p38α MAP kinase-deficient mouse embryonic stem cells can differentiate to endothelial cells, smooth muscle cells, and neurons



p38 MAP kinase α (p38α) regulates various cellular processes in adult cells, but little is known about its function in stem cells. We investigated the potential of wild type and p38α deficient mouse embryonic stem cells (ESCs) to differentiate into endothelial cells (ECs), smooth muscle cells (SMCs), and neurons. Our differentiation methods allowed simultaneous development of all these cell types. ECs formed monolayers similar to mature ECs and could assemble into vessel-like structures. SMCs had well-organized actin filaments with morphology similar to adult SMCs. Neurons exhibited well-developed cell bodies and elongated axons. Deletion of the p38α gene did not significantly compromise ESC differentiation since p38α−/− cells could express cell-specific markers and displayed similar overall morphology to the cells differentiated from p38α+/+ ESCs. Although p38α regulates various cellular activities of adult SMCs, ECs, and neurons, our data demonstrate that p38α is not essential for ESC differentiation to these cell types. Developmental Dynamics 236:3383–3392, 2007. © 2007 Wiley-Liss, Inc.


Mitogen-activated protein (MAP) kinases are widely expressed protein kinases that regulate a variety of cellular activities. They are activated by growth factors, cytokines, or environmental stresses. Activated MAP kinases transduce signals by phosphorylating their target enzymes or transcription factors, which in turn regulate subsequent cellular processes. Three distinct subtypes of MAP kinases have been identified. Extracellular signal-regulated kinases (ERK) are strongly activated by growth factors; thus they are critical for cell proliferation (Cobb et al.,1991). Two other subtypes, known as stress-activated protein kinases, are c-Jun N-terminal protein kinases (JNK) and p38 MAP kinases (p38) (Karin,1998). They are strongly activated by cytokines and cellular stresses (Guo et al.,1998; Kyriakis and Avruch,2001). Therefore, their activation generally inhibits cell growth or induces apoptosis (Kyriakis and Avruch,1996; Xia et al.,1995). However, further studies have extended physiological functions of MAP kinases far beyond those that were originally proposed (Nebreda and Porras,2000). Therefore, MAP kinases represent a group of signaling molecules that regulate a wide variety of cellular functions. On the other hand, their developmental roles in embryogenesis are not well understood (Cuenda and Rousseau,2007).

Four isoforms of the p38 subfamily have been identified in mammalian cells: p38α, p38β, p38γ, and p38δ. Depending on cell types and the nature of stimuli, different isoforms have redundant, specific, or even opposite functions (Nebreda and Porras,2000; Ono and Han,2000). p38α is broadly expressed and the most abundant p38 family member in most cell types. Although it was initially characterized in inflammation and stress responses, it is now recognized that many cellular activities regulated by p38α seem to be unrelated to stress or inflammation (Nebreda and Porras,2000). The pivotal role of p38α, in the regulation of growth and development is further demonstrated in animal models. p38α knockout (p38α−/−) mice had defective vessel structures in placentas and are embryonic lethal (Tamura et al.,2000; Mudgett et al.,2000; Allen et al.,2000). Furthermore, they displayed abnormal angiogenesis in the embryo proper, head region, and visceral yolk sac, indicating a general role for p38α signaling in embryonic angiogenesis (Mudgett et al.,2000). However, further studies revealed that the defective placenta resulted from p38α knockout could be rescued by tetraploid complementation (Adams et al.,2000); under this condition, the p38α−/− mouse embryos could develop to term with normal appearance. This finding led to the hypothesis that the embryos died mainly due to the defect in placental organogenesis. It would also suggest that p38α may not be essential for mouse development once the embryos are independent of the placenta for nutrient supply. This is surprising since it is difficult to envision that such a widely expressed kinase in various cells, including embryonic stem cells (ESCs) (Guo and Yang,2006), is dispensable for embryo development except for placental formation. The embryonic lethality limits further in-depth analysis of the developmental role of p38α at the cellular level in animal models. The generation of p38α−/− ESCs (Roach et al.,1995; Kim et al.,2005) provides a valuable alternative system. Our recent study showed that p38α −/− ESCs displayed altered cell adhesion to different extracellular matrix proteins (Guo and Yang,2006). Other investigators reported that p38α−/− ESCs were able to differentiate into myeloid-like cells, but that these cells failed to produce IL-6 in response to IL-1 stimulation, as opposed to cells derived from wild type cells (Allen et al.,2000). Similarly, p38α was not essential for lymphocyte development (Kim et al.,2005), but p38α −/− Th1 cells were defective in IFN-γ secretion stimulated by IL-12/IL-18 (Berenson et al.,2006). These studies demonstrate the potential of using p38α −/− ESCs as a genetically defined tool to study this enzyme in the regulation of ESC activities and to provide results that otherwise are not obtainable from studies using fully differentiated adult cells.

