Venetoclax in Relapsed/Refractory Acute Myeloid Leukemia: Are Supporting Evidences Enough?

Simple Summary Venetoclax (VEN) is a potent oral inhibitor of the anti-apoptotic molecule BCL2, approved for adults with chronic lymphocytic leukemia (CLL), and recently for naïve acute myeloid leukemia (AML) unfit for intensive chemotherapy. Therefore, VEN is not approved for relapsed/refractory (R/R) AML patients; consequently, this use falls within the so-called off-label use. This review provides evidence of the role of VEN-based therapy in R/R AML patients, including data from clinical trials and from retrospective studies. Abstract Despite the progress in the development of new therapeutic strategies, relapsed/refractory (R/R) acute myeloid leukemia (AML) still represents a high unmet medical need. Treatment options in this setting include enrollment into clinical trials, allogeneic stem cell transplantation and/or targeted therapy. Nevertheless, it is associated with poor outcomes. Thus, the development of new treatments, which could ameliorate the prognosis of these patients with a good safety profile are highly demanded. Recently, venetoclax (VEN) has been approved for naïve AML patients unfit for intensive chemotherapy. In this regard, regimens including VEN could represent a valuable treatment option even in those with R/R disease and several studies have been conducted to demonstrate its role in this clinical setting. This review aims to summarize the current evidence on the use of VEN regimens in the treatment of R/R AML.


Introduction
Venetoclax (VEN) is a potent oral inhibitor of the anti-apoptotic molecule BCL2, used to treat adults with chronic lymphocytic leukemia (CLL), in association with obinutuzumab in patients who have not previously been treated or with rituximab in patients who have received at least one previous treatment [1].
It is also used as monotherapy in patients with 17p deletion or TP53 mutation who cannot be treated with or have failed a B-cell receptor pathway inhibitor or in the absence of these genetic changes in adult patients who have failed both chemo-immunotherapy and a B-cell receptor pathway inhibitor [1].
In the last years, this drug demonstrated to be safe and effective in patients with other hematological diseases, in particular acute myeloid leukemia (AML). The median age of AML patients at diagnosis is 68 years and this population has often limited effective treatment options due to ineligibility for intensive chemotherapy [2,3]. Low-dose cytarabine (LDAC) or hypomethylating agents (azacitidine, decitabine) can be used in this population but have been associated with poor response (complete remission, CR, plus CR with incomplete blood count recovery, CRi, rates less than 30% and median survival <6 months) [4][5][6]. Preliminary results and subsequent confirmatory data showed that the association of VEN + hypomethylants/LDAC was associated with higher response rate and better overall survival [7][8][9].
Therefore, an extension of indication has been granted for VEN by the Food and Drug Administration (FDA) for the use in combination with azacitidine or decitabine or LDAC for the treatment of newly diagnosed AML who are unfit for intensive chemotherapy due to age or comorbidities [10,11]. Moreover, this use was approved in 2021 in Europe, only in combination with a hypomethylating agent (HMA) for the treatment of adult patients with newly diagnosed AML ineligible for intensive chemotherapy [12].
However, despite the progress in the development of new therapeutic strategies, relapsed/refractory (R/R) disease is associated with poor outcomes and still represents an unmet medical need [13][14][15]. Treatment options for these patients include enrollment into clinical trials, allogeneic stem cell transplantation (the only potential curative option), targeted therapy (such as gilteritinib for AML with an FLT3 mutation).
Thus, the development of new treatment modalities, which could ameliorate the prognosis of these patients with a good safety profile, is highly demanded.
In this regard, regimens including VEN could represent a valuable treatment option even in younger patients with R/R AML and in recent years several studies have been conducted to demonstrate its role in this clinical setting [15].
This review aims to analyze the current evidence on the use of VEN-containing regimens in the treatment of R/R AML.

