MAP3Ks as central regulators of cell fate during development



The cytoplasmic serine/threonine kinases transduce extracellular signals into regulatory events that impact cellular responses. The induction of one kinase triggers the activation of several downstream kinases, leading to the regulation of transcription factors to affect gene function. This arrangement allows for the kinase cascade to be amplified, and integrated according to the cellular context. An upstream mitogen or growth factor signal initiates a module of three kinases: a mitogen-activated protein (MAP) kinase kinase kinase (MAPKKK; e.g., Raf) that phosphorylates and activates a MAP kinase kinase (MAPKK; e.g., MEK) and finally activation of MAP kinase (MAPK; e.g., ERK). Thus, this MAP3K-MAP2K-MAPK module represents critical effectors that regulate extracellular stimuli into cellular responses, such as differentiation, proliferation, and apoptosis all of which function during development. There are 21 characterized MAP3Ks that activate known MAP2Ks, and they function in many aspects of developmental biology. This review summarizes known transduction routes linked to each MAP3K and highlights mouse models that provide clues to their physiological functions. This perspective reveals that some of these MAP3K effectors may have redundant functions, and also serve as unique nexus depending on the context of the signaling pathway. Developmental Dynamics 237:3102–3114, 2008. © 2008 Wiley-Liss, Inc.


The cytoplasmic serine/threonine kinases transduce extracellular signals into regulatory events that impact the response of cells to such stimuli. The induction of one kinase triggers the activation of several downstream kinases, leading to the regulation of transcription factors to affect gene function. This arrangement allows for the kinase cascade to be amplified, modulated and integrated according to the cellular context. The classic and basic arrangement starts with a G-protein activating upstream of a signaling module of three kinases: a mitogen-activated protein (MAP) kinase kinase kinase (MAPKKK; e.g., Raf) that phosphorylates and activates a MAP kinase kinase (MAPKK; e.g., MEK) and finally activation of MAP kinase (MAPK; e.g., ERK). Thus, this MAP3K-MAP2K-MAPK module represents critical intermediate effectors that either positively or negatively propagate extracellular stimuli into cellular responses, such as differentiation, proliferation, and apoptosis all of which function during development. The MAP3Ks (Mitogen activated kinase kinase kinases) provide specificity for stimulus-dependent activation of MAP2K-MAPK pathways through unique protein–protein interactions and phosphorylation of signaling effectors.

There are 21 characterized MAP3Ks that activate known MAP2Ks (Table 1), and they arrange into various clusters based on protein homology as shown in a dendrogram (Fig. 1A). When only the kinase domains are compared from each of these MAP3Ks, specific MAP3Ks shift in the dendrogram (Fig. 1B), but continue to form distinct groups. For example, full-length MEKK4 is similar to MEKK1/2/3, but the kinase domain of MEKK4 is closer to that of ASK1/2. Also interesting, full-length TAK1 identifies with the mixed lineage kinases (MLKs) being the most similar to DLK and LZK, but its kinase domain is diverse from the other MLKs and less similar to DLK and LZK. This suggests that variation in regulatory regions outside the kinase domain may confer structural and functional differences to each of these proteins. This review will summarize the activity and physiological functions of MAP3Ks in development as gained from mouse genomic models. This perspective highlights that some of these effectors may have redundant functions but they may also serve as unique nexus depending on the context of the signaling pathway.

Figure 1.

Relationship of MAP3Ks based on protein homology. Dendogram showing the relationship between full-length MAP3Ks (A). Dendogram of MAP3K kinase domains (B). The 22 MAP3K sequences were located on the National Center for Biotechnology Information (NCBI) website and entered into ClustalW2 ( for alignment (Larkin et al.,2007). Tree data was copied from ClustalW2 into Phylodendron ( to create the relationships (Gilbert,1997).

Table 1. Common and Alternative Names of MAP3 Kinases
Common nameAlternative name


The Raf family of protooncogenes are cytoplasmic serine/threonine kinases which not only are related to oncogenesis but play important physiological roles in cell growth and development. Each Raf isoform has a distinct expression profile in tissues which suggests that they may carry nonredundant functions. C-Raf (Raf-1) is widely expressed while A-Raf and B-Raf are restricted to select tissues; A-Raf is expressed mostly in urogenital and gastrointestinal tissues (Luckett et al.,2000) and B-Raf is expressed in neural, testicular, splenic, and hematopoietic tissues (Jaiswal et al.,1996). Thus, the expression patterns suggest distinct physiological functions in development and disease for the Raf molecules. Several independent genetic studies in mouse models provide insight into the physiological roles for each Raf isoform as described below.



During development, A-Raf transduces signals initiated by a variety of mitogens and growth factors, especially in the urogenital and gastrointestinal tracts (Mahon et al.,2005). As shown in Figure 2, active A-Raf induces the activation of the MEK/ERK pathway, although its affinity for MEK proteins is significantly lower than that of other Raf kinases (Marais et al.,1997). Additionally, A-Raf appears to selectively activate MEK1, unlike B-Raf and C-Raf which activate both MEK1 and MEK2 (Wu et al.,1996; Sridhar et al.,2005).

Figure 2.

