The human cardiac and skeletal muscle proteomes defined by transcriptomics and antibody-based profiling (2024)

The transcriptomic analysis

A transcriptomic analysis was performed using deep sequencing (RNA-seq) on five skeletal muscle samples. These were compared with our previously reported transcriptomic analysis on 27 fresh frozen tissue types including four heart muscle samples [4]. This resulted in altogether 100 samples corresponding to 28 tissue types. The normalized mRNA abundance was calculated as FPKM-values with a cutoff of FPKM = 1, roughly corresponding to one mRNA molecule per average cell in the sample [5]. The most abundant genes in muscle tissues were mitochondrial genes expressed in all tissue types, with the highest mRNA abundance (FPKM = 20,737) observed by the mitochondrial gene MT-CO1 in cardiac muscle. An analysis of the global expression profiles using hierarchical clustering and a correlation heat map from all the 100 tissue samples are displayed in Fig.1. The heat map shows pairwise correlations between the 28 tissue types based on mRNA expression levels of all 20,314 genes. Testis and brain were identified as outliers, while a clear connectivity and similarities in global expression was observed between samples from the GI-tract, the hematopoietic tissues, as well as the two striated muscle tissues, respectively. Cardiac and skeletal muscle showed highest correlations with one another as compared to other tissue types.

The variance between muscle samples from different individuals was analyzed by pairwise correlations of the expression level of all protein coding genes (Fig.2a–d). A Spearman correlation of 0.97 was observed between cardiac muscle samples from two different individuals (Fig.2a), while corresponding correlation for the skeletal muscle samples was also 0.97 (Fig.2b). The correlation between average values from the two striated muscles was somewhat lower (Fig.2c), but still at a high value (0.89). Low inter-individual variations in the genome-wide expression patterns across muscle samples indicate both high technical reproducibility and low biological variance. Pairwise comparisons between muscle tissues and other tissue types showed, as expected, higher variance. The lowest correlation was found between skeletal muscle and testis (0.69, data not shown), while a somewhat surprisingly high correlation (0.90) was observed between heart muscle and adipose tissue (Fig.2d), as well as skeletal muscle and adipose tissue (0.86, data not shown). The similarity of global expression between adipose tissue and the striated muscle is interesting and should be explored further.

The cardiac muscle transcriptome

Data from the transcriptomics analysis across 28 tissue types was used for categorization of all the protein-coding genes (n = 20,314) into various classes of tissue-specific expression. Figure2e shows the classification for heart muscle. Altogether 64% of the genes were expressed in the heart, with the largest class of genes (n = 8760) representing “housekeeping genes” expressed in all tissues, out of which 641 genes encoded mitochondrial proteins. Only 283 genes were defined as elevated in heart as compared to the other tissues (Additional file 1: Table S1), including genes enriched in a group of tissues. The 283 genes elevated in heart were classified into three subcategories as defined in Table1, with 32 tissue enriched genes, 143 group enriched genes and 108 tissue enhanced genes. A GO-analysis of these genes indicated an over-representation of genes related to contraction (37 genes), ion transport (23 genes), regulation of membrane potential (22 genes) and actomyosin structure organization (18 genes).

Of the 283 genes elevated in cardiac muscle, 47 genes were differentially expressed between the two auricular samples and the two ventricle samples (Additional file 2: Table S2), based on mean values and a threshold of at least a five-fold difference in FPKM values. Twenty-one genes showed a higher average expression in the ventricle samples as compared to the auricle samples, while 26 genes were higher expressed in the auricle samples as compared to the ventricle samples. A majority of the differentially expressed genes encode proteins associated with developmental processes or regulatory functions such as control of blood volume, blood pressure, heart rate and response to cardiac overload. The list also includes several genes with unknown or only partly characterized functions in cardiac muscle.

An analysis of the expression levels of each gene made it possible to calculate the relative mRNA pool in heart for each category. The analysis (Fig.2f) showed that 32% of the mRNA molecules present in heart correspond to mitochondrial proteins and only 14% of the mRNA pool represents genes elevated in heart. Thus, a large fraction of the transcriptional activity in the heart relates to generation of energy with almost one third of the transcripts encoding proteins implicated in energy metabolism.

The skeletal muscle transcriptome

Figure2g shows the classification for skeletal muscle. Altogether 55% of the genes were expressed in skeletal muscle, with 8760 “housekeeping genes”. Only 371 genes (Additional file 3: Table S3) were defined as elevated in skeletal muscle as compared to other tissues, including 110 tissue enriched genes, 149 group enriched genes and 112 tissue enhanced genes. A GO-analysis of these genes indicated an over-representation of genes involved in contraction (48 genes), actin filament-based processes (37 genes), muscle organ development (24 genes) and muscle filament sliding (23 genes).

Analysis of the relative mRNA pool in skeletal muscle (Fig.2h) showed that 28% correspond to mitochondrial proteins and 28% represent genes elevated in skeletal muscle. Thus, most of the transcriptional activity in the skeletal muscle also relates to energy generation, but with a higher fraction of mRNA molecules encoding genes classified as tissue elevated as compared to heart. Together the two striated muscles have by far the highest fraction of transcripts encoding mitochondrial proteins of all analyzed tissues (data not shown).

