Droplet Microfluidics reverse transcription and PCR ...1151596/FULLTEXT01.pdf · Droplet...

Droplet Microfluidics reverse transcription and PCR towards Single Cell and Exosome Analysis

Lovisa Söderberg

Royal Institute of Technology School of Biotechnology

Stockholm 2017


2

© Lovisa Söderberg Stockholm 2017 Royal Institute of Technology School of Biotechnology Science for Life Laboratory SE-171 65 Solna Sweden Printed by Universitetsservice US-AB Drottning Kristinas väg 53B 114 28 Stockholm Sweden ISBN 978-91-7729-577-8 TRITA BIO-Report 2017:15 ISSN 1654-2312 Cover image: Fluorescence image of RT-PCR reaction performed on exosomes in microfluidic droplets

Lovisa Söderberg

3

Public defense of dissertation This thesis will be defended on the 17th of November 2017 at 10:00 in Air & Fire, Tomtebodavägen 23 A, Science for Life Laboratory, Solna, for the degree of “Teknologie Doktor” (Doctor of Philosophy, PhD) in Biotechnology.

Respondent Lovisa M. Söderberg, MSc in Biotechnology Division of Proteomics and Nanobiotechnology, School of Biotechnology, KTH – Royal Institute of Technology, Stockholm, Sweden Faculty Opponent: Dr. Valérie Taly, CNRS Research Director Université Paris Descartes, France Evaluation committee: Prof. Bjørn T Stokke, Division of Biophysics and Medical Technology, NTNU – Norwegian University of Science and Technology, Trondheim, Norway Prof. Dang Duong Bang, National Food Institute, DTU – Technical University of Denmark Assoc. Prof Christelle Prinz Division of Solid State Physics, Lund University, Lund, Sweden Chairman: Prof. Per-Åke Nygren Division of Protein Technology, School of Biotechnology, KTH – Royal Institute of Technology Main Supervisor: Prof. Helene Andersson Svahn Division of Proteomics and Nanobiotechnology, School of Biotechnology, KTH – Royal Institute of Technology, Stockholm, Sweden Co-supervisor: Asst. Prof Håkan Jönsson Division of Proteomics and Nanobiotechnology, School of Biotechnology, KTH – Royal Institute of Technology, Stockholm, Sweden


4

Lovisa Söderberg

5

Abstract Miniaturization of biological analysis is a trend in the field of biotechnology aiming to increase resolution and sensitivity in biological assays. Decreasing the reaction volumes to analyze fewer analytes in each reaction vessel enables the detection of rare analytes in a vast background of more common variants. Droplet microfluidics is a high throughput technology for the generation, manipulation and analysis of picoliter scale water droplets an in immiscible oil. The capacity for high throughput processing of discrete reaction vessels makes droplet microfluidics a valuable tool for miniaturization of biological analysis. In the first paper, detection methods compatible with droplet microfluidics was expanded to include SiNR FET sensors. An integrated droplet microfluidics SiNR FET sensor device capable of extracting droplet contents, transferring a train of droplets to the SiNR to measure pH was implemented and tested. Papers II-IV focus on increasing the resolution of biological assays performed in microfluidic droplets. In paper II, a workflow was developed for scalable and target flexible multiplex droplet PCR using fluorescently color-coded beads for target detection. The workflow was verified for concurrent detection of two microorganisms infecting poultry. The detection panel were increased to multiple targets in one assay by the use of target specific capture probes on color-coded detection beads. A cells RNA can be used to differentiate cell types and provide biomarkers for cell states. For analysis, RNA is reverse transcribed to cDNA, which can be amplified and sequenced. Droplet microfluidics has been successfully applied to single cell processing, demonstrated in paper III, where reverse transcription was performed on 65000 individually encapsulated mammalian cells. cDNA yield was approximately equivalent for reactions performed in droplets and in microliter scale. This workflow was further developed in paper IV to perform reverse transcription PCR in microfluidic droplets for detection of exosomes based on 18S RNA content. The identification of single exosomes based on RNA content can be further developed to detect specific RNA biomarkers for disease diagnostics. Droplet microfluidics has great potential for increasing resolution in biological analysis and to become a standard tool in disease diagnostics and clinical research. Keywords: Droplet microfluidics, Reverse transcription, Droplet PCR, High Throughput biology, Single cell Analysis, Exosomes


6

List of publications and manuscripts This thesis is based on the following articles and manuscripts referred to in the text by their Roman numerals (I-IV). The full version of the articles can be found in the end of the thesis.

I. Roodabeh Afrasiabi*, Lovisa M. Soderberg*, Haakan N. Joensson, Per Björk, Helene Andersson Svahn and Jan Linnros. Integration of a Droplet-Based Microfluidic System and Silicon Nanoribbon FET Sensor, Micromachines 2016, 7, 134.

II. Prem Kumar Periyannan Rajeswari, Lovisa M. Soderberg, Alia Yacoub,

Mikael Leijon, Helene Andersson Svahn and Haakan N. Joensson. Multiple pathogen biomarker detection using an encoded bead array in droplet PCR, Journal of Microbiological Methods 2017, 139, 22-28.

III. Lovisa M. Soderberg, Haakan N. Joensson and Helene Andersson Svahn.

Parallel cDNA synthesis from thousands of individually encapsulated cancer cells – towards large scale single cell gene expression analysis, 17th International Conference on Miniaturized Systems for Chemistry and Life Sciences 2013, 1737-1739.

IV. Lovisa M. Soderberg, Pedro Fonseca, Theocharis Panaretakis and Haakan

N. Joensson. Detection of single exosomes in microfluidic droplets by RT-PCR amplification of 18S RNA content, Manuscript.

*These authors contributed equally All paper are reproduced with the permission of the copyright holders

Lovisa Söderberg

7

Contribution to the publications Paper I. Contributed in designing the device. Performed all experimental work together with Roodabeh Afrasiabi. Wrote the paper together with co-authors. Paper II. Contributed in method development of the droplet-based PCR and writing the paper. Paper III. Performed and planned the experimental work. Prepared figures and wrote the major part of the paper. Paper IV. Performed and planned the experimental work. Prepared figures and wrote the major part of the paper.


8

Lovisa Söderberg

9

Populärvetenskaplig sammanfattning När man blir sjuk så beror det oftast på att några celler inte beter sig som de ska eller att man har blivit infekterad av patogener, främmande celler, som inte hör hemma i kroppen. Läkare försöker hitta dessa celler genom att ta prover av exempelvis vävnad, blod eller urin. Att diagnosticera en sjukdom på ett tidigt stadium när inga tydliga symptom uppvisats är som att leta efter en nål i en höstack av friska celler, där de flesta mätningar visar att det bara finns hö i stacken. Om man skulle kunna separera höet så att man kan titta på varje strå för sig så hittar man tillslut nålen, eller i mitt fall de sjukdomsorsakande cellerna. Vi letar igenom stora mängder celler genom att partitionera cellerna i små droppar med droppmikrofluidik där varje droppe fungerar som ett avskilt reaktionskärl som kan separat manipuleras och analyseras. Med droppmikrofluidik kan man generera miljontals droppar på bara några minuter så att mängder av reaktioner kan utföras parallellt. För att identifiera olika celltyper och celler relaterade till sjukdomar kan man analysera deras genetiska material dvs. DNA och RNA. När vi analyserar DNA eller RNA letar vi efter specifika sekvenser som fungerar som en markör för en viss patogen eller ett sjukdomsförlopp. Signalen från dessa markörer måste förstärkas för att kunna identifieras, ofta med en PCR-reaktion som framställer miljoner kopior av samma sekvens. Traditionellt så gör man denna typ av reaktioner i plattor med 96 reaktioner per platta. Genom att utföra DNA- och RNA-analyser i droppar så kan man utföra ett mycket större antal reaktioner vilket motsvarar att analysera en större del av höstacken. Artiklarna i den här avhandlingen åskådliggör utvecklingen av nya metoder som syftar till att utöka vilka analyser som droppmikrofluidik kan användas till. Vi har dels utvecklat en metod för att amplifiera flera olika DNA markörer i samma prov. Denna metod tillämpas för att identifiera skadliga mikroorganismer i DNA-prover från tamfåglar. Vi har även utvecklat en metod för att detektera material, kallat exosomer, som utsöndras i blod från cancerceller. Exosomer är små avknoppningar från celler som innehåller bland annat RNA. Vi hoppas på att kunna vidareutveckla den här tekniken för t.ex. cancerdiagnostik genom att identifiera intressanta RNA markörer.


10

Table of contents Public defense of dissertation .................................................................................... 3Respondent.............................................................................................................................................................3

Abstract ................................................................................................................... 5List of publications and manuscripts ................................................................... 6Contribution to the publications ........................................................................... 7Populärvetenskaplig sammanfattning .................................................................. 9Table of contents .................................................................................................. 10Chapter I – Introduction ....................................................................................... 13

Miniaturizing biological analysis – aims and advantages ......................................... 13

Chapter II – Droplet Microfluidics ....................................................................... 15Lab-on-a-chip and Microfluidics ............................................................................ 15Droplet Microfluidics ............................................................................................. 15

Introduction ............................................................................................................................... 15 Stability and biocompatibility .................................................................................................... 16 Unit operations .......................................................................................................................... 17

Droplet generation ................................................................................................................. 17 Particle encapsulation ............................................................................................................ 18 Incubation ............................................................................................................................. 19 Splitting ................................................................................................................................. 20 Fusion .................................................................................................................................... 20 Picoinjection .......................................................................................................................... 21 Sorting ................................................................................................................................... 22

Different detection methods ....................................................................................................... 22 Nanowires for label free detection .......................................................................................... 23

Device fabrication ...................................................................................................................... 24

Chapter III - Transcriptome and genome analysis ............................................ 27Introduction ........................................................................................................... 27

DNA .......................................................................................................................................... 27 RNA .......................................................................................................................................... 27 Mutations .................................................................................................................................. 28 Reverse transcription .................................................................................................................. 28 PCR ........................................................................................................................................... 29 Digital PCR ............................................................................................................................... 30 Alternative DNA amplification techniques ................................................................................. 31 Sequencing ................................................................................................................................. 31

Transcriptome/genome analysis in microfluidic droplets ........................................ 32 Droplet PCR .............................................................................................................................. 32 Droplet RT-PCR ....................................................................................................................... 34

Commercial products for droplet digital PCR ........................................................................ 34 Sequencing and droplet microfluidics ......................................................................................... 34

Single cell analysis .................................................................................................. 36

Lovisa Söderberg

11

High throughput screening methods for single cell analysis ........................................................ 36 Microtiter plates and microwell devices .................................................................................. 36

Single cell analysis in microfluidic droplets ................................................................................. 37 Exosome analysis .................................................................................................... 39

Chapter IV - Present investigation ...................................................................... 41Paper I: Integration of a Droplet-Based Microfluidic System and Silicon Nanoribbon FET Sensor ............................................................................................................. 42Paper II: Multiple pathogen biomarker detection using an encoded bead array in droplet PCR ........................................................................................................... 44Paper III: Parallel cDNA synthesis from thousands of individually encapsulated cancer cells – towards large scale single cell gene expression analysis ........................ 48Paper IV: Detection of single exosomes in microfluidic droplets by RT-PCR amplification of 18S RNA content .......................................................................... 50

Conclusion and future perspectives ................................................................... 53Abbreviations ........................................................................................................ 55Acknowledgement ............................................................................................... 56References ............................................................................................................ 59Publications and manuscripts ............................................................................. 69


12

Lovisa Söderberg

13

Chapter I – Introduction

Miniaturizing biological analysis – aims and advantages A miniaturizing trend has been evident in the field of biotechnology over the last decades. Large flasks have been replaced with 1.5 ml test tubes that were later reduced to microtiter plate with 96 wells of 100µl each. When the microtiter plate format was no longer sufficient new technologies were developed such as microarrays, where the analytes were attached to a surface in small clusters, flow cytometers, which makes it possible to screen single cells, and microfluidics, with channel systems on chip to be able to manipulate liquid volumes too small to pipette. The decrease in reaction volume ensures that more reactions can be performed with the same amount of reagents. Apart from the decreased cost of reagents per reaction there are motivations behind this trend of miniaturizing biology. Research for clinical applications often focus on finding rare targets in a vast background of competing biomaterial. A disease state is commonly initiated by a small population of cells not behaving in their supposed manner or by a small number of pathogens entering the body. This forces clinicians to try to locate and treat the disease-initiating cells in a sea of healthy cells. To diagnose a disease samples are acquired from the patient in a biopsy of a tissue or a sample of a bodily fluid (e.g. blood, saliva, urine). One illustrative example would be circulating tumor cells (CTCs) that are only present in a few cells per milliliter of blood but play a major part in the metastasis of cancer. When only a few cells are affected by a disease the response in a diagnostic assay is often weak. If we then add the background signal from all the healthy cells in the body it seems almost impossible to identify a disease at an early stage. There are however ways of overcoming this obstacle; one is to develop new more sensitive diagnostic assays with increased resolution and another is to reduce the background noise from the surrounding cells.


14

Figure 1.1 Schematic illustrating the increased resolution gained from digitizing biology. When reducing the reaction volume, cells or molecules are partitioned into individual reaction vessels so that rare variants can be identified.

The decreased reaction volume allows for fewer cells or molecules to be analyzed in the same reaction vessel, which would allow us increased resolution to identify rare events that otherwise would be hidden in the background of more common variants (Figure 1.1). Digitizing biology by analyzing one cell or molecule at a time, gives a binary read-out with healthy vs. unhealthy cells and by determining the ratio between them could indicate the severity of the disease and potentially predict outcome. This will reduce the background signal from healthy cells but it will also yield a majority of negative reactions, not containing a disease cell, increasing the need for methods with the potential to run many reactions in short period of time. A technique with the potential for single molecule analysis at high throughput is droplet microfluidics, which will be the focus of this thesis.

Lovisa Söderberg

15

Chapter II – Droplet Microfluidics

Lab-on-a-chip and Microfluidics The idea of Lab-on-a-chip technology (LoC) is to fit the actions performed in a macro scale lab onto a micro scale chip format. Important benefits of using LoC technology rather than conventional methods are the decreased reaction volumes and with that decreased cost per reaction and increased resolution and sensitivity [1]. The properties of micro scale fluidics, such as laminar flow in micro sized channels [2][3], allows for precise control over concentrations of molecules. The need to manipulate small volumes in order to mix, incubate and analyze different reactions has driven the technical development of microfluidic research to create different functionalities such as valves and micro pumps [4][5]. Microfluidics has become a vast field of different technologies over the last couple of decades. This thesis will focus on applications utilizing droplet microfluidics, which will be further introduced in the following chapter.

Droplet Microfluidics

Introduction Droplet microfluidics is a technology where monodisperse water droplets are generated and processed in a continuous phase of immiscible oil and stabilized by a surfactant. The droplets are in the size range from femtoliters to a nanoliters and can be generated at kHz frequencies [6]. Microfluidic droplets can be regarded as discrete reaction vessels comparable to a test tube with the advantage of automated handling in number of unit operations, which in turn can be assembled to build workflows that mimic assays performed at the macroscale. Depending on the content of the water phase, different particles or analytes can be encapsulated in droplets to perform a variety of biochemical assays at high throughput [7]. The monodispersity of the microfluidic droplets allows for quantitative analysis and comparisons between droplets in the same assay.


