Brk/Protein Tyrosine Kinase 6 Phosphorylates p27KIP1, Regulating the Activity of Cyclin D–Cyclin-Dependent Kinase 4

Cyclin D and cyclin-dependent kinase 4 (cdk4) are overexpressed in a variety of tumors, but their levels are not accurate indicators of oncogenic activity because an accessory factor such as p27Kip1is required to assemble this unstable dimer. Additionally, tyrosine (Y) phosphorylation of p27 (pY88) is required to activate cdk4, acting as an “on/off switch.” We identified two SH3 recruitment domains within p27 that modulate pY88, thereby modulating cdk4 activity. Via an SH3-PXXP interaction screen, we identified Brk (breast tumor-related kinase) as a high-affinity p27 kinase. Modulation of Brk in breast cancer cells modulates pY88 and increases resistance to the cdk4 inhibitor PD 0332991. An alternatively spliced form of Brk (Alt Brk) which contains its SH3 domain blocks pY88 and acts as an endogenous cdk4 inhibitor, identifying a potentially targetable regulatory region within p27. Brk is overexpressed in 60% of breast carcinomas, suggesting that this facilitates cell cycle progression by modulating cdk4 through p27 Y phosphorylation. p27 has been considered a tumor suppressor, but our data strengthen the idea that it should also be considered an oncoprotein, responsible for cyclin D-cdk4 activity.

Cyclin D1 Maintains Mantle Cell Lymphoma Through CDK4-Independent Regulation Of DNA Replicative Checkpoints

Mantle cell lymphoma (MCL) is rarely curable and therapy resistance often leaves few viable treatment options for patients. Previous studies have identified the importance of cyclin D1 (CCND1) translocation and overexpression in MCL pathogenesis, which leads to increased cyclin-dependent kinase 4 (CDK4) activity and accelerated cell cycle progression. However, targeting this abnormal cell cycle control, mainly through CDK4 inhibition causes only G1-phase growth arrest without significant cell death (Marzec et al. 2006). In contrast, prolonged inhibition of CCND1 with RNA interference induces apoptosis in MCL cell lines (Weinstein et al. 2012), suggesting an essential function of CCND1 independent of CDK4 activity. The mechanism of this non-catalytic role of CCND1 in maintaining MCL cell survival is largely unknown. To clarify the cell cycle role of CCND1 in addition to its CDK4-dependent function, we compared the effects of CCND1 and CDK4 silencing on MCL cell survival. MCL cell lines co-expressing GFP and doxycycline-inducible shRNA targeting CCND1 or CDK4 were generated. Cells with similar GFP expression levels were FACS sorted to normalize for shRNA expression. Both CCND1 and CDK4 silencing resulted in G1-phase arrest, but only CCND1-silenced cells demonstrated a marked increase in apoptosis. Investigation of the potential cause of apoptosis revealed significant accumulation of DNA double-strand breaks following CCND1 ablation, as measured by nuclear gamma-H2AX focus formation. Interestingly, CCND1-silenced cells exhibited a significant increase in 53BP1+ nuclear bodies in G1-phase, reminiscent of 53BP1 foci observed by Lukas and colleagues in cells undergoing aphidicolin-induced replication stress (Lukas et al. 2011). Analysis of replication fork movement in CCND1-depleted cells showed substantially reduced fork speed and increased frequency of origin firing, both of which are indicative of replication stress. In contrast, knockdown of CDK4 did not result in slower forks or increase in the frequency of origin firing. Genomic instability associated with replication stress was also apparent in CCND1-silenced cells, including increased micronucleus formation and recurrent chromatid gaps or breaks detected by cytokinesis-block assay and karyotyping, respectively. Analysis of DNA replicative and damage checkpoints revealed that both ATR-CHEK1 and ATM-CHEK2 pathways were activated by phosphorylation following CCND1 silencing in MCL cell lines, a xenograft animal model, and primary tumor samples, but not in non-MCL tumors. Interestingly, this activation (with the exception of ATM phosphorylation) was unsustainable over time and did not cause down-regulation of the downstream targets CDC25 and CDK1/2 but, instead, we observed an increase in CDC25A/B protein levels and CDK1/2 activity, indicating defective cell cycle checkpoints. Exposing CCND1-silenced cells to replication stress-inducing or DNA-damaging agents such hydroxyurea, aphidicolin, etoposide or ionizing radiation further amplified the checkpoint defects seen in unperturbed cells. We did not observe any significant difference in this checkpoint signaling in control and CDK4 knockdown cells under these conditions. Furthermore, CCND1-deficient cells were more sensitive to pharmacological inhibition of ATR and CHEK1 but not ATM, confirming a constitutive role of CCND1 in the ATR-CHEK1 pathway. In conclusion, these studies revealed an unexpected CDK4-independent role of CCND1 in maintaining DNA replicative checkpoints to prevent replication stress and genome instability in MCL cells. As most cancer treatments rely on agents that create DNA replication stress, targeting this function of CCND1 could provide a rational approach to overcome resistance to conventional therapies in MCL. Disclosures: No relevant conflicts of interest to declare.