Numerous studies have shown that p38α is critical for cellular function of endothelial cells (ECs) (Rousseau et al.,1997) and smooth muscle cells (SMCs) (Gerthoffer,2005), the two major cell types in blood vessels. However, whether p38α plays a role in ESC differentiation to these cell lineages during embryogenesis is not known. In this study, we investigated the differentiation potential of normal and p38α deficient ESCs. Our results revealed that p38α−/− ESCs could differentiate into ECs, SMCs, and neurons, which showed no apparent difference in morphology from the cells derived from p38α+/+ ESCs. Together with the results reported by other investigators (Allen et al.,2000; Kim et al.,2005), our findings may suggest that p38α is critical in mediating various signals in differentiated cells but is not essential for the ESC differentiation process, at least in cell types that have been tested thus far.


Differentiation of Wild Type and p38α Deficient ESCs

Wild type and p38α mutant mouse ESCs generated by two laboratories were used in this study. p38α+/+ and p38α−/− ESCs described by Allen et al. (Roach et al.,1995; Allen et al.,2000) were designated as p38α +/+P (+/+P) and p38α −/−P (−/−P) (kindly provided by Dr. Christopher Gabel, Pfizer Central Research). p38α +/+ (clone 3), p38α+/− (clone 4), and p38α −/− (clone 1) ESCs described by Kim et al. (2005) were designated as p38α +/+A (+/+A), p38α +/−A (+/−A), and p38α −/−A (−/−A) (kindly provided by Dr. Barry P. Sleckman, Washington University School of Medicine at St. Louis), respectively. Generation and characterization of these cell lines have been described in detail in the given references. The genotypes of ESCs were first analyzed for the expression of p38α at both mRNA and protein level by RT-PCR (reverse transcription-polymerase chain reaction) and Western blot. As shown in Figure 1, p38α was abundantly expressed in wild type ESCs (p38α +/+P and 38α +/+A), and it was completely eliminated in p38α −/−P and p38α −/−A ESCs. As expected, the intermediate level of p38α was detected in heterozygous ESCs (p38α +/−A). The expression of β-actin, a ubiquitously expressed isoform of actin family, was unaffected by p38α deletion and used as a reference marker for the gene expression analysis.

Figure 1.

Expression of p38α in p38α+/+, p38α+/−, and p38α−/− ESCs. A: P ESC lines. Left panels, Western blot analysis of p38α expression. Cell lysates were analyzed with anti-p38α antibodies and anti-β-actin antibodies. Right panels: RT-PCR analysis of p38α expression. The mRNA levels of p38α and β-actin (Actin) were determined by RT-PCR products that were analyzed in an agarose gel stained with ethidium bromide. B: A ESC lines, analysis of p38α expression by Western blot as described in A. In both Western blot and RT-PCR analysis, expression of β-actin, which does not change in all genotypes of ESCs, was used as a control. The data are representative of the results from at least three independent experiments.

Mouse ESCs are routinely maintained as monolayers in cell culture dishes. They are kept in an undifferentiated state when cultured in the presence of leukemia inhibitory factor (LIF), a cytokine that inhibits ESC differentiation (Niwa et al.,1998). When deprived of cell adhesion and cultured in suspension, ESCs grow in aggregates and form structures called embryoid bodies (EBs). In the absence of LIF, EBs undergo spontaneous differentiation to different cell types (Wobus and Boheler,2005). Using these protocols, we generated EBs from wild type and mutant ESCs and differentiated them in the presence of 15% fetal bovine serum (FBS). As shown in Figure 2 (top), 5-day EBs derived from p38α +/+P and p38α−/−P ESCs displayed overall similar morphology. Further differentiation was promoted by either incubating EBs for an additional 5 days (10-day EBs) or allowing 5-day EBs to attach to a cell culture dish. In the latter case, cells within EBs grew out and developed into a monolayer (Fig. 2, middle and bottom). After further differentiation for 5 days, various types of cells with different morphology could be seen in EB-derived monolayers (Fig 2, bottom panels show enlarged areas of middle panels). Similar results were obtained using p38α +/+A, p38α +/−A, and p38α −/−A ESCs (data not shown). Ten-day EBs and EB-derived monolayers were used to identify differentiated cells and for further analysis.

