Understanding Apoptotic Blocks in "Double Hit" and Other Non-Hodgkin Lymphomas By BH3 Profiling

Background: The anti-apoptotic protein BCL2 is expressed in chronic lymphocytic leukemias (CLL), 85% of follicular lymphomas (FL) and 60% of diffuse large B cell lymphomas (DLBCL). It is associated with an inferior survival when it is co-expressed with MYC, an oncogene inducing cellular proliferation, especially when MYC and BCL2 expression is caused by translocations, as in "double hit lymphomas" (DHL). Venetoclax, a selective BCL2 inhibitor, is effective in CLL, but has modest activity in other BCL2-expressing non-Hodgkin lymphomas (NHL). CLL cells are considered "primed" to die and BCL2-dependent, i.e. they are closer to the apoptotic threshold because they express high levels of pro-apoptotic BIM, which is sequestered by BCL2. Venetoclax binds to BCL2 allowing BIM to initiate mitochondrial outer membrane depolarization (MOMP) and apoptosis. We hypothesize that other NHLs may not be "primed" (e.g. low BIM), express other anti-apoptotic proteins not targeted by venetoclax (eg. MCL1, BCLXL, BFL1, BCLW) or have dysfunctional effector proteins BAX/BAK.

Methods: We performed BH3 profiling on 60 NHL and 5 tonsil primary specimens to determine 1) the degree of priming, 2) the response to venetoclax and 3) the dependence on other anti-apoptotic proteins. BH3 profiling measures MOMP by flow cytometry, using cytochrome c release as a marker, after exposure to synthetic pro-apoptotic BH3 peptides that have different affinities to anti-apoptotic proteins. We used DHL cell lines to determine if chemotherapy could "prime" cells and change the levels of anti-apoptotic proteins or MYC.

Results: Tonsilar B and T cells were "unprimed" and insensitive to venetoclax, whereas most NHLs were considered "primed": 93% of CLLs (13/14), 81% of FLs (17/21), 63% of DHL (5/8) and 47% of DLBCLs (8/17). Interestingly, 2 DLBCL samples taken at the time of relapse were unable to undergo MOMP after exposure of BIM or BID, indicating dysfunctional BAX/BAK. Defining a venetoclax response as > 50% of cells undergoing MOMP, CLLs and DHLs had the highest responses to venetoclax (7/14 and 4/8 samples, respectively) followed by FL and DLBCL (4/21) and (1/15). Among the "primed" samples, CLLs and DHLs were primarily BCL2-dependent, whereas DLBCLs and FLs depended on more than one anti-apoptotic protein. Increased MOMP was achieved when combining venetoclax to an MCL1 specific inhibitor MS1, suggesting that inhibiting more than one anti-apoptotic protein may be necessary in some samples. We tested if chemotherapy could "prime" DHL and change the levels of anti-apoptotic proteins. Of 5 DHL cell lines tested, OCI-LY8 and Toledo had the lowest responses to venetoclax. Pre-treating OCI-LY8 with vincristine increased MOMP after exposure to venetoclax, from 8.9 ± 10.82 to 98.41 ± 0.81%. In Toledo, vincristine and doxorubicin increased the venetoclax response from 3.73 ± 4.97% to 45.47 ± 28.66 and 71.74 ± 26.92%, respectively. Furthermore, vincristine decreased the MYC and MCL1 levels in both cell lines, but didn't change BCL2, BCL-XL and BCL-W levels. Doxorubicin exposure also decreased MYC, but had no effect on the levels of anti-apoptotic proteins. Exposure to cyclophosphamide, bendamustine, and steroids did not affect priming, sensitize cells to venetoclax or change levels of anti-apoptotic proteins or MYC.

Conclusions: DHL and CLL appear more sensitive to venetoclax because they are more "primed" and BCL2 dependent compared to FL and DLBCL. RCHOP and venetoclax may be effective to treat DHL because doxorubicin and vincristine could "prime" cells and decrease levels of MCL1 and MYC. However, this would not be effective in NHLs that depend on BCLXL, BCLW or have dysfunctional BAX/BAK, the latter being an uncommon feature that was primarily seen at the time of relapse.