The G1 phase CDKs, (CDK2, CDK4, and CDK6) display specific functions in terms of development and redundant functions in terms of cell cycle regulation [23, 24, 83]. The redundant nature of the G1 phase CDKs is best illustrated by work showing that cells can proliferate in the absence of CDK2 and CDK4, and that CDK1 activity is required for cell cycle progression [84, 85]. However, injured hepatocytes do not require CDK1 to proliferate . Another important function of the CDKs is in DNA repair. Specifically, cyclin D1, CDK4 [58, 59, 87], cyclin A1, and CDK2 have been implicated in DNA repair [88–92]. In fact, the activities of all CDKs are required for efficient DNA repair . Even though CDK2 and CDK4 are largely dispensable in the entry and progression through S phase, they are important mediators of mammary tumorigenesis in mice, as ablation of either CDK2 or CDK4 can prevent mouse mammary tumorigenesis triggered by HER2/Neu [94, 95]. Also, CDK2 and CDK4 can differentially mediate centrosome amplification depending on the oncogenic or cellular context . While cyclin D1/CDK4 specifically mediates centrosome amplification triggered by H-RasG12D, H-RasG12D&c-Myc , and in HER2+ breast cancer cells , silencing of cyclin E or CDK2 have no impact on centrosome amplification in the HER2/Ras systems. In contrast, both CDK2 and CDK4 mediate centrosome amplification and chromosome instability in p53 deficient cells .
Overall, the ability of CDK2 and CDK4 to mediate various malignant phenotypes that play important roles in cancer biogenesis has prompted the development of CDK2 or CDK4 specific inhibitors. Those inhibitors have proven to be effective in a variety of breast cancer cell lines. For example, ER+ luminal cells that overexpress cyclin D1 and Rb and display downregulated p16 respond well to cell cycle inhibition by the CDK4/CDK6 specific inhibitor PD-0332991, while non-luminal/basal cells are refractive to growth inhibition . It has been reported that the Rb status dictates the early responses to that inhibitor . Even though the CDK4/CDK6 inhibitor has shown promise in preclinical trials, cells acquire resistance through the upregulation of CDK2 activity . A major limitation of PD-0332991 is that while it is effective in extinguishing mammary tumors, its ability to block the cell cycle interferes with chemotherapy [99, 100].
In the present study, we addressed whether the G1/S cell cycle regulatory machinery influences radioresistance in MCF10A, ER-PR-HER2-, and ER-PR-HER2+ cells. There is a precedent for the role of ectopic CDK2 and CDK4 activities in imparting resistance or sensitivity to radiotherapy. For example, knockdown of cyclin E synergizes with doxorubicin to enhance radioresistance in breast cancer cells lines . Induction of cyclin A has previously been observed in cells treated with UV or irradiation [91, 102]. Cyclin A null cells are radiosensitive and display impaired double strand break repair; cyclin A/CDK2 is involved in DNA repair following irradiation by phosphorylating KU-70 .
Various mechanisms have been postulated to explain how cyclin D1/CDK4 leads to radioresistance or radiosensitivity of various cells. Inducible expression of cyclin D1 in MCF7 breast cancer cells leads to radiosensitivity through activation of the p53 pathway . Low dose irradiation promotes the free, cytoplasmic cyclin D1 accumulation in human keratinocytes, correlating with radioresistance; in this context, low level radiation disrupted the interaction of cyclin D1 with 14-3-3ζ . Cytoplasmic cyclin D1 then interacts with Bax, suppressing the ability of Bax to induce apoptosis. In another example, long term fractionated irradiation of the human cancer cell lines HepG2 and HeLa induced radioresistance . This radioresistance correlated with upregulation of cyclin D1 due to the stabilization of cyclin D1 by prevention of its proteolysis, achieved by the DNA PK/AKT/GSK3β pathway. In addition, they first proposed that ectopic expression of cyclin D1 leads to increased radioresistance in breast cancer cells by enhancing DNA repair in the radioresistant cells, as downregulation of cyclin D1 in this system abrogated enhanced DNA repair, resulting in radiosensitivity. This group showed using chemical inhibitors that cyclin D1 dependent radioresistance is reversed by AKT or CDK4 chemical inhibitors. A recent report also demonstrated that enhanced cyclin D1 results in enhanced DNA repair and radioresistance . This report postulates that cyclin D1 promotes DNA repair by interacting with RAD51 within DNA repair foci; this interaction is promoted by radiation.
