TargetedProteomics Matrix ... · 2020. 8. 31. · TECHNICAL BRIEF TargetedProteomics...

TECHNICAL BRIEFTargeted Proteomics www.clinical.proteomics-journal.com

Matrix-Assisted Laser Desorption Ionization MassSpectrometry Imaging of Key Proteins in Corneal Samplesfrom Lattice Dystrophy Patients with TGFBI-H626R andTGFBI-R124C Mutations

Anandalakshmi Venkatraman, Guillaume Hochart, David Bonnel, Jonathan Stauber,Shigeto Shimmura, Lakshminaryanan Rajamani, Konstantin Pervushin,and Jodhbir S. Mehta*

Scope: The purpose of this study is to identify and visualize the spatialdistribution of proteins present in amyloid corneal deposits of TGFBI-CDpatients using Mass Spectrometry Imaging (MSI) and compare it with healthycontrol cornea. Corneal Dystrophies (CD) constitute a group of geneticallyinherited protein aggregation disorders that affects different layers of thecornea. With accumulated protein deposition, the cornea becomes opaquewith decreased visual acuity. CD affecting the stroma and Bowman’smembrane, is associated with mutations in transforming growth factorβ-induced (TGFBI) gene.Methods: MALDI-Mass Spectrometry Imaging (MSI) is performed on 2patient corneas and is compared with 1 healthy control cornea using a7T-MALDI-FTICR. Molecular images obtained are overlaid with congo-redstained sections to visualize the proteins associated with the corneal amyloidaggregates.Results: MALDI-MSI provides a relative abundance and two dimensionalspatial protein signature of key proteins (TGFBIp, Apolipoprotein A-I,Apolipoprotein A-IV, Apolipoprotein E, Kaliocin-1, Pyruvate Kinase and Rasrelated protein Rab-10) in the patient deposits compared to the control. Thisis the first report of the anatomical localization of key proteins on cornealtissue section from CD patients. This may provide insight in understandingthe mechanism of amyloid fibril formation in TGFBI-corneal dystrophy.

Dr. A. Venkatraman, Dr. L. Rajamani, Dr. J. S. Mehta11 Third Hospital Avenue, 168751, SingaporeE-mail: [email protected]. A. Venkatraman, Dr. K. PervushinSchool of Biological SciencesNanyang Technological University637551, SingaporeG. Hochart, Dr. D. Bonnel, Dr. J. StauberImaBiotechParc Eurasante885 Avenue Eugene Avinee, 59120, Loos, France

DOI: 10.1002/prca.201800053

The human cornea is a transparent avas-cular tissue that forms the front partof the eye. There are five different lay-ers of cornea, namely epithelium, Bow-man’s membrane, stroma, Descemet’smembrane, and endothelium. Cornealdystrophy (CD) is a genetically inher-ited, bilateral disorder that can affectcorneal clarity[1] with deposition of ag-gregated mutant protein. Stromal andBowman’s membrane CD is most com-monly associated with mutations intransforming growth factor beta-induced(TGFBI) gene. The protein product TGF-BIp is also known as keratoepithelin orBIGH3/βigh3, which is an extracellu-lar matrix protein and the second mostabundant protein in the human cornea.[2]

The protein is 68 kDa and has fourfasciclin-like (FAS1) domains. There aremore than 65 mutations reported inTGFBI that alter the proteolytic process-ing of mutant proteins, contributing totheir abnormal deposition.[3,4]

There have been several studiesthat have investigated and reportedthe protein composition of corneal

Dr. S. ShimmuraDepartment of OphthalmologyKeio University35 Shinanomachi, 160-8582, Shinjuku Tokyo, JapanDr. L. Rajamani, Dr. J. S. MehtaOphthalmology and Visual Sciences Academic Clinical ProgramDuke-NUS Graduate Medical School169857, SingaporeDr. J. S. MehtaSingapore National Eye Centre11 Third Hospital Avenue, 168751, Singapore

Proteomics Clin. Appl. 2019, 13, 1800053 C© 2018 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim1800053 (1 of 5)

http://www.clinical.proteomics-journal.com

mailto:[email protected]

www.advancedsciencenews.com www.clinical.proteomics-journal.com

Figure 1. Distribution of TGFBIp peptide (m/z 1315.7) in both patients and healthy control. The overlay of molecular images and Congo red stainingalso shows an abundance of TGFBIp in patient corneal aggregates.

aggregates from dystrophic patients with different TGFBImutations.[5] Our group has reported the protein composition ofcorneal aggregates from lattice corneal dystrophy (LCD) patientswithH626R and R124Cmutations using liquid chromatography-mass spectrometry/mass spectrometry (LC-MS/MS) methods.[6]