Figure 2.

Embryoid body (EB) formation and ESC differentiation. A: Five-day EBs derived from p38α +/+P (+/+P) and p38α −/− (−/−P) ESCs were photographed with a phase contrast objective lens (CAch N 10×/0.25). Scale bars = 100 μm. B: Monolayers developed from 5-day EBs were stained with 1% toluidine blue and photographed with a phase contrast objective lens (CAch N 10×/0.25). Scale bars = 100 μm. C: Selected areas (around the boxed areas shown in B) were photographed with a phase contrast objective lens (CAch N 20×/0.4). Scale bars = 50 μm. The results are representative of the same experiments repeated at least three times.

Differentiation of ESCs to Endothelial Cells (ECs) and Vasculogenesis

To analyze the differentiation potential of ESCs to ECs and vascular development, we examined the expression of two EC makers, PECAM-1 and VE-cadherin, by using their specific antibodies that have been used to identify mouse ECs (Redick and Bautch,1999). The 10-day EBs were fixed and immunostained with anti-PECAM-1 antibodies. The primitive vascular structures were detected as intensive vascular networks as revealed by PECAM-1 staining (Fig. 3A, a). The formation of these structures remarkably resembled the process of vasculogenesis. The overall morphology and the vascular structures were similar among the EBs derived from p38α+/+A, p38α +/−A, and p38α −/−A ESCs, suggesting that p38α deletion may not significantly affect the early vasculogenesis. We further analyzed endothelial differentiation and vasculogenesis in EB-derived monolayers. PECAM-1 positive cells were arranged in two major types of structures: sheet-like structures similar to monolayers of mature ECs in culture (Fig. 3A, b) and vascular networks assembled from PECAM-1 positive cells with different lengths and diameters (Fig. 3A,c). Individual cells in the vascular structures underwent extensive remodeling. As seen at higher magnification, developing vessels contained empty spaces that appeared to be tubular structures (Fig. 3A, d, indicated by arrows). Immunostaining with anti-VE-cadherin antibodies revealed similar structures (Fig. 3A, e). The same experiments were performed using P ESC lines. Figure 3B shows the EC monolayers and the vessel-like structures differentiated from p38α +/+P and p38α −/−P ESCs.

Figure 3.

Differentiation of ESCs to ECs and vasculogenesis. Ten-day EBs or EB-derived monolayers were immunostained with anti-PECAM-1 (A, a–d; B, and C) antibodies or anti-VE-cadherin (A, e) antibodies followed by FITC-conjugated secondary antibodies. The samples were either analyzed with an Olympus fluorescence microscope (UplanFl 10×/0.3 lens or a UplanFl 100×/1.30 oil lens, A and B) or with a LSM 510 laser-scanning confocal microscope (Plan-Apochromat 63×/1.4 oil DIC lens, C). A: Experiments performed with A ESCs. a: Vascular networks detected in10-day EBs. b: Sheet-like EC structures in EB-derived monolayers. c: Vascular networks in EB-derived monolayers. d: Lumen structures (indicated by arrows) in EB-derived monolayers. e: Vessel structures in EB-derived monolayers identified by anti-VE-cadherin antibodies. Scale bars = 100 μm in a, c; 10 μm in b, d, e. B: Experiments performed with P ESCs. b': Sheet-like EC structures in EB-derived monolayers. d': Lumen structures (indicated by arrows) in EB-derived monolayers. b' and d' are equivalent structures of b and d shown in A, respectively. Scale bars = 10 μm. C: Confocal microscopic analysis of tubular structures formed in EB-derived monolayers from p38+/+A (+/+A) and p38−/−A (−/−A) ESCs. A 2-μm Z stack section for each sample is shown. Scale bars = 20 μm.

To further confirm that the vascular networks are indeed tubular structures, we further examined the EB-derived monolayers with a laser-scanning confocal microscope. Figure 3C illustrates two segments of vascular structures derived from p38α +/+A and p38α −/− A ESCs and detected with anti-PECAM-1 antibodies. In both cases, the inner spaces were continuous throughout most of the vascular structures, which could be more than several hundred micrometers. They were apparently formed from multiple cells and likely represented the primitive vessels (supplemental data). Similar structures derived from ESCs were observed and defined as vessels with lumens in a recent study by Zeng et al. (2007). Quantitative analysis of the vessel structures developed from wild type and p38α −/− ESCs was difficult due to heterogeneity in length, diameter, and morphology. Therefore, it is not certain at present if there are subtle differences between ECs and vascular structures developed from p38α +/+ and p38α −/− ESCs, but we do not see fundamental differences from the results shown in Figure 3.