16

Stability and biocompatibility One key feature that makes droplet microfluidics such a powerful and versatile tool is the stability of the generated droplets. This is achieved by adding a surfactant molecule to the continuous oil phase. Surfactant molecules are amphiphilic, where different parts of the molecule have an affinity to one of the two immiscible phases, water or oil. Surfactants partition to the interphase between the continuous and dispersed phase creating a layer surrounding each droplet with the hydrophobic tail in the oil and the hydrophilic head in the water phase [8][9] (Figure 2.1).

Figure 2.1 Schematic of surfactant molecules. Surfactant molecules consists of a hydrophilic head, that will partition to the water phase inside the droplet, and one or more hydrophobic tails, which will partition to the surrounding liquid. Surfactant molecules will therefore form a layer surrounding each droplet.

When surfactants accumulate at the water/oil interface surface tension is decreased which stabilizes the droplets and prevent them from merging [8]. The stability of the emulsion increases with higher surfactant concentration until the critical micelle concentration (CMC) of the surfactant is reached [10]. Merging of droplets is further prevented by the steric repulsion of surfactant molecules as well as the Marangoni effect which prevents the depletion of surfactant molecules in the continuous phase between two droplets [8]. Oswald ripening is another effect to consider which drives smaller droplets to merge with larger droplets, due to their lower Laplace pressure, and would then form a polydisperse emulsion [11]. To prevent this, one has to select a continuous phase in which the dispersed phase is not soluble [10][12]. Therefore when designing experiments one has to consider which combination of oil and surfactant to use. In droplet microfluidics the continuous oil phase is most commonly either a hydrocarbon or fluorinated oil. Hydrocarbon oils have the limitation that hydrophobic compounds can partition into the oil phase. These are therefore mainly used together with hydrophilic molecules such as DNA [13]. Fluorinated oils, such as 3Ms Novec HFE-7500 and Fluorinert FC-40, are the most commonly used oils in droplet microfluidics applications

Lovisa Söderberg

17

due to the fact that most organic compounds are insoluble in them and thus will remain inside the droplet [8]. One surfactant that can be used to stabilize water droplets in fluorinated oils is the Krytox-based surfactant presented by Holtze et al. [9]. This surfactant is a non-ionic fluorosurfactant and a block copolymer with a PFPE chain as hydrophobic tail and a PEG molecule as a hydrophilic head. Another example is the EA surfactant produced by RainDance Technologies that is used in many droplet microfluidic applications [14][15]. Detergents such as Tween, SDS and Triton are also examples of commercially available surfactants and are in many applications added to the aqueous phase to increase droplet stability and to ensure particle separation prior to droplet formation [8]. As early as the 1980s, cells were shown to grow at the interface between water and fluorinated oils [16]. Biocompatibility is an important requirement for performing biochemical reactions inside microfluidic droplets. This implies that the droplet interface should not significantly interfere with the droplet content. A good basic measure of biocompatibility is the ability to cultivate cells in microfluidic droplets. This has been demonstrated for bacteria [17][18], yeast [9], mammalian cells [19][20] and even entire microorganisms such as C. Elegans [21].

Unit operations Over the past two decades a variety of methods for droplet manipulation have been developed to create a toolbox from which customized protocols can be assembled. This toolbox consists of microfluidic circuits for precise liquid handling, which are often implemented as PDMS (polydiemthylsiloxane) channel systems bonded to a glass slide. By creating different channel geometries and by adding specific features such as electrodes, the liquid’s movement can be controlled for specific operations.

Droplet generation There are two main methods used to form micro droplets. The first type is performed in a microfluidic chip with controlled liquid flows to generate monodisperse droplets. The second method uses macro scale shaking to induce droplet formation in an oil and water mixture [22]. Emulsions generated by shaking are polydisperse with a wide range of droplet sizes but with the advantage that the droplet formation can be performed in minutes for a wide range of sample volumes. In this thesis, droplets produced using the microfluidic chip-based approach is used exclusively, and referred to as microfluidic droplets.


18

Figure 2.2 Schematic illustration of different modes for droplet generation in microfluidic devices. Droplet generation can occur in two different flow regimes; A1) dripping and A2) jetting, where jetting generates smaller droplets. The different designs are A) a flow focusing device B) a T-junction device and C) a capillary device.

Microfluidic droplet formation can be induced by different techniques with the commonality that droplet size is dependent on the nozzle dimensions, flow rates and surface tension between the two liquid phases [23]. Droplet generation can be performed in three channel geometries: flow focusing, cross-flowing (T-junction) and co-flowing (capillary) (figure 2.2). Each channel geometry has two different flow regimes where droplets are generated either by dripping or jetting [24] where the jetting regime will generate much smaller droplets than dripping. By utilizing a T-junction geometry, droplet generation with a size diversity of 3 % in diameter have been achieved [23]. Different water-phase components can be combined on-chip before droplet generation using devices with dual aqueous injection inlets (Figure 2.3).

Particle encapsulation Particle encapsulation in droplets is generally a random event where the average particle occupancy is dependent on the concentration of particles in the dispersed phase prior to droplet generation. The generated population of droplets will have a Poisson distributed particle occupancy [17][19].

𝑓 𝑘, 𝜆 =𝜆!𝑒!!

𝑘!

Where 𝑓 𝑘, 𝜆 is the fraction of droplets containing 𝑘 cells and 𝜆 is the average number of cells/droplet. For example if 𝜆 = 0.1 then 90% of all droplets will be empty, 9% will contain 1 particle and less than 0.5% will contain more than 1 particle. Particle concentrations in this range are used for single molecule and single cell studies where a high

Lovisa Söderberg

19

degree of single occupancy and low double occupancy is necessary. With a droplet generation frequency of 10kHz, 1000 cells can be individually encapsulated per second at 𝜆 = 0.1. This demonstrates the high throughput capacity of droplet microfluidics and why it is such an attractive tool for single cell heterogeneity studies where large populations of cells are to be screened.

Figure 2.3 Schematic illustration of the droplet unit operations; generation, dual-flow generation with two water inlets, droplet incubation on-chip, splitting, fusion, picoinjection and droplet sorting. Black thick lines illustrate on-chip electrodes.

Incubation Many biochemical reactions require an incubation step where samples are exposed to a specific environment for a set time. For this purpose it is a necessity that microfluidic droplets can be stored while preserving a stable emulsion (Figure 2.3). Factors to consider while incubating droplets are liquid evaporation and instability at an air interface, due to the chemical structure of the surfactant [8]. Depending on the time interval and sample size, the incubation can be performed either on- or off-chip. On-chip approaches are usually utilized for shorter incubation periods and smaller emulsion volumes and include


20

long [25] or wide channels [26]. These strategies entail the issue that different droplets travel through the channels at different speed based on the velocity difference over a microfluidic channel [3], which leads to vastly different incubation times for different droplets in a population incubated under flow in a channel substantially wider than the droplet diameter. One solution to this problem was realized by including constrictions of narrower and shallower parts to allow a randomization of the droplet positions in the channel, which reduced the variability of incubation times for different droplets [26]. The constrictions also allow the possibility to probe droplets one-by-one for real-time detection [27]. On-chip incubation is also advantageous for continuous analysis of individual droplets where droplets are either immobilized by stopping the flow in the channels using valves [28] or are captured in traps and imaged over a long period of time [29]. To avoid evaporation the chip can be saturated with and submerged in water during the process [30]. Off chip incubation is best suited for larger sample volumes and requires that the emulsion be transferred to an external vessel that is sealed to prevent contact with air. The most common approach used is to store the emulsion in a glass or plastic syringe. Such syringes are easily connected to the microfluidic setup [14]. A second approach is to collect the emulsion in a PCR tube sealed with a PDMS plug to prevent an emulsion-air interface. This is advantageous for applications including multiple thermal incubation steps where the emulsion is collected in a PCR tube that can be fitted in a conventional thermocycler [31][32]. For off-chip incubation of enzymatic reactions active at room temperature, the time required for collection and reinjection before analysis has to be considered since this could generate a diversity in a droplet population due differences in incubation time.

Splitting Microfluidic droplets can be split passively by introducing a junction in the device’s channel system (Figure 2.3). The channels’ hydrodynamic resistance after the junction determines the relative volume of the split droplets [33]. This feature has e.g. been applied in an assay to reduce the droplet size prior to PCR amplification [34].

Fusion In many applications, adding new reagents is a necessary operation. One method to do this is by controllably fusing droplets together (Figure 2.3). Droplet fusion can be controlled by electrocoalescence where the surfactant interface is temporarily destabilized by applying an electric field over the microfluidic channel [35]. For fusion to occur, the two droplets must be in contact as they pass through the electric field. Droplet synchronization can be achieved by using droplets of different sizes. In pressure driven flow the smaller droplet will travel faster through the channel and catch up with the larger droplet [36]. A widening of

Lovisa Söderberg

21

the channel, to form a small chamber, is a more frequently used method, where the slower flow in the wider channel detains the first droplet enough that the second droplet catches up and merges as they enter the electric field [37]. Droplets can also be fused passively without an electric field as demonstrated by Pekin et al. where a winding channel was used to add PCR reagents to droplets [31]. Common to all droplet fusion applications are the need for the droplets to be organized in a microfluidic channel in an alternating pattern along the channel. By matching the driving flows of the two droplet populations and combining them in a T-junction this synchronization can be achieved [34].

Figure 2.4 Micrographs of the droplet unit operations a) picoinjection and b) droplet sorting. Reprinted with permission from ref [38] and ref [39], respectively.

Picoinjection An alternative technique for adding reagents to droplets is picoinjection, which was first introduced by Abate et al. in 2010 [39]. In picoinjection every droplet is individually injected with a precise volume from a continuous aqueous phase inlet (Figure 2.4). This unit operation is used to add reagents to droplets without the need to synchronize the delivery rate of droplets from two emulsions as required in droplet fusion. To perform picoinjection, additional oil is added to evenly space and separate the droplets in a microfluidic channel. The reagents to be injected into the droplets are then introduced through a T-junction with a narrow nozzle. As an electric field is applied over the intersection, the droplet’s surfactant layer is momentarily destabilized, and the flow of the injected reagents merge with the droplets as they pass by. The time it takes the droplet to pass by the intersection, as well as the flow rate of the reagents to be injected into the droplets determines the volume injected into each droplet. By varying the flow rates of the injected reagents, the continuous oil phase and the emulsion the volume injected into each


22

droplet can be precisely controlled [39]. Further examples of applications utilizing picoinjection has been demonstrated for example to study enzyme kinetics [27] and to add reverse transcription reagents prior to a reverse transcription-PCR [40].

Sorting In addition to droplet fusion and picoinjection where an electric field is used to destabilize droplets, electric fields can also be used to move droplets by dielectrophoresis. The aqueous liquid in the droplets have a higher relative permittivity than the surrounding continuous oil and will be pulled towards the direction of increasing gradient of the electric field [41]. A microfluidic device for droplet sorting has two or more possible outlets where the default outlet has a lower hydrodynamic resistance so that all droplets will go to the default outlet unless the electric field is applied (Figure 2.4). As the electric field is activated the droplets will be displaced and exit through the non-default outlet [42]. Fluorescently activated droplet sorting (FADS) was first introduced by Baret et al. in 2009 [41]. In FADS the droplet’s fluorescence intensity is used to trigger the electric field which actuates droplet sorting. Labview FPGA code was implemented such that if the fluorescent signal from a droplet exceeded a set threshold the electric field would be turned on and the droplet was sorted. The advantage of FADS over the more commonly known fluorescently activated cell sorting (FACS) [43] is the possibility to enrich based on the properties of the entire droplet volume, not just a single cell or particle. This provides the opportunity to sort based on a secreted enzyme [14] or a reaction performed inside the droplet, such as a PCR amplification of genetic material [32]. Sorting of droplets into different classes by using fluorescent intensity intervals instead of a single threshold has also been demonstrated [44]. Size-based droplet sorting is an alternative method to FADS and can be achieved with deterministic lateral displacement (DLD) in a tilted pillar array. Droplets of different sizes will have different streamlines through the pillar array where droplets with a diameter below a critical diameter flow straight whereas larger droplets follow the tilt angle of the array and can be collected from a different outlet [45].

Different detection methods Analysis of droplet microfluidic assays can be performed either by breaking the emulsion and performing the analysis in bulk or by keeping the droplets intact to analyze each droplet individually. By performing the analysis in bulk, conventional detection methods can more easily be coupled to the workflow. However, when breaking the emulsion the benefits of increased resolution during detection is often lost.

Lovisa Söderberg

23

To perform the analysis on intact droplets, several detection methods have been modified to ensure compatibility with the lab-on-a-chip format. By far the most commonly used analysis method for droplet microfluidics is fluorescence detection since it is compatible with high throughput analysis. Laser induced fluorescence is suitable for on-line detection where droplets are illuminated with a laser and the emitted fluorescent light is collected by a photomultiplier tube (PMT) [27]. Fluorescence microscopy has also been used for detection of immobilized droplets either side-by-side on a glass slide [46] or in droplet traps [29]. Fluorescent analysis requires that a fluorescent molecule be coupled to the object or reaction one wish to detect. This could be achieved by labeling particles, such as cells, inside droplets or by having free fluorophores spread throughout the droplet [6]. There are alternative detection methods successfully coupled with droplet microfluidics, many of which have the advantage of label free detection. Raman spectroscopy is based on measuring vibrations and rotations in molecular bonds upon interactions with photons, so called Raman scattering, to characterize molecular compositions. Applications presented with droplet microfluidic based assays with Raman spectroscopy readout are e.g. polymer conversion during photo polymerization of monomers [47] and lipid production in algae [48]. Mass spectrometry has also been coupled to droplet microfluidics either by immobilizing droplets in a matrix for Matrix-Assisted Laser Desorption/Ionization Mass Spectrometry (MALDI-MS) [49] or by directly ionizing the droplet content on-chip by electrospray ionization through a “spyhole” in the microfluidic channel [50]. A change in electrical properties inside a droplet has also been used as a feature for detection. Live cells can be detected inside droplets in a label-free fashion by measuring the change in impedance between two electrodes compared to an empty droplet [51]. Droplet content has also been analyzed with capacitance measurements where droplets with different ethanol concentration and therefore different relative permittivity could be distinguished [52].

Nanowires for label free detection As described in the previous section, several label-free detection methods have been developed as an alternative to fluorescence analysis, which requires that a fluorescent identifier be coupled to the reaction. Another approach to label-free detection is by using electrochemical biosensors such as Silicon Nanowire Field Effect Transistors (SiNW FET). Field Effect Transistors (FETs) are potentiometric sensors with a structure including a


24

source, a drain and a gate where the conductance in the channel between the source and the drain is controlled by the voltage applied at the gate. In a bio sensing SiNW FET the source and the drain are separated from the gate by an isolating layer and the source and the drain are connected through a nanowire, located at the bottom of a microfluidic channel where the sample liquid will be injected (Figure 2.5). Analytes in the sample liquid that interact with the nanowire will alter the surface potential and with that the conductance of the nanowire and this will function as an additional gate voltage [53]. These sensors have great appeal due to their high surface-to-volume ratio and their small size that offers measurements with high sensitivity and the interaction of only a few analytes will induce a change in conductance in the entire structure regardless of where on the nanowire they bind [54][55].

