Uncovering the “Hidden Fibrosis” In Human Aortic Valve Development and Disease via Novel

Extracellular Matrix ApproachesCassandra L. Clift1, David Bichell2, Yan Ru Su2, Jennifer Bethard1, Susana Comte-Walters1, Lauren E. Ball1, Anand Mehta1, Richard R. Drake1, Peggi M. Angel1

1. Dept of Cell and Molecular Pharmacology & Experimental Therapeutics, Medical University of South Carolina, Charleston, SC

2. Dept of Medicine, Division of Pediatric Cardiology, Vanderbilt University, Nashville, TN

Hypothesis and Clinical Application

Introduction: The Hidden Fibrosis of Congenital Aortic Valve Stenosis

Acknowledgements➢CLC is supported by a T32 from

NHLBI (HL007260). This work was

supported by 16GRNT31380005

(American Heart Association), P20

GM103542 (NIH/NIGMS).

Methods: LC-MS/MS Proteomics and MALDI Imaging Mass Spectrometry

➢Hypothesis: Hydroxylated prolines at very specific sites in collagen sequences are determinants of the

progression of valvular dysfunction that results in pediatric congenital aortic valve stenosis.

➢ Clinical Application: A main challenge in fibrosis research is accessing the extracellular matrix (ECM) to

understand how ECM proteins yield feedback information promoting disease progression. Preferentially

accessing the ECM via these novel proteomic and MALDI-IMS techniques may lead to simplified

pathological evaluations in the clinic as well as identification of 1) novel fibrosis biomarkers; 2) novel

upstream regulators of fibrosis; and 3) potential targets for pharmacotherapeutic intervention.

➢ Formalin Fixed Paraffin Embedded (FFPE) aortic valve tissues from 26 pediatric patients were obtained through Vanderbilt Core

Laboratory for Cardiovascular Translational and Clinical Research and through the National Disease Research Interchange and

used for research at MUSC. De-identified tissues are linked with preoperative valve function defined via echocardiogram.

A: Aortic Valve Tissue Cohort

Conclusions and Future Directions

➢ LC-MS/MS analysis was done via

data dependent acquisition on an

Orbitrap Elite mass spectrometer

equipped with a LC Packings

U3000 nano-LC system (Thermo).

➢ CID Tandem mass spectra were

searched using MaxQuant version

1.6.3.3 (7). A subset database was

created to re-search MS/MS

spectra for the dynamic

modification of Oxidation (P). A

peptide probability of 99% was

used.

➢ MALDI-IMS analysis was via a 7.0

Tesla solariX™ Legacy FT-ICR

(Bruker Scientific, LLC) operated in

positive ion broadband mode

spanning m/z 500-4000.

➢ Images were analyzed using

SCiLS (Bruker Daltonics).

➢ MALDI-IMS data and LC-MS/MS

data were linked by accurate mass

within ±5 ppm mass accuracy.

B: MALDI IMS and High-Resolution Accurate Mass Proteomics (LC-MS/MS)

HRAM Proteomics

A

Figure 2. It is currently unknown how hydroxylation of

collagen sequences influences aortic valve development

and contributes to collagen misalignment in CAVS.

?Collagen triple

helix

Tissue

preparation

ECM Digest

Acquire

ECM Spectra

(6) Shoulders, M., et al., Ann. Rev. of Biochemistry 2009, 78:929-58

➢ ECM proteins and their PTMs are preferentially accessed using a novel proteomic method.

➢ Proteomic work reported unique changes in hydroxyproline HYP) of collagen in CAVS compared to normal AV.

COL3a1 HYP mapping showed the normal valve contains 13% more HYP sites than in CAVS valve. This is

indicative of reduced collagen stability in CAVS and may effect cell-ECM interactions.

➢ Novel information on localization of collagen HYP by MALDI Imaging Mass Spectrometry is anticipated to drive

new studies on signaling factors of collagen organization and deposition in congenital aortic valve stenosis, as

well as elucidate mechanisms behind pediatric end-stage CAVS.

➢ These and future studies aim to 1) report the regulation of collagen types in human AV development and

childhood disease; 2) establish foundational work for studying in situ collagen PTMs in fibrotic cardiovascular

disease; and 3) generate new research directions in the treatment of human heart valve disease.

