In this report we describe a novel role for ACF in the regulation of apoB
mRNA export from the nucleus and the importance of this process in animal models for modulating the expression of ApoB protein. The function of ACF to date has been limited to its role as an auxiliary factor required for apoB
mRNA editing in mammalian small intestine and in the liver of many species [49
]. Our new findings provide an intriguing potential hypothesis for why ACF expression has been retained in human and non-human primate livers despite the loss of APOBEC-1 expression and lack of apoB
mRNA editing activity [49
An important implication of the findings from this study was that ACF may function early in the regulatory pathways that control lipoprotein synthesis. Prior to our study, research on ACF focused on its role in apoB
mRNA editing [50
] and studies of ApoB focused on lipoprotein assembly and secretion [51
]. Early work from this lab showed that ACF trafficking was regulated by insulin-dependent serine phosphorylation [27
]. We reasoned that Ob/Ob mice might enable an evaluation of ACF trafficking and ApoB expression in the context of an insulin-resistant disease model. We observed that the levels of hepatic acf
message and protein were markedly increased. This change stood out for the magnitude of change in acf
mRNA levels relative to that of related RNA binding proteins.
The progression to obesity in Ob/Ob mice begins immediately after birth due to hyperphagia. This manifests in an increased abundance of adipose tissue, altered response to insulin stimulation and exposure of the liver to increased levels of serum insulin, adipokines (leptin [42
] and resistin [43
]), pro-inflammatory cytokines (Il-6 and TNFα [41
]), and free fatty acid flux [44
]. In testing several of these factors individually, under the defined conditions of primary hepatocyte culture, we found that only leptin had the ability to alter acf
mRNA abundance, causing its suppression in a dose-dependent manner. Supporting a role for leptin in ACF expression was the finding that insulin resistance and obesity were not sufficient to change ACF expression in the Db/Db mice that retained leptin signaling in their livers.
The regulation of ACF abundance by leptin is only the second example of hormonal regulation of acf
message. Previously, thyroid hormone was shown to increase the abundance of ACF in neonatal Pax8−/−
]. Treatment with thyroid hormone resulted in a mobilization of hepatic triglycerides, which the authors hypothesized was due to increases in the proportion of ApoB48 availability [12
]. It is of interest that the PPARα agonist, ciprofibrate, also reduced ACF abundance in Low Density Lipoprotein Receptor (LDLR) knockout animals [52
]. In these studies, the amount of edited apoB
mRNA varied in direct proportion with ACF abundance. In contrast, we observed no significant changes in the basal level of apoB
mRNA editing in Ob/Ob mice relative to that of the lean control. These findings together with the data showing that Ob/Ob hepatocytes failed to up-regulate ACF retention in the nucleus underscore the importance of insulin for controlling the function of ACF in the lean physiological state.
Our studies employed two animal models for obesity. Ob/Ob mice express leptin receptors, but do not express leptin and consequently become hyperphagic and obese. Db/Db mice express leptin but lack the leptin receptor in their CNS and become hyperphagic and obese. Obesity led to insulin resistance in both animal models. We found that hepatic ACF nuclear retention and apoB
mRNA editing activity were insulin-resistant in both animal models. Db/Db hepatic acf
mRNA and ACF protein abundances were not elevated above control whereas Ob/Ob acf
mRNA and ACF protein levels were markedly elevated. We attribute this difference to the expression of leptin and leptin signal in Db/Db liver but not in the Ob/Ob liver [45
]. There are several splice variants of the leptin receptor that differ in the length of their cytoplasmic domain required for docking of Janus kinase and STAT proteins, through which leptin signals [53
]. Signaling through the long form of leptin receptor (Ob-Rb) is unlikely to have contributed to acf
mRNA suppression as only the truncated leptin receptor (Ob-Ra) is expressed [46
]. The Ob-Ra receptor lacks the consensus box 1 and 2 motifs utilized for docking of Janus kinase and therefore its signaling does not go through a STAT3 dependent pathway [55
]. Signaling through the Ob-Ra receptor has been described through MAPK activation [55
]. Functional Ob-Ra receptor signaling has been shown through induced expression of c-fos, c-jun
, and jun-B
]. Leptin signaling in the Db/Db liver may have remained responsive and therefore continued to the suppression of acf
mRNA abundance. The regulatory factors that normally modulate ACF expression will be an important area for future research.
