Cellular and genetic approaches to myocardial regeneration Tuyn, J. van

Cellular and genetic approaches to myocardial regeneration

Tuyn, J. van

Citation

Tuyn, J. van. (2008, January 9). Cellular and genetic approaches to myocardial

regeneration. Department of Cardiology and Department of Molecular Cell Biology (MCB), Faculty of Medicine, Leiden University Medical Center (LUMC), Leiden University.

Retrieved from https://hdl.handle.net/1887/12548

Version: Corrected Publisher’s Version

License: Licence agreement concerning inclusion of doctoral thesis in the Institutional Repository of the University of Leiden

Downloaded from: https://hdl.handle.net/1887/12548

Note: To cite this publication please use the final published version (if applicable).

Chapter

7

John van Tuyn Joost P.G. Sluijter Jim Swildens Marie-José Goumans Pieter A. Doevendans Arnoud van der Laarse Martin J. Schalij Douwe E. Atsma Antoine A.F. de Vries

Under preparation

Gene expression profiling of myocardin- induced cardiomyogenic differentiation in

human cardiac progenitor cells

150

Abstract

Cardiac diseases are the leading cause of death in the Western world. Healing of ischemically or otherwise damaged myocardium is problematic due to the limited regenerative capacity of the heart. The use of cardiomyogenic transcription factors (cTFs) to induce cardiomyocyte (CM) differentiation of non-muscle cells in the heart may be a novel approach to accomplish myocardial regeneration. In addition, stimulation of (stem) cells with cTFs prior to their transplantation may increase the effectiveness of cell therapy to treat the heart. In this study, we used DNA microarray-based gene expression profiling of human cardiac progenitor cells (CMPCs) transduced with the cTF myocardin to investigate the cardiomyogenic potential of myocardin. Upon culture in 5’-azacytidine-containing medium, isolated CMPCs quantitatively differentiate into beating CMs providing an ideal reference sample for the transcriptome analysis. We also compared the transcriptional programs induced by the heart muscle enriched (CM-MYOCD) and by the smooth muscle-enriched isoform of myocardin (SM-MYOCD). For selected genes, the DNA microarray data were validated by quantitative reverse transcription- polymerase chain reaction analysis and immunofluorescence microscopy.

Forced CM-MYOCD gene expression stimulated transcription of a large repertoire of heart muscle-specific genes in CMPCs but failed to induce their complete differentiation into CMs. LV.CM-MYOCD infection of CMPCs did not induce expression of sarcomeric (regulatory) proteins such as titin cap and cardiac myosin-binding protein C and certain genes encoding ion channel components, like ryanodine receptor 2 which are upregulated in cultures of CMPC-derived CMs. Both transduction with CM-MYOCD and differentiation with 5’-azacytidine induced expression of the MYH11 gene in CMPCs at both the RNA and protein level extending previous findings that (primitive) CMs contain smooth muscle proteins. Comparison of transcriptional programs induced by CM-MYOCD and SM-MYOCD revealed that both isoforms of myocardin potently and to similar extent stimulated the expression of a large set of smooth muscle genes and a number of cardiac muscle genes.

However, CM-MYOCD could activate an additional set of cardiac muscle genes and also activated a subset of CM-related genes encoding mainly sarcomeric components to a much greater extent. This provides an explanation for the underestimation of myocardin’s cardiomyogenic potential reported in previous studies that used N-terminally truncated versions of this cTF.

151

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

Introduction

Ischemic heart muscle injury result in permanent cardiomyocyte (CM) loss and myocardial scar formation, which both contribute to a decline in cardiac function¹. Currently, new therapeutic strategies aiming to repair damaged myocardium by transplantation of autologous (stem) cells with myogenic differentiation capacity are being explored². Generally experimental treatment modalities focus on inhibition of cardiac fibrosis, which is a major factor in the long-term deleterious remodeling process that often takes place in damaged hearts³. A novel concept that combines elements of both therapeutic approaches is the in situ conversion of myocardial scar fibroblasts (MSFs) towards striated muscle cells. This approach was first tested by Murry and colleagues using the myogenic transcription factor (TF) MyoD. After injection of a first-generation adenovirus serotype 5 (Ad5) vector encoding murine MyoD in the healing wounds of cryogenically damaged rat hearts, these authors found a fraction of resident MSFs to obtain characteristics of embryonic skeletal muscle cells⁴. However, as skeletal muscle cells differ from CMs in electrical, mechanical and physiological properties, they may display serious shortcomings as therapeutic agents for myocardial regeneration⁵. In an attempt to improve on the original concept, we directed our attention to the recently discovered cardiac and smooth muscle-specific TF myocardin⁶ to induce cardiomyogenic differentiation.

Myocardin gene transcripts appear during the earliest stages of cardiac and smooth muscle development⁶ and are required for normal myogenesis^6-8. Forced expression of myocardin in non-muscle cells induces expression of various cardiac and smooth muscle genes^6,8-16. However, these cells lack organized sarcomeres and do not display spontaneous contractions. In addition double immunostaining has shown that these cells co-express cardiac and smooth muscle-specific genes, raising questions concerning their phenotype^10,11,17. Myocardin does not bind DNA directly, but exerts its transcription-promoting effects by interacting with other TFs, including the MADS box-containing TFs serum response factor (SRF) and myocyte enhancer factor 2 (MEF2;^6,13). All known muscle-specific isoforms of myocardin contain an SRF binding site, but the smooth muscle-enriched variants of the protein lack a MEF2 interaction domain^6,13. Accordingly, SM-MYOCD and CM- MYOCD are expected to differ in their gene activation patterns.

In this study we compared the gene expression programs induced by the main cardiac and smooth muscle isoforms of myocardin, to investigate their contribution to the differential gene expression between heart and smooth muscle cells (SMCs) and to determine the extent to which forced CM-MYOCD gene expression induces CM differentiation in primary human cells.

To most effectively test the limits of myocardin mediated activation of the CM gene expression program we explored the effects of myocardin in human cardiac progenitor cells (CMPCs). Populations of CMPCs are present in fetal and adult human heart, can be expanded in vitro and efficiently differentiated into spontaneous beating CMs (Goumans et al., under preparation). CMPCs can therefore provide a reference sample for the transcriptome analysis and should theoretically be highly permissive to myocardin-mediated induction of cardiomyogenesis.

Gene expression was analyzed using DNA microarrays, reverse transcription (RT)-

152

polymerase chain reactions (PCRs) and immunofluorescence microscopy (IFM).

Our study shows that CM-MYOCD and SM-MYOCD are potent activators of a large variety of cardiac and smooth muscle genes. While both myocardin isoforms were equally effective in stimulation expression of their smooth muscle target genes, CM-MYOCD activated a subset of CM-specific genes much stronger than SM- MYOCD. However, also CM-MYOCD-transduced CMPCs displayed no cross-striation or spontaneous contractions, in contrast to CMPC-derived CMs. Comparison of their gene expression profiles resulted in the identification of the cardiac muscle genes that were and were not upregulated by CM-MYOCD, providing new insights into the position of myocardin within the genetic network governing CM differentiation.