Our studies demonstrate that knockdown of CDK4 acts as a potent radiosensitizer, independently of the breast cancer molecular subtype. In our studies, we tested whether radiosensitization was a consequence of cell cycle inhibition by silenced CDK4, altered DNA break repair, or through the activation of an apoptotic program. We used various assays to show that knockdown of CDK4 did not influence the cell cycle of non-irradiated or irradiated cells. This is consistent with our previous observations that in MCF10A cells expressing control pLKO.1, H-RasG12V, H-RasG12V&c-Myc, or in HER2+ breast cancer cells, knockdown of CDK4 does not greatly influence DNA replication or cell cycle progression [76, 97]. This is also consistent with our data showing that ablation of CDK4 does not alter the cell cycle of p53−/− mouse embryonic fibroblasts . In contrast, co-inhibition of CDK4 and CDK6 using a chemical inhibitor interferes with the cell cycle, leading to protection against chemotherapy [99, 100]. Our studies conducted with the PD0332991 CDK4/CDK6 inhibitor revealed a similar story, as treatment of MCF10A or MDA-MB-231 protected against radiation-induced apoptosis. This is contrary to our studies revealing that knockdown of CDK4 synergizes with radiation to induce apoptosis in breast cancer cells and MCF10A controls. The only cell in which radiation synergized with inhibition of CDK4/CDK6 was MDA-MB-468, a cell that is insensitive to growth inhibition by PD0332991. However, we do not think that insensitivity to growth inhibition is the sole cause of the sensitivity to apoptosis triggered by PD0332991 treatment, as we treated all cell lines exactly at IC50 to prevent complete growth inhibition. One potential explanation for the differences between knockdown of CDK4 and inhibition with PD0332991 is that while the chemical inhibits CDK4/CDK6 activity, the shCDK4 may disrupt the protein:protein interactions of CDK4 with cyclin D, p21, p27, p15, p16, p18 and p19, leading to modulation of the activities of other CDKs.
In contrast to studies showing that increased cyclin D1 activity results in radioresistance through increasing cellular DNA repair capacity, all breast cancer cells tested in this study efficiently repaired DNA breaks irrespective of CDK4 levels, as indicated by the efficient clearance of H2AX foci after radiation. A recent study is consistent with our results, as chemical inhibition of CDK4 and CDK6 in breast cancer cells did not influence double-strand break repair after irradiation . We speculate that the differences between cyclin D1 and CDK4/CDK6 inhibition in relation to DNA repair is because cyclin D1 displays functions that are independent of CDK4. For example, it can bind the promoters of genes involved in chromosome instability .
The presence of increased levels of cleaved caspase-3 and of cleaved PARP in all irradiated cells down regulated for CDK4 indicated that apoptosis was a significant driver of radioresistance in mammary epithelial and breast cancer cells. Although apoptosis is high in unirradiated MCF10A cells silenced for CDK4, levels of apoptosis rise significantly in those and in the breast cancer cells silenced for CDK4 upon irradiation. This suggests synergy between downregulation of CDK4 and irradiation in regards to apoptosis. To establish how absence of CDK4 cooperates with radiation to enhance cell death, we screened various Bcl family members. No changes in the levels of phosphorylation of the antiapoptotic members Bcl2 or Mcl1 were observed in the cells. In contrast, levels of phosphorylated Bad ser136 were sharply diminished Even though dephosphorylation of Bad in serine 136 is associated with apoptosis, we observe increased phosphorylation in irradiated shCDK4 MCF10A cells. The regulation of apoptosis is complex, involving various protein/protein interactions, as well as the post-translational modification of numerous proteins. Further studies must be performed in order to establish why MCF10A cells silenced for CDK4 display phosphorylated Bad Ser 136, and still apoptose. Bad ser136 phosphorylation is reversed by various phosphatases, including PP2A and PP1α [81, 82]. In fact, silencing of CDK4 resulted in higher levels of PP2A, suggesting that it specifically dephosphorylates Bad ser136 upon irradiation PP2A is a protein complex with many cellular functions, including the regulation of apoptotic and mitogenic pathways [104–106], as well as the modulation of DNA repair . Inhibition of PP2A by DNA tumor viruses is tightly linked to cellular transformation . Also, the dephosphorylation of Bcl2 by PP2A promotes its anti-apoptotic activity; this event may enhance Bcl2’s oncogenic potential . In wide contrast, PP2A’s proapoptotic activities have been mapped to its ability to dephosphorylate Bad [81, 82] and Bax . We did not observe significant dephosphorylation of Bcl2 in any of the cell lines, suggesting that upregulation of PP2A in breast cancer cells exclusively dephosphorylates Bad ser136 without significantly affecting other Bcl family members. Thus, our observation that downregulation of CDK4 resulted in decreased phosphorylation of Bad ser136 suggests that this event is primarily responsible for apoptosis and radioresistance. Additional experiments are required to address how CDK4 regulates PP2A levels; that regulation may result from increased protein stability, increased degradation, or enhanced de-repression of E2F target genes upon CDK4 silencing. Another important future area of exploration is whether and how the apoptotic program triggered by downregulation of CDK4 in conjunction with radiation is rooted in Rb/E2F dependent transcription.
Unfortunately, we were unable to show that ectopic expression of PP2Ac radiosensitizes breast cancer cells, or that its inhibition modulates radiation induced apoptosis in breast cancer cells. We conclude that apoptosis in cells knocked down for CDK4 is independent of PP2A. Further experiments are required to establish pathways signaling apoptosis in irradiated breast cancer cells knocked down for CDK4.