Genes with shared enriched expression in cardiac and skeletal muscle

A network plot of the tissue and group enriched genes in heart and skeletal muscle, respectively, was generated in order to illustrate genes commonly expressed between different tissue types (Fig.3). The two tissues had 81 simultaneously enriched genes, suggesting common functions. GO-analysis of these genes showed an overrepresentation of genes associated with contraction (18 genes), actin filament-based processes (15 genes), actomyosin structure organization (10 genes) and muscle filament sliding (8 genes). Several genes were also shared between striated muscles and cerebral cortex, and these genes are dominated by genes for which the corresponding proteins are involved in ion transport. Moreover, a number of genes were simultaneously enriched in striated muscles, adipose tissue and/or liver, mainly represented by genes implicated in metabolic processes and enzymatic activities.

Proteins with elevated expression in cardiac muscle

The proteins elevated in heart or skeletal muscle identified by the transcriptomics analysis were further studied by antibody-based profiling using the publicly available resource of TMA data within the Human Protein Atlas (www.proteinatlas.org) portal cross-referenced with the Cardiac Organellar Protein Atlas Knowledge base (www.heartproteome.org) [6] in order to determine heart and skeletal muscle specificity and spatial and cellular distribution of the expression. In total, 239 of the 283 genes elevated in cardiac muscle, and 313 of the 371 genes elevated in skeletal muscle had available antibodies as part of the Human Protein Atlas project and were studied with antibody-based profiling using immunohistochemistry. Immunohistochemistry was performed on TMA cores from 44 different normal tissues, including heart and skeletal muscle samples from three different individuals, respectively. The analyzed proteins were divided into categories based on specificity of the expression, such as proteins selectively expressed in either heart or skeletal muscle, or proteins found to be expressed in both heart and skeletal muscle. Examples of proteins with various functions and expression patterns in striated muscles are described below.

An interesting analysis is to investigate proteins elevated in heart, particularly in relation to skeletal muscle. Examples of eight proteins related to contraction and with highest expression in heart muscle are displayed in Fig.4a. The primary structural proteins in the cardiomyocytes are myosin and actin filaments, related to contraction and forming a striated pattern observed in electron microscopy. Examples of members of the myosin and actin families mainly expressed in heart muscle include MYH6, MYL4, MYL7 and ACTC1. Another protein family related to muscular contraction is the troponin family, regulating the binding of myosin to actin via conformation differences dependent on calcium ion concentration in the cells [7]. Two troponin family members solely expressed in heart are TNNI3 and TNNT2. Other examples include the myosin binding protein MYBPC3, influencing contraction by cross-bridging in the sarcomere, and the calcium channel protein CACNA1C, regulating contraction via mediation of calcium ion entry into excitable cells.

A large number of proteins identified as highly expressed in heart were implicated in homeostasis and regeneration/repair, and some of these are displayed in Fig.4b. In order to retain balanced levels of various substances in the body, heart plays an important role in homeostasis. One such example is the atrial natriuretic peptide NPPA, controlling extracellular fluid volume and electrolyte homeostasis. Mutations in NPPA are thought to be responsible for the development of atrial fibrillation (arrhythmia) [8]. Another example is AKAP6, involved in controlling energy homeostasis and glycolysis. Heart regeneration and repair have been extensively studied as cardiomyocyte deficiency is the underlying cause of most heart failures, and it is still debated as to what extent cardiomyocytes are replaced in adult hearts [9]. Examples of proteins expressed in heart suggested to play a role in regeneration and repair include the stem cell protein NES and the cell adhesion molecule BVES. NES is involved in brain and eye development and has been suggested to be involved in reparative vascularization and differentiation of cardiac stem cells into striated cells following myocardial infarction [10, 11]. NES displayed distinct positivity in endothelial cells of both heart and other tissue types. BVES is involved in formation and regulation of tight junction permeability and limited literature suggests involvement in muscle regeneration [12].

Intercalated discs are defined as the connections between two adjacent cardiomyocytes, aiding in contraction of multiple cardiomyocytes simultaneously as a unit, which is necessary for proper heart function. Examples of eight proteins distinctly expressed in intercalated discs are displayed in Fig.4c. Four of these (ATP1A3, CDH2, MYZAP and PKP2) are well characterized in literature and involved in functions related to plasma membrane ion exchange, cell adhesion, cell junctions and desmosomes, respectively. For the remaining four proteins (ART3, FILIP1, POPDC2 and RAB9B) no or only limited literature exists on function in heart. ART3 has been suggested as an arginine-specific protein with function in testis [13], distinctly stained also in the present investigation, however, the additional high expression in intercalated discs has previously not been described. FILIP1, with evidence only at transcript level, is suggested to play a role in neuronal development and showed highest expression in heart and brain, but the exact function is unknown [14]. The transmembrane protein POPDC2 has in few studies on mice and zebrafish been suggested to play a role in proper sinus node function [15], but no previous studies suggest expression in intercalated discs. RAB9B has limited literature on membrane trafficking, but no data on expression in heart.

