Enhancing venetoclax activity in hematological malignancies

Abstract

Introduction

Targeting anti-apoptotic pathways involving the BCL2 family proteins represents a novel treatment strategy in hematologic malignancies. Venetoclax, a selective BCL2 inhibitor, represents the first approved agent of this class, and is currently used in CLL and AML. However, monotherapy is rarely sufficient for sustained responses due to development of drug resistance and loss of dependence upon the targeted protein. Numerous pre-clinical studies have shown that combining venetoclax with other agents may represent a more effective therapeutic strategy by circumventing resistance mechanisms. In this review, we summarize pre-clinical data providing a foundation for rational combination strategies involving venetoclax.

Areas covered

Novel combination strategies in hematologic malignancies involving venetoclax, primarily at the pre-clinical level, will be reviewed. We emphasize novel agents that interrupt complementary or compensatory pro-survival pathways, and particularly mechanistic insights underlying synergism. PubMed, Cochrane, EMBASE, and Google scholar were searched from 2000.

Expert opinion

Although venetoclax has proven to be an effective therapeutic in hematologic malignancies, monotherapy may be insufficient for maximal effectiveness due to the development of resistance and/or loss of BCL2 addiction. Further pre-clinical and clinical development of combination therapies may be necessary for optimal outcomes in patients with diverse blood cancers.

1. Introduction: BCL2 blockade in hematological malignancies

Mitochondrial outer membrane permeabilization (MOMP) is an irreversible step ultimately culminating in apoptotic cell death, and B-cell lymphoma 2 (BCL2) and related anti-apoptotic proteins, such as BCLXL, myeloid cell leukemia sequence 1 (MCL1) regulate this process [1]. These anti-apoptotic proteins bind to and sequester pro-apoptotic BCL2 family members, such as multimeric proteins BAK and BAX which are required for MOMP. These relationships are summarized in Fig. 1. Once activated, these proteins undergo conformational change, leading to mitochondrial permeabilization, release of pro-apoptotic factors, such as cytochrome c into the cytoplasm, culminating in caspase activation and cell death. Cell fate is ultimately determined by interactions between these competing cell death-related proteins, and it has generally been understood that a predominance of anti-apoptotic signaling by this protein family represents a seminal characteristic of the neoplastic state. However, it is currently unknown whether induction of apoptosis reflects release of pro-apoptotic proteins e.g., BIM from the inhibitory effects of their anti-apoptotic counterparts (e.g., BCL2), or the reverse e.g., disruption of anti-apoptotic protein function by pro-apoptotic family members.

Fig 1.

The identification of BCL2 as an important determinant of cell survival [2]e.g., in CLL prompted attempts to target this protein as a therapeutic strategy. Initial efforts to inhibit BCL2 involved antisense oligonucleotides, such as G3139, but these were unsuccessful, presumably due to the long BCL2 half-life.

Despite these setbacks, pre-clinical studies indicated that BCL2 inhibition through development of small molecules that mimic the BCL2 homology domain 3 (BH3) found in all pro-apoptotic BCL2 family proteins represented a potentially more effective approach. ABT-737, developed through fragment-based design, was the first-in- class BH3 mimetic to be developed that binds BCL2, BCLXL and BCLW with high affinity as well as weakly to other BCL2 family members including MCL1 [3]. In in vivo studies employing primary cancer cells, including those of hematopoietic origin, ABT-737 effectively induced apoptosis [3,4], However, limited bioavailability and the need for parenteral administration hindered further clinical development. Subsequently, navitoclax, (ABT-263), an orally bioavailable derivative of ABT-737, was then developed. Navitoclax showed efficacy in early phase clinical trials alone or in combination with Rituximab [5–7]. However, thrombocytopenia (grade 3 or higher) was reported in 18–33% of patients treated with navitoclax [6], presumably due to on-target effects related to BCLXL inhibition [8].

These adverse characteristics made navitoclax a less attractive option in the treatment of hematological malignancies and prompted development of ABT-199 (venetoclax), a highly selective small molecule BH3 mimetic with an even greater affinity for BCL2 but a much lower affinity for BCLXL [9] compared to navitoclax. In pre-clinical studies, venetoclax demonstrated in vivo activity comparable to navitoclax in CLL and MYC-driven lymphoma [9,10] and also in AML xenograft models [11]. Further clinical trials showed striking monotherapy efficacy in CLL [12] which led to accelerated approval in 2016 and modest efficacy as monotherapy in AML [13] which was significanlty enhanced when combined with hypomethylating agents [14]. Venetoclax is currently approved in CLL as monotherapy or combined with rituximab or obinutuzumab, and in combination therapy with azacitidine, decitabine, or low-dose cytarabine in AML [15].