Materials and Methods Lentivirus vectors

The generation of vesicular stomatitis virus G protein (VSV-G)-pseudotyped self- inactivating human immunodeficiency virus type I vectors (HIV) encoding the largest cardiac isoform of the human myocardin protein (i.e. CM-MYOCD) extended at its C terminus with an epitope of the human influenza virus A hemagglutinin (HA) or specifiying a nuclear-targeted version of Escherichia coli β-galactosidase has been previously described¹¹. To make a similar vector directing the synthesis of a C-terminally HA-tagged version of the major myocardin species found in human vascular smooth muscle tissue, the SM-MYOCD-coding sequence was amplified from human atrial RNA by RT-PCR employing the same primers and reaction conditions as were used for the cDNA cloning of CM-MYOCD¹⁰. Next, the SM-MYOCD cDNA was inserted into a derivative of plasmid pLV-CMV-IRES.eGFP¹⁸ to yield the lentivirus vector shuttle construct pLV.SM-MYOCD. The latter plasmid was used for the production in 293T cells of recombinant HIV1 vector particles as described before¹¹. For the titration of lentivirus vector stocks we employed HeLa cells using previously described methods¹¹. The 293T and HeLa cells were both cultured at 37°C in Dulbecco’s modified Eagle’s medium containing 10% fetal bovine serum (both from Invitrogen) in a humidified air-5% CO₂ atmosphere. The nucleotide sequences of pLV.SM-MYOCD has been deposited in GenBank under accession number EU000249.

Isolation, ex vivo propagation and cardiomyogenic differentiation of CMPCs

CMPCs were isolated, cultured and differentiated into CMs using 5’-azacytidine as previously described (Goumans et al., under preparation). In brief, fetal hearts were collected after elective abortion followed by Langendorff perfusion with Tyrode’s solution, collagenase and protease. After cardiomyocyte depletion of the cell suspension, CMPCs were isolated by magnetic cell sorting (MACS, Miltenyl Biotec, Sunnyvale, CA) using Sca-1-coupled beads, following the manufacturers protocol. Sca-1+-like cells were eluted from the column by washing with PBS supplemented with 2% FBS. CMPCs were expanded on gelatin coated culture dishes in growth medium consisting of equal volumes of DMEM and M199, supplemented

153

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

with 10% FBS, 10 ng/ml basic Fibroblast growth factor (bFGF), 5 ng/ml epithelial growth factor (EGF) and non-essential amino acids (all from Invitrogen). To induce differentiation, CMPCs were treated with 5 μM 5’-azacytidine (Sigma) for 72 hours in differentiation medium (Iscove’s Modified Dulbecco’s Medium /Hams F12 (1:1) (Gibco) supplemented with L-Glutamine (Gibco), 2% horse serum, non-essential amino acids, Insulin-Transferrin-Selenium supplement, and 10^-4 M Ascorbic Acid (Sigma)). After induction, the medium was changed every 3 days. For gene expression analysis, CMPCs were either mock-infected or infected with LV.C-PGK.

nls-bGal (negative control vector;¹¹), LV.CMV.myocL-HA (¹¹; hereafter referred to as LV.CM-MYOCD) or LV.SM-MYOCD. One day before transduction, 2×10⁵ CMPCs were seeded in porcine gelatin (Sigma-Aldrich)-coated 10-cm² culture dishes and placed overnight in a CO₂ incubator at 37°C. The next morning, the CMPCs were infected with 1 HeLa cell-transducing unit (HTU) of lentivirus vector per cell. To facilitate vector-host cell interaction, the recombinant HIV1 particles were diluted in CMPC growth medium containing 20 μg/ml of diethylaminoethyl-dextran sulfate (GE Healthcare). After 4 hours, the inoculum was removed, the cells were washed thrice with phosphate-buffered saline (PBS) and fresh CMPC growth medium was added. All infection experiments were repeated three times except for the transduction of CMPCs with LV.CM-MYOCD, which was carried out in quadruplicate.

At 7 days post-infection, when essentially all cells stained positive for the HA tag and for sarcomeric

α

-actinin (sACTN) indicating myocardin-induced activation of cardiac muscle genes, total cellular RNA was extracted from each sample using the NucleoSpin® RNA II purification kit (Macherey Nagel). The CMPC cultures that were converted into CMs using 5’-azacytidine were processed for RNA extraction after 21 days. At this time, CMPCs were spontaneously contracting and displayed immunoreactivity for sACTN (Goumans et al., under preparation).

Oligonucleotide microarray (OMA) analysis

Biotin-labeled cRNA was synthesized from 500 ng RNA per sample, employing the Illumina® TotalPrep™ RNA amplification kit (Ambion). The biotin-labeled cRNA was hybridized to HumanRef-8 v2 Expression BeadChips (Illumina) and analyzed following the recommendations of the manufacturer using the BeadStation 500 microarray reader and BeadStudio v2 data analysis software (both from Illumina).

Expression data were normalized using a global variance stabilization normalization (VSN) procedure¹⁹ and differential expression was analyzed using the linear models for microarray analysis (LIMMA) software package in R²⁰. Gene classification was performed using the panther classification system (www.pantherdb.org);

significant over- or underrepresentation of gene classes was determined using the binominal test, as described²¹.

PCR and immunofluorescence analyses

Human myocardin RNA with or without exon 2a was amplified by conventional RT-PCR using primers 5’-CAAAGAAGGACCCAGGAACA-3’ and 5’- GCAGTTGGAATGGACCTCTC-3’, an annealing temperature of 60 °C and a total number of 30 PCR cycles¹⁰. Quantitative RT-PCR (qPCR) was performed in

154

quadruplicate using a previously described procedure¹¹. The input material consisted of aliquots from RNA samples that were also used for the OMA analysis. The primer collection for the qPCR was extended with oligonucleotides specific for cardiac

α

-actin (ACTC; 5’-CTGTGCCAAGATGTGTGACG-3’ and 5’- TGGGATACTTCAGGGTCAGG-3’), cardiac myosin-binding protein C (MYBPC3;

5’-CTGGACGTCCCTATCTCTGG-3’ and 5’-TGTAGACGCCCTCATCTTCC-3’), embryonic myosin light chain 4 (MYL4; 5’-CACTCCTCTGCCAAAGATCC-3’ and 5’-CCGGTCAAACAATGAAAAGG-3’), regulatory myosin light chain 7 (MYL7; 5’- GGTTCTTCCAACGTCTTTTCC-3’ and 5’-CTGTCCCATTGAGCTTCTCC-3’) and titin cap (TCAP; 5’-GATCTGACACTGTCCACACG-3’ and 5’-CCTCACGCTCCTCCTTGG-3’) RNA and the formerly used qPCR primers for analyzing natriuretic peptide precursor A (NPPA) transcripts were replaced by oligonucleotides 5’-TGTACAATGCCGTGTCCAAC- 3’ and 5’-TCTTCATTCGGCTCACTGAG-3’. qPCRs carried out on cDNA derived from human total cellular RNA of the left cardiac ventricle, the left internal mammary artery and the vastus lateralis muscle served as positive controls and PCRs in which the cDNA was replaced by water were included as negative controls.

Immunofluorescent labeling of CMPCs was performed as described before¹⁰ with the addition of monoclonal antibody MF20 (Developmental Studies Hybridoma Bank at the University of Iowa), which binds sarcomeric myosin heavy chain (sMHC).

