The ERK pathway
Engagement of the gp130
cytokine receptor subunit by the LIF generates two intracellular signalling
pathways: on one hand, the JAK-STAT3 pathway, which is required for ES
self-renewal, and, on the other hand, the ERK pathway. Surprisingly and in
contrast to most cultured cell lines, undifferentiated ES cells do not require
the ERK pathway for normal cell cycle, proliferation and self-renewal [20-22].
In
fact, inhibition of this pathway actually promotes self-renewal of murine ES
cells. These observations
are in agreement with the study of K-Ras-/- ES cells, the absence of
this gene leading to downregulation of the ERK pathway. These cells display
LIF-independent capacity to grow undifferentiated [23]. Therefore, it appears that the
self-renewal signal downstream of LIF is a finely tuned balance of positive
(via STAT3 proteins) and negative (via the ERK pathway) effectors [24].
The dominant role of the ERK pathway becomes
apparent upon differentiation, both in vivo and in vitro.
Interfering with the ERK signalling pathway, for example by knock-out of the
upstream activator Grb2, leads to inhibition of primitive endoderm [25, 26] and trophectoderm formations [27]. These in vivo observations have
also been reproduced in cultures of the corresponding deficient ES cells.
Surprisingly, activation of the ERK pathway by an oncogenic Ras also interferes
with extraembryonic endoderm differentiation of embryonic stem cells [28]. It is likely that a critical level of
activation of the ERK pathway is necessary for this early commitment and that up
or down variations of this threshold level have deleterious consequences.
Beside these early functions in commitment, not much was known about the role
of ERK in ES cell differentiation. Recently, our laboratory showed that the
treatment of ES cells by retinoic acid, which is required for induction of
neurogenesis and adipogenesis, activates the ERK pathway. Inhibition of ERK
activation, using specific chemical inhibitors during this period, results in a
strong inhibition of adipocyte formation without affecting neurogenesis [29]. Furthermore, by studying knock-out
animals and fibroblasts, the role of ERK in adipogenesis appears to be limited
to ERK1, with no role for ERK2 in this process [30]. To gain genetic evidence of the role of
ERK1 in adipocyte differentiation of ES cells and to analyse its function in
other cell lineages, it would be of interest to generate and test the
differentiation capacities of ERK1-/- ES cells.
Interestingly, while ERK1-/- mice
are viable and fertile [31], disruption of ERK2 is embryonic lethal
due to defective placenta formation, trophectoderm and mesoderm
differentiations [32, 33]. These findings confirm the direct role of the ERK
pathway early during the embryonic development. Furthermore, they
demonstrateERK1 and ERK2 have distinct biological functions.
While, as expected, ERK2
disruption does not interfere with proliferation of undifferentiated ES cells,
no apparent mesoderm-derived lineages can be observed upon ES commitment (B.
Binetruy and F. Bost, personal observation), suggesting that ERK2 is necessary
at an early step of ES cell commitment. Since ERK-/-mice present normal mesoderm
differentiation -except for adipocyte formation-, it is likely that the defect
of ERK2-/- ES cells in mesoderm commitment takes place earlier than
the defective adipogenesis of ERK1-/- cells.
Many defined ERK
substrates are transcriptional regulators (reviewed in [34]), but what specific molecular mechanisms are
controlled by each of the ERK isoforms during mesoderm formation remains to be
determined. Regarding
adipocyte differentiation, the positive role of the ERK pathway is blurred by
works demonstrating that the adipocyte-specific transcription factor PPARg is a substrate of ERK and that this
phosphorylation decreases its transcriptional activity and inhibits adipocyte
differentiation [35, 36]. Owing to the known role of ERK in cell
proliferation, one could reconcile these contradictory results by hypothesizing
that the function of ERK in adipogenesis is dictated by the window of time
during which it is activated. Early in the program, ERK activity is increased
for a proliferative step, while later, it must be inactivated to prevent PPARg phosphorylation. This model is supported
by the fact that the expression of MKP-1, the phosphatase that inhibits ERK, is
augmented in mature adipocytes [37].
No comments