All studies have identified several amyloidogenic, non-amyloidfibrillar proteins and TGFBIp to be the major components of thecorneal aggregates and suggested the possibility of differences inproteolytic procession of mutant protein compared to wild type(WT).[3,5]

A few matrix-assisted laser desorption/ionization-massspectrometry imaging (MALDI-MSI) reports in the past havestudied regions of the eye like the human lens capsule andevaluated the drug-penetration properties in various regionsusing animal models.[7] In the current study, we optimizedprotocols by MALDI-Fourier-transform ion cyclotron resonance(FTICR) MSI that combined high-resolution power, molecularimaging, and proteomics with histology correlation capabilitiesto provide more information on corneal tissue proteins ofpatients with CD.[8] We observed MALDI-MSI key proteins

associated with corneal deposits from patients with R124Cand H626R mutations that are highly prevalent in SoutheastAsia.[9] The results are in agreement with previously reportedLC-MS/MS data and support the presence of localization ofkey proteins to the amyloid aggregates in LCD patients, withan advantage of easy sample preparation, using fewer tissuesections, and an understanding of the spatial distribution of themolecules.Corneal tissues were collected from two patients with LCD

(TGFBImutationsH626R [n= 1,male, age= 45] and R124C [n=1, female, age = 62]) who were undergoing corneal transplanta-tion at the Singapore National Eye Centre. Written informed con-sent was obtained from all patients prior to surgery. Transplant-grade control tissue (n = 1, male, age = 54) from healthy controlwas obtained from the Lions Eye Institute, Tampa, FL, USA. TheSingHealth Institutional Review Board granted ethical approvalfor the collection of patient’s cornea. The corneal sections werestained with Congo red, and the presence of dark red staining inthe corneal stroma indicated the presence of amyloid aggregatesin the patient samples (Figure S1A, Supporting Information).


http://www.advancedsciencenews.com



Figure 2. Distribution of proteins in corneal samples based on specific peptide sequence according to BLAST database.

The patient with R124C mutations had fewer amyloid depositscompared to the patient with H626R mutation, as visualized byCongo red staining.The corneas were fixed on solid Tissue-Tek Optimal Cut-

ting Temperature chemical (VWR, Fontenay-sous-Bois, France),sectioned at −35 °C with a cryomicrotome (Microm HM560,Thermo Scientific, Brignais, France). Sections, 10μm thick, weremounted on Indium-tin-oxide (ITO) glass slides (Delta Technolo-gies, Loveland, CO, USA), cryodesiccated for 1 h, and dried un-der vacuum in a desiccator for 20 min. Additional serial sec-tions were prepared on SuperFrost slides for staining with Congored, hematoxylin, and eosin. A six-step washing was appliedto the tissue sections on the ITO slides to enhance proteindetection.[10]

Porcine trypsin (Promega, Charbonnieres-les-Bains, France)at 20 μg mL−1 in 50 mm ammonium bicarbonate was sprayedusing SunCollect (SunChrom, Friedrichsdorf, Germany) (25 ˚C,10 μL min−1, 15 cycles) onto the tissue sections and standard

samples (human recombinant TGFBIp; Abcam, Paris, France).The ITO slide was then incubated overnight at 37 °C and driedunder vacuum for 15 min before matrix deposit. MALDI matrix(CHCA 10 mg mL−1, acetonitrile/0.2% TFA 125 mm, and am-monium sulfate 7:3 v/v were sprayed on the tissue sections us-ing the automatic TM sprayer device (nose temperature at 60 °C,100 μL min−1, 10 cycles) (HTX Imaging, Chapel Hill, North Car-olina). The presence of corneal protein deposits from patientswere visualized with Congo red staining (Figure S1, SupportingInformation) on serial sections according to procedures men-tioned in Table S1, Supporting Information, followed by mount-ing the slide with aqueous medium Aquatex and drying beforehigh-definition scanning ×20 with a HD pannoramic 250 flashII scanner (3DHistech, Budapest, Hungary). H&E staining waseither performed (as per Table S2, Supporting information) afterMALDI imaging of same sections or visualized on serial sections.MALDI-MSI of three human corneas (two patient samples

and one healthy control) was performed at high spatial resolution





Figure 3. Distribution of mutation-dependent proteins in patient corneas. The characteristic expression of proteins specific for each type of TGFBImutation.