For checking transduction and/or myogenic conversion levels, CMPC cultures were used that underwent exactly the same pretreatment as those employed for the RNA analysis. However, to show the intracellular distribution of specific cardiac or smooth muscle marker proteins, low-density cultures of CMPCs were employed and for the LV.CM-MYOCD-treated CMPCs, the vector dose was reduced to 0.5 HTU per cell.

Online datasupplement

Table S1-S4 are available as online supplements.

Table 1. Muscle genes activated in CMPCs after forced CM-MYOCD expression transcript fold indunction P full name

ACTC 1788.1 5.09E-15 actin, alpha, cardiac muscle

TNMD 969.2 3.07E-18 tenomodulin

MYH11 820.2 8.67E-11 myosin, heavy chain 11, smooth muscle MYL7 590.9 3.07E-18 myosin, light chain 7, regulatory NPPA 519.7 2.75E-18 natriuretic peptide precursor type A EFHD1 452.6 1.71E-16 EF-hand domain family, member D1

MYL4 344.9 2.82E-18 myosin, light chain 4, alkali; atrial, embryonic CASQ2 343.1 1.42E-16 calsequestrin 2 (cardiac muscle)

TNNT2 281.5 2.61E-17 troponin T type 2 (cardiac) NPPB 238.6 5.21E-12 natriuretic peptide precursor type B

MYOCD 238.1 3.74E-11 myocardin

TNNC1 227.6 1.10E-13 troponin C type 1 cardiac / slow skeletal

MYH3 115.0 4.00E-14 myosin, heavy chain 3, skeletal muscle, embryonic CMYA1 103.2 8.73E-17 cardiomyopathy associated 1

ACTG2 95.2 9.28E-09 actin, gamma 2, smooth muscle, enteric SYNPO2L 93.4 2.24E-06 synaptopodin 2-like

MYOZ1 89.3 4.18E-18 myozenin 1

155

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

APOBEC2 86.2 1.09E-13 apolipoprotein B editing complex 2

KCNMB1 72.4 8.26E-17 potassium large conductance calcium-activated channel, subfamily M, beta member 1

MYOM1 67.0 3.12E-08 myomesin 1 (skelemin) TMOD1 59.7 1.11E-05 tropomodulin 1

MYLK 40.8 3.17E-07 myosin, light chain kinase

ITGB1BP2 38.3 6.67E-11 integrin beta 1 binding protein (melusin) 2 LDB3 33.7 1.85E-15 LIM domain binding 3

CAST1 31.3 4.58E-14 calpastatin

NKX2-2 28.4 7.12E-14 NK2 transcription factor related, locus 2 (Drosophila) IL19 20.2 4.81E-13 interleukin 19

HAND1 19.4 1.93E-12 heart and neural crest derivatives expressed 1 CKB 18.1 1.23E-13 creatine kinase, brain

ACE2 17.2 2.72E-13 angiotensin I converting enzyme (peptidyl-dipeptidase A 2) BMP2 16.3 1.58E-05 bone morphogenetic protein 2

PPP1R14A 15.9 2.22E-12 protein phosphatase 1, regulatory (inhibitor) subunit 14A CKM 15.8 1.30E-13 creatine kinase, muscle

POPDC2 12.1 6.30E-07 popeye domain containing 2 MYOM2 12.0 3.69E-08 myomesin (M-protein) 2 CNN1 10.1 4.83E-07 calponin 1, basic, smooth muscle DLX5 9.2 3.84E-12 distal-less homeobox 5

LMOD1 9.2 2.00E-09 leiomodin 1 (smooth muscle) DUSP2 9.0 2.45E-11 dual specificity phosphatase 2 CRYAB 7.8 2.39E-02 crystallin, alpha B

BVES 7.4 2.57E-08 blood vessel epicardial substance TAGLN 6.4 4.90E-09 Transgelin (SM22)

CSRP2 5.2 6.00E-07 cysteine and glycine-rich protein 2 DTNA 5.1 3.35E-12 dystrobrevin, alpha

CACNA1C 5.1 3.35E-06 calcium channel, voltage-dependent, L type, alpha 1C subunit

DMD 5.1 1.49E-11 Dystrophin

MYL1 4.9 5.12E-09 myosin, light chain 1, alkali; skeletal, fast MYH6 4.8 5.61E-08 myosin, heavy chain 6, cardiac muscle, alpha TNC 4.3 2.72E-03 tenascin C (hexabrachion)

PDLIM4 3.8 3.47E-07 PDZ and LIM domain 4 FHL1 3.8 1.81E-07 four and a half LIM domains 1

TTN 3.6 2.95E-08 Titin

ADAM19 3.5 5.53E-10 a disintegrin and metalloproteinase domain 19 CACHD1 3.4 1.73E-08 cache domain containing 1

ADAMTS1 3.4 2.98E-02 ADAM metallopeptidase with thrombospondin type 1 motif MYL9 3.0 4.51E-09 myosin, light chain 9, regulatory

BOC 3.0 3.03E-04 Boc homolog (mouse)

FHL2 3.0 1.78E-06 four and a half LIM domains 2

CAMK2B 3.0 3.28E-05 calcium/calmodulin-dependent protein kinase (CaM kinase) II beta

EDNRA 2.9 5.28E-03 endothelin receptor type A

TG 2.8 1.12E-07 Thyroglobulin

ITPR1 2.8 7.15E-03 inositol 1,4,5-triphosphate receptor, type 1 EDNRB 2.6 7.81E-07 endothelin receptor type B, transcript variant 2 MYO18B 2.6 4.53E-05 myosin XVIIIB

KCNMA1 2.6 1.19E-02 potassium large conductance calcium-activated channel, subfamily M, alpha member 1

PDLIM5 2.5 2.31E-05 PDZ and LIM domain 5 TPM1 2.5 2.85E-03 tropomyosin 1 (alpha)

MYL2 2.4 8.41E-06 myosin, light chain 2, regulatory, cardiac, slow

UCN2 2.3 6.56E-08 urotensin 2

MYH7B 2.2 1.57E-04 myosin, heavy chain 7B, cardiac muscle, beta C20orf35 2.2 3.07E-04 chromosome 20 open reading frame 35

MEF2A 2.2 2.01E-03 MADS box transcription enhancer factor 2, polypeptide A (myocyte enhancer factor 2A)

TPM2 2.1 2.18E-06 tropomyosin 2 (beta) FHL5 2.0 6.64E-06 four and a half LIM domains 5

156

Results

CM-MYOCD induces expression of both cardiac and smooth muscle genes in CMPCs

To investigate the cardiomyogenic potential of the largest isoform of human myocardin, CMPCs were infected with the lentivirus vector LV.CM-MYOCD. OMA analysis of these cells showed that like in other primary human cells^10,11,17, forced expression of CM-MYOCD in CMPCs induced the transcription of a large variety of cardiac muscle genes encoding, among others, cardiac channel proteins, hormones and sarcomeric components (Table 1 and Table S1 in the data supplement). CM- MYOCD also upregulated the expression of several typical smooth muscle genes, such as those encoding enteric smooth muscle γ-actin (ACTG2) and smooth muscle