Mechanism of Action and Resistance in AML
The BCL-2 family consists of various pro-apoptotic and anti-apoptotic molecules, which regulate the intrinsic apoptotic pathway and have been implicated in the cell survival but also tumorigenesis in many hemato-oncological malignancies [16]. This pathway is activated in response to stress or DNA damage, and leads to the formation of pores in the outer mitochondrial membrane through effector proteins (BAX and BAK), resulting in the release of cytochrome C into the cytosol, caspase-9 activation, and proteolytic cell death.
The overexpression of BCL-2 has been associated with cell survival and apoptosis escape, but also with therapy resistance [17].
In preclinical studies, the oral BH3 mimetic highly selective for BCL-2, VEN, exhibited potent anti-leukemic activity in AML cell lines, xenograft murine models and patient samples [18]. The drug acts by binding BCL-2 causing the release of sequestered proapoptotic signaling proteins BAX and BAK [29,30]. Alternatively, it might also lead to cell death by destabilizing the proton gradient across the mitochondrial inner membrane [31]. Moreover, the combination of VEN + azacitidine inhibits amino acid metabolism, which has been recognized as fundamental to leukemia stem cell survival [32]. This inhibition reduces oxidative phosphorylation and induces cell death, in particular in de novo patients [19,33]. In the R/R setting, an upregulation of fatty acid metabolism has been recognized as a potential compensatory metabolic pathway; therefore, the change in metabolic phenotype of relapsed disease may be responsible for lower VEN efficacy [34,35].
Whatever the mechanism of action, the ultimate effect is the mitochondrial outer membrane permeabilization (MOMP), release of cytochrome C to the cytoplasm and decreased ATP production, formation of cytosolic apoptosome complexes, and subsequent caspase activation and apoptosis [36,37]. Notably, as also reported in multiple myeloma and CLL [38,39], VEN sensitivity is strongly and inversely correlated with the BCL-2/MCL-1 ratio, with a loss of AML cell sensitivity when high levels of MCL-1 are expressed [18,[40][41][42][43].
It is noteworthy that patients with IDH1/2-mutations treated with VEN combinations showed higher and durable responses, and longer survival compared to other subgroups. Indeed, IDH1/2 mutations induce BCL-2 dependence, making AML cells particularly sensitive to VEN, as a single agent or in combination with other agents [7,[58][59][60][61].
On the contrary, FLT3-ITD mutation may produce primary resistance to VEN by enhancing the expression of other anti-apoptotic molecules such as BCL-XL and MCL-1 [47,61,62] and responses to VEN-based combinations have been lower and short-lived [7,58,59]. The combination of VEN with FLT3-tyrosine kinase inhibitor (TKI) induced more durable tumor regression in FLT3-mutant AML cell lines [62].

Clinical Trials
Currently, only few clinical trials (phase I/II) with VEN-based regimens have been performed in R/R AML patients (Table 1).
A phase II study showed only a modest activity (CR/CRi = 19%) of VEN monotherapy administered in R/R AML patients [46].
Exploratory subgroup analyses among 83 previously treated patients (also including R/R patients) showed a CR or CRi rate of 37% (95% CI:  in those having received at least four previous cycles of HMAs, of 36%  in those with previous intensive chemotherapy, and of 27%  in those with previous SCT. The CR or CRi rate was 48% (95% CI 30-67) for patients receiving first salvage therapy for AML and 40%  for those receiving second salvage therapy. Twenty-three patients underwent SCT after a median of three cycles of treatment [interquartile range (IQR) 2-3]. Responding patients transitioning to SCT had the most durable responses with median OS not reached (95% CI: 13.0-not reached) for treatment-naïve patients and 22.1 months (95% CI: 6.8-not reached) in previously treated groups.
In summary, VEN with 10-day decitabine achieved an excellent response both in untreated AML patients and in those previously treated (including R/R AML), and showed an acceptable tolerability, with expected TEAEs (infections with grades 3 or 4 neutropenia and febrile neutropenia) and a low 30-day and 60-day mortality.