B

Result A: ECM peptides are discretely regulated amongst pediatric and adult CAVS

Result B: Collagen HYP is reduced in CAVS; Novel upstream regulator identified

Collagenase III DigestDetaches ECM peptides

Incubate overnight37°C, shaking

STAGE Tip CleanupSolid phase extraction of ECM

peptides

Acquire Mass Spectra

(LC-MS/MS)

PNGase F Digest and

DeglycosylationDetaches & removes N-glycans

Heat Induced Epitope

Retrieval (HIER)10mM Tris pH 9

Formalin fixed paraffin

embedded tissue (FFPE)Heat 1 Hr 60°C

Dewax

MALDI-Imaging Mass Spec

Collagenase III applicationDetaches ECM peptides

Incubation 5 hrs>80% RH, 37°C

MALDI matrix application

Image data acquisition

MALDI-FTICR

(High mass accuracy & sensitivity)

Heat Induced Epitope Retrieval

(HIER)10mM Tris pH 9

Formalin fixed paraffin embedded

tissue (FFPE)Heat 1 Hr 60°C

Dewax

PNGase F Digest and

DeglycosylationDetaches & removes N-glycans

VIM Vimentin

ECM Proteins Upregulated in Disease vs. Normal ANOVA Significant Peptides Clustered to Patient Category

Figure 1. A healthy valve has a trilayer structure of

elastin, GAGs, and collagen. These proteins are

abundantly post-translationally modified and crosslinked.

Contact Info

Cassandra Cliftcliftc@musc.edu

Ph.D. Candidate

Biomedical Sciences

Mentors: Dr. Peggi Angel, Chair

Dr. Richard Drake, Co-chair

Medical University of South Carolina

Normal

CAVS

PV by Ross

Adult FAVS

AVI

Hydroxylated Proline

Age

Created with BioRender

Tricuspid

Bicuspid (CAVS)

Stratified ECM

Mixed ECM

Ao

Ao

120/10

200/25

120/80

110/70

➢Congenital aortic valve stenosis (CAVS)

occurs at a rate of over 13.9 in 1000 U.S.

births (1). Bicuspid AV thickens with ECM in

pediatrics, altering hemodynamics, leading to

heart failure (Fig. 1).

➢ A hallmark of pediatric CAVS is ECM

deregulation (Fig. 1, 2b), however

translational and post-translational regulation

of these critical proteins remains mostly

unknown (2).

➢ A PTM of interest in this study is

hydroxylation of proline (HYP). HYP is critical

to proper stabilization of collagen’s triple

helical structure (Fig. 2a).

➢Here, we use a novel proteomic approach

that allows us to target the collagen

proteome (4,5).

➢Understanding how collagen PTMs are

deregulated at pediatric end-stage CAVS

may identify new non-surgical therapeutic

targets that inhibit disease progression,

especially in pediatric cases where surgical

intervention is limited.

Aortic Valve Extracellular Matrix

2 Mo

2 Yr11 Yr

8 Mo

Healthy AV

DevelopmentPediatric Endstage

CAVS

Movat’s Pentachrome Collagen GAGs Elastin Nuclei

Bar = 200 µm

Muscle

2 Mo

2 Yr11 Yr

8 Mo

Healthy AV

DevelopmentPediatric Endstage

CAVS

Movat’s Pentachrome Collagen GAGs Elastin Nuclei

Bar = 200 µm

MuscleElastin Collagen GAGs Nuclei Muscle

C: COLase3 Proteomics: Protein Identifications and ECM ClassificationCOL1A1 Collagen alpha-1(I)

COL1A2 Collagen alpha-2(I)

COL3A1 Collagen alpha-1(IIII)

COL4A1 Collagen alpha-1(IV)

COL4A2 Collagen alpha-2(IV)

COL5A1 Collagen alpha-1(V)

COL5A2 Collagen alpha-2(V)

COL6A1 Collagen alpha-1(VI)

COL6A2 Collagen alpha-2(VI)

COL6A3 Collagen alpha-3(VI)

COL6A6 Collagen alpha-6(VI)

COL11A1 Collagen alpha-1(XI)

COL12A1 Collagen alpha-1(XII)

COL14A1 Collagen alpha-1(XIV)

COL15A1 Collagen alpha-1(XV)