CMPCs + CM-MYOCD

ANFMYH11

sACTNsMHC

*

ACTB GAP

DHACTC CAC

NA1 C

MYH11 MYL4

MYL 7

NPPATNNT2 0.011000.1101

100001000 100000 10000001.0×10⁰⁷

Fold induction

ACTB GAPDH ACTC

CAC NA1C

MYH 11MYL4

MYL7 NPP

A TNNT2 ACTB

GAP DHACTC

CAC NA1

C MYH11

MYL4 MYL7

NPPA TNNT2

* ***** *** *

*** *

CMPCs + CM-MYOCD ventricular CMs vascular SMCs

CMPCs + mock CMPCs + CM-MYOCD CMPCs + mock

.$ .$

Figure 1. (A) qPCR analysis of cardiac and smooth muscle transcripts in CM-MYOCD- transduced CMPCs, human ventricular CMs and human vascular SMCs. Gene expression was normalized to both ß-actin (ACTB) and glyceraldehyde-3-phosphate dehydrogenase (GAPDH) transcript levels. Induction factors are relative to mock-treated CMPCs. Both means (n=4) and standard errors are shown; *indicates transcript levels that are significantly upregulated (p<0.05), ND indicates transcript levels below detection limit. (B) Immunolabeling of CM- MYOCD- or mock-transduced CMPCs with antibodies specific for the cardiac muscle proteins sACTN, NPPA and sMHC or the smooth muscle marker MYH11. Nuclei were stained with the blue-emitting fluorochrome Hoechst 33342. Due to the use of a relatively low multiplicity of infection (see Materials and Methods) not all LV.CM-MYOCD-treated cells were positive for the marker proteins.

A

B

157

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

myosin heavy chain (MYH11) (Tables 1 and S1). The induction in CM-MYOCD- transduced CMPCs of ACTC, L-type calcium channel

α

1C (CACNA1C), MYH11, MYL4, MYL7, NPPA and cardiac troponin T (TNNT2) was confirmed by qPCR (Figure 1A). Comparison of mRNA induction levels to those in reference samples from ventricular cardiac tissue and left interior mammary artery derived smooth muscle tissue confirmed that transcription of ACTC, CACNA1C, MYL4, MYL7, NPPA and TNNT2 is specific for cardiomyocytes, while MYH11 transcripts are detectable in both tissues, but are more abundant in smooth muscle tissue. The stimulatory effect of CM-MYOCD on sACTN, MYH11 and NPPA gene expression in CMPCs was also evident by IFM (Figure 1B). Moreover, consistent with the upregulation of especially the embryonic myosin heavy chain (MYH3) gene observed in the OMA analysis, the CM-MYOCD-transduced cells but not the mock-treated CMPCs stained Table 2. Cardiac muscle genes upregulated in CMPC-derived CMs but not in CM-MYOCD- transduced CMPCs.

transcript fold induction genbank full name

TCAP 80.6 NM_003673 titin-cap (telethonin)

MYBPC3 55.6 NM_000256 myosin binding protein C 3 (cardiac) RYR2 41.6 NM_001035 ryanodine receptor 2 (cardiac) TNNI3 41.4 NM_000363 troponin I 3 (cardiac)

CKMT2 21.6 NM_001825 creatine kinase mitochondrial 2 (sarcomeric)

NEB 6.8 NM_004543 Nebulin

POSTN 6.3 NM_006475 periostin; osteoblast specific factor 2 (fasciclin I-like) SEMA3C 4.1 NM_006379 sema domain immunoglobulin domain short basic domain

secreted (semaphorin) 3C

KRT19 4.1 NM_008471 keratin 19

CRIP2 3.7 NM_001312 cysteine-rich protein 2

MYOZ3 3.7 NM_133371 Myozenin 3

MEF2C 3.6 NM_002397 MADS box transcription enhancer factor 2 polypeptide C TNNT3 3.0 NM_006757 troponin T3 (skeletal fast)

TPM3 3.0 NM_152263 tropomyosin 3

UNC93A 2.8 NM_018974 unc-93 homolog A (C. elegans)

CAMK2A 2.7 NM_171825 calcium/calmodulin-dependent protein kinase II alpha transcript variant 2

SCN7A 2.6 NM_002976 sodium channel voltage-gated type VII alpha DUSP16 2.5 NM_030640 dual specificity phosphatase 16

DMRT2 2.5 NM_181872 doublesex and mab-3 related transcription factor 2 MYF5 2.5 NM_005593 myogenic factor 5

MYOM3 2.5 NM_152372 myomesin 3

CAMK1 2.4 NM_003656 calcium/calmodulin-dependent protein kinase I BMP10 2.4 NM_014482 bone morphogenetic protein 10

LMO7 2.4 NM_005358 LIM domain only 7

POPDC3 2.4 NM_022361 popeye domain containing 3

TRDN 2.3 NM_006073 Triadin

ADORA1 2.3 NM_000674 adenosine A1 receptor

JARID2 2.3 NM_004973 jumonji homolog (mouse) (JMJ).

NKX2-8 2.2 NM_014360 NK2 transcription factor related locus 8 (Drosophila) ADRA1A 2.2 NM_000680 adrenergic alpha-1A- receptor transcript variant 1.

CACNB2 2.2 NM_201590 calcium channel voltage-dependent beta 2 subunit transcript variant 3.

CHRND 2.2 NM_000751 cholinergic receptor nicotinic delta polypeptide C3AR1 2.2 NM_004054 complement component 3a receptor 1

TTF1 2.1 NM_007344 transcription termination factor RNA polymerase I VAMP2 2.1 NM_014232 vesicle-associated membrane protein 2 (synaptobrevin 2) MAK 2.1 NM_005906 male germ cell-associated kinase

SEMA6D 2.0 NM_153618 sema domain transmembrane domain and cytoplasmic domain (semaphorin) 6D transcript variant 4.

158

positive for sMHC (Figure 1B). These data confirm the role of myocardin as a regulator of CM differentiation. However, the absence of spontaneous contraction, the lack of properly organized sarcomeres (Figs. 1B, 2B and 4B) and the expression of typical smooth muscle genes in LV.CM-MYOCD-infected CMPCs (Table 1, Figs. 1B and 4B) suggest that cardiomyogenic differentiation of these cells is incomplete.