Despite these successes, as in the case of all targeted agents, resistance to venetoclax may occur de novo or may be acquired following drug administration. The basis for venetoclax drug resistance may be multi-factorial, but often involves up-regulation of anti-apoptotic proteins (e.g., BCLXL or MCL1) to which venetoclax binds weakly. Notably, up-regulation of these anti-apoptotic proteins has also been identified as a resistance mechanism for ABT-737 or navitoclax [16–18]. Alternatively, resistance could also theoretically reflect down-regulation of pro-apoptotic proteins, or potentially increased ratios of BCL2 to BIM [19]. The latter may reflect increased binding of BCL2 to BIM, thereby neutralizing the ability of the latter protein to induce apoptosis. Regardless of the mechanism by which resistance to venetoclax occurs, such findings stimulated the search for agents capable of lowering the threshold for venetoclax-induced cell death e.g., by reducing the expression of resistance-conferring BCL2 family proteins, particularly MCL1. Indeed, synergistic interactions between such agents and venetoclax have been reported in diverse neoplastic cell types, particularly hematopoietic malignancies. In this review, we will summarize strategies designed to optimize venetoclax efficacy and overcome resistance in hematopoietic malignancies through rational combinations with other novel agents, focusing on mechanisms responsible for potential therapeutic interactions. To review the relevant literature, PubMed, Cochrane, EMBASE and Google Scholar were searched from 2010 using following terms; Antiapoptotic pathway, BCL2, MCL1, BCLXL, venetoclax and references within each article were also reviewed to identify further studies.

2. Combinations with DNMT (DNA methyltransferase) inhibitors

Inhibitors of DNA methyltransferase1 (e.g., 5-azacytidine and deoxyazacytidine) modify chromatin structure and the expression of genes implicated in leukemic cell differentiation and cell death. They have been approved for patients with the myelodysplastic syndrome (MDS). The combination of venetoclax with DNMT1 inhibitors has shown promise in pre-clinical AML models as well as in early-phase clinical trial in older patients with AML [14]. Notably, synergistic induction of mitochondrial apoptosis with concomitant DNMT1 and BCL2 inhibition in AML cells was described prior to venetoclax approval, and this was initially attributed to MCL1 downregulation by DNMT inhibitors [20,21]. For example, Jin and colleagues showed that azacitidine induces NOXA upregulation through non-epigenetic mechanisms in AML cells while NOXA neutralizes MCL1, thereby priming cells for venetoclax [22].

Subsequently, other mechanisms underlying synergism have been identified. BCL2 inhibition has been shown to eradicate leukemia stem cells (LCS) through disruption of oxidative phosphorylation(OXPHOS) [23]. Following the initial promise of regimens combining venetoclax with azacitidine in early phase clinical trials, Pollyea and colleagues investigated the role of OXYPHOS disruption in LSCs ablation examining pre- and post-treatment patient samples [24]. Combined treatment resulted in the rapid eradication of blasts within 6 days of treatment and this rapid effect was also observed in putative LSCs (CD34+CD38-CD123+) while sparing normal hematopoietic lineages. This rapid reduction had not previously been observed in specimens obtained from patients who underwent intensive standard chemotherapy induction. The specificity of blast and LSC elimination as well as rapid eradication was confirmed using single cell RNA sequencing methods. Based on the observed inhibitory effects of BCL2 inhibition on OXPHOS downregulation in LSCs, this group hypothesized that the striking effect of the venetoclax/azacitidine regimen reflected inhibition of LSC OXPHOS. Indeed, metabolomic analyses of leukemic specimens from patients pre- and 24h post-treatment with venetoclax + azacitidine, a significant reduction in OXPHOS was observed but was generally not seen in patients treated with induction chemotherapy or azacitidine alone. Metabolomic analyses from treated patients also revealed diminished alpha-ketoglutarate, fumarate, malate and increased succinate and citrate, collectively suggesting disruption of the tricarboxylic acid cycle through inhibition of electron transport chain II. Further investigation of patient samples following short exposures to venetoclax and azacitidine in vitro revealed decreased glutathione, electron transport complex II activity, OXPHOS, and ATP levels which were not seen with single-agent treatment. Of note, these effects were partially rescued by addition of cell permeable glutathione prior to treatment. Collectively, these findings suggest that combined BCL2/DNMT1 exposure selectively and synergistically eradicates AML blasts and LSCs through disruption of electron transport chain complexes in a glutathione-dependent manner [24].

