Characterizing the Anti-Apoptotic Dependencies of T-Cell Prolymphocytic Leukemia Identifies HDAC and JAK/STAT Pathway Inhibitors As Promising Combination Partners to Augment Bcl-2 Targeted Killing By Venetoclax

Introduction:

Response to conventional therapies for patients with T-cell prolymphocytic leukemia (T-PLL) is usually poor and is associated with short survival. The BCL-2 antagonist venetoclax was recently found to have some clinical activity in this disease (B Boidol et al., Blood, 2017); however, these early data suggest that this drug will not provide prolonged response when given as monotherapy. Several other drug classes have demonstrated preclinical activity in T-PLL, including HDAC inhibitors (HDACi), JAK/STAT inhibitors (JAK/STATi), and TCR pathway inhibitors (TCRi), particularly ITK inhibitors. To determine which drug(s) may be the optimal combination partner(s) for venetoclax in T-PLL, we utilized a functional approach known as BH3 profiling. This assay measures how close a cell is to the threshold of apoptosis ("priming") and identifies which anti-apoptotic proteins a cell depends on for survival. We also utilized a variant known as dynamic BH3 profiling (DBP) to measure early changes in pro-apoptotic signaling after various drug treatments.

Methods:

Clinically annotated primary T-PLL patient samples were obtained from the French Innovative Leukemia Organization network after informed consent. Peripheral blood mononuclear cells were isolated by Ficoll and viably frozen and later thawed for the experiments. Baseline BH3 profiling to measure cytochrome C (cyto-C) release was performed as per Ryan et al., Methods, 2013, and DBP as per Montero et al., Cell, 2015. Viability was assessed by AnnexinV/Hoechst staining. Ex vivo drug treatments included: BH3 mimetics (BCL-2i: venetoclax (VEN), MCL-1i: AZD5991, S63845), HDACi (belinostat = BEL), JAK/STATi (ruxolitinib = RUX) and TCRi (PRN694 = PRN). Protein expression was assessed by standard Western Blot. Primary CLL cells were used in some experiments as a comparator. To mimic the lymph node microenvironment, DBP and viability assays were performed in co-culture with the stromal cell line NK.tert. Tumoral DNA was also extracted, and we performed NGS on a panel of 29 genes, including ATM and TP53, as well as Sanger sequencing to assess for IL2R, JAK1, JAK3, STAT5B mutations. Statistical analyses were by unpaired and paired t-test with a two-tailed nominal p ≤ 0.05 considered as significant.

Results:

Samples were evaluated from 31 T-PLL patients. Baseline BH3 profiling revealed that, compared to CLL cells, T-PLL cells are less primed for apoptosis but have comparable dependency on MCL-1. BCL-2 dependency was found to be significantly lower in T-PLL than CLL (cyto-C release 48.8%; 62.7% p=0.0005), and to decrease further in the presence of stroma (Figure A, cyto-C release from 72.6% to 36.2%, p = 0.01). Consistent with our BH3 profiling results, the degree of BCL-2 dependency in T-PLL cells was strongly associated with the amount of apoptotic cell death induced by VEN (R² -0.58, p=0.004), whereas MCL1 dependency was strongly associated with the cell death induced by the MCL1 inhibitors S63845 and AZD5991 (R² -0.59, p=0.002 and R² -0.68, p=0.0005 respectively, Figure B).

We next performed DBP to assess the changes in apoptotic priming in T-PLL cells induced by HDACi, JAK/STATi and TCRi. To utilize doses similar to what can be achieved in patients, we assessed BEL 1mM, RUX 1mM and PRN 1mM. BEL and RUX increased overall T-PLL cell priming and BCL2 dependency (delta cyto-C release of 26.8%, p=0.004 and 14.8%, p=0.01 respectively Figure C), with no effect on MCL1 dependency. PRN had no significant effect on priming. Consistent with the DBP data, our viability assays showed that BEL and RUX induced significantly more cell death when combined with VEN compared to PRN (Figure D). Mutations in ATM, TP53, and JAK/STAT pathway genes were observed in cells from 35%, 6%, and 53% of patients, respectively, and did not impact the ex vivo activity of these drugs.

Conclusion:

We report the first data for BH3 profiling in T-PLL. We found that this disease is heterogeneously dependent on both BCL-2 and MCL-1, and that the lymph node microenvironment may decrease BCL-2 dependency. HDACi and JAK/STATi both enhance BCL-2 dependence, thereby sensitizing T-PLL cells to VEN. Ongoing studies will help further define the mechanism underlying these promising new combinations for T-PLL.

View largeDownload slide

View largeDownload slide

Close modal