CM-MYOCD does not transactivate all cardiac muscle genes in CMPCs

In a previous study, Goumans and colleagues have shown that upon exposure to 5’-azacytidine-containing differentiation medium, CMPCs differentiate into beating CMs in vitro (Goumans et al., under preparation). This allowed us to compare, on the same cellular background, the gene expression profile of functional heart muscle cells with that imposed by CM-MYOCD. Table S2 lists the genes whose expression was >2-fold upregulated in CMPC-derived CMs as compared to undifferentiated CMPCs, but which were not activated in CM-MYOCD-transduced CMPCs. Among the genes upregulated in CMPC-derived CMs but not in CM-MYOCD-transduced CMPCs, 37 encoded muscle-associated polypeptides (Table 2) and 54 specified ion channel proteins (Table 3). Of the four CM-MYOCD-unresponsive genes that were most potently upregulated in CMPC-derived CMs, MYBPC3 and TCAP encode proteins involved in the regulation of myofibrillogenesis, RYR2 specifies the cardiac sarcoplasmic reticulum calcium release channel and TNNI3 encodes cardiac troponin I, which controls actin-myosin interactions. qPCR analysis confirmed that MYBPC3 and TCAP were not significantly induced by CM-MYOCD indicating that other factors are responsible for the increase in their expression in CMPC-derived CMs and ventricular CMs as compared to in mock- or LV.CM-MYOCD-infected CMPCs (Figure 2A). In contrast, the levels of ACTC, CACNA1C, MYL4, MYL7, NPPA and TNNT2 mRNA, did not significantly differ between LV.CM-MYOCD-infected CMPCs, CMPC-derived CMs and ventricular CMs (Figure 2A). Although CM-MYOCD stimulated the expression of genes encoding the major components of the cardiac sarcomeric filament systems (Tables 1 and S1), cross-striation was only observed in CMPC-derived CMs and not in LV.CM-MYOCD-infected CMPC after labeling of representative CMPC cultures with sACTN-specific antibodies (Figure 2B). As expected by the absence of properly organized sarcomeres, the CM-MYOCD- transduced CMPCs did not show spontaneous contractions. Apart from showing remarkable similarities in cardiac gene expression, ventricular CMs, CMPC-derived CMs and CM-MYOCD-transduced CMPCs also expressed various smooth muscle genes (e.g. MYH11; Figs. 1, 2, 4b and compare Tables S1 and S2). In double immunolabeling experiments using antibodies directed against sACTN and MYH11, all LV.CM-MYOCD-infected and 5’-azacytidine-treated CMPCs stained positive for both marker proteins, indicating co-expression of striated and smooth muscle genes in single cells (Figure 2B).

CM-MYOCD induces cardiac muscle genes more effectively than SM-MYOCD

RT-PCR analysis of total RNA extracted from the aorta, heart and lungs of a mouse has shown that primary myocardin transcripts are alternatively spliced in murine cardiac and smooth muscle tissue¹³, giving rise to cardiac and smooth muscle-

159

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

Table 3. Ion channels and transporter genes upregulated in CMPC-derived CMs but not in CM-MYOCD-transduced CMPCs

transcript fold induction genbank full name

RYR2 41.6 NM_001035 ryanodine receptor 2 (cardiac) KCNK1 17.8 NM_002245 potassium channel subfamily K member 1 CLIC5 13.2 NM_016929 chloride intracellular channel 5

SLC26A7 9.0 NM_052832 solute carrier family 26 member 7 transcript variant 1.

KCNA5 8.5 NM_002234 potassium voltage-gated channel shaker-related subfamily member 5 SLC4A3 6.2 NM_005070 solute carrier family 4, anion exchanger, member 3 transcript variant 1,.

SLC26A9 6.0 NM_134325 solute carrier family 26 member 9 transcript variant 2.

FXYD1 5.9 NM_021902 FXYD domain containing ion transport regulator 1 (phospholemman) transcript variant b.

KCNMB4 5.7 NM_014505 potassium large conductance calcium-activated channel subfamily M beta member 4

IF 5.2 NM_000204 I factor (complement)

CLIC6 5.0 NM_053277 chloride intracellular channel 6 ACP5 5.0 NM_001611 acid phosphatase 5 tartrate resistant

KCTD9 4.7 NM_017634 potassium channel tetramerisation domain containing 9 SLCO2B1 4.4 NM_007256 solute carrier organic anion transporter family, member 2B1 MRS2L 4.3 NM_020662 MRS2-like magnesium homeostasis factor (S. cerevisiae) KCNJ8 4.1 NM_004982 potassium inwardly-rectifying channel subfamily J member 8 SLC16A9 3.9 NM_194298 solute carrier family 16 (monocarboxylic acid transporters), member 9 GABRB2 3.5 NM_021911 gamma-aminobutyric acid A receptor beta 2 transcript variant 1.

KCNJ14 3.5 NM_013348 potassium inwardly-rectifying channel subfamily J member 14 transcript variant 1.

UCP3 3.4 NM_003356 uncoupling protein 3 (mitochondrial proton carrier)

ATP6V1G3 3.4 NM_133262 ATPase H+ transporting lysosomal 13kDa V1 subunit G isoform 3 transcript variant 1.

TSC 3.3 NM_017899 thiazide-sensitive na-cl cotransporter (SLC12A3) SLC9A3 3.3 NM_004174 solute carrier family 9 isoform 3

RYR3 3.3 NM_001036 ryanodine receptor 3

SLC22A4 3.1 NM_003059 solute carrier family 22 member 4

TM4SF11 3.0 NM_015993 transmembrane 4 superfamily member 11 (plasmolipin) ATP6V0A2 2.8 NM_012463 ATPase, H+ transporting, lysosomal V0 subunit a2 ACCN3 2.8 NM_020322 amiloride-sensitive cation channel 3 transcript variant 3.

VDAC3 2.7 NM_005662 voltage-dependent anion channel 3 CNGB1 2.7 NM_001297 cyclic nucleotide gated channel beta 1 SCN7A 2.6 NM_002976 sodium channel voltage-gated type VII alpha

CUTC 2.6 NM_015960 copper transporter protein

SLC9A6 2.5 NM_006359 solute carrier family 9 isoform 6 GABRA1 2.5 NM_000806 gamma-aminobutyric acid A receptor alpha 1

CLCNKB 2.5 NM_000085 chloride channel Kb

KCNB2 2.4 NM_004770 potassium voltage-gated channel Shab-related subfamily member 2 SLC17A8 2.4 NM_139319 solute carrier family 17 member 8

KCNK4 2.3 NM_033311 potassium channel subfamily K member 4 transcript variant 3.

SLC10A2 2.3 NM_000452 solute carrier family 10 member 2 GABRA5 2.3 NM_000810 gamma-aminobutyric acid A receptor alpha 5 ATP6V1D 2.3 NM_015994 ATPase H+ transporting lysosomal 34kDa V1 subunit D

TPCN1 2.3 NM_017901 two pore segment channel 1

CHRND 2.2 NM_000751 cholinergic receptor nicotinic delta polypeptide RSC1A1 2.2 NM_006511 regulatory solute carrier protein family 1 member 1 VDAC1 2.2 NM_003374 voltage-dependent anion channel 1

CLCN2 2.1 NM_004366 chloride channel 2

KCNQ2 2.1 NM_004518 potassium voltage-gated channel KQT-like subfamily member 2 transcript variant 3.

CLNS1A 2.1 NM_001293 chloride channel nucleotide-sensitive 1A

KCNC2 2.1 NM_139137 potassium voltage-gated channel Shaw-related subfamily member 2 transcript variant 2.

ATP6V0A4 2.1 NM_130841 ATPase H+ transporting lysosomal V0 subunit a isoform 4 transcript variant 3.

PBP 2.1 NM_002567 prostatic binding protein

SLC1A7 2.1 NM_006671 solute carrier family 1 (glutamate transporter) member 7

CAPS2 2.1 NM_032606 calcyphosphine 2

SLC17A3 2.1 NM_006632 solute carrier family 17 member 3

160

CMPCs + mock CMPCs + CM-MYOCD

sACTN MYH11 merge merge

untreated CMPCs CMPCs + 5’-azacytidine

ACTBGAPDH ACTC CAC

NA1C MY

BPC MY

H11 MY

L4 MY

L7

NPPA TCAP TNNT2 0.1

1 10 100 1000 10000 100000 1000000 1.0×10⁰⁷

CMPCs + CM-MYOCD CMPCs + 5’-azacytidine ventricular CMs

Fold induction



 

 



* *

Figure 2. (A) qPCR analysis of cardiac and smooth muscle transcripts in CM-MYOCD- transduced CMPCs, 5’-azacytidine-treated CMPCs and human ventricular CMs. Gene expression was normalized to both ACTB and GAPDH transcript levels. Induction factors are relative to mock-treated CMPCs. Both means (n=4) and standard errors are shown;

#indicates transcript levels that are not significantly upregulated (p>0.05) and *denotes significantly higher transcript levels in CMPC-derived and ventricular CMs as compared to those in LV.CM-MYOCD-infected CMPCs (p<0.05). (B) Double immunolabeling of CM-MYOCD- transduced or 5’-azacytidine-treated CMPCs with antibodies directed against sACTN (green) and MYH11 (red). Nuclei were stained with the blue-emitting fluorochrome Hoechst 33342.

