support over the years and always pushing me to be better. I want to thank Chona De Mesa and Lynn Keithlin for their support with all the complications of graduate school. Christine Balolong has been an in-valuable resource. Her dedication and impressive efficiency have gotten me out of many problematic situations and for that I am truly grateful. I want to acknowledge all the other staff in the Department, including the custodians, the staff at the shops, the Coffee lab, HR and building management. Everyone has been so nice and helpful and they made me feel welcome!

I was very lucky to meet my Thriving in Science crew: Ryan, Jeanne, Allison, Sara and Andrew. Thank you, Sara, for always encouraging us and being positive! Jeanne, Allison and Andrew – you have been great thesis buddies and made the writing process 100 times better. I will miss our thesis parties, with the dance and coffee breaks. I am very grateful to the BPEP team: Naresh, Michaela, Lance, Mark, Antoine, Tyler and Robert. You guys have introduced me to the world of entrepreneurship and we have shared won-derful moments together.

My college friends have been of paramount importance to maintaining my sanity during graduate school. They have always been next to me and for that I am truly grateful. Janesha, Sonal, Kevin, Michael, Abhinav and Best are some of the most caring, genuine, fun-loving, strong, intelligent, adventurous people I know and I am very blessed to have you in my life. Crina has been my friend for 14 years. She is tough, sharp, hard working, candid and a wonderful friend. Thank you for always being by my side! My friends from boxing and salsa have kept me balanced during the tumultuous times in grad school. Thank you, Victor, for teaching me how to be strong and compassionate, driven and mindful. Svetlana has been my mentor and source of wisdom and I am very grateful for her time, the wonderful brunches and the honest insight. Jeremy has brought much joy into my life. We share many great memories and I hope we will make many more. I am grateful for all the times he calmed me down and helped me find solutions to my prob-lems, as well for pushing me outside of my comfort zone and helping me see things from a different perspective.

ix

My family is the main pillar of my life. No matter how far I am from home, I feel their support and their love. I would not be here without them and I am eternally grateful for their help. Mama, Tata, Radu, Oana, Bunica si Unchiu’ Radu, va multumesc si va iubesc mult!

My graduate school experience would not have been as enlightening and enjoyable without the guidance from Prof. Matthew Francis. Matt has been the best mentor I could have asked for. He gave me the liberty to explore new avenues and advice on what to do when I got stuck. He has been kind and supportive and created a great environment to work in. My favorite thing about Matt is his ability to put things in perspective. I would walk into his office disillusioned and upset with Science and he knew just what to say to remind me of the complexity of the project, the progress we had made and what the important parts are. I would walk out of his office with new wind under my wings, ready to tackle a new experiment. I am extremely grateful for him taking a chance on me and helping me become a more experienced scientist.

Thank you to everyone who helped me grow, learn and have fun!

I believe that’s all, folks. Now let’s get to the Science!

1

Chapter 1: Protein-based nanoparticles

Abstract

Over the past several decades, significant advances have been made in the funda-mental understanding of human diseases. As a result of this wealth of information, the field of nanotechnology emerged, and maintains its role as a promising source of novel therapies and diagnostic tools. Numerous types of nanoparticles have been synthesized with a wide array of properties, each with its own advantages and disadvantages. Pro-tein-based nanoparticles present numerous benefits, including structural homogeneity, ease of production and modification, and biocompatibility. This work will present methods for modification of protein scaffolds, with the goal of developing imaging and drug delivery platforms with translational potential.

2

1.1 Nanoparticles in medicine

Advances in modern medicine have led to the development of novel treatments and improvements in patient quality of life. Nevertheless, the small molecules currently used as therapeutics for these afflictions are non-specific to the region of interest, and can lead to systemic toxicity.1 The field of nanomedicine was born out of the necessity to increase the control over the distribution of small molecule drugs, in the hopes of reducing their detrimental effects on healthy tissue. Drug carriers can also alleviate undesired proper-ties of therapeutics such as poor solubility, rapid degradation, and fast clearance.2

Nanoparticles have gained a lot of popularity, especially in the field of cancer ther-apeutics. It is hypothesized that certain nanoparticles may preferentially accumulate in tumor tissues due to the enhanced permeability and retention (EPR) effect.3 It has been shown that in certain tumor models the permeability of the vasculature increases, such that carriers are able to pass through the endothelial barrier and accumulate in the tumor environment. In some cases, large spaces (600 – 800 nm) between endothelial cells can be observed, in contrast to the normal blood vessels that present a tight packing of endo-thelial cell.4 Additionally, the lymphatic drainage in the tumor vicinity is not very efficient, leading to reduced clearance of particles.5 Therefore, passive targeting of drug delivery systems smaller than 200 nm can result in up to a ten-fold increased accumulation as compared to the free small molecule drugs.4

The ability to synthesize and characterize structures on the nanometer scale has opened up numerous opportunities to create new materials with useful properties. These

nm

10-1 1 10 102 103 104 105

Doxorubicin Protein DNA Virus Cell

Polymer LiposomeGold nanoparticle

Figure 1.1 Size comparison of biomaterials in the nanometer size range. The small molecule drug doxoru-bicin and a cell were included for reference. Nanoparticles with dimensions of 1-1000 nm include polymers, inorganic nanoparticles (such as gold nanoparticles), viruses and liposomes. Graphics created with Chem-Draw, MolView and PDB software. PDB IDs: 5H9W and 3J31.

3

nanoscale materials have been used in a variety of applications, including as research tools, drug delivery, and diagnostic imaging.5-7 The composition of these materials can vary greatly, only to be limited by the creativity of the researchers and the stability of the resulting construct (Figure 1.1). Nanoparticle drug delivery systems can be composed of lipids (liposomes, micelles, emulsions) or polymers (polymer-drug conjugates, polymer microspheres).8, 9 Doxil is one of the most well-known examples of nanoparticles used in the clinic. It comprises a doxorubicin small molecule therapeutic encapsulated in a PEGylated liposome.9 Another prevalent example is N-2-hydroxypropyl methacrylamide copolymer-linked doxorubicin, which can be administered in higher doses than free doxo-rubicin and thus is a more efficient treatment.10 Other nanoscale therapeutic delivery sys-tems being explored include: inorganic nanoparticles (silica, iron, gold), polysaccharide colloids, polyamidoamine (PAMAM) dendrimer clusters, hydrogel dextran nanoparticles and latex beads.7, 8 For example, PAMAM dendrimers conjugated with methotrexate (an anti-cancer drug) and folic acid (for cell targeting) demonstrated in vitro cytotoxicity to cell lines known to be methotrexate-resistant.11 This result suggests that nanoparticle delivery of currently used anti-cancer agents can result in beneficial properties, including the eva-sion of therapeutic resistance.

When comparing the advantages and disadvantages of drug delivery systems, sever-al factors have to be taken into account: the ability to scale their synthesis, their capacity to carry cargo, their distribution in an in vivo system and the ease of introducing cell-spe-cific targeting capability. Liposomes can carry many therapeutic molecules (on the order of thousands), but the nature of the nanoparticle core limits the types of compounds that can be entrapped.4 Polymeric micelles can deliver water insoluble therapeutics due to their hydrophobic cores.12 Polymeric delivery systems can be modified with a variety of molecules and can be used to control the release of the drug cargo by selecting an ap-propriate linker.4, 13 Dendrimers have been used to encapsulate anticancer agents, but the release rates are usually very slow.14 Additionally, the cytotoxicity of PAMAM dendrimers has been observed to increase with increasing size of the scaffold, although PEGylation can reduce the toxicity.13

In addition to synthetic scaffolds, self-assembled multimeric structures, such as heat shock proteins and viral capsids, have also been investigated for the development of next generation imaging and drug delivery agents.15 Each of the protein cage scaffolds has unique attributes - size, structure, solvent accessibility, chemical and temperature stability, structural plasticity, and assembly and disassembly parameters - all of which can be leveraged for particular applications.16 Based on the information available about their structure and composition, scientists have employed genetic and chemical modifica-tions to alter their properties and endow them with additional functionality for therapeutic purposes or for creating novel biomaterials. The interior cavities and multiple attachment sites of these protein cage scaffolds allow them to house a large amount of imaging or therapeutic agents, leading to the enhancement of the signal intensity and ability to deliver multiple copies of drug molecules.16 Housing therapeutics within protein cage ar-chitectures may restrict their bioavailability until their release is triggered by a particular cellular- or tissue-specific signal. In order to achieve selective detection or delivery, these vehicles have been modified with targeting agents.17 Various chemical bioconjugation

4

techniques have played crucial roles in the development of these targeted protein cage nanoparticles using different types of targeting groups, including small molecules,18, 19 nucleic acid aptamers,20 peptides,21 glycans,22 or antibodies.23-26

1.2 Viral capsids as protein-based nanoparticles

Viruses come in a wide range of sizes, shapes and surface morphologies (Figure 1.2).15, 16 The structures of viral capsids are composed of hundreds of copies of one or a few proteins. Despite their complexity, many capsids form homogenous populations of precise geometrical assemblies. This is important in medicine, where a large emphasis is placed on homogeneity for evaluating drug performance and obtaining federal approval, and nanotechnology, where the placement of the components can have a great impact on the properties of the final material. Although they are very diverse, all viruses consist of an encapsulated nucleic acid genome within the protein coat assembly, or capsid. Vi-ruses play a great role in human diseases and the ecosystem. Thus, over the last century, scientists have studied their life cycle, composition, and assembly in detail. The shapes and sizes of viral capsids are very well-defined (such as icosahedral or rod-shaped), with protein-protein and protein-nucleic acid interactions directing all the components into their precise location, like a puzzle being perfectly completed each time.27 Some capsids are resistant to the removal of their genome and form virus-like particles (VLPs), that can no longer infect the host organism. Applications of viral capsids in biotechnology include vaccines,28 gene delivery,29 and epitope display methods.30

1.2.1 Viral capsid p roduction

Production of viral capsids can be accomplished by infecting the native host organ-ism or by recombinant expression of the coat protein. Choosing the appropriate method depends on the desired application and case-specific safety precautions or challenges in

Figure 1.2 Structural diversity of viruses from different origins. Graphics created with the Protein Data Base (PDB) viewer. Capsids are color coded by protein chain. The viruses depicted (and their PDB IDs) are as follows: MS2 bacteriophage T=1 (4ZOR), MS2 bacteriophage T=3 (2MS2), HK97 bacteriophage (1OHG), Cowpea Mosaic Virus (CPMV, 5FMO), Brome Mosaic Virus (BMV, 3J7N), Partitivirus (3ES5), Hepatitis B Virus (HBV, 1QGT), Human Papillomavirus (HPV, 3J6R), Human Immunodeficiency Virus capsid without its viral envelope (HIV, 3J3Q).

HK97 CPMVMS2T=1

BMV HBV HPV

bacteriophages plant viruses animal viruses

HIV virionPartitivirusMS2T=3

5

the production protocol. Some viral capsids, such as the polyomavirus, the hepatitis B vi-rus (HBV), or the cowpea chlorotic mottle virus (CCMV), can be obtained by using E. coli recombinant expression, even though their native host is eukaryotic.31-33 Recombinant expression is also compatible with the insertion of artificial amino acids, leading to the production of capsids with novel properties.34 Infection of the host organisms can be used to obtain large amounts of viral capsid, as seen for the tobacco mosaic virus (TMV)35 and cowpea mosaic virus (CPMV).36 For some applications, these native capsids need further modification (such as genome removal) before use.

1.2.2 Properties of viral capsids

Viruses can vary greatly in their size, structure, sensitivity to pH conditions, and the propensity to assemble without their native genome. The size of viruses is usually on the scale of nanometers, with some of the smallest known capsids, the porcine circovirus 2 (PCV2), measuring only 17 nm across,37 while the Megaviridae is nearly 1 μm in diam-eter.38 Capsid size is an important parameter for applications such as drug delivery and templated synthesis of nanoparticles. Some capsids, such as HBV, CCMV, polyomavi-ridae, TMV, and brome mosaic virus (BMV) can be disassembled and reassembled by varying the pH and ionic strength of their environment.39-41 Control over the porosity of some viral capsids can be attained at high temperatures42 or in the presence of metal che-lators.43 The large variety of viral capsids and the knowledge about their properties can be used to produce new materials with features difficult to access using purely synthetic materials.

Taking advantage of the propensity to self-assemble around negative charges, sci-entists have used negatively charged polymers of different sizes to produce BMV capsid assemblies of 18 nm, 25 nm, or 28 nm diameter, depending on the assembly conditions.44,

45 Rod-shaped viruses have also been studied over several decades. TMV is one of the most well-known systems and the intermediates in its assembly process have been used for different nanotechnology applications.35, 40 TMV mutants can form rods of various lengths,46 disks47 and even spheres.48 Another example of engineering the properties of a viral capsid comes from filamentous bacteriophage. The length of this type of virion is dic-tated by the size of its genome. Recent work has reduced the size of the genome, which resulted in reducing the length of the phage from ~900 nm to only 50 nm.49 These new particles are expected to have vastly different interactions with cells and biodistribution properties as compared to the wild type filamentous bacteriophage. As we understand more about the self-assembly of viral coat proteins, we will be able to use protein engi-neering and mixed protein assemblies to form nover and complex protein nanostructures.

1.3 Chemical modifications of protein materials

The properties of the exterior surface can be altered to target capsids to specific tis-sues, change the overall in vivo biodistribution, or scaffold other molecules for precise spatial localization. Many approaches use lysine chemistry to append new functionality to the surfaces of capsids. Although this is often the most facile method, the lack of speci-ficity can negate the advantages of using a homogeneous capsid structure. Researchers have incorporated unnatural amino acids for bioorthogonal modification by using Amber

6

stop codon suppression or replacement of methionine residues.50, 51 These strategies can lead to site-selective modification of the capsid surface with a wide variety of useful moi-eties, including peptides and small molecules such as folate.

1.3.1 Modification of native amino acids

Taking advantage of the know amino acid sequence and the 3D structure of proteins, scientists can strategically pick the site and degree of modification. Over the last decade, numerous reactions have been developed to be selective for certain reactive handles installed on proteins (Figure 1.3).52 The most commonly used modifications target the nucleophilic side chains of lysine and cysteine, with carboxylic acid and aromatic residue modification strategies occurring less frequently in literature. Lysines can be modified with activated esters, sulfonyl chlorides, and iso(thio)cyanates to form amides, sulfonamides, or (thio)ureas, respectively.53-56 Reacting the amine moiety with aldehydes followed by reductive amination leads to a secondary amine, which minimally perturbs the iso electric point of the protein.57 Although lysines are highly abundant on proteins and usually sol-vent accessible, making them a good target for modification, the position and number of modifications are hard to control. Additionally, the N-terminal amine could also be modi-

Tyr (3.2%)

Trp (1.4%)

amines and carbodiimides

Glu, Asp (11.6%)

diazonium salts Mannich

p-anisidines

π-allylation

cyclic diazo-dicarboxamide

phenylenediamines rhodium carbenoid

NHS esters sulfonyl chlorides iso(thio)cyanates reductive amination

Lys (5.9%)

Aminoacid (abundance) Chemical modification

Cys (1.9%)

maleimidesiodoacetamidesdehydroalanine

thiol-enedisulfides

Figure 1.3 Chemical modifications of native amino acid side chains. Adapted from reference 52.

NH O

NH2

NH O

SH

NH O

COO-

1-2

NH O

OH

NH O

NH

N

O

O

O

OCl S

O

OX C N

H

ONaBH3CNor [Cp*Ir(bipy)(H2O)]SO4NaHCO2

RS S I

O

NH

N

O

O light or AIBN2. HS

H2N NC

N N

N2+

NH2

H

O

R

AcOPd(OAc)2

N

O

N

NN

O O

Et2N

HN

ROOC

N2

= cargo

R = (CH2CH2O)3CH3Rh(OAc)4

(NH4)2Ce(NO3)6

X=O or S

SO

OO

NH2

1.

7

fied under most conditions used for lysine modification.

The low abundance of cysteine residues (1.9%) make cysteine modification a popular strategy.58 The side chain can be reversibly modified using a disulfide reagent that can be later removed using reductants (such as glutathione, dithiothreitol, β-mercaptoethanol, or tris(2-carboxyethyl) phosphine).56, 58, 59 Cysteines are usually modified with alkylhalides (e.g. iodoacet amides) or with α,β-unsaturated carbonyl compounds (such as maleimid-es).59 Modification of native cysteines can be also be accomplished using thiol-ene chem-istry. At high pH (10-11), cysteine can be transformed into dehydroalanine using O-mes-itylenesulfonylhydroxylamine that can then be modified with thiol reagents, resulting in a thioether linkage.60

The carboxylic acid side chains of aspartate and glutamate are more abundant on protein surfaces and thus offer more opportunities for modification.58 The acid moiety can be coupled to amines using activating agents such as 1-eth yl-3-(3-dimethylaminopropyl)carbodiimide (EDC).56 One major drawback of this strategy is the possibility of undesired cross-linking of the native amine moieties with the now activated acid groups.

The aromatic amino acid tyrosine can be modified by taking advantage of electrophilic aromatic substitution with iodine and nitrous acid.58 Diazonium salts can modify the phe-nol moiety to form an azo linkage.61 Our group has developed a three-component Man-nich reaction between tyrosine, aldehydes, and anilines.62 Tyrosine can also be modified in the presence of palladium63 or ceric ammonium nitrate (CAN) and anisidines containing electron-rich aromatic rings.64 Cyclic diazodicarboxamides can also be used for the selec-tive modification of tyrosine.65, 66

Ce(IV) can be used to modify either tyrosine and tryptophan resi dues with N,N-di-alkyl-p-phenylenediamines.64 Tryptophan residues can be modified with rhodium car-benoids generated in situ from diazo compounds and [Rh2(OAc)4] in the presence of N-(tert-butyl)hydroxylamine.65, 66

The methods presented above allow for facile modification of native amino acid resi-dues on proteins. In order to achieve site-specific modification and homogeneous product populations, it is often necessary to introduce a uniquely reactive residue, as presented in the following section.

1.3.2 Unnatural amino acids for site-selective protein modification

The options for chemical modifications of proteins were greatly increased by the ability to introduce unnatural amino acids into proteins. These unnatural residues can provide an impressive variety of chemical handles to be used for site-specific modifications.69, 70 Given the importance of precise control over protein modification, a large effort was dedi-cated to the optimization of in vivo incorporation of artificial amino acids in E. coli, as well as in mammalian cells.51, 70, 71

The methods for site-selective incorporation of non-canonical amino acids usually leverage the amber stop codon (UAG). Schultz and co-workers have evolved orthog-onal aaRSs for the unnatural amino acids and tRNAs that can recognize the amber

8

stop codon. By co-expressing the aaRS, tRNA, and the gene of interest containing the amber stop codon, the artificial amino acid can be inserted into the protein of in-terest.50,72

To achieve site-specific protein mod-ification, the reactive handle targeted must be present in a single copy and the chemistry must be unique to that function-al group, such that it does not modify any native residues. Several approaches have been attempted towards this goal,69 and some of the most commonly used bio-orthogonal reactions are listed in Figure 1.4.

One of the earliest bioorthogonal re-actions for protein modification used car-bonyl groups incorporated as unnatural amino acids to generate uniquely elec-trophilic sites. Subsequent reactions with alkoxyamines and hydrazides resulted in the formation of stable products.73, 74 Bertozzi and co-workers later developed an aminooxy reagent that undergoes oxime formation followed by a Pictet-Spengler type reaction to form a product that has superior stability.75

In recent times, one of the most commonly used strategies for biomolecule modifi-cation is the incorporation of azide moieties, which are relatively small in size and do not perturb the protein structure. The Cu(I) catalyzed reaction between an azide and an alkyne (usually referred to as the “click reaction”76, 77) has been widely adopted and used in a large array of applications.78 Nevertheless, given the toxic nature of Cu(I), the copper catalyzed cycloaddition has limitations when working with live cells. Bertozzi and co-workers developed a variant of the azide-alkyne cycloaddition that proceeds without the presence of Cu(I), by using strain to promote a cycloaddition between the azide moi-ety and a cyclooctyne.26, 27 To note, some of these strained alkynes are not selective for the azide moiety and can also react with thiols.79

Recently, tetrazines have gained a lot of popularity for their use in rapid bioconjuga-tion reactions. Tetrazines can react with strained alkenes (trans-cyclooctene, norbornene, 1,3-disubstituted cyclopropenes), as well as unstrained alkenes.80-82 After the discovery of strain-promoted bioorthogonal reactions, numerous groups have used this strategy to synthesize well-defined protein-based structures.69

Our group has developed bioorthogonal reactions for the modification of aniline groups with electron-rich aromatic rings (Figure 1.4) in the presence of an oxidizing agent (sodium periodate or potassium ferricyanide).21, 34, 83-85 In addition to being carried out in mild aqueous conditions, a particular advantage of these reactions is that they can be

Figure 1.4 Chemical modifications of proteins with unnatural amino acid residues. Adapted from refer-ence 52.

Oxime formation

“Click chemistry”

Tetrazine ligation

Oxidative coupling

O

R

NO

R

ONH2

N3 NN N

Cu(I)

NN

NN

NN

NH2

OH

NH2 OOH

N

= cargo

NaIO4or K3[Fe(CN)]6

9

performed in low micromolar concentrations of both protein and aminophenol coupling partners in short reaction times . The oxidative coupling has been used to attach peptides, DNA, and PEG to aniline containing proteins and current efforts are dedicated to the de-velopment of improved versions of this reaction.86

1.4. Conclusions

Many nanocarrier systems have been developed, each having its own inherent ad-vantages and disadvantages. Among these, protein-based scaffolds have gained much interest due to their biodegradability and ability to carry a large amount of cargo. Recent improvements in genetic modifications, synthetic schemes and bioconjugation reactions have laid the foundation for elaborate designs with enhanced properties and the area of targeted therapeutic and imaging agent delivery continues to be a rapidly growing field. Hopefully the next few decades will bring about the development of nanoparticle systems able to provide earlier disease diagnostics and more efficient treatments for patients in need.

The work presented in the next few chapters focused on the use of protein-based nanoparticles as imaging agents and drug delivery vehicles. Several modification strat-egies are employed, taking advantage of the composition of the nanoparticles and their physicochemical properties. The efficiency of the nanoparticle conjugates to deliver cargo to mammalian cells in culture and mouse models of breast and brain cancer is evaluated.

1.5 References

1. Boussios, S., Pentheroudakis, G., Katsanos, K. and Pavlidis, N., Systemic treat-ment-induced gastrointestinal toxicity: incidence, clinical presentation and management. Ann Gastroenterol. 2012, 25, 106-118.

2. Devalapally, H., Chakilam, A. and Amiji, M.M., Role of nanotechnology in pharmaceu-tical product development. J Pharm Sci. 2007, 96, 2547-2565.

3. Maeda, H., Nakamura, H. and Fang, J., The EPR effect for macromolecular drug de-livery to solid tumors: Improvement of tumor uptake, lowering of systemic toxicity, and distinct tumor imaging in vivo. Adv Drug Deliv Rev. 2013, 65, 71-79.

4. Allen, T.M. and Cullis, P.R., Drug delivery systems: entering the mainstream. Science. 2004, 303, 1818-1822.

5. Brannon-Peppas, L. and Blanchette, J.O., Nanoparticle and targeted systems for can-cer therapy. Adv Drug Deliv Rev. 2004, 56, 1649-1659.

6. Holzinger, M., Le Goff, A. and Cosnier, S., Nanomaterials for biosensing applications: a review. Front Chem. 2014, 2, 63.

7. Steichen, S.D., Caldorera-Moore, M. and Peppas, N.A., A review of current nanoparti-cle and targeting moieties for the delivery of cancer therapeutics. Eur J Pharm Sci. 2013, 48, 416-427.

10

8. De Jong, W.H. and Borm, P.J., Drug delivery and nanoparticles:applications and haz-ards. Int J Nanomedicine. 2008, 3, 133-149.

9. Alberts, D.S. and Garcia, D.J., Safety aspects of pegylated liposomal doxorubicin in patients with cancer. Drugs. 1997, 54 Suppl 4, 30-35.

10. Vasey, P.A., Kaye, S.B., Morrison, R., Twelves, C., Wilson, P., Duncan, R., et al., Phase I clinical and pharmacokinetic study of PK1 [N-(2-hydroxypropyl)methacrylamide copolymer doxorubicin]: first member of a new class of chemotherapeutic agents-drug-polymer conjugates. Cancer Research Campaign Phase I/II Committee. Clin Cancer Res. 1999, 5, 83-94.

11. Gurdag, S., Khandare, J., Stapels, S., Matherly, L.H. and Kannan, R.M., Activity of dendrimer-methotrexate conjugates on methotrexate-sensitive and -resistant cell lines. Bioconjug Chem. 2006, 17, 275-283.

12. Kwon, G.S., Polymeric micelles for delivery of poorly water-soluble compounds. Crit Rev Ther Drug Carrier Syst. 2003, 20, 357-403.

13. Kobayashi, H., Kawamoto, S., Saga, T., Sato, N., Hiraga, A., Ishimori, T., et al., Positive effects of polyethylene glycol conjugation to generation-4 polyamidoamine dendrimers as macromolecular MR contrast agents. Magn Reson Med. 2001, 46, 781-788.

14. Nanjwade, B.K., Bechra, H.M., Derkar, G.K., Manvi, F.V. and Nanjwade, V.K., Den-drimers: emerging polymers for drug-delivery systems. Eur J Pharm Sci. 2009, 38, 185-196.

15. Koudelka, K.J., Pitek, A.S., Manchester, M. and Steinmetz, N.F., Virus-Based Nanopar-ticles as Versatile Nanomachines. Annu Rev Virol. 2015, 2, 379-401.

16. Glasgow, J. and Tullman-Ercek, D., Production and applications of engineered viral capsids. Appl Microbiol Biotechnol. 2014, 98, 5847-5858.

17. Yildiz, I., Shukla, S. and Steinmetz, N.F., Applications of viral nanoparticles in medi-cine. Curr Opin Biotechnol. 2011, 22, 901-908.

18. Destito, G., Yeh, R., Rae, C.S., Finn, M.G. and Manchester, M., Folic acid-mediated targeting of cowpea mosaic virus particles to tumor cells. Chem Biol. 2007, 14, 1152-1162.

19. Ren, Y., Wong, S.M. and Lim, L.Y., Folic acid-conjugated protein cages of a plant vi-rus: a novel delivery platform for doxorubicin. Bioconjug Chem. 2007, 18, 836-843.

20. Tong, G.J., Hsiao, S.C., Carrico, Z.M. and Francis, M.B., Viral capsid DNA aptamer conjugates as multivalent cell-targeting vehicles. J Am Chem Soc. 2009, 131, 11174-11178.

21. Behrens, C.R., Hooker, J.M., Obermeyer, A.C., Romanini, D.W., Katz, E.M. and Fran-cis, M.B., Rapid chemoselective bioconjugation through oxidative coupling of anilines and aminophenols. J Am Chem Soc. 2011, 133, 16398-16401.

11

22. Rhee, J.K., Baksh, M., Nycholat, C., Paulson, J.C., Kitagishi, H. and Finn, M.G., Glycan-targeted virus-like nanoparticles for photodynamic therapy. Biomacromolecules. 2012, 13, 2333-2338.

23. ElSohly, A.M., Netirojjanakul, C., Aanei, I.L., Jager, A., Bendall, S.C., Farkas, M.E., et al., Synthetically Modified Viral Capsids as Versatile Carriers for Use in Antibody-Based Cell Targeting. Bioconjug Chem. 2015, 26, 1590-1596.

24. Aanei, I.L., ElSohly, A.M., Farkas, M.E., Netirojjanakul, C., Regan, M., Taylor Murphy, S., et al., Biodistribution of Antibody-MS2 Viral Capsid Conjugates in Breast Cancer Mod-els. Mol Pharm. 2016, 13, 3764-3772.

25. Brown, W.L., Mastico, R.A., Wu, M., Heal, K.G., Adams, C.J., Murray, J.B., et al., RNA bacteriophage capsid-mediated drug delivery and epitope presentation. Intervirolo-gy. 2002, 45, 371-380.

26. Flenniken, M.L., Willits, D.A., Harmsen, A.L., Liepold, L.O., Harmsen, A.G., Young, M.J., et al., Melanoma and lymphocyte cell-specific targeting incorporated into a heat shock protein cage architecture. Chem Biol. 2006, 13, 161-170.

27. Mateu, M.G., Assembly, stability and dynamics of virus capsids. Arch Biochem Bio-phys. 2013, 531, 65-79.

28. Roldao, A., Mellado, M.C., Castilho, L.R., Carrondo, M.J. and Alves, P.M., Virus-like particles in vaccine development. Expert Rev Vaccines. 2010, 9, 1149-1176.

29. Kay, M.A., State-of-the-art gene-based therapies: the road ahead. Nat Rev Genet. 2011, 12, 316-328.

30. Bratkovic, T., Progress in phage display: evolution of the technique and its application. Cell Mol Life Sci. 2010, 67, 749-767.

31. Zhao, X., Fox, J.M., Olson, N.H., Baker, T.S. and Young, M.J., In vitro assembly of cowpea chlorotic mottle virus from coat protein expressed in Escherichia coli and in vi-tro-transcribed viral cDNA. Virology. 1995, 207, 486-494.

32. Teunissen, E.A., de Raad, M. and Mastrobattista, E., Production and biomedical ap-plications of virus-like particles derived from polyomaviruses. J Control Release. 2013, 172, 305-321.

33. Stahl, S., MacKay, P., Magazin, M., Bruce, S.A. and Murray, K., Hepatitis B virus core antigen: synthesis in Escherichia coli and application in diagnosis. Proc Natl Acad Sci U S A. 1982, 79, 1606-1610.

34. Carrico, Z.M., Romanini, D.W., Mehl, R.A. and Francis, M.B., Oxidative coupling of peptides to a virus capsid containing unnatural amino acids. Chem Commun. 2008, 1205-1207.

35. Alonso, J.M., Gorzny, M.L. and Bittner, A.M., The physics of tobacco mosaic virus and virus-based devices in biotechnology. Trends Biotechnol. 2013, 31, 530-538.

12

36. Sainsbury, F., Canizares, M.C. and Lomonossoff, G.P., Cowpea mosaic virus: the plant virus-based biotechnology workhorse. Annu Rev Phytopathol. 2010, 48, 437-455.

37. Khayat, R., Brunn, N., Speir, J.A., Hardham, J.M., Ankenbauer, R.G., Schneemann, A., et al., The 2.3-angstrom structure of porcine circovirus 2. J Virol. 2011, 85, 7856-7862.

38. Philippe, N., Legendre, M., Doutre, G., Coute, Y., Poirot, O., Lescot, M., et al., Pan-doraviruses: amoeba viruses with genomes up to 2.5 Mb reaching that of parasitic eu-karyotes. Science. 2013, 341, 281-286.

39. Pfeiffer, P. and Hirth, L., Aggregation states of brome mosaic virus protein. Virology. 1974, 61, 160-167.

40. Durham, A.C., Finch, J.T. and Klug, A., States of aggregation of tobacco mosaic virus protein. Nat New Biol. 1971, 229, 37-42.

41. Adolph, K.W. and Butler, P.J., Studies on the assembly of a spherical plant virus. I. States of aggregation of the isolated protein. J Mol Biol. 1974, 88, 327-341.

42. Jiang, W., Li, Z., Zhang, Z., Baker, M.L., Prevelige, P.E., Jr. and Chiu, W., Coat protein fold and maturation transition of bacteriophage P22 seen at subnanometer resolutions. Nat Struct Biol. 2003, 10, 131-135.

43. Sherman, M.B., Guenther, R.H., Tama, F., Sit, T.L., Brooks, C.L., Mikhailov, A.M., et al., Removal of divalent cations induces structural transitions in red clover necrotic mosaic virus, revealing a potential mechanism for RNA release. J Virol. 2006, 80, 10395-10406.

44. Lucas, R.W., Kuznetsov, Y.G., Larson, S.B. and McPherson, A., Crystallization of Brome mosaic virus and T = 1 Brome mosaic virus particles following a structural transi-tion. Virology. 2001, 286, 290-303.

45. Krol, M.A., Olson, N.H., Tate, J., Johnson, J.E., Baker, T.S. and Ahlquist, P., RNA-con-trolled polymorphism in the in vivo assembly of 180-subunit and 120-subunit virions from a single capsid protein. Proc Natl Acad Sci U S A. 1999, 96, 13650-13655.

46. Bruckman, M.A., Soto, C.M., McDowell, H., Liu, J.L., Ratna, B.R., Korpany, K.V., et al., Role of hexahistidine in directed nanoassemblies of tobacco mosaic virus coat pro-tein. ACS Nano. 2011, 5, 1606-1616.

47. Dedeo, M.T., Duderstadt, K.E., Berger, J.M. and Francis, M.B., Nanoscale protein assemblies from a circular permutant of the tobacco mosaic virus. Nano Lett. 2010, 10, 181-186.

48. Atabekov, J., Nikitin, N., Arkhipenko, M., Chirkov, S. and Karpova, O., Thermal transi-tion of native tobacco mosaic virus and RNA-free viral proteins into spherical nanoparti-cles. J Gen Virol. 2011, 92, 453-456.

49. Rakonjac, J., Bennett, N.J., Spagnuolo, J., Gagic, D. and Russel, M., Filamentous bacteriophage: biology, phage display and nanotechnology applications. Curr Issues Mol Biol. 2011, 13, 51-76.

13

50. Wang, L., Brock, A., Herberich, B. and Schultz, P.G., Expanding the genetic code of Escherichia coli. Science. 2001, 292, 498-500.

51. Link, A.J., Mock, M.L. and Tirrell, D.A., Non-canonical amino acids in protein engi-neering. Curr Opin Biotechnol. 2003, 14, 603-609.

52. Obermeyer, A.C., Development and application of oxidative coupling bioconjugation reactions with ortho-aminophenols. Dissertation, U.C. Berkeley, 2013.

53. McFarland, J.M. and Francis, M.B., Reductive alkylation of proteins using iridium cat-alyzed transfer hydrogenation. J Am Chem Soc. 2005, 127, 13490-13491.

54. Kalkhof, S. and Sinz, A., Chances and pitfalls of chemical cross-linking with amine-re-active N-hydroxysuccinimide esters. Anal Bioanal Chem. 2008, 392, 305-312.

55. Sletten, E.M. and Bertozzi, C.R., Bioorthogonal chemistry: fishing for selectivity in a sea of functionality. Angew Chem Int Ed Engl. 2009, 48, 6974-6998.

56. Basle, E., Joubert, N. and Pucheault, M., Protein chemical modification on endoge-nous amino acids. Chem Biol. 2010, 17, 213-227.

57. Gildersleeve, J.C., Oyelaran, O., Simpson, J.T. and Allred, B., Improved procedure for direct coupling of carbohydrates to proteins via reductive amination. Bioconjug Chem. 2008, 19, 1485-1490.

58. Tilley, S.D., Joshi, N.S., and Francis, M.B. Proteins: Chemistry and Chemical Reactiv-ity. Wiley Encyclopedia of Chemical Biology. 2008.

59. Chalker, J.M., Bernardes, G.J., Lin, Y.A. and Davis, B.G., Chemical modification of proteins at cysteine: opportunities in chemistry and biology. Chem Asian J. 2009, 4, 630-640.

60. Bernardes, G.J., Chalker, J.M., Errey, J.C. and Davis, B.G., Facile conversion of cys-teine and alkyl cysteines to dehydroalanine on protein surfaces: versatile and switchable access to functionalized proteins. J Am Chem Soc. 2008, 130, 5052-5053.

61. Jones, M.W., Mantovani, G., Blindauer, C.A., Ryan, S.M., Wang, X., Brayden, D.J., et al., Direct peptide bioconjugation/PEGylation at tyrosine with linear and branched poly-meric diazonium salts. J Am Chem Soc. 2012, 134, 7406-7413.

62. Joshi, N.S., Whitaker, L.R. and Francis, M.B., A three-component Mannich-type re-action for selective tyrosine bioconjugation. J Am Chem Soc. 2004, 126, 15942-15943.

63. Tilley, S.D. and Francis, M.B., Tyrosine-selective protein alkylation using pi-allylpalla-dium complexes. J Am Chem Soc. 2006, 128, 1080-1081.

64. Seim, K.L., Obermeyer, A.C. and Francis, M.B., Oxidative Modification of Native Pro-tein Residues Using Cerium(IV) Ammonium Nitrate. Journal of the American Chemical Society. 2011, 133, 16970-16976.

65. Ban, H., Gavrilyuk, J. and Barbas, C.F., 3rd, Tyrosine bioconjugation through aqueous

14

ene-type reactions: a click-like reaction for tyrosine. J Am Chem Soc. 2010, 132, 1523-1525.

66. Ban, H., Nagano, M., Gavrilyuk, J., Hakamata, W., Inokuma, T. and Barbas, C.F., 3rd, Facile and stabile linkages through tyrosine: bioconjugation strategies with the tyro-sine-click reaction. Bioconjug Chem. 2013, 24, 520-532.

67. Antos, J.M. and Francis, M.B., Selective tryptophan modification with rhodium carben-oids in aqueous solution. J Am Chem Soc. 2004, 126, 10256-10257.