(20 μm) using a 7T-MALDI-FTICR (SolariX, Bruker Daltonics,Bremen, Germany) with a Smartbeam II laser in minimummode with a repetition rate of 2000 Hz. Full scan positive modewithin 300–2000 Da mass range was selected for imaging withonline calibration using matrix peaks (mass accuracy <1 ppm).The mass spectrum obtained for each position of the images cor-responded to the averaged mass spectra of 300 consecutive lasershots on the same location. FTMS Control 2.0 and FlexImaging4.1 software packages (Bruker Daltonics, Bremen, Germany)were used to set and control imaging and mass spectrometerparameters. Standard t-test (SCiLS) andWelch t-tests (Multimag-ing) between different regions were carried out using imagingdata sets according to the workflow described in Figure S2,Supporting Information. Molecular images were overlaid withCongo red images and regions of interest (ROI) were selected.To identify major differences between regions, a ratio above 3between conditions and p-value < 0.001 were imposed to selectbest hits.The on-tissue digestion was checked with peptide mass

fingerprint (PMF) spectra. The spectrum profile indicated agood digestion of tissue proteins (Figure S3, Supporting Infor-mation). A differential approach was implemented to identifymajor differences between regions from control and diseasedcorneas, leading to a list ofm/z of interest (Figure S3, SupportingInformation). Based on this list and expected protein digestsfrom complementary LC-MS/MS investigations, a mass accu-racy below 10 ppm was used for querying the databases (Mascot,BLAST). Peptide sequences were correlated and most of the se-quenceswere found to be specific for proteins listed in the BLAST(Figure S4, Supporting Information) and UniProtKB/Swiss-Prot

database. The implementation of all the peptide m/z in Mascotthough did not allow obtaining significant scores as per TGFBIpin Figure S5, Supporting Information. The untargeted approachto analyzing differential expression of proteins between CDpatient samples and control samples resulted in several sig-nificant molecular differences between the two groups. Thespectrum profiles readily matched five specific proteins thatwere confirmed by BLAST analysis. The m/z values of two otherpeptides that showed a significant difference between the patientand control groups (m/z 1094.514 and m/z 894.533) could notbe assigned to known proteins from the MASCOT database.Five proteins that matched the specific sequences (Figure 2

A–O) were TGFBIp (specific m/z 838.489 [ALPPRER], m/z787.474 [VIGTNRK], m/z 929.508 [IPSETLNR], m/z 1179.579[GIFPVLCK], m/z 1315.802 [LTLLAPLNSVFK]), apolipoproteinE (specific m/z 1131.614 [EQVAEVR] and m/z 830.438 [TR-DRLDEVK]), pyruvate kinase (specific m/z 876.5 [GIFPVLCK]and m/z 1235.616 [EMIKSGMNVAR]), apolipoprotein A-IV(specific m/z 1058.6 [KNAEELKAR]), and apolipoprotein A-I(specific m/z 603.3 [KLNTQ]). TGFBIp protein was confirmedby fragmentation using MS/MS using quadrupole collisioninduced dissociation (qCID) fragmentation mode and com-pared with human TGFBIp standard. The overlay of molecularimages generated by MSI and Congo red staining for TGFBIppeptide (m/z 1315.792) showed that TGFBIp (Figure 1 D–F)was expressed uniformly in control, whereas in the patientwith H626R mutation, it was expressed more toward anteriorstroma, and for R124C patient, it was localized toward thecentral and posterior stroma. In most places, the protein wasfound to be concentrated with amyloid deposits, highlighting





the heterogeneity of the peptide distribution in different regionsof the corneal sections in these small histological regions asdemonstrated by the magnified image for the TGFBIp peptideat m/z 1315.792 (Figure 1G,H).MSI images showed that apolipoprotein E (m/z 1131.614) was

absent in control and expressed only in patient samples. Forthe R124C corneal samples, the peptide overlapped with Congored stained areas, confirming the presence of apolipoprotein Ein the amyloid deposits. The MSI images for pyruvate kinase(m/z 876.5), apolipoprotein A-IV (m/z 1058.6), and apolipopro-tein A-I (m/z 603.3) showed that the peptide was uniformlydistributed in control, whereas in patients, it was found inabundance around the area of amyloid deposits (Figure 2G–O). Two proteins, kaliocin-1 (m/z 820.4) and ras-related pro-tein Rab-10 (m/z 801.446) showed characteristic expression inonly one type of the mutations (Figure 3). Kaliocin was specif-ically expressed in the patient with TGFBI-H626R mutation,and ras-related protein Rab-10 was specific to TGFBI-R124Cmutation.This is the first study to optimize and report MALDI-MSI

on corneal tissue samples from CD patients. However, thenumber of patient corneas available for research limits the study.Notwithstanding the limitation, we identified the differentialexpression of seven proteins in two patients with differentmutations, which colocalized with the amyloid protein depositsusing a single tissue section from each patient. We have alsoreported the presence of proteins specific to a type of TGFBImu-tation. The specific expression of proteins related to a particularmutation needs further validations on a number of patients withthe same mutation. The presence of the differential expressionof these proteins was validated by mass accuracy and the speci-ficity of their digests (peptides) according to the databases. Theconfirmed expression of these proteins in the corneal tissue maybe validated in future by additional fragmentation data and im-munohistochemistry techniques if highly specific antibodies areavailable.

Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe author.

AcknowledgementsV.A. and G.H. contributed equally toward the publication. The au-thors would like to acknowledge the funding support from SNEC-HREFR1265/71/2015 and SNEC-HREF R1376/62/2016 to J.S.M.

Conflict of InterestThe authors declare no conflict of interest.

Keywordscorneal dystrophy,matrix-assisted laser desorption/ionization,mass spec-trometry imaging, TGFBIp

Received: March 9, 2018Revised: October 23, 2018

Published online: November 15, 2018

[1] a) G. K. Klintworth,Orphanet J. Rare Dis. 2009, 4, 7; b) A. J. Aldave, B.Sonmez, Arch. Ophthalmol. 2007, 125, 177.

[2] a) G. K. Klintworth, Front. Biosci. 2003, 8, d687; b) G. K. Klintworth,Z. Valnickova, J. J. Enghild, Am. J. Pathol. 1998, 152, 743; c) T. F. Dyr-lund, E. T. Poulsen, C. Scavenius, C. L. Nikolajsen, I. B. Thogersen,H. Vorum, J. J. Enghild, J. Proteome Res. 2012, 11, 4231.

[3] E. Korvatska, J. Biol. Chem. 2000, 275, 11465.[4] H. Karring, K. Runager, Z. Valnickova, I. B. Thøgersen, T. Møller-

Pedersen, G. K. Klintworth, J. J. Enghild, Exp. Eye Res. 2010, 90, 57.[5] a) H. Karring, K. Runager, I. B. Thøgersen, G. K. Klintworth, P. Højrup,

J. J. Enghild, Exp. Eye Res. 2012, 96, 163; b) H. Karring, E. T. Poulsen,K. Runager, I. B. Thøgersen, G. K. Klintworth, P. Højrup, J. J. Enghild,Mol. Vis. 2013, 19, 861; c) E. T. Poulsen, K.Runager, M. W. Risør, T.F. Dyrlund, C. Scavenius, H. Karring, J. Praetorius, H. Vorum, D. E.Otzen, G. K. Klintworth, J. J. Enghild, Proteomics: Clin. Appl. 2014,8, 168; d) D. G. Courtney, E. Toftgaard Poulsen, S. Kennedy, J. E.Moore, S. D. Atkinson, E. Maurizi, M. A. Nesbit, C. B. T. Moore, J.J. Enghild, Invest. Opthalmol. Vis. Sci. 2015, 56, 4653; e) E. T. Poulsen,N. S. Nielsen, M. M. Jensen, E. Nielsen, J. Hjortdal, E. K. Kim, J. J.Enghild, Proteomics 2016, 16, 539.

[6] A. Venkatraman, B. Dutta, E. Murugan, H. Piliang, R. Lakshmi-naryanan, A. C. Sook Yee, K. V. Pervushin, S. K. Sze, J. S. Mehta, J.Proteome Res. 2017, 16, 2899.

[7] a) K. J. Grove, V. Kansara, M. Prentiss, D. Long, M. Mogi, S. Kim, P. J.Rudewicz, Y. Chen, J. V. Jester, D. M. Anderson, S. A. Marchitti, K. L.Schey, D. C. Thompson, V. Vasiliou, P. A.-O. Vinciguerra, R.Mencucci,V. A.-O. Romano, E. A.-O. Spoerl, F. I. Camesasca, E. Favuzza, C. A.-O.Azzolini, R. A.-O. Mastropasqua, R. A.-O. Vinciguerra, D. M. Drexler,L. Tannehill-Gregg Sh Fau-Wang, B. J. Wang L Fau-Brock, B. J. Brock;M. Ronci, S. Sharma, T. Chataway, K. P. Burdon, S. Martin, J. E. Craig,N. H. Voelcker, J. Proteome Res. 2011, 10, 3522; b) D.M. Drexler, S. H.Tannehill-Gregg, L. Wang, B. J. Brock, J. Pharmacol. Toxicol. Methods2011, 63, 205.

[8] R. M. Caprioli, T. B. Farmer, J. Gile, Anal. Chem. 1997, 69, 4751.[9] H. M. Chau, Br. J. Ophthalmol. 2003, 87, 686.[10] M. Gessel, J. M. Spraggins, P. Voziyan, B. G. Hudson, R. M. Caprioli,

J. Mass Spectrom. 2015, 50, 1288.




TargetedProteomics Matrix ... · 2020. 8. 31. · TECHNICAL BRIEF TargetedProteomics...

Documents

Transcript of TargetedProteomics Matrix ... · 2020. 8. 31. · TECHNICAL BRIEF TargetedProteomics...