Comprehensive Myocardial Proteogenomics ... - ACS Publications

Subscriber access provided by EDINBURGH UNIVERSITY LIBRARY | @ http://www.lib.ed.ac.uk

Article

Comprehensive Myocardial Proteogenomics Profiling Reveals C/EBP# as the Key Factor in the Lipid Storage of ARVC Liang Chen, Fan Yang, Xiao Chen, Man Rao, Ningning Zhang, Kai Chen, Haiteng Deng, Jiangping Song, and Shengshou Hu J. Proteome Res., Just Accepted Manuscript • DOI: 10.1021/acs.jproteome.7b00165 • Publication Date (Web): 30 Jun 2017 Downloaded from http://pubs.acs.org on July 2, 2017

Just Accepted “Just Accepted” manuscripts have been peer-reviewed and accepted for publication. They are posted online prior to technical editing, formatting for publication and author proofing. The American Chemical Society provides “Just Accepted” as a free service to the research community to expedite the dissemination of scientific material as soon as possible after acceptance. “Just Accepted” manuscripts appear in full in PDF format accompanied by an HTML abstract. “Just Accepted” manuscripts have been fully peer reviewed, but should not be considered the official version of record. They are accessible to all readers and citable by the Digital Object Identifier (DOI®). “Just Accepted” is an optional service offered to authors. Therefore, the “Just Accepted” Web site may not include all articles that will be published in the journal. After a manuscript is technically edited and formatted, it will be removed from the “Just Accepted” Web site and published as an ASAP article. Note that technical editing may introduce minor changes to the manuscript text and/or graphics which could affect content, and all legal disclaimers and ethical guidelines that apply to the journal pertain. ACS cannot be held responsible for errors or consequences arising from the use of information contained in these “Just Accepted” manuscripts.

Journal of Proteome Research is published by the American Chemical Society. 1155 Sixteenth Street N.W., Washington, DC 20036 Published by American Chemical Society. Copyright © American Chemical Society. However, no copyright claim is made to original U.S. Government works, or works produced by employees of any Commonwealth realm Crown government in the course of their duties.

Page 1 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Journal of Proteome Research

Comprehensive Myocardial Proteogenomics Profiling Reveals C/EBPα as the Key Factor in the Lipid Storage of ARVC Liang Chen1‡; Fan Yang2‡; Xiao Chen1‡; Man Rao1; Ning-Ning Zhang1; Kai Chen1; HaiTeng Deng2; Jiang-Ping Song1*; Sheng-Shou Hu1*

1. State Key Laboratory of Cardiovascular Disease, Fuwai Hospital, National Center for Cardiovascular Diseases, Chinese Academy of Medical Sciences and Peking Union Medical College, Beijing, China. 2. MOE Key Laboratory of Bioinformatics, School of Life Sciences, Tsinghua University, Beijing, China

ABSTRACT

Arrhythmogenic right ventriclar cardiomyopathy (ARVC) is hereditary cardiomyopathy characterized by the fibro-fatty replacement of the myocardium. A small number of noncomprehensive profiling studies based on human cardiac tissues have been conducted and reported; consequently, ARVC’s gene expression pattern characteristics remain largely undocumented. Our study applies large scaled, quantitative proteomics based on TMT-labeled LC-MS/MS to analyze the left and right ventricular myocardium of four ARVC and four DCM

ACS Paragon Plus Environment

1


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 2 of 49

explanted hearts to compare them with normal hearts. Our objective is to reveal the characteristic proteome pattern in ARVC compared to DCM as well as Non-diseased Heart. We also conducted the RNA sequencing of 10 right ventricles from ARVC hearts paired with four nondiseased donor hearts to validate the proteome results. In a manner similar to that of the welldefined DCM heart failure model, the ARVC model demonstrates the downregulation of mitochondrial function proteins, and the effects of many heart failure regulators such as TGFB, RICTOR and KDM5A. In addition, the inflammatory signaling, especially the complement system, was activated much more severely in ARVC than in DCM. Our most significant discovery was the lipid metabolism reprogramming of both ARVC ventricles in accordance with the upregulation of lipogenesis factors such as FABP4 and FASN. We identified the key upstream regulator of lipogenesis as C/EBPα. Transcriptome profiling verified the consistency with proteome alterations. This comprehensive proteogenomics profiling study reveals that an activation of C/EBPα, along with the upregulation of its lipogenesis targets, accounts for lipid storage and acts as a hallmark of ARVC.