68. Antos, J.M., McFarland, J.M., Iavarone, A.T. and Francis, M.B., Chemoselective tryp-tophan labeling with rhodium carbenoids at mild pH. J Am Chem Soc. 2009, 131, 6301-6308.

69. Lang, K. and Chin, J.W., Cellular incorporation of unnatural amino acids and bio-orthogonal labeling of proteins. Chem Rev. 2014, 114, 4764-4806.

70. Chin, J.W., Expanding and reprogramming the genetic code of cells and animals. Annu Rev Biochem. 2014, 83, 379-408.

71. Link, A.J. and Tirrell, D.A., Cell surface labeling of Escherichia coli via copper(I)-cata-lyzed [3+2] cycloaddition. J Am Chem Soc. 2003, 125, 11164-11165.

72. Mehl, R.A., Anderson, J.C., Santoro, S.W., Wang, L., Martin, A.B., King, D.S., et al., Generation of a bacterium with a 21 amino acid genetic code. J Am Chem Soc. 2003, 125, 935-939.

73. Dirksen, A. and Dawson, P.E., Rapid oxime and hydrazone ligations with aromatic aldehydes for biomolecular labeling. Bioconjug Chem. 2008, 19, 2543-2548.

74. Dirksen, A., Hackeng, T.M. and Dawson, P.E., Nucleophilic catalysis of oxime ligation. Angew Chem Int Ed Engl. 2006, 45, 7581-7584.

75. Agarwal, P., van der Weijden, J., Sletten, E.M., Rabuka, D. and Bertozzi, C.R., A Pic-tet-Spengler ligation for protein chemical modification. Proc Natl Acad Sci U S A. 2013, 110, 46-51.

76. Saxon, E. and Bertozzi, C.R., Cell surface engineering by a modified Staudinger re-action. Science. 2000, 287, 2007-2010.

77. Saxon, E., Armstrong, J.I. and Bertozzi, C.R., A “traceless” Staudinger ligation for the chemoselective synthesis of amide bonds. Org Lett. 2000, 2, 2141-2143.

78. Kolb, H.C., Finn, M.G. and Sharpless, K.B., Click Chemistry: Diverse Chemical Func-tion from a Few Good Reactions. Angew Chem Int Ed Engl. 2001, 40, 2004-2021.

79. van Geel, R., Pruijn, G.J., van Delft, F.L. and Boelens, W.C., Preventing thiol-yne ad-dition improves the specificity of strain-promoted azide-alkyne cycloaddition. Bioconjug Chem. 2012, 23, 392-398.

80. Yang, J., Seckute, J., Cole, C.M. and Devaraj, N.K., Live-cell imaging of cyclopropene

15

tags with fluorogenic tetrazine cycloadditions. Angew Chem Int Ed Engl. 2012, 51, 7476-7479.

81. Devaraj, N.K., Weissleder, R. and Hilderbrand, S.A., Tetrazine-based cycloadditions: application to pretargeted live cell imaging. Bioconjug Chem. 2008, 19, 2297-2299.

82. Blackman, M.L., Royzen, M. and Fox, J.M., Tetrazine ligation: fast bioconjugation based on inverse-electron-demand Diels-Alder reactivity. J Am Chem Soc. 2008, 130, 13518-13519.

83. Obermeyer, A.C., Jarman, J.B., Netirojjanakul, C., El Muslemany, K. and Francis, M.B., Mild bioconjugation through the oxidative coupling of ortho-aminophenols and ani-lines with ferricyanide. Angew Chem Int Ed Engl. 2014, 53, 1057-1061.

84. Obermeyer, A.C., Jarman, J.B. and Francis, M.B., N-terminal modification of proteins with o-aminophenols. J Am Chem Soc. 2014, 136, 9572-9579.

85. Hooker, J.M., Esser-Kahn, A.P. and Francis, M.B., Modification of aniline containing proteins using an oxidative coupling strategy. J Am Chem Soc. 2006, 128, 15558-15559.

86. ElSohly, A.M. and Francis, M.B., Development of oxidative coupling strategies for site-selective protein modification. Acc Chem Res. 2015, 48, 1971-1978.

16

Chapter 2: MS2 bacteriophage as an imaging agent

The work presented herein was performed in collaboration with Prof. Michelle Farkas and Dr. James O’Neil. Parts of this work were published in “PET Imaging and Biodis-tribution of Chemically Modified Bacteriophage MS2” (DOI: 10.1021/mp3003754) and “Biodistribution of Antibody-MS2 Viral Capsid Conjugates in Breast Cancer Models” (DOI: 10.1021/acs.molpharmaceut.6b00566) and were reproduced with permission.

Abstract

Synthetic platforms, such as polymers, dendrimers, liposomes, and virus-like particles (VLPs), have been investigated for their potential to serve as in vivo delivery vehicles for imaging tracers. In particular, protein cages are promising candidates due to their biocompatibility and capacity to carry large numbers of tracers. Herein, we describe a nanoscale imaging platform based on the bacteriophage MS2, the VLPs of which self-as-semble from 180 copies of a single coat protein into monodisperse, 27 nm icosahedral capsids. MS2 capsids are biodegradable, stable under a variety of temperature, pH, and solvent conditions, and easily synthesized and purified in large quantities. The capsids were successfully modified with macrocyclic chelators for the 64Cu radioisotope to obtain nanoparticles with high specific activity. The MS2 VLPs showed excellent serum stability and accumulated in tumors in mouse models of breast cancer, proving their potential as imaging agents.

17

2.1 Current imaging techniques

Imaging plays a crucial role in detecting the location of cancerous formations, as well as in monitoring the response of a tumor to a specific treatment course.1 Several molec-ular imaging approaches have been used for tracking the distribution of agents with ther-apeutic potential, including positron emission tomography (PET), single photon emission computed tomography (SPECT), magnetic resonance imaging (MRI), computed tomog-raphy (CT), ultrasound, bioluminescence, and fluorescence imaging (Figure 2.1).1, 2 Each imaging modality has strengths and weaknesses, and the appropriate technique should be selected based on the application of interest. For example, CT and MRI provide a high degree of spatial resolution and are well suited for gaining information about the location of the tumor. Radionuclide-based imaging methods, such as SPECT and PET, have been a particular focus in biomedical research due to advantages such as high sensitivity (pi-comolar level) and excellent tissue penetration by the signal. These techniques can be used to monitor tumor burden, progression and metastasis, but they lack anatomical in-formation to localize the signal, which is better fulfilled by MRI and CT. The application of in vivo optical imaging is still modest, due to the limited penetration of light into tissues.3 However, due to recent advances in intravital fluorescence microscopy, information about subcellular biomarker distribution can be obtained in living animals.4 An increasing num-ber of facilities now have multimodal scanners in which two imaging modes can be ap-plied simultaneously, thus benefiting from the advantages of both techniques.1

Recent advances in nanotechnology are providing new opportunities for biomedical imaging with great promise for the development of novel imaging agents with unique optical, magnetic, and chemical properties.5 This new class of imaging probes could pro-vide better contrast enhancement, increased sensitivity, controlled biodistribution, better spatial and temporal information, and the ability to perform multimodal imaging. These features could translate to earlier detection, real time assessment of disease progression and personalized medicine.

An ideal nanoparticle imaging probe for clinical use should be biodegradable, non-tox-

PET SPECT MRI CT Optical Ultrasound

Imaging time

Resolution

Sensitivity

Depth

Cost

30 min

1 - 8 mm

10-11-10-12 M

no limit

high

30 min

1 - 2 mm

10-9-10-11 M

no limit

high

30 - 60 min

10-100 μm

10-3-10-4 M

no limit

medium

15 - 60 min

50 μm

N/A

no limit

medium

< 5 min

1 - 3 mm

10-15 M

1 - 10 cm

low

30-60 min

50 μm

N/A

mm

medium

Figure 2.1 Imaging technologies and their advantages. The features of the most common imaging method-ologies are compared. Images adapted from reference 1.

18

ic, and produce a strong imaging signal. There are numerous examples of nanoparticles used for MRI, optical or PET imaging applications, that have achieved with various de-grees of success.6 For example, bacteriophage P227 and TMV8 modified with Gd3+ have been used as MRI contrast agents. These designs benefitted from the presence of multi-ple paramagnetic molecules that can increase the signal, and the fact that the large size of the capsid leads to long tumbling time in solution, which is important for the agent’s relaxivity. Recently, Duconge et al. reported a novel bifunctional probe for fluorescence and nuclear imaging based on quantum dots covalently labeled with 18F, a popular pos-itron-emitting isotope.9 This combination of PET and fluorescence imaging was used to monitor the in vivo distribution of the probes, from the whole body to the cellular level. Although these results are very promising, the quantum dots present limitations, namely the negative effects of bioaccumulation and toxicity.10

Herein, we present the development of a biodegradable, protein-based PET imaging agents with high specific activity (amount of radioactivity per mass of carrier) and excel-lent stability. The agents accumulate in the tumor in murine breast cancer models and have a long circulation time in vivo.

2.2 The MS2 bacteriophage

The bacteriophage MS2 is an icosahedral, single-stranded RNA virus that infects the bacterium E. coli and other related bacteria. The MS2 viral capsid has a diameter of 27 nm (Figure 2.2) and is composed of 180 monomers.11 Recombinant expression of the coat protein yields non-infectious virus-like particles (VLPs) with remarkable stabil-ity to pH, temperature, assembly/disas-sembly and chemical modifications.12, 13 The genome-free MS2 capsid has been used for a variety of applications, includ-ing the synthesis of high-relaxivity MRI con trast agents,14 delivery of chemother-apeutic agents,15 photodynamic therapy,16 enzyme nanoreactors,17 and the study of the effect of proteins and encapsulated gold nanoparticles on the fluorescence lifetime of dyes.18 In this work, we set out to further understand the properties of the MS2 scaffold, such as its stability in phys-iological conditions, its biodistribution and efficiency as an imaging agent to detect cancerous formations in vivo.

Previous work in our group determined that the native Lys and Cys residues of the MS2 capsid were not very reactive. A Cys residue was introduced at position 87 (N87C mutation) for modification with thiol-reactive probes, taking advantage of the 2 nm pores that have been shown to allow for small molecules to diffuse into the interior cavity. An unnatural amino acid, p-aminophenylalanine (paF), was incorporated at position 19

N87C

T19paF

Figure 2.2 MS2 bacteriophage. (a) Cut-away struc-ture showing the interior and exterior surfaces avail-able for modification. (b) MS2 protein coat dimer with the unnatural amino acid p-aminophenylalanine (paF) at position 19 (facing the exterior of the surface) and an engineered Cys at position 87 (facing the interior of the capsid).

27 nm

2 nm pores

exterior surface

interior surface

19

(T19paF), on the exterior surface, to serve as a handle for oxidative coupling partners, as described in Chapter 1.19 The two mutations afford an orthogonal modification strategy for the interior and exterior surface, as previously reported (Figure 2.3).20 Using site-specific modifications at these two locations, a large array of agents was synthesized for in vitro and in vivo fluorescent imaging, binding assays, and drug delivery applications.

2.3 Evaluating the stability of MS2 capsids

In preparation for in vivo imaging experiments, we sought out to determine the stability of MS2 capsids in physiologically-relevant conditions that best represent an in vivo sys-tem. Ideally, the capsids would remain intact to protect the cargo from the serum proteins until accumulated into tumor. After uptake by cells in the tumor environment, the capsids should fall apart to release the cargo and get cleared from the system.

Initial studies looked at the stability of fluorescently-labeled capsids when incubated with mouse serum at 37 ºC for 24 h. A high-performance liquid chromatography (HPLC) system with a size exclusion chromatography (SEC) column was used. Serum autoflu-orescence and low signal from the dyes (data not shown) prevented us from drawing conclusions regarding the assembly state of the capsids upon incubation. We therefore proceeded with a more sensitive method of detection, based on radioactivity.

For investigation of capsid stability at time points on the order of hours, a radioisotope with a long half life was needed. The 64Cu isotope of copper was chosen for its known positron emitting properties and half life of 12.7 h.21, 22 Utilizing chelators that form stable complexes is critical in the accurate determination of the stability of the capsids, as well as minimizing non-selective interactions of the 64Cu with other biomolecules. Recent stud-ies show that macrocyclic molecules have a much more stable interaction with 64Cu than

pH 7, RT, 1 hSH N

O

O

NH2

MS2C87

MS2paF19

MS2paF19

+

+HO

HN

O

NH2

NH

O

HN

O

CN

O

N

O

O

S

cargoPEG or

targeting group

pH 7, RT, 2 min

NaIO4

polymerspeptides

DNA aptamersantibodies

fluorescent dyeschelators for radioisotopes

small molecule drugs

(a)

(b)

Figure 2.3 General strategy for modification of the MS2 capsid. (a) Interior modification can be achieved by leveraging an engineered Cys residue at position 87 and reacting with maleimide moieties to install cargo such as fluorescent dyes, chelators for radioisotopes, and small molecule drugs. (b) Exterior modification of the p-aminophenylalanine (paF) unnatural amino acid residues at position 19 can be achieved using oxidative coupling in the presence of sodium periodate.

20

acyclic ligands.21

MS2 capsids were modified at the internal Cys87 residues with macrocycle chelators bearing maleimide moieties. We tested four possible chelators (NOTA, NOTA-GA, DOTA and DOTA-GA, Figure 2.4a) to compare the stability of the complexes they form with the 64Cu radioisotope ions. Up to 180 copies of the chelators (one per monomer) were installed on the interior surface of each capsid (Figure 2.4b). Chelation with 64Cu was performed at room temperature for 2 h (Figure 2.5a). Optimized radiolabeling protocols yielded >95% pure radiolabeled capsids with radiochemical yields between 65 and 85% and total synthesis time between 3 and 4 h. Compared to other approaches to radiolabel-ing nanoparticles, this strategy is more facile and requires fewer steps.23

A HPLC-SEC setup with a radioactivity channel was used to determine the reference retention times, the purity of the materials and monitor the species present at each time point (Figure 2.5b). The radioactivity traces indicated that most of the radioactive 64Cu iso-tope (more than 70%) remained associated with the intact capsids even upon incubation in mouse serum at physiological temperature (Figure 2.5c). The second peak (shoulder at retention time ~6.3 min) can be attributed to the dissociation of 64Cu from the chelator and non-specific interactions with other serum proteins or to the partial disassembly of the MS2 capsids. Similar stability over time was observed for all the chelators investigated, with NOTA showing a larger amount of signal associated with the intact capsid peak at longer time points. Based on these results, we decided to proceed with the NOTA-MS2 conjugates for in vivo experiments.

2.4 Positron Emission Tomography and biodistribution of MS2 capsids

Currently, most PET tracers are small molecules modified with radioisotopes such as 18F, 68Ga or 15O2.

1 These probes usually suffer from fast clearance and low specific activ-ity. Nanoparticles can have a longer residence time in the body and possess sites that

DOTA-maleimide (MW 526.5)

N N

NN

O

NH

O

OH

O

OH

O

HO

N

O

O

DOTA-GA-maleimide (MW 598.6)

N N

NN

OO OH

OOH

OHO

OHO

NHN

O

O

NOTA-maleimide (MW 425.4)

N

N

N

OHO

HNO

O

OHNO

O

NOTA-GA-maleimide (MW 497.5)

N

N

N

OHO

O

OH

HN

OHO

O

NO

O

MS2 paF/Cys13778.8

MS2-DOTA 14305.2

MS2-DOTA-GA 14377.3

MS2-NOTA 14204.2

1400013800 14200 14400 1460013600

MS2-NOTA-GA 14276.2

mass (Da)

(a) (b)

Figure 2.4 Conjugation with chelators. (a) Structures and molecular weights (MW) of the four chelator molecules tested for the stability of their interaction with 64Cu. (b) ESI LC-MS of the modification of the MS2 capsid with the four chelators. Almost complete modification is observed in all cases, suggesting that close to 180 copies of the macrocyles are available for chelating with the radioisotope, thus leading to an increased specific activity (amount of radioactivity per mole of capsid).

21

can be used for derivatization with radionuclides and targeting moieties. In addition, they have the potential to attain high specific activity, which is extremely important in order to achieve high-quality images at low doses of radioactivity. For our system, the average specific activity was 1.1 GBq/mg capsid, as compared to 0.37-0.74 GBq/mg Fe reported for 64Cu-DOTA-iron oxide nanoparticles, and 15 GBq/mg reported for polymer nanoparti-cles.23

By leveraging the paF residue at position 19 and the oxidative coupling reaction de-scribed in Chapter 1, we decorated the capsids with polyethylene glycol (PEG) chains with a molecular weight of 5 kDa. PEG has been shown to alter the circulation profile of several nanoparticles in in vivo settings24 and can provide a reduction in the immune response mounted against the viral capsid.25, 26 We therefore investigated the in vivo be-havior of both PEGylated and non-PEGylated capsids in mouse models with or without a breast cancer tumor.

We used PET as a non-invasive way to gain information about the biodistribution of the MS2 conjugates in real time and at different time points while limiting the number of animals used. Dynamic imaging was performed on one animal per group from the time of injection until 1 h post-injection, and static scans were collected at 4, 24, and 48 h post-in-jection (Figure 2.6). Throughout the experiment, a significant amount of signal was ob-served in the liver, the lower abdominal area and the heart. Based on the signal in the heart and the biodistribution results, we concluded that the unmodified MS2 capsid and its PEGylated derivative remain in circulation for a prolonged time. At 48 h post-injection,

free 64Cu

MS2 monomer

MS2-DOTA-64Cuafter purification

74 6 8 95

64Cu-MS2-DOTA-GAR

adio

activ

ity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0 *

t=01 h

4 h24 h

48 h

64Cu-MS2-NOTA-GA

Rad

ioac

tivity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0

*

t=01 h

4 h24 h

48 h

64Cu-MS2-DOTA

Rad

ioac

tivity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0 *

t=01 h

4 h24 h

48 h

64Cu-MS2-NOTA

Rad

ioac

tivity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0

*

t=01 h

4 h24 h

48 h

Retention time (min)

(c)

(a) (b)

Figure 2.5 Evaluating the stability of MS2-chelator-64Cu constructs. (a) Protocol for radiolabeling of MS2 capsids. (b) The retention time for free (not chelated) 64Cu, MS2 monomers (obtained from disassembly of capsids in acidic conditions) and intact radiolabeled MS2 capsid were determined on an SEC HPLC system. The normalized signal from the radioactivity channel shows that pure 64Cu-labeled intact MS2 capsids can be observed as a peak with a retention time of ~5.7 min. (c) Comparison of 64Cu-chelator inter-action stability in mouse serum at 37 ºC. The normalized signal from the radioactivity channel during SEC HPLC runs of aliquots taken at 0 to 48 h time points indicate that the predominant species is the 64Cu-la-beled intact MS2 capsid. The secondary peak at ~6.3 min (*) could be attributed to partial disassembly or dissociation of 64Cu.

radiolabeled MS2

S

1. DOTA-maleimidepH 7.2, 4 h, RT.

2. 64CuCl2, pH 6.2, 2 h, RT

64Cu

SH=

N N

NN

O

NH

O

OH

O

OH

O

HO

N

O

O

22

the decay-corrected signal was significantly weaker overall, as result of the agents being cleared from the system.

For biodistribution studies, groups of 3 mice were injected with each agent and sacri-ficed at either 1 or 24 h time points. The blood and major organs were harvested and their activity was measured using a gamma counter. The tissue samples were weighed and the radioactivity signal was normalized as percent injected dose per gram of tissue (% ID/g, Figure 2.7). Similar biodistribution profiles were obtained from mice with no tumor or with a breast cancer (MCF7 Clone18 tumor). The tissues with the largest degree of ac-cumulation were the liver, the spleen (the usual clearance pathways for nanoparticles),27 and the large intestine. A large amount of signal (7-17% ID/g) was observed in the blood, suggesting a long circulation time. Increased agent accumulation was observed in the tumor from 1 h (1-2% ID/g) to the 24 h time point (3-8% ID/g). No significant differences were observed between the PEGylated and not PEGylated capsids, indicating that the polymer coating does not drastically change the clearance of the MS2 VLP.

To determine the stability of the MS2 capsids in vivo, we collected blood samples from mice injected with radiolabeled capsids. Upon separating the cellular content, plasma was injected onto the HPLC-SEC system with radioactivity detector. As seen in Figure 2.8, even after circulating in a mouse for 24 to 48 h, most of the radioactive signal is asso-ciated with the intact capsids (retention time 5.8 min). This observation is in large contrast

MS2

MS2

-PEG

0 h 0.5 h 1 h 4 h 24 h 48 h

25

%ID/cc

%ID/cc

022

0

tumor

liver

heart

bladder

Figure 2.6 Positron Emission Tomography - Computed Tomography (PET-CT) imaging with 64Cu-NO-TA-MS2 scaffolds in nude mice with MCF7 Clone18 breast cancer cell line implanted in the mammary fat pad (MFP). A dynamic scan was performed over the first 60 min, followed by scans obtained at 4, 24, and 48 h. At earlier time points, the signal is predominantly in the heart (circulatory system), the bladder and the liver. At later time points, more signal is localized in the internal organs and the tumor. All images have been decay-corrected and normalized. The scale is reported as percent injected dose per cubic centimeter (% ID/cc).

23

with information reported about other viral particles, such as the Cowpea Mosaic Virus (CPMV) which locates mainly to the liver, gets cleared from the system within hours and does not maintain its infectivity after circulating in a mouse system, suggesting disassem-bly of the capsids.28, 29

2.5 Conclusions

We have efficiently radiolabeled the genome-free MS2 capsid with 64Cu radioisotopes and investigated the stability of the capsids in physiological conditions, both in in vitro assays and after circulation in a mouse model. MS2 agents with or without a PEG coating showed similar biodistributions and a large presence in the blood pool at the 24 h time point, indicative of a long circulation time. These results lay the foundation for developing the MS2 scaffold into an efficient imaging agent for cancer detection.

2.6 Materials and methods

Unless otherwise noted, all chemicals and solvents were of analytical grade and used as received from commercial sources. Polyethylene glycol (NH2-PEG5kDa-OMe) was purchased from Laysan Bio, Inc. (Arab, AL). Mouse serum was obtained from

% ID

/g

% ID

/g

MCF7 clone 18 MFP

MCF7 clone 18 MFP

no tumor

% ID

/g

blood

bone

brain fat he

artkid

ney

lg inte

stinelive

rlun

gmus

cle

panc

reas

sm int

estinesp

leen

stomac

htail

0

5

10

15

20

25

30

35

40

45

50

blood

bone

brain fat he

artkid

ney

lg inte

stinelive

rlun

gmus

cle

panc

reas

sm int

estinesp

leen

stomac

htail tum

or0

5

10

15

20

25

30

35

40

45

50

muscle tumor0

2

4

6

8

10

MS2 1 hMS2-PEG5kDa 1 hMS2 24 hMS2-PEG5kDa 24 h

Figure 2.7 Biodistribution studies of 64Cu-labeled MS2 scaffolds in nude mice with or without MCF7Clone18 breast cancer cell line implanted in the mammary fat pad (MFP). At 1 h and 24 h post-injection, organs were harvested, counted on a gamma counter and weighed. The percent injected dose per gram of tissue (%ID/g) was calculated for each organ and is represented as an average of biological triplicates (3 mice per group).

no tumors MCF7 clone 18 tumors

*

Retention time (min)

t=01 h

24 h

0

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10

64Cu-MS2-PEG5kDa

Rad

ioac

tivity *

Retention time (min)

t=01 h

24 h

0

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10

64Cu-MS2

*64Cu-MS2-PEG5kDa

Retention time (min)

t=0

24 h

48 h

00.20.40.60.81.0

0 2 4 6 8 10

*

Retention time (min)

t=0

24 h

48 h

0

0.2

0.4

0.6

0.8

1.0

0 2 4 6 8 10

64Cu-MS2

Figure 2.8 In vivo stability of MS2 with or without PEG5kDa modification. Agents were injected into the tail vein of nude mice with or without MCF7 clone 18 tumors. Blood samples were collected at several time points post-injection. The plasma was separated via centrifugation and was injected onto a HPLC equipped with a radioactivity detector. HPLC traces are normalized to the maximum intensity from each run.

24

Gibco/Thermo Fisher (Waltham, MA). Maleimide-DOTA (1,4,7,10-tetraazacyclodo-decane-1,4,7-tris-acetic acid-10-maleimidoethylacetamide) and pSCN-Bn-NOTA (2-S-(4-isothiocyanatobenzyl)-1,4,7-triazacyclononane-1,4,7-triacetic acid) were pur-chased from Macrocyclics (Dallas, TX). Maleimide-NOTA (2,2’-(7-(2-((2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)amino)-2-oxoethyl)-1,4,7-triazonane-1,4-diyl)diacetic acid), maleimide-NOTA-GA (2,2’-(7-(1-carboxy-4-((2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)amino)-4-oxobutyl)-1,4,7-triazonane-1,4-diyl)diacetic acid) and maleimide-DO-TA-GA (2,2’,2’’-(10-(1-carboxy-4-((2-(2,5-dioxo-2,5-dihydro-1H-pyrrol-1-yl)ethyl)ami-no)-4-oxobutyl)-1,4,7,10-tetraazacyclododecane-1,4,7-triyl)triacetic acid) were purchased from CheMatech (Dijon, France). Ethylenediaminetetraacetic acid (EDTA) disodium salt and potassium phosphate dibasic were purchased from EMD Chemicals Inc. (Darmstadt, Germany). Copper-64 was purchased from the Medical Cyclotron Laboratory, University of Wisconsin. Saline (sodium chloride) solution was Injection USP, 0.9% from APP Phar-maceuticals (Schaumburg, IL). Water (dd-H2O) used as reaction sol vent was deionized using a Barnstead NANOpure purification system (ThermoFisher, Waltham, MA). NAP desalting columns were purchased from GE Healthcare (Marlborough, MA). Spin concen-trators with different molecular weight cutoffs (MWCO) were from Millipore (Billerica, MA).

Liquid chromatography mass spectrometry (LC-MS) analysis. Acetonitrile (Optima grade, 99.9%, Thermo Fisher), formic acid (99+%, Pierce, Rockford, IL), and dd-H2O were used to prepare mobile phase solvents for LCMS. Electrospray ionization mass spectrometry (ESI-MS) of proteins was performed using an Agilent 1260 series liquid chromatograph outfitted with an Agilent 6224 time-of-flight (TOF) LCMS system (Santa Clara, CA). The LC was equipped with a Poroshell 300SB-C18 (5 µm particles, 1.0 mm × 75 mm, Agilent) analytical column. Solvent A was water with 0.1% formic acid and solvent B was acetonitrile with 0.1% formic acid (v/v). For each sample, approximately 15 to 30 picomoles of analyte were injected onto the column. Following sample injection, a gradi-ent of solvent B was run at a flow rate of 0.55 mL/min. Data was collected and analyzed using Agilent MassHunter Qualitative Analysis B.05.00.

High performance liquid chromatography (HPLC): Synthetic PEG derivatives were pu-rified by reverse phase semi-preparative HPLC on a C18 Gemini column (5 µm particles, 250 mm×10 mm, Phenomenex, Torrance, CA) using a gradient of water and acetonitrile with 0.1% TFA as eluent at a flow rate of 3 mL/min.

Gel Analyses: For protein analysis, sodium dodecyl sulfate-polyacrylamide gel elec-trophoresis (SDS-PAGE) was carried out on Tris gels in a with Mini-Protean apparatus from Bio-Rad (Hercules, CA) or on Bis-Tris gels in a Mini Gel Tank apparatus (Thermo), following the protocol from the manufacturer. The protein electrophoresis samples were heated for 10 min at 95 °C in the presence of β-mercaptoethanol to ensure reduction of any disulfide bonds. Gels were run for 35-60 min at 150-200 V in 2-(N-morpholino)eth-anesulfonic acid (MES) - SDS buffer to allow good separation of the bands. Commercially available markers (Bio-Rad) were applied to at least one lane of each gel for assignment of apparent molecular masses. Visualization of protein bands was accomplished by stain-ing with Coomassie Brilliant Blue R-250 (Bio-Rad). Quantification of the degree of mod-ification was obtained by imaging on a Gel Doc™ EZ imager (Bio-Rad) and subsequent

25

optical densitometry using the ImageJ (National Institutes of Health, Bethesda, MD) or Image Lab (Bio-Rad) software.

Procedure for expression and purification of T19paF N87C MS2 viral capsid: The expression and purification of genome-free bacteriophage MS2 has been previously re-ported. A yield of 5-10 mg/L culture was obtained for T19paF N87C MS2 following three rounds of purification.

Synthesis of nitrophenol-NHS ester. 3-(4-hydroxy-3-nitrophenyl)propanoic acid (1 equiv.) and N-hydroxysuccinimide (NHS, 1.1 equiv.) were dissolved in DCM at room tem-perature and 1-ethyl-3-(3-dimethylaminopropyl)carbodiimide (EDC, 1.1 equiv.) was add-ed. The reaction was stirred at room temperature for 3 h, at which point the crude mixture was poured into water and extracted with DCM twice. The combined organic extracts were dried over Na2SO4, filtered, and concentrated. Purification by silica gel chromatogra-phy (ethyl acetate 10% to 40% gradient in hexanes) gives the nitrophenol-NHS (NP-NHS) product as a bright yellow solid (~70% yield).

Synthesis of aminophenol-PEG5kDa-OMe: Amine-PEG5kDa-methoxy (NH2-PEG5kDa-OMe) (1 equiv.) was dissolved in DCM and NP-NHS ester (5 equiv.) and triethylamine (10 equiv.) were added sequentially. The reaction mixture was stirred at room tempera-ture overnight. The solvent was removed under high vacuum and excess NP-NHS was precipitated by the addition of 500 µL of water. The resulting precipitate was removed by filtration through a 0.45 µm spin filter (PALL, Port Washington, NY) and the filtrate was concentrated and purified using a 3 kDa MWCO spin concentrator. Subsequent reduction with Na2S2O4 (at a final concentration of 10 mM) for 20 min at room temperature afforded the aminophenol PEG conjugate. Excess reducing agent was removed by using 3 kDa MWCO spin concentrators.

Attachment of chelator molecules. To a 100 µM sample of protein (based on capsid monomer) in 10 mM potassium phosphate buffer, pH 7.2, was added 10 equiv. of maleim-ide-chelator from a 10 mM stock in DMSO. The reaction was allowed to proceed for 4 h at room temperature and was purified using a Nap-10 size exclusion column equilibrated with phosphate buffer, pH 6.5 and 100 kDa MWCO spin concentrators. The amount of modification was quantified using ESI LC-MS.

Modification with PEG5kDa. After modification of the internal Cys residues, MS2 capsids were reacted with aminophenol-PEG5kDa-OMe. The reaction took place at room tempera-ture for 5 min, in pH 7.2 buffer. The final concentrations were 100 µM for the MS2 mono-mer, 500 µM PEG and 1 mM of NaIO4. The reaction mixture was loaded onto a Nap-5 desalting column and eluted with phosphate buffer pH 7.2. Subsequent concentration and additional removal of the excess reagents was obtained using 100 kDa MWCO spin concentrators.

Radiolabeling of MS2 conjugates. The 64Cu copper stock (20-30 mCi, ~200 µL) was diluted with 1 mL of 0.1 M ammonium citrate buffer, pH 6.2 to generate a final volume of ~1200 µL at pH 5.5 (determined by pH paper). Each reaction tube was then charged with 100-400 µL of diluted 64Cu solution and 50-150 µL of the MS2 conjugates (200 µM in

26

capsid monomer). The complexation reactions were allowed to proceed for 2 h at room temperature, then purified using Nap-5 or Nap-10 columns. Samples were subsequently concentrated using 100 kDa MWCO spin concentrators. Centrifugation was performed at 10,000 rpm for 5 min per round of concentrating until the desired volume was reached.

In vitro stability studies. Concentrated 64Cu-chelator-MS2 conjugates (~300 µCi in 20 µL) were added to PBS or 100% mouse serum to a final volume of 200 µL (1:9 v/v). The samples were then incubated at 37 °C for a predetermined time using a temperature-con-trolled heat block (VWR, Radnor, PA). Aliquots were drawn from the samples at 1, 4, 8, 24 and 48 h time points, and injected onto a PolySep GFC-P5000 (Phenomenex, Torrance, CA) size exclusion chromatography column (300 x 7.8 mm, 5 µm particle size, 500 Å pore size; column flow rate 1.5 mL/min of 10 mM KH2PO4 containing 1 mM disodium EDTA, pH 7.2). The HPLC system consisted of a 590 HPLC pump (Waters, Milford, MA), UV de-tector operating at 280 nm (Linear Systems, Fremont, CA), Model 105S-1 high-sensitivity radiation detector with 1 cm3 CsI (T1) scintillating crystal coupled to a 1 cm2 Si PIN pho-todiode/low-noise preamplifier (Carroll-Ramsey Associates, Berkeley, CA), and fluores-cence detector (Spectra system FL3000, Thermo Separation Products, St. Peters, MO). Chromatography traces were collected using PeakSimple data system and software (SRI Instruments, Las Vegas, NV), and analyzed using the Gaussian multi-peak fitting feature of the OriginPro software v. 8.6.0 (OriginLab, Northampton, MA).

PET/CT and biodistribution studies. All animal procedures were performed according to a protocol approved by the UCSF Institutional Animal Care and Use Committee (IA-CUC). Six-week old female nu/nu (nude) mice weighing 18-23 g were purchased from Charles River Laboratories (Hollister, CA). For tumor inoculation, the cells were implanted in the number 4 mammary fat pad (β-estradiol pellets were also implanted subcutaneous-ly in the flank for MCF7 Clone 18 cells) or in the right flank for subcutaneous models. The imaging and biodistribution experiments were started approximately two weeks following implantation, when the tumors were ~1 cm in diameter.

Tumor-bearing nude mice in sets of 3 animals per study group were injected in the tail vein (intravenous, IV) with 150-250 µCi (5.5-9.25 MBq) of 64Cu-labeled MS2 conjugates in 100 µL of sterile saline. One animal from each group was selected for imaging with micro-PET/CT (Inveon microPET docked with microCT, Siemens, Washington, D.C.). Dynamic imaging was performed from the time of injection to 1 h post injection, followed by a 20 min static scan at 4, 24 and 48 h. CT scans were performed after each PET scan to pro-vide anatomical localization of radionuclide data, as well as photon attenuation map for attenuation-corrected PET reconstruction. Images were reconstructed using the AMIDE software v.1.0.4.

Mice were sacrificed at 1 h or 24 h post-injection and the blood, tumor, and ma-jor organs were harvested and weighed. The radioactivity present in each sample was measured using a Wizard gamma-counter (Perkin Elmer, Waltham, MA). All values were decay corrected, and the percentage injected dose per gram (% ID/g) was calculated for each tissue sample. Means and standard deviations were calculated within each group using Excel software (Microsoft, Redmond, WA).

27

To determine the assembly state of the agents after circulating in vivo, blood samples were allowed to coagulate (30 min at room temperature) and serum was collected by spinning at 10,000 g for 10 min and removing the supernatant. Serum aliquots were in-jected into an SEC HPLC system and the radioactivity channel was monitored for peaks corresponding to the assembled MS2 capsid, its monomer units or 64Cu associated with other serum proteins.

2.7 References

1. de Jong, M., Essers, J. and van Weerden, W.M., Imaging preclinical tumour models: improving translational power. Nat Rev Cancer. 2014, 14, 481-493.

2. Kherlopian, A.R., Song, T., Duan, Q., Neimark, M.A., Po, M.J., Gohagan, J.K., et al., A review of imaging techniques for systems biology. BMC Syst Biol. 2008, 2, 74.

3. Hillman, E.M., Amoozegar, C.B., Wang, T., McCaslin, A.F., Bouchard, M.B., Mansfield, J., et al., In vivo optical imaging and dynamic contrast methods for biomedical research. Philos Trans A Math Phys Eng Sci. 2011, 369, 4620-4643.

4. Pittet, M.J. and Weissleder, R., Intravital imaging. Cell. 2011, 147, 983-991.

5. Chapman, S., Dobrovolskaia, M., Farahani, K., Goodwin, A., Joshi, A., Lee, H., et al., Nanoparticles for cancer imaging: The good, the bad, and the promise. Nano Today. 2013, 8, 454-460.

6. Steinmetz, N.F., Viral nanoparticles as platforms for next-generation therapeutics and imaging devices. Nanomedicine-Nanotechnology Biology and Medicine. 2010, 6, 634-641.

7. Qazi, S., Liepold, L.O., Abedin, M.J., Johnson, B., Prevelige, P., Frank, J.A., et al., P22 Viral Capsids as Nanocomposite High-Relaxivity MRI Contrast Agents. Molecular Phar-maceutics. 2013, 10, 11-17.

8. Bruckman, M.A., Hern, S., Jiang, K., Flask, C.A., Yu, X. and Steinmetz, N.F., Tobacco mosaic virus rods and spheres as supramolecular high-relaxivity MRI contrast agents. J Mater Chem B Mater Biol Med. 2013, 1, 1482-1490.

9. Duconge, F., Pons, T., Pestourie, C., Herin, L., Theze, B., Gombert, K., et al., Fluo-rine-18-labeled phospholipid quantum dot micelles for in vivo multimodal imaging from whole body to cellular scales. Bioconjug Chem. 2008, 19, 1921-1926.