Interestingly, two proteins previously not characterized in heart (SBK2 and SHD) were at the mRNA level selectively expressed in both samples from the atrium and not in the ventricular wall or in any of the other analyzed tissues. Immunohistochemically stained images of SBK2 and SHD are displayed in Fig.4d. SBK2 has evidence of existence only inferred from hom*ology, while SHD is suggested to be an adapter protein with previous evidence only at transcript level.

Proteins with elevated expression in skeletal muscle

Similar to heart muscle, the main function of skeletal muscle is contraction. Four proteins related to contraction with exclusive expression in skeletal muscle are displayed in Fig.5a. Two of those (MYH2 and TNNT1) are members of the myosin and troponin families. Other examples include the myosin binding protein MYBPC1 and the actin cross-linking protein MYOT, involved in myofibril stability and assembly. MYH2 was expressed in fast (type II) fibers [16] and TNNT1 in slow (type I) fibers [17], while expression of MYOT was concentrated to Z-lines of the muscle fibers.

In both heart and skeletal muscle, contraction is dependent on the level of intracellular calcium. However, in contrast to cardiomyocytes, where calcium release is regulated via binding of external calcium ions to voltage gated calcium channels, skeletal myocytes store calcium in the sarcoplasmic reticulum until a neuronal impulse triggers calcium influx along the myofilaments. In Fig.5b, four proteins related to calcium function with selective expression in skeletal muscle are displayed. Two of the main proteins responsible for this function are RYR1, the ryanodine receptor acting as the calcium release channel, and CASQ1, essential for calcium storage in the sarcoplasmic reticulum. CAPN3 is an intracellular protease regulated by changes in calcium levels and plays a role in skeletal muscle degradation and sarcomeric remodeling, with defects resulting in muscular dystrophy [18]. Another protein involved in calcium function is JPH1, aiding in the functional cross-talk between cell surface and intracellular calcium release channels.

Contraction is an energy demanding process and it is acknowledged that skeletal muscle is a tissue with high metabolic activity. Skeletal muscle is one of the body’s major sites for glucose disposal and is implicated in metabolic diseases such as type II diabetes. Thus, enzymatic activity is an important function in skeletal muscle, relating to various processes such as glucose metabolism, glycogen storage and regeneration. Indeed, a large number (72 out of 371) of the elevated proteins in skeletal muscle are related to different metabolic processes, i.e. present in the HMR2.0 metabolic reaction library [19, 20]. Genes from several enzymatic pathways were elevated in skeletal muscle, including membrane transport, protein modification and purine metabolism. The skeletal muscle specific transport reactions involved calcium transport (ATP2A1, ATP2B2, nucleoside uptake (SLC29A2) and glucose transport (SLCA4). Additional examples of proteins implicated in enzymatic activities with selective expression in skeletal muscle are displayed in Fig.5c. These examples include (i) AMPD1, an enzyme involved in the purine nucleotide cycle, (ii) PYGM, an enzyme essential for carbohydrate metabolism and glycogenolysis, (iii) CA3, a metalloenzyme catalyzing hydration of carbon dioxide, and (iv) ENO3, an isoenzyme suggested to play a role in muscle development and regeneration, with mutations associated with glycogen storage disease [21]. At the protein level, ENO3 was highly expressed in type II fibers [22].

Four examples of proteins selectively expressed, but with unknown function in skeletal muscle are displayed in Fig.5d. The putative protein FAM166B showed additional positivity in adrenal gland and ciliated cells in the present investigation. KLHL40 is suggested to be required for skeletal muscle development [23]; however, the exact function is unknown. Another example is OR10AB1P, with evidence of existence only at the transcript level according to Uniprot. Despite low FPKM values, OR10AB1P was selectively expressed in striated muscles as compared to other tissues, with strongest immunoreactivity observed in skeletal muscle. PHKG1 is the catalytic subunit of the phosphorylase b kinase (PHK), which is responsible for hormonal regulation of glycogen breakdown [24]; however, PHKG1 only has evidence of existence at the transcript level according to Uniprot and the exact function is unknown. FAM166B, OR10AB1P and PHKG1 displayed a heterogeneous expression pattern possibly representing type I or type II muscle fibers.

Proteins with elevated expression in both cardiac and skeletal muscle

Although heart and skeletal muscles differ in arrangement of cells and response to signals, the two muscle types have many similarities. Several proteins related to contraction are shared between the two organs, as displayed in Fig.6a. Some of these are members of the myosin and troponin families, such as MYH7, MYL3 and TNNC1. Another example is TPM3, a member of the tropomyosin family essential for contraction. TPM3 binds to actin and is associated with the troponin complex. In skeletal muscle, all four proteins were higher expressed in type I muscle fibers.

In Fig.6b, four proteins involved in calcium transport and storage expressed in both heart and skeletal muscle are shown. CAMK2B is involved in synapse formation and regulation of sarcoplasmic reticulum calcium transport [25]. Additional strong staining of CAMK2B was observed in the CNS. Other examples include CASQ2, responsible for the internal calcium store in muscle [26], HRC, implicated in regulation of calcium sequestration and release [27], and the calcium transport regulating protein SRL, with evidence of existence only at the transcript level according to Uniprot.