KEYWORDS: Arrhythmogenic right ventricular cardiomyopathy; proteomics; proteogenomics; C/EBPα; lipid storage


2

Page 3 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


INTRODUCTION Arrhythmogenic right ventricular cardiomyopathy (ARVC) is a type of primary cardiomyopathy that exhibits the main features of right ventricular-based progressive fibrofatty substitution and malignant ventricular arrhythmia.1 ARVC is mainly an autosomal dominant disease caused by mutations in the desmosomal genes including: junction plakoglobin (JUP), desmoplakin (DSP), plakophilin 2 (PKP2), desmoglein 2 (DSG2), and desmocollin 2 (DSC2), which account for approximately 50% of probands; some other non-desmosomal genes like TMEM43, LMNA, PLN are also reported to be responsible for ARVC. 2 Although ARVC has large genetic heterogeneity, it has the characteristic, pathological change in fibro-fatty substitution which is considered to be the final common hallmark of this disease.3 Marian et al. first proposed the potential mechanisms of fibro-fatty replacement i.e., through different ways to inhibit the classical Wnt/β-catenin pathway, thereby activating fibrillogenic and lipogenic pathways,4,

5

while Kim et al. considered the activation of the

peroxisome proliferator-activated receptor gamma (PPARγ) pathway as the core mechanism.6 In addition, inflammation and apoptosis have been confirmed as participants in the pathogenic mechanism of ARVC/D.

6, 7

Previous studies, however, have mostly used in vivo mouse models

or an in vitro model for ARVC research that could provide insight in the mechanism and genetic pathogenicity to some extent, but have rarely reported on the specimens from ARVC patients. Only a few targeted genes profiling studies based on human ARVC myocardium, which screened several genes expression altered specific to ARVC and independent of the underlying genetic mutations, have been reported before.8, 9 Our group recently also reported on the first microRNA array of ARVC compared with normal donor heart.10 Thus, the molecular phenotypes as well as comprehensive gene expression pattern of ARVC myocardium are poorly understood.


3


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 4 of 49

Quantitative proteomics based on tandem mass tags (TMT) and LC-MS/MS are widely used to precisely detect the proteomes of cells, tissues, and body fluids with high sensitivity.11, 12 Because of their higher stability in the storage and sampling process, proteins are more reliable than RNAs when reflecting the molecular phenotype of the disease, but proteomics cannot detect sufficient targets like RNA profiling can, especially for low abundant proteins like transcription factors. Thus, a comprehensive proteomics profiling, combined with RNA sequencing method is necessary to reveal the pathophysiological phenotype of ARVC. Dilated cardiomyopathy (DCM) is the most common form of primary cardiomyopathy. It occurs mainly due to the formation of left ventricle-based heart failure, and sometimes, even the bi-ventricular heart failure is caused by complicated heredity and environmental factors and requires heart transplantation.13, 14 Because the explanted hearts (from both ARVC and DCM patients) are end-stage cardiac tissue, they share some common pathophysiological manifestations of heart failure. To distinguish and illustrate the common pathways of heart failure and characteristic changes of ARVC, we used the DCM explanted hearts as a positive control for heart failure, and non-diseased donor hearts as the normal control. Our study compared the similarities and differences in the results of the protein profiling of the left and right ventricular myocardium from ARVC, DCM, and donor hearts, to establish disease networks and identify the ARVC-related characteristic molecular phenotype and key pathways.