10. Liu, W., Zhang, S., Wang, L., Qu, C., Zhang, C., Hong, L., et al., CdSe quantum dot (QD)-induced morphological and functional impairments to liver in mice. PLoS One. 2011, 6, e24406.

11. Valegard, K., Liljas, L., Fridborg, K. and Unge, T., The three-dimensional structure of the bacterial virus MS2. Nature. 1990, 345, 36-41.

28

12. Cargile, B.J., McLuckey, S.A. and Stephenson, J.L., Jr., Identification of bacterio-phage MS2 coat protein from E. coli lysates via ion trap collisional activation of intact protein ions. Anal Chem. 2001, 73, 1277-1285.

13. Kovacs, E.W., Hooker, J.M., Romanini, D.W., Holder, P.G., Berry, K.E. and Fran-cis, M.B., Dual-surface-modified bacteriophage MS2 as an ideal scaffold for a viral cap-sid-based drug delivery system. Bioconjug Chem. 2007, 18, 1140-1147.

14. Garimella, P.D., Datta, A., Romanini, D.W., Raymond, K.N. and Francis, M.B., Multi-valent, high-relaxivity MRI contrast agents using rigid cysteine-reactive gadolinium com-plexes. J Am Chem Soc. 2011, 133, 14704-14709.

15. Wu, W., Hsiao, S.C., Carrico, Z.M. and Francis, M.B., Genome-free viral capsids as multivalent carriers for taxol delivery. Angew Chem Int Ed Engl. 2009, 48, 9493-9497.

16. Stephanopoulos, N., Tong, G.J., Hsiao, S.C. and Francis, M.B., Dual-surface modified virus capsids for targeted delivery of photodynamic agents to cancer cells. ACS Nano. 2010, 4, 6014-6020.

17. Glasgow, J.E., Capehart, S.L., Francis, M.B. and Tullman-Ercek, D., Osmolyte-medi-ated encapsulation of proteins inside MS2 viral capsids. ACS Nano. 2012, 6, 8658-8664.

18. Capehart, S.L., Coyle, M.P., Glasgow, J.E. and Francis, M.B., Controlled integration of gold nanoparticles and organic fluorophores using synthetically modified MS2 viral capsids. J Am Chem Soc. 2013, 135, 3011-3016.

19. Behrens, C.R., Hooker, J.M., Obermeyer, A.C., Romanini, D.W., Katz, E.M. and Fran-cis, M.B., Rapid chemoselective bioconjugation through oxidative coupling of anilines and aminophenols. J Am Chem Soc. 2011, 133, 16398-16401.

20. Carrico, Z.M., Romanini, D.W., Mehl, R.A. and Francis, M.B., Oxidative coupling of peptides to a virus capsid containing unnatural amino acids. Chem Commun (Camb). 2008, 1205-1207.

21. Anderson, C.J., Wadas, T.J., Wong, E.H. and Weisman, G.R., Cross-bridged mac-rocyclic chelators for stable complexation of copper radionuclides for PET imaging. Q J Nucl Med Mol Imaging. 2008, 52, 185-192.

22. Zinn, K.R., Chaudhuri, T.R., Cheng, T.P., Morris, J.S. and Meyer, W.A., Jr., Production of no-carrier-added 64Cu from zinc metal irradiated under boron shielding. Cancer. 1994, 73, 774-778.

23. Liu, Y. and Welch, M.J., Nanoparticles labeled with positron emitting nuclides: advan-tages, methods, and applications. Bioconjug Chem. 2012, 23, 671-682.

24. Storm, G., Belliot, S.O., Daemen, T. and Lasic, D.D., Surface Modification of Nanopar-ticles to Oppose Uptake by the Mononuclear Phagocyte System. Advanced Drug Deliv-ery Reviews. 1995, 17, 31-48.

25. O’Riordan, C.R., Lachapelle, A., Delgado, C., Parkes, V., Wadsworth, S.C., Smith,

29

A.E., et al., PEGylation of adenovirus with retention of infectivity and protection from neu-tralizing antibody in vitro and in vivo. Hum Gene Ther. 1999, 10, 1349-1358.

26. Caliceti, P. and Veronese, F.M., Pharmacokinetic and biodistribution properties of poly(ethylene glycol)-protein conjugates. Adv Drug Deliv Rev. 2003, 55, 1261-1277.

27. Yu, M. and Zheng, J., Clearance Pathways and Tumor Targeting of Imaging Nanopar-ticles. ACS Nano. 2015, 9, 6655-6674.

28. Rae, C.S., Khor, I.W., Wang, Q., Destito, G., Gonzalez, M.J., Singh, P., et al., System-ic trafficking of plant virus nanoparticles in mice via the oral route. Virology. 2005, 343, 224-235.

29. Singh, P., Prasuhn, D., Yeh, R.M., Destito, G., Rae, C.S., Osborn, K., et al., Bio-dis-tribution, toxicity and pathology of cowpea mosaic virus nanoparticles in vivo. Journal of Controlled Release. 2007, 120, 41-50.

30

Chapter 3: MS2-antibody conjugates for cancerimaging applications

The work presented in this chapter was performed in collaboration with Dr. Adel El-Sohly, Dr. Chawita Netirojjanakul, Prof. Michelle Farkas, Kevin Bond and Prof. Martin Jarrold at Indiana University, and Astraea Jager, Prof. Gary Nolan and Prof. Sean Bendall at Stanford University. Parts of this work were published in “Synthetically Modified Viral Capsids as Versatile Carriers for Use in Antibody-Based Cell Targeting” (DOI: 10.1021/acs.bioconjchem.5b00226) and “Biodistribution of Antibody-MS2 Viral Capsid Conjugates in Breast Cancer Models” (DOI: 10.1021/acs.molpharmaceut.6b00566) and were repro-duced with permission.

Abstract

In the last three decades, scientists have explored the prospect of endowing nanopar-ticles with targeting moieties that could increase their homing to a region of interest. Targeted nanoparticles are particularly valuable as diagnostic and therapeutic carriers because they can increase the signal-to-background ratio of imaging agents, improve the efficacy of drugs, and reduce adverse effects by concentrating the therapeutic molecule in the region of interest. A large number of nanoparticle and targeting moiety combina-tions have been investigated, with various degrees of success. Although encouraging results have been observed in in vitro settings, there is a paucity of information about the distribution of these targeted agents in vivo. In this chapter, the synthesis and character-ization of MS2-antibody conjugates is described, and the in vitro targeting of cancer cells is evaluated. Charge detection mass spectrometry is used as a novel technique to deter-mine the amount of modification upon attaching the full length antibodies. Additionally, we investigate the stability and immunogenicity of MS2-antibody constructs and determine the biodistribution of these agents in murine cancer models.

31

3.1 The tumor microenvironment

To design treatments with high specificity for the tumor over the healthy tissue, a thor-ough understanding of the tumor environment must be attained. A tumor forms from a single cell that undergoes a set of mutations, causing it to proliferate uncontrollably.1 After the tumor reaches a certain size, it begins the recruitment of host blood vessels or the for-mation of new blood vessels. The tumor vasculature is highly abnormal, with blood ves-sel density varying greatly from the tumor-host interface to the central part of the tumor. There are at least six distinctly different types of tumor blood vessels: feeding arteries, mother vessels, glomeruloid microvascular proliferations (GMP), vascular malformations, capillaries and draining veins (Figure 3.1a).2 Nanoparticles are most likely to cross the vascular barrier and enter the tumor environment through mother vessels. These vessels are hyperpermeable to plasma proteins because they are only lined by a thin layer of endothelial cells, with little or no pericyte coverage and have disrupted basement mem-branes.3, 4 The predominant belief in the nanomedicine community is that nanoparticles take advantage of gaps (100–500 nm in size, depending on the tumor type and stage) that form between adjacent endothelial cells on the wall of these blood vessels (Figure 3.1b).5-7 Tumors also have a reduced ability to drain fluid and waste from the interstitial space through the lymphatic system.7 The combination of hyperpermeability and poor drainage was labeled as the enhanced permeability and retention effect (EPR) and has been used by researchers for the last few decades to explain the passive accumulation of nanoparticles in the tumor environment.8 An alternative hypothesis is that nanoparticles

Tumor cell

Dendritic cell

Macrophage

Endothelial cell

Smooth musclecell

Nanoparticle

Pericyte

Host vessel

Mother vessel Vascularmalformation

Draining veinFeeding artery

Pores

Intercellular Glomeruloid microvascular proliferation

Capillary

Figure 3.1 Tumor microenvironment and blood vessels. (a) The tumor consists of cancerous cells, as well as macrophages and dendritic cells. Their interaction with nanoparticles can differ based on the properties of the nanoparticle and the targeting groups used. The different types of blood vessels feeding the tumor can explain the variability in tumor permeability. For example, mother vessels are highly permeable not only to small molecules but also to large plasma proteins. Mother vessels may also include pores (100-500 nm) through which macromolecules can extravasate (b). Mother vessels are unstable and, over time, differenti-ate into glomeruloid microvascular proliferations, vascular malformations and capillaries. These smaller, more irregular vessels, are less permeable to nanoparticles and transcellular transport (b) is more likely than intercellular transport (c). Adapted from reference 2.

Vesicle transport

Transcellular(a) (b) (c)

32

can extravasate into tumors via a transcellular pathway (Figure 3.1c). Endothelial cells contain structures called vesiculo–vacuolar organelles that span the cytoplasm from the lumen to the albumen and can transport materials across the cell.9, 10 It is unclear which pathway dominates the extravasation of nanoparticles in tumors; however, the scarcity of the endothelial gaps might suggest that investigating the transcellular route would be a worthwhile endeavor to guide the design of novel nanoparticles.

Once the nanoparticles cross the vascular barrier, they have to navigate through the tumor to reach the cells of interest. The tumor is a highly heterogeneous environment consisting of the tumor stroma (fibroblasts, pericytes and immune cells) and parenchyma (malignant tumor cells).2, 11 Additionally, extracellular matrix components, such as col-lagen, fibronectin, hyaluronan, fibrin and proteoglycans, are present in the tumor envi-ronment, leading to a high interstitial fluid pressure.11 This pressure can influence the transport and distribution of chemotherapeutics, imaging agents, macromolecules and nanoparticles in the tumor.12, 13

Nanoparticles have been synthesized with numerous surface modifications to attempt to increase their interactions with the tumor cells. Active targeting approaches rely on functionalizing the surface of nanoparticles with ligands that can target tumor-specific markers. In passive targeting, the nanoparticle surface is coated only with stabilizing agents. The physicochemical properties of the particles greatly influence their distribution within the tumor, with a general consensus that smaller nanoparticles penetrate more deeply into the tumor extracellular matrix, whereas their larger counterparts are restricted to the immediate vicinity of the vascular extravasation point.14, 15 Threfore, the physico-chemical properties may determine whether nanoparticles coated with tumor-cell-target-ing ligand, such as aptamers, antibodies, peptides or small molecules, get the chance to interact with the tumor cells of interest.16

3.2 Current examples of targeted virus-like particles

Although great strides have been made in the development of new chemotherapeutic drugs, their toxicity to healthy tissues still poses limitations in clinical applications.17, 18 To address these issues, significant efforts have been dedicated to the design of carriers that can deliver the desired cargo selectively to tumor sites.1, 7, 19, 20 Nanoparticles have been decorated with different types of targeting groups (small molecules,21, 22 aptamers,23, 24 peptides,25-27 glycans,28 antibodies29-32) with the hope of increasing their specificity to the cancerous tissue.

Numerous studies have investigated the effect of active targeting on tumor uptake of virus-like particles (VLPs).2 Unfortunately, most of these studies only examined the interaction with cultured mammalian cells. Some examples include the use of the small molecule folic acid to target the folate receptor, overexpressed on a variety of cancer types. Folic acid has been conjugated to Cucumber mosaic virus (CMV),33 Hibiscus chlo-rotic ringspot virus (HCRSV),21 and adenovirus,34 and in each case was shown to lead to uptake by target cells.35 Lactobionic acid and doxorubicin have been attached to rotavirus VP6 for delivery to hepatic cancer cells, with encouranging in vitro results.36 Larger tar-geting groups, such as peptides, nucleic acid aptamers, carbohydrates, and proteins, can

33

be attached to capsid surfaces. The red clover necrotic mottle virus (RCNMV) modified with doxorubicin and a 16 amino acid, CD46-binding peptide, was shown to be cytotoxic to HeLa cells.37 Transferrin was used as a targeting group on HK97 capsids to deliver drug-like fluorophores,38 as well as on the MS2 capsid to deliver ricin A.39 Antibodies have gained a lot of popularity as targeting groups, given their widespread availability, high specificity and excellent affinity for their targets. For example, immunoliposomes are some of the most advanced nanoparticles on the way to the clinic and have demonstrated encouraging results in treating several types of cancer.40

Previous work in the Francis Group has shown successful application of the MS2 viral capsid as a targeted delivery vehicle for imaging agents and therapeutics using peptides or DNA aptamers as targeting moieties.23, 25-27, 41, 42 Here, we expand the classes of targeting groups to include antibodies, the targeting moiety with the greatest variety of targets, and the best binding specificity and affinity. In this chapter, we investigate the use of MS2-antibody conjugates for mass cytometry (CyTOF)43 and in vivo positron emission tomography (PET) imaging applications.44 The enhancement in signal intensity gained by the attachment of >100 copies of metal chelators in one agent could be particularly useful in both applications.

3.3 Synthesis of MS2-antibody conjugates

The general modification strategy for obtaining targeted MS2 capsids that can deliver imaging tracers or drug molecules is to attach the cargo on the interior surface (protected from interacting with other proteins) and the targeting moiety on the exterior surface of the capsid, to allow for unobstructed interaction with the targeted receptor (as discussed in Chapter 2). We investigated the use of anti-EGFR and anti-HER2 antibodies as target-ing moieties, given their prevalent overexpression in several types of cancer, including breast, pancreatic and non-small cell lung cancer.45 Constructs with anti-CD20, anti-CD3, anti-CD44 and anti-VCAM antibodies were also synthesized and will be mentioned briefly in this chapter and Chapter 4.

3.3.1 Interior surface modification

As described in Chapter 2, the internal surface modification leveraged the presence of 2 nm pores that allow for the diffusion of small molecules and of the engineered Cys res-idue at position 87. The lack of other solvent-accessible Cys residues afforded site spe-cific modification via maleimide-thiol chemistry. For cell binding assays, fluorescent dyes with maleimide moieties were used. For CyTOF experiments and imaging/biodistribution studies, chelators for lanthanides or 64Cu radioisotopes, respectively, were attached to the interior surface.

3.3.2 Exterior surface modification

Given the large size of antibodies and the MS2 viral capsid, we selected an oxidative coupling of these biomacromolecules as the conjugation reaction of choice. This reaction has been previously shown to couple two biomolecules efficiently under mild conditions and short reaction times by using aniline and aminophenol coupling partners. MS2 viral capsids containing an aniline in the form of the non-canonical amino acid p-aminophe-

34

nylalanine (paF) on the exterior surface was obtained using an amber codon/tRNA syn-thetase system, as described previously. The nitrophenol was attached to antibodies via non-site-specific lysine modification using a nitrophenol-NHS ester, followed by reduction to the corresponding aminophenols. The coupling reaction was then performed in the presence of NaIO4 at room temperature for 5 min (Figure 3.2a).

Initial studies designed to investigate the possibility of generating MS2-Ab conjugates used an anti-human IgG mouse monoclonal antibody as a model substrate. To find the

MWladder 1 2 3

50

37

25

20

75100150250

15

10

kDa

HC + 2 MS2 HC + 1 MS2 HC LC + 2 MS2 LC + 1 MS2

MS2 monomer

LC MS2-PEG5kDa

MS2-PEG2kDa

4 5

NaIO4

AP-αEGFR + – + – –

AP-PEG5kDa-αEGFR – – – + +

– – + + +

MeO-PEG2kDa-AP – – – +–

paF-MS2 + + + +–

light chain

heavy chain

+0+1

+2

+0

+1

+2

+3+4

49000 49500 50000 50500

4932

1 4951

4

4970

7

4990

0

5009

4

23000 23500 24000 24500

2349

7

2369

0

2388

3

mass (Da)

ratio Ab:MS2 3:1 5:1 10:1

1 2 3 4 5 6

20:1

HC + 2 MS2 HC + 1 MS2 HC LC + 2 MS2 LC + 1 MS2

MS2 monomer

LC

NaIO4

AP-αEGFR + – +

– – + + +

paF-MS2 + + + +–

+ +

+

+

+

MWladder

(a)

50

37

2520

75100150250

15

10

kDa

(c) (d)

Antibody (Ab)

NaIO42. Na2S2O420 min, RT

1. , 2.5 h

2 min, RT

MS2-Ab

MS2

Ab-AP

(b)

Figure 3.2 Synthesis and characterization of MS2-antibody conjugates (MS2-Ab). (a) Synthetic scheme for the generation of MS2-Ab. First, nitrophenol (NP) groups were attached to antibodies via lysine modification using a NP-NHS ester. The nitrophenol groups were then reduced to yield aminophenol-Ab conjugates (AP-Ab) by addition of Na2S2O4. The resulting AP-Ab was attached to paF MS2 via oxidative coupling using NaIO4. (b) LC-MS analysis of humanized anti-HER2 antibodies after lysine modification with 5 equiv. of NP-NHS. The light chains (LC) were either unmodified (59%) or modified with one (34%) or two (6%) NP groups. The heavy chains (HC) were modified with 0 (16%), 1 (40%), 2 (31%), 3 (12%) or 4 (1%) NP groups. (c) SDS-PAGE analysis of MS2-anti-EGFR conjugates using 3, 5, 10, and 20 equiv. of αEGFR-aminophenol (AP-αEGFR) with respect to the MS2 capsid concentration. The gel showed conjuga-tion of one or two MS2 monomers to the light chain (LC) and heavy chains (HC) of antibodies. More equiva-lents of Ab resulted in a higher intensity of the modified bands. (d) SDS-PAGE analysis of MS2-Ab conju-gates using 3 equiv. of AP-αEGFR in excess with respect to the MS2 capsid concentration. The gel showed conjugation of one or two MS2 monomers to the LC and HC of antibodies. The PEG2kDa and PEG5kDa modifi-cations are observed as higher molecular weight bands.

OHNO2

O

NHS

OHH2N

O

HN

35

optimal number of aminophenol coupling partners to be attached to Ab, the number of NP-NHS in the first step was varied from 5 to 100 equivalents. LCMS analysis revealed that with 5 equivalents of NP-NHS, ~30% of the light chains were modified with one NP and ~75% of the heavy chains had one, two, or three NPs (Figure 3.2b). More nitrophenol groups were appended on both the heavy and light chains as the number of NP-NHS in-creased, and increased amounts of crosslinked products were produced in the presence of NaIO4 when a large excess of reagent was used (data not shown). Therefore, 5 equiv-alents of NP-NHS were used for the generation of AP-Ab in all the following experiments.

When we varied the number of antibody equivalents per capsid (3, 5, 10, or 20), we observed increasing amounts of conjugation, as analyzed by optical densitometry per-formed on SDS-PAGE gels (Figure 3.2c). The light chains were usually modified with one or two MS2 monomer proteins. The heavy chains were mainly modified with one MS2 monomer. For some of the heavy chain modifications, band diffusivity limited the analysis of the conjugates.

3.3.3 Linkers between antibody and MS2

We hypothesized that upon attachment, the conjugated antibodies may be hindered in their interaction with the target receptor. As such, we investigated the use of a 5 kDa polyethylene glycol chain (PEG5kDa) linker between the capsid and antibody to allow for

MWladder 1 2 3

50

37

2520

75100

150

250

15

10

kDa

4 5 6 7

αEGFR

+ – + + +

NP-PEG5kDa-αEGFR – – – +– –– +

–paF-MS2

+ + + ++ –

+

–+

Spin concentration – –– + +

befo

re

afte

r

befo

re

afte

r

HC

MS2 monomer

LC

LC + PEG5kDa

HC + PEG5kDa

0 3 6 9 12 150

5

10

15

20

unconjugated Ab

0 3 6 9 12 150

5

10

15

20

25

A28

0

Retention time (min)

A28

0

Retention time (min)

Na2SO3 and reduced

periodate

(a) (b)

Figure 3.3 Removal of excess antibody. (a) SDS-PAGE analysis of the efficiency of removing unreacted antibodies using spin concentrators with a molecular weight cutoff of 100 kDa. The MS2 capsids were incubated with αEGFR (lanes 3 and 4) or PEG5kDa-αEGFR (lanes 5 and 6) in a 3:1 antibody:capsid stoichi-ometry in the absence of NaIO4. Half of the mixture was loaded onto a spin concentrator with molecular weight cutoff of 100 kDa and washed 5 times with 10 mM phosphate buffer pH 7.2. The other half was kept separately as reference. The MS2 viral capsid was loaded at a high concentration of coat protein, thus some higher molecular weight impurities are visible. The modification of the light and heavy chains of the antibody with PEG polymers can be observed in lane 7. (b) SEC-HPLC of the oxidative coupling of human-ized AP-αEGFR and paF-MS2 when a 3:1 ratio of antibody:capsid was used indicates complete consump-tion of the antibody. The top trace was taken before the addition of sodium periodate, and the bottom trace was taken after addition of sodium sulfite to quench the reaction after 6 min (peak at ~11.5 min).

36

unobstructed interactions with the targets. The PEG chain could also have a shielding effect from the immune system and, therefore, was used in most in vivo experiments.46 We also explored constructs with a small number of antibodies linked to the MS2 capsids through PEG5kDa linkers and additional surface coating with PEG2kDa (i.e. “backfilling” via a subsequent oxidative coupling reaction) in order to shield the surface of the viral capsid while still allowing the antibodies to interact with their targets (Figure 3.2d).

3.3.4 Removal of unreacted antibody

We investigated the efficient removal of the excess antibody by two methods: SDS-PAGE and HPLC SEC (Figure 3.3). The MS2 capsids were incubated with αEGFR or PEG5kDa-αEGFR in a 3:1 antibody:capsid stoichiometry as used in conjugation reactions, but in the absence of the periodate oxidant. Half of the mixture was loaded onto a spin concentrator with molecular weight cutoff of 100 kDa and washed 5 times with 10 mM phosphate buffer pH 7.2. The other half was kept separately as reference. SDS-PAGE analysis shows that most of the antibody was efficiently removed from solution, indicating that the 100 kDa molecular weight cutoff (MWCO) spin concentrators are able to remove the unbound antibody (Figure 3.3a). Additionally, SEC-HPLC of the oxidative coupling of humanized AP-αEGFR and paF-MS2 when a 3:1 ratio of antibody:capsid was used indi-cates complete consumption of the antibody. The top trace was taken before the addition of sodium periodate, and the bottom trace was taken after addition of sodium sulfite to quench the reaction after 6 min (peak at ~11.5 min in Figure 3.3b).

3.4 Characterization of MS2-antibody conjugates

3.4.1 SDS-PAGE

SDS-PAGE followed by optical densitometry was the routine analysis used to charac-terize the conjugates formed between the MS2 capsid and the antibodies of interest. To note, some limitations of this method are the different interaction of the Coomassie stain with the smaller molecular weight species versus the higher molecular weight species, given the lower number of positive residues that the (negatively-charged) dye can interact with. Additionally, the diffuse character of some of the bands made it difficult to determine the percentage of that particular modification in the reaction mixture.

3.4.2 Dynamic Light Scattering

Dynamic light scattering (DLS) (Figure 3.4a) suggested homogeneous populations of MS2 derivatives, with the hydrodynamic radii increasing from 27 nm to 30.7 nm upon conjugation of antibodies, and to 28.9 nm when PEG5kDa was used as a linker. In the back-filling case, the measured diameter was 33.7 nm. We hypothesize that, when the two car-bon linker is used, the antibodies are attached tangentially to the capsid surface, explain-ing the relatively small increase in diameter. Given the surface of the MS2 bacteriophage (~2300 nm2) and the number of PEG chains attached (100-130 per capsid), the density of PEG chains is likely insufficient to extend the polymers to the “brush” conformations. This expectation is supported by the fact that the predicted length of the fully extended PEG chain is ~16 nm for PEG2kDa and ~35 nm for PEG5kDa, but only small increases in nanoparticle diameter are observed by DLS.

37

3.4.3 Transmission Electron Microscopy

Transmission Electron Microscopy (TEM, Figure 3.4b) revealed that the antibody-mod-ified capsids maintained their morphology and had very similar size to the unmodified MS2 VLPs. Some of the MS2-Ab capsids appeared to have a thicker protein shell in the micrographs, but the difference in thickness was not found to be statistically significant when compared to the MS2 capsids.

3.5. Charge Detection Mass Spectrometry for determining the number of antibod-ies attached to the MS2 capsids

The quantitative determination of the number of antibodies bound to the surface of the capsids was a challenging task. As discussed previously, SDS-PAGE is unable to provide quantitative information about the modifications and, since the capsids are dena-tured, only the average number of modifications can be observed. ESI-MS is limited by the denaturing that takes place in the organic solvent and by the large ratio of the signal coming from the unmodified coat proteins as compared to the antibody fragment-coat protein conjugates. Fortunately, recent advances in mass spectrometry,47 particularly na-tive mass spectrometry, make it possible to analyze these complex biological systems with full length antibody modifications.

In native mass spectrometry, electrospray ionization from a non-denaturing, volatile solution allows large non-covalently bound assemblies to be transferred from solution

0

20

40

60

80

100

Nor

mal

ized

Cou

nts

Mass (MDa)

(a) (b)

RF hexapole

RF only hexapole and quadrupole

Dual hemispherical deflection energy analyzer

Cylindrical ion trap

Ion lens and deflectors

Ion lens

SID/CID cellQuadrupole

Ion funnel

Nanoelectrospray

source

MS2MS2-3Ab

0 1 2 3 4 5 6

Figure 3.5 Charge Detection Mass Spectrometry (CDMS). (a) Diagram of the instrument used to collect CDMS data, courtesy of the Jarrold lab. (b) CDMS spectra of MS2 untreated (black trace) and conjugated with 3 equiv. anti-EGFR antibody (red trace). The peak for the MS2 capsid was observed at a higher mass than expected (3.9 MDa versus 2.4 MDa). The width of the peak did not allow for resolving species with different number of antibodies attached.

MS2-αEGFR MS2-PEG5kDa-αEGFR MS2

Sample PEG modification PEG strands Diameter (nm)(a) (b)

MS2-PEG2kDa/5kDa-αEGFR 33.7 ± 0.6 50 nm 50 nm50 nm

MS2 27.1 ± 1.4MS2-αEGFR 30.7 ± 0.9

~2~130

~1% PEG5kDa-αEGFR ~70% PEG5kDa-OMe

MS2-PEG5kDa-αEGFR 28.9 ± 1.4~2~1%MS2-PEG2kDa-OMe 31.2 ± 0.9~130~70%MS2-PEG5kDa-OMe 35.6 ± 1.3~130~70%

Figure 3.4 Physical characterization of MS2-antibody conjugates. (a) DLS measurements of the diameter of MS2-Ab conjugates. The percentage of PEG modification is estimated from gel densitometry measure-ments of SDS-PAGE gels using ImageJ software. (b) TEM micrographs of MS2-Ab conjugates.

38

into the gas phase.48 Nevertheless, as size increases, it becomes increasingly difficult to resolve the charge states.49 Charge detection mass spectrometry (CDMS, Figure 3.5a) is an alternative method for measuring the masses of ions in the kDa and MDa mass regime that circumvents the ongoing challenge of charge state resolution.50-52 CDMS is a sin-gle-particle technique where the masses of individual ions are determined from simulta-neous measurement of each ion’s mass-to-charge ratio (m/z) and charge.50 CDMS does not have a theoretical upper mass limit, making it particularly well-suited to the analysis of high mass analytes, such as virus capsid conjugates.53

Initial CDMS experiments of the unmodified capsid indicated a mass peak of 3.82 MDa (Figure 3.5b), which was significantly higher than expected for the empty viral cap-sid of 2.4 MDa (180 times the molecular weight of each monomer, 13.8 kDa). This error in mass cannot be accounted for by instrumental error. We hypothesized that this obser-vation was caused by adventitious RNA or DNA that was encapsulated in the viral capsid during the assembly process. The width of the peak (>500 kDa) was too large to allow for the discrimination of the modification with different numbers of antibodies, which each

MS2 untreated

MS2 alkaline

disassembly

nucleic acid precipitation reassembly

around RNA

MS2 hybrid

circular and linear DNA

E. coli RNA

pH 11.8RNA removal

Yeast RNA

pH 11.8RNA removal

untreated alkaline hybrid

0.2

0.4

0.6

0.8

1

1.2

220 240 260 280 300 320 340

untreatedalkalinehybrid

0

Wavelength (nm)

Abs

orba

nce

Num

ber (

%)

0

10

20

30

40

Diameter (nm)

(a) (b)

(c)

(d)

Figure 3.6 Synthesis and characterization of MS2 capsids without encapsidated nucleic acid material. (a) Diagram of the protocol used to remove DNA and RNA from MS2 expressed recombinantly in E. coli. (b) UV-VIS spectra of MS2 capsids upon alkaline and hybrid treatments, showing a reduction in the A260 peak corresponding to the presence of nucleic acids. (c) Transmission Electron Micrographs (TEM) of capsids after the removal of nucleic acids, showing intact morphology of the capsids. (d) Dynamic Light Scattering (DLS) spectra showing the retention of the assembled state of the capsids after alkaline treatment and successful reassembly during the hybrid treatment. (e) Upon treatment of the MS2 capsids with the alkaline or the hybrid treatment, a significant shift in mass is observed. In the case of the hybrid treatment, the peak is much narrower as compared to the untreated capsid. To note, there is an additional peak at 3.2 MDa in the hyrbid treatment trace, which could be assigned to incomplete removal of the nucleic acids.

0.1 1 10 100 1000 10000

untreatedalkalinehybrid

untreatedalkalinehybrid

(e)

0

20

40

60

80

100

Nor

mal

ized

Cou

nts

Mass (MDa)0 1 2 3 4 5 6

39

add 150 kDa to the mass of the capsid. Therefore, several sample preparation protocols were attempted to remove genetic material and narrow the peaks to a width that would allow resolution of the antibody modifications (Figure 3.6b).

3.5.1 Obtaining empty MS2 capsids

Taking advantage of the sensitivity of RNA to high pHs and the stability of the MS2 capsid in these conditions, we first attempted to remove the RNA by using a solution of phosphate buffer at pH 11.8. This “alkaline method” (Figure 3.6a) was able to greatly reduce the 260 nm peak usually observed in the UV spectra of MS2 expressed recombi-nantly in E.coli, indicating efficient removal of the genetic material (Figure 3.6b). The pro-tocol leads to intact capsids as observed by TEM (Figure 3.6c) and DLS (Figure 3.6d). However, the mass of the capsid was still higher than expected, at 3.15 MDa (Figure 3.6e). We hypothesized that this extra mass could come from DNA fragments that are resistant to hydrolysis under high pH conditions.

In order to fully remove the remaining genetic material from inside the capsid, we de-veloped a “hybrid method” including a disassembly step, followed by reassembly of the MS2 coat protein around yeast RNA and subsequent alkaline treatment (Figure 3.6a). This method also leads to the formation of homogeneous capsids of the expected size, as shown by TEM and DLS (Figure 3.6c and d), with a relatively narrow mass peak as

hybrid

hybrid-3Ab

hybrid-10Ab

MS2 untreated MS2 alkaline MS2 hybrid

antibody attachment

1 2 3 4 5 6

HC + 2 MS2 HC + 1 MS2 HC LC + 2 MS2 LC + 1 MS2

MS2 monomer

LC

50

37

2520

75100150250

15

10

kDa

untreated alkaline hybrid

MW

(a)

(c)

(b)

Figure 3.7 Synthesis and characterization of MS2-antibody conjugates without nucleic acid material. (a) The untreated, alkaline and hybrid treated capsids were conjugated to anti-EGFR antibodies as described in Figure 3.2, leading to a mixed population of conjugates. (b) SDS-PAGE analysis of the antibody modifica-tion of the MS2 capsids. Lanes 1,3 and 5 contain the starting material for each type of capsid. Lanes 2, 4 and 6 show additional bands, corresponding to the light chain (LC) and the heavy chain (HC) of the antibod-ies and their conjugates with MS2 monomers. To note, the impurities observed in the untreated sample are removed during the hybrid method protocol. (c) Upon conjugation with 3 or 10 equiv. of antibody, there is a small shift in the diameter of the capsids, as indicated in the DLS spectrum.

Num

ber (

%)

0

10

20

30

40

Diameter (nm)0.1 1 10 100 1000 10000

40

measured by CDMS (Figure 3.6e).

3.5.2 CDMS analysis of the MS2-Ab conjugates

Antibodies were attached to the capsids treated with the alkaline or hybrid method by the protocol described in Section 3.3.2. The characterization of the resulting conjugates is presented in Figure 3.7b. For the capsids treated with the hybrid method, 3 and 10 equiv. of antibody were used in the coupling reaction, and a small change in diameter was observed by DLS (Figure 3.7c).

Samples were buffer ex-changed into a 100 mM ammoni-um acetate solution to minimize the presence of salt adducts during electrospray. Hybrid MS2 samples modified with 0, 3, and 10 equiva-lents of antibody were analyzed by CDMS, and the results are shown in Figure 3.8. Gaussian curves equally spaced at the molecular weight of an antibody (0.15 MDa) were used to fit the observed spectrum. A trend of increased amount of antibody at-tached to the capsids was observed with an increased equivalence of antibody added to solution in the coupling step. Additionally, there was a smaller percentage of cap-sids from the 10 equiv. sample with zero antibodies attached, indicating an increased amount of coupling at a higher antibody concentration (Ta-ble 3.1).

In summary, CDMS was unique-ly suited to answer a very difficult question regarding the number of antibodies attached to each MS2 capsid. Initial experiments were limited by the peak width of the recombinant-ly-expressed MS2 capsids. The removal of nucleic acids lead to much narrower peaks and the modification with antibodies was successfully quantified using CDMS.

3.6 In vitro interactions with cancer cell models

Fluorescently-labeled MS2 conjugates were used in flow cytometry and confocal mi-

Nor

mal

ized

Cou

nts

Mass (MDa)MS2 (hybrid)MS2 (total fit)MS2 (individual fits)

Mass (MDa)

MS2-Ab, 3 equiv (total fit)MS2 (total fit)MS2-Ab, 3 equiv (individual fits)

MS2-Ab, 3 equiv (hybrid)

Mass (MDa)

MS2-Ab, 10 equiv (total fit)MS2 (total fit)MS2-Ab, 10 equiv (individual fits)

MS2-Ab, 10 equiv (hybrid)

(a) (b) (c)

Figure 3.8 Charge Detection Mass Spectrometry of viral capsids without nucleic acids. (a) The shape of the peak of the hybrid-treated MS2 capsids (black trace) was fit to a series of Gaussian curves (gray traces). The sum of the curves matches the observed spectrum (red trace). The sum of the fits was used to determine the species present after modifica-tion with (b) 3 or (c) 10 equiv. of antibodies. The red trace was used at the fixed distance of 0.15 MDa, corresponding to the addition of a full-length antibody (gray traces). The sum of all the fits is represented as the blue trace and nicely fits the observed spectra (black traces).

0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6

No. Abattached untreated alkaline

51.7 150 26.931.9 18.91 20.42.9 36.42 15.69.7 12.93 11.81.7 12.94 9.42.1 3.9>4 15.9

hybrid

Table 3.1 The distribution of species modified with different amounts of antibodies was determined by integrating the curves from CDMS spectra of samples with a 3:1 initial ratio of Ab:capsid.

Percentage of population (%)

Nor

mal

ized

Cou

nts

Mass (MDa)MS2 (hybrid)MS2 (total fit)MS2 (individual fits)

Mass (MDa)

MS2-Ab, 3 equiv (total fit)MS2 (total fit)MS2-Ab, 3 equiv (individual fits)

MS2-Ab, 3 equiv (hybrid)

Mass (MDa)

MS2-Ab, 10 equiv (total fit)MS2 (total fit)MS2-Ab, 10 equiv (individual fits)

MS2-Ab, 10 equiv (hybrid)

(a) (b) (c)

Figure 3.8 Charge Detection Mass Spectrometry of viral capsids without nucleic acids. (a) The shape of the peak of the hybrid-treated MS2 capsids (black trace) was fit to a series of Gaussian curves (gray traces). The sum of the curves matches the observed spectrum (red trace). The sum of the fits was used to determine the species present after modifica-tion with (b) 3 or (c) 10 equiv. of antibodies. The red trace was used at the fixed distance of 0.15 MDa, corresponding to the addition of a full-length antibody (gray traces). The sum of all the fits is represented as the blue trace and nicely fits the observed spectra (black traces).

0 1 2 3 4 5 6 0 1 2 3 4 5 6 0 1 2 3 4 5 6

No. Abattached untreated alkaline

51.7 150 26.931.9 18.91 20.42.9 36.42 15.69.7 12.93 11.81.7 12.94 9.42.1 3.9>4 15.9

hybrid

Table 3.1 The distribution of species modified with different amounts of antibodies was determined by integrating the curves from CDMS spectra of samples with a 3:1 initial ratio of Ab:capsid.