Differentiation of ESCs to Smooth Muscle Cells (SMCs)

Like ECs, SMCs are among the early-differentiated cell types during embryogenesis (Wobus and Boheler,2005). We analyzed the expression of smooth muscle α-actin (SMA), a marker that is widely used to identity SMCs. In EB-derived monolayers that were immunostained with specific antibodies against SMA, the positive cells were scattered sparsely in many locations, but a large population was detected around the edges of the EB outgrowths (data not shown). Figure 4A and B shows a group of these cells differentiated from p38α +/+A and p38α −/−A ESCs. They were double stained with FITC-phalloidin to detect F-actin (green) and anti-SMA antibodies to detect SMA (red). The intense F-actin and SMA filaments were clearly identifiable in wild type cells (A+/+) as well as in p38α −/− cells (A−/−). They were arranged in similar patterns and were mostly overlapped as shown in merged images (yellow). The same experiments were performed with p38α +/+P and p38α −/−P ESCs. A single SMA-positive cell with intense SMA filaments derived from each type of ESCs is shown in Figure 4C. It should be noted that the size, morphology, and arrangement of actin filaments vary to some degree among SMA-positive cells. It is likely that these subtle differences represent different differentiation stages and/or physiological states of individual cells, rather than different genotypes since these variations could be seen among the cells derived either from p38α +/+A or from p38α −/−A ESCs (Fig. 4A and B).

Figure 4.

Identification of SMCs in EB-derived monolayers. A,B: EB-derived monolayers formed from p38+/+A (A) and p38α −/−A ESCs (B). The cells were labeled with: a: FITC-phalloidin to detect F-actin (green); b: anti-SMA antibodies to detect SMA (red); c: Hoescht 33258 to detect nuclei (blue). d: The merged image. The cells were examined and photographed with a LSM 510 confocal microscope (Plan-Apochromat 40×/1.3 oil DIC lens). Scale bars = 20 μm. C: Identification of SMCs in EB-derived monolayers formed from p38+/+P (+/+P) and p38α −/− (−/−P) ESCs. The cells were immunostained with anti-SMA antibodies followed by rhodamine-conjugated secondary antibodies. The cells were examined and photographed with an Olympus fluorescence microscope (UplanFI 100×/1.30 oil lens). Scale bars = 10 μm.

Differentiation of ESCs to Neurons

The results from the above experiments indicate that p38α may not be critical for ESC differentiation to ECs and SMCs, two cell types derived from mesoderm germ layer during embryonic development. We further tested the effect of p38α deletion on ESC differentiation to neurons, which develop from the ectoderm germ layer of the embryo. Neurons were identified in EB-derived monolayers by the expression of β-tubulin-3, which has been used as a neuron marker (Mujtaba et al.,1999). Immunostaining with an anti-β-tubulin-3 antibody revealed that some individual positive cells with neuron morphology were scattered throughout EB-derived monolayers, but numerous positive cells were detected in clusters (Fig. 5). At higher magnification, these cells displayed typical neuron morphology with long axons and expanded cell bodies. There were no apparent differences in morphology of neurons derived from the three genotypes of ESCs (Fig. 5A +/+A, +/−A, and −/−A, respectively). Similar results were obtained when the same experiment was performed with P ESC lines (Fig. 5B).

Figure 5.

Identification of neurons in EB-derived monolayers. A: EB-derived monolayers from p38α +/+A (+/+A), p38α +/−A (+/−A), and p38α −/−A (−/−A) were immunostained with anti-β-tubulin-3 antibodies and followed by FITC-conjugated secondary antibodies. The cells were examined and photographed with an Olympus fluorescence microscope (UplanFI 10×/0.13 lens or UplanFI 100×/1.30 oil lens). Scale bars = 100 μm (top panels), 10 μm (bottom panels). B: EB-derived monolayers from p38α +/+P (+/+P) and p38α −/−P (−/−P) ESCs were immunostained with anti-β-tubulin-3 antibodies and followed by rhodamine-conjugated secondary antibodies. The cells were examined and photographed with an Olympus fluorescence microscope (UplanFl 100×/1.30 oil lens). Scale bars = 10 μm.