Major signaling pathways activated by Raf isoforms. Unlike B-Raf and C-Raf, A-Raf activates MEK1 but not MEK2 in response to growth factors and mitogens. Mice lacking the A-Raf gene and kept in a C57 BI/6 genetic background die between days 7 and 21 postpartum due to neurological and intestinal abnormalities. Several studies link B-Raf activity to the mobilization of ERK1/2 and potentially ERK3. Mouse embryos with a disruption in B-Raf die between E10.5 and E12.5 due to severe vascular abnormalities. C-Raf (Raf-1) regulates apoptosis by means of the MEKK5, ERK1/2 and NFκB signaling pathways. Cardiac specific C-raf knockout mice survive to adulthood but display excessive apoptosis of cardiomyocytes leading to cardiac dysfunction.

Mice with conventional gene disruption of the A-Raf gene exhibit intestinal and neurological defects in a predominantly C57 BI/6 genetic background (Pritchard et al.,1996). Mutant mice are born and appear normal at birth but become significantly smaller than wild-type littermate controls by postpartum day 3. They also display athetotic movements, abnormal propioception, rigidity of musculature, and pronounced distension of the colon. These mice die between days 7 and 21 postpartum, likely due to the range of neurological and intestinal problems. However, this lethal phenotype can be rescued by maintaining the mutated allele on a 129/OLA background. In this case, A-Raf–deficient mice survive to adulthood, are fertile, do not display intestinal defects and show few neurological abnormalities (Pritchard et al.,1996). These observations suggest that genetic modifiers may impact A-Raf activity and that other Raf kinases may compensate for A-Raf during development of urogenital and gastrointestinal organ systems.



Dominant-activating mutations in B-Raf are linked to a wide range of cancers including melanoma, colorectal, and glioma malignancies (Basto et al.,2005; Poynter et al.,2006; Barault et al.,2008). The conventional gene-disruption of B-Raf in mice revealed an unexpected function in vascular development. B-Raf–deficient mice die in utero between embryonic days E10.5 and E12.5 due to severe vascular abnormalities. Mutant embryos show excessive apoptosis of endothelial cells resulting in irregularly shaped blood vessels which leads to reduced blood circulation and hemorrhage into major body cavities (Wojnowski et al.,1997). The B-Raf–deficient phenotype is similar to that observed in mice lacking Tie2, a receptor tyrosine kinase involved in vasculogenesis (Dumont et al.,1994). This suggests that B-Raf may be activated by Tie2 signaling. This link is supported by studies showing that binding of Tie2 to its ligand angiopoietin-1 induces the activation of MEK1/2 and ERK1/2, signaling molecules downstream of B-Raf (Yoon et al.,2003). Although ERK1-deficient mice are viable, ERK2−/− embryos have decreased vascular patterning in the placenta contributing to lethality by midgestation. This phenotype is very similar to that observed in MEK1 knockout embryos (Bissonauth et al.,2006). Therefore, B-Raf may act as the central MAP3K regulating placenta vascular patterning. In addition to activating ERK1/2, recent evidence suggests that B-Raf may also activate ERK3 (Fig. 2) (Hoeflich et al.,2006). Although little is known about the functions of ERK3, this MAPK is markedly up-regulated during mouse development and myogenic differentiation (Turgeon et al.,2000; Coulombe et al.,2003). These findings suggest a role for B-Raf during cellular differentiation and embryogenic vascular formation.

C-Raf (Raf-1)


C-Raf or Raf-1 is the prototypic member of the Raf family being the most ubiquitously expressed. Originally believed to be a potent effector for cell proliferation it is now well characterized for its function as a regulator of apoptosis (Slater et al.,2003; Lamberti et al.,2007). C-Raf can phosphorylate and thus inactivate the proapoptotic protein BAD (Kebache et al.,2007). Active C-Raf also induces the MEK/ERK pathway which potentiates BCL-2-mediated resistance to apoptosis (Fig. 2) (Pardo et al.,2002; Kumar et al.,2007). Conventional C-Raf knockout mice appear normal during the first stages of development up to E11. At this stage, the liver enlarges and takes over as the main source of hematopoeitic activity from the yolk sac. However, as embryogenesis progresses C-Raf −/ − embryos become growth retarded and die between days E11.5 and E13.5. This timing in death of C-Raf−/ − embryos coincides with the period in which the developing liver becomes a well differentiated organ. The mutant mice display poorly vascularized placentas and increased apoptosis in embryonic tissues, especially in the liver (Mikula et al.,2001). To examine the role of C-Raf in the heart, Yamaguchi et al. (2004) created a cardiac-specific deletion of C-Raf (C-Raf CKO) by use of Cre-recombinase driven by α-myosin heavy chain promoter. C-Raf CKO mice develop to adulthood, are fertile and have a relatively normal lifespan. However, there is an induction of apoptosis in cardiomyocytes from mice between 3 and 5 weeks of age. This increased apoptosis results in cardiac muscle enlargement and fibrosis with subsequent left ventricular dysfunction (Yamaguchi et al.,2004). C-Raf CKO mice also display an increase in the activity of MEKK5 (ASK1), which has been shown to play an important role in the induction of neuronal and cardiac apoptosis (Kanamoto et al.,2000; Yamaguchi et al.,2003). Ablation of MEKK5 in C-Raf CKO mice rescues the enhanced-apoptosis phenotype, which suggests that C-Raf promotes cell survival in the heart through modulation of MEKK5 activity. Collectively, these observations highly implicate C-Raf in the maintenance and function of cardiomyocytes.