Percentage of population (%)

41

croscopy experiments to investigate their interaction with breast cancer cell lines. These samples were prepared by reacting cysteine residues introduced at position 87 with Oregon Green 488 (OG488) maleimide dyes. Mass spectrometry analysis shows that near-complete modifications can be obtained, resulting in agents with close to 180 copies of the small molecule attached (Figure 3.9a).

3.6.1. Flow cytometry experiments

13800 14000 14200 1440013600

13778 *

14241MS2-OG488

MS2

MS2paF/Cys

+Na++H2O

mass (Da)

13778

100 101 102 103 1040

20

40

60

80

100 untreated

MS2

MS2-αEGFR

MS2-PEG2kDa/5kDa-αEGFR

MS2-PEG5kDa-αEGFR

100 101 102 103 1040

20

40

60

80

100

MDA-MB-231 HCC1954

% o

f Max

Fluorescence Fluorescence

Fluorescence Fluorescence

% o

f Max

10 0 101 10 2 103 10 40

20

40

60

80

1000.1 µM

410 0 101 10 2 103 100

20

40

60

80

1001 µM

untreated

MDA-MB-231

MS2-PEG5kDa-αEGFR 3 equiv

MS2-PEG5kDa-αEGFR 20 equiv

MS2-αEGFR 20 equiv

MS2-αEGFR 3 equiv

[agent] (µM)

norm

aliz

ed m

edia

n flu

ores

cenc

e

1E-7 1E-6 1E-5 1E-4 1E-3 0.01 0.1-0.2

0.0

0.2

0.4

0.6

0.8

1.0

1.2

(a) (b)

(c) (d)(e)

Kd (nM)

0.410 ± 0.067

0.138 ± 0.030

0.199 ± 0.033

αEGFR

MS2-αEGFR

MS2-PEG5k-αEGFR

Figure 3.9 Flow cytometry binding studies. (a) LC-MS spectra of the MS2 T19paF N87C mutant and the modification of the interior surface Cys residues with Oregon Green 488 maleimide (OG488). (b) Flow cytometry analysis of the binding of MS2-αEGFR conjugates to EGFR positive cell lines MDA-MB-231 and HCC1954. Direct conjugation to the capsid surface, as well as the used of a PEG linker between the capsid and the antibody were investigated. A PEG2kDapolymer was used to “backfill” the surface of the capsid in order to cover the surface better and possibly lower the immune response to the conjugates. (c) Incubation of MDA-MB-231 cells with MS2-αEGFR conjugates leads to a dose-dependent response as observed by flow cytometry fluorescence shifts. The equivalents of antibodies used (3 versus 20) and the size of the linker (2 carbon atoms or PEG5kDa) do not have a large impact on the amount of agent bound. (d) Compari-son of binding affinities of unmodified αEGFR antibody, MS2-αEGFR, and MS2-PEG5k-αEGFR conjugates. Flow cytometry was used to measure the median fluorescence of MDA-MB-231 cell population after incubation with the samples. The data were fit to a single site binding model. The Kd values and the stan-dard error of each fit are listed in table (e). (f) MS2-antibody agents before and after incubation with mouse serum for 24 h at 37 ºC were added to MDA-MB-231 cells. The small differences between the “before” and “after” samples indicate that the serum proteins have little-to-no detrimental effects on the binding abilities of these agents.

Fluorescence

% o

f Max

10 0 10 1 10 2 10 3 10 40

20

40

60

80

100beforeserum

Fluorescence

% o

f Max

10 0 10 1 10 2 10 3 10 40

20

40

60

80

100afterserum

MDA-MB-231

MS2-PEG5kDa/5kDa-αEGFR

MS2-PEG2kDa/5kDa-αEGFR

MS2-PEG5kDa

MS2-αEGFR

MS2 + αEGFR mix

untreated

(f)

42

Incubation of fluorescently-labeled MS2-antibody conjugates samples with two breast cancer cell lines (HCC1954, MDA-MB-231) indicated that the targeted agents bind to the cells more than the untargeted capsids (Figure 3.9b). Additional experiments with cell lines not expressing the receptor of interest show that this interaction is specific (data not shown).

The addition of PEG linkers to extend the antibody further from the capsid surface did not seem to improve the binding to the targeted receptor (Figure 3.9b,c). Flow cytometry experiments (Figure 3.9c) did not reveal any clear advantage of having more than 3 cop-ies of the antibody per capsid, and thus for cost efficiency and ease of purification, conju-gates prepared from three equivalents of antibodies were used for all subsequent studies.

In order to investigate whether the conjugation of the antibody to the capsid affects its binding ability, a binding curve was obtained by incubating free anti-EGFR antibody, MS2-αEGFR and MS2-PEG5kDa-αEGFR with MDA-MB-231 cells at different concentra-tions (Figure 3.9d). Data was fit to a single site binding model, and Kd values were ob-tained for MS2-Ab agents. We observed similar Kd to free Ab (Figure 3.9e), indicating that the binding is unaffected by the reaction conditions or conjugation to MS2.

A wealth of literature shows that the physicochemical properties of nanoparticles change when they enter a living system.54 The nanoparticles usually interact with serum proteins and are coated with a “protein corona” consisting of ions, opsonins and other proteins. The identity of the protein corona will dictate the in vivo fate of the nanoparticles, such as clearance and tumor accumulation. In order to investigate ability of the MS2-agents to bind to their targets in the presence of the protein corona, we performed a flow cytometry experiment where we pre-incubated the agents with mouse serum, then allowed them to interact with cells expressing the receptor of interest. Pre-incubation of the agents in mouse serum did not inhibit their ability to bind the cells (Figure 3.9f), suggesting that serum proteins are not detrimental to the binding of the MS2-Ab conjugates. Ad-ditionally, we attempted to investigate the identity of the proteins that coat the MS2 capsid upon incubation in serum by per-forming a protein-A pull down experiment. Unfortunately, the amounts of protein ob-served on subsequent protein gel analysis were too low to draw any conclusions (Ap-pendix).

3.6.2 Live-cell confocal microscopy.

Specific binding of the antibody-tar-geted agents to the receptors of interest was also confirmed through fluorescent microscopy experiments. HCC1954 cells

MS2

MS2-αEGFRMS2-PEG5kDa-αEGFR

MS2-αEGFR

brightfield DAPI OG488 merged

Figure 3.10 Confocal microscopy studies of MS2-Ab binding to live HCC1954 cells. Images were taken after 1 h of incubation at 37 ºC with OG488-contain-ing MS2, MS2-αEGFR, or MS2-PEG5kDa-αEGFR (green channel). DAPI was used to stain the cell nuclei (blue channel). The scale bars represent 20 µm.

43

were incubated for 1 h at 37 ºC with OG488-labeled MS2, MS2-αEGFR and MS2-PEG5k-

Da-αEGFR in DPBS with 1% FBS. Fluorescent signal was observed on the cell surfaces and, at later time points, inside the cells, in several vesicles (Figure 3.10). These results indicate that, upon binding to the surface receptors, the constructs are internalized.

3.6.3 Mass cytometry analysis of binding ability.

Mass cytometry43, 55, 56 has emerged as a powerful tool in studying cell signaling events through receptor/protein profiling. This technology uses inductively coupled plasma time-of-flight (ICPTOF) mass spectrometry for the detection of isotopes, thus allowing as many as 100 parameters to be measured simultaneously (Figure 3.11a). This number rep-resents a dramatic increase over traditional fluorescence based flow cytometry, which is limited to ∼15 channels due to spectral overlap. The isotopes more commonly used are

Mass

Cell 3

N N

N N

HO2C

HO2C CO2H

CO2H

HNO

N

O

OLn3+

DOTA-GA maleimide

1. pH 4.5 acetate,

2. pH 7.5 bisTRIS, paF/N87C MS2

AP-Ab,NaIO4

13778MS2

14377MS2-DOTA-GA

14538MS2-DOTA-GA-Ho

13600 13800 14000 14200 14400 14600 14800 15000mass (Da)

up to 180 metal ions MS2-Ab (Ln)

MS2(Eu)

MS2-αCD20(Tb)

MS2-αEGFR(Ho)

MDA-MB-231 (EGFR+)

100 101 102 103 104 100 101 102 103 104

Signal intensity

Even

ts

Signal intensity

Even

ts

MCF7 Clone 18 (EGFR –)

ICP-MSElemental Analysis

Nebulize Single-Cell Droplets

Antibodieslabeled with

elemental isotopes

Cell 1

Cell 2

(a)

(b)

(c) (d)

Figure 3.11 Mass cytometry analysis of binding interactions of lanthanide-chelated agents with cells. (a) Diagram of the protocol for mass cytometry analysis. Adapted from reference 43. (b) Synthetic scheme summarizing synthesis of lanthanide-containing MS2-Ab conjugates. In situ chelation and conjugation of the lanthanide with DOTA-GA maleimide and subsequent oxidative coupling delivers the desired conjugate. (c) LC-MS characterization of the MS2-lanthanide constructs demonstrating that near quantitative conver-sion occurs. All spectra are reconstructed, and the values represent [M + H]+ ions. (d) Mass cytometry using MCF7 Clone 18 (EGFR and CD20 negative) and MDA-MB-231 (EGFR positive, CD20 negative) costained with a panel of agents. Graphs are plotted as channel overlays of histograms with an arcsinh of 45 applied to all data sets. High levels of staining were only observed for agents displaying EGFR antibodies with EGFR positive cells.

44

lanthanides, which occur in very small amounts in nature, therefore circumventing the problem of background signal from endogenous sources. The high signal-to-noise ratio provides excellent sensitivity over a large dynamic range.

To evaluate the interaction of MS2-based agents with mammalian cells in more depth, we generated a series of MS2-lanthanide mass cytometry staining reagents, including MS2(Ho)-αEGFR, MS2(Ho)-PEG5kDa-αEGFR, MS2(Eu) (untargeted control), and MS2(T-b)-αCD20 (negative control). These constructs were prepared by initial chelation of the lanthanide ion of interest with DOTA-GA-maleimide followed by in situ conjugation to the interior cysteines of the MS2 capsid (Figure 3.11b,c). The use of the DOTA-GA chelator proved optimal (as compared to DOTA-maleimide) in the development of this one-pot procedure due to the increased solubility of the lanthanide-chelated species in aqueous solutions. Subsequent conjugation of the antibodies provided the desired panel of agents. For these studies, we used the MDA-MB-231 cell line, which is known to overexpress the EGFR and not the CD20 receptor. The MCF7 Clone 18 cell line was used as the HER2- negative control. Cells were stained with a mixture of 3 agents that included ei-ther MS2(Ho)-αEGFR or MS2(Ho)-PEG5kDa-αEGFR as the targeted reporter in addition to MS2(Eu) and MS2(Tb)-αCD20 as controls. The simultaneous treatment with multiple agents dramatically reduces the number of samples to be examined and limits the vari-ability among them.

Consistent with our fluorescence-based flow cytometry results, we observed specific binding of the targeted agents to the cells expressing EGFR (Figure 3.11d). In addition, these reporters provided signal on par with the positive control antibody−polymeric che-lators when treated at similar concentrations (data not shown). Continuing studies are focused on providing substantial signal increases through encapsulation of lanthanide nanoparticles inside targeted MS2 capsids to detect single binding events within individ-ual cells in complex cell populations.

3.7 In vitro stability of MS2-antibody conjugates

Radiolabeled conjugates (obtained as described in Chapter 2) were analyzed for pu-rity and assembly state using SEC HPLC. The radioactivity traces indicated that most of the radioactive 64Cu isotope remained associated with the intact MS2-antibody conju-gates even upon incubation in mouse serum at physiological temperature (Figure 3.12).

64Cu-MS2 64Cu-MS2-αEGFR 64Cu-MS2-PEG5kDa-αEGFR 64Cu-MS2-PEG2kDa/5kDa-αEGFR

Figure 3.12 Stability of MS2-NOTA-64Cu antibody conjugates in mouse serum. The normalized signal from the radioactivity channel during SEC HPLC runs of aliquots taken at time points from 0 to 48 h indicates that the predominant species is the 64Cu-labeled intact MS2 capsid. The secondary peak (*) is thought to be due to partial disassembly or loss of 64Cu due to its interaction with other serum proteins.

Rad

ioac

tivity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0

Rad

ioac

tivity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0

Rad

ioac

tivity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0

Rad

ioac

tivity

Retention time (min)0 2 4 6 8 10

00.20.40.60.81.0

t=01 h

4 h24 h

48 h

t=01 h

4 h24 h

48 h

t=01 h

4 h24 h

48 h

t=01 h

4 h24 h

48 h

45

This high level of intact capsid is encouraging for drug delivery applications, for which it is important to allow enough time for specific accumulation in the region of interest before releasing the active molecule. By housing the cargo on the inside of the capsid, the side effects of systemic administration of chemotherapeutics could be mitigated.

3.8 In vivo distribution of MS2-antibody

3.8.1 Positron Emission Tomography

As mentioned in Chapter 2, several molecular imaging approaches have been used for tracking the distribution of agents with therapeutic potential, including positron emission tomography (PET), single photon emission computed tomography (SPECT), magnetic resonance imaging (MRI), computed tomography (CT), ultrasound, bioluminescence, and fluorescence imaging. We used PET as a noninvasive way to gain information about the biodistribution of the MS2 conjugates in real time and at different time points while limiting the number of animals used.

Tumor-bearing nude mice in sets of 3 animals per study group were injected in the tail vein (intravenous, IV) with 150−250 μCi of 64Cu-labeled MS2 conjugates in 100 μL of sterile saline. One animal from each group was selected for imaging with PET. Dynamic

(b)

(a) αEGFR MS2 MS2-PEG2kDa

tumorintestines

liver

heart

bladder

tumor

MS2-PEG2kDa/5kDa-αEGFR

0.5

1.5

0.6

3.0

% ID

/g%

ID/g

tumortumor

tumor

tumor tumor tumor

tumor

Figure 3.13 Positron Emission Tomography (PET) with radiolabeled MS2 agents in a mouse model. In vivo distribution of 64Cu-labeled MS2 scaffolds in nude mice with HCC1954 orthotopic breast cancer tumors (mammary fat pad). Mice were injected with 150−250 μCi radiolabeled agents. 64Cu-labeled antibodies were used as positive controls. (a) PET scans were acquired at 24 h post-injection. A maximum intensity projection (MIP) of the PET signal within a 50 mm slice is presented overlaid with a CT slice of 1 mm from the coronal plane. The horizontal blue dashed line represents the slice used for the transverse plane image in panel b. (b) A transverse plane through the tumor is shown to illustrate the agent distribution within the tumor.

46

imaging was performed from the time of injection until 1 h post-injection, and a 20 min static scan was collected at 24 h post-injection (Figure 3.13). The dynamic PET images taken from time of injection to 1 h post-injection were used to determine the clearance profile of the injected agents. A 3D region of interest (ROI) was chosen to overlap with the left ventricle of the heart to estimate the radioactivity present in the blood pool. The signal showed a decay profile corresponding to a two compartment model, as evidenced by the shape of the of the log10 plot of the signal within the ROI against the midpoint of each scanning frame (Appendix).

3.8.2 Biodistribution studies

The blood and major organs were harvested, and their activity was measured using a gamma counter. The tissue samples were weighed, and the radioactivity signal was nor-malized as percent injected dose per gram of tissue (% ID/g). Overall tumor uptake for the MS2-based agents varied between 2 and 5 % ID/g at 24 h post-injection (Figure 3.14).

0

5

10

15

20

25

30%

ID/g

HCC1954 SC

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or0

5

10

15

20

25

30

% ID

/g

HCC1954 MFP

MS2

MS2-PEG2kDa/5kDa-αEGFRMS2-PEG2kDa

αEGFRMS2

MS2-PEG2kDa/5kDa-αEGFRMS2-PEG2kDa

αEGFR

Agent

14.25 ± 5.713.70 ± 1.044.18 ± 1.094.62 ± 1.26

23.83 ± 12.122.55 ± 1.633.52 ± 1.794.91 ± 2.72

αEGFRMS2MS2-PEG2kDa MS2-PEG2kDa/5kDa-αEGFR

T/M (MFP) T/M (SC)

Table 3.2 The tumor-to-muscle ratio (T/M) of the normalized radioactive signal per unit weight (% ID/g) was calculated as a measure of non-specific accumu-lation of the agents (T/M = 1 indicates no specificity for the tumor over the muscle).

0

5

10

15

20

25

30

35 αEGFR MS2 2x MS2 10x MS2-αEGFR 2x MS2-αEGFR 10x

% ID

/g

L3.6pl (EGFR+) SC

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or0

5

10

15

20

25

30

35

% ID

/g

αHER2 MS2 1x MS2 10x MS2-αHER2 1x MS2-αHER2 10x

MCF7Cl18 (HER2+) MFP

Figure 3.14 Biodistribution studies of 64Cu-labeled MS2 scaffolds: untargeted, HER2-targeted and EGFR-targeted in nude mice in mammary fat pad (MFP) or subcutaneous (SC) tumor models. 64Cu-labeled antibodies were used as positive controls. At 24 h post-injection, organs were harvested, counted on a gamma counter and weighed. The 1x, 2x, 10x notation refers to a multiple of a standard dose of 2 nmoles MS2 monomer.

47

The tissues with the largest degree of accumulation were the liver, the spleen (the usual clearance pathways for nanoparticles, as discussed in Chapter 2) and the large intestine. The high signals in the blood even after 24 h of circulation validate the long half-lives ob-served from the analysis of the dynamic PET scans.

To investigate the non-specific tissue uptake of the agents, we calculated the tu-mor-to-muscle ratios by dividing the % ID/g in the tumor to the signal from the muscle tissue (Table 3.2). The numbers greater than 1 indicate specific uptake of the agents into the tumor. The free anti-EGFR antibody used as positive control had the highest tu-mor-to-muscle ratio. Notably, the amount of antibody injected was 3-4 times more than the equivalent amount of antibodies attached to the MS2 capsids. The MS2 agents had tumor-to-muscle ratios in the range of 2.5-4.9, indicating specific uptake in the tumor compared to muscle tissue.

At 24 h post-injection, a similar amount of activity is present in the tumor for both the targeted and the untargeted viral capsids, suggesting that the extrav-asation from the blood vessel to the tu-mor might be the limiting step. Some recent studies have shown that a spherical structure is not ideal for crossing the gaps in the endothelial cell layer of tumor blood vessels.57, 58 Our results also indicate that targeting does not provide a significant improvement in the efficiency of the agent to accumulate in the region of interest, which has also been noted in the case of gold nanoparticles of similar size.2, 59 Similar results were observed in the case of liposomes,60 although further studies investigating the delivery of drug molecules indicated a clear benefit to the attachment of the targeting moiety, presumably due to in-creased uptake once the cancer cells were reached.61

3.9 In vivo assessment of immunogenicity of MS2-Ab constructs

Balb/c mice with intact immune systems were administered MS2-based agents through tail vein injection in two doses, at t=0 and 3 weeks. The agents tested were: MS2, MS2 coated with ~130 strands of methoxy-terminated PEG2kDa and MS2 decorated with both antibodies with PEG5kDa spacers and PEG2kDa backfilling. To facilitate the ability to distinguish the immune response caused by the capsids from the immunogenicity of the humanized anti-EGFR antibody, a mouse IgG antibody was used for this study.

After the first dose, the amount of total IgM and IgG produced by animal groups treat-ed with the three agents was similar (0.5-2 mg/mL). Upon injection of the second dose, the amount of IgG produced was 3-4 times higher than the IgM (Figure 3.15). Brown et al. observed a similar ratio of IgG to IgM antibodies, indicating that upon multiple admin-istrations of MS2 (albeit subcutaneous in their case), the bulk of the response was in the form of IgG.31

Anti-MS2 coat protein antibodies were present in the serum after the initial dosing and

0

5

10

15

20

25

30

% ID

/g

HCC1954 SC

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or0

5

10

15

20

25

30

% ID

/g

HCC1954 MFP

MS2

MS2-PEG2kDa/5kDa-αEGFRMS2-PEG2kDa

αEGFRMS2

MS2-PEG2kDa/5kDa-αEGFRMS2-PEG2kDa

αEGFR

Agent

14.25 ± 5.713.70 ± 1.044.18 ± 1.094.62 ± 1.26

23.83 ± 12.122.55 ± 1.633.52 ± 1.794.91 ± 2.72

αEGFRMS2MS2-PEG2kDa MS2-PEG2kDa/5kDa-αEGFR

T/M (MFP) T/M (SC)

Table 3.2 The tumor-to-muscle ratio (T/M) of the normalized radioactive signal per unit weight (% ID/g) was calculated as a measure of non-specific accumu-lation of the agents (T/M = 1 indicates no specificity for the tumor over the muscle).

0

5

10

15

20

25

30

35 αEGFR MS2 2x MS2 10x MS2-αEGFR 2x MS2-αEGFR 10x

% ID

/g

L3.6pl (EGFR+) SC

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or

blood bo

nebra

in fathe

art

kidne

y

lg int

estin

eliv

erlun

g

muscle

panc

reas

sm in

testin

e

splee

n

stomac

htai

ltum

or0

5

10

15

20

25

30

35

% ID

/g

αHER2 MS2 1x MS2 10x MS2-αHER2 1x MS2-αHER2 10x

MCF7Cl18 (HER2+) MFP

Figure 3.14 Biodistribution studies of 64Cu-labeled MS2 scaffolds: untargeted, HER2-targeted and EGFR-targeted in nude mice in mammary fat pad (MFP) or subcutaneous (SC) tumor models. 64Cu-labeled antibodies were used as positive controls. At 24 h post-injection, organs were harvested, counted on a gamma counter and weighed. The 1x, 2x, 10x notation refers to a multiple of a standard dose of 2 nmoles MS2 monomer.

48

a small increase was observed after the second dose (Figure 3.15). Using a two sample unpaired Student t-test, we determined that the amounts of antibodies produced upon injection of the MS2 agents were significantly higher (p < 0.05) than the PBS control. The addition of PEG chains did not seem to have a significant effect on the amount of antibody produced. This phenomenon might be explained by the fact that even if there are about 130 strands of PEG per capsid, they might not reach the “brush regime” and could be less efficient at shielding the viral capsid from the immune system. Mastico et al.62 have also observed the production of anti-MS2 antibodies in mice, which was amplified by the addi-tion of a nonapeptide derived from haemagglutinin. Although there is a paucity of studies on the immune response caused by IV administration of viral capsid as a delivery agent (and not as a vaccine), a study by Kaiser et al. noted that antibody production against CCMV and Hsp was also observed.63

3.10 Efforts to remove endotoxins

We hypothesized that some of the immunogenicity of the MS2 agents could be due to the presence of endotoxins in the samples. Endotoxins consist of fragments of bac-terial cell wall that can be shed during the recombinant expression process and are not removed during the subsequent purification steps. Endotoxins can produce a very strong immune response in mammals, even leading to septic shock.64

Initial measurements of endotoxin levels using a commercial assay indicated very high levels of endotoxins present in the samples. The efforts to remove these contamina-tions included using endotoxin removal resins, anion exchange resins, extractions with detergents (Triton X 114) and syringe filters.65-67 These methods were able to reduce the endotoxin level 10-1000 times, but the protein recovery was very low, limiting the feasibil-ity of introducing them into the MS2 purification protocol.

3.11 Conclusions

In vitro results indicated specific binding of the MS2-Ab targeted agents to the re-ceptors of interest expressed on the surface of cancer cell models. The agents can be

0

0

0.4

0.8

1.2

1.6

2

IgM

(mg/

mL)

days post-injection0 14 28 42 56

0

1

2

3

4

5

6

7

IgG

(mg/

mL)

days post-injection

anti

MS2

(log

10

titer

)

days post-injection

IgG Titer IgM Titer

14 28 42 56 0

0

1

2

3

4

5

6

7

anti-MS2 Titer

14 28 42 56

MS2-PEG2kDaMS2-PEG2kDa/5kDa-IgGInjection

MS2PBS

Figure 3.15 Immunogenicity assay. The amount of IgG, IgM and anti-MS2 coat protein antibodies in the serum of Balb/c mice was measured at different time points post-injection using ELISA assays. Although the amount of antibodies produced for all of the MS2 samples was statistically higher than the PBS negative control (black line), the differences between the agents were not significant (unpaired t-test with equal variance and a two-tailed p value, p > 0.05).

49

radiolabeled to obtain high specific-activity constructs that maintain structural integrity over 48 h when incubated in mouse serum at 37 °C. Upon intravenous injection in mouse breast cancer models, the MS2 viral capsid conjugates show moderate tumor uptake, comparable to targeted inorganic nanoparticles. No significant difference in tumor uptake is observed between the antibody-targeted and untargeted capsids.

The work presented herein is one of the few examples of antibody-targeted pro-tein-based nanoparticles to be investigated for tumor homing. The particles exhibit sur-prisingly long circulation profiles and accumulate within the tumor environment. However, no significant difference is observed between the targeted and untargeted viral capsids, indicating that the diffusion through the endothelial gaps might be the limiting factor in determining tumor accumulation of nanoparticles. Our observations add evidence to the case that other factors, such as size, shape and overall charge, might play a bigger role in the homing of nanoparticles to the tumor, while the targeting groups can increase spec-ificity to tumor cells within the tumor environment. This hypothesis can help guide the design of nanoparticle carriers with various physical properties, while the synthetic plan and battery of analyses performed in this work can provide a framework for the thorough characterization of new scaffolds for drug delivery.

This work suggested that it might be beneficial to leverage the long circulation time of the MS2-based agents for cardiovascular disease imaging applications by targeting markers accessible to the circulatory system. Given the potential of the MS2 VLPs to car-ry up to 180 copies of drug molecule cargo, the large amount of agent accumulated in the tumor could provide a significant therapeutic effect. As such, the efficiency of MS2-anti-body conjugates for delivering drug cargo and inhibiting tumor growth will be investigated.

3.12 Materials and methods

Unless otherwise noted, all chemicals and solvents were of analytical grade and used as received from commercial sources. Anti-EGFR human IgG1 monoclonal antibody was obtained from Eureka Therapeutics, Inc. (Emeryville, CA). All polyethylene glycol (PEG) derivatives were purchased from Laysan Bio, Inc. (Arab, AL). Trimethylamine N-oxide (TMAO) and yeast tRNA were purchased from Sigma (St. Louis, MO). All cell culture reagents, including normal mouse serum were obtained from Gibco/Thermo Fisher (Waltham, MA) unless otherwise noted. Maleimide-NOTA (2,2’-(7-(2-((2-(2,5-dioxo-2,5-di-hydro-1H-pyrrol-1-yl)ethyl)amino)-2-oxoethyl)-1,4,7-triazonane-1,4-diyl)diacetic acid was purchased from CheMatech (Dijon, France). Ethylenediaminetetraacetic acid (EDTA) di-sodium salt and potassium phosphate dibasic were purchased from EMD Chemicals Inc. (Darmstadt, Germany). Copper-64 was purchased from the Medical Cyclotron Laboratory, University of Wisconsin. Saline (sodium chloride) solution was Injection USP, 0.9% from APP Pharmaceuticals (Schaumburg, IL). NAP desalting columns were purchased from GE Healthcare (Marlborough, MA). Spin concentrators with different molecular weight cutoffs (MWCO) were from Millipore (Billerica, MA).

Liquid chromatography mass spectrometry (LC-MS) analysis, high performance liquid chromatography (HPLC) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) were performed as described in Chapter 2. The protocol for the purification

50

of MS2 T19pAF/N87C was described in Chapter 2.

Dynamic Light Scattering (DLS). DLS measurements were obtained using a Malvern Instruments Zetasizer Nano ZS (Malvern Instruments Ltd., Worcestershire, UK). MS2 samples were prepared at a concentration of 20 µM MS2 monomer in 10 mM pH 7.2 phosphate buffer and filtered through a 0.22 µm centrifugal filter unit (Millipore) prior to data collection. Data plots are shown as size distribution by number, which weighs large and small particles equally. Diameters were calculated from an average of three mea-surements.

Transmission Electron Microscopy (TEM). TEM images were taken at the U.C. Berke-ley Electron Microscope Laboratory using a FEI Tecnai 12 transmission electron micro-scope (FEI, Hillsboro, OR) with 120 keV accelerating voltage. Samples were prepared by pipetting 5 µL onto Formvar-coated copper mesh grids (400 mesh, Ted Pella, Redding, CA) for 5 min, followed by exposure to 8 µL of a solution of uranyl acetate (15 mg/mL in dd-H2O) for 2 min as a negative stain. Excess stain was then removed and the grids were allowed to dry in air for 10 min.

Synthesis of aminophenol-antibody conjugates. To an antibody solution in 25 mM phosphate buffer pH 8, NHS-NP was added (5 equiv) from a concentrated stock solution in DMSO. The mixture was incubated at room temperature for 1 h 15 min, after which the nitrophenol moiety was reduced to aminophenol by addition of sodium dithionite for 20 min at room temperature. The reaction was stopped by using 0.5 mL spin concentrators with a MWCO of 30 kDa and washing with 10 mM pH 7.2 phosphate buffer. The amin-ophenol-antibody (AP-Ab) was used immediately in subsequent conjugation reactions.

Generation of MS2-antibody conjugates. A solution of MS2 was added to the above solution of AP-Ab conjugates at a 3:1 ratio of antibody to capsid. In general, the final concentration of the capsid was ~200-300 nM (corresponding to ~50 µM MS2 monomer), and the concentration of Ab was adjusted accordingly. To initiate the oxidative coupling, NaIO4 was added to a final concentration of 500 µM, and the reaction was performed at room temperature for 5-8 min. Excess NaIO4 was removed by using NAP-5 Sephadex size exclusion columns equilibrated with 10 mM pH 7.0 phosphate buffer, by sequential spin concentration using a MWCO of 100 kDa.

Synthesis of nitrophenol-PEG5kDa-COOH. Amine-PEG5kDa-pentanoic acid (NH2-PEG5k-

Da-COOH) (1 equiv.) was dissolved in DCM and NP-NHS ester (5 equiv.) and triethylamine (10 equiv.) were added sequentially. The reaction mixture was stirred at room tempera-ture overnight. The solvent was removed under high vacuum and excess NP-NHS was precipitated by the addition of 500 µL of water. The resulting precipitate was removed by filtration through a 0.45 µm spin filter (PALL, Port Washington, NY) and the filtrate was concentrated and purified using a 3 kDa MWCO spin concentrator.

Synthesis of nitrophenol-PEG2kDa-OMe. Amine-PEG5kDa (NH2-PEG2kDa-OMe) (1 equiv.) was dissolved in DCM and NP-NHS ester (5 equiv.) and triethylamine (10 equiv.) were added sequentially. The reaction mixture was stirred at room temperature overnight. The solvent was removed under high vacuum and excess NP-NHS was precipitated by the

51

addition of 500 µL of water. The resulting precipitate was removed by filtration through a 0.45 µm spin filter (PALL). The filtrate was reduced with 10 equiv. Na2S2O4 at room tem-perature to give the aminophenol conjugate (AP-PEG2kDa-OMe). Excess dithionite was removed by purification with three Sephadex size exclusion columns. The purified amin-ophenol PEG was lyophilized and resuspended in 10 mM phosphate buffer, pH 7.2. The concentration was adjusted to 1 mM by measuring the aminophenol absorbance at 290 nm. The final solution of o-aminophenol PEG was stored in single-use aliquots at -20 °C until use.

Synthesis of aminophenol-PEG5kDa-antibody conjugates. Nitrophenol-PEG5kDa-NHS ester was pre-formed by mixing NHS, EDC, and NP-PEG5kDa-COOH in water in a 1:1:1 ratio and incubating at room temperature for 30 min. The NP-PEG5kDa-NHS ester solution was added to a 1 mg/mL solution of antibody (~6.6 µM in PBS). The reaction was allowed to proceed at room temperature for 3 h. The NP-PEG5kDa-Ab conjugates were reduced to aminophenol (AP) to yield AP-PEG5kDa-Ab conjugates by adding sodium dithionite in 100 mM pH 6.5 phosphate buffer (final concentration 10 mM) for 15 min at room temperature. All the excess small molecules in the reaction mixture were removed by using 0.5 mL spin concentrators with a MWCO of 30 kDa and washing with 10 mM pH 7.2 phosphate buffer.

Generation of MS2-PEG5kDa-antibody conjugates. A solution of MS2 was added to the above solution of AP-PEG5kDa-Ab conjugates at a 3:1 ratio of antibody to capsid. In general, the final concentration of the capsid was ~200-300 nM (corresponding to ~50 µM MS2 monomer), and the concentration of AP-PEG5kDa-Ab was adjusted accordingly. To initiate the oxidative coupling, NaIO4 was added to a final concentration of 500 µM, and the reaction was performed at room temperature for 5-8 min. For the “backfilling” experiments, 5 equiv. of AP-PEG2kDa-OMe were added and fresh NaIO4 was added to a final concentration of 1 mM. Excess NaIO4 was removed by using NAP-5 Sephadex size exclusion columns equilibrated with 10 mM pH 7.2 phosphate buffer, by sequential spin concentration using a MWCO of 100 kDa.

Alkaline treatment. A solution of MS2 was diluted to a concentration of 1 mg/mL (~72 μM monomer concentration) with 10 mM potassium phosphate buffer to a final volume 4V. A solution of 500 mM sodium phosphate at pH 11.8 in a 1:4 volume ratio to the MS2 solution was added (V) and the mixture was incubated at room temperature for 2 h. An equal volume as the total reaction mixture (5V) of saturated ammonium sulfate solution was added and the mixture was incubated at 4 °C for 30 min. The protein was then precip-itated by spinning at 15,000g for 10 min at 4 °C. The resulting precipitate was resuspend-ed in a volume of 5V and the steps presented above were repeated twice. After the final resuspension, the concentration of the protein was determined by measuring the A280 on a NanoDrop spectrometer and using an extinction coefficient of ε= 18450 M-1cm-1.

Hybrid treatment. A 10 mg/mL solution of MS2 coat protein (~720 μM monomer con-centration) in ST buffer was mixed 2:1 vol/vol with glacial acetic acid pre-chilled on ice. The reaction mixture was incubated on ice for 30 minutes and white precipitate was ob-served. The solution was spun down using a microcentrifuge at 14,000 g for 20 min at 4 °C. The supernatant was loaded onto a NAP-25 desalting column equilibrated with chilled

52

1 mM acetic acid. The protein was eluted with 1 mM acetic acid and the fractions contain-ing protein (by A280 absorbance) were pooled. TMAO was added to a final concentration of 0.25 M and yeast tRNA was added to a final concentration of 1 mg/mL. The volume of the reaction mixture was adjusted using ST buffer such that the final MS2 monomer concentration was 15 μM. The reassembly was allowed to proceed at 4 °C without stir-ring for 48 h. The solution was spin filtered through a 0.22 μm centrifugal filter unit then concentrated using a spin concentrator with a MWCO of 100 kDa. The final concentration was determined by measuring the A280 absorbance on a NanoDrop spectrometer and using an extinction coefficient of ε= 18450 M-1cm-1. For complete removal of the genetic material, subsequent alkaline treatment (as described above) was performed.

Collection of mass spectra using CDMS. Prior to CDMS analysis, samples were buffer exchanged into 100 mM ammonium acetate (Sigma, St. Louis, MO) by size exclusion col-umns (100 kDa MWCO). Samples were prepared at a final concentration of about 1 mg/mL protein (72 μM MS2 monomer). Electrospray ionization was used to introduce ions into the gas phase. A single ion was captured in the electrostatic cone trap for 95 ms.

CDMS antibody fitting procedure. The spectrum for a unmodified capsid (no antibody) was measured and the peak was fit using a series of Gaussians in Origin2015 (OriginLab, Northampton, MA). The parameters of the fit, such as the peak center and width, were used to create a custom peak fitting function in Origin. The custom fitting function fit the experimental data with an R-squared value of 0.99953, essentially a perfect fit to the ex-perimental data. This custom peak fitting function was then used to analyze the spectra from the 3 and 10 antibody equiv. samples. The reasoning for using the same parameters from the bare spectrum to analyze the antibody spectra was that since the antibody con-jugates are produced from the same bare MS2 capsids they will have the same intrinsic heterogeneity, and hence the same peak shape. We believe this is a valid assumption based on TEM images that show that the antibody-capsid conjugates have similar shape and morphology compared to the bare capsids. We imposed the parameters found from the bare MS2 spectrum on the peaks from the MS2-antibody spectra and enforced the peak center to be at integer number of additions of Ab mass to the bare capsid mass (MWMS2 + n x MWAb). We then used a multiple peak fitting function to determine the distri-bution of the number of antibodies on each capsid by allowing the adjustment of the area under the curve for each peak that was assigned to the spectrum.

Conjugation of Oregon Green 488 to the interior of T19paF N87C MS2. The cysteine residues on the interior of MS2 were alkylated with Oregon Green 488-maleimide (Ther-mo Fisher) by adding 10 equiv. dye (from a 100 mM stock solution in DMF) to a solution of T19paF N87C MS2 (final monomer concentration of 100 µM) in 10 mM pH 7.2 phosphate buffer. The reaction mixture was incubated at room temperature for 4 h. Upon completion, the excess dye was removed using NAP-10 size exclusion columns equilibrated with 10 mM pH 7.2 phosphate buffer and concentrated with 100 kDa MWCO spin concentrators. LCMS and absorbance measurements confirmed that complete modification of the inte-rior cysteines was achieved.