Effects of p38 MAP Kinase Inhibitor SB203580 on ESC Differentiation and p38 Activation

The above experiments suggest that p38α deletion did not affect ESC differentiation to the three cell types that have been tested. To further confirm this conclusion, the differentiation process was carried out with wild type ESCs in the presence or absence of SB203580 (SB), a pyridinyl imidazole derivative widely used as a selective inhibitor of p38α and p38β (Gum et al.,1998). As shown in Figure 6A, SB treatment did not result in detectable morphological changes in ECs, SMCs, or neurons in comparison with the control cells (-SB), which is similar to the results obtained from the cells differentiated from p38α +/+ and p38α −/− cells (Figs. 3, 4, and 5).

Figure 6.

Effect of SB203580 (SB) on ESC differentiation and p38 activation. A: EB-derived monolayers were developed from p38+/+ ESCs in the absence (−SB, control) or presence of 10 μM SB (+SB). SMCs (a), ECs (b), and neurons (c) were detected by immunostaining with anti-SMA, anti-PECAM-1 antibodies, and anti-β-tubulin-3, respectively. The cells were examined and photographed with an Olympus fluorescence microscope (UplanFl 100×/1.30 oil lens). Scale bars = 10 μm. B: Inhibition of p38 by SB during EB differentiation. EBs were generated from p38α +/−A ESCs in the presence (+SB) or absence (−SB) of SB. 2-, 5-, and 7-day EBs (2d, 5d, and 7d, respectively) were collected. The cell lysates were prepared for Western blot analysis. Activated p38 (phosphorylated p38, pp38) was detected with the antibodies that recognize only the phosphorylated form (pp38). Activation of p38 was also indicated by the phosphorylation of HSP27 detected with anti-pHSP27 antibodies (pHSP). SB inhibited phosphorylation of both p38 and HSP27 without affecting the HSP27 protein level (HSP), which was detected by an antibodies that recognized total HSP27 protein.

We further tested p38 activity change during EB differentiation and the effect of SB under the experimental conditions described in Figure 6A. Activation of p38 requires phosphorylation of a threonine and a tyrosine residue in their active sites (Ono and Han,2000). Thus, phosphorylation of p38 on these two residues, which can be detected by Western blotting using the antibodies that specifically recognize phosphorylated forms of p38, has been widely used to indicate their activation (Kang et al.,2006). Figure 6B shows p38 activation levels (indicated by phosphorylated p38, pp38) in 2-, 5-, and 7-day-old EBs. Low levels of pp38 were detected in 2-day EBs. A significant increase of pp38 was observed in 5-day EBs and to a lesser extent in 7-day EBs, indicating that p38 was activated during EB differentiation. Incubation of EBs with 10 μM SB reduced p38 phosphorylation nearly to the basal level. The activation of p38 and the effect of SB were further assessed by the phosphorylation of heat shock protein 27 (HSP27), a downstream component of the p38 pathway. The phosphorylated HSP27 (pHSP) was detected at the time points corresponding to p38 activation (pp38, 5- and 7-day EBs). SB treatment completely reversed pHSP27 to the basal level while HSP27 (HSP, detected by an antibody that recognized total HSP27 protein) was not affected. Together, these results demonstrated that SB could effectively inhibit p38 activation during ESC differentiation (Fig. 6B), but it did not affect ESC differentiation to the three types of cells tested (Fig. 6A).

Quantitative Analysis of EC, SMC, and Neuron Marker Expression During ESC Differentiation

We have confidently identified ECs, SMCs, and neurons by microscopic analysis of cell morphology and immunodetection of cell markers. However, quantitative analysis of the number of each type of cells in EBs or EB-derived monolayers was difficult, especially ECs in vessel networks (Fig. 3) and neurons in clusters (Fig. 5). Therefore, we analyzed cell marker expression in 5- and 10-day EBs by quantitative real time-PCR (qRT-PCR). Figure 7A shows the relative mRNA levels of two markers for each cell type; SMA and SM22α for SMCs, PECAM-1, and VE-cadherin for ECs, and β-tubulin-3 and nestin for neurons. Among the genes tested, VE-cadherin was significantly decreased in 10-day p38α −/− EBs in comparison with wild type EBs. It also appeared that the expression of β-tubulin-3 and nestin was upregulated in both 5- and 10-day p38α −/− EBs, but the increases were not statistically significant. We noticed the relatively large variations (standard deviations) of the data in different experiments, which may be attributed to the nature of long-term EB differentiation and the heterogeneity environment within EBs as recently described by Carpenedo et al. (2007). It is also possible that the physiological state of p38α +/+ and p38α −/− ESCs before EB differentiation could affect the results. To eliminate this factor, we performed the experiments with p38α +/+ ESCs in the presence of SB to inhibit p38 activity during differentiation. As shown in Figure 7B, the expression levels of β-tubulin-3 and nestin were higher in SB-treated EBs than in the control EBs. Although statistically insignificant, this change reflected a similar tendency observed in p38α −/− cells (Fig. 7A). These results are in agreement with those reported by Aouadi et al. (2006) who showed that inhibition of p38α promoted ESC differentiation to neurons. The lack of statistical significance in our study could be related to the differentiation conditions used.