Figure 2.5 Schematic of a SiNW FET sensor. The silicon gate is separated from the gold-coted source and drain by an isolating silicon oxide layer. The source and the drain are connected through the silicon nanowire, which is in contact with the sample liquid in the microfluidic channel. A constant voltage is applied at the drain and the sample liquid in the microfluidic channel will affect the conductance of the nanowire.

pH-changes in the sample liquid induces protonation or deprotonation on the nanowire surface that will alter the conductance in the nanowire [53]. Biological reactions that cause a pH shift that have been analyzed using SiNW FET are e.g. an enzyme-linked immunosorbent assay (ELISA) [56] and an assay to analyze enzyme-substrate interactions to determine enzyme kinetics [57].

Device fabrication The microfluidic devices used for the experiments in this thesis was fabricated by soft lithography. Soft lithography is a technique where a master mold is created with the inverse pattern of the desired channel system. A polymeric material is poured onto the master and cured. The cured polymer can then be peeled of and bound to a flat surface to seal the channels (Figure 2.6). This is a rapid prototyping technique where a new design can be fabricated in under a week and a large number of devices can be cast on the same master.

Lovisa Söderberg

25

In addition, the technique do not require a clean room setting after the master has been produced. A number of different polymers with different properties and curing mechanisms, such as heat or UV-light, can be used for soft lithography. The microfluidic devices used for the projects presented in this thesis as well as many other microfluidic applications were made out of polydimethylsiloxane (PDMS). PDMS is an elastomer that can be molded to create geometries at a micrometer scale. After curing with heat, these structures can be peeled off the silicon master without deforming the pattern. The material is non-toxic to mammalian cells, gas permeable and optically transparent which enables a variety of biological assays and detection methods [58]. Another advantage of using PDMS is that it can be irreversibly bonded to e.g. itself, glass, silicon and polystyrene by oxidizing both materials with oxygen plasma treatment.

Figure 2.6 Schematic illustration on soft lithography fabrication of a microfluidic device. a) a negative photo resist is spin coated onto a silicon wafer which is b) illuminated with UV-light through a mask with the desired channel pattern. c) The wafer is put in developing liquid to remove the uncured photo resist and d) is then hard baked in an oven to complete the master, which now has the inverse pattern of the desired channel design. e) PDMS is cured on top of the silicon master before being peeled of and f) holes are cut for inlets and outlets. g) The PDMS slab is then bonded to a glass slide using oxygen plasma treatment. h) The channels are surface treated and electrodes are created. i) Finished device.


26

To manufacture a device with soft lithography (Figure 2.6) the channel design is initially drawn in CAD software and printed in a high-resolution printer to create a mask with transparent parts for the desired design [59]. A photoresist is then spin-coated on a silicon wafer to the thickness corresponding to the device depth. A commonly used photoresist is SU-8 which cures when exposed to UV-light [60]. The wafer is subsequently illuminated with UV-light through the printed mask to achieve the desired pattern. Developing liquid washes away the uncured photoresist and the result is a silicon wafer with the inverse channel design in cured photoresist that can be used for molding devices multiple times. To create multiplayer devices, with different channel depths or design features at different topographical levels, this procedure can be repeated with multiple masks [4]. To perform rapid prototyping a polymeric material, e.g. PDMS, is cured on top of the silicon master and subsequently peeled off. This creates a negative impression of the silicon master in the polymer and forms the channel structure. Inlets and outlets to the channels can be cut using a biopsy needle to get holes with precise diameters. To seal the channel system, the polymeric device is bonded to a glass slide by oxygen plasma treatment. As the two plasma activated surfaces are pressed together covalent siloxane bonds are formed which will irreversibly bind the materials together [61]. To prevent wetting, where the aqueous phase adheres to the channel walls, the device is surface treated with e.g. Aquapel (PPG Industries) to make the PDMS surface more hydrophobic and fluorophilic. For applications requiring active manipulations with electric field, electrodes can be added to the chip. Channels are then included in the device design, which are later filled with a conductive material, either a low melting point solder, that solidify in the channels, or a high salt solution [62]. These are then in turn connected to a voltage source to create the electric field. Polymeric devices manufactured by soft lithography are a good choice in research laboratories where production is performed at small scale and where rapid prototyping is essential. For larger scale commercial productions more favorable techniques are hot embossing of plastics or chemical etching in glass.

Lovisa Söderberg

27

Chapter III - Transcriptome and genome analysis

Introduction The central dogma of biology describes the flow of information within a cell where deoxyribonucleic acid (DNA) is transcribed into ribonucleic acid (RNA) and translated into proteins [63]. These three categories of molecules together form the basis of cellular life. In this thesis the focus will be on analysis of the first two; DNA and RNA.

DNA DNA is the building blocks of the genetic code. This macromolecule is arranged as two chains coiling around each other [64] where each chain is constructed out of a backbone of deoxyribose sugars with four different nucleotide bases; adenine (A), thymine (T), cytosine (C) and guanine (G). These nucleotides form distinct pairs of A and T and C and G, respectively, that links the two chains together so that one strand of the DNA molecule is complementary to the other. The human genome is organized into 2 sets of 23 chromosomes, which taken together holds over 3 billion nucleotide bases. The sequence of the nucleotides holds the information that translates into specific proteins. The genome is divided into genes, which are traditionally defined as a specific region of the genome that is translated into a protein. As our understanding of the genome and its transcription expands this definition is likely to be modified since protein coding regions on the genome has been proven to overlap and the resulting protein differ depending on post-transcriptional modifications of the RNA [65].

RNA The function of RNA is to mediate the genetic information from DNA to the rest of the cell. The structural difference of RNA compared to DNA is that it contains ribose sugars in its backbone instead of deoxyribose and the nucleotide thymidine (T) is replaced by uracil (U). RNA molecules are mainly single stranded but commonly forms secondary structures. There are different classes of RNA molecules divided based on their function in the cell. Ribosomal RNA (rRNA) is the most abundant RNA in the cell and is involved in protein synthesis. Messenger RNA (mRNA), also called protein-coding RNA, functions as a template for protein synthesis. Transfer RNA (tRNA) recruits amino acids for the protein synthesis. There are several other classes of RNA that are involved in signaling, regulation


28

of transcription and translation or RNA modifications e.g. small nuclear RNA (snRNA), long non-coding RNA (lncRNA), microRNA (miRNA) and small interfering RNA (siRNA). The RNA content of a cell is called the transcriptome and in comparison to DNA, where the content is overall the same in most cells in an organism, the transcriptome is much more dynamic. The level of different transcripts varies over time and in different cell types. Following translation the RNA molecule can be spliced at different positions. Depending on the splicing and other post-transcriptional modifications, different functional mRNA molecules are formed [65]. RNA molecules are usually present in a cell in multiple copies, but one quarter of all mRNAs are not present at all in a specific cell type or present in one single copy at any given time [66].

Mutations Although the major part of the genetic code is the same for all humans, there are variations between individuals [67]. These variants can be present in the form of single nucleotide variants (SNV), larger structural variations (SV) or copy number variations (CNV). Not all of these variations give rise to a phenotypic change e.g. 100 genes have been presented that could be completely deleted without any apparent phenotypic affect on the individual [68]. Some mutations will enhance that individual, other mutations give rise to or increase the likelihood of disease or other deleterious effects. If we can find the mutations that are related to a disease we can use them as biomarkers to diagnose and to predict therapeutic response in patients [69]. Identifying such mutations will also add to the knowledge of these diseases and help in the development of treatments. To study genetic molecules there are several techniques that have been developed. A description of some of these techniques will be presented below.

Reverse transcription Compared to DNA, RNA is a rather unstable molecule. To analyze the transcriptome, RNA is therefore often reverse transcribed back to complimentary DNA (cDNA), which can be amplified and sequenced. The first evidence that RNA could be transferred back to DNA was published in 1970s when it was demonstrated that an enzyme in RNA viruses could synthesize DNA from an RNA template, as a part of the virus’ replication [70][71]. This enzyme was later isolated and used in vitro together with short DNA primer sequences of poly(dT)s to target mRNA transcripts [72][73]. The reverse transcription reaction generates an RNA-DNA complex that can be made into double stranded DNA by first

Lovisa Söderberg

29

digesting the RNA strand with RNase H, followed by the application of a DNA polymerase to form the second DNA strand. The double stranded cDNA can then processed using standard DNA analysis methods. Reverse transcription can also be performed with target specific primers to target specific RNAs or with random primers to cover the entire transcriptome.

PCR Polymerase Chain Reaction (PCR) is a critical technology in molecular biology since it enables the amplification of specific DNA segments from a few copies to millions of copies. The DNA sequence to be amplified is defined by primer sequences, which are designed to be complimentary to the beginning and end of the DNA target. After the primer anneals to the DNA strand a heat stable DNA polymerase extends a complimentary strand from the 3’ to the 5’ end. The reaction is cycled through temperature steps to initiate DNA denaturation, primer annealing and DNA extension. The amplification reaction is exponential since the number of templates is doubled after each cycle. A PCR reaction can be coupled with reverse transcription to make a reverse transcription PCR (RT-PCR) using RNA as starting material. This can either be performed as a two-step reaction, where the reactions are performed one after the other or as a one-step reaction where all reagents are added at the same time. A one-step reaction is at risk of lower amplification efficiency but no starting material is lost in the transfer from one reaction to the next. Visualization of PCR product during the amplification has also become a valuable tool termed real-time PCR. There are several techniques used to track the amplification reaction, one is to utilize a quenched fluorescent dye that bind to the amplified double stranded DNA and then becomes fluorescent, examples of such dyes are SYBR Green or EvaGreen. Another technique is to use a TaqMan probe, which is a short DNA fragment with a hybridized quenched fluorophore. The probe anneals to the DNA strand and as the DNA polymerase elongates the DNA strand it hydrolyzes the probe and releases the fluorophore. A third option is to have fluorescent molecules attached to the PCR primers, making the PCR product fluorescent [74]. To approximate a quantitative measurement of the amplified PCR product a dilution series of known DNA concentrations, a standard series, is included in the experiment as a reference. This approach is called quantitative PCR (qPCR) where the real-time amplification of samples with unknown concentrations are compared to the standard series


30

by the number of PCR cycles it takes for the reaction to reach a set fluorescence intensity value. The number of cycles required to reach the set fluorescent intensity is referred to as the cycle threshold (Ct) for that sample. PCR can be performed in multiplex reactions where several DNA targets are amplified in the same vessel. Each individual target has its own primer pair to select the desired target region and for real-time PCR the different targets are detected using fluorophores that fluoresce in different wavelengths. However, multiplex PCR reactions require carefully designed primers and extensive optimization to ensure that each target is specifically amplified with equal efficiency. A 5-plex assay using different fluorophores to distinguish the targets have been demonstrated [75]. The number of targets that can be distinguished by fluorescence are however limited mainly due to spectral overlap in the selected fluorophores.

Digital PCR A method to more accurately quantify the number of DNA and RNA molecules is by digital PCR where the starting material is diluted so that each reaction vessel contains maximum one template molecule. The ratios between PCR positive and negative reactions give a quantitative measurement of the number of targets in the sample. This technique was initially performed in microtiter plates of 96- or 384-wells [76], which required a vast amount of PCR reagents and substantial hands-on-time. Decreasing the reaction volume and automating the process by performing digital PCR in nano- to picoliter-sized microfluidic droplets reduces both the volume of reagents needed and the hands-on time. This technology has been made available as several commercial instrument have been developed for the application [77], which will be further presented later in this chapter. Commercial systems are also available for well-based digital PCR. The QuantaStudio 3D Digital PCR system from ThermoFisher Scientific uses devices with 20 000 nano-scaled wells each for digital PCR [78]. Fluidigm also has a chamber-based system, Biomark HD, and offer different device design with varying numbers of reaction chambers. The fluid handling is automated with different number of assay and sample inlets to be combined in the reaction wells, depending on the device. The maximum number of reaction wells on such a device is 9216 with 96 assay and sample inlets each [79].

Lovisa Söderberg

31

Alternative DNA amplification techniques There are several other established DNA amplification techniques but none of them are as extensively used as PCR. There are however drawbacks to PCR such as the need for a thermocycler, amplification bias towards commonly present targets and the fact that early amplification errors will be accumulated during amplification. Isothermal amplifications techniques e.g. Rolling Circle Amplification (RCA), Multiple Displacement Amplification (MDA) and Loop-mediated isothermal amplification (LAMP) are alternatives to PCR and have the advantage of not requiring a thermocycler. Instead of using heat to denature DNA these techniques uses polymerases with strand displacing activity. RCA uses a circular DNA template and a strand displacing polymerase so that a long single stranded DNA product is generated with the repeated sequence complimentary to the target DNA [80]. MDA is an amplification technique often used for whole genome amplification (WGA) where random hexamer primers are used together with a strand displacing DNA polymerase. The hexamers anneal to the target DNA and as the DNA polymerase elongates the DNA strand it also displaces the two strands allowing new primers to anneal. This form of amplification generates a hyper-branched DNA network with a very low error rate due to the proof reading capabilities of the often used Phi 29 DNA polymerase [81]. MDA has also been performed in microfluidic droplets [82][83]. In a LAMP reaction, 4 primers are designed to specifically select 6 regions in the target sequence, which makes for a highly specific target selection. LAMP requires extensive primer design but once specific primers have been created it has a short assay time [84]. This technique has also been performed in microfluidic droplets [85].

Sequencing Sequencing of DNA is the reading of the genetic code and has been a rapidly advancing technology with great clinical potential. There are several commercial instruments with different fragment read length, amount of data output, error rate and run time. Illumina is the dominant provider on the sequencing market with several commercial instruments from the MiniSeq with a maximum output of 7.5 Gb to the HiSeq X with a maximum output of 1800 Gb per run, both with a maximum read length of 2x150 bp per DNA molecule [86]. For Illumina sequencing the DNA has to be fragmented into pieces to accommodate the short read length. Some alternative methods to avoid fragmenting the DNA have been developed for long single molecule sequencing such as Pacbio sequencing


32

with an average read length of 10 000 bp [87] and Oxford Nanopore sequencing with a read length of up to 1 million bp. Nanopore sequencing also have the advantage of not requiring pre PCR amplification [88]. However, both Pacbio and Oxford Nanopore sequencing suffer from higher error rates than Illumina.

Transcriptome/genome analysis in microfluidic droplets Droplet microfluidics is a technology with the potential for nucleic acid analysis with increased resolution and throughput, which are qualities that are highly sought-after in clinical research. Several conventional methods have been adapted to a droplet microfluidic format where some critical features has to be to considered to succeed, such as droplet stability during heat incubations and droplet manipulations as well as the lack of a method to wash the droplet content. For this reason the developed methods have to take care in the handling of the emulsion as well as be mindful of everything that is included in the reaction, as it will remain in the droplet and potentially effect downstream protocol steps. In the following section a selection of applications of nucleic acid analysis performed in microfluidic droplets will be presented.

