TAK1/Map3k7 enhances differentiation of cardiogenic endoderm from mouse embryonic stem cells.

Abstract

Specification of the primary heart field in mouse embryos requires signaling from the anterior visceral endoderm (AVE). The nature of these signals is not known. We hypothesized that the TGFβ-activated kinase (TAK1/Map3k7) may act as a cardiogenic factor, based on its expression in heart-inducing endoderm and its requirement for cardiac differentiation of p19 cells. To test this, mouse embryonic stem (ES) cells overexpressing Map3k7 were isolated and differentiated as embryoid bodies (EBs). Map3k7-overexpressing EBs showed increased expression of AVE markers but interestingly, showed little effect on mesoderm formation and had no impact on overall cardiomyocyte formation. To test whether the pronounced expansion of endoderm masks an expansion of cardiac lineages, chimeric EBs were made consisting of Map3k7-overexpressing ES and wild type ES cells harboring a cardiac reporter transgene, MHCα::GFP, allowing cardiac differentiation to be assessed specifically in wild type ES cells. Wild type ES cells co-cultured with Map3k7-overexpressing cells had a 4-fold increase in expression of the cardiac reporter, supporting the hypothesis that Map3k7 increases the formation of cardiogenic endoderm. To further examine the role of Map3k7 in early lineage specification, other endodermal markers were examined. Interestingly, markers that are expressed in both the VE and later in gut development were expanded, whereas transcripts that specifically mark the early definitive (streak-derived) endoderm (DE) were not. To determine if Map3k7 is necessary for endoderm differentiation, EBs were grown in the presence of the Map3k7 specific inhibitor 5Z-7-oxozeaenol. Endoderm differentiation was dramatically decreased in these cells. Western blot analysis showed that known downstream targets of Map3k7 (Jnk, Nemo-like kinase (NLK) and p38 MAPK) were all inhibited. By contrast, transcripts for another TGFβ target, Sonic Hedgehog (Shh) were markedly upregulated, as were transcripts for Gli2 (but not Gli1 and Gli3). Together these data support the hypothesis that Map3k7 governs the formation, or proliferation of cardiogenic endoderm.

1. Introduction:

It is well established that the earliest stages of heart formation in mouse [1–4], chick [5,6], and amphibian [7–9] embryos, depends on signals from the adjacent endoderm. In the mouse, this is referred to as the anterior visceral endoderm (AVE). Similarly, cardiac differentiation is enhanced when differentiating mouse and human pluripotent cells are co-cultured with AVE-like cell lines [10–12]. Despite the acknowledged importance of the AVE, the nature of its cardiogenic signals is poorly understood.

We previously identified Map3k7 as a potential cardiogenic factor expressed in AVE-like cell lines [10]. Map3k7 is a member of the Mitogen-Activated Kinase Kinase Kinase family of signaling molecules. Signaling through this pathway is essential for cardiac differentiation in p19 cells [13, 14], and mice lacking its expression die during embryogenesis with severe heart and gut defects [15]. Mice with cardiac-specific expression of a dominant interfering form of Map3k7 die shortly after birth due to conduction system abnormalities [16]. In addition, we recently showed that myocardial differentiation within Map3k7-overexpressing EBs is directed toward the sinoatrial node (SAN) lineage [17]. Taken together, these data suggest a role for Map3k7 in both the establishment and patterning of heart tissue, but its mechanism of action is unclear. We hypothesized that Map3k7 may function in the endoderm to mediate its heart-inducing action, either by expanding the AVE, or by increasing signaling from the AVE. To examine this, a lentivirus in which Map3k7 and green fluorescent protein (GFP) are driven under the control of the ubiquitous human promoter, hPGK (Sin18 hPGK::Map3k7-IRES2-GFP), was stably transduced into mouse ES cells [17].

Here we report that Map3k7 overexpression upregulates transcription of both general endodermal markers and AVE-specific markers. Interestingly, transcripts that specifically mark the early, streak-derived endoderm were unaffected by Map3k7 overexpression. In addition, transcripts for the early mesoderm markers, T/Brachyury and FGF8, and the early neural marker Sox2, were only slightly increased by Map3k7 overexpression.

