This report is the first to demonstrate PPARβ/δ-dependent inhibition of cell proliferation in HRAS-expressing cells by increasing G2
/M arrest. This is consistent with previous work that showed inhibition of skin carcinogenesis and inhibition of proliferation in keratinocyte cell lines with Hras
). The current studies revealed that induction of G2
/M arrest caused by ligand activation of PPARβ/δ specifically targets cells with higher expression of HRAS. That the PPARβ/δ-dependent induction of G2
/M arrest causes selection against cells with higher expression of HRAS is consistent with results showing that cells treated with chemicals that induce G2
/M arrest also cause selection against cells expressing higher levels of HRAS. Interestingly, human cancer cell lines expressing oncogenic RAS were previously shown to be more sensitive to mitotic perturbations than normal cells (36
), an observation also noted in the present studies. For example, as HRAS activity increased in keratinocytes (as the result of viral transduction) or 308 cells (with an activating mutation in HRAS), the efficacy of GW0742 with respect to induction of G2
/M arrest increased. Moreover, ligand activation of PPARβ/δ inhibits mitosis in skin tumors, and this phenotype is also associated with reduced expression of HRAS in these tumors. It was also shown that cells expressing oncogenic RAS exhibit a disadvantage with respect to cell proliferation following knockdown of many mitotic genes (36
). This is important because PPARβ/δ-dependent repression of many of these mitosis-related genes was also observed in HRAS-expressing keratinocytes, 308 cells, and skin tumors in the present study.
Among the mitosis-related genes that were repressed by ligand activation of PPARβ/δ in HRAS-expressing cells, Cdk1
are of great interest. While some of these changes were relatively modest, this could have been due to the presence of an endogenous high-affinity agonist that prevents alterations in expression of greater robustness. An active cyclin B1-CDK1 complex is a trigger to enter mitosis, whereas depletion of cyclin B1-CDK1 can cause a block in mitosis concomitantly with repeated rounds of S phase, leading to cells with polyploidies in both fission yeast and human cells (11
). CHEK1 is required for spindle checkpoint function (specifically at the tension-sensing branch of the checkpoint), and Chek1−/−
cells can exit mitosis in the presence of paclitaxel and undergo endoreduplication, leading to polyploidies (66
). Ligand activation of PPARβ/δ decreased expression of CDK1 and CHEK1 in HRAS-expressing cells, and the observed phenotype, including delayed entry into mitosis, retarded exit from mitosis, and increased polyploidy cell numbers, is similar to the phenotype of Cdk1
-null and Chek1
-null cells. Combined, these observations suggest that the PPARβ/δ-dependent decrease in expression of CDK1 and CHEK1 alone in HRAS-expressing cells may largely underlie the observed mitosis block following ligand activation of PPARβ/δ.
Inhibition of cell cycle kinetics induced by ligand activation of PPARβ/δ could have been due in part to direct regulation of target genes by PPARβ/δ, which was not examined in the present study. However, results from these studies also establish that PPARβ/δ-dependent inhibition of mitosis in cells with an activating Hras mutation can also inhibit cell cycle progression and is mediated by a mechanism (H) that involves (i) PPARβ/δ directly binding with p107/p130 proteins; (ii) translocation of PPARβ/δ to the nucleus in response to ligand activation, leading to increased nuclear hypophosphorylated p130 and p107; (iii) ligand bound-PPARβ/δ maintaining p130 in a hypophosphorylated state; and (iv) heightened nuclear p107/p130 causing increased recruitment of the p130/p107/E2F4 complex to the promoters of mitosis-related genes and inhibition of their transcription, i.e., of genes with repressor E2F4 binding sites (such as Cdk1 and E2f1) that are repressed directly by this complex. Because of this PPARβ/δ-dependent downregulation of E2F1 expression, decreased E2F1 recruitment to promoters of other genes preferentially regulated by activator E2Fs (such as Chek1) is secondarily influenced by this regulation. Thus, it is not surprising that expression of CHEK1 is also downregulated by ligand activation of PPARβ/δ. Alternatively, it remains possible that the PPARβ/δ/p130/p107/E2F4 complexes exhibit differential affinities for binding sites on chromatin or that their presence leads to differences in recruitment of transcriptional corepressors. Further studies are needed to examine these ideas. Since cells with RAS mutations are more sensitive to mitotic perturbations than normal cells, the present studies focused more on the regulation of mitosis genes. However, it is also noteworthy that expression of many E2F target genes involved in DNA replication and DNA repair was also reduced by ligand activation of PPARβ/δ. This change in gene expression is reflected by the decrease in cells in the S phase in HRAS-expressing keratinocytes following ligand activation of PPARβ/δ.