Droplet PCR During the development of droplet PCR the amplification was performed either in a microfluidic chip where the droplets were immobilized and placed on a temperature cycling hot plate [89][46] or by having the droplets move in a microfluidic channel that transported the droplets between different thermal zones to induce the shift in temperature [25]. Both of these approaches are limited in the number of droplets that can be processed in parallel and has largely been replaced by methods where the droplets are collected in a vessel off the chip and subsequently incubated in a standard thermocycler [31][90]. The partitioning of targets into individual droplets is a key feature that ensures increased sensitivity and decreased PCR bias in droplet PCR. This was demonstrated in an assay where mutated DNA was detected in a vast background of wild type DNA. In this assay genomic DNA, compartmentalized in picoliter droplets together with PCR reagents and two types of TaqMan probes specific to mutated or wild type target coupled to different fluorophores. The DNA molecules were at low enough concentration that the majority of the droplets contained one or no target. After PCR amplification the droplets were dark or fluorescent in green, red or both green and red, indicating no target, mutated KRAS target, wild type KRAS target or both mutated and wild type, respectively (Figure 2.1). Mutated targets were detected at ratios as low as 1 mutated to 200 000 wild type KRAS genes [31].

Lovisa Söderberg

33

This assay was later modified to detect 7 common mutations in KRAS oncogene from circulating tumor DNA in a two-panel assay of either 4 or 5 targets per assay. The different targets were distinguished by the use of two different fluorophores and by varying the concentration of TaqMan probe added to the reaction so that droplets with different targets would have different fluorescent intensity [91]. The approach to multiplex droplet PCR assay by varying the TaqMan probe concentration for different targets has also been demonstrated [92].

Figure 2.1. PCR in microfluidic droplets for detection of mutated and wild type KRAS genes using target specific TaqMan probes hybridized with a NED or 6-FAM fluorophores, respectively. Droplets containing a) wild-type DNA or b) mutated DNA generates a red and green fluorescent signal, respectively after PCR amplification. In c) a mixture of mutated and wild type DNA were encapsulated in microfluidic droplets inducing either red or green droplets depending on target content. Yellow droplets correspond to double target occupancy of both wild type and mutated DNA. Image reprinted from [31] with permission from The Royal Society of Chemistry.

An alternative approach for multiplex PCR readout is to use a flow cytometer. The amplified PCR products can be labeled with different fluorophores, to identify the different targets, and attached to agarose beads using target specific probes [93][94]. To be able to perform further analysis on the content of a droplet with a target of interest the droplet PCR can be coupled to fluorescently activated droplet sorting (FADS) [41]. The droplets with increased fluorescence intensity after PCR amplification can be sorted to a separate outlet by dielectrophoresis (see droplet sorting in chapter 2). The selected droplets are enriched for the target of interest and can then be further analyzed by e.g sequencing [32]. The droplet PCR technology is making its way into clinical research. The previously mentioned assay to detect 7 common KRAS mutations was performed on clinical plasma samples to detect circulating tumor DNA [91]. Droplet PCR has also been applied for sensitive detection of rare mutations in the SETBP1 gene to determine poor prognosis in juvenile myelomonocytic leukemia [95].


34

Droplet RT-PCR Droplet microfluidics can also be used to analyze RNA by including a reverse transcription. Similarly to the PCR case, RT-PCR amplification in microfluidic droplets was initially demonstrated on chip with immobilized droplets on a thermal plate and a fluorescent readout to detect droplets initially containing an RNA target molecules [28]. An RT-PCR reaction was also used to illustrate the potential of including picoinjection in a workflow [40]. In this protocol droplets where generated with target RNA and PCR-reagents. The RT-reagents were later picoinjected into all droplets to demonstrate that only the droplets where RT-reagents had been added gave an increased fluorescent signal.

Commercial products for droplet digital PCR Several commercial instruments have been released to perform droplet digital PCR (ddPCR). These commercial instruments are built for laboratory use by clinicians without an engineering background. Bio Rad offers a ddPCR instrument in two parts. The first is a droplet generator where 8 samples of 20µl reaction mix each is loaded into a cartridge and partitioned into approximately 20 000 nanoliter-sized droplets. The emulsion is transferred into a PCR tube to be processed in a standard thermocycler. The second step is a droplet reader, which analyses the amplified PCR product in the droplets by measuring fluorescence. The droplet reader can analyze two fluorescent channels optimized for FAM and HEX or VIC [96][97]. RainDance Technologies takes a similar approach to Bio Rad with a two-part instrument called RainDrop. The sample of 25-50µl is generated into 5-10 million droplets of approximately 5 pl each which gives RainDrop a higher resolution and target throughput compared to Bio Rad. The emulsion is processed in a standard thermocycler and the droplet analyzing part of the instrument is optimized for two fluorescent channels [98].

Sequencing and droplet microfluidics There have been attempts to sequence DNA molecules in microfluidic droplets [99]. The main effect that droplet microfluidics have had on sequencing thus far was achieved when including droplet PCR in the workflow for pre-amplification in library preparations. Droplet PCR has a reduced amplification bias and allows for a more even amplification over different targets compared to a standard PCR. Amplification by droplet PCR before sequencing has been demonstrated to generate a more uniform coverage for targeted sequencing as well as a higher variant detection [100]. The same effect has been

Lovisa Söderberg

35

demonstrated where WGA performed with MDA in microfluidic droplets achieves a more even amplification and that the species distribution was better preserved compared to reactions performed in bulk [83]. For Illumina sequencing, the DNA has to be fragmented, due to increases in error rate when sequencing long reads. However, when sample DNA is fragmented, the complete sequence of each individual molecule is lost. Droplet PCR has been used for barcoding DNA prior to sequencing so that the sequence of fragmented long DNA templates are more accurately reassembled (figure 2.2). An assay was developed where long DNA targets are encapsulated and amplified in droplets prior to fragmentation in droplets. Droplets containing fragmented DNA were then merged with droplets containing unique DNA barcodes and in a second droplet PCR the barcodes were attached to the fragmented DNA. After sequencing the sequence of each DNA molecule could be reassembled to its entirety by the use of the barcodes [101]. There is also a commercially available instrument from 10X Genomics that utilizes microfluidic droplets to barcode long molecules before sequencing, which allows for haplotype calling [102].

Figure 2.2. Droplet microfluidics for single molecule deep sequencing. Long DNA templates are encapsulated into microfluidic droplets and amplified by PCR before fragmentation. These droplets are then individually merged with droplets containing unique DNA barcode, which are attached to the fragmented DNA in a second PCR step. The content of all droplets are then sequenced before clustering the acquired data using the barcodes to reassembling the sequence of the initial DNA templates. Reprinted from ref [101] licensed under Creative Commons Attribution 4.0.


36

Single cell analysis Traditional methods for DNA and RNA analysis require more starting material than what can be extracted from a single cell. Analysis is therefore performed on large samples containing hundreds to thousands of cells. This will only represent the average cell since when a large number of cells are analyzed together in the same reaction, the majority often obscures the effects of low abundant cell types. As analysis techniques have become more sensitive, fewer cells are needed as starting material to achieve a measurable signal until analysis can be performed on a single cell level. Essentially all cells in an organism share the same genotype, but different cell types will have diverse gene expression [103]. Finding these rare cell types or variants has increasing diagnostic relevance and single cell gene expression analysis has become a powerful tool to distinguish different subpopulations [104]. The diversity in a cell population is referred to as its heterogeneity. A tissue contains many different cell types, tumor tissues have been demonstrated to contain heterogeneous cell populations [105] and blood in metastatic cancer patients can contain circulating tumor cells [106]. Even clonal populations of bacteria will express differences in phenotype [107].

High throughput screening methods for single cell analysis In the pursuit to identify rare cell variants, there is often a need to divide cells into different reaction vessels prior to analysis. The number of cells that needs to be screened to give a true representation of a cell population depends on the diversity of each population, a characteristic which is hard to predict beforehand. The method used for detailed single cell analysis must therefore have the capability to analyze cells at high throughput, meaning it should be possible to process many reactions in a short period of time. To access the RNA or DNA in a cell it often has to be lysed. It is therefore essential that the material from different cells are kept separate when performing analysis on the nucleic acid content of single cells. Previously single cell gene expression analysis has mainly been performed at low throughput of 10-100 cells. In the next section a few techniques for single cell gene expression analysis with increasing throughput will be presented.

Microtiter plates and microwell devices In early single cell gene expression analysis assays the cells where physically separated in individual reaction tubes or wells in microtiter. The method for isolation was based on picking individual cells by mouth pipetting, glass capillaries or laser-assisted microdissection, all of which are highly time consuming techniques and requires high precision [103]. To accommodate a higher number of cells in the same assay the reaction

Lovisa Söderberg

37

vessels where decreased to micrometer-sized wells on slides which could contain hundreds to thousands such wells [108]. A commercial microwell-based system for automated single cell genomics processing is the C1 from Fluidigm. It performs automatic loading of up to 800 cells per device that ensures that only a single cell is loaded into each well. The C1 then proceeds with cell lysis, wash steps, reverse transcription and PCR amplification and can be used for sequencing library preparations [109][110]. Microwell devices have also been used to barcode the transcriptome of 10 000 single cells by randomly partitioning the cells into microwells together with barcoded beads. The beads where coupled to oligos containing a cell label, unique for each bead, a molecular index, unique for each oligo and an oligo(dT) tail. The cells where lysed before performing reverse transcription of the mRNA molecules and the cDNA product was captured on the beads using the oligo(dT) tail. The barcoded transcriptomes could then be further analyzed by sequencing [111]. Cells can also be actively sorted into different wells by Fluorescent Activated Cell Sorting (FACS), which is a high throughput tool for single cell analysis based on detection of size and expressed proteins [43]. Cells are engineered to display fluorescent surface proteins or labeled with specific antibodies coupled to fluorophores to achieve detection sensitive enough to detect a single target cell among millions [112].

Single cell analysis in microfluidic droplets Microfluidic droplets have great potential for single cell gene expression analysis due to the high throughput of the technology and since the content of each cell will be contained within the droplet even after cell lysis. The picoliter scale reaction volume in microfluidic droplets has however presented an obstacle for PCR assays on single cells since cell lysate at high concentrations acts as an inhibitor for PCR amplification [113][114]. Eastburn et al. overcame this obstacle and performed RT-PCR on 50 000 single cells by splitting off a large part of the droplet after cell lysis and merging the remaining part of the droplet with a water droplet, to decrease the concentration of cell lysate, before adding the RT-PCR reagents by picoinjection (Figure 2.3)[90]. This enabled a successful PCR amplification but the analysis is limited to a small part of the cell content and it is likely that many rare transcripts were lost in the process. This assay was later coupled with droplet sorting for PCR-activated enrichment of target cells for further downstream analysis [115].


38

Figure 2.3. Ultrahigh-Throughput RT-PCR in Microfluidic droplets of 50 000 single mammalian cells. The device design is illustrated in image a-c the following images show the different step of the workflow where d) cells are encapsulated in droplets together with lysis buffer. After cell lysis e) the droplets are reinjected on-chip and f) paired with a water droplet before g) merging under the influence of an electric field. h) A large part of the droplets is split of, to ensure droplet stability during the downstream PCR amplification, before i) RT-PCR reagents are picoinjected into each droplet. Reprinted with permission from [90]. Copyright (2013) American Chemical Society.

Droplet microfluidics has also been used to generate hydrogel beads to encapsulate single cells. The cells were lysed within the bead, which was functionalized with primers to capture target specific mRNA molecules. The capturing of the mRNA targets allows for washing of the beads to remove cell debris that could interfere with down-stream reactions. Next, reverse transcription was performed of the target mRNA molecules prior to amplification with RCA. The RCA products can be counted to acquire the number of transcribed targets in the cell [116]. Barcoding of single cells before sequencing ensures a way to keep the content from each cell intact. Different methods have been demonstrated to prepare barcoded-sequencing libraries for high throughput single cell RNA-sequencing (Figure 2.4). The barcodes for both techniques contain a cell-tag for each droplet as well as a unique molecular identifier (UMI) to be able to count the number of barcoded molecules. These workflows have the capacity to process thousands of cells per assay but are limited by the sequencing cost for deep sequencing [117][118]. The method to barcode single cells in microfluidic droplets before sequencing has lead to two commercialized systems by 10X Genomics [119] and Illumina-BioRad [120].

Lovisa Söderberg

39

Figure 2.4. Two methods for barcoding single cells in microfluidic droplets for high throughput RNA-sequencing. For barcoding of single cells either a) DNA-barcoded microparticles or b) hydrogel beads containing barcoded primers are used. The barcodes are attached to the mRNA by reverse transcription before library preparation and sequencing. Image reprinted with permission from the respective publishers [117][118].

Exosome analysis Over the last few years a new research field have emerged focusing on small extracellular vesicles (EV) released from cells. Their definition is still evolving but currently these vesicles are divided into micro vesicles (MV) and exosomes based by their origin in the cell. Micro vesicles are in the size range of 100-1000 nm in diameter and are released by budding from the cytoplasmic membrane. In contrast, exosomes are in the size of 40-150nm and are released from the cells when a multivesicular endosomes (MVE) fuse with the cytoplasmic membrane to release their vesicle content from the cells instead of merging with a lysosome for degradation [121][122]. The function of exosomes is not yet clearly known but they have been demonstrated to participate in cell-to-cell signaling as they are transferred between cells and influence the


40

behavior of the receiving cell [123]. The content of exosomes reflects the origin cell and can consist of proteins as well as RNA and genomic DNA [124]. It is not yet known if there is a structured packing mechanism influencing what content goes into exosomes but some molecules have been seen to be at elevated levels compared to their originating cells [121][125]. It has also been demonstrated that the contents of exosomes can influence the behavior of the receiving cell since the mRNA content in the exosomes can be translated into proteins [126]. Exosomes released from cancer cells have also been seen to induce metastatic behavior and spread cancer [123][127]. Exosomes can be released from many different cell types and have been found in a number of bodily fluids such as blood, urine, saliva, semen, breast milk and cerebrospinal fluid [121]. The blood of a normal human is estimated to contain around 2×1015 exosomes [122] and this number can be significantly increased in cancer patients [128][129]. There have also been approaches to use exosomes isolated from liquid biopsies in disease diagnostics [129]. Since exosomes are being released from all cell types, both from healthy and disease tissues, diagnostics based on exosomes extracted from bodily fluids contains a heterogeneous sample of multiple types of exosomes. For cancer diagnostics the use of liquid biopsies will provide a sample containing information on the heterogeneity in the tumor that could not be provided by a single needle biopsies [122]. So for the same reasons to perform single cell gene expression analysis, exosome should also be analyzed on a single vesicle level so that different and rare types of exosomes can be identified. The exosome vesicles are too small (<150nm) to be analyzed with a standard flow cytometer so single exosome analysis has so far been limited to imaging with light scattering or electron microscope. The exosomes can instead be separated into different reaction vessels by dilution.