Figure 7.

Expression of EC, SMC, and neuron markers during ESC differentiation and the effect of SB. Total RNA was extracted from 5- and 10-day EBs generated from p38+/+ and p38−/− ESCs (A) or from 5- and 10-day EBs generated from p38+/− ESCs in the presence (+SB) or absence of 10 μM SB (−SB) (B). The mRNA level of each tested gene was determined by qRT-PCR in duplicates and was normalized to β-actin mRNA. The mRNA of each gene determined from p38+/+ EBs (+/+P, A) or EBs without SB treatment (−SB, B) was used as the control and was set as 100 (arbitrary unit). Results are means ± SD of four independent experiments. Statistical analysis was performed with the Student's t-test. The P < 0.05 (*) was considered statistically significant.


Generation of various cell types from ESCs may not only provide a promising source for cell-based therapy, but may also represent a powerful approach to study gene function in developmental biology (Keller,2005). For instance, the roles of VEGF signaling pathways in vasculogenesis and angiogenesis have been elegantly illustrated in recent studies using in vitro ESC differentiation systems (Schuh et al.,1999; Zeng et al.,2007; Vittet et al.,1996), which support and complement the findings from in vivo studies (Drake et al.,1998,2000). The objective of this study was to assess an overall role of p38α in ESC differentiation. Therefore, we induced ESC differentiation with serum without additional factors that promote specific cell type differentiation. The methods described in this study were simple yet allowed for simultaneous differentiation of ESCs into different cell types.

One of the major phenotypic features of p38α −/− embryos is overall underdeveloped vascular structures (Mudgett et al.,2000). However, it is not clear whether this is a direct effect of p38α knockout or secondary to insufficient nutrient and oxygen supply due to the defective placental organgenesis. Therefore, in vitro differentiation methods, in which nutrient and oxygen supply is not a limiting factor, are useful models to clarify the uncertain role(s) of p38α in vessel development. We first analyzed the differentiation of wild type and p38α mutant ESCs to vascular cells. Our results clearly demonstrated that, like wild type ESCs, p38α −/− ESCs could differentiate into ECs and SMCs, the two major cellular components of the vascular system. Together with the p38 inhibitor studies, our data provided direct evidence at the cellular level that p38α is not critical for SMC and EC differentiation. This is consistent with a hypothesis derived from in vivo studies that the defective vessel formation and the lethality of p38α knockout embryos are due to abnormal placental development, but not a direct effect of p38α knockout. This hypothesis is based on the findings that placental defects in p38α −/− embryos can be rescued by genetic complementation thus the embryo lethality can be avoided (Adams et al.,2000; Okada et al.,2007). Gene knockout is a powerful tool to identify gene function in animal models; however, the secondary death of embryos resulting from defective placental organgensis can be a major limitation. The present study using in vitro ESC differentiation provides useful knowledge to complement the data obtained from p38α knockout studies.

