Potent Efficacy of BCL2 Inhibition with ABT-199 in High-Risk Aggressive B-Lymphoma Models When Combined with Knockdown of MCL1

Avoiding apoptosis is a hallmark of cancer. Diffuse large B-cell lymphoma (DLBCL) is the most common lymphoma and carries a poor prognosis in cases at high-risk of failing up-front R-CHOP (rituximab, cyclophosphamide, doxorubicin, vincristine, and prednisone). Frequent expression of the anti-apoptotic protein BCL2 is well-described in DLBCL in numerous studies and is clear a negative prognostic marker when co-expressed with the oncogenic transcription factor c-MYC. Expression of the anti-apoptotic protein MCL1 also is found in about half of cases. BCL2 and MCL1 have redundant function in protecting cells from apoptosis.

Direct inhibitors of MCL1 are not clinically available, but its short half-life permits knock-down through inhibition of cyclin-dependent kinase 9 (CDK9), which regulates transcriptional elongation. Older multi-CDK inhibitors have anti-tumor activity from loss of MCL1 but are not approved clinically due to off-target toxicities. Dinaciclib is a more potent and specific multi-CDK inhibitor with activity against CDK9. We tested dinaciclib against a panel of >20 DLBCL cell lines and found high potency, with IC₅₀ < 20 nM in most lines. Both in vitro and in vivo, dinaciclib results in rapid loss of MCL1 protein and corresponding induction of apoptosis. Interestingly, both sensitive and resistant lines show loss of MCL1 in response to the compound. Thus, we hypothesized BCL2 activity compensates for loss of MCL1 in resistant lines. Correspondingly, over-expression of BCL2 in sensitive cells renders them completely insensitive to dinaciclib without effecting MCL1 knockdown. In 59 DLBCL cases with known BCL2 status, we assessed MCL1 protein by immunohistochemistry and found no significant difference in MCL1 expression between BCL2 positive (66%, 10/15) and negative (57%, 25/44) cases (p=0.5576).

Expression of MCL1, BCL2, or both in DLBCL and the proteins’ redundant function led to the hypothesis that knockdown of MCL1 combined with direct BCL2 inhibition would synergize in the killing of high-risk DLBCL tumors. ABT-199 is a third-generation BH3 mimetic direct inhibitor of BCL2, which has shown remarkable clinical activity in chronic lymphocytic leukemia but less activity in DLBCL and other more aggressive lymphomas. We found ABT-199 combines potently and synergistically with dinaciclib in DLBCL cell lines with none of 23 lines resistant to the combination. We confirmed this in vivo using the line U2932, which is resistant in vitro to both drugs as single agents. U2932 xenografts showed dramatic reduction of tumor burden in response to the combination, a response far superior to either drug alone. We next evaluated a genetically defined immunocompetent mouse model of MYC-BCL2 double-hit lymphoma, based on MYC expression in the VavP-Bcl2 transgenic model, replicating the genetics, pathology, and aggressive clinical behavior of the human disease. Tumors from this model in vitro, interestingly, show little response to single-agent ABT-199, but the combination with dinaciclib is again synergistic. Treatment of tumor-bearing mice in vivo showed animals treated with either drug alone had no significant survival difference from vehicle-treated controls, while those treated with the combination had dramatically improved survival by Kaplan-Meier analysis (p<0.0001).

Finally, we assessed the effect of combining ABT-199 with standard lymphoma chemotherapy drugs that are thought to affect MCL1 protein levels due to global effects on transcription. Doxorubicin, etoposide, and cytarabine all result in loss of MCL1 at peak in vivo attainable concentrations and synergize with ABT-199 to kill DLBCL cells otherwise resistant to the single agents. In sum, we propose therapeutic strategies combining direct inhibition of BCL2 with knockdown of MCL1 expression will be effective and tolerable for poor-prognosis lymphomas such as high-risk DLBCL and double-hit lymphoma.