The present study, showing that disruption of the ARF6 gene leads to almost complete lethality starting at midgestation (Table ), demonstrates that ARF6 is essential for mouse development. The most obvious defect of ARF6−/− embryos is in liver development. The primary effect of ARF6 deficiency on liver development appears to be interference with hepatic cord formation (Fig. ). This observation is supported by the successful modeling of the abnormal phenotype in an in vitro culture system for hepatic cord-like structure formation as induced by HGF using fetal hepatocytes (Fig. ). Moreover, ARF6 in cultured fetal hepatocytes is activated in response to HGF stimulation (Fig. ). Collectively, these results lead us to conclude that ARF6 functions at least in part as a signaling molecule in the HGF-dependent signaling pathway coupled to hepatic cord formation.
Although the molecular mechanism by which ARF6 regulates the cascade of signaling required for hepatic cord formation remains to be clarified, it is plausible that ARF6 functions by regulating membrane morphology and/or cell migration through controlling actin cytoskeletal reorganization, which is a well-known role for it in model systems (1
). This idea is derived from the observation that elongation and branching of the E10.0 hepatic epithelial sheets were abnormal in the ARF6−/−
embryonic liver (Fig. ). Alternatively, membrane trafficking regulated by ARF6 might also be critical for hepatic cord formation, as suggested by reports that ARF6-regulated internalization of E-cadherin enhances the motility of epithelialized cells (14
) and that the level of E-cadherin decreases in migrating fetal hepatocytes during liver development (24
Disruption of HGF-dependent cord-like structure formation in ARF6−/−
fetal hepatocytes in the in vitro assay system was not complete (Fig. ), suggesting that HGF utilizes another signaling system, as well as the ARF6-mediated signaling pathway, to couple HGF signaling to cord-like structure formation. If this is true and HGF is the critical physiological factor for hepatic cord formation, then disruption of hepatic cord formation in ARF6−/−
embryos should be less severe than that in HGF
knockout embryos. However, the severity of the liver developmental defect in ARF6−/−
embryos was almost the same as that reported for HGF
knockout embryos (21
). These observations lead us to speculate that, in addition to HGF, other hepatic cord formation-promoting factors most likely exist. Supporting this assumption, it has previously been reported that epidermal growth factor and transforming growth factor β also induce hepatic cord-like structure formation in vitro (13
). In addition, it has been reported that transforming growth factor β type III receptor-deficient mouse embryos display defects in the ultrastructure of the liver (25
). Finally, previous reports demonstrating that ARF6 is absolutely required for these types of growth factor-induced cell functions in other settings (4
) also support the hypothesis described above.
In the present study, we demonstrated that the smaller sizes and hypocellularity of the ARF6−/−
embryonic livers were attributable to the progression of liver cell apoptosis (Fig. ); proliferation of fetal hepatocytes in vitro and in vivo was not impaired by ARF6 deficiency (Fig. and ), inconsistent with prior reports that ARF6 is essential for cytokinesis (22
). This apparent discrepancy suggests the existence of compensatory or redundant mechanisms that promote cytokinesis not only in hepatocytes but in most if not all other cell types. Moreover, we would emphasize that the induction of apoptosis in the ARF6−/−
embryonic liver, which is not cell lineage specific (Fig. ), does not seem to be the primary consequence of ARF6 deficiency. This conclusion is supported by the observation that abnormal liver architecture (Fig. ), but not apoptosis (Fig. ), was observed at earlier stages (E10.0) of embryonic liver development, and cultured ARF6−/−
fetal hepatocytes or hematopoietic cells prepared from E13.5 embryos did not exhibit increased apoptosis (Fig. ). Instead, we would propose that the induction of apoptosis of fetal hepatocytes and of hematopoietic cells that would normally come to coreside in the cord niche might be attributable secondarily to an aberrant fetal liver microenvironment that is unsupportive for liver cell survival, due to incomplete hepatic cord formation. Such an aberrant microenvironment could cause the activation of caspase 3, resulting in the induction of apoptosis, although the details of this death response remain to be elucidated.
embryos were frequently found to be anemic, which might be responsible for the mid- to late-gestational lethality (data not shown). Considering that we observed progressive and substantial apoptosis of hematopoietic cells in the embryonic liver (Fig. ), it is quite reasonable to speculate that anemia caused in ARF6−/−
embryos is attributable to progressive apoptosis of hematopoietic cells that is triggered by the defect of liver formation, as was suggested for HGF
knockout mice (21
). We cannot rule out, however, an additional effect on hematopoiesis as well. Nonetheless, the lethality observed for ARF6−/−
embryos during mid- to late gestation appears attributable to the observed defects in liver formation. This idea is consistent with the observation that there was some variation in the stage of lethality of ARF6−/−
embryos (Table ), since variation was also observed in the severity of the liver formation defect (data not shown).
In conclusion, this report provides evidence for the first time that ARF6 physiologically functions in liver development by regulating hepatic cord formation. In addition to this function, ARF6 may also be involved in other physiological and pathological events, such as development and functions of postnatal tissues and metastasis of tumor cells that, like fetal hepatocytes during liver development, require actin cytoskeletal reorganization. ARF6 could also be involved in many other settings, e.g., pathogenesis of the Vibrio cholerae
bacterium-induced diarrhea through the promotion of ADP-ribosylation of Gsα (10
) by cholera toxin in intestinal epithelial cells. To clarify these functions, it will be necessary to generate conditional knockout mice, since ARF6
-null mice invariably exhibit embryonic lethality.