Examples of proteins involved in metabolism and myofibrillar organization with expression shared between heart and skeletal muscle are displayed in Fig.6c. CKM is a creatine kinase involved in energy transduction, and is also an important serum marker for myocardial infarction [28]. Another enzyme is PGAM2, related to glycolytic pathways, with mutations causing glycogen storage disease [29]. LDB3 is involved in sarcomere organization, with mutations associated with myopathy [30]. MURC aids in myofibrillar organization and has also been shown to induce NPPA transcription [28, 31].

Four proteins with selective expression but unknown function in heart and skeletal muscle are displayed in Fig.6d. The nuclear protein ADSSL1 has limited data suggesting contribution to glycolysis, while vague literature on HHATL mentions heart specificity [32], however, the exact function is unknown and neither ADSSL1 nor HHATL has previously been characterized on the protein level. Another protein only characterized on the transcript level is C1ORF170, which regulates the expression of target genes associated with energy transfer and contractile function [33]. SPTB is known to be involved in cytoskeletal structure of erythrocyte plasma membrane [34], and erythrocytes were distinctly stained also in the present investigation, however, the clear membranous expression in striated muscle cells has previously not been described.

Tissue samples

Tissues used for RNA extraction was based on biopsies from four heart muscle samples and five skeletal muscle samples. The heart material was composed of two samples from the left ventricle (both adult females) and two samples from the left auricles (one female and one male, both adults). One left ventricle sample was obtained by surgery for asymmetrical hypertrophy, with the sample showing slight fibrosis, but not diagnostic for hypertrophic cardiomyopathy. The other left ventricle sample was obtained from an individual diagnosed with post partum cardiomyopathy, with the material taken in connection with temporary use of left ventricle device (heart mate). The sample had some dilation and slight endocardial increased thickness. The histology of the sample was mainly normal, but also showed areas with increased fibrotic tissue and compensatory hypertrophy, as well as a slight probably unspecific inflammation. Both auricle samples were obtained from individuals operated for atrial fibrillation, with normal histological picture without amyloid depositions. The five skeletal muscle samples (two males and three females) were derived from the vastus lateralis muscle in adult healthy volunteers [39]. In addition, samples from 26 other tissues and organs were analyzed, as previously described [4].

Transcript profiling (RNA-seq)

The use of human tissue samples was approved by the Uppsala Ethical Review Board (Reference #2011/473). Tissues samples were embedded in Optimal Cutting Temperature (O.C.T.) compound and stored at −80°C. A hematoxylin-eosin (HE) stained frozen section (4μm) was prepared from each sample using a cryostat and each slide was examined by a pathologist to ensure proper tissue morphology and sampling of representative normal tissue. Three sections (10μm) were cut from each frozen tissue block and collected into a tube for subsequent RNA extraction. The tissue was hom*ogenized mechanically using a 3mm steel grinding ball (VWR). Total RNA was extracted from cell lines and tissue samples using the RNeasy Mini Kit (Qiagen, Hilden, Germany) according to the manufacturer’s instructions. For the skeletal muscle samples, RNA was extracted using the Trizol® method (Invitrogen, Carlsbad, CA, USA). The extracted RNA samples were analyzed using either an Experion automated electrophoresis system (Bio-Rad Laboratories, Hercules, CA, USA) with the standard-sensitivity RNA chip or an Agilent 2100 Bioanalyzer system (Agilent Biotechnologies, Palo Alto, USA) with the RNA 6000 Nano Labchip Kit. Only samples of high-quality RNA (RNA Integrity Number ≥7.5) were used in the following mRNA sample preparation for sequencing. mRNA sequencing was performed on Illumina HiSeq2000 and 2500 machines (Illumina, San Diego, CA, USA) using the standard Illumina RNA-seq protocol with a read length of 2 × 100 bases. The samples were sequenced multiplexed 15 per lane, producing an average of 18 million mappable read pairs per sample.

Analysis of data

The raw reads obtained from the sequencing system were trimmed for low quality ends with the software Sickle (3), using a phred quality threshold of 20. All reads shorter than 54bp after the trimming were discarded. The processed reads were mapped to the GRCh37 version of the human genome with Tophat v2.0.3 (4). Potential PCR duplicates were eliminated using the MarkDuplicates module of Picard 1.77 (5). To obtain quantification scores for all human genes and transcripts, FPKM (fragments per kilobase of exon model per million mapped reads) values were calculated using Cufflinks v2.0.2 (6), which corrects for transcript length and the total number of mapped reads from the library to compensate for different read depths for different samples. The gene models from Ensembl build 69 (7) were used in Cufflinks. Data was analyzed using R Statistical Environment (8) with the addition of package ‘gplots’ (9). A network analysis was performed using Cytoscape 3.0 (10). For analyses where a log2-scale of the data was used, pseudo-counts of +1 were added to the values.