Note the lack of cross-striation in LV.CM-MYOCD-infected CMPCs.

A

B

161

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

enriched isoforms of myocardin. To test whether human myocardin exhibits similar splicing, RNA extracted from human cardiac, skeletal and smooth muscle tissue was subjected to RT-PCR using PCR primers flanking the differentially spliced exon 2a (Figure 3A). This experiment demonstrated that, like in mice, the majority of myocardin transcripts in human vascular smooth muscle tissue contain exon 2a while most myocardin-encoding RNAs in human ventricular myocardium lack this exon (Figure 3B). Nucleotide sequence analysis of full-length cDNA representative of the predominant smooth muscle-specific splice variant of human myocardin confirmed the presence of exon 2a and showed that inclusion of this exon introduces an additional stop-codon. As a result the smooth muscle-enriched isoforms of myocardin are 79 residues shorter at their amino terminus compared to the heart muscle-enriched myocardin species (Figure 3A). To investigate whether the cardiac and smooth muscle-enriched isoforms of myocardin exhibit different patterns of transactivation, we compared the gene expression profile of LV.CM-MYOCD-infected CMPCs (Table S1) with that of SM-MYOCD-transduced CMPCs (Table S3). At equal myocardin transcript levels (compare Table S1 with S3), CM-MYOCD significantly upregulated 475 genes more than 2-fold as compared to SM-MYOCD (Table S4). Functional classification of these significantly overexpressed genes revealed, significant overrepresentation of genes involved in muscle contraction (p=1.89×10^-5), muscle development (p=6.07×10^-4) and cell cycle control (p=3.36×10^-23). As is evident from Tables 4 and 5, 63 out of 475 genes (i.e. 13%) encoded muscle-associated proteins and 95 out of 475 genes (i.e. 20%) encoded proteins involved in cell cycling. qPCR analysis confirmed that the heart muscle-specific genes MYL4, NPPA and TNNT2 were more efficiently activated by CM-MYOCD than by SM-MYOCD, whereas several other CM-related genes (i.e. ACTC, CACNA1C, MYL7) and the smooth muscle-specific gene MYH11 were equally well induced by both isoforms of myocardin (Figure 4A). In keeping with the RNA analyses, both LV.CM-MYOCD- and LV.SM-MYOCD-infected CMPCs stain equally positive for MYH11, but only LV.CM-MYOCD-treated CMPCs stain positive for NPPA (Figure 4B). SM-MYOCD significantly induced 383 genes more than 2-fold over CM-MYOCD, however this group of genes did not contain any significant overrepresentation of functional classes.

Discussion

In this study, OMA analyses and complementary techniques were used to compare the gene expression profile of unstimulated CMPCs with that of CMPCs subjected to myocardin- or 5’-azacytidine-induced cardiomyogenesis. We found that forced human myocardin expression in CMPCs initiates a transcriptional program that strongly resembles the gene expression program of CMPC-derived CMs. However, transcripts for several crucial sarcomeric components and channel proteins present in CMPC-derived CMs are underrepresented or absent in myocardin-transduced CMPCs. This might explain why these cells do not contract and/or display the cross-striation characterizing true striated muscle cells. We also found that the CM-enriched isoform of myocardin was a more potent inducer of a subset of cardiac muscle genes than the smooth muscle-enriched variant of the protein.

162

Myocardin is a key regulator of smooth muscle gene expression

Although myocardin is generally accepted to play an important role in CM differentiation and homeostasis, its ability to transactivate smooth muscle genes has been given much more emphasis than its stimulatory effect on the activity of cardiac muscle genes. This may relate to the fact that in an early study in mouse embryonic fibroblasts (i.e. 10T1/2 cells), an N-terminally truncated version of murine myocardin (trMYOCD; containing residues 128 through 935) activated all smooth muscle genes that were tested but failed to activate NPPA

ACTAATCCAAGCTACAGCAGCCACATACTTACTGCAAAACCTTG stop

exon 2a

ATG1 ATG2

MEF2 SAP LZ TAD

SM-MYOCD (41 + 907 aa) CM-MYOCD (986 aa)

exon 1-2 exon 3-14

2a

K

GAPDH SM-MYOCD CM-MYOCD NC smooth musclecardiac muscleskeletal muscle

F R

SRF Q

Figure 3. (A) Schematic representation of the cardiac and smooth muscle-enriched splice variants of the human myocardin gene and their translation products. While the CM-enriched myocardin mRNAs are assumed to encode one protein of 986 (+ exon 11) or 938 (- exon 11) amino acids, the smooth muscle-enriched splice variants of myocardin potentially code for two proteins of 41 and either 907 (+ exon 11) or 859 (- exon 11) amino acids. MEF2, MEF2 binding domain; SRF, SRF binding domain; K, lysine-rich region; Q, proline-rich region; SAP, SAF-A, Acinus, PIAS domain; LZ, leucine zipper motif; TAD, transactivation domain. (B) RT-PCR analysis of myocardin splice variants with and without exon 2a. In ventricular heart muscle tissue, myocardin mRNAs lacking exon 2a (named CM-MYOCD) are most abundant whereas in vascular smooth muscle tissue exon 2a-containing myocardin transcripts (named SM-MYOCD) dominate. Consistent with our previous results10, no myocardin transcripts were detected in human skeletal muscle tissue. The binding sites of the PCR primers were located in exons 2 and 5 as indicated in panel A (F and R). GAPDH transcripts were used as internal control and a sample containing water instead of cDNA template served as negative control (NC).

A

B

163

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

and cardiac

α

-myosin heavy chain (MYH6)¹⁴. On the basis of these and other findings myocardin was variably called a key or master regulator of smooth muscle gene expression^14,15. In agreement with these reports^9,14,15 SM-MYOCD stimulated the expression in CMPCs of MYH11, ACTG2, myosin light chain kinase (MYLK), desmin (DES), h-calponin (CNN1), SM22 (TAGLN), h-caldesmon (CALD1), and smooth muscle

α

-actin (ACTA2). These smooth muscle genes share the presence of functional SRF binding sites (i.e. CArG boxes;^22,23) in their promoter regions.

Other typical smooth muscle genes that –in our study- were potently induced following the transduction of CMPCs with SM-MYOCD included the β subunit of the maxi-K_Ca channel (KCNMB1) and nephroblastoma overexpressed protein (NOV).

