|Home | About | Journals | Submit | Contact Us | Français|
Autophagy plays a critical role in the initiation and progression of tumors. The exact nature of this role, however, is complex. Autophagy is suppressive to tumor initiation, and reduces genomic instability. Genes with key roles in autophagy are mutated in human cancer, and knock-out mice for certain autophagy genes are predisposed to cancer. Conversely, established tumors appear to utilize autophagy in order to survive periods of metabolic or hypoxic stress. Consistent with this, small molecule inhibitors of autophagy like chloroquine are effective anti-cancer agents for certain tumor types. The consensus appears to be that autophagy suppresses tumor initiation, but promotes the survival of established tumors. But this premise may be over-simplified. Several groups have recently shown that the ARF tumor suppressor can induce autophagy. While some groups have found that ARF-mediated autophagy is cytotoxic to tumor cells, we have shown that ARF’s autophagy function may promote the survival and progression of certain tumors. We have previously shown that silencing ARF limits autophagy and the development of p53-null lymphomas. In this addendum, we show this is not true for primary p53-null sarcoma cells. Rather, ARF-silencing enhances sarcoma development. These data suggest that the survival-benefit of ARF, and possibly also of autophagy, may be restricted to certain tumor types.
There is ample biochemical and genetic evidence to indicate that autophagy is suppressive to tumor initiation. Autophagy acts to suppress tumor development by removing damaged organelles and reducing chromosome instability.1–3 Heterozygous knock-out mice for the autophagy gene beclin 1 are predisposed to multiple tumor types, including lymphoma and liver cancer. In tumors from these mice, the wild-type allele of beclin 1 is not lost, so beclin 1 is a haplo-insufficient tumor suppressor gene.4,5 beclin 1 is also mono-allelically deleted in a subset of tumors of the breast, ovary and prostate.5,6 Two Beclin 1 binding proteins that both play important roles in the induction of autophagy are UVRAG and Bif-1; knockout mice for these proteins are predisposed to multiple spontaneous cancers.7,8 Like Beclin 1, UVRAG is mono-allelically mutated in human cancer.7 The combined data indicate that autophagy suppresses tumor initiation. It is interesting to note, however, that autophagy genes are never bi-allelically deleted or mutated in cancer, suggesting that a basal level of autophagy may be incompatible with life, or may be required for tumor survival. Consistent with this latter premise, there are also compelling data that the pathway of autophagy promotes the survival of established tumors.
Autophagy allows cells to recoup ATP and essential building blocks for biosynthesis when they are starved of nutrients or when they are exposed to environmental stresses such as hypoxia. As tumor cells are exposed to such conditions frequently, it is perhaps not surprising that several groups have found that established tumors may rely on autophagy in order to survive. Specifically, the autophagy inhibitors chloroquine and 3-methyladenine are very effective anti-tumor drugs for Burkitts lymphoma and chronic myelogenous leukemia.9–11 Additionally our group and others found that silencing beclin 1 inhibits tumor growth in vivo.12,13 The combined data indicate that autophagy promotes the survival of established tumors. A useful consensus might be that inhibiting autophagy can lead to tumor formation because this process normally removes damaged mitochondria and misfolded proteins, which would otherwise lead to increased reactive oxygen species and mutagenic damage. In contrast, established tumors encounter increasing metabolic stress, and autophagy is critically necessary to survive such episodes.
Some of the complexities of the role of autophagy in tumor development are exemplified by recent studies on the role of the p14ARF tumor suppressor (p19ARF in mouse, and hereafter referred to as ARF). ARF is a bona fide tumor suppressor gene; it is frequently mutated in human cancer, it suppresses tumor development when overexpressed, and germ-line mutations in ARF that target the ARF reading frame but not that of the overlapping p16ink4a gene are present in familial cancer kindreds (for a review see ref. 14). A key critical tumor suppressor function of ARF is mediated by its ability to signal to and activate p53 following mutational activation of oncogenes like Myc and Ras. It is also clear that ARF has tumor suppressor functions that are independent of p53.14 Recently we and others have demonstrated that ARF can induce autophagy, in a p53-independent manner.15–17 Two of these groups found that transfected ARF induces autophagy and is cytotoxic, raising the possibility that ARF’s autophagy role may contribute to its p53-independent tumor suppressor function.