Cell culture was conducted using standard techniques, using culture-treated flasks

53

(Corning, Tewksbury, MA). HCC1954 cells (ATCC, Manassas, VA) were grown in Roswell Park Memorial Institute (RPMI) Medium 1640 supplemented with 10% (v/v) fetal bovine serum (FBS, Omega Scientific, Tarzana, CA) and 1% penicillin/streptomycin (P/S, Sig-ma). MDA-MB-231 (ATCC) and MCF7 Clone 18 (Preclinical Therapeutics Core Facility, UCSF) cells were grown in Dulbecco’s Modified Eagle Medium (DMEM) supplemented with 10% (v/v) FBS and 1% P/S. All cell lines were grown at 37 °C in 5% CO2 atmosphere.

Flow cytometry analysis with MS2-Ab conjugates. The cell lines of interest (HCC1954 and MDA-MB-231) were washed with Dulbecco’s phosphate buffered saline (DPBS) and trypsinized at 37 °C for 5 min, followed by the addition of the appropriate complete media to stop trypsinization. Cells were then pelleted and resuspended in binding buffer (DPBS with 1% FBS) to the density of 3 x 106 cells/mL. Aliquots of 100 µL containing 3 x 105 cells were incubated with MS2-Ab-OG488 at 1 µM final concentration of MS2 monomer (5.5 nM capsid) for 1 h. The cells were washed twice with 500 µL binding buffer, resuspended in 200 µL of binding buffer, and analyzed by flow cytometry. For each sample, 10,000 cells were counted. A FACSCalibur flow cytometer (BD Biosciences, San Jose, CA) equipped with 488 nm laser was used for all flow cytometry measurements, usage courtesy of Prof. Carolyn Bertozzi (U.C. Berkeley). Data was analyzed using FlowJo 8.0 (TreeStar, Ash-land, OR).

Confocal microscopy with MS2-Ab conjugates. HCC1954 cells were washed with DPBS and trypsinized, then the trypsin was quenched with complete growth media, as described above. Cells were then resuspended in the growth media at a concentration of 25,000 cells/mL, and 2 mL of the cell suspension were added to each 35 mm glass bottom dish (MatTek Corp., Ashland, MA). Cells were allowed to grow at 37 °C with 5% CO2 for 48-72 h. Media was removed from the dishes, and the cells were washed once with 1 mL DPBS, then samples were added to each well at a final concentration of 1 µM MS2 monomer (~ 5 nM capsid) in 150 µL of binding buffer. The dishes were incubated at 37 °C with 5% CO2 for 1 h, then the cells were washed three times with DPBS (1 mL each), and 1 mL of phenol red-free media with 10% FBS was added to the cells. 4’,6-di-amidino-2-phenylindole (DAPI, Thermo) was added to 1 µM final concentration prior to acquisition of the final images. Images were acquired on a Zeiss 510 NLO Axiovert 200M Tsunami microscope equipped with a 488 nm laser, usage courtesy of Prof. Christopher Chang (U.C. Berkeley).

Mass cytometry analysis. Adherent cell lines (HCC1954, MCF7 Clone 18 and MDA-MB-231) were trypsinized at 37 °C for 5 min, followed by the addition of complete media (RPMI + 10% FBS in the case of HCC1954, or DMEM + 10% FBS for all the other cell lines) to stop trypsinization. Cells were then pelleted and resuspended in binding buffer (DPBS + 1% FBS) to the density of 107 cells/mL. Aliquots of 100 μL containing 106 cells were incubated with MS2-Ab conjugates containing different lanthanide metal ions at a final concentrations of 1 μM monomer for 1 h on ice. Free antibodies (anti-CD20 and anti-EGFR) were conjugated to polymeric chelators and lanthanide ions according to the manufacturer instructions (MaxPar antibody labeling kit, Fluidigm, San Francisco, CA). The cells were washed once by adding 3 mL binding buffer, spinning and resuspending in 1 mL of fixing solution (1.6% paraformaldehyde, 0.02% saponin, 1:5000 Ir DNA inter-

54

calator in PBS). Upon fixation for 20 min at room temperature, the cells were washed once with 2 mL of binding buffer and twice with 2 mL PBS. The cells were then injected onto a CyTOF or CyTOF2 instrument (Fluidigm), and 20,000 events were recorded. The cell-length versus DNA graph was used for gating whole cells, and histograms for the dif-ferent lanthanide channels were obtained using CytoBank software (CytoBank Inc., www.cytobank.org).

Radiolabeling of MS2 conjugates. To a 100 µM sample of protein (based on capsid monomer) in 10 mM potassium phosphate buffer, pH 7.2, was added 10 equiv. of maleim-ide-chelator from a 10 mM stock in DMSO. The reaction was allowed to proceed for 4 h at room temperature and was purified using a NAP-10 Sephadex size exclusion column equilibrated with phosphate buffer, pH 6.5 and 100 kDa MWCO spin concentrators. For the Ab-NOTA samples, the antibodies were first reacted with 200 equiv. pSCN-Bn-NOTA at 37 °C for 24 h. The reaction was stopped by using a NAP-5 desalting column and the purified antibody was concentrated to 10-20 µM using a 30 kDa spin concentrator.

The 64Cu copper stock (20-30 mCi, ~200 µL) was diluted with 1 mL of 0.1 M ammoni-um citrate buffer, pH 6.2 to generate a final volume of ~1200 µL at pH 5.5 (determined by pH paper). Each reaction tube was then charged with 100-400 µL of diluted 64Cu solution and 50-150 µL of the MS2 conjugates (200 µM in capsid monomer). The complexation reactions were allowed to proceed for 2 h at room temperature, then purified using NAP-5 or NAP-10 columns. Samples were subsequently concentrated using 100 kDa MWCO spin concentrators. Centrifugation was performed at 10,000 rpm for 5 min per round of concentrating until the desired volume was reached.

In vitro stability studies. Concentrated 64Cu-chelator-MS2 conjugates (~300 µCi in 20 µL) were added to PBS or 100% mouse serum to a final volume of 200 µL (1:9 v/v). The samples were then incubated at 37 °C for a predetermined time using a temperature-con-trolled heat block (VWR, Radnor, PA). Aliquots were drawn from the samples at 1, 4, 8, 24 and 48 h time points, and injected onto a PolySep GFC-P5000 size exclusion chroma-tography column, as described in Chapter 2.

PET/CT imaging. All animal procedures were performed according to a protocol ap-proved by the UCSF Institutional Animal Care and Use Committee (IACUC). Six-week old female nu/nu (nude) mice weighing 18-23 g were purchased from Charles River Lab-oratories (Hollister, CA). For tumor inoculation, the cells were implanted in the number 4 mammary fat pad (β-estradiol pellets were also implanted subcutaneously in the flank for MCF7 Clone 18 cells) or in the right flank for subcutaneous models. The imaging and biodistribution experiments were started approximately two weeks following implantation, when the tumors were ~1 cm in diameter.

Tumor-bearing nude mice in sets of 3 animals per study group were injected in the tail vein (intravenous, IV) with 150-250 µCi (5.5-9.25 MBq) of 64Cu-labeled MS2 conjugates in 100 µL of sterile saline. One animal from each group was selected for dynamic imaging from the time of injection to 1 h post injection, followed by a 20 min static scan at 24 h, as described in Chapter 2.

55

Biodistribution studies. Mice were sacrificed at 24 h post-injection and the blood, tu-mor, and major organs were harvested and weighed. The percent injected dose per gram of tissue (%ID/g) was determined as described in Chapter 2.

Immunogenicity assays. In order to investigate the immune response that the capsids might trigger in a living system, we contracted an outside company, Pacific Bio Labs, to test the amount of antibody produced upon tail vein injections of MS2 conjugates. Mice with intact immune systems (female Balb/c, 3 mice per group) were injected intravenously (IV) with 100 µL of agent in sterile saline at t = 0 and with a second dose at 3 weeks. Blood samples were obtained before the first injection and at 2, 4, and 8 weeks. The amount of antibody present was determined using commercially available ELISA kits for IgM and IgG quantification.

Titer measurements for IgG/IgM. Mouse serum samples were diluted 1:100, then se-rially diluted from 100 to 10,000-fold for IgM determination or by 5 to 50,000-fold dilution for IgG determination. All samples were assayed using the mouse IgM or IgM ELISA kits, respectively, according to the instructions from the manufacturer (Bethyl Labs, Montgom-ery, TX). The absorbance at 450 nm was read and corrected for the blank, and the anti-body titer was determined.

Anti-MS2 antibody measurement. High binding 96-well enzyme immunoassay plates (Corning Inc., Corning, NY) were coated with 50 µL/well of MS2 at 1 µg/mL in 1x DPBS overnight at 2-8 ˚C. Blocking buffer (Meso Scale Discovery, Rockville, MD) containing 5% goat serum was then added (150 µL/well) and the plates were incubated for 2 h at room temperature with shaking at 300 rpm. Blocking buffer with 0.05% w/v Tween 20 (Sigma) was used to wash the wells upon incubation (washing buffer). Each sample was diluted 1:200, followed by three-fold serial dilution. The plate was incubated 1.5 h at room tem-perature at 300 rpm, then each well was washed 3 times. A solution of biotin-F(ab’)2 goat anti-mouse IgG Fc gamma (Jackson ImmunoResearch, West Grove, PA) in washing buf-fer (final concentration 0.25 μg/mL) was added to each well and the plate was incubated for 1 h at room temperature with shaking at 300 rpm. After washing the wells 3 times, 50 µL of streptavidin-Horse Radish Peroxidase solution (1 mg/mL, Jackson ImmunoRe-search, final concentration 6.25 ng/mL, diluted in washing buffer) was added to each well, incubated for 1 h at room temperature with shaking at 300 rpm. A solution of 3,3’,5,5’-te-tramethylbenzidine (TMB, BioRad) substrate (100 µL of 1:9 solution A : solution B) was added and the plate was incubated for 30 min at room temperature in the dark, without shaking. The reaction was stopped by adding 100 µL of 2 N sulfuric acid solution to each well. The absorbance was then read at 450 nm and the titer of anti-MS2 antibodies was determined.

Endotoxin removal efforts. Endotoxin removal resin (Thermo) was used according to manufacturer’s instructions. Potassium phosphate buffer at pH 7.2 was prepared using endotoxin-free water (Sigma) and used to pre-equilibrate the resin. The samples were incubated for 1 h at 4 °C, then spun out and stored in pyrogen-free conical tubes until the time of measurement. The LAL endotoxin quantitation kit was used as indicated by the supplier. Serial dilutions were made using endotoxin-free water. The absorbance was

56

measured at 405 nm using an Infinite 200 Pro (Tecan, Switzerland) plate reader and the absorbance of the blank sample was subtracted from each well.

Clear coliTM BL21(de3) electrocompetent cells (Lucigen, Middleton, WI) were used to express the MS2 T19pAF and MS2 T19pAF N87C mutants. Initial attempts to transform the plasmid encoding for the MS2 coat protein gene and the tRNA synthetase at the same time were unsuccessful. A stepwise approach was attempted next. In the first step, 25 μL electrocompetent cells were electroporated in the presence of 1 μL pBAD AmpR encoding for the MS2 monomer. The cells were rescued with 950 μL Expression Recovery Media (Lucigen) and incubated at 37 °C with shaking for 2 h. The cells were plated on LB-agar plates containing 50 μg/ml carbenicillin (same antibiotic resistance gene as ampicillin) and incubated at 37 °C overnight. Colonies were selected and inoculated in LB with 50 μg/ml carbenicillin. After 16 h incubation at 37 °C, the cultures were diluted to 1 L LB with 50 μg/ml carbenicillin final concentration. After 12 h, the cells were pelleted at 3000 rpm for 20 min at 4 °C. The pellet was resuspended and washed with 200 mL water twice. The pellet was resuspended in 40 mL 10% glycerol and spun at 1000 g for 20 min at 4 °C. The cells were resuspended in 1 mL 10% glycerol, aliquotted and stored at -80 °C until used in the next step. Cells containing the pBAD plasmid were electroporated in the presence of the pDULE TetR tRNA synthetase plasmid. The cells were rescued with 450 μL Expres-sion Recovery Media and incubated at 37 °C with shaking for 2 h. The cells were plated on LB-agar plates with 50 μg/ml carbenicillin and 15 μg/ml tetracycline. The overnight cultures obtained from the colonies selected were diluted with 500 mL autoinduction me-dia as described above. The cells were pelleted and stored at -20 °C. The MS2 protein was purified as described in Chapter 2. The yield of the protein expression was very low as compared to the expression in E. coli of the corresponding mutant. The identity of the MS2 mutants was confirmed by ESI-MS and oxidative coupling reactions with aminophe-nol-PEG (data not shown). The amount of endotoxin present was measured using the LAL kit as described above. Surprisingly, higher levels of endotoxin were measured for the protein purified from Clear coli than from E. coli (about 10 x difference, 430,000 EU/mL versus 3027 EU/mL for 50 μM protein solutions).

A HiTrap Q FF strong anion exchange column (GE Healthcare, was loaded with a MS2 solution and eluted with 1 mM Tris, 50 mM ammonium sulfate, pH 7.5 at 1 mL/min. Fractions were tested for endotoxin presence. No significant reduction in endotoxin levels was observed.

A solution of MS2 protein in 10 mM phosphate buffer pH 7.2 was mixed with TritonX 114 (Sigma) at a final concentration of 1%. The solution was kept on ice for 5 min, then at room temperature for 5 min, followed by centrifugation at 10,000 g for 1 min. The super-natant was moved to a new tube and the process was repeated twice. A partition between two phases could be observed upon spinning. The final supernatant was loaded onto a NAP-25 desalting column and eluted with endotoxin free phosphate buffer. The fractions containing the protein were concentrated using a 100 kDa MWCO spin concentrator. The overall protein recovery was 29%. Traces of detergent were still present, as indicated by the UV spectrum measured by NanoDrop. Endotoxin levels were slightly reduced, but the probability of having TritonX 114 still present in the samples limited the use of this method

57

for preparing samples for in vivo studies.

Acrodisc® Syringe filters with Mustang® E membranes (PALL) were used in attempt to remove endotoxin. Protein recovery levels were very low, making this method not a feasible option for removing endotoxins from samples of interest.

3.13 References

1. Steichen, S.D., Caldorera-Moore, M. and Peppas, N.A., A review of current nanoparti-cle and targeting moieties for the delivery of cancer therapeutics. Eur J Pharm Sci. 2013, 48, 416-427.

2. Wilhelm, S., Tavares, A.J., Dai, Q., Ohta, S., Audet, J., Dvorak, H.F., et al., Analysis of nanoparticle delivery to tumours. Nature Reviews Materials. 2016, 1.

3. Nagy, J.A., Dvorak, A.M. and Dvorak, H.F., Vascular hyperpermeability, angiogenesis, and stroma generation. Cold Spring Harb Perspect Med. 2012, 2, a006544.

4. Pettersson, A., Nagy, J.A., Brown, L.F., Sundberg, C., Morgan, E., Jungles, S., et al., Heterogeneity of the angiogenic response induced in different normal adult tissues by vascular permeability factor/vascular endothelial growth factor. Lab Invest. 2000, 80, 99-115.

5. Nagy, J.A., Benjamin, L., Zeng, H., Dvorak, A.M. and Dvorak, H.F., Vascular perme-ability, vascular hyperpermeability and angiogenesis. Angiogenesis. 2008, 11, 109-119.

6. Nagy, J.A., Chang, S.H., Shih, S.C., Dvorak, A.M. and Dvorak, H.F., Heterogeneity of the tumor vasculature. Semin Thromb Hemost. 2010, 36, 321-331.

7. Brannon-Peppas, L. and Blanchette, J.O., Nanoparticle and targeted systems for can-cer therapy. Adv Drug Deliv Rev. 2004, 56, 1649-1659.

8. Maeda, H., Nakamura, H. and Fang, J., The EPR effect for macromolecular drug de-livery to solid tumors: Improvement of tumor uptake, lowering of systemic toxicity, and distinct tumor imaging in vivo. Adv Drug Deliv Rev. 2013, 65, 71-79.

9. Hobbs, S.K., Monsky, W.L., Yuan, F., Roberts, W.G., Griffith, L., Torchilin, V.P., et al., Regulation of transport pathways in tumor vessels: role of tumor type and microenviron-ment. Proc Natl Acad Sci U S A. 1998, 95, 4607-4612.

10. Dvorak, A.M. and Feng, D., The vesiculo-vacuolar organelle (VVO). A new endothelial cell permeability organelle. J Histochem Cytochem. 2001, 49, 419-432.

11. Pickup, M.W., Mouw, J.K. and Weaver, V.M., The extracellular matrix modulates the hallmarks of cancer. EMBO Rep. 2014, 15, 1243-1253.

12. Heldin, C.H., Rubin, K., Pietras, K. and Ostman, A., High interstitial fluid pressure - an obstacle in cancer therapy. Nat Rev Cancer. 2004, 4, 806-813.

58

13. Jain, R.K. and Stylianopoulos, T., Delivering nanomedicine to solid tumors. Nature Reviews Clinical Oncology. 2010, 7, 653-664.

14. Perrault, S.D., Walkey, C., Jennings, T., Fischer, H.C. and Chan, W.C., Mediating tu-mor targeting efficiency of nanoparticles through design. Nano Lett. 2009, 9, 1909-1915.

15. Albanese, A., Lam, A.K., Sykes, E.A., Rocheleau, J.V. and Chan, W.C.W., Tumour-on-a-chip provides an optical window into nanoparticle tissue transport. Nature Commu-nications. 2013, 4.

16. Bazak, R., Houri, M., El Achy, S., Kamel, S. and Refaat, T., Cancer active targeting by nanoparticles: a comprehensive review of literature. J Cancer Res Clin Oncol. 2015, 141, 769-784.

17. Boussios, S., Pentheroudakis, G., Katsanos, K. and Pavlidis, N., Systemic treat-ment-induced gastrointestinal toxicity: incidence, clinical presentation and management. Ann Gastroenterol. 2012, 25, 106-118.

18. Devalapally, H., Chakilam, A. and Amiji, M.M., Role of nanotechnology in pharmaceu-tical product development. J Pharm Sci. 2007, 96, 2547-2565.

19. Allen, T.M. and Cullis, P.R., Drug delivery systems: entering the mainstream. Science. 2004, 303, 1818-1822.

20. De Jong, W.H. and Borm, P.J., Drug delivery and nanoparticles:applications and haz-ards. Int J Nanomedicine. 2008, 3, 133-149.

21. Ren, Y., Wong, S.M. and Lim, L.Y., Folic acid-conjugated protein cages of a plant vi-rus: a novel delivery platform for doxorubicin. Bioconjug Chem. 2007, 18, 836-843.

22. Destito, G., Yeh, R., Rae, C.S., Finn, M.G. and Manchester, M., Folic acid-mediated targeting of cowpea mosaic virus particles to tumor cells. Chem Biol. 2007, 14, 1152-1162.

23. Tong, G.J., Hsiao, S.C., Carrico, Z.M. and Francis, M.B., Viral capsid DNA aptamer conjugates as multivalent cell-targeting vehicles. J Am Chem Soc. 2009, 131, 11174-11178.

24. Liang, C., Guo, B., Wu, H., Shao, N., Li, D., Liu, J., et al., Aptamer-functionalized lipid nanoparticles targeting osteoblasts as a novel RNA interference-based bone anabolic strategy. Nat Med. 2015, 21, 288-294.

25. Behrens, C.R., Hooker, J.M., Obermeyer, A.C., Romanini, D.W., Katz, E.M. and Fran-cis, M.B., Rapid chemoselective bioconjugation through oxidative coupling of anilines and aminophenols. J Am Chem Soc. 2011, 133, 16398-16401.

26. Obermeyer, A.C., Capehart, S.L., Jarman, J.B. and Francis, M.B., Multivalent viral capsids with internal cargo for fibrin imaging. PLoS One. 2014, 9, e100678.

27. Carrico, Z.M., Romanini, D.W., Mehl, R.A. and Francis, M.B., Oxidative coupling of

59

peptides to a virus capsid containing unnatural amino acids. Chem Commun (Camb). 2008, 1205-1207.

28. Rhee, J.K., Baksh, M., Nycholat, C., Paulson, J.C., Kitagishi, H. and Finn, M.G., Glycan-targeted virus-like nanoparticles for photodynamic therapy. Biomacromolecules. 2012, 13, 2333-2338.

29. ElSohly, A.M., Netirojjanakul, C., Aanei, I.L., Jager, A., Bendall, S.C., Farkas, M.E., et al., Synthetically Modified Viral Capsids as Versatile Carriers for Use in Antibody-Based Cell Targeting. Bioconjug Chem. 2015, 26, 1590-1596.

30. Aanei, I.L., ElSohly, A.M., Farkas, M.E., Netirojjanakul, C., Regan, M., Taylor Murphy, S., et al., Biodistribution of Antibody-MS2 Viral Capsid Conjugates in Breast Cancer Mod-els. Mol Pharm. 2016, 13, 3764-3772.

31. Brown, W.L., Mastico, R.A., Wu, M., Heal, K.G., Adams, C.J., Murray, J.B., et al., RNA bacteriophage capsid-mediated drug delivery and epitope presentation. Intervirolo-gy. 2002, 45, 371-380.

32. Flenniken, M.L., Willits, D.A., Harmsen, A.L., Liepold, L.O., Harmsen, A.G., Young, M.J., et al., Melanoma and lymphocyte cell-specific targeting incorporated into a heat shock protein cage architecture. Chem Biol. 2006, 13, 161-170.

33. Zeng, Q.B., Wen, H.B., Wen, Q., Chen, X.H., Wang, Y.G., Xuan, W.L., et al., Cucum-ber mosaic virus as drug delivery vehicle for doxorubicin. Biomaterials. 2013, 34, 4632-4642.

34. Banerjee, P.S., Ostapchuk, P., Hearing, P. and Carrico, I., Chemoselective attachment of small molecule effector functionality to human adenoviruses facilitates gene delivery to cancer cells. J Am Chem Soc. 2010, 132, 13615-13617.

35. Low, P.S. and Kularatne, S.A., Folate-targeted therapeutic and imaging agents for cancer. Curr Opin Chem Biol. 2009, 13, 256-262.

36. Zhao, Q., Chen, W., Chen, Y., Zhang, L., Zhang, J. and Zhang, Z., Self-assembled virus-like particles from rotavirus structural protein VP6 for targeted drug delivery. Biocon-jug Chem. 2011, 22, 346-352.

37. Lockney, D.M., Guenther, R.N., Loo, L., Overton, W., Antonelli, R., Clark, J., et al., The Red clover necrotic mosaic virus capsid as a multifunctional cell targeting plant viral nanoparticle. Bioconjug Chem. 2011, 22, 67-73.

38. Huang, R.K., Steinmetz, N.F., Fu, C.Y., Manchester, M. and Johnson, J.E., Transfer-rin-mediated targeting of bacteriophage HK97 nanoparticles into tumor cells. Nanomedi-cine (Lond). 2011, 6, 55-68.

39. Wu, M., Brown, W.L. and Stockley, P.G., Cell-specific delivery of bacteriophage-en-capsidated ricin A chain. Bioconjug Chem. 1995, 6, 587-595.

40. Paszko, E. and Senge, M.O., Immunoliposomes. Curr Med Chem. 2012, 19, 5239-

60

5277.

41. Hooker, J.M., O’Neil, J.P., Romanini, D.W., Taylor, S.E. and Francis, M.B., Ge-nome-free viral capsids as carriers for positron emission tomography radiolabels. Mol Imaging Biol. 2008, 10, 182-191.

42. Wu, W., Hsiao, S.C., Carrico, Z.M. and Francis, M.B., Genome-free viral capsids as multivalent carriers for taxol delivery. Angew Chem Int Ed Engl. 2009, 48, 9493-9497.

43. Bendall, S.C., Simonds, E.F., Qiu, P., Amir el, A.D., Krutzik, P.O., Finck, R., et al., Single-cell mass cytometry of differential immune and drug responses across a human hematopoietic continuum. Science. 2011, 332, 687-696.

44. de Jong, M., Essers, J. and van Weerden, W.M., Imaging preclinical tumour models: improving translational power. Nat Rev Cancer. 2014, 14, 481-493.

45. Howe, L.R. and Brown, P.H., Targeting the HER/EGFR/ErbB family to prevent breast cancer. Cancer Prev Res (Phila). 2011, 4, 1149-1157.

46. O’Riordan, C.R., Lachapelle, A., Delgado, C., Parkes, V., Wadsworth, S.C., Smith, A.E., et al., PEGylation of adenovirus with retention of infectivity and protection from neu-tralizing antibody in vitro and in vivo. Hum Gene Ther. 1999, 10, 1349-1358.

47. Keifer, D.Z. and Jarrold, M.F., Single-molecule mass spectrometry. Mass Spectrom Rev. 2016.

48. Fenn, J.B., Mann, M., Meng, C.K., Wong, S.F. and Whitehouse, C.M., Electrospray ionization for mass spectrometry of large biomolecules. Science. 1989, 246, 64-71.

49. Heck, A.J., Native mass spectrometry: a bridge between interactomics and structural biology. Nat Methods. 2008, 5, 927-933.

50. Fuerstenau, S.D., Benner, W.H., Thomas, J.J., Brugidou, C., Bothner, B. and Siuzdak, G., Mass Spectrometry of an Intact Virus. Angew Chem Int Ed Engl. 2001, 40, 9822.

51. Benner, W.H., A gated electrostatic ion trap to repetitiously measure the charge and m/z of large electrospray ions. Analytical Chemistry. 1997, 69, 4162-4168.

52. Fuerstenau, S.D. and Benner, W.H., Molecular weight determination of megadalton DNA electrospray ions using charge detection time-of-flight mass spectrometry. Rapid Commun Mass Spectrom. 1995, 9, 1528-1538.

53. Pierson, E.E., Keifer, D.Z., Selzer, L., Lee, L.S., Contino, N.C., Wang, J.C., et al., Detection of late intermediates in virus capsid assembly by charge detection mass spec-trometry. J Am Chem Soc. 2014, 136, 3536-3541.

54. Walkey, C.D., Olsen, J.B., Song, F., Liu, R., Guo, H., Olsen, D.W., et al., Protein co-rona fingerprinting predicts the cellular interaction of gold and silver nanoparticles. ACS Nano. 2014, 8, 2439-2455.

55. Tanner, S.D., Baranov, V.I., Ornatsky, O.I., Bandura, D.R. and George, T.C., An intro-

61

duction to mass cytometry: fundamentals and applications. Cancer Immunol Immunother. 2013, 62, 955-965.

56. Spitzer, M.H. and Nolan, G.P., Mass Cytometry: Single Cells, Many Features. Cell. 2016, 165, 780-791.

57. Albanese, A., Tang, P.S. and Chan, W.C., The effect of nanoparticle size, shape, and surface chemistry on biological systems. Annu Rev Biomed Eng. 2012, 14, 1-16.

58. Barua, S., Yoo, J.W., Kolhar, P., Wakankar, A., Gokarn, Y.R. and Mitragotri, S., Particle shape enhances specificity of antibody-displaying nanoparticles. Proc Natl Acad Sci U S A. 2013, 110, 3270-3275.

59. Melancon, M.P., Lu, W., Yang, Z., Zhang, R., Cheng, Z., Elliot, A.M., et al., In vitro and in vivo targeting of hollow gold nanoshells directed at epidermal growth factor receptor for photothermal ablation therapy. Mol Cancer Ther. 2008, 7, 1730-1739.

60. Kirpotin, D.B., Drummond, D.C., Shao, Y., Shalaby, M.R., Hong, K., Nielsen, U.B., et al., Antibody targeting of long-circulating lipidic nanoparticles does not increase tumor localization but does increase internalization in animal models. Cancer Res. 2006, 66, 6732-6740.

61. Mamot, C., Drummond, D.C., Noble, C.O., Kallab, V., Guo, Z., Hong, K., et al., Epider-mal growth factor receptor-targeted immunoliposomes significantly enhance the efficacy of multiple anticancer drugs in vivo. Cancer Res. 2005, 65, 11631-11638.

62. Mastico, R.A., Talbot, S.J. and Stockley, P.G., Multiple presentation of foreign pep-tides on the surface of an RNA-free spherical bacteriophage capsid. J Gen Virol. 1993, 74 ( Pt 4), 541-548.

63. Kaiser, C.R., Flenniken, M.L., Gillitzer, E., Harmsen, A.L., Harmsen, A.G., Jutila, M.A., et al., Biodistribution studies of protein cage nanoparticles demonstrate broad tissue dis-tribution and rapid clearance in vivo. Int J Nanomedicine. 2007, 2, 715-733.

64. Davies, B. and Cohen, J., Endotoxin removal devices for the treatment of sepsis and septic shock. Lancet Infect Dis. 2011, 11, 65-71.

65. Petsch, D. and Anspach, F.B., Endotoxin removal from protein solutions. J Biotechnol. 2000, 76, 97-119.

66. Hirayama, C. and Sakata, M., Chromatographic removal of endotoxin from protein solutions by polymer particles. J Chromatogr B Analyt Technol Biomed Life Sci. 2002, 781, 419-432.

67. Aida, Y. and Pabst, M.J., Removal of endotoxin from protein solutions by phase sep-aration using Triton X-114. J Immunol Methods. 1990, 132, 191-195.

62

Chapter 4: MS2 as a cardiovascular imaging agent

Abstract

Atherosclerosis is a cardiovascular disease characterized by the formation of lipid-rich plaques within the walls or large arteries, which can detach and cause serious complica-tions, such as strokes or heart attacks. At the moment, there is no effective way to detect the presence of atherosclerosis in patients until it has reached a relatively advanced stage. Additionally, increasing evidence suggests that the pathobiological behavior of plaques is determined mainly by their composition, and not their size, which is the pa-rameter usually monitored with current imaging techniques. We set out to synthesize a panel of agents that can target multiple markers of atherosclerotic plaque development. These agents could be used for imaging, diagnosis and drug delivery purposes. Initial experiments focused on targeting the Vascular Cell Adhesion Molecule (VCAM1), a pro-tein that plays a crucial role in the disease progression. In vivo experiments with murine atherosclerosis models indicated that the targeted protein nanoparticles were successful in detecting plaques of various sizes in the descending aorta and the aortic arch.

63

4.1 Imaging techniques for cardiovascular diseases

Cardiovascular diseases are currently the leading cause of death worldwide,1 most commonly manifesting through the prevention of blood flow to critical organs, such as the heart (myocardial infarction) or the brain (stroke).2 The main contributor to these afflic-tions is atherosclerosis (Figure 4.1a), which is a systemic disease that produces both the localized narrowing of the arterial lumen (Figure 4.1b)3 and the development of fibrous caps that can rupture to cause blockages.4 It is now appreciated that atherosclerosis is a complex inflammatory disease of the blood vessels involving several distinct stages.5 Traditional imaging methods (Figure 4.1c) yield little information about the molecular composition of plaques, their corresponding stages in atherosclerosis, and thus the levels of risk that are associated with them.5 Specific imaging agents that can detect markers for the different stages of atherosclerosis could also provide valuable tools for shedding light on the molecular mechanisms of plaque development and for monitoring the effects of treatments on disease progression.

500 μm

healthy artery atherosclerotic artery

lesion

lesion

MRIUltrasound

(a) (b)

(c)

lipid richnecrotic core

thrombus

macrophage accumulationTissue

Artery

Time

thrombus

endothelial-cell activation

adhesionmolecules

inflammation

macrophages

lipid core and fibrouscap formation

monocyte macrophage foam cell fibrinplateletLDL VCAM1, ICAM1and selectins

smooth muscle cell

adhesion molecules

Process

Target

Stage early middle late

(c)

Figure 4.1 Progression of atherosclerosis. (a) The development of an atherosclerotic lesion is shown in a simplified form, developing from a normal blood vessel (far left) to a vessel with an atherosclerotic plaque and superimposed thrombus (far right). In early stages, endothelial cell activation leads to overexpression of adhesion molecules such as VCAM1 and ICAM1. Monocyte recruitment is followed by macrophage accumulation and foam cell formation. In advanced stages, the plaque develops a thrombus that can dettach and block smaller blood vessels, leading to stroke or heart attack. Potential targets for molecular imaging at each stage are also listed. ICAM, intercellular cell adhesion molecule; LDL, low density lipopro-tein; tomography; VCAM, vascular cell adhesion molecule. Figure adapted from reference 2. (b) Histologi-cal analysis of ex vivo tissue from a atherosclerotic patient shows clear changes in the diameter of the artery, as well as the increase of the number of cells present at the site of the lesion (cell nuclei are stained with hematoxylin blue, proteins are stained with eosin pink). Figure adapted from reference 3. (c) Current imaging techniques such as ultrasound and MRI can detect significant changes in the diameter of blood vessels, but cannot provide information about the molecular composition of the plaque and therefore are limited in diagnosing the stage of atherosclerosis development. Figure adapted from reference 2.

64

A number of groups have developed PET,6, 7 MRI8-10 and ultrasound methods11, 12 that can detect specific components or events that occur in arterial plaques. While significant progress has been made (e.g. 18F-fluorodeoxyglucose for macrophage detection with PET,13 CT scans for detecting calcification,14 and the measurement of lipid accumulation using MRI15), the majority of these techniques are applicable to a single stage of athero-sclerotic plaque development and are not readily adaptable for use with other targets. In re cent years, targeted approaches to the imaging of mo lecular markers have been at-tempted through the use of peptides and antibodies modified with tracers.16 The efficiency of most of these constructs could be augmented by the use of nanoscale carriers that can present multiple cop ies of the targeting groups and numerous copies of the reporters. In addition to increasing the binding affinity through multivalency, such platforms could produce improved levels of signal and could allow multimodal detection as tracers can be combined in various ways.

4.2 Using the MS2 capsid to target atherosclerotic plaques

As discussed in Chapters 2 and 3, the MS2 system can incorporate up to 180 copies of any reporter group inside the protein shell with minimal to no effect on biodistribution. From experiments described in previous chapters, we had information about the biodistri-bution properties of several conjugates of the MS2 capsids, as well as the stability of the capsids in vivo and in vitro. We observed that a large amount of signal was still present in the blood at 24 h and even 48 h post-injection. We have also obtained evidence of intact capsids in blood samples isolated after 24 h, as determined using analytical SEC with radiochemical detection. Together, these experiments show that MS2 capsids are surprisingly resistant to clearance from the blood stream. This is in sharp contrast to other viral cap sids that have been evaluated, which are typically cleared completely in minutes to a few hours. Therefore, we wanted to explore the potential of MS2 to target markers present in the cardiovascular system.

We envisioned developing this system into a versatile platform that can image specific molecules associated with early, middle, and late stages of atherosclerosis, such as adhe-sion molecules, macrophages and thrombi, respectively (Figure 4.1a). Once established, the new system could provide a major step forward in early-stage cardiovascular imag-ing, diagnosis, and treatment monitoring. Initial studies focused on the early markers of atherosclerosis, such as the Vascular Cell Adhesion Molecule 1 (VCAM1 or CD106).17,18 The installation of target ing groups on the external surface of the MS2 capsid allows for active targeting, which is more specific and informative than passive, metabolism-based approaches such as 18F-fluorodeoxyglucose (FDG).19

As discussed in Chapter 3, our group has used a powerful series of oxidative coupling reactions to install targeting groups on the capsid surface (~135 cyclic RGD peptides, up to 100 DNA aptamers and several copies of full-length antibodies). We have demonstrat-ed that the oxidative coupling reaction conditions are compatible with the targeting groups themselves, as well as complex imaging or drug cargo. We have recently found that MS2 capsids bearing ~90 copies of a fibrin-binding peptide associate with aggregating fibrin polymers, and thus inhibit clot formation in in vitro assays.20 MS2 capsids bearing control

65

peptides did not show inhibition, nor did the fibrin binding peptide itself at an equivalent concentration. This result indicated that the capsids can interact with polymerized fibrin, and that multivalency effects are enhancing the binding. In infrared imaging experiments, MS2 capsids bearing fibrin peptides were also observed to bind to clot specimens formed in vitro.

In order to evaluate the targeting ability of the agents, we used commercially avail-able models of atherosclerosis. The most commonly used model consists of mice with a knockout of the gene encoding for Apolipoprotein E (ApoE), a protein important for the clearance of cholesterol and triglyceride-rich lipoprotein particles from the blood.21 In hu-mans, mutations in ApoE are associated with several hereditary hyperlipidemic disorders, including familial hypercholesterolemia and type III hyperlipidemia, and patients with these conditions have increased susceptibility to atherosclerosis. ApoE-deficient mice show elevated plasma cholesterol levels, and develop atherosclerotic plaques either spontaneously (after 3-8 months) or more predictably when fed a high fat diet (plaques form after 14 weeks).21 After confirming that the agents are specific to their targets and localize within lesions, in vivo imaging and other targets (such as the CD68 receptors on macrophages22) could be pursued. Given the versatility of the MS2 platform, agents tar-geting other atherosclerotic markers should be easily obtained.