In humans, mutations in the C-Raf gene have been implicated in developmental disorders such as LEOPARD and Noonan syndromes (Pandit et al.,2007). Noonan syndrome is an autosomal dominant congenital disease characterized by short stature, facial dismorphia and cardiac abnormalities such as pulmonary stenosis, hypertrophic cardiomyopathy and atrioventricular septal defects (Sharland et al.,1992; Allanson,2007). Approximately 50% of all Noonan syndrome cases have mutations in the PTPN11 gene, which encodes the nonreceptor tyrosine phosphatase Shp2 (Tartaglia et al.,2004). This enzyme is a key component in several signaling pathways that control developmental processes such as cardiac valvulogenesis and hematopoietic cell differentiation (Neel et al.,2003). In individuals with Noonan syndrome that do not have mutations in the PTPN11 gene, the development of hypertrophic cardiomyopathy (CMH; OMIM#611553) has been linked to mutations in the C-Raf gene (Pandit et al.,2007). However, only lesions that make the C-Raf gene product over-active result in cardiomyopathy while C-Raf loss-of-function mutations have not been associated with CMH. This highlights the importance of properly regulated C-Raf activity for normal cardiovascular formation and function.

Many of the functions of C-Raf, however, appear to be performed in concert with other members of the Raf family, at least during the initial developmental stages. In mouse mutants with only one of four copies of B-Raf and C-Raf (B-Raf−/−/C-Raf−/+ or B-Raf−/+/C-Raf−/−) there is a greater impairment of growth compared with C-Raf single knock outs. The compound mutants exhibit significant underdevelopment of the brain, heart and limbs. This leads to death of 90% of the embryos before E10.5. The complete absence of B-Raf and C-Raf further exacerbates the developmental delay due to a lack of differentiation of embryonic lineages which results in death of the embryos before E 8.5 (Wojnowski et al.,2000). These findings suggest that B-Raf and C-Raf display a significant redundancy during early embryonic development. However, the activity of both Raf molecules is indispensable for normal development beyond the blastocyst stage.

TAK1 (TAK1a, TAK1b, and TAK1c isoforms)


TGFβ Activated Kinase-1 (TAK1) is expressed as three different isoforms from which TAK1a is the most commonly found. TAK1 is a downstream effector of TGFβ, which are cytokines that play an essential role in cell proliferation and differentiation during embryonic development (Wagner and Siddiqui,2007). Activation of TAK1 by TGFβ mobilizes p38 and JNK by means of the phosphorylation of MKK3/6 and MKK4/7, respectively (Fig. 3) (Dempsey et al.,2000). TAK1 is also found downstream of tumor necrosis factor α (TNFα) and toll-like receptor (TLR) ligands during cell survival and pro-inflammatory responses. Activation of TAK1 by TNFα and TLR ligands induces the release of NFκB from its cytoplasmic inhibitory complex and its subsequent translocation to the nucleus (Sakurai et al.,1998). NFκB is a key transcription factor whose target genes are important for inflammation responses and cell survival especially in epithelial to mesenchymal transition processes during organogenesis. The importance of TAK1 in the transduction of survival signals has been demonstrated in vivo through gene inactivation studies. Interferon-inducible Cre-recombinase mediated deletion of TAK1 in adult mice results in bone marrow and liver failure due to increased apoptosis of hematopoietic cells and hepatocytes (Tang et al.,2008). Additionally, cells lacking TAK1 fail to activate the NFκB and JNK cascades suggesting that TAK1 does promote cell survival by regulating apoptosis by means of these pathways (Liu et al.,2006; Tang et al.,2008).

Figure 3.

TAK-1 and its downstream signaling effectors. TAK-1 activates the p38 MAPK and JNK signaling pathways. Conventional disruption of the TAK-1 gene in mice leads to lethality by day E11.5 due to severe vascular defects and abnormal growth. Conditional deletion of the TAK-1 gene in mice results in massive apoptosis of hepatocytes and hematopoietic cells, suggesting a crucial role for TAK-1 in cell survival.

Conventional disruption of the TAK1 gene in mice causes a lethal phenotype by day E11.5 due to severe vascular defects and substantial delayed growth. These embryos have a small head, retarded posterior development, and an enlarged pericardium consistent with a failing cardiovascular system. They also exhibit extreme dilation and decreased complexity of vascular branching in the yolk sack as well as an absence of vascular smooth muscle (Jadrich et al.,2006). Interestingly, loss-of-function mutations in the TGFβ type I receptor Alk1 and the auxiliary TGFβ receptor endoglin display defects in vascular development that are similar to those observed in TAK1 null mice (Arthur et al.,2000; Urness et al.,2000). These findings, together with studies showing that TAK1 is activated by TGFβ (Yamaguchi et al.,1995) indicate that TAK1 may be a major downstream effector of TGFβ signals to regulate vascular development. Also, TAK1 activates the SAPK/JNK and to a lesser extent p38 mitogen-activated protein kinase (MAPK) pathways, which in turn regulate the transcription of genes involved in angiogenesis (Fig. 5; Moriguchi et al.,1996; Shirakabe et al.,1997). Furthermore, increased p38 levels have been shown to inhibit TAK1 activity, suggesting that a negative feedback mechanism contributes to the regulation of the TAK1/p38 pathway (Cheung et al.,2003). Similar to TAK1-deficient mice, MAPK p38α knockout mice also exhibit abnormalities in vascular development, although yolk sac angiogenesis does not appear to be affected in the latter (Adams et al.,2000; Mudgett et al.,2000). These differences in phenotype between the TAK1 and p38α null mice may be due to redundant activity of other p38 isoforms in the yolk sac which may compensate for the lack of p38α in this tissue (Ihle,2000). Collectively, these studies highlight the importance of TAK1 in mediating TGFβ signaling especially for vascular development and likely maintenance of the cardiovasculature in adults.

