Secondary Acute Myeloid Leukemia transformed from chronic myelomonocytic leukemia is strongly resistant to venetoclax but demonstrates sensitivity to anti-GM-CSF lenzilumab immunotherapy Free

Introduction:Secondary acute myeloid leukemia arising from prior myelodysplastic syndrome or chronic myelomonocytic leukemia accounts for 20-30% of AML cases. Published reports and our South Australian Registry revealed poorer median overall survival of sAML (3.5-3.7 months) and CR/CRi rates (0-33%) compared to newly-diagnosed previously untreated AML receiving azacitidine and venetoclax therapy (median OS 14.7 months, CR/CRi 66.4%). Decreased sensitivity to VEN in monocytic phenotypes has been reported, thus we sought to establish if sAML transformed from CMML demonstrated resistance to VEN and test alternative agents. We have recently observed a durable response to a humanized anti-GM-CSF antibody therapy, Lenzilumab (LENZ; Taran Therapeutics, NJ) in combination with AZA in patients with high blast count CMML and increased risk of AML transformation in the PREcision Approach to CHronic Myelomonocytic Leukemia clinical trial (PREACH-M; ACTRN 1262100022383). We therefore aimed to investigate if sAML may respond to LENZ and in combination with VEN-based therapies.

Method:Mononuclear cells were cultured in IMDM with 10% FBS with VEN (48 hours) or LENZ (5 days) and analysed by flow cytometry following staining with AnnexinV-PE/DAPI, CD45 and CD33 antibodies. Colony-forming assays were performed by seeding 0.5-1x10⁵ cells/mL in methylcellulose-based media without cytokines in the presence of VEN, AZA or LENZ, on their own and in combination. Colonies were enumerated after 10-14 days incubation.

Results:6 of 7 sAML (CMML>AML) tested showed VEN IC₅₀ greater than 500 nM (range 545-3314 nM; mean IC₅₀ 1855.3±517.9 nM vs. de novo NPM1 mutant AML 11.8±1.0 nM; P=0.04) confirming strong resistance to VEN inhibition and consistent with clinical experience. We assessed the colony-forming ability of sAML and found that 4 of 6 were able form colonies in the absence of exogenously supplied cytokines. Crucially, 4 of 4 sAML that were able to form spontaneous colonies were inhibitable by 20 µg/mL LENZ (mean 67.1±7.8% reduction vs. IgG, P=0.003) suggestive of an autocrine source of GM-CSF driving sAML proliferation targetable by LENZ. We developed a short-term liquid culture assay for in vitroLENZ sensitivity testing with rapid turnaround time and found that the viability of CD45+ mononuclear cells (excluding B and T lymphocytes) was significantly reduced following 5 days treatment with LENZ (56.3±10.6% reduction vs. IgG, P=0.006). This result was consistent with the colony-forming assay (65.1±14% reduction vs. IgG; P=0.0001) of the same patient sample. We further tested the efficacy of LENZ compared to VEN and AZA as monotherapies and in combination, as well as to a triple combination therapy of VEN/AZA/LENZ. Colony suppression by LENZ was significantly greater than VEN or AZA alone or in combination (52.7±7.0 colonies vs. VEN 120.0±23.6; P=0.0004 or AZA 126.0±23.1; P=0.0002, VEN/AZA 98.7±17.2; P=0.007), but comparable to VEN/AZA/LENZ triple therapy. Based on this data, we treated a patient with sAML transformed from CMML (resistant to VEN/AZA after 8 cycles) with VEN/AZA/LENZ triple therapy (100 mg PO VEN D1-D28, AZA 75 mg/m² D1-D7, LENZ 552 mg D1 and D15 given as a 35 day cycle). The subject showed durable response following 10 cycles of triple therapy and remains on treatment with clinical improvement in weight (73 to 76 kg) and quality of life with no acute admissions over 11 months. Triple therapy was well tolerated and significantly, the subject remained transfusion independent throughout all 10 cycles. Bone marrow blast percentage decreased from 20% at baseline to 7% after cycle 3 and 13% after cycle 6. Improvements in platelet count and hemoglobin were observed (26, 87, 64, 119 x10⁹/L and 73, 92, 89, 88 g/L respectively, evaluated at baseline, cycle 3, 6 and 10) and lactate dehydrogenase levels remained normal.

Conclusion: The ability of secondary AML to form colonies may be driven by an autocrine source of pro-inflammatory cytokine production, revealing a vulnerability that can be targeted by precision immunotherapy. We report for the first time that venetoclax resistant secondary AML transformed from CMML show sensitivity to antibody-directed GM-CSF blockade. Indeed, encouraging real-world clinical data of a patient with secondary AML safely receiving Lenzilumab in combination with standard-of-care venetoclax/azaciditine therapy for AML points towards a possible advancement in the management of venetoclax resistant AML.