Alternative mechanisms have been identified to explain synergistic interactions between DNMT1 and BCL2 inhibitors in AML cells. For example, the cytotoxicity of DNMT1 inhibitors has been found to be partly dependent on the formation of reactive oxygen species(ROS) [25–27]. Furthermore, aberrant activation of the Nrf2 transcription factor has been shown to mediate resistance to such agents by preventing oxidative injury [28,29]. Nguyen and colleagues reported that venetoclax disrupted DNMT inhibitor-induced nuclear translocation of Nrf2 and induction of down-stream antioxidant enzymes. Venetoclax also induced dissociation of the BCL2 protein from the Nrf2/Keap-1 complex, thereby targeting Nrf2 for ubiquitination and proteasomal degradation. Consequently, combined venetoclax and DNMT inhibitor treatment induced significantly greater mitochondrial ROS induction and apoptosis in AML cells compared to treatment with a DNMT1 inhibitor alone [30]. Finally, these two mechanisms are not mutually exclusive, and it is possible that both disruption of OXPHOS as well as Nrf2-mediated antioxidant effects both contribute to DNMT1/BCL2 inhibition-related anti-leukemic activity.

3. Combination strategies involving MCL1 inhibition

During the development of navitoclax, high expression of MCL1 was shown to represent a frequent resistance mechanism [31], and the concept of circumventing resistance by targeting anti-apoptotic proteins (e.g., MCL1 and BCLXL) was also found to be applicable to venetoclax [32]. MCL1 is known to be important for the survival of AML [33], lymphoma [34–36] and MM (multiple myeloma) [37] cells. Consequently, targeting MCL1 in combination with venetoclax represents a rational approach to increase venetoclax activity and overcome resistance. Targeting MCL1 can be accomplished through both direct and indirect inhibition, e.g., by agents that bind to and inactivate MCL1, or by agents that disrupt MCL1 synthesis or stability. Al-harbi and colleagues also studied the effect of BCLXL antagonism and micro RNA targeting BCLXL (mir-377) in venetoclax resistant cell lines as well as previously treated CLL patient samples. mir-377 was decreased both in resistant cell lines and previously treated patient samples while BCLXL expression was increased. In addition, BCLXL was significantly correlated with patient disease-free survival, arguing that BCLXL may play an important role in treatment strategies in CLL [38].

3.1. Direct MCL1 inhibitors

Through conditional gene knockout studies, MCL1 was shown to be essential not only for malignant hematopoietic cell expansion but also for the survival of hematopoietic stem cells [39]. S63845 was the first potent, selective MCL1 inhibitor shown to exhibit considerably lower affinity to other anti-apoptotic proteins e.g., BCL2 and BCLXL. This compound induces apoptosis by interfering with BAK and BAX binding to MCL1 [40]. S63845 displayed single agent activity in myeloma, leukemia and lymphoma cells both in vitro and in vivo, and was well tolerated in mice at effective doses [40]. Notably, its affinity to human MCL1 was six-fold higher than that for murine MCL1. This tolerability was evaluated further in humanized mice, and demonstrated that S63845 did not induce sustained toxicity at maximally tolerated doses [41]. Combination with a BCL2 inhibitor and S63845 in vivo displayed synergistic efficacy in a BAK/BAX dependent manner in AML, ALL (Acute Lymphoblastic Leukemia), MM and NHL (Non Hodgkin Lymphoma) cells [42–45].

Other selective MCL1 inhibitors have recently been developed e.g., (AZD5991 [46], AMG176 [47]). These agents have also been shown to interact synergistically with venetoclax in malignant hematopoietic cells (e.g., AML and MM [46])

There are several early phase clinical trials ongoing designed to evaluate combinations of venetoclax and MCL1 inhibitors in AML (NCT03672695) or relapsed/refractory hematological malignancies (NCT03218683). However, a trial has been suspended (NCT03797261) involving this combination strategy due to safety assessments, and the feasibility of this strategy remains to be determined. Further approaches to predict/maximize responses with acceptable toxicity for this strategy, including the use of biomarkers to minimize adverse events may be required.

3.2. Indirect MCL1 inhibition involving the PI3K/AKT/mTOR pathway

The PI3K/AKT/mTOR pathway is involved in many cellular functions, including protein synthesis, cell cycle progression, cell survival, apoptosis, angiogenesis and drug resistance [48] (Fig. 2) and its activation is frequently observed in diverse hematological malignancies including AML [49,50], ALL [51], NHL [52], MM [53,54], CML [55] and CLL [56]. Inhibition of this pathway has been an attractive strategy in both hematological as well as non-hematologic malignancies, and idelalisib (a PI3KΔ inhibitor in CLL), copanlisib (PI3KαΔ inhibitor in follicular lymphoma) and duvelisib (PI3K Δ/γ inhibitor for CLL) are currently approved and in clinical use for these hematological malignancies. Choudhary and colleagues initially developed venetoclax resistant NHL cell lines and found activation of AKT accompanied by up-regulation of MCL1/BCLXL in these cells [57]. Of note, down-regulation of these proteins, particularly AKT and MCL1, with siRNA sensitized resistant cells to venetoclax. Furthermore, co-treatment with idelalisib and venetoclax decreased MCL1 expression significantly and increased cell death in association with AKT-mediated BAX activation [57]. Various other groups subsequently demonstrated synergism between venetoclax and inhibitors of the PI3K/AKT/mTOR pathway, such as copanlisib and duvelisib [58–61], in malignant hematopoietic cells. The mechanism of synergism in these studies appeared to proceed primarily through down-regulation of MCL1 and/or BCLXL. For example, in AML cells, PI3K and AKT inactivation resulted in GSK3α activation, leading to increased MCL1 degradation [60] Based in part upon these pre-clinical findings, clinical trials combining venetoclax with copanlisib (NCT03886649; relapsed/refractory B-cell lymphoma), duvelisib (NCT03534323 for R/R CLL or Richter’s transformation) and idelalisib (NCT03639324 for R/R CLL) are currently ongoing.