Specificity classification

The average FPKM value of all individual samples for each tissue was used to estimate the gene expression level and classify the 20,314 genes into one of six categories: (1) “Not detected”, <1 FPKM in tissue; (2) “Tissue enriched”, at least five-fold higher FPKM level in tissue compared to all other 27 tissues; (3) “Group enriched”, at least five-fold higher FPKM level in a group of 2–7 tissues compared to all other tissues; (4) “Expressed in all”, detected in all 28 tissues; (5) “Tissue enhanced”, at least five-fold higher FPKM level in tissue compared to the average FPKM value of all 28 tissues and (6) “Mixed”, genes expressed in 1–27 tissues and in none of the above categories. A tissue-specific score [4] was defined as the heart or skeletal muscle FPKM respectively, divided by the maximum FPKM value in any of the other 27 tissues. Genes categorized as tissue enriched, group enriched or tissue enhanced were together defined as tissue elevated, while genes in the “expressed in all” category were further classified into a “mitochondrial” category based on a list of mitochondrial proteins obtained from the MitoProteome database [40].

Gene ontology analysis

A gene ontology (GO) [41] analysis was performed using the GOrilla tool [42] in order to determine overrepresented GO categories in the gene set of tissue elevated genes. The number of genes for each term was counted, allowing a gene to be associated with more than one term. A list of all genes analyzed in this study was used as the background list in GOrilla.

Tissue profiling

Tissue microarrays (TMAs) were generated as previously described [43], containing 1mm cores of 44 different normal tissues in triplicates. Formalin-fixed and paraffin-embedded samples were received from the Department of Pathology, Uppsala University Hospital, Sweden, approved by the local Research Ethics Committee (Uppsala, Sweden, Ups 02–577). One section (4μm) was cut from each donor block, stained with HE and reviewed by a pathologist in order to select the appropriate area for sampling. Cylindrical cores with a diameter of 1mm cylindrical tissue core were removed from the donor blocks and placed into recipient paraffin blocks, using an TMA Grand Master (3DHistech, Budapest, Hungary).

Four-micrometer sections were cut from the TMA blocks, mounted on adhesive slides and baked at 60°C for 45min. TMA slides were then deparaffinised in Neo-Clear® (Merck Millipore, Darmstadt, Germany), followed by hydration in graded alcohols and blocking for endogenous peroxidase in 0.3% hydrogen peroxide. For antigen retrieval, slides were immersed and boiled in Citrate buffer®, pH6 (Lab Vision, Freemont, CA) for 4min at 125°C and then allowed to cool to 90°C. Automated IHC was performed essentially as previously described [43], using an Autostainer 480 instrument® (Lab Vision). Primary antibodies and a dextran polymer visualization system (UltraVision LP HRP polymer®, Lab Vision) were incubated for 30min each at room temperature and slides were developed for 10min using Diaminobenzidine (Lab Vision) as chromogen. All incubations were followed by rinse in Wash buffer® (Lab Vision) for 5min. Slides were counterstained in Mayers hematoxylin (Histolab) and cover slipped using Pertex® (Histolab) as mounting medium. Incubation with PBS instead of primary antibody served as negative control. The dilution of the primary antibody was optimized by testing the antibody on a selection of normal tissues and comparison with previously published literature and gene characterization data. The AperioScanScope XT Slide Scanner (Aperio Technologies, Vista, CA) system was used to capture digital whole slide high-resolution images with a 20× objective.

Antibodies

The IDs of all antibodies used in the immunohistochemically stained images are summarized here. All antibodies are from the Human Protein Atlas project, unless otherwise stated. Proteins elevated in cardiac muscle: MYH6 (HPA001239), MYL4 (HPA051884), MYL7 (HPA013331), ACTC1 (61075, PROGEN Biotechnik), TNNI3 (HPA063258), TNNT2 (HPA015774), MYBPC3 (HPA040147), CACNA1C (HPA039796), NPPA (HPA058269), AKAP6 (07–087, Upstate), NES (HPA026111), BVES (HPA014788), ATP1A3 (HPA045367), CDH2 (sc-8424, Santa Cruz Biotechnology), MYZAP (HPA039827), PKP2 (HPA014314), ART3 (HPA011268), FILIP1 (HPA055754), POPDC2 (HPA024255), RAB9B (HPA061932), SBK2 (HPA030631), SHD (HPA017955). Proteins elevated in skeletal muscle: MYH2 (M4276, SIGMA), TNNT1 (HPA058448), MYBPC1 (HPA021004), MYOT (HPA037733), RYR1 (HPA056416), CASQ1 (HPA026823), CAPN3 (HPA040052), JPH1 (HPA009413), AMPD1 (HPA028080), PYGM (HPA056003), CA3 (H00000761-M02, Abnova), ENO3 (HPA000793), FAM166B (HPA045540), OR10AB1P (HPA044992), PHKG1 (HPA012057). Proteins elevated in both cardiac and skeletal muscle: MYH7 (sc-53090, Santa Cruz Biotechnology), MYL3 (sc-47719, Santa Cruz Biotechnology), TNNC1 (NCL-TROPC, Novocastra), TPM3 (HPA047089), CAMK2B (HPA026307), CASQ2 (HPA027285), HRC (HPA004833), SRL (HPA041535), CKM (HPA047859), PGAM2 (HPA060173), LDB3 (HPA048955), MURC (HPA020973), ADSSL1 (HPA052621), HHATL (HPA018174), C1ORF170 (HPA031712), SPTB (HPA003398).