The roles of p38α in the regulation of SMC and EC activities have been intensively studied in adult somatic cells. In SMCs, p38α participates in a wide variety of cellular activities such as contraction, migration, and differentiation (Gerthoffer,2005). Recent studies (Matsumoto et al.,2002; Sweeney et al.,2003), including our own (Yang et al.,2004), have demonstrated that the p38 pathway regulates EC morphorgenesis during in vitro vessel assembly. Using a 3D cell culture system, we showed that inhibition of p38 activity caused EC elongation, but ECs failed to assemble into tube structure (Yang et al.,2004). Thus, we expected that p38α would play an important role in ESC differentiation to vascular cells. In this regard, it was surprising to find that p38α deletion did not affect ESC differentiation to SMCs and ECs. A similar conclusion can be extended to neuron differentiation despite the fact that p38 regulates diverse neuronal functions (Takeda and Ichijo,2002). Together with the findings from lymphocyte (Kim et al.,2005) and myeloid differentiation (Allen et al.,2000), our data imply that p38α may have less important roles in ESC differentiation process per se although it is critical in mediating various signals in differentiated cells. A possible explanation for the dispensability of p38α for ESC differentiation is that its function may be compensated for by other p38 isoforms, namely p38β p38γ or p38δ. In this aspect, one would speculate that p38β which shares the highest homology with p38α (Ono and Han,2000), is the best candidate. However, the fact that SB did not affect ESC differentiation to SMCs, ECs, and neurons argues against the requirement of p38β. Whether p38γ or p38δ is responsible for compensating p38α function in p38α −/− cells remains to be investigated.

In a recent study, Aouadi et al. (2006) reported that inhibition of p38 promoted neuron differentiation but reduced the number of cardiomyocytes under the induced differentiation with retinoic acid. In agreement with this finding, our data indicate a similar tendency, that the deletion or inhibition of p38α increased the expression of neuron markers. Therefore, the p38α activity level may affect the commitment of ESCs to develop into different cell lineages. In another study, Matsuo et al. (2006) reported that that the rate of metastasis of lung tumor cells injected into p38α +/− mice was markedly decreased compared to that in wild-type mice. Further analysis revealed that the p38α +/− mice have a lower expression of E- and P-selectin and a reduced interaction between tumor cells and ECs. This study demonstrates that although p38α +/− mice have a normal appearance, they have altered responses to pathological conditions. At the cellular level, the consequences of p38α deletion seem to be also primarily associated with functional defects, as found in p38α −/− ESC-derived myeloid cells (Allen et al.,2000) and lymphocytes (Kim et al.,2005). In both cases, no defects were found in the differentiation process per se. This appears to be the case for EC, SMC, and neuron differentiation although at present we do not know how the functions of these cells are affected by p38α deletion. Together, these studies suggest that p38α deletion may not significantly compromise the potential of ESC differentiation to certain cell types with expected morphology, but it may affect the function of differentiated cells and/or the commitment of ESCs to different cell lineages.

The differentiation methods used in this study allow for the cell–cell interaction in a three-dimensional EB structure, which not only promotes cell differentiation, but also the assembly of vessel-like structures. Therefore, EBs represent a similar environment in the developing embryo. However, the heterogeneous nature of differentiated cells limits functional analysis of a specific cell type. Although we have shown that p38α −/− ESCs displayed altered cell adhesion to different extracellular matrix proteins (Guo and Yang,2006), it is not clear how this alteration is reflected in differentiated p38α −/− cells. Since the main objective of this study was to investigate the overall differentiation capacity of p38α −/−ESCs, the simultaneous generation of multiple cell types induced by serum is desirable. However, a potential shortfall is that serum may contain undefined factors that might be able to compensate for the lost function of p38α. Thus, the importance of p38α could be underestimated. The solutions to these problems will depend on the development of defined differentiation methods and the eventual isolation of each type of cells for further analysis. While recognizing the advantages and limitations of the methods used in this study, we can conclude that p38α−/−ESC can differentiate to ECs, SMCs, and neurons without apparent defect in cell morphology.



Recombinant murine leukemia inhibitory factor (LIF) was purchased from Chemicon International. p38 inhibitor SB203580 was purchased from Calbiochem. Monoclonal antibodies against smooth muscle α-actin (SMA, clone 1A4), β-actin (clone AC-15), and β-tubulin-3 (clone SDL.3D10), and SYBR green jumpstart TAQ readymix were purchased from Sigma Chemical Co. Anti-p38α and anti-phopsho-p38 (pp38) antibodies were purchased from Cell Signaling Technology Inc. Anti-HSP27 and anti-pHSP27 antibodies were purchased from Stressgen. Anti-PECAM-1 antibodies (MEC 13.3) and anti-VE-cadherin antibodies (11D4.1) were purchased from BD Biosciences.

ESC Culture

ESCs were maintained in knockout-DMEM (Invitrogen) containing 15% fetal bovine serum (FBS), 0.2 mM L-glutamine, 0.1 mM 2-mercaptoethanol, 0.1 mM MEM nonessential amino acids, and 1,000 U/ml LIF. ESCs were initially expanded on embryonic fibroblast feeder cells. For experiments, fibroblasts were removed from ESCs by incubating cell suspension on bacterial dishes for 60 min. Non-adherent ESCs in suspension were separated from attached fibroblasts. ESCs were then routinely grown in cell culture dishes coated with 0.1% gelatin at 37°C in a humidified atmosphere at 5% CO2 as previously described (Guo and Yang,2006).