Upregulation of NOV in response to trMYOCD has been reported previously⁹. However a previous study in which smooth muscle transcripts induced by trMYOCD were compared to those present after retinoic acid-induced SMC differentiation of A404 cells, has demonstrated that not all smooth muscle genes are activated by myocardin¹⁶. In agreement with these results, several smooth muscle genes including aortic carboxypeptidase-like protein (AEBP1), aortic preferentially expressed protein 1 (APEG1), elastin (ELN), histidine-rich calcium-binding protein (HRC), integrin

α

8 (ITGA8), NOTCH3, focal adhesion kinase-related nonkinase (PTK2) and smoothelin (SMTN) were not upregulated after transduction of CMPCs with SM-MYOCD (Table S3).

Collectively, these data indicate that (i) the failure of myocardin to activate the complete smooth muscle transcriptome in past experiments was not caused by the use of an N-terminally truncated version of SM-MYOCD, (ii) smooth muscle gene activation by SM-MYOCD is largely restricted to SRF-dependent genes, and (iii) SM-MYOCD is a key, but not a master regulator of smooth muscle gene expression.

Myocardin is a key regulator of cardiac muscle gene expression

In addition to activating CArG-box-containing smooth muscle genes, SM-MYOCD also potently stimulated transcription in CMPCs of a wide diversity of heart muscle genes encoding among others metabolic enzymes, sarcomeric components, channel proteins and transcription factors (Table S3). Among the genes which are most strongly activated by myocardin are ACTC, NPPA, NPPB, whose myocardin- depended activation has been observed previously in a variety of cell types^8-11,16,17. In addition, we identified many other novel CM-related target genes of SM-MYOCD including myozenin 1 (MYOZ1), tropomodulin 1 (TMOD1), Z-band alternatively spliced PDZ motif-containing protein (LDB3), desmuslin (DMN) and junctophilin 2 (JPH2).

Creemers and co-workers recently demonstrated that the N-terminus of the full- length myocardin protein contains a MEF2 binding region and that as a result of tissue-specific alternative splicing the predominant myocardin species in smooth muscle differ from the CM-enriched isoforms of myocardin in lacking this interaction domain¹³. These authors also showed for three different MEF2-dependent promoters (derived from myosin light chain 1/3F [MYL1], SET and MYND domain containing protein 1 [SMYD1] and serine/arginine-rich protein specific kinase 3 [SRPK3]) that they were activated to a much greater extent by a combination of MEF2C

164

and the CM-enriched isoform of myocardin than by either of these TFs alone.

This synergistic effect was lost after mutation of the MEF2 binding sites in these promoters. We found that splicing of human myocardin closely resembled that of murine myocardin (Figure 3B) and were therefore interested whether the cardiac and smooth muscle-enriched isoforms of myocardin display different gene induction profiles. A comparison of Tables S1 and S3 revealed that, with the exception of CALD1, all genes encoding smooth muscle marker genes²³ were induced equally

Table 4. Fold difference in muscle gene expression levels between CM-MYOCD- and SM- MYOCD-transduced CMPCs

transcript fold change P full name

APOBEC2 59.0 1.33E-11 apolipoprotein B mRNA editing enzyme, catalytic polypeptide- like 2

TNNC1 44.1 7.53E-13 troponin C, cardiac/slow skeletal SYNPO2L 33.5 8.13E-13 synaptopodin 2-like

CMYA1 16.5 8.13E-13 cardiomyopathy associated 1 CCL3 16.1 1.34E-12 chemokine (C-C motif) ligand 3 TXNIP 14.4 1.47E-13 thioredoxin interacting protein CCL3L3 13.6 2.51E-12 chemokine (C-C motif) ligand 3-like 3 CCL3L1 13.0 1.52E-11 chemokine (C-C motif) ligand 3-like 1 MYOM1 12.7 7.53E-13 myomesin 1 (skelemin)

ALDH1A1 12.6 8.13E-13 aldehyde dehydrogenase 1 family, member A1 CKM 12.4 7.69E-12 creatine kinase, muscle

TGFB3 8.7 1.06E-11 transforming growth factor, beta 3 KIF20A 8.2 1.73E-12 kinesin family member 20A SMPX 8.0 6.38E-11 small muscle protein, X-linked NPPA 7.4 1.94E-11 natriuretic peptide precursor type A

FABP3 6.7 6.12E-09 fatty acid binding protein 3, muscle and heart HSPB3 6.2 3.89E-11 heat shock 27kDa protein 3

UNC45B 5.9 1.01E-09 unc-45 homolog B (C. elegans) CRYAB 5.9 1.28E-12 crystallin, alpha B

ASPM 5.8 8.24E-10 asp (abnormal spindle) homolog, microcephaly associated (Drosophila)

SPARCL1 5.8 9.69E-08 SPARC-like 1 (mast9, hevin) SMYD1 5.5 2.29E-07 SET and MYND domain containing 1 CALD1 5.3 6.23E-09 caldesmon 1

TNNT2 5.1 1.31E-09 troponin T type 2 (cardiac) CASQ2 5.1 2.03E-09 calsequestrin 2 (cardiac muscle) MYOM2 5.0 3.65E-08 myomesin (M-protein) 2

GRIN2C 5.0 1.05E-09 glutamate receptor, ionotropic, N-methyl D-aspartate 2C ACTN2 5.0 1.30E-09 actinin alpha 2

MYL4 4.9 8.29E-11 myosin, light chain 4, alkali; atrial, embryonic MYH3 4.9 3.12E-10 myosin, heavy chain 3, skeletal muscle, embryonic UCP2 4.7 6.22E-08 uncoupling protein 2 (mitochondrial, proton carrier) MYL1 4.6 1.90E-08 myosin, light chain 1, alkali; skeletal, fast

EFHD1 4.6 2.02E-09 EF-hand domain family, member D1 LMCD1 4.5 3.09E-10 LIM and cysteine-rich domains 1 TAGLN3 4.4 2.39E-10 transgelin 3

ITGB1BP2 4.4 1.74E-11 integrin beta 1 binding protein (melusin) 2

MYH6 3.7 2.09E-06 myosin, heavy chain 6, cardiac muscle, alpha (cardiomyopathy, hypertrophic 1)

CENPF 3.7 1.74E-10 centromere protein F, 350/400ka (mitosin)

SEMA3A 3.3 1.72E-07 sema domain, immunoglobulin domain, short basic domain, secreted, (semaphorin) 3A

LMO3 3.3 2.66E-07 LIM domain only 3 (rhombotin-like 2)

MYOZ1 3.3 8.98E-11 myozenin 1

HAND1 3.3 9.85E-07 heart and neural crest derivatives expressed 1

TTN 3.2 6.04E-07 titin

P2RX1 3.1 2.02E-09 purinergic receptor P2X, ligand-gated ion channel, 1

165

Chapter 7Gene expression profiling of myocardin-induced cardiomyogenic differentiation in human cardiac progenitor cells

well by SM-MYOC and CM-MYOCD (<2-fold difference in induction). Since CALD1 has been detected in adult cardiac muscle of a variety of mammalian species, it should not be considered a very specific marker of SMCs²⁴.