Despite, the expansion of markers for the cardiogenic AVE, overall cardiac differentiation, as assessed by the percentage of cells expressing the cardiac epitope MF20, was not enhanced in Map3k7-overexpressing EBs, as compared to the parent ES cell line (R1). However, given the expansion of endoderm markers in these EBs, it is possible that expansion of cardiac cells is obscured by increases in other cell types. To address this possibility, Map3k7-overexpressing ES cells were co-cultured with wild type ES cells harboring a transgene in which green fluorescent protein (GFP) is driven by the cardiac-specific promote, MHCα (MHCα::GFP). This allows cardiac differentiation to be assessed specifically in the wild type population. Wild type cells that were co-cultured in EBs with Map3k7-overexpressing cells exhibited a 4-fold increase in the number of GFP(+) cells that formed as compared to those co-cultured with control cells.

We had previously showed that Map3k7 activates a negative feed back loop downregulating its own expression. To understand the nature of this feedback mechanism better, several known downstream targets of Map3k7 (NLK, p38 and Jnk), were analyzed by western blot. All three of these key downstream components were downregulated. By contrast, transcription of Shh and Gli2 was dramatically upregulated in Map3k7-overexpressing cells.

We had previously reported that cardiac differentiation in Map3k7-overexpressing cells was biased to the SAN, however in co-culture experiments cardiac cells derived from wild type ES cells did not beat rapidly, suggesting that Map3k7 acts cell-autonomously in cardiac cells, directing them to the SAN fate, but non-cell-autonomously in the endoderm to enhance overall cardiac differentiation.

2. Materials and Methods:

2.1. Cell Culture

Mouse (R1 or CGR8) ES cell lines (ATCC) comprise the parent cell lines for these studies. To make cardiac reporter lines, cells were transduced with the cardiac promoter reporter, MHCα::GFP. This reporter was previously described, and rigorously tested for cardiac specific expression [10]. To make an early endodermal reporter cells were transduced with Afp::GFP (A gift from Kat Hadjantonakis). These were previously validated [10, 18]. Map3k7-overexpressing cell lines were transduced with the overexpression vector, hPGK::Map3k7-IRES-GFP. Clonal lines were established and checked for constitutive overexpression of Map3k7 by qRT-PCR both in undifferentiated ES cells and in differentiating EBs [17]. All lines were maintained in ES cell growth medium (10% ES-qualified FBS, 2mM glutamine, 0.1mM non-essential amino acids, 1mM sodium pyruvate, 10⁻⁴ mM β-mercaptoethanol, 1x pen/strep, and 1000 units LIF). For differentiation studies, ES cells were passaged off of mouse embryonic fibroblasts (MEFs) and differentiated as EBs using the hanging-drop method, as previously described [10]. 5z-7-oxozeaenol was added from day 1 to day 4 at a concentration of 1μM diluted in 0.1% DMSO.

2.2. Construction of the hPGK::Map3k7 expression vector

The open reading frame of mouse Map3k7 was amplified by PCR from the pRK5mMap3k7-WT-flag vector (a gift from Hiroshi Shibuya) and directionally cloned into the Sin18-pre.hPGK-IRES2.eGFP.PB vector (a gift from Mark Mercola), which drives expression of both the inserted gene and GFP from the ubiquitous human promoter, hPGK. Virus was produced using a second-generation lentiviral expression system [19], as described in Bajpai and Terskikh [20].

2.3. Real-time Quantitative Reverse-Transcription PCR

EBs were grown 25 EBs to a plate, on 60 mM plates that were gelatin-coated to encourage attachment. EBs were collected on specified days of differentiation. Total RNA was extracted with RNeasy Mini Kit (Qiagen), and 1 μg was used to synthesize first strand cDNA, using the QuantiTect Reverse Transcription Kit (Qiagen). Quantitative PCRs were performed with SybrGreen Master Mix (Roche), using 40 ng template/reaction on a Roche LightCycler® 480 Real-Time PCR Instrument, and analyzed with the LightCycler 480 software package (version 1.5.0). Crossing-point data were first adjusted to reflect the efficiency of primer pairs by comparison to standard curves established by dilution series for positive control cDNAs. Subsequently these data were normalized to the crossing-point data for the ubiquitously expressed transcript GAPDH. Data represent averages ± SEM of 3 independent experiments. Further analysis was carried out using GraphPad Prism (version 7.0).