Data availability

FPKM values for samples from all analyzed tissues except skeletal muscle are available for download without any restrictions (www.proteinatlas.org/about/download). The skeletal muscle data will be added to the next release of the Human Protein Atlas. The primary data (reads) are available through the Array Express Archive (www.ebi.ac.uk/arrayexpress/) under accession number E-MTAB-1733, as well as at Gene Expression Omnibus under the accession numbers GSE58387 and GSE58608. Transcript profiling data for each gene in each tissue type, as well as the original immunohistochemically stained images are also available on the Human Protein Atlas (www.proteinatlas.org).

Additional files

Additional file 1: Table S1.^{(53K, docx)}

List of the 283 genes elevated in cardiac muscle. The list is sorted after tissue-specific score which is the FPKM value in cardiac muscle divided by the maximum FPKM in all other tissues.

Competing interests

The authors declare that they have no competing interests.

Authors’ contributions

MU and FP conceived and designed the study. MU and CL wrote the manuscript. CL and JL performed the protein expression analysis and comparison. LF and BMH analyzed the RNA-Seq data and performed statistical analyses. CK supervised the staining and annotation of samples. CJS, ML, MH, HC, DL, PP, LV, AM, JN and EL contributed analysis, text and comments to the paper. All authors read and approved the final manuscript.

1. Ounzain S, Micheletti R, Beckmann T, Schroen B, Alexanian M, Pezzuto I, Crippa S, Nemir M, Sarre A, Johnson R, et al.. Genome-wide profiling of the cardiac transcriptome after myocardial infarction identifies novel heart-specific long non-coding RNAs. Eur Heart J. 2015;36:353-68a. [PMC free article] [PubMed]

2. Barth AS, Chakir K, Kass DA, Tomaselli GF. Transcriptome, proteome, and metabolome in dyssynchronous heart failure and CRT. J Cardiovasc Transl Res. 2012;5:180–7. doi:10.1007/s12265-011-9339-2. [PubMed] [CrossRef] [Google Scholar]

3. Thygesen K, Alpert JS, Jaffe AS, Simoons ML, Chaitman BR, White HD, Task Force for the Universal Definition of Myocardial I Third universal definition of myocardial infarction. Nat Rev Cardiol. 2012;9:620–33. doi:10.1038/nrcardio.2012.122. [PubMed] [CrossRef] [Google Scholar]

4. fa*gerberg L, Hallstrom BM, Oksvold P, Kampf C, Djureinovic D, Odeberg J, Habuka M, Tahmasebpoor S, Danielsson A, Edlund K, et al. Analysis of the human tissue-specific expression by genome-wide integration of transcriptomics and antibody-based proteomics. Mol Cell Proteomics. 2014;13:397–406. doi:10.1074/mcp.M113.035600. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

5. Hebenstreit D, Fang M, Gu M, Charoensawan V, van Oudenaarden A, Teichmann SA. RNA sequencing reveals two major classes of gene expression levels in metazoan cells. Mol Syst Biol. 2011;7:497. doi:10.1038/msb.2011.28. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

6. Zong NC, Li H, Li H, Lam MP, Jimenez RC, Kim CS, Deng N, Kim AK, Choi JH, Zelaya I, et al. Integration of cardiac proteome biology and medicine by a specialized knowledgebase. Circ Res. 2013;113:1043–53. doi:10.1161/CIRCRESAHA.113.301151. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

7. Gunning P, O’Neill G, Hardeman E. Tropomyosin-based regulation of the actin cytoskeleton in time and space. Physiol Rev. 2008;88:1–35. doi:10.1152/physrev.00001.2007. [PubMed] [CrossRef] [Google Scholar]

8. Roberts JD, Davies RW, Lubitz SA, Thibodeau IL, Nery PB, Birnie DH, Benjamin EJ, Lemery R, Ellinor PT, Gollob MH. Evaluation of non-synonymous NPPA single nucleotide polymorphisms in atrial fibrillation. Europace. 2010;12:1078–83. doi:10.1093/europace/euq161. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

10. Berry SE, Andruszkiewicz P, Chun JL, Hong J. Nestin expression in end-stage disease in dystrophin-deficient heart: implications for regeneration from endogenous cardiac stem cells. Stem Cells Transl Med. 2013;2:848–61. doi:10.5966/sctm.2012-0174. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

11. El-Helou V, Chabot A, Gosselin H, Villeneuve L, Clavet-Lanthier ME, Tanguay JF, Enikolopov G, Fernandes KJ, Jasmin JF, Calderone A. Cardiac resident nestin(+) cells participate in reparative vascularisation. J Cell Physiol. 2013;228:1844–53. doi:10.1002/jcp.24345. [PubMed] [CrossRef] [Google Scholar]

12. Vasavada TK, DiAngelo JR, Duncan MK. Developmental expression of Pop1/Bves. J Histochem Cytochem. 2004;52:371–7. doi:10.1177/002215540405200308. [PubMed] [CrossRef] [Google Scholar]

13. Friedrich M, Grahnert A, Paasch U, Tannapfel A, Koch-Nolte F, Hauschildt S. Expression of toxin-related human mono-ADP-ribosyltransferase 3 in human testes. Asian J Androl. 2006;8:281–7. doi:10.1111/j.1745-7262.2006.00125.x. [PubMed] [CrossRef] [Google Scholar]