Fig 2.

3.3. Indirect MCL1 inhibition involving the Ras/Raf/MEK/ERK pathway

The Ras/Raf/MEK/ERK signaling pathway is one of the most frequently dysregulated signaling cascades in cancer. Activating mutations of Ras and Raf occur frequently in both solid tumors and hematologic malignancies, leading to activation of their downstream targets MEK1/2 and ERK1/2 [62,63]. In addition, inactivation of MEK1/2 can trigger up-regulation of the pro-apoptotic protein BIM, thereby promoting cell death [64]. In AML, this signaling pathway is frequently activated through up-stream protein mutation [65]. In pre-clinical models of AML, MEK inhibitors induced apoptosis in a p-4E-BP1 and MCL1 down-regulation- dependent manner [66]. However, the clinical benefit of single-agent MEK1/2 inhibitors has been limited in AML [67] and in multiple myeloma [68]. More recently, a pre-clinical study demonstrated that combining the MEK1/2 inhibitor cobimetinib with venetoclax interacted synergistically in AML cell models through a mechanism involving down-regulation of MCL1 accompanied by BIM release [69]. In addition, Best and colleagues showed that the MEK inhibitor binimetinib acted in a cytostatic manner as a single agent in CLL cells but potentiated the activity of venetoclax through a similar mechanism [70].

A phase II trial combining venetoclax with cobimetinib in R/R AML patients not eligible for intensive therapy is underway (NCT02670044).

4. Combination involving CDK (Cyclin Dependent Kinases) inhibitors

CDKs represent a family of serine/threonine protein kinases whose activity depends on binding to and inhibiting the function of specific cyclin partners [71]. Currently, there are principally two categories of CDKs e.g., those with direct functions in cell cycle regulation including CDK1, 2, 3, 4 and 6, and those with transcriptional regulatory and other functions, such as CDK 7 and 9. CDK function is frequently perturbed in human cancers, including hematological malignancies, and inhibition of CDKs has been the subject of significant clinical interest [72,73]. CDK6 is considered oncogenic in MLL-rearranged leukemia [74] and chromosomal translocation involving cyclin D1 in mantle cell lymphoma [73]. CDK4/6 inhibitors (palbociclib, abemaciclib, ribociclib) are currently in clinical use and have been approved when combined with aromatase inhibitors in patients with breast cancer. Pre-clinical studies investigating CDK4/6 inhibitor interactions with venetoclax in hematopoietic malignancies are largely lacking.

In contrast to other CDKs, CDK7 and 9 interact with RNA polymerase II, wherein CDK7 facilitates transcription initiation whereas CDK9 is required for efficient transcription elongation [75]. The therapeutic activity of pan-CDK inhibitors are believed to be at least partially due to CDK9 inhibition [76,77] which blocks RNA Pol II activation, disrupts transcription, and triggers down-regulation of several short lived pro-survival proteins, such as c-MYC, XIAP (X linked inhibitor of apoptosis protein) and MCL1 [78] (Fig. 2). In pre-clinical studies, CDK inhibitors with significant CDK9 inhibitory activity have been evaluated in combination with venetoclax. Li and colleagues evaluated dinaciclib, which inhibits CDK1, 2, 5, 9, and found this agent to be more effective in DLBCL (Diffuse Large B-Cell Lymphoma) cell lines exhibiting higher MCL1 expression relative to BCL2. Moreover, exposure to venetoclax induced upregulation of MCL1. When co-administered, venetoclax and dinaciclib displayed significant synergistic anti-lymphoma activities both in vitro and in vivo and MCL1 upregulation was countered [79]. This MCL1 downregulation may stem from CDK2 inhibition as well [80]. Dey and colleagues also observed synergism between venetoclax and voruciclib, a CDK9 inhibitor with excellent bioavailability in several DLBCL models [81]. More recently, it has been shown that that CDK9 inhibition by the pan-CDK inhibitor flavopiridol (alcovidib) synergistically enhanced venetoclax activity in multiple myeloma cells through MCL1 down regulation [82]. Notably, this interaction also occurred in high-risk multiple myeloma cells otherwise resistant to venetoclax (e.g., t (4;14). Bogenberger and colleagues combined venetoclax with alvocidib in AML and proposed that down-regulation of MCL1 and/or increased BIM and NOXA expression represented mechanisms underlying synergism [83]. Xie and colleagues showed MCL1 and XIAP down regulation and induction of apoptosis by LS-007, another potent CDK9 inhibitor, in ALL cell lines. BCL2 expression was not affected by CDK9 inhibition and the combination of LS-007 with venetoclax was synergistic in these models [84]. Most recently, Phillips et. al. studied another CDK9 inhibitor A-1592668 and showed synergism with venetoclax both in vitro and in vivo associated with satisfactory tolerability [85]. Based on these findings, there are currently several early phase trials underway combining venetoclax and CDK9 inhibitors, e.g., NCT04017546, NCT03441555 and NCT03484520 in R/R AML or CT03739554 in CLL.

