These results indicate that ARC, previously recognized as a regulator of cell death solely in terminally differentiated cells (e.g. cardiac myocytes), promotes multiple aspects of breast carcinogenesis. This conclusion is supported by experiments in both genetically manipulated mice and xenografts of human breast cancer cells. Importantly, our studies emphasize genetic loss of function to provide insight into the importance of endogenous ARC levels.
Our previous discovery that ARC abundance is markedly increased in human breast cancers is recapitulated in mammary tumors of MMTV-PyMT transgenic mice (). Based on prior studies (18
), we presume that ARC levels are increased, in part, through upregulation of Ras signaling by the PyMT transgene (35
). The persistence of ARC from primary tumors through metastases in this model is in accordance with our results demonstrating involvement of ARC in pathogenesis at multiple stages. In fact, ARC expression is even higher in the invading cell population than in corresponding primary mammary tumors of PyMT mice (unpublished data, S. Goswami and J. Condeelis).
Onset and multiplicity of PyMT tumors are unaffected by the loss of ARC suggesting that ARC is not necessary for tumor initiation. Rather, our data indicate that ARC strongly influences tumor growth (). Despite its known function as an apoptosis inhibitor, the absence of ARC does not significantly increase apoptosis in this system (). Although the explanation for this is not clear, we hypothesize the presence of redundant cell death inhibitors. Surprisingly, the absence of ARC reduces cellular proliferation (). This proliferation function of ARC has not been previously described, although we have recently observed that ARC knockdown/knockout also impairs the proliferation of benign MCF-10A breast epithelial cells (unpublished data, C. Medina-Ramirez and R. Kitsis) and hypoxic pulmonary arterial smooth muscle cells in vivo
). The mechanism of ARC-induced cellular proliferation is not known. While the ability of ARC to inactivate p53 may be involved (11
), this explanation is less likely in PyMT tumors in which activated Akt (35
) would be predicted to have already decreased p53 levels through MDM2-mediated degradation (36
ARC does not influence the kinetics with which lung metastases develop in the PyMT model. Rather, the number of lung metastases in PyMT mice lacking ARC is markedly reduced suggesting decreases in the efficiency of this process. The mechanisms by which ARC affects invasion and metastasis appear complex. Absence of ARC reduces local invasion, blood burden (reflecting the balance between intravasation and extravasation), and metastasis by similar amounts (40–44%) (), implying that the most proximal event, invasion, may play the rate-limiting role. Of note, significant effects of ARC on invasion were observed in all systems studied (PyMT (), met-1 (), and LM2 ()). While the mechanisms by which ARC facilitates invasion are unclear, emerging evidence suggests that other cell death regulators (e.g., procaspase-8, IAPs, Bcl-2) may promote invasion by effects independent of their apoptosis functions (37
). Procaspase-8, in particular, promotes migration and metastasis through interactions with a focal adhesion kinase (FAK) complex (40
). Since ARC and procaspase-8 interact (8
) we tested the possibility that the effects of ARC on invasion involve procaspase-8, but found no evidence in support of this model (K. Bacos and D. Stupack, not shown). In addition to the effects of ARC on invasion, administration of cancer cells by tail vein to bypass invasion and intravasation steps still demonstrates that ARC increases the efficiency of lung metastasis (). A large proportion of injected cells are known to die rapidly in the circulation and/or the lung parenchyma (42
). This lack of persistence limits the ability of these cells to form macrometastases. Accordingly, it is likely that ARC confers a post-intravasation survival advantage that increases the efficiency lung metastasis. These data suggest that ARC regulates invasion and metastasis at multiple steps.
The effects of ARC on tumor growth, invasion and metastasis observed in the PyMT model were reproduced using LM2 xenografts (). The significance of this is two-fold. First, these experiments translate our findings in mouse cells to a humanized model of breast cancer. Second, these data indicate that ARC exerts cell autonomous effects in mammary carcinogenesis, a conclusion that could not be drawn from PyMT mice that lack ARC in non-tumor, as well as, tumor cells.
We have shown that ARC confers resistance to killing by doxorubicin using human Hs578T (12
) and LM2 () cells. The LM2 xenografts provide a useful preclinical model to test the contributions of ARC to chemoresistance in vivo
. Endogenous levels of ARC by themselves are sufficient to render cells of the primary tumor resistant to doxorubicin-induced death (). In this situation, the anti-apoptosis function of ARC is clearly involved. The increased chemoresistance of the invading cell population (32
) poses an particularly challenging barrier to effective treatment. Our experiments, in which cells captured in the process of invading are tested for their sensitivities to doxorubicin and tamoxifen, demonstrate that ARC is a critical factor in the chemoresistance of this population ().
In conclusion, ARC promotes multiple aspects of breast carcinogenesis including tumorigenesis, invasion, metastasis, and chemoresistance. Given these multiple roles, ARC may provide a particularly potent target for future therapies directed against breast cancer.