Primers used in this study are as follows:

Primer	Forward	Reverse
Gapdh	AAT GGA TAC GGC TAC AGC	GTG CAG CGA ACT TTA TTG
mAfp	AGCTGACAACAAGGAGGAGTG	TTAATAATGGTTGTTGCCTGGA
mT-Brachyury	AGC TTC GTG ACG GCT GAC AA	CGA GTC TGG GTG GAT GTA G
mCXCR4	CCGCCTTTACCCCGATAGC	ACCCCCAAAAGGATGAAGGAG
mFGF8	GCTCATTGTGGAGACCGATAC	TTGCTCTTGGCAATTAGCTTC
mGata4	CAT CAA ATC GCA GCC T	AAG CAA GCT AGA GTC CT
mGPC1	TGGTGCTCATCACTGACAAGTTC	GGATGACCTTAGCTGTGAGTGTGT
mHNF4	CGA ACA GAT CCA GTT CAT CAA G	ATG TGT TCT TGC ATC AGG TGA G
mTm4sf2	CCAGTTGCTGCATGAACGAA	CACCAGATCATAACAGCCCTTCT
mTtr	CTGGGATTGGGTGTGTCAG	CACCGTATTGGCAAGCAG
mSox17	GCTAGGCAAGTCTTGGAAGG	CTTGTAGTTGGGGTGGTCCT
mSox2	GAGGGCTGGACTGCGAACT	TTTGCACCCCTCCCAATTC
mDkkl	ATCTGTCTGGCTTGCCGAAAGC	GAGGAAAATGGCTGTGGTCAGAG
mHhex	CGGTCAAGTGAGGTTCTCCAAC	CTCGGCGATTCTGAAACCAGGT

2.4. Western blot analysis

EBs were grown as described above on gelatin-coated plates to promote attachment. They were harvested by scraping cells from plates after 2, 5, 9, 12 and 16 days of differentiation. This is done to preserve surface proteins that might be disrupted by enzymatic isolation of cells. After centrifugation, EB pellets were stored at −80 °C. EB pellets were resuspended in lysis buffer (150 mM NaCl, 2% SDS, 1% Triton X−100, 1% sodium deoxycholate, 40 mM NaF, 50 mM Tris-HCl, pH 7.4), mechanically disrupted with a mini-pestle (Argos), heated 10 minutes at 95 °C, sheared through a 26 gauge needle and centrifuged 10 minutes at 10,000 x g. Supernatants were collected and total protein was measured using the BCA method (Thermo Scientific). Proteins were resolved by SDS-PAGE (NuPAGE 12% Bis-Tris gels; Invitrogen) and transferred to PVDF membranes. Membranes were immune blotted with the following primary antibodies: total JNK (Cell Signal Technology 9252), JNK2 (CST 9258), JNK3 (CST 2305), phospho-JNK (CST 4671), p38 MAPK (CST 9212), phosph-p38 MAPK (CST 9211), NLK (CST 94350), all used at 1:1000. Immunoblots were visualized by chemiluminescence (Millipore) and digital images were captured with a GeneGnome imager (Syngene). Densitometry analysis was performed in ImageJ (NIH link). Statistical analysis was performed in Prism 7 (GraphPad) using multiple unpaired two-tailed t-test, and the Holm-Sidak method to correct for multiple comparisons. P values < 0.05 were considered significant.