Figure 5.

Major signaling pathways activated by MEKK4 and MEKK5/6. MEKK4 and MEKK5/6 (ASK1/2) activate the p38 and JNK pathways in response to cytokines and/or stress. MEKK4 knockout mice and kinase inactive mice exhibit neural tube, skeletal and skin defects. MEKK5 knockout mice exhibit dilation of the heart and increased apoptosis in maturity. MEKK5 is necessary for left ventricular remodeling after myocardial infarct.



MEKK1 appears to transduce signals for activation of JNK and ERK1/2 by pro-inflammatory stimuli and growth factors (Fig. 4; Yujiri et al.,1998; Xia et al.,2000). MEKK1 is a key MAP3K regulating cell migration. Developmentally, MEKK1-knockout mice exhibit altered migration of epithelial cells resulting in defective eyelid closure. Despite this minor defect, these animals are fertile and viable under normal conditions (Yujiri et al.,2000). In regards to developmental signals, MEKK1-induced epithelial cell movement during eyelid closure is activated by TGFβ and activin, two important developmental factors (Zhang et al.,2003). Further studies with these mice show that MEKK1 through JNK MAPK activity protects from pressure overload in the heart (Sadoshima et al.,2002). In this regard, one heterotrimeric G-protein, Gαq, contributes to the onset of cardiac hypertrophy triggered by seven-transmembrane receptors (Minamino et al.,2002). Some factors, including Angiotensin II, endothelin, and adrenergic agonists are known to induce this type of hypertrophy (Simpson,1983; Shubeita et al.,1990; Sadoshima and Izumo,1993; Glennon et al.,1995). MEKK1 knockout mice and cardiac myocytes derived from these animals are resistant to Gαq-mediated hypertrophic response (Minamino et al.,2002). This indicates that MEKK1 is critically involved in the development of hypertrophic heart disease. Phenylephrine-induced hypertrophy is mediated by MEKK1 activation; however, inhibition of Ras was able to partially block the effect in an in vitro system (Aoki et al.,2000) suggesting that MEKK1 is not a sole player in hypertrophic responses. Also, MEKK1 interacts directly with c-Raf, MEK1, and ERK2 (Karandikar et al.,2000), all of which are important to cell survival and hypertrophic responses (Ueyama et al.,2000; Harris et al.,2004). It is also known that MEKK1 signaling through JNK and ERK MAPKs increases cell survival in embryonic fibroblasts (Yujiri et al.,1998). Moreover, embryonic cardiac myocytes lacking MEKK1 are highly sensitive to oxidative stress-induced apoptosis. MEKK1 promotes myocyte survival by negatively regulating TNFα production (Minamino et al.,1999). Thus, manipulation of MEKK1 activity may be useful in protecting the heart from ischemic damage and reperfusion injury. In addition to hypertrophy and cell survival, there is evidence showing the involvement of MEKK1-JNK in epithelial cell migration for re-epithelialization of injured tissue, as observed with MEKK1 mutant mice with a deletion of the kinase domain (MEKK1KD/KD; Deng et al.,2006). In another study, Gallagher et al. (2007) found that mice deficient in MEKK1 kinase activity have diminishing B-cell numbers. Specifically, B-cell proliferation is positively regulated by CD40-dependent mobilization of MEKK1-JNK/p38 signal indicating MEKK1 activity contributes to the development of clonal populations of B cells. Definitive evidence for the role of MEKK1 in cell motility and dissemination is reported by Cuevas et al. (2006) showing MEKK1 controlling the onset of mammary tumor metastasis in polyoma middle T antigen mice (MMTV-PyMT). MEKK1−/− PyMT-mice have impaired degradation of basement membrane ECM by urokinase-type plasminogen activator (uPA), which retards the onset of tumor cell migration and metastasis. Although MEKK1 is not required for the growth of primary PyMT-tumors, this MAP3K promotes tumor cell migration and dissemination to distant sites to form metastatic foci. Collectively, these observations highlight the function of MEKK1 to directly and indirectly promote cellular migration. This activity affects prevention of apoptosis, promotion of hypertrophy, proliferation of B-cells, and dissemination of breast cancer cells.

Figure 4.

Major signaling pathways activated by MEKK1-3. MEKK1 can activate ERK1/2 and JNK. MEKK1 knockout mice have defects in eyelid formation and epithelial cell migration. MEKK2 can activate ERK5 and JNK. MEKK2 knockout mice do not have developmental defects, but alterations in T-cell response. MEKK3 can activate ERK1/2, ERK5, p38, and JNK. MEKK3 knockout mice have defects in developmental angiogenesis and heart formation, particularly related to endothelial cell function.



