Venetoclax combination therapy induces deep AML remission with eradication of leukemic stem cells and remodeling of clonal haematopoiesis

Venetoclax combination therapy induces deep AML remission with eradication of leukemic stem cells and remodeling of clonal haematopoiesis Romain Vazquez, Claire Breal, Loria Zalmai, Chloe Friedrich, Carole Almire, Adrien Contejean, Sylvain Barreau , Eric Grignano, Lise Willems, Benedicte Deau-Fischer, Patricia Franchi, Marguerite Vignon, Justine Decroocq, Rudy Birsen, Lauriane Goldwirt, Sophie Kaltenbach, Lucile Couronne, Michaela Fontenay, Olivier Kosmider, Didier Bouscary and Nicolas Chapuis 1,3

Dear Editor, Treatment of acute myeloid leukemia (AML) in older patients remains challenging 1 . Promising results recently emerged in patients treated with the BCL-2 inhibitor venetoclax associated with low dose cytarabine (LDAC) or hypomethylating agents (HMAs) [2][3][4] . This clinical benefit may be due to leukemic stem cells (LSCs) eradication 5,6 . The association is well-tolerated with manageable toxicity but different mechanisms of resistance 7-10 and heterogeneous responses in accordance with molecular patterns were observed 11 . We report here our single institution experience of venetoclax combination therapy for 19 consecutive patients with previously untreated AML ineligible to intensive chemotherapy. Clonal hematopoiesis and LSCs follow-up monitored by high throughput sequencing (HTS) and multiparametric flow cytometry (MFC) were analyzed.
Characteristics of patients are reported in Supplementary  Tables 1 and 2 Fig. 1).
Twelve (63%) patients received a 400 mg daily dose of venetoclax whereas 4 (21%) and 3 (15.8%) patients were treated with a dose of 200 mg and 100 mg, due to concomitant posaconazole administration from the beginning of the treatment. For 11 patients (57.9%), the starting dose of venetoclax was decreased when posaconazole was introduced and/or in accordance with monitoring of the venetoclax plasma concentration. After dose adjustment, plasmatic concentration was always in line with expectations ( Supplementary Fig. 2).
Although well-tolerated, the association was difficult to maintain over a prolonged period, probably due to the fragility of the underlying haematopoiesis. Indeed, the median number of completed venetoclax cycles for patients achieving CR/CRi was 2,5 (range, 1-17). Venetoclax was discontinued for 11 patients (57.9%) after a median of 2 completed venetoclax cycles (range, 1-8) because of hematologic toxicity, leading to azaciditine monotherapy. Nine patients (56.3%) relapsed in a median time of 3.9 months (range, 1.3-8.9) and five of them were still treated at this time with venetoclax. Interestingly, the DOR of patients who stopped venetoclax was not significantly decreased [9 months (95% CI, 8.9-NR) vs 3.7 (95% CI, 2.3-NR); p = 0.0916, Supplementary Fig. 3], suggesting that early venetoclax discontinuation did not increase the relapse risk. In terms of overall survival (OS), the median follow-up was 10.7 months (range, 1.8-20.4) and the median OS for all patients was NR (95% CI, 8.6-NR) (Fig. 1b). The median OS for adverse AML cases was significantly decreased [10.1 months (95% CI, 5.3 -NR) vs NR; p = 0.0244, Fig. 1b]. Eight patients (42.1%) died in a median time of 7.2 months (range, 1.8-11.9) due to disease progression.
Minimal residual disease based on LAIP identification (LAIP-MRD) was quantified by MFC at each BM evaluation during follow-up (Supplementary Figs. 4 and 5). When patients achieved CR/CRi, the LAIP-MRD was negative in 9 of the 11 patients tested (81.8%) (Fig. 2a) attesting therefore to the depth of the response. Given that venetoclax specifically targets LSCs, we monitored LSCs of 14 patients during treatment. LSCs became undetectable during follow-up in 12 patients (85.7%) and all of them achieved CR/CRi (Fig. 2b). Disappearance of LSCs was observed from the first course of venetoclax combination therapy in eight patients. On therapy, LSCs-MRD and/or LAIP-MRD were evaluable for 13/16 patients who obtained a CR/CRi. For four of the six patients with durable response, the MRD remained always negative (Fig. 2c). In contrast, for 6/7 patients who relapsed, the treatment regimen led to an initially negative MRD that became detectable again during follow-up (Fig.  2d). Finally, the DOR for patients with at least one positive MRD assessment at any time during therapy (n = 9) was significantly shorter [4.6 months (95% CI, 3.7-NR) vs NR; p = 0.0378, Fig. 2e]. The OS was also decreased for these patients but without statistical significance (p = 0.2) (Supplementary Fig. 6). These results indicate therefore that venetoclax combination therapy induces deep responses and eradicates LSCs.
Clonal hematopoiesis was also investigated by HTS for all patients achieving CR/CRi. The mutation patterns detected at AML diagnosis were compared to the molecular features harbored by 6/7 patients with a prior MDS. For three patients who harbored at least one additional mutation at the time of MDS leukemic transformation, the combination therapy eradicated the sub-clonal population which raised at AML diagnosis and allowed obtaining CR/CRi (Supplementary Fig. 7a). For the three other patients, the molecular pattern did not change at AML diagnosis compared to the prior MDS (Supplementary Fig. 7b). Furthermore, we observed that 5/7 c Patients were categorized into those who achieved CR/CRi without relapse (group 1; n = 7) or with relapse at any time during the study (group 2; n = 9) and patients who only achieved partial remission (PR) or with a refractory disease (RD) (group 3; n = 3). The presence of study ID number, indicated mutations, complex karyotype, percentage of LSCs among the total blast population and best response are shown for each case. patients (71%) of group 1 experienced a strong decrease in most variant alleles frequencies (VAF) of the mutations detected at diagnosis (Fig. 2f). Interestingly, mutations for which the VAF only slightly decreased or even increased were DTA mutations which do not correlate with an increased relapse rate in AML patients treated with chemotherapy (Fig. 2f) 15 . In contrast, a strong decrease in most VAFs with a persistence or slight increase evident only in DTA mutations was observed for only three of the nine patients (33.3%) of group 2 (Fig. 2g). Furthermore, for these patients, the treatment clearly led to the selection of an initial sub-clonal population with IDH2, CBL or NRAS mutations (Fig. 2g). As expected, HTS performed at relapse in five patients confirmed the raise of the emergent mutations observed under treatment in two cases (IDH2 and CBL) ( Supplementary Fig. 8). Thus, the type of mutations (DTA vs non-DTA) which persist or increase under treatment clearly affected patient's outcomes.
Overall, our current study confirms the high efficacy and safety profile of the venetoclax combination therapy in older AML patients. The clinical benefits of this combination are associated with the efficient targeting of LSCs but we also show that a high amount of LSCs (>1%) at diagnosis may limit its efficacy. Finally, monitoring of MRD and clonal hematopoiesis in these patients seems useful approaches for predicting patient's outcomes.