2.5. Flow Cytometry

Wild type EBs containing the MHCα::GFP reporter and Map3k7-overexpressing ES cells (and combinations of these) were grown for 15–17 days. EBs were collected and dissociated into single cell suspensions using 1 mg/ml Collagenase D in DMEM. Cells were resuspended in 1X PBS and filtered through an 80μm sieve. Cells were then centrifuged and resuspended in 1X PBS containing 0.5% BSA and 2 units/ml DNAseI (Invitrogen). Flow cytometry was performed with a Becton-Dickinson (B-D) FACScan, and data were acquired using the B-D CellQuest software. Dot plots were created with FL2-H (red fluorescence) and FL1-H (green fluorescence) on the x- and y-axes respectively, and compensation control samples were prepared from EBs made of the parent cell lines (R1 or CGR8) that were grown and prepared under identical conditions but that did not harbor a lineage specific promoter reporter. These compensation controls were used to establish background fluorescence in differentiating EBs. Gates for these compensation controls were set at 0% GFP(+) cells for controls and 0.1% for Map3k7-overexpressing cell lines. Data are represented as fold difference compared to untreated wild type EBs. Bars represent the average difference between wild type and Map3k7-overexpressing cells over four separate trials. Error bars represent standard error. Statistical significance was determined by unpaired, two-tailed t-test. In figure 1, Map3k7-overexpressing cells did not possess the MHCα::GFP promoter. Therefore cardiac cells were identified by expression of the MF20 epitope. Cells were washed with 1X PBS and fixed with 4% PFA at RT for 15 minutes. Permeabilization was carried out for 10 minutes in 1X PBS consisting of 0.5% BSA, and 0.1% Triton X-100 (permeabilization buffer). After permeabilization, cells were incubated for 1 hour in permeabilization buffer containing mouse MF20 antibody (Developmental Studies Hybridoma Bank) diluted at 1:100. Compensation control samples were grown and prepared under identical conditions, without the addition of MF20. Cells were then washed with permeabilization buffer and incubated for 1 hour with a FITC-conjugated goat anti-mouse IgG secondary antibody (abcam) diluted at 1:1000. Analysis from this point was identical to samples containing the GFP reporter.

Figure 1.

Cartoon of the Map3k7 overexpression construct used in these studies. B. Immunocytochemistry showing near absence of Tak1/Map3k7 protein in the parent cell line R1 but strong, ubiquitous expression in the Tak1/Map3k7 clonal overexpression cell line. No fluorescence was observed in a no-primary antibody control. All images were taken at the same exposure and at the same magnification (TAK1/Map3k7, red; DAPI, blue). C. Quantitative Real-Time PCR data showing a general up regulation of the general endoderm markers Afp, Gata4, Hnf4, and Ttr, in Map3k7-overexpressing EBs (red) as compared to control EBs (blue line). Error bars represent standard error.

2.6. Immunocytochemistry

ES cells were dissociated by trypsinization and replated on gelatin-coated slide wells. After 24 hours, cells were fixed for 15 minutes in 4% paraformaldehyde, followed by 3 washes in PBS. Cells were then permeablized in blocking buffer (1% FBS, 0.1% BSA and 0.1% Triton-X-100). After permeablization cells were washed 3 times in PBS then incubated 1 hour at room temperature in permeablization buffer containing the TAK1 antibody (Sigma) at a dilution of 1:200. Cells were then washed 3 times in PBS and the incubated in secondary antibody, Alexa-594-conjugated anti-rabbit IgG (Invitrogen) at a dilution of 1:1,000).

3. Results:

3.1. Map3k7 overexpression causes a marked increase in endoderm formation in mouse ES cells.

To examine the role of Map3k7 in development, ES cell lines were transduced with a lentivirus in which Map3k7 is driven under control of the human PGK promoter (Fig. 1A). We previously demonstrated that clonal lines exhibit a continuous 2-fold up regulation of Map3k7 transcripts during EB differentiation [17]. To confirm this, and to demonstrate that transcripts are translated into protein we also assessed Map3k7 expression by immunocytochemistry. Cells were trypsinized and re-plated in slide wells. Cells were then immunostained using an anti-TAK1 antibody and an Alexa-conjugated secondary antibody. Cells were counterstained with DAPI. All images were taken at the same exposure. Map3k7 protein was expressed at low levels in undifferentiated R1 mouse ES cells but was strongly, and ubiquitously, expressed in the Map3k7-overexpressing ES cells. Since expression in the R1 cells was low, a no-primary control was added to insure that the low level expression observed was not due to background fluorescence. No red fluorescence was observed in a no-primary control (Fig. 1B).

We previously showed that Map3k7 was expressed in cardiogenic endoderm [21]. To determine if it plays a role in establishing the cardiogenic endoderm, Map3k7-overexpressing EBs were analyzed by qRT-PCR for expression of various endodermal markers. Map3k7-overexpressing ES cells were differentiated as EBs using the hanging drop method and groups of 25 EBs collected at various time points during EB differentiation for up to 16 days. Total RNA was isolated at each time point and analyzed by qRT-PCR. This was compared to qRT-PCR data from control EBs. Transcripts for general endodermal markers, Transthyretin (Ttr), α-feto-protein (Afp), Hnf4 and Gata4 [18, 22], were all upregulated in the Map3k7-overexpressing cells as compared to control EBs (Fig. 1C).