MEKK2 activity appears to be selective for activating JNK and ERK5 (Fig. 4; Chayama et al.,2001; Hammaker et al.,2004; Kesavan et al.,2004). MEKK2 knockout mice are viable and fertile without any apparent developmental malformations. T-cell responses in MEKK2−/− mice are enhanced in part due to elevated JNK activation suggesting that T-cell receptor modulation may be the major role for this MAP3K (Guo et al.,2002). Cytokine production in response to IgE is also inhibited by disrupting MEKK2 (Garrington et al.,2000). Interestingly, ERK5 can be activated by MEKK2 and MEKK3 (Nakamura and Johnson,2003). ERK5 knockout embryos fail to properly remodel the vasculature to form normal blood vessel patterning during development (Regan et al.,2002). However, the lack of cardiovascular defects in MEKK2-deficient mice rules out a compensatory role for MEKK2 in vascular development. The analysis of MEKK2−/− on other MEKK-deficient backgrounds may reveal unrecognized functions for MEKK2 in the maturation and function of cellular immunity.



MEKK3 is a MAP3K that regulates p38, ERK1/2 and JNK (Fig. 4) (Deacon and Blank,1997; Uhlik et al.,2003; Fritz et al.,2006). More recent evidence suggests that it also regulates activities of ERK5 (Chao et al.,1999) and the transcription factor, NFκB (Zhao and Lee,1999; Yang et al.,2001). MEKK3 knockout mice (MEKK3−/−) die in utero at approximately E10.5 due to abnormalities in extraembryonic vasculature of the yolk sac, the embryonic vasculature and the heart. Although primary vessels form, angiogenesis fails to occur (Yang et al.,2000). Additionally, the endocardium does not adhere to the cardiac myocytes of the ventricle (Yang et al.,2000). These defects are partially due to increased endothelial cell apoptosis (Deng et al.,2007). Additionally, the MEKK3-MEK5-ERK5 pathway is implicated in cardiovascular developmental programming, as mice deficient for MEK5 or ERK5 have similar cardiovascular defects including failure to form primary vasculature and the endocardium to mature (Regan et al.,2002; Wang et al.,2005b). On the cellular level, decreased proliferation and increased apoptosis are also observed in the hearts of Mek5 null embryos (Wang et al.,2005b). An increase in apoptosis is also observed in the cephalic mesenchyme of Erk5 knockout embryos (Yan et al.,2003). In addition, fibroblasts from MEK5-deficient embryos fail to induce Myocyte enhancing factor 2 (MEF2) activity, which is a downstream mediator of ERK5. Importantly, overexpression of MEK5 to the heart in mice leads to hypertrophy and eventually dilated cardiomyopathy (Nicol et al.,2001). Activation of p38, JNK, ERK1/2, and NFκB signaling pathways are necessary for epithelial to mesenchymal transition (EMT) with each having an effect during mesenchyme production (Compton et al.,2006; Rivera-Feliciano et al.,2006; Santibanez,2006; Grund et al.,2008). Of the MAP kinases, only ERK5 has yet to be examined in terms of developmental EMT. For these reasons, MEKK3 may act as a master MAP3K in regulating signaling effectors required for developmental EMT during cardiovascular and nervous system formation.



MEKK4 activates JNK and p38 MAP kinases in response to stress stimuli such as osmotic shock (Fig. 5; Gerwins et al.,1997; Takekawa and Saito,1998; Bettinger and Amberg,2007). Both MEKK4 knockout and MEKK4 kinase inactive mutant mice are born, but most animals die shortly after birth. These mice exhibit neural tube and skeletal malformations (Abell et al.,2005; Chi et al.,2005). Neural tube defects in MEKK4-deficient or kinase inactive embryos include spina bifida and exencephaly and appear due to increases in apoptosis during neural development. MEKK4 is observed to be strongly expressed in the neuroepithelium, which further supports its role in developmental epithelial biology. In addition to neural tube malformations, there is a disruption in skeletal patterning in MEKK4 kinase inactive mice. Because MEKK4 is an upstream activator of JNK, it is worth noting that JNK1/JNK2 double knockout mice also have neural tube abnormalities (Sabapathy et al.,1999). The neural tube defect is highly penetrant in the MEKK4−/− mice suggestive of a role in developmental epithelial biology. There is enhanced apoptosis in MEKK4 mutant mice at the time coincident with closure of the neural tube, but no differences in cell proliferation (Abell et al.,2005). Because cardiac chamber partitioning and valve formation also require appropriate timing of cellular apoptosis, alterations in programmed cell death may also lead to heart defects. In addition, mice deficient in c-Jun, a transcription factor downstream of JNK, exhibit outflow tract defects in the heart (Eferl et al.,1999). The outflow tract tissue is partially derived from neural crest. The migration of the cardiac neural crest is essential for parasympathetic innervation of the heart as well as completion of aortic–pulmonary septation (Kirby et al.,1983). Additionally, the JNK pathway has been implicated downstream of WNT signaling (Yamanaka et al.,2002), and WNT factors are critical to outflow tract formation. Knockout mice of Dishevelled 2, a modulator of the WNT pathway, exhibit neural tube defects as well as malformations of the cardiac outflow tract (Hamblet et al.,2002). WNT signaling can operate through a JNK-dependent mechanism, which involves MEKK4 (Luo et al.,2003). Because the neural crest is an important contributor to the outflow tract (Jiang et al.,2000; Porras and Brown,2008), it could be suggested that a proportion of MEKK4-deficient mice will likely have outflow tract malformations. In this regard, MEKK4 was recently shown to regulate developmental EMT in cardiac cushion tissue or the prevalvular regions of the embryonic heart. However, MEKK4 is probably acting to regulate proliferation, apoptosis, and/or cell migration, because MEKK4 alone is not sufficient to cause endocardial differentiation to mesenchyme (Stevens et al.,2006). Additional studies have shown that MEKK4 functions in both apoptosis and cell migration in other systems (Chi et al.,2005; Sarkisian et al.,2006), and these MEKK4-regulated processes will have to be explored during heart morphogenesis to determine its function in valvulogenesis, chamber partitioning, and formation of the coronary vasculature. Interestingly, Angiotensin II (AngII), a key regulator of the cardiovascular system, leads to tyrosine phosphorylation of MEKK4 by means of a calcium-Pyk2–dependent mechanism in rat smooth muscle cells. This phosphorylation causes MEKK4 to activate its downstream target, MKK6 (Derbyshire et al.,2005). As MKK6-deficient mice have deficits in T cell signaling and apoptosis (Tanaka et al.,2002), collectively, these observations suggest that MEKK4 may be a central mediator of calcium-mediated signaling. Calcium is crucial for the MEKK4-MKK6-p38Mapk transduction route and also for calcium-regulated phosphatase, calcineurin, which may regulate MEKK4 or visa verse. Both calcineurin and MKK6 are key effectors in T cell signaling and activation. Thus, additional studies may reveal a role for MEKK4 in T cell development and function.