Lovisa Söderberg

41

Chapter IV - Present investigation The aim of the research included in this thesis was to expand the detection mechanisms, number of detectable targets and the samples targeted in gene expression analysis to single cells and exosomes in droplet microfluidics In paper I a new label-free detection method is developed where droplet microfluidics is coupled to a Nanoribbon FET sensor for pH sensing. The next three papers are focused on developing droplet microfluidic applications towards increased resolution in nucleic acid and single cell analysis. In paper II we demonstrate the use of fluorescence barcoded beads for multiplex detection of nucleic acid targets in microfluidic droplets with a scalable approach using an easily exchangeable target panel. Paper III focuses on increasing the processing throughput of single cells for reverse transcription by performing the reaction in individual microfluidic droplets. The droplet format enables thousands of cells to be processed in parallel and keep the resulting cDNA from individual cells in separate reaction vessels. In paper IV reverse transcription is coupled with a PCR amplification to amplify RNA biomarkers from single exosomes. This enables exosome detection based on RNA content rather than surface biomarkers.


42

Paper I: Integration of a Droplet-Based Microfluidic System and Silicon Nanoribbon FET Sensor Droplet microfluidics have long been dependent on fluorescence detection for analysis. In many applications fluorescence detection is beneficial since it allows for single droplet measurements at high throughput. However it does require that the reaction output is fluorescent to enable readout. In this project we have investigated an alternative label-free detection approach by integrating a droplet device with a Silicon Nanoribbon Field Effect Transistor (SiNR FET) sensor [53][54].

Figure 4.1 Schematic of the integrated device consisting of a droplet microfluidic module for droplet injection and droplet extraction to a continuous water phase connected with a second microfluidic part with a SiNR FET sensor.

An SiNR FET is an electrochemical sensor that measures changes in the surface potential of the nanoribbon. The Nanoribbon’s oxide layer is affected by the proton concentration in the surrounding liquid as protonation/deprotonation reactions occur at the surface and changes the conductivity of the Nanoribbon. The conductivity of the Nanoribbon could therefore be used to determine pH in the surrounding liquid.

Figure 4.2 a) Image of droplets extraction to a continuous water phase. b) Schematic of the droplets extraction module. The droplet content is illustrated in blue, the oil phase in yellow and the continuous water phase in white.

Lovisa Söderberg

43

We designed a microfluidic device that allows for pH measurement of microfluidic droplet content. The device consists of three parts; a droplet injection part, a droplet extraction part and the sensing SiNR FET (Figure 4.1). The droplet extraction part has the function of transferring droplets from the oil phase to a continuous water phase for sensing with SiNR FETs (Figure 4.2). The oil phase must be removed prior to sensing since it otherwise will function as an insulator covering the sensor. This functional module could also be used as a stand-alone microfluidic module to extract droplets containing materials such as beads or cells. Droplet extraction occurs at a channel juncture where an interface between the oil and continuous water phase is formed. By applying an electric field across the channel, droplets are transferred from the oil phase to the water phase, which is then passed to the SiNR FET sensor. The electric field is applied in 20 second pulses actuating the extraction of a train of approximately 6000 droplets which are subsequently combined to a single measurement by the SiNR FET. The conductivity of the wire is affected by the shift in pH induced by the addition of the droplet contents to the continuous water phase.

Figure 4.3 a) Detection of droplets with different pH by measuring the current passing through the nanoribbon. The electric field is applied in 20 s pulses, during which droplets are transferred to the continuous water phase, and as the droplet content reaches the sensor the current in the SiNR FET sensor decreases. When the droplets content has passed the sensor, the current returns to the baseline. Three consecutive droplet transfers are performed for pH 9 followed by three droplet transfers with pH 6. In b) pH measurements of droplets with different pH measured in the same device on two different days are presented.

In this proof-of-concept project we were able to differentiate between droplet populations containing buffers with one pH-units difference for pH 6, 7, 8 and 9 (Figure 4.3). In future applications this device could be used to analyze reactions performed in droplets that induce a pH-shift such as loop-mediated isothermal amplification (LAMP) or as a glucose sensor where an enzymatic reaction oxidizes glucose releasing hydrogen peroxide. The developed device also allows for several measurements of the same sample, using multiple wires in series, which is difficult to achieve by fluorescent measurements.


44

Paper II: Multiple pathogen biomarker detection using an encoded bead array in droplet PCR In paper II, a technique is presented to increase the capabilities of fluorescent readout in microfluidic droplets to detect multiple targets using fluorescently barcoded Luminex beads. In droplet PCR, it is possible to perform multiplex reactions to amplify several targets in the same assay by designing sequence specific PCR primers for each target. Multiplex PCR assays then requires extensive optimization to ensure specific and unbiased amplification of all targets. To detect multiple targets with fluorescent readout, sequence specific probes coupled to different fluorophores are also included in the assay to generate a reporter signal for the amplification of each individual target. However, the number of distinguishable fluorophores is limited by spectral overlap and therefore limits the number of detectable targets in the same assay. An alternative approach to multiplex detection in droplet PCR is to couple target specific capture probes to beads [93][94]. In this paper we present a novel detection strategy for droplet PCR were we use commercially available fluorescently barcoded Luminex beads available in 500 color-coded bead sets which can be individually distinguished. In our assay, the target DNA is initially amplified by droplet PCR with fluorescent primers followed by target hybridization to sequence specific capture probes coupled to a uniquely barcoded Luminex bead set for each target. The specific capture probe on each bead together with the fluorophore on the hybridized target and the spectral signature of the Luminex bead ensure specific target detection. The Luminex bead array allows for a scalable assay with a flexible detection panel requiring limited optimization for each additional target compared to a convectional multiplex PCR. Rapid and sensitive detection strategies are not only needed for human diseases but also in the food industry to ensure proper food safety. We therefore verified the developed workflow by detecting nucleic acid biomarkers for three common poultry pathogens: avian influenza virus (AIV), infectious laryngotracheitis virus (ILTV) and Campylobacter jejuni.

Lovisa Söderberg

45

Figure 4.4 Schematic of the workflow where pathogen target DNA was encapsulated in droplets together with PCR reagents and fluorescent primers and collected in a PCR tube before thermocycling. Lambda exonuclease solution and maximum 1 detection bead/droplet, with target specific capture probes, was picoinjected into each droplet. The detection beads were recovered before analysis in a Luminex instrument to detect beads with attached target. Reprint from [130] with permission from Elsevier.

The workflow (Figure 4.4) was initially validated for each target individually. The nucleic acid targets were partitioned at a concentration of approximately 1 target/droplet together with PCR reagents and primers before collection in a PCR tube sealed with a PDMS plug. The emulsion was thermocycled for target amplification and subsequently re-injected onto a microfluidic chip where a lambda exonuclease and detection bead/droplet were picoinjected into each droplet. The lambda exonuclease treatment is necessary to ensure the availability of single stranded targets that can hybridize to the detection beads. The detection beads were then recovered and analyzed in a Luminex instrument. The classification of positive beads, with hybridized target, was based on the fluorescent reporter signal from non-target detection beads where a threshold was set to include 5% of non-target beads with the highest fluorescent signal, to allow comparisons between experiments. The results of the hybridization assay with 1 target/droplet on average are presented in histograms showing the number of beads with different fluorescent intensities for each bead set (Figure 4.5). 67%, 80% and 60% of the beads were classified as positive for the AIV, ILTV and Campylobacter jejuni target respectively, which demonstrates high assay specificity.


46

Figure 4.5 Histograms showing the hybridization assay signal of detection beads analyzed with a Luminex instrument. The percentage of positive detection beads reached with an average concentration of 1 target/droplet for each experiments was a) 67.2% for AIV, b) 80% for ILTV and c) 60.5% for Campylobacter jejuni. Reprint from [130] with permission from Elsevier.

The workflow was further extended to include three different bead sets for concurrent detection of two of the targets (Figure 4.6a). A mixture of AIV and ILTV target DNA, at a concentration of 0.5 and 5 targets/droplet respectively, was mixed on chip with three serially injected aqueous plugs, containing PCR reagents and detection beads for AIV, ILTV and Campylobacter jejuni target respectively, immediately prior to droplet encapsulation. After target amplification, a lambda exonuclease solution was picoinjected into each droplet to ensure the availability of single stranded DNA target. The Campylobacter jejuni detection beads were used as a reference for signal normalization following the same thresholding scheme applied for singleplex target detection. 20,8% and 98,7 % of the AIV and ILTV detection beads respectively, were classified as positive which proves that dual targets can be detected in the same assay with the developed readout strategy (Figure 4.6b).

Lovisa Söderberg

47

Figure 4.6 a) A schematic of the workflow for concurrent detection of two targets in one assay. A solution containing the AIV and ILTV DNA targets was injected and mixed on chip with three separate aqueous plugs, containing target-specific PCR primers and detection beads for one target each (AIV, ILTV or Campylobacter jejuni), before encapsulation in droplets. The emulsion was collected in a PCR-tube and thermocycled before lambda exonuclease solution was picoinjected into each droplet. The detection beads were recovered and analyzed in a Luminex instrument. B) Histogram of the hybridization signal from the recovered detection beads. The AIV and ILTV targets were encapsulated at an average concentration of 0.5 and 5 targets/droplets, respectively. The percentage of positive detection beads for the two targets shows this trend with 20.8% for AIV and 98.7% for ILTV. Reprint from [130] with permission from Elsevier.

In this paper we have developed a detection strategy for droplet PCR utilizing a color-coded suspension bead array for a potentially scalable assay due to the use of Luminex beads that are available with 500 different spectral signatures. Since droplet PCR is performed in multiple singleplex reactions, the detection panel can be more easily exchanged due to the limited optimization needed compared to a conventional multiplex PCR assay.


48

Paper III: Parallel cDNA synthesis from thousands of individually encapsulated cancer cells – towards large scale single cell gene expression analysis Droplet Microfluidics is a highly advantageous technique to use for single cell analysis due to the fact that cells can be individually compartmentalized into picoliter-sized reaction vessels at high throughput. Because of the extremely high throughput, large cell populations can be screened and the contents of each cell can be kept together after lysis. This advantage was explored in paper III where reverse transcription was performed on the RNA of 65 000 individually encapsulated mammalian cells.

Figure 4.7 Schematic illustration of the workflow for reverse transcription performed on single mammalian cells encapsulated in individual microfluidic droplets. Cells were compartmentalized in microfluidic droplets together with reverse transcription reagent and random hexamer primers. The emulsion was than collected and incubated at 65°C for 10 min for heat lysis and 50°C for 50 min for reverse transcription of the RNA released from the lysed cells. The resulting cDNA yield was analyzed using qPCR. Reprint from [131] with permission from the chemical and Biological Microsystem society (CBMS).

In the presented workflow (Figure 4.7) A549 mammalian lung cancer cells are encapsulated in microfluidic droplets (Figure 4.8a) together with reverse transcription reagents and subsequently collected in a syringe. The cells were lysed during a 10 min incubation at 65°C. Cell lysis was verified by Calecin AM/Ethidium Homodimer live/dead staining, after which the reverse transcription occurred during an incubation for 50 min at 50°C. The heat incubation of 65°C is enough to lyse the cells without substantially affecting the performance of the

Lovisa Söderberg

49

Figure 4.8 A) Picture of single cell encapsulation during droplet generation in a flow focusing microfluidic device. B) qPCR data presenting the starting cDNA concentration after reverse transcription of RNA from mammalian cells in two replicates performed in microfluidic droplets, D1 and D2, a control performed in bulk, Bulk, and a negative control, NC, without and cells present. Reprint from [131] with permission from the chemical and Biological Microsystem society (CBMS).

reverse transcriptase enzyme. The emulsion was subsequently broken and qPCR was performed on the droplet contents targeting the B2M, a housekeeping gene, to determine the resulting cDNA yield. Reverse transcription in microfluidic droplets was performed in two replicates and compared to a control reaction performed at microliter scale (labeled Bulk) with the same amount of cells as starting material as wells as a negative control without any cells (Figure 4.8b). We could demonstrate that the reverse transcription performed in microfluidic droplets gave a similar or slightly higher yield than when performed in microliter scale reactions. Reverse transcription in microfluidic droplets has later been included in a number of applications, including single cell RT-PCR [90] and single cell RNA sequencing [117][118]. This method was also further developed to perform RT-PCR on single exosomes presented in paper IV.


50

Paper IV: Detection of single exosomes in microfluidic droplets by RT-PCR amplification of 18S RNA content Single cell analysis is performed to resolve sample heterogeneity and to distinguish different cell types and rare variants in a sample. In the same manner as cells have different properties so does vesicles, such as exosomes, released from these cells. Exosomes are small (40-150nm) extracellular vesicles released from cells and are involved in cell-to-cell signaling. These vesicles share some characteristics with their origin cell and can contain proteins, RNA as well as DNA [121][122]. It has been demonstrated that the content of exosomes can influence the behavior of a receiving cell [126] as well as promote metastatic behavior in cancer [125][123]. Exosome research has lately become a field of interest due to the diagnostic potential of exosomes and the possibility to isolate them from bodily fluids (e.g. blood or saliva) in minimally invasive procedures.

Figure 4.9 Schematic illustration of the workflow for droplet microfluidic RT-PCR for single exosome detection based on RNA content. The exosomes are encapsulated in microfluidic droplets together with proteinase K and collected in a modified PCR-tube. The emulsion is incubated to allow proteinase K digestion of the exosomes before heat inactivation of the proteinase. RT-PCR reagents together with EvaGreen DNA dye are then picoinjected into each droplet individually before collection in a second modified PCR-tube for thermocycling. The droplets are finally collected as a monolayer in an imaging device before analysis with fluorescent microscopy.

In paper IV we demonstrate a workflow for detection of exosomes based on their 18S RNA content (Figure 4.9). To access the RNA content of the exosomes without mixing the content of different exosomes, the exosomes are first individually encapsulated in microfluidic droplets together with proteinase K, which degrades proteins in the exosome, which disrupts the membrane. After heat inactivation of the proteinase K, RT-PCR

Lovisa Söderberg

51

reagents together with 18S specific PCR primers and EvaGreen DNA dye are picoinjected into each droplet. The emulsion is collected in a PCR-tube and thermocycled before being reinjected into an imaging device to immobilize the droplets for fluorescence imaging. Droplets that initially contained and exosome will have increased fluorescence intensity, due to the PCR amplified DNA product, and are detected using Icy, open source bioimage informatics software, for automated imaging analysis.

Figure 4.10 Fluorescence images showing RT-PCR product stained with EvaGreen DNA dye in microfluidic droplets on samples with decreasing exosome concentrations as follows a) undiluted sample, b) 5-fold dilution, c) 25-fold dilution and d) negative control without exosomes.

The workflow was validated by performing the reaction on 5-fold serially diluted exosome samples. We demonstrate that the number of PCR-positive droplets decrease with lower exosome concentration (Figure 4.10). A negative control without exosomes was also included and a few bright droplets can be seen in the negative control, most likely due to contamination. A substantial difference in number of fluorescent droplets is observed between the negative controls and the lowest exosome concentration sample.