3.2. Map3k7 specifically upregulates the expression of markers for the cardiogenic endoderm.

In mouse embryos, Ttr, Afp, Hnf4, and Gata4, are expressed in both the VE, and the gut endoderm [24, 25]. VE arises from the primitive endoderm and surrounds the embryo prior to gastrulation, whereas DE is comprised of cells from both the VE and DE, which migrates out of the primitive streak during gastrulation [23] To clarify if the increase in these endodermal markers is the result of expansion of the VE or DE, transcripts specific to each of these lineages were assessed by qRT-PCR (Fig. 2A). First, markers for the AVE were assessed, as this is the source of cardiogenic signals in the embryo. Transcripts for both Cerberus and hHex were significantly upregulated, suggesting that a Map3k7-overexpressing EBs do show a general expansion of AVE. Another AVE marker, Dkk1, was not upregulated, suggesting that its expression in the AVE may be controlled by another pathway. Sox17, a marker for all early endodermal lineages, including the VE and DE was also upregulated but only transiently. Transcripts that are specifically expressed in the DE, but not in the VE, including mCXCR4, mTM4sf2, and GPC, were not changed in Map3k7-overexpressing EBs (Figure 2A). Finally, we assessed the early neural marker Sox2. Its transcription was also not significantly altered as compared to controls (Fig. 2A). From this we conclude that Map3k7 dramatically and consistently causes an increase in markers for the VE and AVE, but does not effect the formation of streak-derived DE.

Figure 2:

A. Quantitative Real-Time PCR data showing transient upregulation of transcripts for the AVE markers Cerberus, and hHex, as well as the pan-endodermal marker Sox17. By contrast transcripts for the AVE marker, Dkk1 were not upregulated. In addition, neither the definitive endoderm markers Tm4sf2, GPC1 and CXCR4, nor the early neural marker Sox2, were upregulated in Map3k7-overexpressing EBs (red) as compared to control EBs (green line). Error bars represent standard error. B. EBs harboring an endoderm specific promoter reporter AFP::GFP were grown in the presence of the Map3k7 inhibitor 5z-7-oxozeaenol from day 1–4 of EB differentiation and photographed at day 10 showed a significant decrease in endoderm formation as assessed by the expression of GFP (+) cells, as compared DMSO-treated controls.

To confirm the role Map3k7 on endoderm formation, we differentiated EBs harboring an endoderm specific promoter reporter, Afp::GFP [24]. These were plated 1 EB per well in ten wells of a 48-well plate. Five wells were treated with the TAK1/Map3k7 specific inhibitor 5z-7-oxozeaenol (OXO)(1μM in DMSO) and an additional five wells treated with DMSO only. EBs were treated with the inhibitor from day 1 to day 4 of EB differentiation and assessed from day 5 and day 10 for expression of the endoderm reporter. At day 5 all of the control EBs showed expression of the AFP::GFP reporter but none of the EBs grown in the presence of OXO had expression of the endodermal reporter. From day 6–9, control EBs showed robust expansion of cells expressing the GFP reporter but no expression was observed in the OXO treated EBS. Finally at day 10, 6 days after cells were released from OXO treatment, small patches of consisting of a few cells expressing the AFP::GFP reporter were seen in 3 out of the 5 EBs (60%) that had been previously been treated with OXO (Fig. 2B). These data show that blocking Map3k7 does not kill cells but rather blocks the differentiation of endoderm.

3.3. Map3k7 overexpression during EB differentiation of mouse ES cells does not affect early mesoderm formation

Given that Map3k7 seems to specifically expand the VE and AVE in the endoderm it was hypothesized that Map3k7-overexpressing EBs might also show an expansion of the cardiac mesoderm. We used qRT-PCR on differentiated EBs to determine if Map3k7 impacted the transcription of early mesodermal markers T/brachyury (T/Bra) and FGF8 (Fig. 3A). There was a small but statistically significant difference between Map3k7-overexpressing EBs and controls, suggesting that mesoderm formation in these cells may be slightly increased.