MEKK5 (ASK1) and MEKK6 (ASK2)

(OMIM#*602448, *604468).

MEKK5 or Apoptosis Signal-regulating Kinase (ASK1) functions, as its name indicates, for regulating apoptosis downstream from stress stimuli (Tobiume et al.,1997). MEKK5-knockout mice do not exhibit developmental defects (Yamaguchi et al.,2003). However, MEKK5 is implicated in left ventricular remodeling by inducing apoptosis after myocardial infarcts (Yamaguchi et al.,2003). The loss of MEKK5 in C-Raf–deficient hearts rescues the cardiac phenotype, characterized by increased apoptosis and dilation of the heart (Yamaguchi et al.,2003). MEKK5 activates the p38 and JNK pathways (Fig. 5), which are involved in development and pathological remodeling of the heart (Ichijo et al.,1997). When MEKK6 interacts with MEKK5 in a heteromeric complex, these MAP3k can induce apoptosis (Takeda et al.,2007; Ortner and Moelling,2007). A MEKK6-deficient mouse model has not been reported yet, but may likely reveal definitive roles for MEKK6 in apoptosis during development. Although no cardiovascular defects are observed with loss of MEKK5 in development, it may provide additive or even compensatory function to MEKK4, which functions through p38 and JNK as well (Gerwins et al.,1997; Abell et al.,2007). MEKK5 is expressed widely during mouse development including high expression in the myocardium of the embryonic heart (Ferrer-Vaquer et al.,2007). Investigating double-knockout mice for MEKK4/MEKK5 or MEKK5/MEKK6 will be informative for determining compensatory effects by these related MAP3Ks during development.

TAO kinase (TAOK1, TAOK2, TAOK3)


Thousand And One kinase (TAOk) primarily activates the p38 pathway in response to cellular stress and DNA damage (Fig. 6; Hutchison et al.,1998; Chen et al.,2003; Raman et al.,2007). The cell cycle protein, Ataxia telangiectasia mutated (ATM), is activated in response to DNA damage, and directly phosphorylates TAOk. Phosphorylated TAOk then causes activation of p38 MAPK leading cells to enter the G2/M damage checkpoint for DNA repair (Raman et al.,2007). To date, three of these MAP3Ks have been characterized: TAOk1, TAOk2, and TAOk3. TAOk1 interacts with and activates MKK3 leading to activation of p38 (Hutchison et al.,1998). TAOk2 negatively regulates TAK1 activation of the NFκB pathway after osmotic stress (Huangfu et al.,2006). TAOk2 associates with MKK3 and MKK6, MAP2Ks upstream of p38 (Chen and Cobb,2001). TAOk may also be the MEKK that inhibits TAK1 through p38 activity. Although the importance of p38 has been demonstrated in cardiovascular development and disease (Liao et al.,2001; Engel et al.,2005), the TAO kinases have not been characterized in development. Knockout mice or other transgenics of TAO kinases will be necessary for future studies in developmental processes.

Figure 6.

Thousand and one kinase (TAOk) and its downstream signaling effectors. TAO kinase activates the p38 pathway and is involved in apoptosis. Developmental phenotypes have not been investigated for TAO knockout and/or transgenic mice.


(OMIM#*600136 [MLK1], *600137[MLK2], *600050[MLK3]).