COL16A1 Collagen alpha-1(XVI)

COL18A1 Collagen alpha-1(XVIII)

COL21A1 Collagen alpha-1(XXI)

COMP

Cartilage oligomeric matrix

protein

EMIL1 EMILIN-1

FBN1 Fibrillin-1

FNC Fibronectin

MFAP4

Microfibril-associated

glycoprotein 4

TNC Tenascin

TNXB Tenascin-X

VTN Vitronectin

BGN Biglycan

DCN Decorin

FMOD Fibromodulin

HSPG2

Basement membrane-

specific heparan sulfate

proteoglycan core protein

LUM Lumican

NID2 Nidogen-2

ASPN Asporin

BGH3

Transforming growth-

factor-beta-induced ig-h3

FBLN1 Fibulin-1

DERM Dermatopontin

DSPP

Dentin

sialophosphoprotein

C1QA

Complement C1q

subcomponent subunit A

LOXL1 Lysyl oxidase homolog 1

MYH10 Myosin-10

SPARC SPARC

Principal Component Analysis

➢ Normal: Non-

diseased

➢ CAVS: Pediatric

CAVS

➢ AVI: Non-bicuspid

Aortic Valve

Insufficiency

➢ FAVS: Adult

Fibrocalcific AVS

(right)

➢ Unbiased clustering

techniques show

pediatric CAVS and

adult FAVS consist of

unique proteomes

➢ Proteomic

differences between

CAVS and normal

may be more subtle,

suggesting the need

for both translational

studies and post-

translational studies

of collagen–type

regulation.

Fibril-Type Collagen 3A1 HYP Mapping

➢ (above) Log2FC scatterplot shows differentially dysregulated collagen peptides in CAVS

have reduced hydroxyproline (HYP) content (13% reduction over all collagen types).

➢ (top right) Loss of HYP in CAVS corresponds to integrin and glycoprotein vi binding sites,

suggesting CAVS HYP reduction may effect cell-ECM interactions and platelet aggregation.

*Glycoprotein VI Binding Site (HYP Required) * Integrin Binding Site (HYP Required)

N

N

C

C

References

(1) Hoffman, J.I.E.; J Amer. College of Cardio. 2002, 39:1890

(2) Cox, T.R. et. al., Dis Model Mech 2011, 4(2): 165-178

(3) Gilkes D.M., et. al., J Biol Chem 2013, 288(15):10819-29

(4) Angel, P.M., et. al., Proteomics Clin Appl 2018, epub

(5) Angel, P.M., et. al., J Proteome Research 2018, 17(1):635-646

(6) Shoulders, M., et al., Ann Rev of Biochemistry 2009, 78:929-58

(7) Cox, J. Mann, M. Nature. Biotechnology. 2008, 26, 1367-1372

BAMBI: Master Regulator of Collagen Interactome Identified

9.5

10.0

10.5

11.0

11.5

12.0

12.5

BAMBINorm

aliz

ed R

ela

tive I

nte

nsitie

s

Normal CAVS

*Protein-Protein

Interaction Network

RNA-Seq Confirmation

Immunohistochemistry Confirmation

➢ BAMBI (BMP and activin membrane bound inhibitor) is a

negative regulator of TGF-beta and is downregulated in

CAVS valves, suggesting a mechanisms for overactive TGF-

beta signaling and a potential target in AV fibrosis research.

Result C: Serial Enzymatic Digest Elucidates the Complete Extracellular Matrisome

Enzyme 1: Chondroitinase (Glycosaminoglycans)

Enzyme 2: PNGaseF (N-Glycans)

Movat’s Pentachrome

Elastin

Collagen

GAGs

Enzyme 3: Elastase (Elastin)

Enzyme 4: Collagenase (Collagen Subtypes and PTMs)

Collagens

ECM

Glycoproteins

ECM

Proteoglycans

Collagen

Homology

Collagen

Chaperone

Intermediate

Filament

ECM

Crosslinking

Actin

Binding

➢ P* indicates a hydroxyproline containing peptide

➢ This multimodal approach may be useful for deep mining of protein

information from difficult to obtain tissues, as one tissue section can be

evaluated for 1) pathology via histology 2) glycosaminoglycans, 3) N-

glycans, 4) Elastin, and 5) Collagen and ECM Peptides.