Recently, the results of a phase Ib/II study evaluating the combination of fludarabine, cytarabine, granulocyte colony-stimulating factor, and idarubicin with VEN in naïve and R/R-AML have been published [64]. Sixty-eight patients have been enrolled, including 39 with R/R disease, with a median age of 46 years (range, 20-73). The ORR was 75% in those enrolled in the phase Ib portion (composite CR = 75%), and 70% in the phase IIb (composite CR = 61%). MRD-negativity was attained in 69% of R/R patients. Median OS was not reached after 12 months. Forty-six percent of R/R-AML patients proceeded to allogeneic HSCT, with a significant improvement in OS (median OS, NR; 1-year OS, 87%). Therefore, this regimen represents an effective treatment even in R/R-AML patients, associated with high rate of remission and the possibility to be a successful bridge to transplantation.
Several clinical trials with regimen including VEN in R/R AML patients are still ongoing. Table S1 shows the details of these studies, in particular the trial design, the number of patients to enroll, the study interventions, the primary outcome measures, and the estimated completion date.

Observational and Real-World Studies
Since FDA approval in 2018 for upfront use in AML, VEN-combination regimens have been widely used in real-world settings as salvage regimens, with several published retrospective reports documenting activity in patients with R/R AML (Table 1).
A first report of 43 R/R myeloid patients [39 (91%) with AML, 2 with MDS, and 2 with blastic plasmacytoid dendritic cell neoplasm]-the majority of which were over age 65 (58%) and ≥2 salvage-treatment (n = 36, 84%), including prior HMA in 77% and prior allogenic SCT in 12%-who received VEN-based salvage therapy, most commonly in combination with HMAs, documented an ORR of 21% (n = 9), with 2 (5%) CR, 3 (7%) CRi and 4 (9%) MLFS [65]. All 9 patients with ORR responded within the first cycle and 2 successfully transitioned to allogenic SCT. The median OS was 3.0 months overall and 4.8 months in the 9 responders (range 0.5-8.0) and the estimated 6-month survival was 24%. Treatment was discontinued in 38 patients due to the lack of response or disease progression (n = 29, 77%), death (n = 7, 16%), or transition to allogeneic SCT (n = 2, 5%). In regard to molecular/cytogenetic risk, 3/11 (27%) patients with IDH mutations, 4/8 (50%) RUNX1-mutated patients and 2/10 (20%) TP53-mutated patients (with a concurrent RUNX1 mutation) achieved an objective response. The most common complications in this heavily pretreated, older, and high-risk R/R population were grade ≥ 3 neutropenia and grade ≥ 3 infections, mainly pneumonia (n = 17, 40%). Several retrospective studies analyzed the therapeutic response according to the specific molecular subgroup. Sub-analysis of specific molecular mutations in the retrospective study by Aldoss et al. [66] showed a response rate of 67% for IDH 1/2 mutations (similarly to previous reports [7,8,46]), 44% for FLT3 mutations, and 67% for TP53 mutations, suggesting a good clinical activity for the combination of VEN + HMAs even in patients who would be expected to respond poorly to conventional chemotherapy. These results are in line with the single-center retrospective analysis of a cohort of 90 adults with R/R AML treated with at least one cycle of VEN + HMA [67]. In this study, different conventional predictors of poor response (as failure of prior HMA therapy or prior allogenic SCT) were not associated with response rate. Moreover, high-risk mutations such as TP53 and FLT3 were not associated with a lower rate of response compared to other mutations. Only TP53 mutation was associated with reduced LFS in univariate analysis (p = 0.01), but not in multivariate analysis (p = 0.074).
The results of a small retrospective study on 12 patients with R/R AML carrying the NPM1 mutation (5 with molecular persistence and 7 with progression or relapsed), showed a high rate of CR with MRD negativity (92%) after 1-2 cycles of VEN + LDAC or azacitidine [86]. Median relapse-free survival was not reached in the 5 responders with previous molecular persistence (median follow-up = 20 months), and none of them experienced disease progression or received an allogenic HSCT for consolidation.