Mixed Lineage Kinases (MLKs) target the JNK and p38 pathways in response to a variety of cellular stresses (Fig. 7; Gallo and Johnson,2002). There are five characterized MLKs (MLK1, 2, 3, 4, and 7). To date, there are no published knockout mouse models for MLK4 or MLK7. A MLK1/MLK2 double-deficient mouse shows no developmental abnormalities (Bisson et al.,2008). In addition, knockout mice for MLK3, a MLK closely related to MLK1 and MLK2, also exhibit normal morphogenesis (Brancho et al.,2005). However, MLK3 knockout mice have reduced JNK activation in response to TNFα (Brancho et al.,2005). B-raf/ERK activation by mitogens is also regulated by MLK3 (Chadee and Kyriakis,2004), but this is apparently dispensable during development. Of the known MLKs, MLK3 and MLK7 are detected in the adult heart (Bloem et al.,2001). MLK7 is expressed in cardiomyocytes and is activated in response to beta-adrenergic stimulation. In transgenic mice with cardiac-specific overexpression of MLK7 through the α-myosin heavy chain promoter, there is increased cardiac myocyte hypertrophy and mortality as a result of beta-adrenergic signaling. Overexpressed MLK7 activates p38 and JNK in cardiac myocytes (Christe et al.,2004). This strongly suggests a role for MLK7 in regulating cardiac function and in stress responses (Christe et al.,2004). In vitro, MLK7 is activated in response to stressors, such as anisomycin and ultraviolet irradiation, and leads to pro-apoptotic pathways through p38 and JNK (Wang et al.,2005a). Saturated free fatty acid (FFA) also causes metabolic stress and activates JNK through MLK3, MLK4, and MLK7, while JNK activity in knockout mouse models of each of these MLKs is decreased in response to FFA. The ultimate consequence of FFA-activated MLK-JNK is an increase in insulin resistance (Jaeschke and Davis,2007). Thus, the targeted regulation of MLKs may be suitable for treatment of metabolic diseases such as diabetes. In total, these findings indicate that MLKs are highly involved in stress responses and further characterization of MLKs may uncover developmental roles for these MAP3Ks.

Figure 7.

Mixed lineage kinases (MLKs) and their downstream signaling effectors. MLKs activate the p38 and JNK pathways. Particulary, MLK7 overexpressing transgenic mice exhibit increased mortality in response to beta-adrenergic agonists showing that MLK7 acts downstream of beta adrenergic activation in the heart.



Tumor progression locus-2 (Tpl2,or MEKK8) is a MAP3K that activates the ERK and JNK pathways in cells of the immune system when triggered by lipopolysaccharide (LPS) and other inflammatory mediators (Fig 8; Dumitru et al.,2000; Das et al.,2005). Knockout mice of Tpl2 develop normally, but exhibit defects in inflammatory responses (Dumitru et al.,2000). Tpl2 is also overexpressed in cancer, mainly in T-cell lymphomas, emphasizing its role in T-cell proliferation (Christoforidou et al.,2004). Tpl2 is also up-regulated predominantly in ventricular heart muscle during development compared with adult tissue (Kim et al.,1998), suggesting that it may have a dispensable role during cardiac morphogenesis. High expression of Tpl2 in cardiac myocytes suggest it may have a role in maintaining the differentiated state of these cells (Kim et al.,1998). Further studies will determine whether Tpl2 is involved in developmental or disease processes outside of the immune system.

Figure 8.

Tpl2 and its downstream signaling effectors. Tpl2 activates the ERK and JNK pathways in response to inflammatory mediators. Defects in inflammatory responses have been observed in Tpl2 knockout mice.


(OMIM#*600447, *604915).

Dual leucine zipper bearing kinase (DLK) is a MAP3K, which activates JNK during neuronal migration in cerebral cortex development (Fig 9). DLK activity is absolutely required for the creation of axon tracts (Hirai et al.,2006). A highly related MAP3K, leucine zipper kinase (LZK), is detected in the human brain, which suggests a role for these MAP3Ks in neural/brain development (Sakuma et al.,1997). Additionally, LZK complexes with JIP-1 and induces activation of JNK in vitro (Fig. 9) (Ikeda et al.,2001). In Drosophila, the protein wallenda is homologous to vertebrate DLK and LZK, and overexpression of wallenda in Drosophila results in an increase in synaptic formation due to increased JNK-Fos activation (Collins et al.,2006). Further understanding of signaling through DLK and LZK will increase our knowledge of these MAP3Ks contribution to neuronal development.

Figure 9.

Major signaling cascades activated by DLK and LZK. DLK activates the JNK pathway necessary for synaptic formation in the nervous system. Knockout mice of DLK show defects in the cerebral cortex. LZK activates the JNK pathway and may also be involved in nervous system development.


With the development of specific genetic knockout models for individual MAP3Ks, new insights have been provided in regards to how these molecules affect crucial processes during embryogenesis. Deregulation of several MAP3Ks has been associated with a wide range of developmental defects while other members of this family appear to be dispensable for the proper formation of the embryo. However, it is possible that the latter, while not exerting overt effects, may lead to increased predisposition to other diseases later in life. Therefore, further detailed studies in animal models that lack specific MAP3Ks and survive to adulthood are necessary to fully understand the overall effect of deleting these apparently unessential MAP3Ks. Conditional knockout mouse models of MAP3Ks will reveal developmental and/or homeostatic roles for these signaling proteins. Furthermore, these knockouts may be interbred to create compound-deficient animals for study into compensatory functions by similar MAP3Ks. In addition, knock-in mice, animals where the wild-type gene is replaced by a mutant form of the gene, will provide more information to the function of each MAP3K. Mutant proteins encoded by these genes may maintain scaffolding ability for protein interactions, but may have altered kinase activity or loss of a specific binding domain. For example, MAP3Ks that are dominant-negative or constitutively active for kinase activity will reveal downstream targets in vivo and phenotypic consequences for loss-of-function or gain-of-function, respectively. In this respect, knock-in mice for a kinase inactive version of MEKK4 exhibit neural tube and skeletal defects (Abell et al.,2005).