14. Nagano T, Yoneda T, Hatanaka Y, Kubota C, Murakami F, Sato M. Filamin A-interacting protein (FILIP) regulates cortical cell migration out of the ventricular zone. Nat Cell Biol. 2002;4:495–501. [PubMed] [Google Scholar]

15. Froese A, Breher SS, Waldeyer C, Schindler RF, Nikolaev VO, Rinne S, Wischmeyer E, Schlueter J, Becher J, Simrick S, et al. Popeye domain containing proteins are essential for stress-mediated modulation of cardiac pacemaking in mice. J Clin Invest. 2012;122:1119–30. doi:10.1172/JCI59410. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

16. Larsson L, Moss RL. Maximum velocity of shortening in relation to myosin isoform composition in single fibres from human skeletal muscles. J Physiol. 1993;472:595–614. doi:10.1113/jphysiol.1993.sp019964. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

17. Barton PJ, Cullen ME, Townsend PJ, Brand NJ, Mullen AJ, Norman DA, Bhavsar PK, Yacoub MH. Close physical linkage of human troponin genes: organization, sequence, and expression of the locus encoding cardiac troponin I and slow skeletal troponin T. Genomics. 1999;57:102–9. doi:10.1006/geno.1998.5702. [PubMed] [CrossRef] [Google Scholar]

18. Richard I, Broux O, Allamand V, Fougerousse F, Chiannilkulchai N, Bourg N, Brenguier L, Devaud C, Pasturaud P, Roudaut C, et al. Mutations in the proteolytic enzyme calpain 3 cause limb-girdle muscular dystrophy type 2A. Cell. 1995;81:27–40. doi:10.1016/0092-8674(95)90368-2. [PubMed] [CrossRef] [Google Scholar]

19. Mardinoglu A, Agren R, Kampf C, Asplund A, Nookaew I, Jacobson P, Walley AJ, Froguel P, Carlsson LM, Uhlen M, Nielsen J. Integration of clinical data with a genome-scale metabolic model of the human adipocyte. Mol Syst Biol. 2013;9:649. doi:10.1038/msb.2013.5. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

20. Mardinoglu A, Agren R, Kampf C, Asplund A, Uhlen M, Nielsen J. Genome-scale metabolic modelling of hepatocytes reveals serine deficiency in patients with non-alcoholic fatty liver disease. Nat Commun. 2014;5:3083. doi:10.1038/ncomms4083. [PubMed] [CrossRef] [Google Scholar]

21. Comi GP, Fortunato F, Lucchiari S, Bordoni A, Prelle A, Jann S, Keller A, Ciscato P, Galbiati S, Chiveri L, et al. Beta-enolase deficiency, a new metabolic myopathy of distal glycolysis. Ann Neurol. 2001;50:202–7. doi:10.1002/ana.1095. [PubMed] [CrossRef] [Google Scholar]

22. Keller A, Demeurie J, Merkulova T, Geraud G, Cywiner-Golenzer C, Lucas M, Chatelet FP. Fibre-type distribution and subcellular localisation of alpha and beta enolase in mouse striated muscle. Biol Cell. 2000;92:527–35. doi:10.1016/S0248-4900(00)01103-5. [PubMed] [CrossRef] [Google Scholar]

23. Ravenscroft G, Miyatake S, Lehtokari VL, Todd EJ, Vornanen P, Yau KS, Hayashi YK, Miyake N, Tsurusaki Y, Doi H, et al. Mutations in KLHL40 are a frequent cause of severe autosomal-recessive nemaline myopathy. Am J Hum Genet. 2013;93:6–18. doi:10.1016/j.ajhg.2013.05.004. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

24. Exton JH. Mechanisms of hormonal regulation of hepatic glucose metabolism. Diabetes Metab Rev. 1987;3:163–83. doi:10.1002/dmr.5610030108. [PubMed] [CrossRef] [Google Scholar]

25. Martinez-Pena y Valenzuela I, Mouslim C, Akaaboune M. Calcium/calmodulin kinase II-dependent acetylcholine receptor cycling at the mammalian neuromuscular junction in vivo. J Neurosci. 2010;30:12455–65. doi:10.1523/JNEUROSCI.3309-10.2010. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

26. Gyorke S, Gyorke I, Terentyev D, Viatchenko-Karpinski S, Williams SC. Modulation of sarcoplasmic reticulum calcium release by calsequestrin in cardiac myocytes. Biol Res. 2004;37:603–7. doi:10.4067/S0716-97602004000400014. [PubMed] [CrossRef] [Google Scholar]

27. Gregory KN, Ginsburg KS, Bodi I, Hahn H, Marreez YM, Song Q, Padmanabhan PA, Mitton BA, Waggoner JR, Del Monte F, et al. Histidine-rich Ca binding protein: a regulator of sarcoplasmic reticulum calcium sequestration and cardiac function. J Mol Cell Cardiol. 2006;40:653–65. doi:10.1016/j.yjmcc.2006.02.003. [PubMed] [CrossRef] [Google Scholar]

28. Morelli RL, Carlson CJ, Emilson B, Abendschein DR, Rapaport E. Serum creatine kinase MM isoenzyme sub-bands after acute myocardial infarction in man. Circulation. 1983;67:1283–9. doi:10.1161/01.CIR.67.6.1283. [PubMed] [CrossRef] [Google Scholar]