Figure 3:

A. Representative qRT-PCR data comparing relative transcript expression for early mesodermal markers T/Bra and Fgf8 in Map3k7-overexpressing EBs (red) as compared to wild type EBs (blue). Error bars indicate standard error from three technical replicates. B. Summary of flow cytometry data comparing cardiac differentiation, as assessed by expression of the MF20 epitope, in wild type (blue) and Map3k7-overexpressing EBs at day 17 of differentiation. Error bars represent standard deviation.

Cardiac formation on day 17 was assessed in these EBs by flow cytometry. EBs were dissociated and immunostained for expression of the MF20 epitope. MF20 marks both skeletal and cardiac muscle, however, we previously showed that in our culture system that there was near perfect overlap between MF20 and other cardiac markers [10], suggesting that EBs differentiated in this manner make little or no skeletal muscle. Cells were analyzed for expression of the MF20 epitope by flow cytometry. Interestingly, there was no statistically significant difference between the percentages of MF20 (+) cells that formed in the Map3k7-overexpressing EBs, as compared to the parent (R1) cell line (Fig. 3B).

3.4. Map3k7 overexpression increases cardiomyocyte differentiation via a non-cell-autonomous mechanism

Map3k7 appears to expand the AVE but does not increase the number of cells expressing the cardiac marker MF20. This is surprising because the AVE is the site of cardiogenic signaling in the early embryo. We hypothesized that expansion of endodermal cell types might mask an expansion of cardiac lineages. To test this, EBs were made that combined Map3k7-overexpressing ES cells with a wild type responder ES cell line. Responder cells, which are otherwise wild type, express a fluorescent cardiac reporter (MHCα::GFP). In this combination, cardiac differentiation can be assessed in the wild type cells specifically. EBs were grown until day 16 and cardiac differentiation assessed by flow cytometry (Fig. 4).

Figure 4:

Examples of flow cytometry analysis of cardiomyoctye differentiation, as assessed by expression of the MHCα::GFP promoter reporter on day 16 of EB differentiation as follows: A. ES cells (R1) that do not possess a cardiac reporter. B. ES cells (CGR8; MHCα::GFP) that possess the MHCα::GFP reporter. C. R1 ES (R1; PGK::Map3k7) cells transduced with the Map3k7 overexpression construct, but no cardiac reporter. D. EBs consisting of equal numbers of R1 and CGR8::MHCα::GFP ES cells. E. EBs consisting of equal number of unlabeled CGR8 and CGR8; MHCα::GFP ES cells and F. EBs consisting of equal number of R1; PGK::Map3k7 and CGR8; MHCα:: GFP ES cells. G. Summary of cardiomyocyte differentiation under the conditions described above. Error bars represent standard error from 4 separate experiments. H. Schematic of experimental design.

To determine background fluorescence in EBs, two different ES cell lines (R1 and CGR8) were differentiated as EBs and examined for autofluorescence (FL2-H) versus GFP expression (FL1-H). A gate above autofluorescence was set that contained 0% GFP (+) cells in the control (Fig. 4A–B). Background fluorescence for the Map3k7-overexpressing EBs was determined separately (Fig. 4C). Since these had higher background fluorescence than the control cell lines, a less stringent gate for fluorescence was used (0.1% GFP(+) cells).

To determine a base line of cardiac marker expression in wild type EBs, unlabeled ES cells of the R1 and CGR8 lineages were combined in a 1:1 ratio with responder cells (Fig. 4D–E). The % of GFP positive cells in the responder population was statistically the same regardless of whether they were co-cultured with R1 or CGR8 cells (Fig. 4G). These mixing experiments are schematically represented (Fig. 4H).

By contrast, when responders were co-cultured with Map3k7-overexpressing cells there was an approximately 4-fold increase in the number of GFP(+) cells (Fig. 4F–G)(p=0.0038 for CGR8 cells and <0.0006 for R1 cells). This suggests that Map3k7-overexpressing cells possess an enhanced ability to direct myocardial differentiation in neighboring cells, and do so by a non-cell-autonomous mechanism.