4

Page 5 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


EXPERIMENTAL METHODS Research Design and Sample Collection We selected the explanted hearts of four ARVC and four DCM patients, as well as 4 abandoned, non-diseased, donor hearts for quantitative proteomic analysis in our study. We also selected a total of 30 right ventricle (RV) specimens from 10 ARVC patients, 10 DCM patients, and 10 non-diseased donors for validation experiment (including the hearts for proteome profiling). In addition, the RVs of 10 ARVC hearts, and 4 out of the 10, non-diseased, donor hearts were used for RNA sequencing to validate the proteomic changes. All the explanted hearts were from the Heart Transplant Center of Fuwai Hospital. The non-diseased, donor hearts were selected from those donor hearts initially considered and processed for transplantation, but subsequently abandoned because of mismatched blood types or heart sizes. All the non-diseased donors had normal left ventricular function according to in-vivo evaluation and no history of myocardial disease. For handling these hearts, we immediately performed ex vivo cardiac perfusions on all the excised hearts with histidine-tryptophanketoglutarate (HTK) cardioplegia, followed by dissection on ice. The dissected full thicknesses of the myocardia from the left ventricles (LVs) and RVs were fixed in 10% formalin and preserved at room temperature. All the study procedures were approved by the Ethics Committee from Fuwai Hospital, Beijing, China. Prospective informed consent was obtained from each patient to use their heart tissue for research before heart transplantation. Each organ donor (or their relatives) also provided informed consent for organ donation. Sample Preparation for Quantitative Proteomics Analysis Quantitative proteomics based on liquid chromatography-tandem mass spectrometry (LCMS/MS) were applied to proteins prepared from the left and right ventricular tissues from the


5


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 6 of 49

ARVC, DCM and normal control hearts. The cardiac tissue was washed twice with PBS, and homogenized in 8 mol/L urea in PBS. The whole tissue extracts were centrifuged at 14,000 g for 30 minutes at 4℃. The protein concentrations were determined using the BCA protein assay kit (SolarBio, Beijing). Equal amounts of proteins from the diseased cardiac tissue and healthy cardiac tissue mixtures (200 µg) were reduced with 5 mM dithiothreitol (DTT) and alkylated with 12.5 mM iodoacetamide (IAM). The protein solutions were diluted to 1.5 M urea with PBS, followed by digestion using sequencing grade modified trypsin (Promega, Fitchburg, WI) of a 1:50 protease/protein ratio at 37℃ overnight. To quantitate the differentially expressed proteins in each sample, tandem-mass-tagging reagent (TMT, Thermo, Pierce Biotechnology, Rockford, IL) was applied to each ARVC and DCM sample, as well as the mixture of 4 normal hearts as the normal standard.15 The samples were desalted using Sep-Pak C18 Vac cartridges (Waters, Milford, MA). The extracts were centrifuged in SpeedVac to dry them completely. The purified peptides were resuspended in 100 µL of 100 mM triethyl ammonium bicarbonate (TEAB). Next, 20 µL of TMT labeling reagent was added to each peptide solution and allowed to react for one hour. In the ARVC RV experiment, peptides from the right ventricles of ARVC patients were labeled using TMT6-127, TMT6-128, TMT6-129, and TMT6-130 reagents (TMT, Thermo, Pierce Biotechnology, Rockford, IL). The peptides from the mixture of normal right ventricle proteins were labeled using TMT6-126 reagent. In the ARVC LV experiment, peptides from the left ventricles of ARVC patients were labeled with TMT6-127, TMT6-128, TMT6-129, and TMT6-130 reagents. The peptides from the mixture of normal left ventricle proteins were labeled using TMT6-126 reagent. In the DCM RV experiment, peptides from the right ventricles of DCM patients were labeled using TMT6-127, TMT6-128, TMT6-129, and TMT6-130 reagents. The peptides from


6

Page 7 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


the mixture of normal right ventricle proteins were labeled with TMT6-131 reagent. In the DCM LV experiment, peptides from the left ventricles of DCM patients were labeled using TMT6-127, TMT6-128, TMT6-129, and TMT6-130 reagent. The peptides from the mixture of normal left ventricle proteins were labeled with TMT6-131 reagent. Next, 5% hydroxylamine was used to quench the reactions for 15 minutes. The TMT-labeled peptides were mixed and desalted using Sep-Pak C18 Vac cartridges. The isobaric tag-labeled peptides were then fractioned using HPLC and sent to LC-MS/MS for analysis.16 The peptide mixture were fractioned by a UPLC 3000 system (Thermo-Fisher Scientific, Waltham, MA) with a XBridge C18 RP column (5 µm, 150 Å, 250 mm × 4.6 mm i.d., (Waters, Milford, MA)). Mobile phase A consisted 98% H2O and 2% acetonitrile (pH 10.0). Mobile phase B consisted 98% acetonitrile and 2% H2O (pH 10.0). Ammonium hydroxide was added to raise the pH to 10. Peptides were separated with the following gradients: 5% to 8% B, 5 min; 8% to 18% B, 25 min; 18%-32% B, 32 min; 32%-95% B 6min; 95%-5% B, 5 min. Peptide peaks were detected by 280 nm absorbance, 48 fractions were collected, combined into 12 fractions after drying by speedvac, and then resolved in 0.1% formic acid. Proteomics analysis based on LC-MS/MS For LC-MS/MS analysis, the TMT-labeled peptides were separated by a 120 min gradient elution at a flow rate 0.250 µL/min with a Thermo-Dionex Ultimate 3000 HPLC system. The analytical column was a homemade fused silica capillary column (75 µm i.d., 150 mm length; Upchurch, Oak Harbor, WA) packed with C18 resin (300Å, 5 µm; Varian, Lexington, MA). Mobile phase A consisted of 0.1% formic acid and mobile phase B consisted of 100% acetonitrile and 0.1% formic acid. An Q-Extractive mass spectrometer was operated in the