In the case of Raf kinases, it is likely that they carry distinct physiological functions as evidenced by the tissue specificity of each isoform. This is highlighted by the fact that B-Raf and not A- or C-Raf contributes to vascular patterning in the placenta. The Raf-kinases may also play redundant roles during the initial stages of embryonic development as suggested by studies using double knockout animals for B-Raf and C-Raf. Thus, many questions remain to be answered in regards to how the different Raf isoforms are spatially and temporally regulated throughout development.

For TAK1, while its requirement for vascular development has been established, the signaling mechanisms through which it exerts its effects remain unclear. TAK1 has been shown to activate the SAPK/JNK-, p38-, and TGFβ-dependent pathways as well as the NFκB cascade. One possible way to explain the wide range of effectors activated by TAK1 is that these signaling events are cell type and context specific. Another possible explanation is that TAK1 may act in concert with several other effectors, thus creating a network of signaling events to bring about specific responses. However, future signaling studies must be undertaken to properly define the complex mechanisms by which TAK1 affects embryonic development.

MEKK3 is essential to cardiovascular and heart development, as indicated by knockout mice of MEKK3 that exhibit severe cardiovascular defects and embryonic lethality (Yang et al.,2000). The MEKK3 → MEK5 → ERK5 pathway (Fig. 10) may represent an essential signaling cascade in the formation of the heart and cardiovascular system, as MEK5- and ERK5-deficient mice have similar developmental defects (Regan et al.,2002; Wang et al.,2005b). Other proteins necessary for early cardiovascular morphogenesis such as Hyaluronan synthetase 2 (Has2), ErbB2, and Heregulin-1 may eventually be connected to this pathway (Camenisch et al.,2000,2002; Meyer and Birchmeier,1995) due to phenotypic similarities with MEKK3, MEK5 and ERK5 knockout mice.

Figure 10.

Two separate MAP3K pathways direct vascular patterning. The B-Raf cascade (left-side) regulates blood vessel formation in the placenta and MEKK3 controls vessel development in the embryo proper.

MAP3Ks that may be involved in mature heart homeostasis such as MEKK1, MEKK5, and MLK7 may act as targets for late stage heart disease due to their roles in cardiac remodeling or adrenergic responses. Other MAP3Ks, such as MEKK4, TAO, and Tpl2, will need further genetic and biochemical studies to determine their functions in cardiovascular development and/or disease.

The MAP3Ks are primarily associated with activation of MAP2Ks; however, it will be of interest to examine other proteins that are regulated through phosphorylation by these kinases. It is noted that TAK1 and MEKK3 can lead to activation of the NFκB. In this regard, it is known that IKKα, an activator of NFκB, is directly phosphorylated by NIK (Ling et al.,1998). It remains to be clearly determined whether TAK1 or MEKK3 directly regulates NIK or IKK in the NFkB activation pathway. Another example is that MEKK3 can activate the calcineurin-NFAT pathway. This requires direct phosphorylation of the phosphatase, calcineurin, by MEKK3 (Abbasi et al.,2005). Future research will identify phosphorylation of additional nonclassical effectors, and define the significance of these events to development.

In addition to their activity as kinases, MAP3Ks contain scaffolding domains for protein–protein interactions. The size and sequence of these domains varies for each of the MAP3Ks. Certain MAP3Ks, such as MEKK4, use the N-terminal domain for autoregulation. When the N-terminal domain of MEKK4 interacts with its C-terminal kinase domain, then it is in a kinase inactive state. Interaction of the MEKK4 N-terminus with GADD45 disrupts this interaction with the C-terminus opening MEKK4 into the kinase active state (Mita et al.,2002). Other potentially important interactions may include protein phosphatases, such as calcineurin and Shp2 (Ptpn11). Shp2 regulates MEKK3 and MEKK4 activities (Lerner-Marmarosh et al.,2003; Halfter et al.,2005). Of interest, mutations in Shp2 are a leading cause of Noonan's syndrome and Leopard syndrome, which are characterized by craniofacial, skeletal, cardiac developmental defects and myeloproliferative disorders (Tartaglia et al.,2001; Legius et al.,2002; Araki et al.,2004; Kontaridis et al.,2006). Additional work will determine the significance of Shp2/MEKK3 and Shp2/MEKK4 interactions in development. Studying the structure–function of the scaffolding domains of MAP3K will be very important in understanding how these proteins are regulated and how they control activities of other signaling effectors.

In this review, we highlight how MAP3Ks are important to development and tissue homeostasis. We also identify two separate MAP3K circuits that direct vascular patterning (Fig. 10). A MEKK3-MEK5-ERK5 cascade appears to be a primary conduit connecting upstream factors to nuclear factors driving cardiovascular development. This MEKK3 pathway appears to specifically regulate vascular patterning in the embryo proper. The B-Raf -MEK1-ERK2 transduction route governs blood vessel patterning in the placenta. Thus, vascularization during embryonic development appears to have two distinct compartmentalized signaling cascades functioning in parallel during embryogenesis. The loss of specific MAP3Ks necessary for early cardiovascular or neuronal development leads to embryonic lethality. However, further studies will reveal whether there are subtle developmental abnormalities in select MAP3K-deficient models. While these may not cause prenatal death, they may predispose for disease later in life. This concept of developmental origins of adult disease is of great interest to many areas of molecular medicine. A prime example of this is that of MEKK1 deficiency leading to the onset of pressure overload and increased apoptosis in the heart. An enhanced understanding of these integrated MAP3K-networks during developmental processes will allow for improved strategies to treat congenital defects and disease.