4.3 Synthesis and characterization of VCAM-targeted agents

The VCAM1 receptor is expressed on activated endothelial cells that line blood ves-sels, as well as on macrophages present in atherosclerotic lesions.17, 21 Its role is to reg-ulate the cell adhesion through interaction with the integrin very late antigen-4 (VLA-4).2 This process only takes place in early stages of atherosclerosis, during the inflammatory phase. Given that the expression of VCAM1 is tightly confined to regions of inflammation and that the marker is accessible from the blood pool, we decided to pursue VCAM as a target for cardiovascular disease imaging agents.

In vivo phage display in ApoE–deficient mice identified a linear peptide affinity ligand, VHPKQHR, homologous to VLA-4.23, 24 This peptide was developed by the Weissleder lab into a multivalent agent detectable by MRI and optical imaging.23, 25 These magnetic nanoparticles allowed noninvasive imaging of VCAM1–expressing endothelial cells and macrophages in atherosclerosis and could be used in the clinic to detect inflammation in early stages of atherosclerosis. Although current efforts are dedicated towards synthesiz-ing iron nanoparticles that are biodegradable, the scaffold used in the study by Nahren-dorf et al. suffers from bioaccumulation issues.26

We have designed a strategy to use the VHP peptide to target a protein-based nanopar-ticle that can be degraded. We installed a PEG5kDa linker in between the MS2 viral capsid and the VHP sequence in order to allow for unrestricted interaction with the VCAM protein (Figure 4.2). For initial studies, we used optical imaging to determine if the VCAM-target-ed nanoparticles can reach their target and are specific for the atherosclerotic plaques. A fluorescent dye with near-IR properties (AlexaFluor680) and excellent photostability was chosen to limit the tissue autofluorescence issues commonly encountered at lower wavelengths. The free peptide was also labeled with a fluorescent dye and served as the

66

positive control (Figure 4.2b, c and d).

In order to test our hypothesis that the lack of increased accumulation for the targeted agents compared to untargeted agents was caused by extravasation being the limiting step (Chapter 3), we decided to also synthesize and investigate the capsid decorated with anti-VCAM antibodies. The synthetic protocol used was similar to the one described in Chapter 3 and is summarized in Figure 4.2a. Good levels of modification were observed

(a)

(b)

(d)

(e)

(f)

(c)

VHP-PEG5kDa-APVHP-SH + Mal-PEG5kDa-NP

MWladder

Wavelength (nm)

1 2 3

50

37

25

20

75

15

kDa

HC + 1 MS2

HC

LC + 2 MS2 LC + 1 MS2

MS2 monomer

LC MS2-PEG5kDa-VHP MS2-PEG5kDa

4 5 6

Figure 4.2 Synthesis and characterization of VCAM-targeted MS2 agents. (a) The VHP peptide (VHP-KQHRGGSKGC) was reacted with a PEG5kDa spacer that allowed for attachment to AlexaFluor680 (AF680) fluorescently-labeled MS2 capsids using the oxidative coupling reaction presented in previous chapters. The lysine residues on the αVCAM antibody were used to couple to a nitrophenol moiety through an NHS-amine reaction. The antibody was then attached to the MS2 VLP as described in previous chapters. NP = nitrophenol, NHS = N-hydroxysuccinimidyl, AP = aminophenol; mal = maleimide. (b) The free peptide and free antibody controls were labeled with AF680 fluorescent dyes (maleimide and NHS, respectively) to allow for direct comparison with the targeted MS2 agents. Unmodified MS2 capsid was used to determine non-specific interactions between the capsid and the plaques. (c) UV-VIS spectra of the agents, showing peaks at 680 nm corresponding to the AF680 modification. (d) ESI-MS spectra of VHP-AF680 showing the +2, +3 and +4 charged states. The (*) represents a small impurity. (e) Modification of the MS2 capsid with AF680-mal can lead to high levels of conversion, as shown in the bottom LC-MS spectrum. (f) SDS-PAGE analysis of VCAM-targeted MS2 agents. Lane 1 shows the MS2 paF/Cys starting material, while lane 2 shows the capsid conjugated to AF680 dye. Lane 3 contains the MS2-αVCAM conjugate and the light chains (LC) and heavy chains (HC), as well as their modification with MS2 can be observed. Lane 4 has MS2-PEG5kDa-VHP capsids. The higher molecular band corresponding to the conjugate (~60% modifica-tion) is higher than the band in lane 5, corresponding to an MS2-PEG5kDa-OMe control, given that the peptide adds about 1 kDa to the mass of the PEG chain. Lane 6 contains an unmodified αVCAM antibody, for reference purposes.

MS2-AF680-PEG5kDa-VHPNaIO4, 5 min, RT

1. overnight, RT

2. Na2S2O4 reduction

MS2-AF680

αVCAM-APαVCAM-NH2 + NHS-NP MS2-AF680-αVCAMNaIO4, 5 min, RT

1. 2.5 h, RT

2. Na2S2O4 reduction

MS2-AF680

MS2-SH + AF680-mal MS2-AF6804 h, RT

αVCAM-NH2 + AF680-NHS αVCAM-AF6802.5 h, RT

VHP-SH + AF680-mal VHP-AF680overnight, RT

220 340 460 580 700

αVCAM VHP MS2MS2-VCAM MS2-PEG-VHP

Abs

orba

nce

1000600 800 1200

594

713

791

1186

+4

+3

+2

*

mass

13778

13600 14000 14400 14800

14759

mass (Da)

MS2

MS2-AF680

VHP-AF680

AF680

67

upon conjugation (Figure 4.2f). Based on the results described in Chapter 3, we decided not to use a PEG5kDa linker in between the MS2 capsid and the antibody. The free antibody labeled with the AlexaFluor680 (AF680) dye served as the control.

4.4 Pathology of ApoE knockout mice

Female mice with a ApoE knockout (ApoE KO) genotype were purchased from The Jackson Laboratories. After weaning at 4 weeks, the mice were fed a high fat (“Western”) diet and aged for 15 weeks. To verify that the conditions led to the formation of atherosclerotic plaques, the descending aorta from ApoE KO mice and control mice (C57B/6J strain, fed normal chow) were harvested and compared. The presence of numerous plaques (opaque white spots) was observed in the case of ApoE mice (Figure 4.3a). The aortic arch also presented many visible plaques, highlighted by red arrows. The blood vessels were cryopreserved and sectioned. Hematoxylin/eosin (H&E) staining was performed to investigate the cellular composition of the blood vessel walls. Oil Red O was used to detect the plaques. In Figure 4.3b, it can be observed that the blood vessels harvested from ApoE KO mice have a thicker layer of cells and Oil Red O staining reveals the presence of a lesion. Therefore, the designed aging process was successful in developing plaques in ApoE KO mice.

4.5 Biodistribution and plaque-targeting ability of VCAM-targeted agents

ApoE KO and control mice were injected intravenously (IV) with fluorescently-labeled MS2 agents, free peptide and antibody controls. At 5 h post-injection, the mice were sacrificed and perfused with phosphate buffered saline (PBS) and neutral buffered 10% formalin. The main internal organs (liver, pancreas, spleen, kidneys, heart and fat) were harvested and imaged using an IVIS Xenogen 50 system. Figure 4.4 shows the overlay of the photographs and fluorescent channel readings, false colored in red. The count

Control ApoE KO Control ApoE KO

Figure 4.3 Comparison of the pathology of control and ApoE KO mice. (a) Photographs of the descending aortas for control (C57BL/6J) and ApoE knockout (ApoE KO) mice showing the presence of lesions (opaque white, highlighted by the red arrows) only in the case of ApoE KO mice fed a high fat diet. (b) Segments of the aorta were embedded in Optimal Cutting Temperature media (OCT) and sliced using a cryotome. The sections were stained using standard protocols for haematoxylin/eosin staining (H&E, top panel) and Oil Red O staining (ORO, bottom panel). The ApoE KO samples show a thicker layer of cells cells (nuclei are stained in blue) and the presence of a lesion (red staining by ORO).

(b)(a)

aortic arch

aorta

100 μm

lesion

68

range was adjusted to limit the autofluorescence signal, based on the PBS-injected control mice. The agents mainly accumulated in the liver, kidneys and spleen. Fat tissue was included as a reference, given that accumulation in the fat tissue is not very likely. The VHP-dye and antibody-dye samples were dosed at the same number of moles as displayed on the MS2 capsids but did not have the same number of dyes attached, and thus the signal was very weak.

After excising the descending aortas and removing the surrounding connective tissue, the blood vessels were imaged using an IVIS Spectrum system. A bin size of 1 was selected for image analysis in order to achieve maximum resolution. The overlay of the photographs and fluorescent channel are shown in Figure 4.5a for representative samples from each group. It is only in the case of VCAM-targeted agents (number 7 and 8) that the red signal overlays with the opaque spots representing the atherosclerotic plaques.

Figure 4.5b shows all the replicates for the MS2-, MS2-PEG5kDa-VHP- and MS2-aVCAM-treated mice. A small amount of signal is colocalized with the plaques in the case of MS2-treated mice. This observation could be explained by the fact that macrophages present in the plaques could interact with the MS2 capsids. In in vitro experiments (data not shown), we observed that activated macrophages can incorporate the capsid after

Figure 4.4 Biodistribution of VCAM-targeted MS2 agents. An IVIS imaging system was used to determine the biodistribution of the fluorescently-labeled agents at the 24 h time point. The photographs are shown overlaid with the signal from the Cy5 channel (same emission/excitation as AF680), false colored in red. All images are adjusted to the same count range, represented on the right side. The range was chosen to limit the amount of signal in the PBS samples, representing organ autofluorescence. Control mice did not possess atherosclerotic lesions, had lower overall body weights and less fat. The main organs where the signal accumulated were the liver (L), kidneys (K) and spleen (S). The pancreas (P), heart (H) and fat (F) had very small amounts of agents still present at 24 h. The VHP-targeted agents seem to have a faster clearance in the control mice (second panel) versus the ApoE KO mice (panel 7). The antibody-targeted agents accumulate more in the spleen and less in the kidneys in the control mice (panel 3) versus the ApoE KO mice (panel 8).

1. PBS

L P S K H F

2. MS2-PEG-VHP 3. MS2-VCAM 4. MS2

8. MS2-VCAM7. MS2-PEG-VHP6. VCAM5. VHP

Control mice Control mice

69

prolonged incubation. The VCAM-targeted agents show excellent localization with the plaques, including the small plaques that formed at the branching of arteries from the main aorta. This result is very encouraging for the potential of MS2-based agents to specifically target markers in the cardiovascular system for use in non-invasive imaging, diagnosis and drug delivery applications.

The blood vessels harvested were cryopreserved and sectioned for fluorescent microscopy analysis (Figure 4.6). Unfortunately, the high level of autofluorescence and low signal limited the conclusions that can be drawn from these studies. We hypothesize that the low signal could be explained by the low dose of agent administered, photobleaching of the dye during IVIS imaging and handling or the diffusion of the agent during the cryopreservation protocol.

(a) 1 2 3 4 5 6 7 8

MS2-PEG5kDa-VHP MS2-αVCAMMS2

(b)

Figure 4.5 Colocalization of agents with atherosclerotic plaques. (a) IVIS imaging of representative descending aortas from control mice (1 and 2) and ApoE KO mice (3-8). The photographs are overlaid with the dye signal, falsely colored in red. A bin size of one was used to obtain the highest resolution. (b) The fluorescent signal is represented in the reverse rainbow color scheme, showing more clearly the intensity of the signal at each location (with red representing higher signal and blue representing lower signal). A bin size of 4 was used to process these images, to allow for detection of weaker signals. All images were processed with the same parameters, including background correction. The first panel shows untargeted MS2 capsids, with limited accumulation within the lesions. This accumulation could be explained by interac-tion with macrophages. For the VCAM-targeted agents, the localization within the lesions is much more pronounced and even small lesions on the branching arteries can be observed. The aortic arch, which is the most likely location for plaque formation, is clearly containing the fluorescently-labeled MS2 agents.

3. PBS

1. control mice + MS2-PEG5kDa-VHP

2. control mice + MS2-αVCAM

4. MS2

8. MS2-VCAM

7. MS2-PEG5kDa-VHP

6. αVCAM

5. VHP

70

4.6 Conclusions

We have synthesized and characterized VCAM-targeted MS2 agents that can be used for the detection of atherosclerotic plaques in early stages of development. The targeted agents accumulated much more in the lesions than the untargeted MS2 scaffold, providing a confirmation for the hypothesis that previous experiments in breast cancer models (Chapter 3) might have been limited by the extravasation into the tumor environment. Given the success of the VCAM-targeted agents to image the plaques, we plan to synthesize agents targeting other stages of atherosclerotic plaque development as well. Using the versatile MS2 platform, a series of agents can be developed for the imaging, diagnosis and treatment of cardiovascular diseases.

4.7 Materials and methods

All reagents were obtained from commercial sources and used without any further purification. Anti-VCAM (CD106, clone M/K2) antibody was purchased from Southern Biotech (Birmingham, AL). VHP peptide (VHPKQHRGGSKGC) was custom ordered from GenScript (Piscataway, NJ) and stored as a lyophilized solid. AlexaFluor680 (AF680) dyes with maleimide or NHS reactive handles and sterile PBS were purchased from Ther-mo Fisher (Waltham, MA). Amine-PEG5kDa-maleimide was purchased from JenKem (Pla-no, TX). N-hydroxysuccinimidyl nitrophenol (NHS-NP) was synthesized as described in Chapter 2. Spin concentrators with 3, 30 and 100 kDa molecular weight cutoffs (MWCO) and sterile spin filters with 0.22 mm pores were purchased from Millipore (Billerica, MA). Neutral buffered 10% formalin (NBF) was purchased from Sigma (St. Louis, MO). Doubly distilled water was obtained from a Millipore purification system.

Electrospray ionization mass spectrometry (ESI-MS) and sodium dodecyl sulfate-poly-acrylamide gel electrophoresis (SDS-PAGE) analysis of proteins was performed as de-scribed in Chapter 2. UV-VIS measurements were taken on a NanoDrop 1000 (Thermo).

MS2 expression and purification. The protocol for MS2 T19paF N87C expression and purification is described in Chapter 2.

MS2-PEG5kDa-VHPPBS MS2-αVCAM

Figure 4.6 Confocal microscopy on sections of blood vessels. High autofluorescence was observed in the PBS sample. The low signal in the VCAM-targeted samples can be explained by prolonged interaction with buffers during preparation for cryopreservation and extended exposure to light during IVIS imaging and manipulation.

71

MS2 modification with AF680. To a solution of MS2 paF/Cys in 25 mM phosphate buffer pH 8 was added AF680-maleimide (0.8 equiv.) from a DMSO stock solution at 10 mM. The reaction mixture was incubated for 3 h at room temperature. Purification of the dye-protein conjugate was performed using 100 kDa MWCO spin concentrators and 3 rounds of washing with 10 mM phosphate buffer pH 7.2.

Synthesis of Mal-PEG5kDa-NP. Maleimide-PEG5kDa-amine (10 mmoles) was dissolved in 1 mL of DCM. To this solution, 35 equiv. of NHS-NP (from a 400 mM DMSO stock solution) and 10 mL triethylamine were added. The reaction was allowed to proceed at room temperature overnight. The solvent was removed using a rotary evaporator. Water was added to the yellow residue and white precipitate was observed. The precipitate was removed using a 0.22 mm syringe filter (PALL Corporation, Port Washington, NY). The filtrate was loaded onto a 3 kDa MWCO spin concentrator and spun at 4700 rpm for 20 min, followed by 3 rounds of washing with water. The retentate was subsequently used in the reaction with the VHP peptide.

Synthesis of VHP-PEG5kDa-AP. To a solution of Mal-PEG5kDa-NP prepared as described above, 10 mmoles of VHP peptide (1.5 mg, TFA salt) were added. Phosphate buffer at pH 8 was added to ensure that the pH of the reaction was appropriate for the Cys-maleimide reaction. The reaction was allowed to proceed overnight at room temperature, with shak-ing. To reduce the nitrophenol moiety to an aminophenol, a solution of 100 mM Na2S2O4 (freshly prepared) was used to a final concentration of 10 mM for 20 min at room tempera-ture. The product was purified using a 3 kDa MWCO spin concentrator and 3 rounds of washing with water. The extent of modification of the MS2 capsid was investigated using SDS-PAGE analysis and subsequent optical densitometry.

Synthesis of VHP-PEG5kDa-MS2-AF680. To a solution of MS2-AF680 obtained as de-scribed above, a solution of VHP-PEG5Da-AP (2 equiv.) was added. Sodium periodate at a final concentration of 1 mM was used as an oxidant. The reaction was allowed to proceed at room temperature for 5 min. The excess oxidant and PEG were removed using a 100 kDa MWCO spin concentrator. Sterile PBS was used to wash the VHP-PEG5kDa-MS2 con-jugate and sterile spin filters with 0.22 mm pores were used to ensure sterility.

Synthesis of VHP-AF680. VHP peptide (0.83 mg, 545 nmoles) was dissolved in 1 mL of 100 mM phosphate buffer pH 7.5. To this, 0.25 equiv. of AF680-maleimide stock solu-tion in DMSO (10 mM) was added and the reaction was left to stand overnight. The condi-tions were selected in the hopes of limiting the side-reactions of the maleimide moiety with the Lys residues in the peptide sequence. The reaction mixture was loaded on a 0.5 cc SepPak C18 column (Waters Corporation, Milford, MA) and the product was eluted with a gradient of 5-30% acetonitrile in 0.1% TFA/water. The fractions containing the desired product were evaporated to dryness and resuspended in sterile PBS. Illustra Microspin G-25 columns (GE Healthcare, Marlborough, MA) were used sequentially to remove any TFA present. Sterile PBS and spin filters with 0.22 mm pores were used to ensure sterility. The purity of the VHP-AF680 product was confirmed using ESI TOF LCMS and MALDI.

Synthesis of αVCAM-AF680. To a solution of antibody at 0.5 mg/mL (~3 M) in borate buffer pH 8 (as provided by the supplier) were added 4 equiv. from a stock of AF680-NHS

72

in DMSO (2 mM). The reaction was performed at room temperature for 2.5 h and the product was purified using 30 kDa MWCO spin concentrators. To note, in our experience, the recovery of antibodies is low when using 100 kDa MWCO spin concentrators. About 2 dyes / antibody were attached using these reaction conditions (based on NanoDrop readings). Sterile PBS was used to wash the product and the retentate was filtered using sterile spin filters with 0.22 mm pores to ensure sterility.

Synthesis of αVCAM-MS2-AF680. To a solution of anti-VCAM antibody in borate buf-fer at pH 8 (as provided by supplier) was added a stock solution of NHS-NP in DMSO (10 equiv.). The reaction was performed at room temperature for 2.5 h. The nitrophenol moiety was reduced using 100 mM Na2S2O4 to a final concentration of 10 mM. Excess reagents were removed using 30 kDa MWCO spin columns and washing 3 times with 10 mM phosphate buffer pH 7.2. To the solution of the antibody-aminophenol obtained was added a solution of MS2-AF680 in a 3:1 antibody:capsid ratio. Sodium periodate at a final concentration of 1 mM was used as oxidizing agent. The reaction proceeded at room tem-perature for 5 min and the product was purified using 100 kDa MWCO spin concentrators. Sterile PBS was used to wash the antibody-MS2 conjugate and sterile spin filters with 0.22 mm pores were used to ensure sterility.

Ex vivo fluorescence imaging. All animal procedures were performed according to a protocol approved by the UCSF Institutional Animal Care and Use Committee (IACUC). Female ApoE knockout (B6.129P2-Apoe<tm1Unc>/J) or control (C57BL/6J) mice were purchased from The Jackson Laboratory (Bar Harbor, ME). The mice were used at the age of 19 weeks, after being fed a high fat diet (Teklad TD.88137, Envigo, Indianapolis, IN) for 15 weeks. The animals were euthanized by cardiac puncture, then perfused with 10 mL each of PBS and NBF. The aorta and main organs (liver, spleen, pancreas, kid-neys, heart and fat) were harvested and imaged using an IVIS Xenogen 50 with Living Image 3.2 software or IVIS Spectrum imaging system with Living Image 4.2 (Perkin El-mer, Waltham, MA). The intestines and stomach were not harvested due to high autoflu-orescence of the tissue and chow.

Excess tissue was removed from the aorta under a dissecting microscope (VisiScope 350, VWR, Radnor, PA), while pinned (Minutien pins, Fine Science Tools, Foster City, CA) on a polymer support. The support was made by mixing a Sylgard 148 silicone elastomer base and curing agent (Dow, Midland, MI) in a 10:1 ratio in a Petri dish (Corning, Corning, NY), removing the bubbles in a vacuum dessicator and curing at 60 °C on a hot plate for 1 h. IVIS was used to collect information about the macroscopic localization of the agents injected. The blood vessels harvested were fixed in neutral buffered 10% formalin, submerged in 30% sucrose at 4 °C for 2-5 days, then segmented into 0.5-1 cm portions and embedded in Tissue Tek Optimal Cutting Temperature medium (Sakura Finetek, Tor-rance, CA). Slices (10 mm thick) were obtained at the UCSF Preclinical Mouse Pathology Core and were stained using standard protocols for Hematoxylin and Eosin (H&E) and Oil Red O (ORO). Confocal images were obtained using a Zeiss LSM 710 AxioObserver confocal microscope with a 635 nm laser.

73

4.8 References

1. Writing Group, M., Mozaffarian, D., Benjamin, E.J., Go, A.S., Arnett, D.K., Blaha, M.J., et al., Heart Disease and Stroke Statistics-2016 Update: A Report From the American Heart Association. Circulation. 2016, 133, e38-360.

2. Libby, P., Inflammation in atherosclerosis. Nature. 2002, 420, 868-874.

3. Yamashita, A. and Asada, Y., A rabbit model of thrombosis on atherosclerotic lesions. J Biomed Biotechnol. 2011, 2011, 424929.

4. Libby, P., Ridker, P.M. and Hansson, G.K., Progress and challenges in translating the biology of atherosclerosis. Nature. 2011, 473, 317-325.

5. Choudhury, R.P., Fuster, V. and Fayad, Z.A., Molecular, cellular and functional imaging of atherothrombosis. Nat Rev Drug Discov. 2004, 3, 913-925.

6. Tatsumi, M., Cohade, C., Nakamoto, Y. and Wahl, R.L., Fluorodeoxyglucose uptake in the aortic wall at PET/CT: possible finding for active atherosclerosis. Radiology. 2003, 229, 831-837.

7. Helft, G., Worthley, S.G., Zhang, Z.Y., Tang, C., Rodriguez, O., Wei, T.C., et al., Non-in-vasive in vivo imaging of atherosclerotic lesions with fluorine-18 deoxyglucose (18-FDG) PET correlates with macrophage content in a rabbit model. Circulation. 1999, 100, 311-311.

8. Botnar, R.M., Stuber, M., Danias, P.G., Kissinger, K.V. and Manning, W.J., Improved coronary artery definition with T2-weighted, free-breathing, three-dimensional coronary MRA. Circulation. 1999, 99, 3139-3148.

9. Botnar, R.M., Stuber, M., Kissinger, K.V., Kim, W.Y., Spuentrup, E. and Manning, W.J., Noninvasive coronary vessel wall and plaque imaging with magnetic resonance imaging. Circulation. 2000, 102, 2582-2587.

10. Fayad, Z.A., Fuster, V., Fallon, J.T., Jayasundera, T., Worthley, S.G., Helft, G., et al., Noninvasive in vivo human coronary artery lumen and wall imaging using black-blood magnetic resonance imaging. Circulation. 2000, 102, 506-510.

11. Hamilton, A.J., Huang, S.L., Warnick, D., Rabbat, M., Kane, B., Nagaraj, A., et al., Intravascular ultrasound molecular imaging of atheroma components in vivo. J Am Coll Cardiol. 2004, 43, 453-460.

12. Zhao, X.Q., Yuan, C., Hatsukami, T.S., Frechette, E.H., Kang, X.J., Maravilla, K.R., et al., Effects of prolonged intensive lipid-lowering therapy on the characteristics of carotid atherosclerotic plaques in vivo by MRI - A case-control study. Arteriosclerosis Thrombosis and Vascular Biology. 2001, 21, 1623-1629.

13. Kubota, K., Furumoto, S., Iwata, R., Fukuda, H., Kawamura, K. and Ishiwata, K., Comparison of 18F-fluoromethylcholine and 2-deoxy-D-glucose in the distribution of tu-mor and inflammation. Ann Nucl Med. 2006, 20, 527-533.

74

14. Lippert, J.A., White, C.S., Mason, A.C. and Plotnick, G.D., Calcification of aortic valve detected incidentally on CT scans: prevalence and clinical significance. AJR Am J Roent-genol. 1995, 164, 73-77.

15. Zhao, X.Q. and Kerwin, W.S., Utilizing imaging tools in lipidology: examining the po-tential of MRI for monitoring cholesterol therapy. Clin Lipidol. 2012, 7, 329-343.

16. Teng, F.F., Meng, X., Sun, X.D. and Yu, J.M., New strategy for monitoring targeted therapy: molecular imaging. Int J Nanomedicine. 2013, 8, 3703-3713.

17. O’Brien, K.D., Allen, M.D., McDonald, T.O., Chait, A., Harlan, J.M., Fishbein, D., et al., Vascular cell adhesion molecule-1 is expressed in human coronary atherosclerotic plaques. Implications for the mode of progression of advanced coronary atherosclerosis. J Clin Invest. 1993, 92, 945-951.

18. Galkina, E. and Ley, K., Vascular adhesion molecules in atherosclerosis. Arterioscler Thromb Vasc Biol. 2007, 27, 2292-2301.

19. Tarkin, J.M., Joshi, F.R. and Rudd, J.H., PET imaging of inflammation in atherosclero-sis. Nat Rev Cardiol. 2014, 11, 443-457.

20. Obermeyer, A.C., Capehart, S.L., Jarman, J.B. and Francis, M.B., Multivalent viral capsids with internal cargo for fibrin imaging. PLoS One. 2014, 9, e100678.

21. Nakashima, Y., Raines, E.W., Plump, A.S., Breslow, J.L. and Ross, R., Upregula-tion of VCAM-1 and ICAM-1 at atherosclerosis-prone sites on the endothelium in the ApoE-deficient mouse. Arterioscler Thromb Vasc Biol. 1998, 18, 842-851.

22. Barros, M.H., Hauck, F., Dreyer, J.H., Kempkes, B. and Niedobitek, G., Macrophage polarisation: an immunohistochemical approach for identifying M1 and M2 macrophages. PLoS One. 2013, 8, e80908.

23. Nahrendorf, M., Jaffer, F.A., Kelly, K.A., Sosnovik, D.E., Aikawa, E., Libby, P., et al., Noninvasive vascular cell adhesion molecule-1 imaging identifies inflammatory activation of cells in atherosclerosis. Circulation. 2006, 114, 1504-1511.

24. Kelly, K.A., Nahrendorf, M., Yu, A.M., Reynolds, F. and Weissleder, R., In vivo phage display selection yields atherosclerotic plaque targeted peptides for imaging. Mol Imaging Biol. 2006, 8, 201-207.

25. Kelly, K.A., Allport, J.R., Tsourkas, A., Shinde-Patil, V.R., Josephson, L. and Weissled-er, R., Detection of vascular adhesion molecule-1 expression using a novel multimodal nanoparticle. Circ Res. 2005, 96, 327-336.

26. Weissleder, R., Stark, D.D., Engelstad, B.L., Bacon, B.R., Compton, C.C., White, D.L., et al., Superparamagnetic iron oxide: pharmacokinetics and toxicity. AJR Am J Roentge-nol. 1989, 152, 167-173.

75

Chapter 5: Viral capsids for glioblastoma detection and treatment

The work described in this chapter was performed in collaboration with: Tiras Lin and Preyas Shah (Shaqfeh lab, Stanford, computational work), Tomoko Ozawa, Raquel Santos and Edgar Lopez-Lepe (Nicolaides lab, UCSF, animal work), Sergio Wong and Tony Huynh (Seo lab, UCSF, PET imaging) and Joel Finbloom, Jenna Bernard, Sarah Klass, Jessica Lee, Alexander Hoepker and Rafi Mohammad (Francis Lab, agent synthesis).

Abstract

Nanoscale carriers present new opportunities to enhance delivery of therapeutic and imag ing cargo to cancer tissue. Potential benefits of such delivery vehicles include: improved efficiency of cancer cell targeting, reduced off-target effects, increased circulation times and reduced metabolic degradation. Using biomaterials is advantageous in drug delivery and imaging applications as these materials are nontoxic and are easily degraded by the body. Despite many advances in developing protein-based carriers, using these agents efficiently requires fundamental knowledge of how carrier size and shape influence in vivo biodistribution properties. To address this need, we prepared and characterized a series of well-defined protein-based agents and evaluated them for use in in vivo delivery applications. The scaffolds selected for investigation are protein assemblies that form spheres, disks and rods. Additionally, a flexible linear protein sequence (intrinsically disordered protein, IDP) was explored as a delivery scaffold. We decorated the external surfaces of these carriers with polyethylene glycol (PEG) chains in order to vary the size and charge of the particles and prolong their circulation in vivo. The carriers were further modified with Positron Emission Tomography (PET) tracers and fluorophores for biodistribution detection. Computational models were also developed to simulate the extravasation flux of these particles. Findings from this study will be applicable to developing many protein-based agents in the field of nanomedicine.

76

5.1 The effect of nanoparticle properties on tumor uptake

Cancer-related deaths are still very common, although the last decades have brought numerous advances in surgical, radiological, and chemotherapeutic treatments.1 Earlier detection of cancerous formations and metastatic sites through more sensitive imaging techniques could greatly improve survival rates.2 As discussed in previous chapters, us-ing carriers that can deliver chemotherapeutic agents selectively to cancerous cells is believed to have the potential to treat identified tumors in a more efficient manner,3-5 by reducing side effects and allowing the use of higher doses of chemoterapeutics.

Numerous interactions must be considered while investigating the behavior of a par-ticle introduced in a living system: serum proteins can coat the nanoparticle and mark it for clearance; the immune system can recognize the particle as foreign; the particle can extravasate through the vascular walls; the interstitial pressure present in the tumor can affect the diffusion of the particle.6 These interactions depend greatly on the physico-chemical properties of the particle and thus, in recent years, more focus has been placed on studying the influence that the size, shape and flexibility have on the performance of a nanocarrier.7-10

5.1.1 Size

Studies investigating the cellular uptake of nanoparticles have observed that there are differences in the rate of uptake based on the size of the particles. For ex ample, Chithrani et al. demonstrated that gold nanoparticles in the regime of 50 nm had optimized tumor cell uptake,11 while Liu and coworkers observed a similar result for liposomes with diam-eters between 100 and 200 nm.12

5.1.2 Shape

An increasing number of scientists are studying the effect of morphology on cellu-lar uptake and biodistribution. In a study by Gratton et al., spherical and rod-shaped PEG-based hy drogels were compared and it was observed that nonspherical particles were more frequently internal ized and endocytosed through different mechanisms than spherical particles.13 Steinmetz and coworkers also compared rod-shaped and spheri-cal nanoparticles and reported enhanced tumor accumulation of the rod-shaped carriers over their spherical counterparts in mouse xenograft models.14 Therefore, while spherical nanoparticles have been studied more extensively, other shapes present great promise for favorable biodistribution profiles.

5.1.3 Flexibility and diffusivity

Engineering flexible drug carriers has been recently attempted as a strategy to alter in vivo behavior. For example, Discher and coworkers prepared filamentous polymeric mi-celles (“filomicelles”) and ob served that they exhibited reduced uptake by macrophages, increased circulation time and enhanced tumor uptake when compared to rigid particles.10 The ability of nanoparticles to diffuse through a viscous environment can also affect their drug delivery efficiency. However, more advances have to be made in the techniques for measuring and controlling the diffusion coefficients of nanoparticles before this property

77

can be leveraged for obtaining the desired biodistribution.

5.1.4 Surface properties and PEGylation

The phar macokinetics, biodistribution, and cellular internalization of nanoparticles are greatly influenced by the overall charge of the carrier, as the charge will determine the interaction with serum proteins and with cell surfaces.15, 16 Most commonly, particles with large positive or negative surface charges interact with serum proteins and get cleared from blood rapidly.17, 18 On the contrary, near-neutrally charged scaffolds have increased circulatory half lives and fewer interactions with serum proteins and mac rophages.19, 20 Polyethylene glycol (PEG) is a ploymer that is frequently used as a strategy to shield the charge of nanoparticles and reduce their interactions with serum proteins and the reticu-loendothelial system.21, 22

5.1.5 Active targeting

The “active targeting” approach for drug deliv ery (discussed in Chapter 3) relies on the interactions between the carriers and certain molecular biomarkers, which are typically surface receptors overexpressed exclusively on cancer cells. By conjugating the appro-priate binding ligands, it is hypothesized that the nanocarriers that passively accumulate in the tumor environment will actively bind to cancer cells and will be endocytosed. Thus, targeted nanocarriers can offer both tumor tissue accumulation and enhanced cell uptake for drug delivery and imaging. In future work, we will investigate the active targeting of protein nanocarriers using well studied targeting ligands such as cyclical RGD (cRGD),23,

24 which have been shown to increase cell uptake via integrin receptor-mediated endocy-tosis.

5.2 Protein scaffolds compared in this study: MS2, TMV, nanophage and IDP

The scaffolds under investigation are the bacteriophage MS2, tobacco mosaic virus (TMV) double disks, nanoscale filamentous phage (nanophage), and intrinsically disor-dered proteins (IDPs) (Figure 5.1). These scaffolds were modified with Positron Emission Tomography (PET) imaging tracers, fluorescent dyes, and hydrophilic polymers (PEG) using bioconjugation strategies25 in order to obtain modular, robust and homogeneous delivery vehicles. The carriers were evaluated for their stability, cytotoxicity and internal-ization in cancer cell lines. Drug cargo was attached to the scaffolds to determine their potential as drug delivery agents. Our previous explorations of genome-free MS2 viral capsids established a well-defined experimental workflow for these studies.26-28

5.2.1 MS2 bacteriophage

As presented in previous chapters, the MS2 bacteriophage has many qualities that make it a good candidate for an imaging agent or a drug delivery scaffold. The assays used to evaluate MS2 were adapted for the other systems and the workflow was greatly enhanced by the experience gained in the experiments presented in previous chapters.

5.2.2 Double disks from the tobacco mosaic virus coat protein

The wild type tobacco mosaic virus (TMV) assembles into a helical rod-shape (300 nm)

78

templated by the genomic RNA.29 TMV rods have been used by other research groups as imaging and drug delivery agents and have been studied in preliminary animal mod-els.30-32 When wild type TMV is recombinantly expressed, several species are observed, including small protein complexes, a 34-monomer double disk intermediate, stacks of double disks, and long helical rods that can be microns in length.29, 33 The proportion of each species is greatly dependend on the chemical environment (such as pH and ionic strength), which can cause complications for in vivo studies. Through previous work in our group, we discovered that the TMVK53R/K68R double mutants form stable double disks (referred to as TMV disks herein).34 We later engineered aditional modification handles, such as a cysteine residue at the 123 position and an AG extension of the N-terminus. The disks are composed of 34 protein monomers (17 in each disk layer) and are 18 nm in diameter and 5 nm in height. As opposed to the wild type TMV, the disk structures are stable in a wide pH range and remain assembled even after chemical modification. Disk-shaped nanocarriers have not been extensively studied and we believe that the TMV disks could have interesting properties that can be beneficial for drug delivery.

5.2.3 Filamentous nanophage

Filamentous phage have gained notoriety for their application in phage display tech-niques, although several labs have investigated their potential to serve as targeted imag-ing agents and drug carriers.35-40 The large size of the phage capsids (>900 nm) might lead to a fast clearance from the system, with reported half-lives on the order of minutes.41 Work performed in the Rakonjac lab developed a “nanophage” scaffold derived from fil-amentous phage by leveraging the “rolling circle” mechanism used by the phage during replication.42, 43 The truncated single stranded DNA is then packaged by the machinery of the full-length phage to produce shorter particles (50 nm). These particles have the same protein types as the wild type phage, but only contain ~100 copies of the pVIII protein, as

bacteriophage MS2(sphere)

50 nmnanophage

TMV mutant(disk)

Nanophage(rod)

IDP(flexible strand)

50 nm(fully extended)

18 nm

5 nm

27 nm

50 nm

Figure 5.1 Protein-based scaffolds with different shapes for in vivo delivery applications. TMV = tobacco mosaic virus, IDP = intrinsically disordered protein. The sites of chemical modification are indicated as follows: red = cysteines, green = N-termini, brown = lysines, purple = artificial amino acids. The rendered protein structures and nanophage cartoon are shown to scale. The TEM images are stained with UO2(OAc)2 and are not shown to scale.

~7 nm

p8

p3

79

opposed to ~2700 for the filamentous phage.

5.2.4 Intrinsically disordered proteins (IDPs)

Intrinsically disordered proteins do not possess a well-defined structure under most conditions due to the abundance of disorder-promoting amino acids such as Gln, Ser, Glu, Lys, and Pro.44-46 We have designed a human-derived IDP sequence as a 166 amino acid-long monodisperse scaffold (AWRGSPWAEA KSPAEAKSPAEVKSPAVA KSPAEVK-SPAEVKSPAEA KSPAEAKSPAEVKSPATV KSPGEAKSPAEAKSPAEV KSPVEAKS-PAEAKSPASV KSPGEAKSPAEAKSPAEV KSPATVKSPVEAKSPAEV KSPVTVKS-PAEAKSPVEV KSPYWCA). This sequence has 25 lysine residues to be used as sites for chemical modification. Various modification degrees and strategies might lead to an array of scaffolds with morphologies that range from a disordered strand to a flexible rod. This would allow us to fine tune the physical properties and access scaffolds with distinct serum stability and circulation times.