29. Burwinkel B, Bakker HD, Herschkovitz E, Moses SW, Shin YS, Kilimann MW. Mutations in the liver glycogen phosphorylase gene (PYGL) underlying glycogenosis type VI. Am J Hum Genet. 1998;62:785–91. doi:10.1086/301790. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

30. Selcen D, Engel AG. Mutations in ZASP define a novel form of muscular dystrophy in humans. Ann Neurol. 2005;57:269–76. doi:10.1002/ana.20376. [PubMed] [CrossRef] [Google Scholar]

31. Ogata T, Ueyama T, Isodono K, Tagawa M, Takehara N, Kawashima T, Harada K, Takahashi T, Shioi T, Matsubara H, Oh H. MURC, a muscle-restricted coiled-coil protein that modulates the Rho/ROCK pathway, induces cardiac dysfunction and conduction disturbance. Mol Cell Biol. 2008;28:3424–36. doi:10.1128/MCB.02186-07. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

32. Soejima H, Kawamoto S, Akai J, Miyoshi O, Arai Y, Morohka T, Matsuo S, Niikawa N, Kimura A, Okubo K, Mukai T. Isolation of novel heart-specific genes using the BodyMap database. Genomics. 2001;74:115–20. doi:10.1006/geno.2001.6527. [PubMed] [CrossRef] [Google Scholar]

33. Cho Y, Hazen BC, Russell AP, Kralli A. Peroxisome proliferator-activated receptor gamma coactivator 1 (PGC-1)- and estrogen-related receptor (ERR)-induced regulator in muscle 1 (Perm1) is a tissue-specific regulator of oxidative capacity in skeletal muscle cells. J Biol Chem. 2013;288:25207–18. doi:10.1074/jbc.M113.489674. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

34. Fowler VM. The human erythrocyte plasma membrane: a Rosetta Stone for decoding membrane-cytoskeleton structure. Curr Top Membr. 2013;72:39–88. doi:10.1016/B978-0-12-417027-8.00002-7. [PubMed] [CrossRef] [Google Scholar]

35. Uhlen M, Oksvold P, fa*gerberg L, Lundberg E, Jonasson K, Forsberg M, Zwahlen M, Kampf C, Wester K, Hober S, et al. Towards a knowledge-based human protein atlas. Nat Biotechnol. 2010;28:1248–50. doi:10.1038/nbt1210-1248. [PubMed] [CrossRef] [Google Scholar]

36. Uhlen M, Bjorling E, Agaton C, Szigyarto CA, Amini B, Andersen E, Andersson AC, Angelidou P, Asplund A, Asplund C, et al. A human protein atlas for normal and cancer tissues based on antibody proteomics. Mol Cell Proteomics. 2005;4:1920–32. doi:10.1074/mcp.M500279-MCP200. [PubMed] [CrossRef] [Google Scholar]

37. Savarese G, Trimarco B, Dellegrottaglie S, Prastaro M, Gambardella F, Rengo G, Leosco D, Perrone-Filardi P. Natriuretic peptide-guided therapy in chronic heart failure: a meta-analysis of 2,686 patients in 12 randomized trials. PLoS One. 2013;8 doi:10.1371/journal.pone.0058287. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

38. Ellis SG, Chew D, Chan A, Whitlow PL, Schneider JP, Topol EJ. Death following creatine kinase-MB elevation after coronary intervention: identification of an early risk period: importance of creatine kinase-MB level, completeness of revascularization, ventricular function, and probable benefit of statin therapy. Circulation. 2002;106:1205–10. doi:10.1161/01.CIR.0000028146.71416.2E. [PubMed] [CrossRef] [Google Scholar]

39. Lindholm ME, Huss M, Solnestam BW, Kjellqvist S, Lundeberg J, Sundberg CJ. The human skeletal muscle transcriptome: sex differences, alternative splicing, and tissue hom*ogeneity assessed with RNA sequencing. FASEB J. 2014;28:4571-81. [PubMed]

40. Guda P, Subramaniam S, Guda C. Mitoproteome: human heart mitochondrial protein sequence database. Methods Mol Biol. 2007;357:375–83. [PubMed] [Google Scholar]

41. Ashburner M, Ball CA, Blake JA, Botstein D, Butler H, Cherry JM, Davis AP, Dolinski K, Dwight SS, Eppig JT, et al. Gene ontology: tool for the unification of biology. The Gene Ontology Consortium. Nat Genet. 2000;25:25–9. doi:10.1038/75556. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

42. Eden E, Navon R, Steinfeld I, Lipson D, Yakhini Z. GOrilla: a tool for discovery and visualization of enriched GO terms in ranked gene lists. BMC Bioinformatics. 2009;10:48. doi:10.1186/1471-2105-10-48. [PMC free article] [PubMed] [CrossRef] [Google Scholar]

43. Kampf C, Olsson I, Ryberg U, Sjostedt E, Ponten F. Production of tissue microarrays, immunohistochemistry staining and digitalization within the human protein atlas. J Vis Exp. 2012;63. [http://www.ncbi.nlm.nih.gov/pubmed/22688270]. [PMC free article] [PubMed]