52

Figure 4.11 Exosome concentrations calculated from droplet-based RT-PCR on 5-fold serially diluted exosome samples in two replicates. The concentration of particles per ml was determined by the number of PCR-positive droplets with the assumption that target occupancy in droplets follows Poisson distribution. A logarithmic trend line is applied for each replicate.

Since exosome encapsulation is considered a random event, which at a population level can be described by the Poisson distribution, the number of exosomes can be estimated by the fraction of PCR-positive droplets. We performed the assay on a 5-fold serially diluted exosome sample in two replicates and by applying Poisson statistics the number of exosomes per unit volume was calculated (Figure 4.11). With this we demonstrate that we can rank the samples based on their exosome concentration. The presented workflow is the first step towards a high throughput method for nucleic acid analysis of single exosomes isolated from liquid biopsies. This method could be developed for diagnostic purposes to detect RNA biomarkers related to diseases such as cancer. The quantitative measurement is also highly relevant due to the fact that the number of released exosomes has been shown to increase in cancer patients [128][129].

Lovisa Söderberg

53

Conclusion and future perspectives Droplet microfluidics is a versatile high throughput technology with several droplet unit operations that can be assembled into complex workflows. The papers presented in this thesis aims to expand the capabilities of droplet microfluidics and to increase the resolution of biological analysis. In paper I a novel detection method is presented based on the integration of droplet microfluidics and SiNR FET sensors. The integration of the two techniques adds to the possible detection methods of droplet microfluidics, which in this case allows for pH sensing of droplet contents. The device created could be used to analyze assays that induce a pH-shift, such as a LAMP amplification of DNA. Functionalization of the Nanowire could expand the range of detection even further e.g. detection of specific DNA hybridization. In paper II a method is developed with the potential to multiplex droplet PCR using fluorescence barcoded beads. Instead of identifying the different targets with fluorophores each target is captured on fluorescence barcoded Luminex beads, available in 500 distinguishable bead sets, to increase the number of targets detected in a single assay. The workflow was demonstrated for concurrent detection of two microorganisms infecting poultry. The method is scalable, to allow detection of many more targets in the same assay, and the workflow is designed so that the targets in the detection panel are easily exchangeable. In paper III a workflow for single cell reverse transcription in microfluidic droplets was developed and demonstrated for 65 000 individually processed mammalian cells. The droplet format ensures the resulting cDNA from each single cell is kept separate for further analysis or processing. This method was expanded in paper IV were we perform RT-PCR from single exosomes in microfluidic droplets to detect exosomes based on their respective RNA content. In paper IV we demonstrated a workflow to determine exosome concentration by detection of 18S RNA, a highly prevalent target. The workflow developed could also be applied to more specific RNA targets for e.g. cancer diagnostics.


54

Droplet microfluidics has proven to be a great addition to technologies that are used for miniaturization of biological analysis to increase resolution and sensitivity. Several different companies have released commercial instruments based on droplet technology during the last couple of years e.g. BioRad, RainDance Technologies and 10x Genomics. This development proves the commercial potential of droplet microfluidics. The commercialization of the droplet microfluidics has made it more accessible and a technology, previously restricted to usage by highly specialized engineers, is now being made available in clinics. As a result, an increasing number of clinical applications are being presented based on droplet microfluidics technology. In the future I believe that this development will continue the research trend of single cell and vesicle analysis to identify rare biological events. With the advent of commercial solutions I further believe that droplet microfluidics will become a standard tool in disease diagnostics.

Lovisa Söderberg

55

Abbreviations CNV Copy number variations DNA Deoxyribonucleic acid dPCR Droplet PCR ddPCR Droplet digital PCR EV Extracellular vesicles FACS Fluorescence-activated cell sorting FADS Fluorescence-activated droplet sorting FET Field Effect Transistors LAMP Loop-mediated isothermal amplification LoC Lab-on-a-chip MDA Multiple displacement amplification mRNA Messenger RNA MV Micro vesicles MVE Multivesicular endosomes PCR Polymerase chain reaction PDMS Polydimethylsiloxane PMT Photomultiplier tube qPCR Quantitative PCR RCA Rolling Circle Amplification RNA Ribonucleic acid RNA-seq RNA sequencing rRNA Ribosomal RNA RT-PCR Reverse transcription PCR SiNW FET Silicon Nanowire Field Effect Transistor SNV Single nucleotide variants SV Structural variations WGA Whole genome amplification


56

Acknowledgement Before I end this thesis I would like to acknowledge a few people. I wouldn’t have gotten far as a PhD-student if it weren’t for all of your help and support. Håkan, tack för att du varit en så bra handledare. Du tar dig alltid tid för att diskutera forskning, oavsett hur lång tid det tar. Du har vänt och vridit på resultaten när jag varit frustrerad och hittade alltid en väg framåt. Jag har lärt mig mycket av att jobba med dig. Helene, tack för att du gav mig möjligheten att doktorera i Nanobioteknikgrupp. Du har alltid kommit med ett nytt perspektiv och uppmuntrande kommentarer om mina projekt vilket har varit väldigt värdefullt. The Knut and Alice Wallenberg foundation, for providing financial support for my research. All of my co-authors and collaborators over the years. It has been such a great experience to try to combine our different fields of research and techniques. Unfortunately, not all of our efforts have ended up in publications, but the discussions we’ve had while trying to implement different biological assays in droplets have been one of the favorite parts of my work and I’ve learnt a lot in the process. Maria, Sara and Håkan for proof-reading my thesis and coming with great comments and suggestions for improvements. Aman-a, thank you for reviewing my thesis and for teaching me when to duck. Former and present Dropies: Prem, Sara, Martin, Staffan, Maria, Petter, Narendar, Yunpeng, Ali, Krzystof and Håkan. Thank you all for creating such a great work environment and for always having time for discussions about research or malfunctioning/nonfunctioning droplet stations. The Nanobio group. This has been such a nice place to work thanks to all of you. Skiing in Romme, Boda Borg, board game nights, so many fun (often disgusting) lunch conversations, dinners and AWs. Ett speciellt tack till mina kontorsgrannar Sahar och Sara som har funnits där med trevligt sällskap när jag behövt en paus.

Lovisa Söderberg

57

To all my conference and course companions: Prem, Philippa, Lara, Jonas, Staffan, Emilie, Sanja, Ida, Athanasia. We had a lot of fun on our trips, like having dumplings for breakfast in China and hiking on the Great Wall, beers in Germany, the memorable road trip in Vermont to the Ben and Jerries’ factory and a Räkmacka in Göteborg. Ken, Anna-Luisa, Joesfine, Tarek, Micke, Anna och alla andra på gamla plan 3 i Albanova. Ni gav mig en fantastisk start på min doktorandtid och jag kan fortfarande tycka att det är väldigt tråkigt att Nanobiogruppen flyttade ifrån er. Mina Höglandstjejer! Maria, Sussie, Sara, Hanna, Anna Erika och Lisa. Jag tycker att det är så kul att vi har hängt ihop under alla år. Tack för att ni stod med mig när allt jag pratade om var celler och bubblor. Jag har aldrig så roligt som när jag är med er på resor, temamiddagar eller festkvällar och jag hoppas att vi hittar på mycket mer kul i framtiden. Och till sist, tack till min familj för att ni alltid finns där och ställer upp för mig. Till Erik och Jennie för alla dagar vi har spenderar försjunkna i ett eller annat brädspel. Mamma och Pappa, tack för att ni alltid tror på mig oavsett vad jag tar mig för. För att ni lyssnar när jag behöver klaga över PCR reaktioner som inte fungerar eller celler som inte beter sig, även om ni inte riktigt förstår vad jag pratar om. Tack för jag har fått följa med på alla era resor runt hela jorden. Och nu på sista tiden för det extra curlandet och allt annat ni gör för mig. Tack!


58

Lovisa Söderberg

59

References

[1] G. M. Whitesides, “The origins and the future of microfluidics.,” Nature, vol. 442, no. 7101, pp. 368–73, 2006.

[2] P. J. A. Kenis, R. F. Ismagilov, and G. M. Whitesides, “Microfabrication Inside Capillaries Using Multiphase Laminar Flow Patterning,” Science (80-. )., vol. 285, no. July, pp. 83–86, 1999.

[3] T. M. Squires and S. R. Quake, “Microfluidics : Fluid physics at the nanoliter scale,” Rev. Mod. Phys., vol. 77, no. July, pp. 977–1026, 2005.

[4] M. A. Unger, H. Chou, T. Thorsen, A. Scherer, and S. R. Quake, “Monolithic Microfabricated Valves and Pumps by Multilayer Soft Lithography,” Science (80-. )., vol. 288, no. April, pp. 113–117, 2000.

[5] S. E. Hulme, S. S. Shevkoplyas, and G. M. Whitesides, “Incorporation of prefabricated screw, pneumatic, and solenoid valves into microfluidic devices,” Lab Chip, vol. 9, pp. 79–86, 2009.

[6] A. B. Theberge, F. Courtois, Y. Schaerli, M. Fischlechner, C. Abell, F. Hollfelder, and W. T. S. Huck, “Microdroplets in microfluidics: an evolving platform for discoveries in chemistry and biology.,” Angew. Chem. Int. Ed. Engl., vol. 49, no. 34, pp. 5846–68, Aug. 2010.

[7] H. Song, D. L. Chen, and R. F. Ismagilov, “Reactions in droplets in microfluidic channels.,” Angew. Chem. Int. Ed. Engl., vol. 45, no. 44, pp. 7336–56, Nov. 2006.

[8] J.-C. Baret, “Surfactants in droplet-based microfluidics.,” Lab Chip, vol. 12, no. 3, pp. 422–33, Feb. 2012.

[9] C. Holtze, a C. Rowat, J. J. Agresti, J. B. Hutchison, F. E. Angilè, C. H. J. Schmitz, S. Köster, H. Duan, K. J. Humphry, R. a Scanga, J. S. Johnson, D. Pisignano, and D. a Weitz, “Biocompatible surfactants for water-in-fluorocarbon emulsions.,” Lab Chip, vol. 8, no. 10, pp. 1632–9, Oct. 2008.

[10] J. Bibette, D. C. Morse, A. T. Witteveen, and D. A. Weitz, “Stability-Criteria for Emulsions,” Phys. Rev. Lett., vol. 69, no. 16, pp. 2439–2442, 1992.

[11] S. Ariyaprakai and S. R. Dungan, “Journal of Colloid and Interface Science Influence of surfactant structure on the contribution of micelles to Ostwald ripening in oil-in-water emulsions,” J. Colloid Interface Sci., vol. 343, no. 1, pp. 102–108, 2010.

[12] P. Taylor, “Ostwald ripening in emulsions,” Colloids and Surfaces, vol. 99, pp. 175–185, 1995.

[13] F. J. Ghadessy, J. L. Ong, and P. Holliger, “Directed evolution of polymerase function by compartmentalized self-replication,” Proc Natl Acad Sci U S A, vol. 98, no. 8, pp. 4552–4557, 2001.

[14] S. L. Sjostrom, Y. Bai, M. Huang, Z. Liu, J. Nielsen, H. N. Joensson, and H. Andersson Svahn, “High-throughput screening for industrial enzyme production hosts by droplet microfluidics.,”


60

Lab Chip, vol. 14, no. 4, pp. 806–13, Jan. 2014.

[15] P. Hammar, S. A. Angermayr, S. L. Sjostrom, J. Van Der Meer, and K. J. Hellingwerf, “Single-cell screening of photosynthetic growth and lactate production by cyanobacteria,” Biotechnol. Biofuels, vol. 8, no. December, pp. 1–8, 2015.

[16] I. Giaever and C. R. Keese, “Behavior of cells at fluid interfaces.,” Proc. Natl. Acad. Sci. U. S. A., vol. 80, no. 1, pp. 219–222, 1983.

[17] a Huebner, M. Srisa-Art, D. Holt, C. Abell, F. Hollfelder, a J. deMello, and J. B. Edel, “Quantitative detection of protein expression in single cells using droplet microfluidics.,” Chem. Commun. (Camb)., vol. 2, no. 12, pp. 1218–20, Mar. 2007.

[18] S. M. Bjork, S. L. Sjostroma, H. Andersson-Svahn, and H. N. Joenssona, “Metabolite profiling of microfluidic cell culture conditions for droplet based screening,” Biomicrofluidics, vol. 9, no. 44128, pp. 1–10, 2015.

[19] S. Köster, F. E. Angilè, H. Duan, J. J. Agresti, A. Wintner, C. Schmitz, A. C. Rowat, C. a Merten, D. Pisignano, A. D. Griffiths, and D. a Weitz, “Drop-based microfluidic devices for encapsulation of single cells.,” Lab Chip, vol. 8, no. 7, pp. 1110–5, Jul. 2008.

[20] P. K. Periyannan Rajeswari, H. N. Joensson, and H. Andersson-svahn, “Droplet size influences division of mammalian cell factories in droplet microfluidic cultivation,” Electrophoresis, vol. 38, pp. 305–310, 2017.

[21] J. Clausell-Tormos, D. Lieber, J. C. Baret, A. El-Harrak, O. J. Miller, L. Frenz, J. Blouwolff, K. J. Humphry, S. Köster, H. Duan, C. Holtze, D. A. Weitz, A. D. Griffiths, and C. A. Merten, “Droplet-Based Microfluidic Platforms for the Encapsulation and Screening of Mammalian Cells and Multicellular Organisms,” Chem. Biol., vol. 15, no. 5, pp. 427–437, 2008.

[22] E. Borgström, D. Redin, S. Lundin, E. Berglund, A. F. Andersson, and A. Ahmadian, “Phasing of single DNA molecules by massively parallel barcoding,” Nat. Commun., vol. 6, no. 7173, pp. 1–6, 2015.

[23] T. Thorsen, R. W. Roberts, F. H. Arnold, and S. R. Quake, “Dynamic Pattern Formation in a Vesicle-Generating Microfluidic Device,” Phys. Rev. Lett., vol. 86, no. 18, pp. 4163–4166, Apr. 2001.

[24] S. L. Anna, N. Bontoux, and H. a. Stone, “Formation of dispersions using ‘flow focusing’ in microchannels,” Appl. Phys. Lett., vol. 82, no. 3, p. 364, 2003.

[25] M. M. Kiss, L. Ortoleva-Donnelly, N. R. Beer, J. Warner, C. G. Bailey, B. W. Colston, J. M. Rothberg, D. R. Link, and J. H. Leamon, “High-Throughput Quantitative Polymerase Chain Reaction in Picoliter Droplets,” Anal. Chem., vol. 80, no. 23, pp. 8975–8981, Dec. 2008.

[26] L. Frenz, K. Blank, E. Brouzes, and A. D. Griffiths, “Reliable microfluidic on-chip incubation of droplets in delay-lines.,” Lab Chip, vol. 9, no. 10, pp. 1344–8, 2009.

[27] S. L. Sjostrom, H. N. Joensson, and H. A. Svahn, “Multiplex analysis of enzyme kinetics and inhibition by droplet microfluidics using picoinjectors.,” Lab Chip, vol. 13, no. 9, pp. 1754–61,

Lovisa Söderberg

61

May 2013.

