“Triple” Combination of Targeting Methyltransferase, BCL-2 and PD-1 Facilitates Therapeutic Responses in Acute Myeloid Leukemia (AML)

Background

The long-term survival of AML, especially of older AML patients, under the established therapies remains poor. A recent breakthrough therapy of selective BCL-2 inhibitor venetoclax with hypomethylating agents (HMA) azacitidine or decitabine is approved to treat older and unfit patients with AML (DiNardo et.al. EHA 2020). This combination induces responses in > 65% of older AML patients, but the responses in relapsed/refractory AML patients are sub-optimal. Upregulation of programmed death-1 (PD-1) in T-cells is associated with AML immune-suppression, and combination of anti-PD-1 with azacitidine has activity in relapsed/refractory AML (Daver et. al. Cancer Discovery 2019). Recent pre-clinical findings indicate that venetoclax preserved T-cells immunity by sparing central memory T-cells and enhanced activity of anti-PD-1 antibodies in immune-competent mouse models in vivo (Mathew et.al. Blood 2018). In this study, we tested the hypothesis that PD-1 inhibition facilitates the depth and duration of response to venetoclax and HMA combination by reactivating T-cells mediated immunity against AML.

Results

We collected peripheral blood (PB) samples from trial patients before and after receiving the treatment of decitabine and venetoclax (Dec_Ven) (NCT03404193) and examined the percentage of AML blasts and T-cells by multi-parameter flow cytometry. The combination of Dec_Ven effectively reduced PB CD33+/CD34+ cells from 46 ± 15% at baseline (BL) to 27 ± 16% at 2.3 days (Day 1 – 3) (time point 1, or T1), and 15 ± 13% at 5.5 days (Day 3 – 8) (T2), and increased CD3+ T-cells: 28 ± 10% (BL), 54 ± 13% (T1) and 59 ± 19 % (T2), (BL vs.T1, CD33/CD34+, p = 0.001; CD3+, p = 0.02, n = 6). The reduction of CD33+/CD34+ positively correlated with the depletion of circulating blasts (R = 0.64, p = 0.0001).

Further analyzing the T-cells subsets among CD4 helper cells and CD8 cytotoxic cells, we found that Dec_Ven therapy promoted an activated T-cells phenotype by upregulating CD69 and PD-1 in both CD4 and CD8 cells in early time point samples, and led to T-cells exhaustion, as indicated by persistent expression of PD-1 at later time-point when CD69 was undetectable (PD-1 in CD3+CD4+ cells: 13 ± 2%, n = 6 (BL), 15 ± 3%, n = 6 (T1) and 22 ± 7%, n = 5 (T2); PD-1 in CD3+CD8+ cells: 14 ± 3%, n = 6 (BL), 27 ± 5%, n = 6, (T1), 29 ± 9%, n =5 (T2), BL vs. T1, p = 0.05). The persistent expression of PD-1 is a hallmark of T-cells exhaustion and immunosuppression (Jia et.al. Blood Cancer J 2018). Mass cytometry CyTOF immune-profiling of various compartments of CD4 and CD8 cells on PB uncovered that Dec_Ven reduced naïve T-cells (CD45RA+CCR7+) but spared or even increased frequencies of the central (CD45RA-CCR7+) and effector memory (CD45RA-CCR7-) CD4 or CD8 cells.

In parallel, we measured the serum cytokines levels, and found that treatment of Dec_Ven increased IL2, IL6 and INFg secretion in PB. IL2 and IL6 are two cytokines that regulate the proliferation and differentiation of effector memory T-cells; and INFg is a critical mediator of treatment-induced immune response.

To further test if targeting PD-1 induced anti-leukemia effect and amplified the effect of Dec_Ven, we treated leukemia cells collected before and after Dec_Ven therapy with anti-CD3 and anti-PD-1 antibodies in vitro. Treatment reduced CD33+/CD34+ cells in 2 of 3 pre-treatment samples (untreated vs. anti-CD3_PD-1: 55% vs. 15%, pt #6; 70% vs. 52%, pt #7) and synergistically reduced/eliminated the residual CD33+/CD34+ cells in AML samples obtained after Dec_Ven therapy (untreated vs. anti-CD3_PD-1: 84% vs. 66% Day 3, pt #3; 18% vs. 8% Day 3, pt #6; 44% vs. 35% Day 1, 5% vs. 2% Day 4, pt #7) (Fig.1).

Conclusions

Our data indicate that Dec_Ven therapy induces PD-1 expression on T-cells but selectively preserves potentially important anti-tumor T-cells immune subsets. The “triple” combination therapy of HMA with BCL-2 and PD-1 antagonists may facilitate anti-leukemia responses by enhancing T-cells function. Clinical trial to test this hypothesis is ongoing in relapsed/refractory AML patients (NCT02397720).