5. Combinations involving HDAC (Histone Deacetylase) inhibitors

HDACs are critical regulators of gene expression that enzymatically remove acetyl group from histones, permitting a more open chromatin structure more conducive to transcription. However, they also exert other effect on various cellular functions through activities involving non-histone proteins [86]. Very early studies revealed that the HDAC inhibitor sodium butyrate hyperacetylated histones and induced new gene expression in leukemia cells [87]. However, there has been an emerging realization that HDAC inhibitors are highly pleiotropic agents which act through a wide variety of disparate and mutually interactive mechanisms. These include down-regulation of anti-apoptotic proteins, such as XIAP and MCL1, up regulation of pro-apoptotic proteins, such as BIM, BMF and NOXA, activation of death receptor pathways, generation of ROS (reactive oxygen species) and induction of DNA damage, disruption of DNA damage repair, disruption of cell cycle checkpoints, signal transducer and activator of transcription (STAT) 5 and 3 inhibition and induction of autophagy [88]. Current FDA approved agent and indications include romidepsin for R/R Cutaneous T Cell Lymphoma (CTCL) and Peripheral T cell Lymphoma (PTCL), vorinostat for R/R CTCL and panobinostat for R/R myeloma.

HDAC inhibitors as single agents have shown only moderate efficacy in diseases other than T-cell lymphomas, which has prompted the development of combination strategies. The most clinically successful approach has been the combination of HDAC inhibitors with proteasome inhibitors. For example, one of the reported HDAC inhibitor resistance mechanisms involves activation of the NF-kB survival pathway and inhibitors of NF-κB potentiate HDAC inhibitor lethality in malignant hematopoietic cells [89]. In this context, proteasome inhibitors prevent degradation of IκBα, which is a negative regulator of NF-kB. Panobinostat in combination with bortezomib and dexamethasone showed significantly longer progression free survival compared with bortezomib and dexamethasone for R/R myeloma patients in PANORAMA 1 trial, and achieved approval by the FDA as well as in Europe [90].

Cyrenne and colleagues investigated synergism between HDAC inhibitors (romidepsin and vorinostat) and venetoclax in T-cell lymphoma cells in vitro. The combination was antagonistic at nanomolar concentration but at micromolar concentrations the combination was highly synergistic in patient samples. The mechanism of synergism was related to a greater than 10- fold increase of BIM expression by HDAC inhibitors, and it was also found that synergism was most marked in cell lines with high BCL2 expression [91]. Ramakrishnan and colleagues studied the combination of panobinostat and MEK inhibitors or venetoclax in multiple myeloma cells. They found increased dissociation of BIM from MCL1 and BCLXL by HDAC inhibition in cells exposed to both agents. Moreover, venetoclax exposure resulted in a further release of BIM from BCL2. This led to increased expression of BIM : BAK and BIM : BAX complexes culminating in the synergistic induction of apoptosis [92]. Andersson and colleagues found combination of five drugs, gemcitabine, busulfan, melphalan, panobinostat and venetoclax synergistically induced apoptosis in vitro, and suggested inhibition of PI3K/AKT/mTOR and activation of unfolded protein response in the endoplasmic reticulum as underlying mechanism of synergism [93]. As of the current writing, there are no active clinical trials combining venetoclax with HDAC inhibitors.