[28] N. R. Beer, E. K. Wheeler, L. Lee-houghton, N. Watkins, S. Nasarabadi, N. Hebert, P. Leung, D. W. Arnold, C. G. Bailey, and B. W. Colston, “Reverse-Transcription PCR in Isolated Picoliter Droplets,” Anal. Chem., vol. 80, no. 6, pp. 1854–1858, 2008.

[29] C. H. J. Schmitz, A. C. Rowat, S. Köster, and D. A. Weitz, “Dropspots: a picoliter array in a microfluidic device,” Lab Chip, vol. 9, no. 1, pp. 44–49, 2009.

[30] M. A. Khorshidi, P. Kumar, P. Rajeswari, H. N. Joensson, and H. A. Svahn, “Automated analysis of dynamic behavior of single cells in picoliter droplets,” Lab Chip, 2014.

[31] D. Pekin, Y. Skhiri, J.-C. Baret, D. Le Corre, L. Mazutis, C. Ben Salem, F. Millot, A. El Harrak, J. B. Hutchison, J. W. Larson, D. R. Link, P. Laurent-Puig, A. D. Griffiths, and V. Taly, “Quantitative and sensitive detection of rare mutations using droplet-based microfluidics.,” Lab Chip, vol. 11, no. 13, pp. 2156–66, Jul. 2011.

[32] D. J. Eastburn, Y. Huang, M. Pellegrino, A. Sciambi, L. J. Ptácek, and a. R. Abate, “Microfluidic droplet enrichment for targeted sequencing,” Nucleic Acids Res., pp. 1–8, 2015.

[33] D. Link, S. Anna, D. Weitz, and H. Stone, “Geometrically Mediated Breakup of Drops in Microfluidic Devices,” Phys. Rev. Lett., vol. 92, no. 5, pp. 1–4, Feb. 2004.

[34] D. J. Eastburn, A. Sciambi, and A. R. Abate, “Ultrahigh-Throughput Mammalian Single-Cell Reverse-Transcriptase Polymerase Chain Reaction in Microfluidic Drops,” Anal. Chem., no. 85, pp. 8016–8021, 2013.

[35] C. Priest, S. Herminghaus, and R. Seemann, “Controlled electrocoalescence in microfluidics: Targeting a single lamella,” Appl. Phys. Lett., vol. 89, no. 13, p. 134101, 2006.

[36] W. L. Olbricht and D. M. Kung, “The interaction and coalescence of liquid drops in flow through a capillary tube,” J. Colloid Interface Sci., vol. 120, no. 1, pp. 229–244, 1987.

[37] E. Brouzes, M. Medkova, N. Savenelli, D. Marran, M. Twardowski, J. B. Hutchison, J. M. Rothberg, D. R. Link, N. Perrimon, and M. L. Samuels, “Droplet microfluidic technology for single-cell high-throughput screening.,” Proc. Natl. Acad. Sci. U. S. A., vol. 106, no. 34, pp. 14195–200, Aug. 2009.

[38] J. J. Agresti, E. Antipov, A. R. Abate, K. Ahn, A. C. Rowat, J.-C. Baret, M. Marquez, A. M. Klibanov, A. D. Griffiths, and D. a Weitz, “Ultrahigh-throughput screening in drop-based microfluidics for directed evolution.,” Proc. Natl. Acad. Sci. U. S. A., vol. 107, no. 9, pp. 4004–9, Mar. 2010.

[39] A. R. Abate, T. Hung, P. Mary, J. J. Agresti, and D. a Weitz, “High-throughput injection with microfluidics using picoinjectors.,” Proc. Natl. Acad. Sci. U. S. A., vol. 107, no. 45, pp. 19163–6, Nov. 2010.

[40] D. J. Eastburn, A. Sciambi, and A. R. Abate, “Picoinjection Enables Digital Detection of RNA with Droplet RT-PCR,” PLoS One, vol. 8, no. 4, p. e62961, 2013.


62

[41] J.-C. Baret, O. J. Miller, V. Taly, M. Ryckelynck, A. El-harrak, L. Frenz, C. Rick, M. L. Samuels, J. B. Hutchison, J. J. Agresti, D. R. Link, D. A. Weitz, and A. D. Griffiths, “Fluorescence-activated droplet sorting ( FADS ): efficient microfluidic cell sorting based on enzymatic activity,” Lab Chip, vol. 9, no. 13, pp. 1850–1858, 2009.

[42] K. Ahn, C. Kerbage, T. P. Hunt, R. M. Westervelt, D. R. Link, and D. a. Weitz, “Dielectrophoretic manipulation of drops for high-speed microfluidic sorting devices,” Appl. Phys. Lett., vol. 88, no. 2, p. 24104, 2006.

[43] S. Becker, H.-U. Schmoldt, T. M. Adams, S. Wilhelm, and H. Kolmar, “Ultra-high-throughput screening based on cell-surface display and fluorescence-activated cell sorting for the identification of novel biocatalysts.,” Curr. Opin. Biotechnol., vol. 15, no. 4, pp. 323–9, Aug. 2004.

[44] D. Frenzel and C. A. Merten, “Microfluidic train station: highly robust and multiplexable sorting of droplets on electric rails,” Lab Chip, vol. 17, no. 6, pp. 1024–1030, 2017.

[45] H. N. Joensson, M. Uhlén, and H. A. Svahn, “Droplet size based separation by deterministic lateral displacement-separating droplets by cell--induced shrinking.,” Lab Chip, vol. 11, no. 7, pp. 1305–1310, 2011.

[46] A. C. Hatch, J. S. Fisher, A. R. Tovar, A. T. Hsieh, R. Lin, S. L. Pentoney, D. L. Yang, and A. P. Lee, “1-Million droplet array with wide-field fluorescence imaging for digital PCR,” Lab Chip, no. 11, pp. 3838–3845, 2011.

[47] S. E. Barnes, Z. T. Cygan, J. K. Yates, K. L. Beers, and E. J. Amis, “Raman spectroscopic monitoring of droplet polymerization in a microfluidic device.,” Analyst, vol. 131, no. 9, pp. 1027–33, 2006.

[48] H. S. Kim, S. C. Waqued, D. T. Nodurft, T. P. Devarenne, V. V. Yakovlev, A. Han, W. Xu, M. Trtilek, J. Xu, and B. Ma, “Raman spectroscopy compatible PDMS droplet microfluidic culture and analysis platform towards on-chip lipidomics,” Analyst, vol. 142, pp. 1054–1060, 2017.

[49] S. K. Küster, S. R. Fagerer, P. E. Verboket, K. Eyer, K. Jefimovs, R. Zenobi, and P. S. Dittrich, “Interfacing droplet microfluidics with matrix-assisted laser desorption/ionization mass spectrometry: Label-free content analysis of single droplets,” Anal. Chem., vol. 85, no. 3, pp. 1285–1289, 2013.

[50] N. Gasilova, Q. Yu, L. Qiao, and H. H. Girault, “On-Chip Spyhole Mass Spectrometry for Droplet-Based Microfluidics,” Angew. Chemie Int. Ed., vol. 53, no. 17, pp. 4408–4412, 2014.

[51] E. W. M. Kemna, L. I. Segerink, F. Wolbers, I. Vermes, and A. van den Berg, “Label-free, high-throughput, electrical detection of cells in droplets,” Analyst, vol. 138, pp. 4585–4592, 2013.

[52] P. K. Isgor, M. Marcali, M. Keser, and C. Elbuken, “Microfluidic droplet content detection using integrated capacitive sensors,” Sensors Actuators B. Chem., vol. 210, pp. 669–675, 2015.

[53] L. Mu, Y. Chang, S. D. Sawtelle, M. Wipf, X. Duan, and M. A. Reed, “Silicon Nanowire Field-Effect Transistors—A Versatile Class of Potentiometric Nanobiosensors,” IEEE Access, vol. 3,

Lovisa Söderberg

63

pp. 287–302, 2015.

[54] J. F. Fennell, S. F. Liu, J. M. Azzarelli, J. G. Weis, S. Rochat, K. A. Mirica, J. B. Ravnsbæk, and T. M. Swager, “Nanowire Chemical/Biological Sensors: Status and a Roadmap for the Future,” Angew. Chemie - Int. Ed., vol. 55, pp. 1266–1281, 2015.

[55] A. Zhang and C. M. Lieber, “Nano-Bioelectronics,” Chem. Rev., vol. 116, no. 1, pp. 215–257, Jan. 2016.

[56] E. Stern, A. Vacic, C. Li, F. N. Ishikawa, C. Zhou, M. A. Reed, and T. M. Fahmy, “A Nanoelectronic Enzyme-Linked Immunosorbent Assay for Detection of Proteins in Physiological Solutions,” Small, vol. 6, no. 2, pp. 232–238, 2010.

[57] L. Mu, I. a Droujinine, N. K. Rajan, S. D. Sawtelle, and M. a Reed, “Direct, Rapid, and Label-Free Detection of Enzyme–Substrate Interactions in Physiological Buffers Using CMOS-Compatible Nanoribbon Sensors,” Nano Lett., vol. 14, no. 9, pp. 5315–5322, Sep. 2014.

[58] J. C. McDonald, D. C. Duffy, J. R. Anderson, D. T. Chiu, H. Wu, O. J. A. Schueller, and G. M. Whitesides, “Fabrication of microfluidic systems in poly(dimethylsiloxane),” Electrophoresis, vol. 21, no. 1, pp. 27–40, 2000.

[59] D. Qin, Y. Xia, and G. M. Whitesides, “Soft lithography for micro- and nanoscale patterning.,” Nat. Protoc., vol. 5, no. 3, pp. 491–502, 2010.

[60] J. Zhang, K. L. Tan, G. D. Hong, L. J. Yang, and H. Q. Gong, “Polymerization optimization of SU-8 photoresist and its applications in microfluidic systems and MEMS,” 2001.

[61] D. C. Duffy, J. C. McDonald, O. J. Schueller, and G. M. Whitesides, “Rapid Prototyping of Microfluidic Systems in Poly(dimethylsiloxane).,” Anal. Chem., vol. 70, no. 23, pp. 4974–84, Dec. 1998.

[62] A. Sciambi and A. R. Abate, “Generating electric fields in PDMS microfluidic devices with salt water electrodes,” Lab Chip, vol. 14, pp. 2605–2609, 2014.

[63] F. H. C. Crick, “Central Dogma of Molecular Biology,” Nature, vol. 227, pp. 561–563, 1970.

[64] F. H. C. Crick, “On protein Synthesis,” Symp Soc Exp Biol, vol. 12, pp. 138–163, 1958.

[65] M. B. Gerstein, C. Bruce, J. S. Rozowsky, D. Zheng, J. Du, J. O. Korbel, O. Emanuelsson, Z. D. Zhang, S. Weissman, and M. Snyder, “What is a gene , post-ENCODE ? History and updated definition,” Genome Res., vol. 17, pp. 669–681, 2007.

[66] S. Djebali, C. A. Davis, A. Merkel, A. Dobin, T. Lassmann, A. Mortazavi, A. Tanzer, J. Lagarde, W. Lin, F. Schlesinger, C. Xue, G. K. Marinov, J. Khatun, B. A. Williams, F. Kokocinski, R. F. Abdelhamid, T. Alioto, I. Antoshechkin, D. Gonzalez, A. Gordon, H. Gunawardena, C. Howald, S. Jha, R. Johnson, P. Kapranov, B. Risk, D. Robyr, M. Sammeth, L. Schaffer, L. See, A. Shahab, and J. Skancke, “Landscape of transcription in human cells,” Nature, vol. 489, pp. 101–108, 2012.

[67] The 1000 Genomes Project Consortium, “A global reference for human genetic variation,”


64

Nature, vol. 526, pp. 68–74, 2015.

[68] M. Zarrei, J. R. Macdonald, D. Merico, and S. W. Scherer, “A copy number variation map of the human genome,” Nat. Publ. Gr., vol. 16, no. 3, pp. 172–183, 2015.

[69] J. C. Chang, E. C. Wooten, A. Tsimelzon, S. G. Hilsenbeck, M. C. Gutierrez, R. Elledge, S. Mohsin, C. K. Osborne, G. C. Chamness, D. C. Allred, and P. O. Connell, “Gene expression profiling for the prediction of therapeutic response to docetaxel in patients with breast cancer,” Lancet, vol. 362, pp. 362–369, 2003.

[70] D. Baltimore, “RNA-dependent DNA polymerase in virions of RNA tumour viruses..pdf,” Nature, vol. 226, pp. 1209–1211, 1970.

[71] H. M. Temin and S. Mizutani, “Viral RNA-dependent DNA Polymerase- RNA-dependent DNA Polymerase in Virions of Rous Sarcoma Virus.pdf,” Nature, vol. 226, pp. 1211–1213, 1970.

[72] I. M. Verma, G. F. Temple, H. Fan, and D. Balitimore, “In vitro Synthesis of DNA Complementary to Rabbit Reticulocyte 10S RNA,” Nature, vol. 235, pp. 163–167, 1972.

[73] D. L. Kacian and S. Spiegelman, “In vitro Synthesis of DNA Components of Human Genes for Globins,” Nature, vol. 235, pp. 167–169, 1972.

[74] I. M. Mackay, K. E. Arden, and A. Nitsche, “Real-time PCR in virology,” Nucleic Acids Res., vol. 30, no. 6, pp. 1292–1305, 2002.

[75] R. Molenkamp, A. Van Der Ham, J. Schinkel, and M. Beld, “Simultaneous detection of five different DNA targets by real-time Taqman PCR using the Roche LightCycler480 : Application in viral molecular diagnostics,” J. Virol. Methods, vol. 141, pp. 205–211, 2007.

[76] B. Vogelstein and K. W. Kinzler, “Digital PCR,” Proc Natl Acad Sci U S A, vol. 96, no. August, pp. 9236–9241, 1999.

[77] M. Baker, “Digital PCR hits its stride,” Nat. Methods, vol. 9, no. 6, pp. 541–544, May 2012.

[78] “QuantStudio 3D Digital PCR System Thermofisher.” [Online]. Available: https://www.thermofisher.com/se/en/home/life-science/pcr/digital-pcr/quantstudio-3d-digital-pcr-system.html. [Accessed: 02-Oct-2017].

[79] “Fluigm Biomark HD.” [Online]. Available: https://www.fluidigm.com/products/biomark-hd-system. [Accessed: 02-Oct-2017].

[80] D. Liu, S. L. Daubendiek, M. A. Zillman, K. Ryan, E. T. Kool, N. York, and R. V August, “Rolling Circle DNA Synthesis : Small Circular Oligonucleotides as Efficient Templates for DNA Polymerases,” pp. 1587–1594, 1996.

[81] F. B. Dean, J. R. Nelson, T. L. Giesler, and R. S. Lasken, “Rapid Amplification of Plasmid and Phage DNA Using Phi29 DNA Polymerase and Multiply-Primed Rolling Circle Amplification,” pp. 1095–1099, 2001.

Lovisa Söderberg

65

[82] Y. Fu, C. Li, S. Lu, W. Zhou, F. Tang, X. S. Xie, and Y. Huang, “Uniform and accurate single-cell sequencing based on emulsion whole-genome amplification,” PNAS, vol. 112, no. 38, pp. 11923–11928, 2015.