7


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 8 of 49

data-dependent acquisition mode using Xcalibur 3.0 software and there was a full-scan mass spectrum in the Orbitrap (300-1800 m/z, 70, 000 resolution) followed by 20 data-dependent MS/MS scans at 32% normalized collision energy (HCD). Ions with charge states of +2, +3，+4, +5 were selected for the high energy collisional dissociation. The mass window for precursor ion selection was 2.4 m/z, the threshold for triggering MS/MS experiments was 1.7×104，and the dynamic exclusion time was 20s. The MS/MS spectra from each LC-MS/MS run was cross-referenced against the UniProt human database (version January 10, 2015; 89105 sequences) using the SEQUEST search engine from Proteome Discoverer Software (version 1.4) (PD).17 The search criteria were as follows: full tryptic specificity was required; two missed cleavages were allowed; the oxidation (M) was set as the variable modification; carbamidomethylation (C) and TMTsixplex (N-terminal and K) were set as the fixed modifications; precursor ion mass tolerance was set at 20 ppm for all MS acquired in an Orbitrap mass analyzer; and the fragment ion mass tolerance was set at 20 mmu for all MS2 spectra acquired. An identified peptide with a confidence value of high was considered to be positive identification and peptide spectral matches (PSMs) were validated. The peptide false discovery rate (FDR) was estimated using the percolator function provided by PD, and the cutoff score was accepted at 1% based on the decoy database. Relative protein quantification was performed using PD software, according to the manufacturer’s instructions, and at the intensity of six TMT reporter ions per peptide. The protein ratios were calculated as the median of all peptide hits belonging to a protein. The proteins were identified through at least two unique peptides. The quantitative precision was expressed as protein ratio variability. Significant, differentially expressed proteins were determined using the two-tailed Student’s t-


8

Page 9 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


test. A p-value of less than 0.05 and a change ratio of greater than 1.3 were considered significant. RNA extraction, cDNA library construction and RNA sequencing Total RNA was prepared using Trizol according to manufacture instruction (Life Technologies, USA). Samples with RIN greater than 7.9 were submitted to prepare cDNA libraries. Library quality was assessed using the Agilent 2100 bioanalyzer followed by sequencing on the platform of Illumina Hiseq 2500. RNA extraction from liquid nitrogen cryoconserved heart tissues (~30mg tissue/sample) was performed with Trizol (Life Technologies Inc.) according to manufacturer’s instruction. RNA was assessed for quality, integrity and concentration measurement using Agilent 2100 Bioanalyzer and Qubit2.0. All samples subjected for RNA sequencing had RNA Integrity Number (RIN) greater than 7.0. Library preparation was carried out in rRNA-depletion combining with strand-specific manner. The brief procedure was described as follows: rRNAdepletion was performed with Ribo-ZeroTM Magnetic kit (Cat. No. MRZG12324); RNA was reverse transcribed into cDNA with reagent containing dUTP; then ends were polished, an “A” was overhanged at both ends and Illumina adaptors were ligated; cDNA were then sheared to an average size of 300~400 bp;. The complementary strand was degraded and 13 cycles of PCR reaction was implemented before loading on sequencer. Fragments were sequenced in a pairends fashion (2×125bp) on Illumina Hiseq2500 platform, with an average reads of 83 million reads per sample. RNA preparation and qRT-PCR