5.3 In silico studies of nanoparticle extravasation

Computer models that explain and, eventually, predict parameters that maximize ac-cumulation in tumor tissue could aid the design of new carriers for cancer delivery. Our collaborators in the Shaqfeh group are working to develop simulations of nanoparticle behavior in blood flow to predict the extravasation of carriers through “pores” in tumor blood vessels.47, 48 Characterizing the scaffolds presented above will provide physical data for nanocarriers with different shapes and dimensions, and these measurements will be incorporated into computational models. Predictions of particle flux into tumor tissue will then be compared to the actual in vivo data with the goal of validating computational

0 0.5 1.0 1.5 2.00

0.5

1.0

1.5

2.0

2.5

3.0

Flow shear−rate (P) →

Flux

(S) →

rods

spheres

rods

spheres

rods

spheres

rods

spheres

d/B = 0.1

d/B = 0.3

d/B = 0.6

d/B = 1.0

Figure 5.2 In silico studies of nanoparticle extravasation. (a) Diagrams of the particles investigated in the preliminary experiments: sphere with the dimensions of the MS2 capsid, a disk representing the TMV and a rod representing the nanophage. Given the high flexibility of the IDP scaffold, modeling its interaction with the pores will be more challenging. (b) Simulated entry of nanoparticles into tumors through pores of differ-ent sizes. The extravasation fluxes of rigid rods (square markers) and spherical particles (circular markers) were compared as a function of flow shear-rate. Four ratios of particle size (d, representing the long axis for rods or the diameter for spheres) to pore diameter (B) were considered. Note that rods significantly outper-form spheres at all flow rates and pore sizes. Graphics courtesy of the Shaqfeh lab.

t

r

blood flow

Sphere pore

t

r

t

r

(a) (b)

t

r

Disk

Rod

80

models, and, later, as a way to explain observed behavioral differences.

The extravasation of nanoparticles through realistic tumor pore geometries is sim-ulated through the development of a series of novel computational models that will in-corporate information about the Brownian motion of the particles.49-51 Our initial goal is to determine whether extravasation through endothelial pores is dominated by thermal motion with cross shear over a pore, or if a convective component into the pore plays a more important role. Pore sizes can range from 10 nm to 2 mm, and vary significantly from mammary carcinomas (largest pores) to brain tumors (smallest pores).52 Initial ex-periments have examined finite-size spheres and rods (corresponding to the MS2 and the na nophage, respectively) extravasating through circular pores of different sizes in a membrane model. For these experiments, the shear was varied over a range of Péclet numbers.53 A depiction of the geometry used in the simulations is shown in Figure 5.2a. The results of these simulations show that there are very strong differences between the fluxes predicted for rods and spheres in the same flow because of the different geometric ratios and Péclet numbers (Figure 5.2b).

5.4 Synthesis and characterization of agents

The agents presented in this chapter can be expressed recombinantly in E. coli. By using an array of chemical reactions, we have modified the scaffolds with dyes, chelators, chemotherapeutic molecules and PEG chains. The MS2 bacteriophage can be readily modified using the T19paF unnatural amino acid residue and the N87C engineered cys-teine, as shown in previous chapters. The TMV scaffold can be modified either at the N-terminus, at the Cys residues or at the Lys residues. The nanophage can be modified by leveraging the Lys residues present on the coat proteins, especially the pVIII protein. In some cases, the amine moieties were transformed into thiols by using SPDP linkers (sulfosuccinimidyl 6-(3’-(2-pyridyldithio) propionamido) hexanoate) or Traut’s reagent (2-iminothiolane).54 The IDP scaffold has 25 Lys residues that can be modified directly or through the use of SPDP and further maleimide modification. In addition, the N-terminus is readily available for modification.

5.5 Positron Emission Tomography in glioblastoma models

Glioblastoma remains one of the most aggressive forms of cancer, with poor patient prognosis and limited treatment options.55 Due to the blood brain barrier, standard meth-ods of delivering drugs to the brain by intravenous infusions of systemic drugs usually result in limited penetration of the central nervous system.56 Recent studies have inves-tigated more localized approaches, such as intratumoral injections, to increase the effi-ciency of delivery to brain tumor cells.57 Convection-enhanced delivery (CED) is a novel delivery technique designed to bypass the blood brain barrier and administer therapeutic agents directly into the targeted brain tissue by administering the therapeutic agents with a microinfusion pump.58-60 As the pressure is maintained during injection, it creates fluid convection to supplement diffusion through the extracellular spaces and enhance the distribution of the drug to the targeted area. Thus, CED has the potential to deliver an agent: homogenously; to a larger volume of brain tissue; at higher drug concentrations; and using molecules that do not normally cross the blood brain barrier.

81

Early detection of brain tumors is important to improve survival rates for patients af-flicted by this disease. Historically, a small fraction of the agents injected IV reach the brain or brain tumors. Therefore, for initial studies on the efficiency of the scaffolds pre-sented above to detect brain tumors, we chose PET as a sensitive, non-invasive imaging technique. To our knowledge, this was also the first radiolabeling instance of the TMV and nanophage scaffolds.

5.5.1 Synthesis of agents for PET

The protein scaffolds MS2, TMV and nanophage were modified with chelators for 64Cu for use in PET imaging experiments and determination of biodistribution. As discussed in Chapter 2, NOTA was chosen as a chelator due to its increased ability to stay associat-ed with the 64Cu radioisotope under physiological conditions. Polyethylene glycol chains (PEG5kDa) were used to increase the circulation time of the TMV disks and nanophage as-semblies, as there was no previous information about their circulation profile in a mouse model. The methods used to modify the three scaffolds are summarized in Table 5.1 The characterization of the agents is presented in Figure 5.3a. In accordance with previous experiments, the radioactive yield for MS2 was 87% and the specific activity (amount of radioactivity per mole of agent) was 479 mCi/nmole or 86 mCi/nmole capsid. TMV and nanophage were labeled for the first time and the protocols were not optimized as in the case of MS2. The radioactive yield for TMV was 65%, with a final specific activity of 357 mCi/nmole or 12.1 mCi/nmole capsid. For nanophage, the radioactive yield was the low-est, with only 13%. The specific activity was 25 mCi/nmole or 2.5 mCi/nmole capsid. Fig-ure 5.3b shows efficient purification of the capsids from free 64Cu. Serum stability studies of radioactively-labeled TMV and nanophage incubated with mouse serum at 37 °C are described in the Appendix.

5.5.2 PET and biodistribution studies

Nude mice were implanted intracranially with U87-Luc glioblastoma cell line. This cell line contains a luciferase reporter that allows for monitoring the size of the tumor by mea-suring the bioluminescence signal when luciferin is injected and was used for all subse-

Chelatorattachment

Sample Chelatorsmod. (%)

PEGper capsid

Chelatorper capsid

PEGattachment

PEGmod. (%)

C87 (mal)MS2 100 126180 paF19 (OC) 70

C123 (mal)TMV 100 834 N-term S (ox, alk) 25

Lys (RSCN)Nanophage 25 1924 N-term (RS, alk) 20

Table 5.1 Synthesis of agents for Positron Emission Tomography. The site of modification and residue used for attachment are listed for each of the nanoparticles. The extent of modification was quantified using MALDI or ESI-TOF for chelator modification (NOTA) and by optical densitometry on SDS-PAGE gels for the PEG attachment. The abbreviations for modification strategies are as follows: mal = maleimide-thiol reac-tion; RSCN = isothiocyanate; OC = oxidative coupling with NaIO4; ox = periodate oxidation of Ser residues to form oximes; alk = alkoxyamine reaction with oxime; RS = transamination with Rapoport’s salt).

82

quent in vivo experiments. When tumors reached a significant size (26 days post-implan-tation, ~109 photons/s total flux), the mice were injected intravenously (IV) with 50-200 mCi radiolabeled agents. One mouse per group was imaged using a PET-CT scanner. A dynamic scan was recorded from time of injection until 1 h post-injection. At 5 h post-in-jection, a static scan was performed and the images are presented in Figure 5.3c after decay correction and normalization to the injection dose. The majority of the signal is lo-calized to the internal organs for all three carriers. Surprisingly, the TMV appears to local-

MS2

Nanophage

TMV

free Cu

blood

heart

lungs liv

er

kidne

ys

splee

n

panc

reas

lg int

estin

e

sm in

testin

e

stomac

h

muscle tai

l

left b

rain

right

brain

tumor

0

5

10

15

20

25

30

35

%ID

/g

MS2TMVNanophage

MS2 TMV Nanophage

Figure 5.3 Positron Emission Tomography with protein-based nanoparticles of different shapes injected intravenously (IV). (a) SDS-PAGE analysis of the protein scaffolds modified with a NOTA chelator and PEG5kDa polymer chains for improved circulation and reduced immunogenicity. Lanes 1 and 2 show the MS2 capsid before and after modification. Lanes 3 and 4 have the TMV disks starting material and PEGylated construct, respectively. Lanes 5 and 6 have the nanophage starting material and PEGylated scaffold. To note, both the p8 and p3 coat proteins are modified using the protocol used. (b) HPLC SEC traces showing the radioactivity channel to monitor the purity of the radiolabeled agents. No peaks corresponding to 64Cu not chelated to the protein scaffolds are observed. (c) Positron emission tomography imaging of mice bear-ing brain tumor models injected with protein scaffolds of different shape. Static images were taken at a 5 h timepoint and a representative coronal plane is shown. All doses were decay corrected and images were normalized to the injected dose. (d) Biodistribution information for radiolabeled PEGylated MS2, TMV and nanophage scaffolds. At 5 h post-injection, organs were harvested and weighed and their radioactivity was measured to determine the percent injected dose per gram tissue (%ID/g). The TMV disks had a large pres-ence in the blood, on par with MS2, indicating a long circulation time. Other organs with high levels of accu-mulation include the liver, spleen and kidneys. The nanophage had a higher uptake in the intestines com-pared to the other two scaffolds. The nanophage also had the largest accumulation in the tumor, but the large variability between the replicates did not lead to statistically significant differences when compared to the MS2 and TMV scaffolds. The fact that the nanoparticles did not accumulate in the normal brain tissue could imply a specific uptake into the tumor environment and is promising for mitigating side effects of drug molecules.

MW 1 2 3

20

15

10

30

40

50

60

3.5

kDa4 5 6

TMV

pVIII

pVIII-PEG5kDa

MS2-PEG5kDa

TMV-PEG5kDa

MS2

pIIIpIII-PEG5kDa

(a) (b)

(d)(c)

Retention time (s)0 200 400 600

tumor

intestines

liver

heart

tumortumor

bladder

Rad

ioac

tivity

1.7

14.5

%ID

/g

coro

nal

83

ize to the heart, suggesting a large presence in the blood pool. A small amount of agents is also localized to the bladder. There is a significant amount of signal present in the brain tumor in the nanophage image set, followed by TMV and MS2. To note, the healthy part of the brain does not seem to have a large amount of agent in either case.

At 5 h post-injection, the organs were harvested and weighed and their radioactivity was measured to determine the percent injected dose per gram (%ID/g). The results are presented in Figure 5.3d. As seen in the PET images, the TMV signal in the blood pool is similar to that of MS2. The liver and spleen also contained large amounta of all agents tested. The nanophage carrier resulted in significant accumulation in the intestines. The rod-shaped nanophage had the highest average amount of signal in the tumor out of the three agents investigated, although there was a large variation between replicates and statistical significance could not be proved.

5.6 Optical imaging in glioblastoma models

In order to gain more information about the localization of the agents within the tumor and the brain, we used dye-protein conjugates and optical imaging. The agents were la-beled with AlexaFluor 680 dyes to limit tissue autofluorescence that is usually observed at lower wavelengths. To note, initial efforts focused on Cy5.5, but issues with solubility and removal of excess dye urged us to reconsider the use of this dye for further experiments. Additionally, when IV administration was used for optical experiments, the signal was very low and did not allow to investigate the distribution of the agents in the tumor and brain tissues.

5.6.1 Synthesis of agents for optical imaging

Table 5.2 summarizes the synthesis and composition of the agents used for optical imaging experiments. A similar amount of dyes per particle was used in order to have a similar range of signal intensities.

5.6.2 Optical imaging after CED administration of fluorescently-labeled carriers

Agents were administered to nude mice bearing U87-Luc intracranial tumors at 28

Dyeattachment

Sample Dyemod. (%)

PEGper capsid

Dyeper capsid

PEGattachment

PEGmod. (%)

C87 (mal)MS2 12 12622 paF19 (OC) 70

C123 (mal)TMV 50 1717 N-term S (ox, alk) 50

Lys (NHS)Nanophage 10 4810 N-term (RS, alk) 50

Table 5.2 Synthesis of agents for optical imaging with fluorescently-labeled agents administered via convection enhanced delivery (CED). The site of modification and residue used for attachment are listed for each of the nanoparticles. The extent of modification was quantified using MALDI or ESI-TOF for dye modi-fication and by optical densitometry on SDS-PAGE gels for the PEG attachment. Modification with dyes was designed to achieve similar fluorescent signal per particle. The abbreviations for modification strategies are as follows: mal = maleimide-thiol reaction; NHS = N-hydroxysuccinimide ester reaction with amines; OC = oxidative coupling with NaIO4; ox = periodate oxidation of Ser residues to form oximes; alk = alkoxyamine reaction with oxime; RS = transamination with Rapoport’s salt).

84

days post-implantation. The agents were dosed at high concentrations (in a vol-ume of 10 mL) through a CED setup. At 5 h post-injection, the brains were harvest-ed and imaged using an IVIS imaging system. The distribution of the agents can be seen in representative images in Figure 5.4. All images are shown at their individual optimal count range. We used a free dye control to mimic the distribu-tion of a drug after CED administration. Preliminary experiments suggest that the dye was distributed throughout the brain tissue, whereas the nanocarriers were primarily retained in the tumor tis-sue. This indicates a faster diffusion to the normal tissue for the small molecule dye as compared to the nanocarriers.

5.7 Drug delivery studies in glioblastoma models

In order to investigate the potential of the carriers to reduce the tumor burden by effi-ciently delivering drugs to the tumor site, we synthesized a series of conjugates with the chemotherapeutic molecule Doxorubicin (Dox). Doxorubicin is an anthracycline antibiotic commonly used for the treatment of many types of cancer. However, its systemic toxicity (especially cardiac toxicity) and poor penetration through the blood-brain barrier limit its use in the treatment of brain tumors. Several different formulations of doxorubicin have been developed, including PEGylated liposomal doxorubicin (Doxil). Unfortunately, none of these agents showed activity in clinical trials investigating its use to treat brain can-cer. We hypothesized that the protein carriers can alter the circulation profile of the Dox molecule (for IV administration) or the diffusion and clearance in the case of CED admin-istration. As seen in the optical imaging experiments, small molecules can diffuse more freely through the brain tissue and reach the healthy part of the brain, possibly inducing serious side effects. By attaching the Dox molecule to the protein scaffolds through an acid-labile linker, the release and distribution of Dox can be controlled and higher doses can be administered.

5.7.1 Synthesis of agents for drug delivery applications

The ketone moiety on Doxorubicin was modified with N-e-maleimidocaproic acid hy-drazide (EMCH), as previously reported (Figure 5.5a). This modification was shown to retain the cytotoxic activity of the Doxorubicin molecule and the resulting hydrazone can be hydrolyzed at pH 4.5-5.5 to allow release of the drug cargo from the carrier. The ma-leimide moiety of Dox-EMCH was reacted with the Cys residues on MS2 (internal Cys at position 87) and TMV (position 123). For the nanophage and IDP scaffolds, the native Lys residues were first converted to thiols using sulfo-SPDP or Traut’s reagent. The strategies employed to synthesize the protein-Dox conjugates are summarized in Table 5.3. PEG

Figure 5.4 IVIS images of the brains of mice injected with fluorescently labeled protein scaffolds by convection enhanced delivery (CED). At 5 h post- injection, the brains were harvested and imaged. Representative images from a cohort of 4 mice are shown. To note, the free dye control is more widely distributed in the brain, likely reaching the healthy brain tissue. Each image is normalized to its maxi-mum. Red represents higher signal, while blue represents a lower signal.

PBS free dye MS2 TMV Nanophage

high

low

injection site

85

chains were added to improve the solubility of the protein constructs, especially at the high concentrations required by the CED mode of administration. LC-MS analysis (Fig-ure 5.5b) of MS2-Dox shows the modification with Dox molecules, as well as the product of hydrolysis in the presence of 0.1% formic acid mobile phase used for ionization. Figure 5.5c shows the SDS-PAGE analysis of the MS2-Dox conjugate, with a clear shift in mo-lecular weight upon addition of the PEG chains. The UV-VIS spectrum (Figure 5.5d) has a small peak at 480 nm, indicating the presence of Dox molecules.

5.7.2 Cytotoxicity of Dox-protein conjugates

The protein conjugates were evaluated for their effects on cell viability. U87MG glioblas-toma cells were incubated with the agents at different concentrations for 72 h and an MTS (3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazoli-um)) viability assay was used to determine the percentage of live cells (Figure 5.5e). All nanocarriers show a similar dose-response curve and have comparable efficiency with the free drug, suggesting that the drug was released and was able to reach the nucleus and induce cell death. The 72 h time point might have been long enough to allow all the carriers to release their cargo, regardless of the amount of uptake into cells, which might have been different for each protein. Therefore, we looked at the effect of MS2-PEG5kDa-Dox at earlier time points (Figure 5.5f). We observed that the protein carrier itself is not toxic and that at early time points (8 and 24 h), free Dox is more efficient at killing cells, presumably since the Dox cargo has not yet been released from the MS2 carrier. Howev-er, at 48 h, the MS2 carrier outperforms free Dox.

5.7.3 Cellular uptake of Dox-protein conjugates

In order to better understand the kinetics of uptake of Dox-conjugates (characterized in Figure 5.6a) in glioblastoma cells, we used an Incucyte live cell imaging system. This setup allows monitoring the cells undisturbed for prolonged periods of time, since the optical setup is inside the incubator. Images were collected every hour for 48 h and the 1,

Doxattachment

Sample Doxmod. (%)

PEGper scaffold

Doxper scaffold

PEGattachment

PEGmod. (%)

C87 (mal)MS2 35 12663 paF19 (OC Fe) 70

C123 (mal)TMV 30 1710 N-term (OC Fe) 50

Lys (Traut, mal)Nanophage 20 1520 Lys (Traut, mal) 15

Lys (SPDP, mal)IDP 5 131 Lys (SPDP, mal) n/a

Table 5.3 Synthesis of agents for delivery of Doxorubicin (Dox). The site of modification and residue used for attachment are listed for each of the scaffolds. The extent of modification was quantified using UV-VIS for Dox modification and by optical densitometry on SDS-PAGE gels for the PEG attachment. The abbrevia-tions for modification strategies are as follows: mal = maleimide-thiol reaction; Traut = appending thiols on the Lys amines using Traut’s reagent; SPDP = appending thiols on the Lys amines using sulfo-SPDP; OC Fe = oxidative coupling with potassium ferricyanide).

86

24 and 48 h time points are presented in Figure 5.6b. The amount of fluorescence inside the cells was quantified by defining “green objects” using the Top Hat algorithm with a radius of 10 mm. We observed that the MS2 had the fastest and largest amount of uptake, followed by TMV, IDP and nanophage.

5.7.4 Tumor growth studies

Nude mice implanted intracranially with 3x105 U87MG-Luc glioblastoma cells were injected with Dox-protein conjugates at a Dox dose of 20 mg/kg mouse (~1 mg per mouse) in 10 mL via CED at 10 days post-implantation. To note, this dose is much smaller than Dox doses usually used for tumor growth inhibition studies (maximum tolerated dose for IV injections is 8 mg/kg). Given that the CED treatment is localized to the tumor site and

8 h 24 h 48 h

13779

13500 14000 14500 15000

13779

14004 14529

mass (Da)

MS2

MS2MS2-EMCH MS2-Dox A

bsor

banc

e

Wavelength (nm)

Doxorubicin (Dox) EMCH Dox-EMCH

Dox absorbance

MS2MS2-PEG5kDa-Dox

0

0.5

1

220 315 410 505 600

1 2 3

MS2 monomer

MS2-PEG5kDaDox

MS2-Dox

2520

15

10

kDaMW

Figure 5.5 Synthesis and characterization of Doxorubicin (Dox) conjugates and evaluation of their cytotox-icity in vitro. (a) Synthetic scheme for obtaining a maleimide-Dox with an acid-cleavable linker. (b) LC-MS analysis of the conjugation of Dox-EMCH to MS2 capsids. Note that the EMCH linker is acid labile and some hydrolysis occured during the LC run. (c) SDS-PAGE analysis of the conjugation of PEG via oxidative coupling with potassium ferricyanide. (d) UV-VIS analysis spectrum of the MS2-PEG5kDa-Dox construct, showing the distinctive 480 nm peak for Dox. (e) Viability assay investigating the cytotoxic effect of the nanocarriers conjugated to Dox after 72 h of incubation with U87 glioblastoma cells. Similar dose-response curves were observed for all three nanocarriers. (f) Investigation of the cytotoxicity of MS2-PEG5kDa-Dox at earlier time points. The protein carrier shows no toxicity to the U87 cells, while the Dox construct shows a smaller effect at early time points (8 and 24 h, blue and red bars), presumably due to the slow release of the drug from the capsid following the hydrolysis of the acid labile linker. All values are normalized to the untreated control cells for each time point.

(b)

(a)

(e) (f)

(c) (d)

Cel

l via

bilit

y (%

)

0

20

40

60

80

100

120

Cel

l via

bilit

y (%

)

0

20

40

60

80

100

120

140

1 10 100 1000 10000

Doxorubicin concentration (nM)

DoxMS2 DoxMS2-PEG5kDa-Dox MS2-PEG5kDa-Dox

TMV-PEG5kDa-DoxNanophage-PEG5kDa-Dox

O

O

O

OH

OH O

OHO

HO

NH2

OH

+ N

O

O

O

NH

NH2TFA, MeOHRT , 24 h

O

O

O

OH

OH O

OHN

HO

NH2

OH

5

HOHN

O

N

O

O5

HO

87

that there was no literature precedence for CED delivery of Dox in mouse models, we chose to start with a conservative dose regiment.

The cohorts for each treatment group contained 9 mice. A large variability was ob-served in the growth profiles for the mice in all groups tested (Figure 5.7). By averaging the tumor sizes for all the mice in one group, the information about this variability would

Figure 5.6 Characterization and cellular uptake of protein-Dox conjugates. (a) Lanes 1, 3, 5 and 7 show the MS2, TMV, nanophage and IDP starting material, respectively. Lanes 2, 4, 6 and 8 show the Dox and PEG conjugated constructs. Note that the high level of modification of the IDP leads to a very faint, broad band at a high molecular weight. (b) U87MG glioblastoma cells were incubated with Dox-protein conjugates at 37 ºC for 48 h. The fluorescence of the Dox molecules was monitored on the green channel of an Incucyte live-cell imaging system. The images were taken at 1, 24 and 48 h of incubation. The uptake of the protein scaffolds was determined as the increase in the number of “green objects” over time in arbitrary units (same scale used for all graphs).

(b)(a) 1 h 24 h 48 h

sign

al In

tens

ity

time

sign

al In

tens

ity

time

sign

al In

tens

ity

time

sign

al In

tens

ity

time

100 μm

MS2

TMV

Nanophage

IDP

MW 1 2 3

2015

10

25

37

50

75100150

250

kDa

4 5 6 7 8

TMV

p8

p8-PEG5kDa

MS2-PEG5kDa TMV-PEG5kDa IDP

IDP-PEG5kDa

MS2

Days post-implantation

PBS1011

109

107

105

6 14 26 34 42

Days post-implantation

Dox1011

109

107

105

6 14 26 34 42

Days post-implantation

MS21011

109

107

105

6 14 26 34 42

Days post-implantation

Nanophage1011

109

107

105

6 14 26 34 42

Days post-implantation

TMV1011

109

107

105

6 14 26 34 42

Days post-implantation

IDP1011

109

107

105

6 14 26 34 42

Days post-implantation

Liposomal Dox1011

109

107

105

6 14 26 34 42

Figure 5.7 Tumor growth after CED treatment with agents. Each curve represents one mouse from the cohort. A large variability between initial tumor sizes (at day 6) and response to treatment can be observed. Mice that benefited from the treatment are highlighted with red arrows.

CED treatment on day 11

Bio

lum

ines

cenc

e (B

LI)

88

have been lost, thus we chose to monitor each mouse individually. Although aver-age tumor sizes are not significantly dif-ferent for the Dox-protein conjugate-treat-ed mice as compared to the PBS control, a few mice (highlighted with red arrows in Figure 5.7) seemed to benefit from the treatment. Additionally, in the case of TMV, 3 mice lived much longer than the rest of the treated mice, including the Dox control group (Figure 5.8). We hypoth-esize that the low dose did not allow for sufficient cancer cell death and that the tumors recovered and continued their growth. Additionally, variability in the site of injection compared to the tumor loca-tion could greatly affect the efficiency of the agents. Future studies will focus on administering multiple, larger doses to achieve a more significant tumor shrink-age and increased survival.

5.8 Conclusions

We have synthesized a panel of agents with various physical properties in order to in-vestigate the effect of size and shape, as well as flexibility and surface properties, on the uptake and drug delivery efficacy of these nanocarriers in tumor models. Computational models are being developed to aid in the understanding of the parameters that are im-portant for the improved flux through tumor pores. A feedback loop between experimental data and the results of the simulations will improve the current models and lead to more efficient carriers.

We have radiolabeled for the first time the TMV disks and nanophage scaffolds with good specific activity and great purity. After IV injection, the nanophage protein cage was able to reach a brain tumor model implanted intracranially better than the TMV or MS2 agents. Variability between biological replicates limited the statistical difference between the three groups. Surprisingly, the TMV disks had a large presence in the blood pool 5 h post-injection, suggesting a long circulation time. As discussed in Chapter 4, this property could be used to target cardiovascular diseases or blood cancers.

Optical imaging experiments revealed that upon CED injection, the MS2, TMV and nanophage agents are localized mainly in the tumor and diffuse less to the healthy brain tissue than a free dye control. This observation has strong implications for the hypothesis that drug carriers could help reduce side effects of chemotherapy by increasing the spec-ificity to the diseased tissue over healthy tissue. In that case, higher doses or more potent therapeutic agents could be used to achieve a more efficient treatment.

Figure 5.8 Kaplan Meier survival plot for tumor growth study in the presence of Dox-protein nanocar-riers. The survival of treated mice was plotted over time. The TMV-treated group showed a promising increase in survival time, but upon statistical analysis, the difference compared to the PBS group was not significant. The shaded areas represent 95% confi-dence intervals for the PBS (blue) and TMV group (orange), respectively. The graphic was made using the Kaplan Meier Survival Curve Grapher.

PBS

Dox

Liposomal Dox

MS2

Nanophage

TMV

IDP

1

10 20 30 40

0.8

0.6

0.4

0

0.2

Prob

abili

ty

Days post-implantation

CED treatment

89

Preliminary drug delivery studies show promising results for the treatment of glioblas-toma models via CED injection. Even if the dose was much lower than in other studies, some mice in the Dox, MS2, TMV and nanophage groups benefited from the treatment. A large variability was observed within most cohorts. This could be explained by the differ-ence in the size/stage of the tumor at the time of the treatment or inconsistencies between the location of the CED injection with respect to the tumor. Future studies will focus on more efficient dose regiments (higher doses, repeated administration), as well as differ-ent PEGyation levels, which might affect the diffusivity and clearance of the particles.

5.9 Materials and methods

Unless otherwise noted, reagents were purchased from Sigma (St. Louis, MO) and used without further purification. AlexaFluor680 NHS, AlexaFluor680 maleimide, sulfo-LC-SPDP were purchased from Thermo (Waltham, MA). pSCN-Bn-NOTA (2-S-(4-Isothio-cyanatobenzyl)-1,4,7-triazacyclononane-1,4,7-triacetic acid) was purchased from Mac-rocyclics (Dallas, TX). NuPage 10% Bis-Tris Gels, MES SDS Running Buffer and Novex Sharp protein standard were purchased from Life Technologies. Water was deionized us-ing the NANOpure purification system (Thermo). DMSO was purchased as an analytical grade and was used without further purification. N‐methylpyridinium‐4‐carboxaldehyde benzenesulfonate hydrate (Rapoport’s salt, RS) was obtained from Alfa Aesar (Ward Hill, MA). NAP desalting columns were purchased from GE Healthcare (Marlborough, MA). Spin concentrators with different molecular weight cutoffs (MWCO) were from Millipore (Billerica, MA). Aminophenol-PEG5kDa-OMe, alkoxyamine-PEG5kDa-OMe and Dox-EMCH were synthesized as reported previously. Protein gel electrophoresis was performed as previously described in Chapter 2. EMEM media for cell culture was purchased from ATCC (Manassas, VA). MTS Assay ((3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxy-phenyl)-2-(4-sulfophenyl)-2H-tetrazolium)) was purchased from Promega (Madison, WI). Doxorubicin was purchased from Pfizer as a 2 mg/mL stock solution in saline. Liposomal Dox was purchased from SunPharma as a 2 mg/mL solution.

UV-VIS spectrophotometer readings were carried out using a Cary 50 Bio Spectro-photometer (Agilent, Santa Clara, CA) or a NanoDrop 1000 (Thermo Scientific). Liquid chromatography mass spectrometry (LC-MS) and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis were performed using the protocols described in Chapter 2. An Incucyte live cell imaging system (Essen Bioscience, Ann Arbor, MI) was used to monitor the cellular uptake of Dox-protein conjugates. An IVIS 50 Lumina imaging system (Perkin Elmer, Waltham, MA) was used to measure in vivo and ex vivo biolumi-nescence and fluorescence.

5.9.1 MS2

Protein expression and purification. The protocols for expression and purification of MS2 mutants are covered in Chapter 2.

Modification with NOTA chelator. A solution of 140 µM MS2 T19paF N87C in 10 mM phosphate buffer pH 7.2 was mixed with NOTA-maleimide from a stock in DMSO (10 equiv.) and 100 mM phosphate buffer pH 8 to reach a final monomer concentration of 100

90

µM. The reaction was incubated for 4 h at room temperature. The excess chelator was removed by using 100 kDa MWCO spin concentrators.

Modification with AF680 dye. A solution of 140 µM MS2 T19paF N87C in 10 mM phos-phate buffer pH 7.2 was mixed with AF680-maleimide from a stock in DMSO (0.15 equiv.) and 100 mM phosphate buffer pH 8 to reach a final monomer concentration of 100 µM. The reaction was incubated for 5 h at room temperature. The excess dye was removed by using 100 kDa MWCO spin concentrators.

Modification with PEG5kDa. After modification of the internal Cys residues, MS2 capsids were reacted with aminophenol-PEG5kDa-OMe (synthesized as described in Chapter 2). The reaction took place at room temperature for 5 min, in pH 7.2 buffer. The final con-centrations were 100 µM for the MS2 monomer, 500 µM PEG and 1 mM of NaIO4. The reaction mixture was loaded onto a Nap-5 desalting column and eluted with phosphate buffer pH 7.2. Subsequent concentration and additional removal of the excess reagents were obtained using 100 kDa MWCO spin concentrators.

Synthesis of MS2-PEG5kDa-Dox. A solution of 140 µM MS2 T19paF N87C in 10 mM phosphate buffer pH 7.2 was mixed with Dox-EMCH from a stock in DMSO (0.75 equiv.). The reaction was incubated for 3 h at room temperature. The excess Dox was removed by using Nap-25 desalting columns and 100 kDa MWCO spin concentrators. For the synthesis of the Dox conjugate, potassium ferricyanide was used to prevent the oxidation of the Dox molecule in the presence of periodate. The final concentrations were 100 µM for the MS2 monomer, 500 µM PEG and 5 mM of K3[Fe(CN)6]. The reaction was allowed to proceed for 1 h at room temperature. The reaction mixture was loaded onto a Nap-5 desalting column and eluted with phosphate buffer pH 7.2. Subsequent concentration and additional removal of the excess reagents were obtained using 100 kDa MWCO spin concentrators. The conjugate was stored frozen in the presence of 500 equiv. of trehalose until use to prevent Dox hydrolysis. Right before use, the stock was thawed and the treha-lose was removed using 100 kDa MWCO spin concentrators. For a study of the stability of MS2 during the freeze/thaw process in the presence and absence of trehalose, please see the Appendix.

5.9.2 TMV

The work presented in this section was performed by Joel Finbloom.

Protein expression and purification. The starting point for the RR-TMV protein was a gene for the coat protein of the TMV U1 strain optimized for the codon usage of E. coli (Genscript, Piscataway, NJ). Site-directed mutagenesis was performed using QuikChange mutagenesis (Stratagene, Santa Clara, CA). The RR-TMV coat protein was expressed to contain K53R, K68R, T104K, S123C, and SSYS N-terminal mutations. BL21 DE3 RIL Codon+ cells were transformed and cultured in Terrific Broth with 100 μg/L ampicillin at 37 °C. When cultures reached optical densities of 0.6 to 0.8, IPTG was added to a final concentration of 30 μM. Cultures were grown 24 h at 30 °C, harvested by centrifugation, and stored at -80 °C. Cells (from a 1 L expression batch) were thawed, resuspended in 20 mL of 20 mM TEA pH 8, and lysed by sonicating with a 2 s on, 4 s off cycle for a total of

91

30 min using a standard disruptor horn at 90% amplitude (Fischer Scientific). The result-ing lysate was cleared by ultracentrifugation for 30 min at 45,000 rpm using a Beckman 45 Ti rotor in an Optima L-80 XP (Beckman Coulter). The clarified lysate was decanted, warmed to room temperature, and stirred while adding a saturated solution of ammonium sulfate dropwise to a final concentration of 50% (v/v). After 5 min, the white precipitate that formed was pelleted by ultracentrifugation (30 min at 45,000 rpm in a Beckman 45 Ti rotor), and resuspended in 20 mM TEA pH 7.2. The resulting protein solution was next loaded onto a diethylaminoethanol (DEAE) Sepharose column and eluted with a 0 – 300 mM NaCl gradient. Purity was confirmed by SDS-PAGE and HPLC-SEC. This prepara-tion provided pure RR-TMV in yields up to 100 mg/L culture.

Modification with NOTA chelator. To a solution of RR-TMV (SSYS N-terminus, S123C, T104K, 100 µM) in 100 mM pH 7 sodium phosphate buffer, 10 equiv. of NOTA-maleimide (5 µL of 100 mM stock solution) was added. The solution was incubated for 1 h at room temperature. The solution was then spin concentrated 3-5 times into 25 mM pH 6.5 sodi-um phosphate buffer using a 30 kDa MWCO spin concentrator.

Modification with AF680 dye. To a solution of RR-TMV (SSYS N-terminus, S123C, T104K, 100 µM) in 50 mM pH 7.2 sodium phosphate buffer, 1 equiv. of AF680-maleimide (5 µL of 10 mM stock solution) was added. The solution was incubated for 2 h at room temperature. The solution was then spin concentrated 3-5 times into 25 mM pH 6.5 sodi-um phosphate buffer using a 30 kDa MWCO spin concentrator.

Modification with PEG5kDa. To a solution of RR-TMV-NOTA (SSYS N-terminus, S123C, T104K, 100 µM) in 25 mM pH 6.5 sodium phosphate buffer, 10.5 equiv. of NaIO4 (3.2 µL of a 100 mM stock solution) was added. The reaction was incubated for 5 min at room temperature. The solution was then spin concentrated 5 times into 25 mM pH 6.5 sodium phosphate buffer with a 30 kDa MWCO spin concentrator to remove excess NaIO4. Fol-lowing spin concentration, 1000 equiv of alkoxyamine-PEG5k (153 mg dissolved in 200 µL reaction buffer) were added. After 18 h of incubation at room temperature, the solution was spin concentrated 7-10 times into 10 mM pH 7 sodium phosphate buffer with a 30 kDa MWCO spin concentrator.

Synthesis of TMV-PEG5kDa-Dox. To a solution of RR-TMV (PAGSYS N-terminus, S123C, T104K, 100 µM) in 20 mM pH 7.5 sodium phosphate buffer, 1 equiv. of Dox-EMCH (2 µL of 100 mM stock solution) was added. The solution was incubated for 1 h at room temperature, at which point an aliquot was taken for LC-MS analysis. The solution was then spin concentrated 3-5 times into 25 mM pH 6.5 sodium phosphate buffer using a 30 kDa MWCO spin concentrator. Aminophenol-PEG5kDa-OMe (5 equiv.) and K3[Fe(CN)6] (50 equiv.) were then added and the reaction mixture was incubated for 30 min at room temperature. The excess reagents were removed using 3-5 rounds of spin concentration with 30 kDa MWCO spin concentrators. The conjugate was stored frozen at -20 °C until use.

5.9.3 Nanophage

The work presented in this section was performed by Jenna Bernard.

92

Protein expression and purification.