Another retrospective study by Aldoss et al. [68], including 32 adult patients all with TP53-mutated AML (16 R/R and 15 newly diagnosed) treated with VEN + HMA, reported a global response rate of 52% (n = 16), with 7 patients experiencing CR and 9 CRi (38% CR/CRi rate in R/R). This represents a good result considering the poor outcomes reported in TP53-mutated AML patients treated with conventional combination chemotherapy [95]. Interestingly, patients with more than one TP53 mutation obtained a higher CR/CRi rate (78% vs. 41%, p = 0.062). Another study by Aldoss et al. [69] revealed a response of 42% among R/R AML patients with FLT3 mutations, better in those who were naïve for FLT3-based TKIs (37% for patients with prior exposure vs. 50% in naïve).
The potential role of mutations as predictive factors for response to VEN-based therapy has been assessed in a cohort of 40 R/R AML patients [70]. NPM1, RUNX1, or SRSF2 mutations have been associated with higher CR/CRi rates and RUNX1 with longer OS. On the contrary, FLT3-ITD, TP53, or DNMT3A mutations resulted into worse outcome.
Piccini et al. recently showed a particularly favorable outcome for HSCT-naïve patients and aged >60 years, even in the multivariate analysis [71]. Median OS was 10.7 months, longer in the group with NPM1 mutation (median not reached) and shorter in the group with both NPM1-FLT3/ITD mutations (2.3 months).
Regarding the use of VEN as salvage therapy post allogenic HSCT, in a cohort of 21 post-transplant relapsed AML patients who received off-label VEN-based regimens (mainly with HMAs), the observed response rate was 42.1% (n = 8, 5 CR and 3 CRi) [73]. None of the four patients with complex karyotype and TP53 mutations responded to therapy. The median OS was 7.8 months (range 0.2-12.1 months), with a significantly longer OS in patients achieving CR/CRi (p = 0.005). Only one patient with CR/CRi progressed prior to the data cut-off, after 9.2 months on therapy. Another retrospective analysis conducted on 29 patients with post allo-HCT AML relapse treated with VEN regimens as salvage therapy [76] showed a 38% response rate, including CR/CRi in 8 subjects. The median duration of response was 7 months (1-11) and median OS was 79 days (2-403), better in responders (403 vs. 55 days, p < 0.0001). A high response rate was 61.5% (including 26.9% of CRi) which has been observed even by Zhao et al. [77] in 26 patients with AML relapsed after allo-HSCT treated with VEN + azacitidine and donor lymphocyte infusion. In the 6 patients relapsed after allogeneic HCT and retrospectively observed by Ram et al. [85], CR/CRi was achieved in 67% of the patients (n = 4) and the median OS was 12.4 months.
Several responders successfully proceeded to allogenic HSCT [13]. In this regard, Zappasodi et al. [84] specifically explored the utility of VEN with azacitidine as a bridge to allogeneic SCT in 10 heavily pretreated patients with refractory AML. ORR was 60% (n = 6), with 4 CR, 1 CRi, and 1 MLFS. HSCT was performed in all 6 responders and in 1 non-responder with a consistent reduction of blasts in the bone marrow and resolution of the thrombocytopenia. Five of the transplanted patients were alive at the time of the analysis and four were in CR. Median OS was 8.9 months (range 2-19) and 11.7 months among transplanted patients.
Sixty-five R/R patients treated with 10-day decitabine and VEN were compared to 130 patients treated receiving intensive chemotherapy [89]. In regard to safety, the most common AEs were neutropenia, and febrile neutropenia, including serious events, which can lead to drug interruption and/or hospitalization. Overall, the drug had an acceptable safety profile, especially in patients with optimal blood count prior to treatment.