6. Combinations involving glutamine metabolism

Cancer cells rely on different metabolic pathways compared to their normal counterpart as they exhibit distinct bioenergetic and biosynthetic requirements. Higher uptake and use of glucose by neoplastic cells with less dependence on OXPHOS but enhanced activity of anaerobic glycolytic pathways (the Warburg effect) has been described known. Pyruvate generated from glycolytic pathways is converted to lactate and used less in the tricarboxylic acid (TCA) cycle. Consequently, neoplastic cells are more dependent on elevated glutaminolysis to maintain a functional TCA cycle and to meet increased requirement for biosynthetic precursors and NADPH. Inhibition of this pathway has recently become a major focus of interest in cancer therapy [94]. Guieze and colleagues studied the mechanism of venetoclax resistance in CLL cell lines, and found that aside from anti-apoptotic protein upregulation, e.g., MCL1, resistant cell lines exhibited deregulated energy metabolism via altered expression of components of the AMPK signaling pathway [95]. Willems and colleagues previously demonstrated that the efficacy of glutamine transporter knockdown in inducing robust antileukemic responses in AML cells in vitro and in vivo [96]. As observed previously, BCL2 inhibition can exert effects on leukemia stem cells through downregulation of OXPHOS [23]. Based on this consideration, venetoclax was combined with the glutaminase 1 inhibitor CB-839, and showed synergistic induction of mitochondrial apoptosis in AML cells [97]. McBrayer and colleagues evaluated the combination of glutamine metabolism inhibition and venetoclax in multiple myeloma cell lines and patient samples in vitro. Although myeloma cells are highly dependent on MCL1 in general, diminished expression of MCL1 was observed following glucose deprivation, but did not uniformly affect cell survival [98]. It was hypothesized that nutrient deprivation may re-configure BCL2 family protein dependence in myeloma cells and enable sensitization to a BCL2 inhibitor like venetoclax. In vitro, both glucose and glutamine deprivation increased BIM binding to BCL2 compared to controls, and co-treatment with venetoclax and 6-diazo-5-oxo-1-norleucine, a glutaminase inhibitor, induced cell death synergistically [19]. Although CB-839 is currently under investigation in combination with azacitidine (NCT03047993), a DNMT1 inhibitor, the combination of venetoclax and an inhibitor of the glutamine metabolism pathway has not yet been evaluated in clinical trials.

7. Other rational combinations involving venetoclax

7.1. NAE (Nedd 8 Activating Enzyme) inhibitors

Pevonedistat (MLN4924) represents a potentially promising new agent in AML treatment. It is a first-in-class inhibitor of protein “neddylation,” a post-translational protein modification that operates in parallel with ubiquitination to allow proteins to be targeted for proteasomal destruction. Pevonedistat blocks degradation of IκBα, resulting in inhibition of NF-kB, upon which LSCs are dependent for survival [99]. Pevonedistat also induces accumulation of the DNA licensing factor CDT1, leading to re -replication and DNA damage [100]. Pevonedistat has shown modest single agent activity in AML [101] and MM [102], and when combined with azacitadine, the overall response rate was 50% in the former [103]. In preclinical studies involving AML cell lines and patient samples, pevonedistat induced accumulation of c-MYC, a substrate of NF-kB, that transactivates NOXA. NOXA specifically binds to and neutralizes MCL1, leading to less BAK binding to MCL1, resulting in apoptosis. Combination of venetoclax and pevonedistat in vitro interacted synergistically to induce apoptosis both in cell lines and patient samples [104]. Based in part upon these findings, there are currently two ongoing early phase clinical trials combining venetoclax, azacitidine and pevonedistat (NCT04172844, NCT03862157).

7.2. BET (Bromodomain and Extra-Terminal) inhibitors

BET proteins, such as BRD2, BRD3, BRD4 and BRDT, bind to acetylated lysins of histones, an initial post-translational modification that allows greater accessibility to DNA, thereby activating transcription of growth-promoting genes, such as MYC and NUT [99,105]. Aberrant expression of the MYC oncogene is seen in various hematological malignancies, such as AML, Burkitt’s lymphoma and mixed-lineage leukemia. Disruption of BET binding reduces cellular proliferation and induces apoptosis in these diseases [106]. JQ1 was one of the first BET inhibitor to be developed that blocks BRD2 and BRD4. Although it is not under clinical development due to short half-life, it demonstrated pre-clinical activity in hematological malignancies in vitro and in vivo through down-regulation of MYC [107–109]. Combination of venetoclax and JQ1 or other BET inhibitors, such as OTX015 was evaluated in T cell ALL [110], CTCL (Cutaneous T Cell Lymphoma) [111], CLL [112], double hit lymphoma [113] and AML [114], and in each case synergism was observed. Such interactions were attributed in part to up-regulation of proapoptotic proteins BIM/BFL-1 by BET inhibition and/or down-regulation of other antiapoptotic proteins e.g., MCL1 and BCLXL.

A phase I clinical trial combining venetoclax and ABBV-075, which binds to BRD2, BRD4, and BRDT is completed and results are pending (NCT02391480).