In contrast, many cardiac muscle genes were much more potently transactivated by CM-MYOCD than by SM-MYOCD (Table 4). Several of the genes that exhibited the largest difference in expression levels between CM-MYOCD- and SM-MYOCD- transduced CMPCs were formerly shown to be regulated by MEF2 proteins including cardiac troponin C (TNNC1;²⁵), cardiomyopathy-associated protein 1 (CMYA1;²⁶), myomesin 1 (MYOM1;²⁷) and muscle creatine kinase (CKM;²⁸). Among the genes more potently activated by CM-MYOCD than by SM-MYOCD were also (1) MYH6 and NPPA, whose transcripts could not be detected in trMYOCD-transduced 10T1/2 cells by RT-PCR¹⁴ and (2) MYL1 and SMYD1, whose promoters could be synergistically activated by MEF2C and the CM-enriched isoform of murine myocardin¹³. The fact that CM-MYOCD induces an extensive repertoire of cardiac muscle genes underlines that myocardin not only plays an important role in smooth muscle gene activation but also is a key regulator of cardiac gene expression.

The cellular environment influences mycoardin’s ability to transactivate genes

The transcriptional program induced by myocardin has been previously investigated in the context of human skeletal myoblasts (SKMs) by Teg Pipes and co-workers⁹. To force myocardin expression in these cells, an Ad5 vector encoding trMYOCD was used. Although there was a clear overlap between the human SKMs and the CMPCs in the gene transcription profiles induced by the smooth muscle enriched isoform of myocardin, the induction factors achieved in the myoblast study were generally much lower than those obtained in our experiments. There are several possible explanations for this apparent discrepancy.

(i) Some myocardin target genes may already be highly expressed in human SKMs but not in CMPCs. Myocardin-dependent upregulation of such genes in human SKMs

FHL2 3.1 5.22E-06 four-and-a-half LIM domains 2

LMO7 2.8 5.19E-06 LIM domain 7

KCNJ8 2.6 1.32E-08 potassium inwardly-rectifying channel, subfamily J, member 8 DLX5 2.6 4.63E-07 distal-less homeobox 5

MYH7B 2.5 1.16E-04 myosin, heavy polypeptide 7B, cardiac muscle, beta ATP2B4 2.5 7.96E-09 ATPase, Ca++ transporting, plasma membrane 4 MYO18A 2.4 1.37E-06 myosin XVIIIA

TMOD1 2.4 7.83E-09 tropomodulin

CACHD1 2.4 3.95E-06 cache domain containing 1 TNNI3K 2.4 1.38E-07 troponin I 3 interacting kinase

SLC8A3 2.3 1.90E-07 solute carrier family 8 (sodium-calcium exchanger), member 3

TNMD 2.3 1.20E-06 tenomodulin

MYCBP 2.3 8.15E-06 c-myc binding protein

MASTL 2.3 2.47E-07 microtubule associated serine/threonine kinase-like MYL2 2.3 8.05E-06 myosin, light polypeptide 2

SLC8A1 2.2 3.81E-04 solute carrier family 8 (sodium/calcium exchanger), member 1 KCNMA1 2.2 2.97E-06 potassium large conductance calcium-activated channel,

subfamily M, alpha member 1

KCNAB1 2.1 2.31E-07 potassium voltage-gated channel, shaker-related subfamily, beta member 1

ADAMTS1 2.0 1.74E-07 ADAM metallopeptidase with thrombospondin type 1 motif, 1

166

would thus only be possible to a limited extent. This situation may for instance apply to ACTC (1.2- vs. 1462-fold upregulation), calsequestrin 2 (CASQ2; 2.1- vs.

68-fold upregulation) and TNNT2 (1.3- vs. 56-fold upregulation; compare Tables S1 and S3 with Table 1 in⁹). However, NPPA is hardly expressed in unstimulated human SKMs⁹ or CMPCs (Figure 1), yet upon forced myocardin expression NPPA is induced much more potently in the latter cell type. This suggests that the epigenetic makeup, microRNA repertoire and/or TF content of a cell may affect the transactivating activity of myocardin. Cardiac, skeletal and smooth muscle cells utilize many related proteins to fulfill their specific functions and undergo

Table 5. Fold difference in the expression levels of cell cycle-related genes between CM- MYOCD- and SM-MYOCD-transduced CMPCs

transcript fold change P Full name

CDCA3 8.4 6.06E-10 cell division cycle associated 3 KIF20A 8.2 1.73E-12 kinesin family member 20A CDCA2 6.8 1.06E-11 cell division cycle associated 2

CDKN3 6.1 3.14E-11 cyclin-dependent kinase inhibitor 3 (CDK2-associated dual specificity phosphatase)

KIF15 6.0 2.67E-08 kinesin family member 15 KIF11 5.9 2.90E-09 kinesin family member 11

CDC20 5.9 4.20E-13 cell division cycle 20 homolog (S. cerevisiae) NEIL3 5.8 1.29E-06 Nei endonuclease VIII-like 3 (E. coli)

CCNA2 5.8 1.13E-10 cyclin A2

ASPM 5.8 8.24E-10 asp (abnormal spindle) homolog, microcephaly associated (Drosophila)

CDC45L 5.7 2.95E-08 CDC45 cell division cycle 45-like (S. cerevisiae) HCAP-G 5.7 5.37E-09 non-SMC condensin I complex, subunit G

BUB1B 5.5 6.40E-11 BUB1 budding uninhibited by benzimidazoles 1 homolog beta (yeast)

TOP2A 5.4 5.05E-12 topoisomerase (DNA) II alpha 170kDa PRC1 5.3 3.85E-12 protein regulator of cytokinesis 1

CCNB1 5.3 2.24E-08 cyclin B1

AURKB 5.3 1.06E-11 aurora kinase B

CCNB2 5.2 5.46E-11 cyclin B2

TTK 5.2 3.43E-08 TTK protein kinase

MAD2L1 5.1 6.42E-08 MAD2 mitotic arrest deficient-like 1 (yeast) CDC2 5.1 5.31E-06 cell division cycle 2, G1 to S and G2 to M UBE2C 4.9 1.26E-10 ubiquitin-conjugating enzyme E2C

PKMYT1 4.9 6.27E-09 protein kinase, membrane associated tyrosine/threonine 1 MYH3 4.9 3.12E-10 myosin, heavy chain 3, skeletal muscle, embryonic STK6 4.8 1.28E-12 aurora kinase A

EXO1 4.7 1.36E-09 exonuclease 1

CDCA1 4.6 9.48E-06 NUF2, NDC80 kinetochore complex component, homolog (S.

cerevisiae) FOXM1 4.6 1.89E-08 forkhead box M1 KIF4A 4.6 1.19E-10 kinesin family member 4A RAD54L 4.5 1.08E-08 RAD54-like (S. cerevisiae)

Pfs2 4.4 3.31E-11 GINS complex subunit 2 (Psf2 homolog)

MCM7 4.2 3.61E-08 MCM7 minichromosome maintenance deficient 7 (S.

cerevisiae)

TPX2 4.2 3.63E-10 TPX2, microtubule-associated, homolog (Xenopus laevis) SGOL2 4.2 2.16E-08 shugoshin-like 2 (S. pombe)

POLQ 4.2 2.32E-10 polymerase (DNA directed) kappa RAD51AP1 4.1 3.00E-08 RAD51 associated protein 1 KNTC2 4.1 2.47E-06 kinetochore associated 2

LOH3CR2A 4.1 1.84E-07 loss of heterozygosity, 3, chromosomal region 2 BUB1 3.9 1.45E-09 BUB1 budding uninhibited by benzimidazoles 1 homolog

(yeast)