All these retrospective data support the role of VEN-based regimen as an effective treatment in R/R-AML patients, even in those heavily pre-treated and relapsed after HSCT. The role of predictors of good response should be confirmed in larger specific clinical trials.

The Regulatory Perspective: The Off-Label Use Regulation
Currently VEN is not approved for R/R AML patients; therefore this use falls within the so-called off-label use [96][97][98][99][100][101][102]. EMA defined off-label use as "situations where the medicinal product is intentionally used for a medical purpose not in accordance with the authorized product information" [72]. The major advantage of off-label use is the potential satisfaction of unmet medical needs; however, it could increase the risk of inappropriate use, medical error and adverse events due to the uncertainty about the drug value in the off-label indication, related to the lack of conclusive data supporting its benefit-risk ratio [103]. In this context, it is crucial and ethically acceptable that patients should always be well-informed about the proposed unauthorized use, about the available evidence and the reasons why it represents the best therapeutic option, so that they can consciously provide their consent to the off-label treatment.
This use is not regulated in Europe, but member states adopted specific regulation in some cases [72]. For example, the Italian Law 94/1998 allows physicians to perform off-label prescriptions in individual and exceptional cases, if the prescriber assumes the responsibility of the prescription, an adequate mandatory informed consent of the patient is provided, and efficacy and safety results derived from at least phase II clinical trials are available [104]. The cost of these prescriptions is not covered by the national health system (NHS). In addition, the Italian Law 648/96 includes the reimbursement of an off-label use supported by evidence from at least phase II trials and recognized by the national regulatory authority. In this case, patient associations, scientific societies, and clinical centers may request to the national competent authority the approval of new therapeutic use of effective and safe medicines beyond the interest of pharmaceutical companies. Finally, the off-label use can be covered in Italy within the "5% Fund" according to Law 326/2003, ensuring access for rare or serious diseases [105].
In France the Recommandations Temporaires d'Utilisation (RTU) ensures nationwide access to off-label drugs according to criteria for appropriate use and monitoring defined in the light of clinical evidence [106,107], as well as the Law 648/1996 in Italy [108].
Since March 2020, the Italian NHS covers the use of VEN in combination with HMAs for adult patients with newly diagnosed AML not eligible for intensive induction chemotherapy [109,110]; on the contrary, the use in R/R is not approved and falls within the Italian Law 94/1998, by which physicians can perform off-label prescriptions (not covered by the NHS) only in individual and exceptional cases [111]. This represents, to date, a limitation due to the exceptionality and not systematicity that should characterize the prescription.
In recent years, VEN has been nominally authorized by the Agenzia Italiana del Farmaco (AIFA) for several patients [71]; however, this can be a time-consuming administrative procedure which can complicate healthcare management.
Based on current evidence including phase II trials, the use of VEN in R/R AML patients could be recognized according to Law 648/96, improving patients' access.

Conclusions
This review provides evidence of the role of VEN-based therapy in R/R AML patients, although most data came from retrospective studies.
Globally, these findings suggest an efficacy of VEN in combinations with HMAs (decitabine and azacitidine) or LDAC not only in the upfront setting but also in R/R AML, with an acceptable safety profile, comparable to those observed in other conditions, like CLL, with persistent cytopenias, infections, and transfusion requirements as major toxicities. VEN regimens can lead to high response rates with moderately durable remission and survival and can bridge patients to transplant. The considerable variability observed regarding clinical and mutational predictors of response and survival might be explained by the heterogeneity in disease biology and prior therapies received in the context of R/R AML patients.
In summary, despite the role of VEN-based combination regimens in the treatment of R/R AML patients should be further confirmed and optimized in additional prospective controlled clinical trials (currently ongoing), these combinations may represent a good salvage option in R/R AML in particular in case clinical trials are not available, other intensive chemotherapy regimens or HSCT have already failed, or in those who cannot tolerate intensive chemotherapy due to their age or performance status.