3.5. Map3k7 does not increase the beat frequency of wild type ES cells in co-culture experiments.

We previously reported that cells overexpressing Map3k7 show myocardial lineage differentiation similar to the SAN [17]. In those studies it is not clear if Map3k7 is required in the cardiac cells themselves, or if it is required in another lineage to direct SAN differentiation. To further address this, the beating frequency of wild type cardiac cells in co-culture was compared to the beating frequency of Map3k7-overexpressing cells in these EBs. Wild type cells co-cultured with Map3k7-overexpressing ES cells did not adopt a rapid beat rate as compared to controls (Fig. 5). However Map3k7-overexpressing cardiomyocytes in these co-cultured EBs did show rapid beat rates consistent to those that we reported previously [17]. Since pacing in the heart comes from the most rapidly beating cells these data are consistent with the notion that Map3k7 expression is required cell-autonomously in cardiac cells that will become pacemakers.

Figure 5.

A. Summary of beat rate data from responder cells alone or in co-cultured with Map3k7-overexpressing cells. In addition, beat rate data from Map3k7 cells in co-culture. Beat rate data in responder cells was not significantly altered by co-culture with Map3k7-overexpressing cells (p = 0.47) but beat rate data for Map3k7 cells in co-culture was significantly different from both (p < 0.0001). Statistics were performed by ANOVA.

3.6. Map3k7 overexpression leads to suppression of JNK, p38 MAPK and NLK activities

We previously noted that constitutive overexpression of Map3k7 lead to decreased expression of its own transcription, possibly through a negative feedback loop. To understand this better, western blot analyses of three kinase effectors downstream of Map3k7 (JNK, p38 MAPK and NLK) (Fig. 6A) were carried out (Fig. 6b). R1 control and Map3k7-overexpressing EBs were differentiated and lysates collected at developmental days 2, 5, 9, 12 and 16 for western analysis. Quantitative densitometry (Fig. 6c) revealed that, relative to R1 controls, Map3k7 overexpression suppresses levels of 52 kDa isoforms of JNK starting at day 9, with a significant reduction observed by 16 (P =0.02). Further qualitative analysis using isoform-specific monoclonal antibodies showed that reductions in total JNK are accounted for, in part, by reductions of both JNK2 and JNK3 isoforms (Fig. 6b). Additionally, downregulation of the 52 kDa JNK isoforms was accompanied by reduced phosphorylation of residues (Thr183/Tyr185) required for full kinase activation (Fig. 6a and b). Unlike JNK, total p38 MAPK expression levels were unaltered in Map3k7-overexpressing EBs relative to R1 controls over the entire developmental time course (Fig. 6c). However, the amount of p38 MAPK activated by phosphorylation at Thr180/Tyr182 was significantly reduced in Map3k7-overexpressing EBs at days 5 (P = 0.047), 9 (P = 0.016), and 15 (P = 0.039), and trended downward at day 12 (P = 0.068, after adjusting for multiple comparisons) (Fig. 6b and c). Finally, NLK expression was found to be significantly downregulated in Map3k7-overexpressing EBs at days 5 (P = 0.036), 12 (P = 0.001) and 16 (P = 0.036), and likely at day 9 (with increased trials) (Fig. 6c). These data support a model in which constitutive Map3k7 overexpression suppresses the kinase activities of particular JNK isoforms, as well as p38 MAPK and NLK, either by downregulation of total protein levels (i.e. JNK, NLK) or by reduction in kinase activation by phosphorylation at the active site (i.e. p38 MAPK)

Figure 6:

Time course western blot analysis of JNK, p38 MAPK and NLK in R1 control versus Map3k7-overexpressing EBs. A. Diagram showing known upstream activators and downstream effectors of TAK1/Map3k7. B. Representative western blots probed with primary antibodies directed against the denoted antigens and detected by chemiluminescence. Equal amount of total protein per well was used as the loading control. Visible MW markers are 50 and 40 kDa. C. Quantitative analysis of western blots using densitometry. Band densities were calculated as a percentage of the total combined density for each experimental group. Time points are mean ± SD (n = 3).

3.7. Map3k7 overexpression leads to up regulation of Shh and Gli2

It has previously been observed that TGFbeta signaling regulates the expression of Sonic Hedgehog (Shh) during gastrulation; for example, in Hensen’s node of avian embryos [27]. To examine whether SMAD-independent TGFbeta signaling might be involved in regulation of Shh, we examined the expression of Shh and several Gli genes by qRT-PCR during EB differentiation. Shh and Gli2, but not Gli1 or Gli3 showed robust overexpression in EBs overexpressing Map3k7 (Fig. 7).