We recently reported that ARF’s autophagy function may be protective for a subset of human tumors.13 Because ARF is potently transcriptionally repressed by p53, many tumors and cell lines in which p53 is deleted or mutated express high levels of ARF.18 We showed that silencing ARF in p53-null mouse embryo fibroblasts results in reduced autophagy and impaired survival in response to nutrient deprivation.13 In contrast, silencing ARF in these cells had no effect on cell proliferation or survival under nutrient-rich conditions. We found that ARF is markedly upregulated in response to nutrient deprivation, consistent with a role for this protein in autophagy. Notably, we showed that silencing ARF in myc-driven lymphoma cells with mutant p53 inhibits autophagy and impairs the progression of these tumors in tail-vein injected mice. These data suggest that ARF has a previously undiscovered tumor-promoting role that is possibly mediated by its autophagy role. Consistent with this premise, ARF is highly expressed in many human tumors that contain mutant p53, and in up to 40% of Burkitts lymphomas.19 These observations seemed to contradict the well-documented function of ARF as a tumor suppressor.
Is it possible that ARF is a regulator with two opposite personalities: a well-known tumor suppressor in one subset of neoplasia and an emerging tumor promoter in another? In addressing this question it is important to note that our data indicate that ARF’s autophagy role is likely to be relevant only to tumors with mutant or deleted p53. What remains unclear is whether ARF-silencing, and the concomitant inhibition of autophagy, can suppress the development of all tumors. Whereas we found that two different B-cell lymphoma lines from the Eμ-myc mouse, as well as a T cell lymphoma from the p53 knockout mouse, all survived more poorly in tail vein-injected mice when ARF was silenced, this study was limited to lymphomas. More recently we isolated a primary sarcoma cell line from p53 knockout mice. Interestingly, ARF-silencing in this tumor cell line did not suppress the development of this tumor when implanted intra-perintoneally into immuno-deficient mice (Fig. 1A). ARF-silencing did reduce autophagy in this cell line, however, as evident by an accumulation of p62/SQSTM1 (data not shown) and the processed form of LC3, LC3 II (Fig. 1B). The analysis of tumor weights from five different control and ARF-silenced tumors indicated that the tumors with ARF-silenced were significantly larger (p=0.008) (Fig. 1C). The above data suggest that the impact of ARF-silencing, and possibly also autophagy inhibition, on tumor development might be cell type- or tumor-specific. These data may also speak to an issue that has confused researchers in the p53 field, regarding the differences in cancer phenotype between p53-null and ARF-null mice.
The majority of p53-null mice develop T cell lymphoma. In contrast, the majority of ARF-null mice develop poorly differentiated sarcomas and carcinomas, with a much lower incidence of lymphomas.20 The reason for this difference in tumor spectrum between p53-null and ARF-null mice has been unclear. Our data suggest that lymphomas may be particularly sensitive to the requirement for autophagy, and thus the high levels of ARF in tumors from p53-null mice may favor the development of lymphoma. Likewise, the lower incidence of lymphoma in ARF-null mice could be attributed to the lack of an essential tumor-promoting activity of ARF in the development of this tumor. In sum, we speculate that the differences in tumor spectrum between ARF-null and p53-null mice may be influenced by differences in the sensitivity of these cell types to the need for autophagy. This premise is highly speculative, however, given the small number of tumor cell lines we have analyzed, and the presently correlative nature between ARF levels and autophagy.
An alternative (but not mutually exclusive) hypothesis also exists. It is possible that distinct tumors have different requirements for the survival function of autophagy, depending upon the driving oncogenic pathway in that tumor. For example, high autophagy might benefit a Myc-driven tumor, perhaps because of the impact of Myc on cell metabolism, but be of limited consequence to the survival of a Ras-driven tumor. To begin to address this issue, we analyzed a series of tumor cell lines for their sensitivity to silencing beclin 1, a key autophagy gene. Examples from two of these lines are depicted in Figure 1D. As depicted in this figure, silencing of beclin 1 markedly enhances the plating efficiency of Saos2 cells. In contrast, silencing beclin 1 markedly impairs the plating efficiency of the U20S osteosarcoma cell line. Neither of these cell lines express ARF, so this effect is ARF-independent. While Saos2 cells are null for p53, and U20S cells have wild-type p53, this effect of beclin 1 silencing is likewise p53-independent; our data in other cell lines indicate this effect is independent of p53 status. Clearly we need to now learn which tumor cell lines are sensitive to autophagy inhibition, and which are not. Because autophagy modulators like chloroquine, rapamycin, and rapamycin analogs are now in the clinic, this becomes an important next issue to be addressed.
We thank Shengkan Jin (University of Medicine and Dentistry of New Jersey, Robert Wood Johnson Medical School) for the Beclin 1 short hairpin and Amanda Frank (Fox Chase Cancer Center) for the p53-null sarcoma cell line. This work was supported by National Institutes of Health Grants CA102184 and CA080854 (Murphy).
Addendum to: Humbey O, Pimkina J, Zilfou JT, Jarnik M, Dominguez-Brauer C, Burgess DJ, Eischen CM and Murphy ME. The ARF tumor suppressor can promote the progression of some tumors. Cancer Research 2008; 68:9608-13. PMID: 19047137