The interaction between p107/p130 and PPARβ/δ is independent of HRAS. Moreover, PPARβ/δ preferentially binds to the hypophosphorylated form of p130 based on data from coimmunoprecipitations. This conclusion is also supported by results showing the more prominent colocalization of p130 and PPARβ/δ in the cytosol of cells, since hypophosphorylated p130 was found primarily in the cytoplasm. In the presence of ligand, PPARβ/δ may inhibit phosphorylation of p130 by CDK4. This suggests that, when p130 is shuttled to the nucleus via random nuclear translocation of PPARβ/δ under normal conditions, p130 is phosphorylated by the CDK4/cyclin D1 complex present in the nucleus, thus losing its repressor activity. This also suggests that ligand activation of PPARβ/δ is essential for repression of mitosis genes, because it maintains p130 in a hypophosphorylated state and chaperones hypophosphorylated p130 into the nucleus in cells with activated HRAS signaling.
Nuclear translocation of PPARβ/δ in HRAS-expressing cells following ligand activation is central to the inhibition of mitosis genes. Nuclear translocation of PPARβ/δ and increased nuclear p107 and p130 levels in normal keratinocytes are typically not observed. In contrast, increased nuclear translocation of PPARβ/δ and concurrent increases of p107 and p130 levels in HRAS-expressing keratinocytes and skin tumors illustrate the essential nature of nuclear translocation of PPARβ/δ following ligand activation in the presence of activated HRAS signaling. The increase in both cytosolic and nuclear PPARβ/δ levels following ligand activation in control keratinocytes is most likely due to the stabilization of the receptor rather than to an increase of protein synthesis and nuclear translocation for the following reasons. First, no increase in the level of PPARβ/δ mRNA was observed following ligand activation. Second, ligand activation of PPARβ/δ is known to prevent its ubiquitin-mediated degradation, thus increasing its half-life (17
). However, the mechanism of nuclear translocation of PPARβ/δ in HRAS-expressing keratinocytes, skin tumors, and confluent 308 cells following ligand activation remains unclear. It is possible that increased HRAS activity activates downstream kinases and alters the phosphorylation status of PPARβ/δ, leading to its nuclear translocation.
/M arrest of cells expressing high levels of HRAS can be achieved by ligand activation of PPARβ/δ, and cotreatment of a PPARβ/δ ligand with various mitosis inhibitors enhances the efficacy of increasing G2
/M arrest. This supports the hypothesis that combining ligand activation of PPARβ/δ with mitosis inhibitors is a feasible approach for treating tumors that express higher levels of RAS. Indeed, oncogenic RAS signaling is increased in a number of human cancers, including lung, colon, pancreas, and melanoma (55
). While the role of PPARβ/δ in some cancers remains controversial (reviewed in references 45
, and 50
), the body of evidence suggesting that PPARβ/δ protects against cancer is increasing. For example, a recent compelling study demonstrated that colorectal cancer patients with relatively low expression of PPARβ/δ in the primary tumor were ~4 times as likely to die from this disease as patients with relatively higher expression of PPARβ/δ in their primary tumors (64
). It is also not disputed that ligand activation of PPARβ/δ inhibits chemically induced skin carcinogenesis (3
). Moreover, preclinical and clinical studies have also shown that ligand activation of PPARβ/δ inhibits or prevents metabolic syndrome, obesity, dyslipidemias, glucose intolerance, and chronic inflammation, characteristics that are positively associated with cancer (43
). Since targeting single molecules for chemoprevention and chemotherapy has not proven highly effective (21
) due in part to the genetic heterogeneity associated with diseases (54
), targeting PPARβ/δ in conjunction with mitosis inhibitors could become a suitable option for development of new multitarget strategies for inhibiting RAS-dependent tumorigenesis.