9


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 10 of 49

From each sample of myocardial tissue preserved in liquid nitrogen, 50 mg was removed and used for total RNA extraction using the Trizol reagent (Invitrogen, Carlsbad, CA). The specific protocols for RNA extraction, detection, and reverse transcription are described in previous study.18 An SYBR green-based PCR kit (Applied Biosystems, Foster City, CA) was used to detect the relative gene expression levels. The ribosomal protein L5 (RPL5) was used as the internal reference gene of the qRT-PCR. The ∆∆Ct method was used for the calculation of relative gene expression levels. Table S9 details the primer sequences for qRT-PCR in our study. Protein extraction and Western Blot Fifty milligrams of myocardial tissue from each sample preserved in liquid nitrogen were used for total protein extraction. The preparation of the proteins, SDS-PAGE, and subsequent procedures were conducted as previously reported.19 The antibodies included: C/EBPα（#2295 ，CST，USA); GAPDH（ab8245，abcam，UK); PPARγ（#2443，CST，USA）; FASN（ #3180，CST，USA）; and FABP4（#3544，CST，USA）. Immunofluorescence The distribution and expression of C/EBPα (ab15047, Abcam, UK) was detected using immunofluorescence by using the formalin-fixed ventricular tissues of the explanted and nondiseased donor hearts in a method previously reported.20 PCM-1(HPA023370, Sigma, USA) was co-stained as the specific marker of the cardiomyocyte nucleus.20, 21 The image was captured with a laser scanning confocal microscope (Zeiss, Germany). Bioinformatics Analysis


10

Page 11 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


To conduct the protein pathway analysis, differentially expressed proteins were analyzed using IPA software (Ingenuity Pathway Analysis, http://www.ingenuity.com, QIAGEN, USA). The canonical pathways, overrepresented biological functions, and molecular networks were generated based on the Ingenuity Knowledge Base. The protein networks were drawn using String software (www.string-db.org). To conduct the RNA sequencing analysis, we first implemented a quality control process, which trimmed the adaptors and filtered out low quality readings, before aligning the readings to the reference genome. The readings were aligned to human reference genome build hg19 with Hisat2 aligner software (version: 2.0.5; http://www.ccb.jhu.edu/software/hisat2/index.shtml).

22

The FPKM of the genes was calculated with Cufflinks software (Version:2.2.1; http://coletrapnell-lab.github.io/cufflinks/).23 The raw counts for each sample were generated with HT-seq software (version: 0.6.1; http://www-huber.embl.de/users/anders/HTSeq/).24 The differentially expressed

genes

were

detected

with

the

Bioconductor

package

DESeq2

(https://bioconductor.org/packages/release/bioc/html/DESeq2.html),25 and finally, to conduct the comparisons between patients and controls, we set the cutoff adjusted p value at < 0.05. The heatmap was generated with R plot software using the FPKM. Genetic Testing The whole genome of 8 cryopreserved, explanted heart tissues from ARVC and DCM patients were extracted and sequenced using targeted next-generation sequencing on Illumina Hiseq2000 platform (Illumina, USA) including PKP2, DSC2, DSG2, DSP, JUP, TMEM43, DES, LMNA, PLN and CTNNA3 according to the ARVC/D genetic variants database (http://www.arvcdatabase.info). The pathogenicity of variant was filtered and evaluated by


11


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 12 of 49

ACMG guideline26. And pathogenic, likely pathogenic and variant uncertain significance (VUS) mutations were identified. Statistical Analysis Continuous variables were presented as mean ± standard deviation. A non-parametric test was used for comparing continuous variables such as the signal intensity of Western Blot and qPCR. SPSS19.0 (SPSS, Inc., Chicago, IL) software was used for the statistical analysis in this study. Graphpad Prism 5.0 (La Jolla, CA) and R studio (R studio, USA) were used to prepare the graphs. All statistical calculations used a two-tailed test. P < 0.05 was considered statistically significant.