Embryoid Body (EB) Formation and Cell Differentiation

EB formation was performed by suspending ESCs in bacterial culture dishes (1 × 105 cells/ml) in which ESCs clumped and formed EBs. After incubation for 24 hr, the medium was changed to LIF-free ESC medium containing 15% FBS to initiate cell differentiation. After incubation for 5 days, EBs were further differentiated in suspension for an additional 5 days; alternatively, 5-day-old EBs were transferred to gelatin-coated cell culture dishes or cover glasses. In the latter case, EBs adhered to the culture dishes where the cells within the EBs grew out to form monolayer-like structures composed of various types of cells, which we defined as EB-derived monolayers in this study. To test the effect of SB203580 on p38 activation and ESC differentiation, 10 μM SB203580 was added to the cell culture medium at the second day of EB formation and refreshed with the medium change every other day. EBs and EB-derived monolayers were used as the sources to identify the differentiated cells and for further analysis.

Immunocytochemistry and Microscopic Analysis

EBs and EB-derived monolayers were fixed with 4% paraformaldehyde and washed with PBS. In some experiments, EB-derived monolayers were stained with 1% toluidine blue in 10 mM Tris-HCl (pH 9.5), 150 mM NaCl. After washing with PBS, the stained cells were analyzed under a bright field microscope. Immunocytochemical analysis was performed according to our published protocols (Guo et al.,2001). Briefly, fixed EBs or EB-derived monolayers were permeabilized with PBS containing 0.25% Triton X-100 for 30 min. After being blocked in 2% bovine serum albumin and 5% preimmune serum, the cells were incubated with primary antibodies as indicated overnight at 4°C. The positive cells were detected by fluorescein isothiocyanate (FITC)- or rhodamine-conjugated secondary antibodies. In some experiments, the nuclei were stained with 10 μ Hoescht 33258 and F-actin was stained with 5 μg/ml FITC-phalloidin. The cells were examined under an Olympus fluorescence microscope (BX60) and photographed with a micropublisher digital camera (Qimaging). The images were processed using Image-Pro Plus software. In some experiments, the cells were examined with a LSM 510 laser-scanning confocal microscope (Zeiss). Image analysis was performed using LSM Image Examiner software (Zeiss). For each experimental condition, at least five fields were examined. The experiments were repeated from two to four times with similar results. The details for each experiment are described in the figure legends.

RNA Extraction and Reverse Transcription-Polymerase Chain Reaction (RT-PCR)

Total RNA was extracted from cells with Tri-reagent and treated with DNase to eliminate residual genomic DNA. cDNA was prepared by reverse transcription using M-MLV reverse transcriptase. The specificity of PCR products was determined by dissociation curve and confirmed by agarose gel electrophoresis. Quantitative real time-PCR (qRT-PCR) was performed using the SYBR green jumpstart TAQ readymix on a MX3000PTM Real-time PCRsystem (Stratagene). Sequences for the primer sets were as follows:




VE-cadhesin, forward: 5′-CACTGCTTTGGGAGCCTTC -3′ reverse: 5′-GGGGCAGCGATTCATTTTTCT-3′




The mRNA level from qRT-PCR was calculated using the comparative Ct method (Pfaffl,2001). β-actin mRNA was used as a calibrator for the calculation of relative mRNA levels of the tested genes.

Cell Lysate Preparation and Western Blot Analysis

ESCs or EBs were lysed in M-PER mammalian cell protein extraction buffer (Pierce) supplemented with a cocktail of protease inhibitors. After being kept on ice for 30 min, the extracts were centrifuged at 15,000g for 15 min at 4°C. The supernatant was designated as the cell lysate and was subjected to SDS-PAGE. Western-blot analysis was carried out as previously described (Guo et al.,2001).


We thank Dr. Christopher Gabel (Pfizer Central Research) and Dr. Barry Sleckman (Washington University School of Medicine at St. Louis) for providing p38 ESCs. We also thank Mississippi Functional Genomics Network for the use of the facility (supported by P20RR016476 from the National Center for Research Resources). This work was supported by The National Institutes of Health grants HL08273 and HL081126 (to Y.L.G.).