7.3. BTK (Bruton’s Tyrosine Kinase) inhibitors

BTK inhibitors, such as ibrutinib, zanubrutinib, acalabrutinib are approved in various hematological malignancies, such as CLL, mantle cell lymphoma, marginal zone lymphoma and Waldenstrom macroglobulinemia. Pre-clinical studies suggested that CLL cells exposed to Ibrutinib displayed an increase in BCL2 dependence, potentially representing a mechanism underlying synergy by of the ibrutinib/venetoclax regimen [115]. Such venetoclax combination regimens are currently in late clinical development, including 5 phase 3 studies involving ibrutinib (NCT02950051, NCT03112174, NCT03462719, NCT03701282, NCT03737981) and two involving acalabrutinib (NCT03836261, NCT03868722).

7.4. SYK (Spleen Tyrosine Kinase) inhibitors

SYK is another component of the BCR pathway that is either phosphorylated by LYN or auto phosphorylated and mediates activation of BTK and PI3K/AKT/mTOR pathway [116]. Given the important role of the BCR pathway in NHL, SYK inhibitors were studied in relapsed/refractory CLL and DLBCL with modest single agent activity and good tolerability [117–119]. Subsequently, combination strategies were further investigated. Bojarczuk et.al. showed that fostamatinib, one of the first SYK inhibitors to be developed, down regulated MCL1 more potently than ibrutinib or idelalisib and overcame venetoclax resistance in CLL cells in vitro. Blunt and colleagues studied the effect of the dual SYK/JAK inhibitor cerdulatinib in CLL cells in vitro. cerdulatinib prevented anti-IgM, IL4 or nurse-like cell mediated up-regulation of MCL1 and BCLXL, and synergistically induced apoptosis in combination with venetoclax. In DLBCL cell lines as well as in vivo, Sasi and colleagues showed fostamatinib induces down-regulation of MCL1 and up-regulation of BIM and HRK, and this agent interacted synergistically with venetoclax both in vitro and in vivo [120]. Currently, there are no active clinical trials combining SYK inhibitors and venetoclax.

7.5. MDM2 (Mouse Double Minute 2) inhibitors

MDM2 is an important negative regulator of the TP53 tumor suppressor and is expressed at high levels in a large proportion of AML patients. As the majority of AML patients express wild type TP53, inhibition of MDM2-p53 interaction can restore the pro-apoptotic TP53 pathway. Idasanutlin is a second-generation, orally bioavailable selective MDM2 antagonist which displayed modest efficacy either alone or in combination with cytarabine in early phase clinical trials of patients with AML [121]. In a pre-clinical study, combination of idasanutlin with venetoclax was evaluated and demonstrated synergism both in vitro and in vivo [122]. Cell lines developed resistance to venetoclax due to prolonged exposure were re-sensitized after treatment with idasanutlin. Up-regulation of p53 led to downregulation of the Ras/Raf/MEK/ERK signaling cascade, activated GSK3 and phosphorylated MCL1, leading to its proteasomal degradation. Interestingly, venetoclax overcame resistance to idasanutlin. One of the resistance mechanisms postulated was induction of pro-survival G1 phase arrest and protection of cells from TP53 induced apoptosis. Venetoclax decreased the apoptotic threshold and switched the cellular response to TP53 activation from pro-survival G1 arrest to apoptosis [123]. There are 3 phase 1/2 studies ongoing to assess clinical efficacy of this combination in AML, follicular lymphoma and DLBCL (NCT02670044, NCT04029688, NCT03135262).

7.6. JAK (Janus Kinase) inhibitors

Janus kinase family is a kinase that links cytokine receptors and the transcription factor STAT which regulates gene expression. As aberrant activation of this JAK/STAT pathway is often seen in myeloproliferative neoplasms (MPNs) including myelofibrosis, the JAK1/2 inhibitor, ruxolitinib, is used clinically in the treatment of these diseases. In pre-clinical studies, Karjalainen and colleagues found ex vivo culture of AML patient samples exhibited resistance to venetoclax in the presence of bone marrow-derived stromal cell cultures, and this effect was reversed in part by JAK inhibitors. Resistance to venetoclax was mediated by up-regulation of BCLXL induced by bone marrow-derived stromal culture conditions, whereas ruxolitinib restored BCL2 dependence and venetoclax sensitivity. In a mouse xenograft model, venetoclax and ruxolitinib interacted synergistically [124]. Senkevitch and colleagues also demonstrated synergism between venetoclax and ruxolitinib in T-ALL cells with interleukin 7 receptor mutations [125]. Currently, a phase 1 trial evaluating the venetoclax + ruxolitinib combination in patients with R/R AML or AML secondary to MDS is underway (NCT03874052).