12

Page 13 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


RESULTS Clinical and genetic Characteristics Four patients (three males, mean age 35.3±8.0 years) were clinically diagnosed as suffering from ARVC which presented with more than two of major criteria listed in the 2010 Task Force Criteria.27 The patients received heart transplantations because of heart failure and malignant ventricular arrhythmias (Table 1, Table S1). By pathological examination, all four ARVC patients were confirmed to have over 50% fibro-fatty replacement in right ventricles according to full-thickness ventricular tissue analysis, instead of endomyocardial biopsy (Figure 1). They were proved to be of the classical RV dominant type with fibro-fatty replacement and dilation (Figure 1). Four DCM patients (two males, mean age 49.25±13.03 years) receiving heart transplantation were also enrolled. These DCM patients mainly presented with LV dilation, or biventricular involvement with fibro infiltration in the myocardial tissue (Figure 1). Through targeted sequencing, we identified likely pathogenic mutations in DSG2 and LMNA in 2 of 4 ARVC patients, respectively, whereas 4 DCM patients had no ARVC related pathogenic or likely pathogenic gene mutation, with one patient carrying VUS mutation in JUP (Table S2). In addition, four non-diseased donor hearts were used as the normal control and their tissue mixture was characterized as the internal reference in the proteomics analysis. Proteomic Analysis Revealed the Integrated View of the Proteome Pattern in Both Ventricles from the ARVC and DCM Patients We performed proteomic analysis on the left and right ventricular tissues of the ARVC and DCM patients, respectively (Figure 2A). The total number of detected proteins from each group was about 6,000, of which 204 to 614 classified as differential expression proteins (DEPs) when


13


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 14 of 49

compared to the normal standard (Figure 2B and Table S3). In a manner consistent with the cluster of the overall protein expression heat map (Figure S1A), the principal component analysis (PCA) indicated that the ARVC and DCM ventricles were completely divided into two groups with different protein expression patterns (Figure 2C). The proteins of the ARVC RV and ARVC LV groups presented similar variation tendencies (p < 0.001, R = 0.68), which indicated that both ventricles of ARVC hearts were undergoing similar pathophysiological alterations, although no remarkable fibro-fatty replacement was observed in the LV of the ARVC group (Figure 2D). Conversely, the proteome patterns between the RV and LV of the DCM were not significantly correlated (Figure 2E). The disease and function enrichment analysis showed similar alteration patterns in DCM and ARVC, with the exception of the lipid metabolism which was enriched in the ARVC, but not in the DCM group. This was in accordance with their pathological phenotypes, which tend to significant fibro-fatty infiltration of the ARVC (Figure S1 B-E). ARVC and DCM Shared Common Pathways of Heart Failure The canonical pathways analysis revealed that the proteome changes in the ARVC RV and DCM LV were commonly enriched in heart failure related pathways, such as mitochondria dysfunction, oxidative phosphorylation dysfunction, and LXR/RXR activation (Table S5). This result demonstrated that the RV from ARVC explanted hearts presented with the same distinct heart failure phenotype as the LV from DCM patients. We also scanned the electronic transport chain (ETC) proteins and observed a significant dysfunction in the ARVC RV and DCM LV, rather than in the ARVC LV and DCM RV (Figure S2). This confirmed that heart failure from classical ARVC mainly involved the right ventricle, while DCM presented left ventricle dominant heart failure and milder changes in the right ventricle.


14

Page 15 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


We performed upstream regulator prediction analysis using IPA. As previously mentioned, the inflammatory regulator (IL6), the pro-fibrotic cytokine (TGFβ1), and histone demethylase [lysine-specific demethylase 5A (KDM5A)] were predicted to be significantly activated in the ARVC RV and DCM LV groups, both of which presented severe dysfunction in the clinical phenotype (Figure 3A and Table S4). It has been reported that IL6 exacerbates adverse LV remodeling through the gp-130 mediated STAT3 pathway.28 We did not observe a significantly elevated expression of IL6 in the ARVC group from both the RNA sequencing, as well as the qPCR results (Figure 3B-C). The TGFB family includes well-known cytokines involved in heart failure.29 The gene expression of TGFB1 was validated as being increased in the ARVC and DCM groups through qPCR (ARVC vs. Normal fold change = 2.31, p < 0.001; DCM vs. Normal fold change = 1.79, p = 0.078) (Figure 3B-C). We also observed the activation of KDM5A in the RV of the ARVC, and LV of the DCM groups. KDM5A is reported to repress mitochondrial function in cancer, but has not been mentioned in heart failure before.30 Transcriptome profiling and qPCR also validated a significant upregulation in ARVC (Figure 3B-C) (ARVC vs Normal fold change = 2.44, p= 0.002, DCM vs Normal fold change = 2.91, p < 0.001). Thus, we inferred that epigenetic modification through KDM5A activation might account for the progression of heart failure. In addition, RICTOR, TSC2, ANGPT2, MAP4K4, MYC, and PPARGC1A, etc. were also enriched and predicted to be altered according to their down-stream targets’ alteration as previously reported (Figure 3A),