We can not rule out that over expression of ACF may have compounded the defect in regulating ACF and apoB
mRNA nuclear export and ApoB translation in the Ob/Ob mouse but our data suggested that insulin resistance was the primary reason for why ACF nuclear retention and apoB
mRNA editing activity were not elevated above the lean control values as would have been anticipated given the hyperinsulinemic condition in these animals. Considering that ACF was hyper-phosphorylated on serine residues in Ob/Ob liver and phosphoACF was restricted in the nucleus, PKC activity on nuclear ACF and factors determining nuclear retention must have retained a basal level of function. However, the possibility remains that insulin resistance involved reduced activation of PKC and/or heightened protein phosphatase I activity. This might explain why the radioactivity per μg of ACF protein in Ob/Ob nuclei was only 2-fold higher than that observed in the lean control when an earlier study showed that nuclear ACF specific radioactivity increased approximately 5-fold in lean mice following a 10 nm insulin injection [27
An equally plausible explanation for our findings is that nuclear protein kinase C activity may have become insufficient for the level of ACF phosphorylate necessary to affect nuclear restriction because this pathway was less responsive to insulin. Given that apoB mRNA editing efficiency was suppressed in Ob/Ob and Db/Db mouse liver compared to lean control liver, our study suggested that the insulin-dependent processes that otherwise regulate ACF assembly into editosomes may have been compromised by the inability to restrict adequate amounts of ACF to the nucleus.
In published studies and in this study, the bulk of total cellular ACF was recovered with cytoplasmic extracts from whole liver or hepatocytes [17
]. Immuno electron microscopy showed that cytoplasmic ACF was predominantly localized along the exterior of the endoplasmic reticulum [25
]. This is also the intracellular site of apoB
mRNA translation where microsomal triglyceride transfer protein facilitates co-translational lipid loading onto nascent ApoB [57
]. Differential centrifugation of liver cytoplasmic extracts yielded microsomal fractions in which the rough endoplasmic reticulum (high density microsomal fraction, HDM), smooth endoplasmic reticulum and Golgi stacks (low density microsomal fraction, LDM) and residual cytoplasm were separated [59
]. The bulk of nascent ApoB protein co-fractionated with the LDM fraction [31
]. Using this protocol, we found that most of the cytoplasmic ACF was not recovered as a soluble protein with the residual cytoplasm fraction but rather, was enriched in the LDM and HDM factions. This characteristic of ACF was conserved in mice and rats. In fact, despite the high level of ACF expression in Ob/Ob mouse liver, the association of ACF with microsomes was maintained. The relative recovery of ACF with the Ob/Ob HDM fraction compared with the LDM fraction, suggested that there was more ACF in the HDM fraction of obese mice compared with HDM of lean control mice. This association is consistent with the hypothesis that ACF maintained it association with apoB
mRNA following nuclear export to the sites of ApoB protein synthesis in the cytoplasm. The importance of ACF abundance and trafficking for ApoB production was underscored by the finding that ApoB secretion was inhibited through RNAi knockdown of ACF. An additional correlation between ACF trafficking and ApoB synthesis and secretion was that nuclear retention of ACF through PP1 inhibition was associated with reduced cytoplasmic apoB
mRNA and increased nuclear apoB
mRNA in primary hepatocytes. Consistent with our hypothesis, intracellular and secreted ApoB protein was reduced in primary hepatocytes treated with Cantharidin. We propose therefore that the inability of the liver in obese animals to appropriately regulate nuclear retention of ACF is what led to dysregulation of ApoB.
More detailed analysis will be necessary to de-convolute the regulation of ACF trafficking in the insulin-resistant state. Future studies will focus on developing genetic models to determine the role of ACF nuclear and cytoplasmic trafficking in mRNA export from the nucleus and identify the signaling pathways that regulate ACF expression and post-translational modification. Specifically, this study has demonstrated using animal models a role for leptin signaling in the regulation of ACF expression and shown that deficiencies in nuclear retention of ACF was part of the obesity-induced, insulin-resistant phenotype. Importantly our studies suggested that independent of the function ACF has in apoB mRNA editing activity (resulting from its interaction with APOBEC-1), ACF may have a primary function as an apoB mRNA recognition factor and chaperone. This function suggested a new hypothesis that may explain why ACF continues to be expressed in nonhuman primate and human livers even though apoB mRNA editing activity is no longer possible. It is also possible that ACF serves to chaperone other mRNAs that are important for development and it is the loss of this function that accounts for why ACF knockouts were not viable. It will be important to develop a conditional ACF knockout animal model to address this possibility.