Expression of the “helper” phage needed to produce nanophage proteins. TG1 cells (with no plasmids) were inoculated in 10 mL 2xYT media and incubated at 37 °C over-night at 250 rpm. TG1 cells transformed with R779 plasmid were inoculated in 5 mL 2xYT media and incubated at 37 °C overnight. The TG1 cells were diluted to 1 L 2xYT and allowed to grow to OD600 = 0.1 (1-3 h). A 1 mL R779 portion of a cell culture expressing phage was added to 2xYT media containing the TG1 cells to be infected and incubation was continued for an additional 12-16 h at 37 °C, 250 rpm. The cells were centrifuged at 12,000 g, 4 °C for 20 min. The pellet was discarded and the supernatant was mixed with solid PEG8kDa and NaCl to achieve a final concentration of 2.5% w/v PEG and 0.5 M NaCl (25 g PEG and 29.2 g NaCl per L of supernanant). The phage was precipitated on ice overnight, then spun at 12,000 g, 4 °C for 20 min. The pellet was resuspended in 10-25 mL PBS and centrifuged at 10,000 g for 10 min to remove any residual bacterial cells. The supernatant was mixed with PEG8kDa and NaCl to achieve 4% w/v PEG8kDa and 0.5 M NaCl and was precipitated on ice for at least 2 h. After centrifugation at 16,000 g, 4 °C for 45 min, the pellet was resuspended in PBS and stored at 4 °C.

Expression of nanophage. K1030 E.coli were transformed with a PNJB07 plasmid and were inoculated in 5 mL 2xYT with 100 µg/mL ampicillin. The cells were allowed to grow at 37 °C, 280 rpm overnight. Several dilutions (10-4-10-8) were plated on LB plates to ensure separated colonies upon overnight incubation at 37 °C. The colonies were inoculated in 10 mL 2xYT with 100 µg/mL ampicillin and allowed to grow at 37 °C, 280 rpm overnight. The cells were then transferred to 1 L 2xYT containing 100 µg/mL ampicillin and grown at 200 rpm until OD600 = 0.2 (2-2.5 h). Note that temperatures below 33-34 °C will prevent the infection of the helper phage. The shaking of the incubator was stopped, and the incuba-tion was continued until the OD600 reached 0.3. At this point, the cells were infected with “helper” phage (from the cells prepared as described above) with multiplicity of infection (m.o.i.) of 50-100 phage/cell and incubation was continued for 15 min without shaking, then for 4 h at 37 °C, 200 rpm. After the OD600 reached 1.8, the cells were centrifuged at 12,000 g, 4 °C for 20 min. The supernatant was transferred into a new container and the helper phage was precipitated using a final concentration of 2.5% w/v PEG8kDa and 0.5 M NaCl on ice overnight. The helper phage was removed by centrifugation at 16,000 g, 4 °C for 45 min. The supernatant was mixed with PEG to achieve 15% w/v PEG8kDa final concentration and precipitated on ice overnight. The nanophage was removed by centrif-ugation at 16,000 g, 4 °C for 60 min, then resuspended in 10-25 mL buffer.

Synthesis of nanophage-PEG5kDa-NOTA. To a solution of nanophage in 250 mM phos-phate buffer pH 8 (25 µM final concentration pVIII), 40 equiv. pSCN-NOTA were added. The reaction mixture was incubated at 37 °C for 18 h. The excess reagent was removed using 100 kDa MWCO spin concentrators. To a solution of 50 µM pVIII 25 mM phosphate buffer pH 8, Rapoport’s salt was added to a final concentration of 100 mM and the solu-tion was incubated for 1 h at 37 °C. To this transaminated protein solution, 1000 equiv. of alkoxyamine-PEG5kDa-OMe were added and the reaction proceeded at pH 6, room tem-perature for 24 h.

93

Synthesis of nanophage-PEG5kDa-AF680. To a solution of nanophage in 100 mM phos-phate buffer pH 8 (50 µM final concentration pVIII), 18 equiv. NHS-AF680 were added. The reaction mixture was incubated at room temperature for 12 h. The excess reagent was removed using 100 kDa MWCO spin concentrators. To a solution of 25 µM pVIII 25 mM phosphate buffer pH 8, Rapoport’s salt was added to a final concentration of 100 mM and the solution was incubated for 1 h at 37 °C. To this transaminated protein solution, 400 equiv. of alkoxyamine-PEG5kDa-OMe were added and the reaction proceeded at pH 6, room temperature for 36 h.

Synthesis of nanophage-PEG5kDa-Dox. To a solution of nanophage in 100 mM phos-phate buffer pH 8 (25 µM final concentration pVIII), 25 equiv. of 2-iminothiolane were add-ed. The reaction mixture was incubated at room temperature for 15 min. Excess reagent was removed using 100 kDa spin concentrators. After the formation of the thiol groups, 5 equiv. Mal-PEG5kDa-OMe (JenKem, Plano, TX) were added and allowed to react at room temperature for 45 min. Excess reagent was removed using 100 kDa spin concentrators. Dox-EMCH was then added (2 equiv.) and incubated with the reaction mixture for 45 min at room temperature. The conjugate was stored frozen until use.

5.9.4. IDP

The work presented in this section was performed in collaboration with Alexander Hoepker, Jessica Lee and Jenna Bernard.

Protein expression and purification. The details regarding the cloning of this difficult to clone, repetitive sequence are described elsewhere. Plasmids were transformed into E. coli BL21 (DE3) competent cells. Starter cultures (20 mL of LB, 50 mg/L Kanamycin) were grown from single colonies, grown overnight at 37 °C, and used to inoculate 1 L of TB media (50 mg/L Kanamycin). Cultures were grown to an optical density (OD) of ca. 0.5, cooled for 20 min at 25 °C, induced with 0.5 mM IPTG, and expressed overnight (ca. 18 h) at 25 °C. Cells were harvested by centrifugation for 15 min at 4,000 rcf at 4 °C.

The cell pellet was transferred to a 50 mL Falcon tube in PBS buffer, and spun down for 10 min at 4,000 rcf. The resulting pellet (ca. 5 g) was lysed in 30 mL of buffer A (20 mM pH 7.5 HEPES, 300 mM NaCl, 10 mM imidazole) supplemented with one tablet of EDTA-free SigmaFast Protease Inhibitor (Sigma), 2 mM phenylmethanesulfonyl fluoride (PMSF, Sigma), and 10 mg of lysozyme (Sigma). The resuspended sample was lysed with an Avestin C3 homogenizer followed by 20 min of centrifugation at 24,000 rcf at 4 °C. The supernatant was filtered through a 40 mm Steriflip filter (Millipore), and loaded onto a 5 mL Ni-NTA column (Protino, Machery Nagel, Düren, Germany) connected to an AKTA FPLC system (GE Healthcare Life Sciences) that was pre-equilibrated with buffer A. The column was washed with 50 mL (10 column volumes, CV) of buffer A containing 10 mM b-mercaptoethanol. The protein was eluted with 20 mM pH 7.5 HEPES, 300 mM NaCl, 250 mM imidazole, 10 mM β-mercaptoethanol. Imidazole was removed by exchanging against 20 mM HEPES (pH=7.5), 100 mM NaCl with a 10DG desalting column (BioRad), and subsequently digested with 1 mg of thrombin protease (high purity from Bovine, MP Biomedicals). Complete digestion was achieved at room temperature after 1 h as con-firmed by LC/MS. The protein mixture was diluted with salt-free 20 mM pH 7.5 HEPES

94

buffer to 50 mL ([NaCl] = ca. 5 mM) before loading it onto a 1 mL HiTrap SP HP cation exchange column connected to an FPLC. The column was pre-equilibrated with 20 mM pH 7.5 HEPES containing 10 mM b-mercaptoethanol, washed with 10 mL (10 CV) of the same buffer and eluted with a gradient from 0-1 M NaCl (50 mL total volume). The final IDP sample was obtained with a final desalting column (10DG, BioRad) to obtain the pro-tein in 20 mM pH 7.5 HEPES, 50 mM NaCl. The protein was >95% pure by SDS-PAGE and LCMS. The purified protein was flash frozen with liquid N2 and stored at -80 °C until use. We observed dimer formation upon storage at room temperature, but it can be re-versed using reducing agents such as b-mercaptoethanol, Tris(2-carboxyethyl)phosphine (TCEP) or dithiothreitol (DTT).

Synthesis of IDP-PEG5kDa-Dox. A solution of 50 µM IDP (1350 µM Lys) was reacted with 5 equiv. sulfo-SPDP in 100 mM phosphate buffer pH 7.5 at room temperature over-night. The pyridine-2-thione group was removed by using 10 equiv. of dithiothreitol for 30 min at room temperature. The excess reagent was removed using 10 kDa MWCO spin concentrators. After the reduction, 1 equiv. of Dox-EMCH and 2.5 equiv. Mal-PEG5kDa-OMe were added and the reaction mixture was incubated at room temperature for 1 h. Excess reagent was removed using 10 kDa MWCO spin concentrators. The conjugate was stored frozen at -20 °C until use.

5.9.5 Radiolabeling of protein capsids conjugates (MS2, TMV and nanophage). The 64Cu copper stock (20-30 mCi, ~200 µL) was diluted with 1 mL of 0.1 M ammonium citrate buf-fer, pH 6.2 to generate a final volume of ~1200 µL at pH 5.5 (determined by pH paper). Each reaction tube was then charged with 100-400 µL of diluted 64Cu solution and 50-150 µL of the protein capsids (150-200 µM in capsid monomer). The complexation reactions were allowed to proceed for 2 h at room temperature, then purified using Nap-5 or Nap-10 columns. Samples were subsequently concentrated using 30 kDa (for TMV) or 100 kDa MWCO spin concentrators (for MS2 and nanophage). Centrifugation was performed at 10,000 rpm for 5 min per round of concentrating until the desired volume was reached. In vitro stability studies for TMV and nanophage are presented in the Appendix.

PET/CT and biodistribution studies. All animal procedures were performed according to a protocol approved by the UCSF Institutional Animal Care and Use Committee (IA-CUC). Five-week old athymic (nude) female mice weighing 18-23 g were purchased from Simonsen Labs (Gilroy, CA). For tumor inoculation, 3x105 U87-Luc glioblastoma cells with luciferase reporter gene were implanted intracranially. The imaging and biodistribution experiments were started 26 days following implantation.

Tumor-bearing mice in sets of 3 animals per study group were injected in the tail vein (intravenous, IV) with 50-200 µCi of 64Cu-labeled conjugates in 100 µL of sterile saline. One animal from each group was selected for imaging with microPET/CT (Inveon microP-ET docked with microCT, Siemens, Washington, D.C.). Dynamic imaging was performed from the time of injection to 1 h post injection, followed by a 20 min static scan at 5 h. CT scans were performed after each PET scan to provide anatomical localization of ra-dionuclide data, as well as photon attenuation map for attenuation-corrected PET recon-struction. Images were reconstructed using the AMIDE software v.1.0.4. More information

95

regarding the circulation profile of the agents based on the PET data is presented in the Appendix.

Mice were sacrificed at 5 h post-injection and the blood, tumor, and major organs were harvested and weighed. The radioactivity present in each sample was measured using a Wizard gamma-counter (Perkin Elmer, Waltham, MA). All values were decay corrected, and the percentage injected dose per gram (% ID/g) was calculated for each tissue sam-ple. Means and standard deviations were calculated for each group. Using Excel software (Microsoft, Redmond, WA), an unpaired t-test with equal variance and a two-tailed p val-ue was performed for organs from different data sets. A result was considered statistically significant if it occurred at the p < 0.05 level.

5.9.6 Optical imaging in glioblastoma models

Mice bearing U87-Luc intracranial tumors in sets of 3 animals per study group were injected via CED with 10 µL of protein-dye conjugates in sterile saline. To obtain the free dye control, AF680-NHS was hydrolyzed with 20 mM NaOH solution for 1 h at room tem-perature. The solution was then neutralized with 100 mM potassium phosphate at pH 5.5. At 5 h post-injections, the animals were imaged using an IVIS 50 Lumina system. The animals were then euthanized and the main organs, including the brain and the tumor were harvested. The tumors were cryopreserved with 4% paraformaldehyde overnight and 30% sucrose, before being embedded in optimal cutting temperature (OCT, Sakura Finetek, Torrance, CA). Slides (6 µm thick) were obtained using a cryostat. Confocal mi-croscopy imaged were obtained using a Zeiss LSM 710 Axio Observer microscope at the Molecular Imaging Center, UC Berkeley (data not shown).

5.9.7 Cell viability assays

U87 cells were trypsinized and diluted to a density of 50,000 cells/mL. An aliquot of 100 µL of cell stock was placed in a well of a 96 well plate (Corning, Corning, NY) for a density of 5000 cells/well. The plate was incubated at 37 °C, 5% CO2 for 2-4 h. Following this, media was removed from the plate and 100 µL of appropriate sample stocks in phe-nol free media was added. The cells were incubated at 37°C, 5% CO2 for a determined about of time (8 h - 3 days). The media containing the sample was removed from the well and 100 µL of MTS media (20% MTS in phenol free media) was added to each well and incubated for 1-3 hours. An Infinite 200 Pro plate reader (Tecan, Switzerland) was used to measure the absorbance at 490 nm.

5.9.8 Cell uptake studies

U87 cells were trypsinized and diluted to a density of 50,000 cells/mL. An aliquot of 200 µL of cell stock was placed in a well of a 96 well plate (Corning) for a density of 10,000 cells/well. The plate was incubated at 37 °C, 5% CO2 for 48 h. Following this, me-dia was removed from the plate and 200 µL of appropriate sample stocks (1 µM Dox) in phenol free media was added. The cells were incubated at 37°C, 5% CO2 for 48 h. The Dox fluorescence was monitored using an Incucyte live cell imaging system. Images were processed by using the Top Hat background subtraction algorithm with a radius of 10 µm.

96

5.9.9 Tumor shrinkage studies in glioblastoma models

Mice bearing U87-Luc intracranial tumors in sets of 9 animals per study group were injected via CED with 10 µL of protein-Dox conjugates in sterile saline, at day 10 post-im-plantation. Tumor size was monitored using the Luciferase reporter system and measur-ing bioluminescence on an IVIS 50 Lumina system.

5.10 References

1. Peer, D., Karp, J.M., Hong, S., FaroKHzad, O.C., Margalit, R. and Langer, R., Nano-carriers as an emerging platform for cancer therapy. Nature Nanotechnology. 2007, 2, 751-760.

2. Chapman, S., Dobrovolskaia, M., Farahani, K., Goodwin, A., Joshi, A., Lee, H., et al., Nanoparticles for cancer imaging: The good, the bad, and the promise. Nano Today. 2013, 8, 454-460.

3. De Jong, W.H. and Borm, P.J., Drug delivery and nanoparticles:applications and haz-ards. Int J Nanomedicine. 2008, 3, 133-149.

4. Allen, T.M. and Cullis, P.R., Drug delivery systems: entering the mainstream. Science. 2004, 303, 1818-1822.

5. Maeda, H., Nakamura, H. and Fang, J., The EPR effect for macromolecular drug de-livery to solid tumors: Improvement of tumor uptake, lowering of systemic toxicity, and distinct tumor imaging in vivo. Adv Drug Deliv Rev. 2013, 65, 71-79.

6. Jain, R.K. and Stylianopoulos, T., Delivering nanomedicine to solid tumors. Nature Re-views Clinical Oncology. 2010, 7, 653-664.

7. Lee, H., Fonge, H., Hoang, B., Reilly, R.M. and Allen, C., The effects of particle size and molecular targeting on the intratumoral and subcellular distribution of polymeric nanopar-ticles. Mol Pharm. 2010, 7, 1195-1208.

8. Perrault, S.D., Walkey, C., Jennings, T., Fischer, H.C. and Chan, W.C., Mediating tumor targeting efficiency of nanoparticles through design. Nano Lett. 2009, 9, 1909-1915.

9. Oltra, N.S., Swift, J., Mahmud, A., Rajagopal, K., Loverde, S.M. and Discher, D.E., Filo-micelles in nanomedicine - from flexible, fragmentable, and ligand-targetable drug carrier designs to combination therapy for brain tumors. Journal of Materials Chemistry B. 2013, 1, 5177-5185.

10. Geng, Y., Dalhaimer, P., Cai, S., Tsai, R., Tewari, M., Minko, T., et al., Shape effects of filaments versus spherical particles in flow and drug delivery. Nat Nanotechnol. 2007, 2, 249-255.

11. Chithrani, B.D., Ghazani, A.A. and Chan, W.C., Determining the size and shape de-pendence of gold nanoparticle uptake into mammalian cells. Nano Lett. 2006, 6, 662-668.

97

12. Liu, D., Mori, A. and Huang, L., Role of liposome size and RES blockade in controlling biodistribution and tumor uptake of GM1-containing liposomes. Biochim Biophys Acta. 1992, 1104, 95-101.

13. Gratton, S.E., Ropp, P.A., Pohlhaus, P.D., Luft, J.C., Madden, V.J., Napier, M.E., et al., The effect of particle design on cellular internalization pathways. Proc Natl Acad Sci U S A. 2008, 105, 11613-11618.

14. Shukla, S., Ablack, A.L., Wen, A.M., Lee, K.L., Lewis, J.D. and Steinmetz, N.F., In-creased tumor homing and tissue penetration of the filamentous plant viral nanoparticle Potato virus X. Mol Pharm. 2013, 10, 33-42.

15. Qiu, Y., Liu, Y., Wang, L., Xu, L., Bai, R., Ji, Y., et al., Surface chemistry and aspect ratio mediated cellular uptake of Au nanorods. Biomaterials. 2010, 31, 7606-7619.

16. Yuan, H., Li, J., Bao, G. and Zhang, S., Variable nanoparticle-cell adhesion strength regulates cellular uptake. Phys Rev Lett. 2010, 105, 138101.

17. Roser, M., Fischer, D. and Kissel, T., Surface-modified biodegradable albumin nano- and microspheres. II: effect of surface charges on in vitro phagocytosis and biodistribu-tion in rats. Eur J Pharm Biopharm. 1998, 46, 255-263.

18. Xiao, K., Li, Y., Luo, J., Lee, J.S., Xiao, W., Gonik, A.M., et al., The effect of surface charge on in vivo biodistribution of PEG-oligocholic acid based micellar nanoparticles. Biomaterials. 2011, 32, 3435-3446.

19. Li, S.D. and Huang, L., Pharmacokinetics and biodistribution of nanoparticles. Mol Pharm. 2008, 5, 496-504.

20. Moghimi, S.M., Hunter, A.C. and Murray, J.C., Long-circulating and target-specific nanoparticles: theory to practice. Pharmacol Rev. 2001, 53, 283-318.

21. Otsuka, H., Nagasaki, Y. and Kataoka, K., PEGylated nanoparticles for biological and pharmaceutical applications. Adv Drug Deliv Rev. 2003, 55, 403-419.

22. Hak, S., Helgesen, E., Hektoen, H.H., Huuse, E.M., Jarzyna, P.A., Mulder, W.J., et al., The effect of nanoparticle polyethylene glycol surface density on ligand-directed tumor targeting studied in vivo by dual modality imaging. ACS Nano. 2012, 6, 5648-5658.

23. Danhier, F., Le Breton, A. and Preat, V., RGD-based strategies to target alpha(v) beta(3) integrin in cancer therapy and diagnosis. Mol Pharm. 2012, 9, 2961-2973.

24. Chen, K. and Chen, X., Integrin targeted delivery of chemotherapeutics. Theranos-tics. 2011, 1, 189-200.

25. Behrens, C.R., Hooker, J.M., Obermeyer, A.C., Romanini, D.W., Katz, E.M. and Fran-cis, M.B., Rapid chemoselective bioconjugation through oxidative coupling of anilines and aminophenols. J Am Chem Soc. 2011, 133, 16398-16401.

26. Farkas, M.E., Aanei, I.L., Behrens, C.R., Tong, G.J., Murphy, S.T., O’Neil, J.P., et al.,

98

PET Imaging and biodistribution of chemically modified bacteriophage MS2. Mol Pharm. 2013, 10, 69-76.

27. ElSohly, A.M., Netirojjanakul, C., Aanei, I.L., Jager, A., Bendall, S.C., Farkas, M.E., et al., Synthetically Modified Viral Capsids as Versatile Carriers for Use in Antibody-Based Cell Targeting. Bioconjug Chem. 2015, 26, 1590-1596.

28. Aanei, I.L., ElSohly, A.M., Farkas, M.E., Netirojjanakul, C., Regan, M., Taylor Murphy, S., et al., Biodistribution of Antibody-MS2 Viral Capsid Conjugates in Breast Cancer Mod-els. Mol Pharm. 2016, 13, 3764-3772.

29. Klug, A., The tobacco mosaic virus particle: structure and assembly. Philos Trans R Soc Lond B Biol Sci. 1999, 354, 531-535.

30. Shukla, S., Eber, F.J., Nagarajan, A.S., DiFranco, N.A., Schmidt, N., Wen, A.M., et al., The Impact of Aspect Ratio on the Biodistribution and Tumor Homing of Rigid Soft-Matter Nanorods. Adv Healthc Mater. 2015, 4, 874-882.

31. Wu, M., Shi, J., Fan, D., Zhou, Q., Wang, F., Niu, Z., et al., Biobehavior in normal and tumor-bearing mice of tobacco mosaic virus. Biomacromolecules. 2013, 14, 4032-4037.

32. Lee, K.L., Hubbard, L.C., Hern, S., Yildiz, I., Gratzl, M. and Steinmetz, N.F., Shape matters: the diffusion rates of TMV rods and CPMV icosahedrons in a spheroid model of extracellular matrix are distinct. Biomater Sci. 2013, 1.

33. Miller, R.A., Presley, A.D. and Francis, M.B., Self-assembling light-harvesting sys-tems from synthetically modified tobacco mosaic virus coat proteins. J Am Chem Soc. 2007, 129, 3104-3109.

34. Dedeo, M.T., Duderstadt, K.E., Berger, J.M. and Francis, M.B., Nanoscale protein assemblies from a circular permutant of the tobacco mosaic virus. Nano Lett. 2010, 10, 181-186.

35. Ghosh, D., Kohli, A.G., Moser, F., Endy, D. and Belcher, A.M., Refactored M13 Bacte-riophage as a Platform for Tumor Cell Imaging and Drug Delivery. Acs Synthetic Biology. 2012, 1, 576-582.

36. Suthiwangcharoen, N., Li, T., Li, K., Thompson, P., You, S.J. and Wang, Q., M13 Bac-teriophage-Polymer Nanoassemblies as Drug Delivery Vehicles. Nano Research. 2011, 4, 483-493.

37. Bar, H., Yacoby, I. and Benhar, I., Killing cancer cells by targeted drug-carrying phage nanomedicines. BMC Biotechnol. 2008, 8, 37.

38. Palaniappan, K.K., Ramirez, R.M., Bajaj, V.S., Wemmer, D.E., Pines, A. and Francis, M.B., Molecular imaging of cancer cells using a bacteriophage-based 129Xe NMR bio-sensor. Angew Chem Int Ed Engl. 2013, 52, 4849-4853.

39. Carrico, Z.M., Farkas, M.E., Zhou, Y., Hsiao, S.C., Marks, J.D., Chokhawala, H., et al., N-Terminal labeling of filamentous phage to create cancer marker imaging agents. ACS

99

Nano. 2012, 6, 6675-6680.

40. Li, K., Chen, Y., Li, S., Nguyen, H.G., Niu, Z., You, S., et al., Chemical modification of M13 bacteriophage and its application in cancer cell imaging. Bioconjug Chem. 2010, 21, 1369-1377.

41. Zou, J., Dickerson, M.T., Owen, N.K., Landon, L.A. and Deutscher, S.L., Biodistribu-tion of filamentous phage peptide libraries in mice. Mol Biol Rep. 2004, 31, 121-129.

42. Rakonjac, J., Bennett, N.J., Spagnuolo, J., Gagic, D. and Russel, M., Filamentous bacteriophage: biology, phage display and nanotechnology applications. Curr Issues Mol Biol. 2011, 13, 51-76.

43. Sattar, S., Bennett, N.J., Wen, W.X., Guthrie, J.M., Blackwell, L.F., Conway, J.F., et al., Ff-nano, short functionalized nanorods derived from Ff (f1, fd, or M13) filamentous bacteriophage. Front Microbiol. 2015, 6, 316.

44. Tompa, P., Intrinsically disordered proteins: a 10-year recap. Trends Biochem Sci. 2012, 37, 509-516.

45. Dunker, A.K., Lawson, J.D., Brown, C.J., Williams, R.M., Romero, P., Oh, J.S., et al., Intrinsically disordered protein. J Mol Graph Model. 2001, 19, 26-59.

46. Wright, P.E. and Dyson, H.J., Intrinsically unstructured proteins: re-assessing the pro-tein structure-function paradigm. J Mol Biol. 1999, 293, 321-331.

47. Zhao, H. and Shaqfeh, E.S., Shear-induced platelet margination in a microchannel. Phys Rev E Stat Nonlin Soft Matter Phys. 2011, 83, 061924.

48. Wang, H., Dommert, F. and Holm, C., Optimizing working parameters of the smooth particle mesh Ewald algorithm in terms of accuracy and efficiency. J Chem Phys. 2010, 133, 034117.

49. Shah, P., Fitzgibbon, S., Narsimhan, V. and Shaqfeh, E.S.G., Singular perturbation theory for predicting extravasation of Brownian particles. Journal of Engineering Mathe-matics. 2014, 84, 155-171.

50. Somasi, M., Khomami, B., Woo, N.J., Hur, J.S. and Shaqfeh, E.S.G., Brownian dy-namics simulations of bead-rod and bead-spring chains: numerical algorithms and coarse-graining issues. Journal of Non-Newtonian Fluid Mechanics. 2002, 108, 227-255.

51. Hoffman, B.D. and Shaqfeh, E.S., Brownian demixing and wall effects in sedimenting suspensions of orientable particles. Phys Rev E Stat Nonlin Soft Matter Phys. 2008, 78, 055301.

52. Hashizume, H., Baluk, P., Morikawa, S., McLean, J.W., Thurston, G., Roberge, S., et al., Openings between defective endothelial cells explain tumor vessel leakiness. Ameri-can Journal of Pathology. 2000, 156, 1363-1380.

53. Huysmans, M. and Dassargues, A., Review of the use of Peclet numbers to determine

100

the relative importance of advection and diffusion in low permeability environments. Hy-drogeology Journal. 2005, 13, 895-904.

54. Hermanson, G.T. Bioconjugate Techniques. 3rd Edition ed: Elsevier, U.S.A.; 2013.

55. Omuro, A. and DeAngelis, L.M., Glioblastoma and Other Malignant Gliomas A Clinical Review. Jama-Journal of the American Medical Association. 2013, 310, 1842-1850.

56. Pardridge, W.M., The blood-brain barrier: bottleneck in brain drug development. Neu-roRx. 2005, 2, 3-14.

57. Lammers, T., Peschke, P., Kuhnlein, R., Subr, V., Ulbrich, K., Huber, P., et al., Effect of intratumoral injection on the biodistribution and the therapeutic potential of HPMA co-polymer-based drug delivery systems. Neoplasia. 2006, 8, 788-795.

58. Xi, G., Robinson, E., Mania-Farnell, B., Vanin, E.F., Shim, K.W., Takao, T., et al., Con-vection-enhanced delivery of nanodiamond drug delivery platforms for intracranial tumor treatment. Nanomedicine. 2014, 10, 381-391.

59. Yamashita, Y., Krauze, M.T., Kawaguchi, T., Noble, C.O., Drummond, D.C., Park, J.W., et al., Convection-enhanced delivery of a topoisomerase I inhibitor (nanoliposomal topotecan) and a topoisomerase II inhibitor (pegylated liposomal doxorubicin) in intracra-nial brain tumor xenografts. Neuro Oncol. 2007, 9, 20-28.

60. Krauze, M.T., Noble, C.O., Kawaguchi, T., Drummond, D., Kirpotin, D.B., Yamashita, Y., et al., Convection-enhanced delivery of nanoliposomal CPT-11 (irinotecan) and PE-Gylated liposomal doxorubicin (Doxil) in rodent intracranial brain tumor xenografts. Neuro Oncol. 2007, 9, 393-403.

101

Appendix

A1. Evaluating the cytotoxicity of MS2 and MS2-PEG5k-OMe

MDA-MB-231 cells were seeded at 5000 cells/well in a flat bottom 96 well plate (Corning) and allowed to adhere overnight. Agents were added at different concentrations and incubated with the cells for 24 h. The cells were washed with DPBS and 20 µL MTS solution (CellTiter 96 Aqueous One, Promega, Madison, WI) was added according to the manufacturer’s protocol. After 4 h of incubation at 37 °C, the absorbance at 490 nm was measured using a Tecan plate reader. The blanks were subtracted and the values were normalized to the untreated cells control.

A2. Protein A pull-down of protein corona associated with MS2

A solution of Protein A agarose beads (Pierce/Thermo) was spun at 1500 g for 3 min and washed twice with 500 µL of PBS. The beads were incubated with rabbit anti-MS2 antibodies for 1 h at 4 °C, then washed twice with PBS. MS2 samples were incubated with mouse serum at 37 °C for 48 h at a final concentration of 10 µM. An equal volume of MS2/mouse serum was mixed with PBS and protein A beads (50 µL each) and incubated at 4 °C for 1 h. The beads were washed twice with PBS. Elution buffer (50 µL 0.1 glycine pH 2.5) was incubated with the beads for 10 min at room temperature. The supernatant was removed and neutralized with a 1 M Tris solution pH 9. The samples were mixed with loading buffer and boiled at 95 °C for 10 min, then analyzed using SDS-PAGE.

A3. Estimating the circulating half life of MS2 agents based on PET data

MS2

MS2-PEG5kDa

140

120

100

80

60

40

20

00.1 1 10 100 1000

MS2 concentration (nM)

Nor

mal

ized

cel

l via

bilit

y (%

)Figure A1 Investigating the cytotoxicity of MS2 agents. MDA-MB-231 cells were incubated with MS2 and MS2-PEG5kDa for 24 h at 37 ºC. Viability was mea-sured using an MTS assay and reported as percent of the untreated control.

1 2 3MW

ladder

50

37

2520

75100150250

15

10

kDa

Figure A2 Protein-A pull down assay of the protein corona associated with MS2 upon incubation with mouse serum. Protein A beads were preincubated with anti-MS2 antibody and MS2 agents. Lane 1 = untreated control; lane 2 = MS2; lane 3 = anti-EG-FR-MS2.

MS2

MS2-PEG5kDa

140

120

100

80

60

40

20

00.1 1 10 100 1000

MS2 concentration (nM)

Nor

mal

ized

cel

l via

bilit

y (%

)

Figure A1 Investigating the cytotoxicity of MS2 agents. MDA-MB-231 cells were incubated with MS2 and MS2-PEG5kDa for 24 h at 37 ºC. Viability was mea-sured using an MTS assay and reported as percent of the untreated control.

1 2 3MW

ladder

50

37

2520

75100150250

15

10

kDa

Figure A2 Protein-A pull down assay of the protein corona associated with MS2 upon incubation with mouse serum. Protein A beads were preincubated with anti-MS2 antibody and MS2 agents. Lane 1 = untreated control; lane 2 = MS2; lane 3 = anti-EG-FR-MS2.

102

The dynamic PET images taken from time of injection to 1 h post-injection were used to approximate the clearance profile of the injected agents. A 3D region of interest (ROI) was chosen to overlap with the left ventricle of the heart to estimate the radioactivity present in the blood pool. The signal showed a decay profile corresponding to a two com-partment model, as evidenced by the shape of the log10 plot of the signal within the ROI against the midpoint of each scanning frame. The breathing motion and the heartbeat of the mouse, as well as the limited resolution of the animal scale PET scanner made it difficult to determine the precise half-lives of the agents. However, the small slope of the curve at later time points (elimination phase) indicates a slow clearance for all agents. Surprisingly, no significant increase in circulation half-life was attained after PEG or anti-body modification. The antibody had a shorter half-life than expected, which could be due to multiple modifications with the chelator leading to an altered clearance, or to dissocia-tion of the 64Cu from the chelator.

A4. Evaluating the cytotoxicity of PEGylated protein cages

U87MG cells were seeded at 5000 cells/well in a flat bottom 96 well plate (Corning) and allowed to adhere overnight. Agents were added at different concentrations and incubated with the cells for 24 h. The cells were washed with DPBS and 20 µL MTS solution (CellTiter 96 Aqueous One, Promega, Madison, WI) was added according to the manufacturer’s protocol. After 4 h incubation at 37 °C, the absorbance at 490 nm was measured using a Tecan plate reader. The blanks were subtracted and the values were normalized to the untreated cells control.

transverse coronal sagittal

0 1200 2400 36000

10

20

30

40

RO

I sig

nal i

nten

sity

Midframe time (s)0 1200 2400 3600

1.1

1.3

1.5

1.7

log

10 R

OI s

igna

l int

ensi

ty

Midframe time (s)

(b) (a)

Region of interest (ROI)

MS2

MS2-PEG2kDa/5kDa-αEGFRMS2-PEG2kDa

αEGFR

Figure A3 Estimating the circulation half life of MS2-based agents using PET data. (a) Regions of interest were drawn around the left ventricle of the heart to approximate the signal present in the blood pool. (b) The signal within an ROI was plotted against the mid frame time. The circulation profile of the agents in a mouse model implanted with HCC1954 breast cancer in the mammary fat pad (orthotopic) presented a two-com-partment profile, as indicated by the exponential decay shape of the log plots.

Free Cu

t = 48 h

t = 24 h

nanophage, t=0

Free Cu

t = 48 h

t = 24 h

TMV, t=0

0 200 400 600Retention time (s)

0 200 400 600Retention time (s)

Figure A5 Serum stability of TMV and nanophage. HPLC SEC traces show the radioactivity signal asso-ciated with the protein cages upon incubation with mouse serum at 37 ºC. At the 24 h and 48 h time points, a large amount of signal was at a different retention time compared to the intact capsid. The scale of the experiment was small and water evapo-ration during the heating process might have affected the assembly state.

0

20

40

60

80

100

120

30 6 1.2 0.24 0.048

MS2 TMV Nanophage

% C

ell v

iabi

lity

Concentration (nM)

Figure A4 Investigating the cytotoxicity of protein cages. U87MG cells were incubated with MS2, TMV and Nanophage for 24 h at 37 ºC. Viability was mea-sured using an MTS assay and reported as percent of the untreated control.

103

A5. Evaluating the stability of TMV and MS2 in mouse serum64Cu-radiolabeled TMV-NOTA-PEG5k-

Da-OMe and Nanophage-NOTA-PEG5k-

Da-OMe were incubated with mouse serum 37 °C for up to 48 h. Aliquots were taken and analyzed on an HPLC SEC equipped with a radioactivity detector. The peaks observed at the 24 h and 48 h time points have different retention times, indicating a partial disassembly of the protein cag-es. The volume of the reaction mixture for this experiment was very small (50-60 µL) and significant water evaporation was observed during the heating process. The reduced volume and increased viscosity might have affected the disassembly pro-cess of the TMV and Nanophage capsids.

A6. Stability of the MS2 capsid in the presence of trehalose as a cryopreservant

A solution of MS2 pAF/Cys at a concentration of 100 µM was incubated with 500 equiv. of trehalose and stored at 4 ˚C or frozen over dry ice/acetone bath and lyophilized. The lyophilized sample was resuspended in water after 3 days. TEM and HPLC SEC were used to investigate the assembly state of the capsid upon lyophilization. A PolySEP GFP5000 column was used with a mobile phase of 1.5 mL/min 10 mM phosphate buffer pH 7.2.

Free Cu

t = 48 h

t = 24 h

nanophage, t=0

Free Cu

t = 48 h

t = 24 h

TMV, t=0

0 200 400 600Retention time (s)

0 200 400 600Retention time (s)

Figure A5 Serum stability of TMV and nanophage. HPLC SEC traces show the radioactivity signal asso-ciated with the protein cages upon incubation with mouse serum at 37 ºC. At the 24 h and 48 h time points, a large amount of signal was at a different retention time compared to the intact capsid. The scale of the experiment was small and water evapo-ration during the heating process might have affected the assembly state.

0

20

40

60

80

100

120

30 6 1.2 0.24 0.048

MS2 TMV Nanophage

% C

ell v

iabi

lity

Concentration (nM)

Figure A4 Investigating the cytotoxicity of protein cages. U87MG cells were incubated with MS2, TMV and Nanophage for 24 h at 37 ºC. Viability was mea-sured using an MTS assay and reported as percent of the untreated control.

0

10

20

30

40

0.1 1 10 100 1000 10000

Num

ber (

%)

Size (nm)

MS2 lyophilizedwith trehalose

MS2MS2

MS2 lyphilized

0 2 4 6 8 10 12

MS2 + trehalose lyophilized

Retention time (min)

MS2 lyphilized / resuspended

no trehalose with trehalose

(a) (b)

(c)

Figure A6 Investigation of the effect of crypreservant on the assembly state of MS2 capsids. VLPs were flash frozen on dry ice/acetone bath in the presence or the absence of trehalose (500 eq), then lyophilized. After 3 days, the capsids were resuspended in water and analyzed via (a) HPLC SEC, (b) DLS and (c) TEM.