29, 31, 32

although some of them were not detected directly

through LC-MS/MS because of their low abundance. Consequently, we screened their gene expression through the FPKM derived from the RNA sequencing, the results of which mostly coincided with the proteomics predicted results (Figure 3B). Inflammation is a Hallmark of ARVC beyond Heart Failure


15


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 16 of 49

We compared the collective results of inflammation-related proteins among the four groups. The results indicated that acute phase response signaling (Figure 4A and Figure S3) and the complement system (Figure 4B and Figure S4) were significantly enriched in both ARVC groups. Notably, a similar but milder alteration of inflammatory factors was also found in the DCM LV, but not the RV (Figure S4). Both ARVC ventricles and the DCM LV presented with the up-regulation of the JNK, p38 MAPK, and NF-kB pathways proteins, while the downstream proteins in the JAK/STAT3 and ERK pathways were specifically upregulated in the ARVC, but not in DCM (Figure 4A and Figure S4). These results suggested that the JAK/STAT3 and ERK pathways might be activated specifically in the end-stage cardiac tissue of ARVC patients, although the phosphorylation of these signaling pathways proteins was not detected in our proteomics profiling. Moreover, most of the complement system proteins, such as C3, C4, C6, C7, C8 and C9, were significantly upregulated in both ARVC ventricles. In contrast, only C4 and C5 were increased in the DCM RV, and none were elevated in the DCM LV (Figure 4B and Figure S5). These results demonstrated that complement system activation is less of a common heart failure pathway, and more of a characteristic molecular phenotype of ARVC. The FPKM results from the RNA sequencing also validated the significantly increased inflammatory response in ARVC, which showed that a majority of the genes involved with acute response reactions, complement systems, and cytokines as well as inflammatory related mediators and receptors, were overexpressed (Figure 4C). These results also confirmed the pathological examination, in which patchy inflammation infiltration was detected in the ARVC ventricles (Figure 1). Lipid Metabolism Was Reprogrammed Specifically in ARVC


16

Page 17 of 49

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60


Figures S1, B and C show that, among the top ten enriched biological functions using IPA, lipid metabolism was especially activated in ARVC, but not in DCM. For further analysis, we conducted a comprehensive protein-protein interaction (PPI) network of ARVC RV (Figure 5A, E), and then we filtered out the commonly regulated proteins in the RV (Figure 5B, F) and LV (Figure 5C, G) of the DCM hearts. Through the PPI network filtered by DCM (Figure 5C), we observed that the lipid metabolism, limited mitochondrial functions and complement systems were enriched. To consider a similar pathophysiological process in the ARVC LV, we also identified the commonly regulated proteins in the RV and LV of ARVC hearts. As a result, the proteins participating in the lipid metabolism and complement systems (four proteins) were specifically remained and highlighted (Figure 5D, H). Thus, we were able to perform an in-depth analysis on the lipid metabolism and it showed that the major proteins participating in fatty acid metabolism (ADIPOQ, FABP4, and ACADVL), transportation of steroids (AEBP1 and SERPINA6), transportation of lipids (APOH and APOM), and transportation of phospholipids (APOA1, APOE, and CD14) were significantly upregulated in the RV of the ARVC groups when compared to the normal standard, while little corresponding evidence was observed in the DCM groups (Table S6). The results indicated that lipid metabolism was the most unique and predominant process in ARVC myocardial tissue, which is consistent with the fibro-fatty replacement in the pathological phenotype (Figure 1). The lipid metabolism networks reconstructed through DEPs in the ARVC RV are shown in Figure 6A. The ARVC LV presented with a similar but milder alteration in the lipid metabolism pattern (Figure S5, Table S6), and the lipid metabolism pathways presented with no significant activation in the DCM groups (Table S6). In a manner similar to the proteomics analysis results, the expression levels of the lipid metabolism-related genes in the


17


1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 51 52 53 54 55 56 57 58 59 60

Page 18 of 49

RNA sequencing results were positively correlated to their protein components (r = 0.36, p

Comprehensive Myocardial Proteogenomics ... - ACS Publications

Recommend Documents