Figure 7:

Quantitative Real-Time PCR data showing a strong, transient up regulation of transcripts for Shh and Gli2. Gli1 showed a small but statistically significant upregulation in Map3k7-overexpressing EBs (red) as compared to control EBs (green line) and Gli3 was not affected. Error bars represent standard error from 3 replicates.

4. Discussion:

4.1. Map3k7 in normal development

TAK1/Map3k7 is a member of the MAP kinase kinase kinase family of signaling molecules. Expression of Map3k7 can be activated by growth factors [25], cytokines [26–28] and various types of chemical and physiological stress (Fig. 6a). Activation of Map3k7 has been shown to activate several downstream kinases including Nemo-like kinase (NLK) [29, 30], p38 [31], Jnk [32], and a pro-inflammatory pathway mediated by NFκB [33]. p38 activity appears to play a role in determining the initial specification of the heart field. For example, Fgf signaling through p38 regulates the initial expression of the cardiac progenitor marker Nkx2.5 [34], whereas strong inhibition of p38 in differentiating ES cells favors neural development, over cardiac differentiation [35]. In our previous work, we showed that forced overexpression of Map3k7 leads to a downregulation of the cardiac progenitor marker Nkx2.5 [17]. At first glance, this appears to be inconsistent with previously published data; however, it has also been shown that Map3k7 expression can activate a negative feedback loop mediated by p38 [36], which is consistent with our previous finding that forced overexpression of Map3k7 leads to a down regulation of endogenous Map3k7 transcripts [17]. This finding is consistent with our finding here, that continuous, forced overexpression of Map3k7 leads to downregulation of Map3k7 targets Jnk, phospho-p38, and NLK.

In addition to the downregulation of active p38, we also saw a marked downregulation of certain Jnk isoforms. This may explain the bias toward VE/AVE expansion, as it has previously been shown that Jnk deficient ES cells, upregulate the VE/gut marker, Hnf4 [35].

Map3k7 deficient mice die during embryogenesis with major defects to the heart and gut endoderm [15], suggesting that it plays important roles in these lineages and, as such, can impact signaling between them that is required for the initial specification, proliferation and patterning of the heart and gut. These data suggest that upregulation of endodermal markers might be mediated either directly by overexpression of Map3k7 or indirectly through inhibition of its known downstream targets.

4.2. Map3k7 has multiple roles in heart formation

To identify cardiac-inducing signals that are present in the endoderm, we performed a transcriptomic analysis comparing gene expression in three VE-like cell lines [24], each with demonstrated ability to enhance myocardial differentiation [10–12,37]. The top pathways expressed by these cell lines, that were not directly related to cell cycle, were the MAP Kinase and TGFβ signaling pathways. This was consistent with various earlier studies pointing to the ability of TGFβ superfamily members to activate myocardial differentiation through both SMAD-dependent and SMAD-independent signaling pathways. Specifically, we noted high expression levels of TAK1/Map3k7 transcripts could also be involved in the ability of endoderm to activate myocardial differentiation. To test whether Map3k7 activity might constitute a portion of the endoderm’s cardiogenic signal, ES cells overexpressing Map3k7 were co-cultured with wild type ES cells expressing a cardiac promoter reporter. In these studies, the number of wild type cells expressing the cardiac reporter increased 4-fold. Interestingly, while the Map3k7-overexpressing cells in these studies showed rapid beat rates, consistent with our previous studies that Map3k7-overexpressing cardiac cells differentiate as SAN cells [17], the wild type cells in co-culture did not show rapid beat rates. Since the pace in the heart is set by the fastest beating population the slower beating wild type cells cannot represent pacemaker cells in this setting. Together these data suggest that Map3k7 plays dual roles in cardiac differentiation, an indirect, non-cell-autonomous effect on primary cardiac specification resulting from increased endoderm differentiation, and a direct, cell-autonomous effect caused by Map3k7 overexpression in myocardial cells, resulting in their preferential differentiation toward SAN fates[17].

5. Conclusions

Our findings here, along with previous studies [10, 24], support a model in which Map3k7 indirectly activates early cardiomyocyte differentiation, by increasing the differentiation of cardiogenic endoderm. Since Map3k7 is a kinase, a further implication of this work is that enhanced differentiation of cardiogenic endoderm might be enhanced by the timely addition of small molecule antagonists of key kinases within the network.