[83] M. Hammond, F. Homa, H. Andersson-svahn, T. J. G. Ettema, and H. N. Joensson, “Picodroplet partitioned whole genome amplification of low biomass samples preserves genomic diversity for metagenomic analysis,” Microbiome, vol. 4, no. 52, pp. 1–8, 2016.

[84] T. Notomi, H. Okayama, H. Masubuchi, T. Yonekawa, K. Watanabe, N. Amino, and T. Hase, “Loop-mediated isothermal amplification ( LAMP ) of gene sequences and simple visual detection of products,” Nucleic Acids Res., vol. 3, no. 5, 2000.

[85] T. Rane, TD; Chen, L; Zec, HC; Wang, “Microfluidic continuous flow digital loop-mediated isothermal amplification (LAMP),” Lab Chip, vol. 15, no. 3, pp. 776–782, 2014.

[86] “Sequencing Platforms - Compare NGS platforms.” [Online]. Available: https://www.illumina.com/systems/sequencing-platforms.html. [Accessed: 16-Sep-2017].

[87] “Pacbio long read sequencing.” [Online]. Available: http://www.pacb.com/smrt-science/smrt-sequencing/read-lengths/. [Accessed: 16-Sep-2017].

[88] “Oxford Nanopore Sequencing.” [Online]. Available: https://nanoporetech.com/applications/dna-nanopore-sequencing. [Accessed: 16-Sep-2017].

[89] N. R. Beer, B. J. Hindson, E. K. Wheeler, S. B. Hall, K. A. Rose, I. M. Kennedy, and B. W. Colston, “On-Chip, Real-Time, Single-Copy Polymerase Chain Reaction in Picoliter Droplets,” Anal. Chem., vol. 79, no. 22, pp. 8471–8475, 2007.

[90] D. J. Eastburn, A. Sciambi, and A. R. Abate, “Ultrahigh-Throughput Mammalian Single-Cell Reverse-Transcriptase Polymerase Chain Reaction in Microfluidic Drops,” Anal. Chem., vol. 85, pp. 8016–8021, 2013.

[91] V. Taly, D. Pekin, L. Benhaim, S. K. Kotsopoulos, D. Le Corre, X. Li, I. Atochin, D. R. Link, A. D. Griffiths, K. Pallier, H. Blons, O. Bouché, B. Landi, J. B. Hutchison, and P. Laurent-Puig, “Multiplex picodroplet digital PCR to detect KRAS mutations in circulating DNA from the plasma of colorectal cancer patients.,” Clin. Chem., vol. 59, no. 12, pp. 1722–31, Dec. 2013.

[92] Q. Zhong, S. Bhattacharya, S. Kotsopoulos, J. Olson, V. Taly, A. D. Griffiths, D. R. Link, and J. W. Larson, “Multiplex digital PCR: breaking the one target per color barrier of quantitative PCR.,” Lab Chip, vol. 11, no. 13, pp. 2167–74, Jul. 2011.

[93] P. Kumaresan, C. J. Yang, S. A. Cronier, R. G. Blazej, and R. A. Mathies, “High-throughput single copy DNA amplification and cell analysis in engineered nanoliter droplets,” Anal. Chem., vol. 80, no. 10, pp. 3522–3529, 2008.

[94] Y. Zeng, R. Novak, J. Shuga, M. T. Smith, and R. A. Mathies, “High-Perfromance Single Cell Genetic Analysis Using Microfluidic Emulsion Generator Arrays,” Anal. Chem., vol. 82, no. 8, pp. 3183–3190, 2010.


66

[95] E. Stieglitz, C. B. Troup, L. C. Gelston, J. Haliburton, E. D. Chow, K. B. Yu, J. Akutagawa, A. N. Taylor-weiner, Y. L. Liu, Y. Wang, K. Beckman, P. D. Emanuel, B. S. Braun, A. Abate, R. B. Gerbing, T. a Alonzo, and M. L. Loh, “Subclonal mutations in SETBP1 confer a poor prognosis in juvenile myelomonocytic leukemia,” Blood, vol. 125, no. 3, pp. 516–525, 2015.

[96] “Biorad - Droplet Digital PCR system.” [Online]. Available: http://www.bio-rad.com/en-cn/product/qx200-droplet-digital-pcr-system. [Accessed: 02-Oct-2017].

[97] A. S. Devonshire, M.-O. Baradez, G. Morley, D. Marshall, and C. a Foy, “Validation of high-throughput single cell analysis methodology.,” Anal. Biochem., vol. 452, pp. 103–13, May 2014.

[98] “RainDance Technologies - RainDrop Digital PCR system.” [Online]. Available: http://raindancetech.com/digital-pcr-tech/raindrop-digital-pcr-system/. [Accessed: 02-Oct-2017].

[99] A. R. Abate, T. Hung, R. a Sperling, P. Mary, A. Rotem, J. J. Agresti, M. a Weiner, and D. a Weitz, “DNA sequence analysis with droplet-based microfluidics.,” Lab Chip, vol. 13, no. 24, pp. 4864–9, Nov. 2013.

[100] R. Tewhey, J. B. Warner, M. Nakano, B. Libby, M. Medkova, P. H. David, S. K. Kotsopoulos, M. L. Samuels, J. B. Hutchison, J. W. Larson, E. J. Topol, M. P. Weiner, O. Harismendy, J. Olson, D. R. Link, and K. a Frazer, “Microdroplet-based PCR enrichment for large-scale targeted sequencing.,” Nat. Biotechnol., vol. 27, no. 11, pp. 1025–1031, 2009.

[101] F. Lan, J. R. Haliburton, A. Yuan, and A. R. Abate, “Droplet barcoding for massively parallel single-molecule deep sequencing,” Nat. Commun., vol. 7, p. 11784, 2016.

[102] “10Xgenomics genome solution.” [Online]. Available: https://www.10xgenomics.com/genome/. [Accessed: 28-Sep-2017].

[103] F. Tang, K. Lao, and M. A. Surani, “Development and applications of single-cell transcriptome analysis.,” Nat. Methods, vol. 8, no. 4 Suppl, pp. S6-11, Apr. 2011.

[104] A. Ståhlberg, V. Rusnakova, and M. Kubista, “The added value of single-cell gene expression profiling.,” Brief. Funct. Genomics, pp. 1–9, Feb. 2013.

[105] P. Dalerba, T. Kalisky, D. Sahoo, P. S. Rajendran, M. E. Rothenberg, A. A. Leyrat, S. Sim, J. Okamoto, D. M. Johnston, D. Qian, M. Zabala, J. Bueno, N. F. Neff, J. Wang, A. A. Shelton, B. Visser, S. Hisamori, Y. Shimono, M. Van De Wetering, H. Clevers, M. F. Clarke, and S. R. Quake, “Single-cell dissection of transcriptional heterogeneity in human colon tumors,” Nat. Biotechnol., vol. 29, no. 12, pp. 1120–1127, 2011.

[106] A. A. Powell, A. H. Talasaz, H. Zhang, M. A. Coram, A. Reddy, G. Deng, M. L. Telli, R. H. Advani, R. W. Carlson, J. A. Mollick, S. Sheth, A. W. Kurian, J. M. Ford, F. E. Stockdale, S. R. Quake, R. F. Pease, M. N. Mindrinos, G. Bhanot, S. H. Dairkee, R. W. Davis, and S. S. Jeffrey, “Single Cell Profiling of Circulating Tumor Cells : Transcriptional Heterogeneity and Diversity from Breast Cancer Cell Lines,” PLoS One, vol. 7, no. 5, p. e33788, 2012.

[107] M. B. Elowitz, A. J. Levine, E. D. Siggia, and P. S. Swain, “Stochastic gene expression in a single cell.,” Science, vol. 297, no. 5584, pp. 1183–6, Aug. 2002.

Lovisa Söderberg

67

[108] S. Lindström and H. Andersson-Svahn, “Miniaturization of biological assays -- overview on microwell devices for single-cell analyses.,” Biochim. Biophys. Acta, vol. 1810, no. 3, pp. 308–16, Mar. 2011.

[109] “Fluidigm C1.” [Online]. Available: https://www.fluidigm.com/products/c1-system#components. [Accessed: 19-Sep-2017].

[110] A. R. Wu, N. F. Neff, T. Kalisky, P. Dalerba, B. Treutlein, M. E. Rothenberg, F. M. Mburu, G. L. Mantalas, S. Sim, M. F. Clarke, and S. R. Quake, “Quantitative assessment of single-cell RNA-sequencing methods,” Nat. Methods, no. october, pp. 1–8, Oct. 2013.

[111] H. C. Fan, G. K. Fu, and S. P. a. Fodor, “Combinatorial labeling of single cells for gene expression cytometry,” Science (80-. )., vol. 347, no. 6222, pp. 1258367–1258367, Feb. 2015.

[112] P. S. Daugherty, B. L. Iverson, and G. Georgiou, “Flow cytometric screening of cell-based libraries,” J. Immunol. Methods, vol. 243, no. 1–2, pp. 211–227, Sep. 2000.

[113] J. Hedman and P. Rådström, “PCR Detection of Microbial Pathogens: Second Edition, Methods in Molecular Biology,” in Methods in Molecular Biology, vol. 943, M. Wilks, Ed. Totowa, NJ: Humana Press, 2013, pp. 17–48.

[114] A. K. White, K. A. Heyries, C. Doolin, M. Vaninsberghe, and C. L. Hansen, “High-Throughput Microfluidic Single-Cell Digital Polymerase Chain Reaction,” Anal. Chem., vol. 85, pp. 7182–7190, 2013.

[115] D. J. Eastburn, A. Sciambi, and A. R. Abate, “Identification and genetic analysis of cancer cells with PCR-activated cell sorting.,” Nucleic Acids Res., vol. 42, no. 16, pp. 1–10, Jul. 2014.

[116] A. Rakszewska, R. J. Stolper, A. B. Kolasa, A. Piruska, and W. T. S. Huck, “Quantitative Single-Cell mRNA Analysis in Hydrogel Beads,” Angew. Chemie - Int. Ed., pp. 1–5, 2016.

[117] E. Z. Macosko, A. Basu, R. Satija, J. Nemesh, K. Shekhar, M. Goldman, I. Tirosh, A. R. Bialas, N. Kamitaki, E. M. Martersteck, J. J. Trombetta, D. A. Weitz, J. R. Sanes, A. K. Shalek, A. Regev, and S. A. McCarroll, “Highly Parallel Genome-wide Expression Profiling of Individual Cells Using Nanoliter Droplets,” Cell, vol. 161, no. 5, pp. 1202–1214, 2015.

[118] A. M. Klein, L. Mazutis, I. Akartuna, N. Tallapragada, A. Veres, V. Li, L. Peshkin, D. A. Weitz, and M. W. Kirschner, “Droplet Barcoding for Single-Cell Transcriptomics Applied to Embryonic Stem Cells,” Cell, vol. 161, no. 5, pp. 1187–1201, 2015.

[119] “10x genomics 3’ single cell solution.” [Online]. Available: https://www.10xgenomics.com/single-cell/. [Accessed: 21-Sep-2017].

[120] “ddseq Illumina Bio-rad.” [Online]. Available: https://info.bio-rad.com/ww-ddseq.html?source_wt=ddseq_surl. [Accessed: 21-Sep-2017].

[121] G. Raposo and W. Stoorvogel, “Extracellular vesicles: Exosomes, microvesicles, and friends,” J. Cell Biol., vol. 200, no. 4, pp. 373–383, 2013.

[122] R. Kalluri, “The biology and function of exosomes in cancer,” J. Clin. Invest., vol. 126, no. 4,


68

pp. 1208–1215, 2016.

[123] A. Zomer, C. Maynard, F. J. Verweij, A. Kamermans, R. Sch??fer, E. Beerling, R. M. Schiffelers, E. De Wit, J. Berenguer, S. I. J. Ellenbroek, T. Wurdinger, D. M. Pegtel, and J. Van Rheenen, “In vivo imaging reveals extracellular vesicle-mediated phenocopying of metastatic behavior,” Cell, vol. 161, no. 5, pp. 1046–1057, 2015.

[124] C. Kahlert, S. A. Melo, A. Protopopov, J. Tang, S. Seth, O. Koch, J. Zhang, J. Weitz, L. Chin, A. Futreal, and R. Kalluri, “Identification of doublestranded genomic dna spanning all chromosomes with mutated KRAS and P53 DNA in the serum exosomes of patients with pancreatic cancer,” J. Biol. Chem., vol. 289, no. 7, pp. 3869–3875, 2014.

[125] J. Skog, T. Würdinger, S. van Rijn, D. H. Meijer, L. Gainche, M. Sena-Esteves, W. T. Curry, B. S. Carter, A. M. Krichevsky, and X. O. Breakefield, “Glioblastoma microvesicles transport RNA and proteins that promote tumour growth and provide diagnostic biomarkers.,” Nat. Cell Biol., vol. 10, no. 12, pp. 1470–6, 2008.

[126] H. Valadi, K. Ekström, A. Bossios, M. Sjöstrand, J. J. Lee, and J. O. Lötvall, “Exosome-mediated transfer of mRNAs and microRNAs is a novel mechanism of genetic exchange between cells,” Nat. Cell Biol., vol. 9, no. 6, pp. 654–659, 2007.

[127] C. Kahlert and R. Kalluri, “Exosomes in tumor microenvironment influence cancer progression and metastasis,” J. Mol. Med., vol. 91, no. 4, pp. 431–437, 2013.

[128] D. D. Taylor and C. Gercel-taylor, “MicroRNA signatures of tumor-derived exosomes as diagnostic biomarkers of ovarian cancer,” Gynecol. Oncol., vol. 110, pp. 13–21, 2008.

[129] S. a. Melo, L. B. Luecke, C. Kahlert, A. F. Fernandez, S. T. Gammon, J. Kaye, V. S. Lebleu, E. a. Mittendorf, J. Weitz, N. Rahbari, C. Reissfelder, C. Pilarsky, M. F. Fraga, D. Piwnica-Worms, and R. Kalluri, “Glypican-1 identifies cancer exosomes and detects early pancreatic cancer,” Nature, vol. 523, no. 7559, pp. 177–182, 2015.

[130] P. K. Periyannan Rajeswari, L. M. Soderberg, A. Yacoub, M. Leijon, H. Andersson, and H. N. Joensson, “Multiple pathogen biomarker detection using an encoded bead array in droplet PCR,” J. Microbiol. Methods, vol. 139, no. March, pp. 22–28, 2017.

[131] L. M. Soderberg, H. Joensson, and H. A. Svahn, “Parallel cDNA synthesis from thousands of individually encapsulated cancer cells - Towards large scale single cell gene expression analysis,” in 17th International Conference on Miniaturized Systems for Chemistry and Life Sciences, 2013, vol. 3, pp. 1737–1739.

Lovisa Söderberg

69

Publications and manuscripts

Droplet Microfluidics reverse transcription and PCR ...1151596/FULLTEXT01.pdf · Droplet...

Documents

Transcript of Droplet Microfluidics reverse transcription and PCR ...1151596/FULLTEXT01.pdf · Droplet...