7.7. CHK1 (Checkpoint Kinase 1) inhibitors

CHK1 is a cell cycle regulatory kinase that plays an important role in preventing cell cycle progression through inhibition of CDKs when damaged DNA is undergoing repair. Disruption of CHK1 results in accumulation of DNA damage and ongoing unscheduled cell cycle progression, culminating in cell death. Zhao and colleagues hypothesized that as DNA damage resulted in diminished expression of MCL1, possibly due to up-regulation of NOXA, CHK1 inhibition might interact synergistically with venetoclax. In fact, synergism was observed both in AML cell lines and patient samples in vitro. Of note, the CHK1 inhibitor, LY2603618 down-regulated MCL1, presumably representing the mechanism underlying synergy [126]. Currently, no clinical trials evaluating venetoclax in combination with CHK1 inhibitors are underway.

8. Conclusions

The intrinsic apoptotic pathway is frequently dysregulated through overexpression of antiapoptotic proteins in diverse forms of cancer, particularly hematologic malignancies. Indeed, BCL2 was shown to be an oncogene in lymphoid leukemia decades ago [127]. Consequently, targeting this pathway in general, and BCL2 in particular, represents a logical strategy in cancer chemotherapy. However, initial attempts to disable BCL2 e.g., with anti-sense oligonucleotides were unsuccessful, most likely due to unfavorable pharmacokinetic properties of available agents and the long BCL2 half-life. While the development of navitoclax, a multi-BCL2 family protein inhibitor with high affinity for BCL2, BCLXL, BCLW represented an alternative approach, thrombocytopenia due to on-target activity against BCLXL made this a sub-optimal strategy in hematopoietic malignancies. On the other hand, venetoclax, which selectively inhibits BCL2, overcame this problem while maintaining clinical efficacy, leading to approval in CLL and AML as well as promising early results in other cancers.

However, as in the case of essentially all other targeted agents, venetoclax monotherapy may not be sufficient to maximize therapeutic activity, and the pre-existence or development of drug resistance invariably supervenes. While the development of assays predicting the activity of venetoclax in individual patients e.g., BH3 profiling [128,129] represents an advance, the interrelated problems of intrinsic or acquired resistance still must be addressed. Although BH3 profiling may be useful in predicting in vitro or short-term responses, dependence on anti-apoptotic proteins represents a dynamic process and can change after treatment with specific inhibitors [130]. Currently, strategies capable of anticipating long-term responses to combination therapies remain to be identified.

Despite these challenges, strategies involving a rational combination of novel agents to overcome or prevent the emergence of resistance and to enhance response rates to venetoclax in hematologic malignancies represents a very attractive approach. For example, in the case of AML, venetoclax monotherapy achieved modest CR rates (e.g., less than 10%) but when combined with DNMT1 inhibitors based upon pre-clinical evidence of anti-leukemic synergism [20,21,23], responses improved to about two thirds of patients. Significantly, the relatively safe toxicity profile of venetoclax provides an incentive to pursue other molecularly rational approaches. As discussed, a wide variety of agents may interact synergistically with venetoclax to induce cell death in malignant hematopoietic cells. Many of these act directly or indirectly to disable proteins that compensate for loss of BCL2 function (e.g., MCL1 or BCLXL). However, other agents (e.g., DNMT1 inhibitors) may similarly interact with venetoclax through unrelated mechanism such as induction of perturbations in redox status. It is likely that other novel agents may also display beneficial interactions with venetoclax in the future. In this context, IDH inhibitors have recently shown significant activity in IDH mutant AML [131]. Moreover, pre-clinical studies suggest that IDH mutant leukemic cells may be particularly dependent upon BCL2 for survival [132]. These findings provide a strong rationale for the concept of combining these agents in patients with IDH-mutant AML. Given the approval of venetoclax either alone (e.g., in CLL) or in combination (e.g., with DNMT1 antagonists in AML), it is highly likely that venetoclax-containing combination regimens will be intensively investigated in the years to come.

9. Expert Opinion

Many of the concepts discussed here are based on the premise that down-regulation of MCL1 represents a key mechanism by which venetoclax activity may be improved. It should be particularly interesting to follow clinical trial results in which venetoclax is combined with direct MCL1 antagonists based on pre-clinical evidence of activity and tolerability of this strategy [133]. However, strategies based on direct antagonism of MCL1 function may need to be pursued with caution, as it remains to be established that all normal host target tissues will be spared by this approach. Whether direct approaches to disabling MCL1 e.g., by small molecule inhibitors or indirect approaches e.g., inhibition of MCL1 synthesis will prove to be the superior therapeutic strategy awaits further clinical information. Determining which of many possible rational combinations involving venetoclax offers the best chance for therapeutic benefit represents a major challenge to the field.

Key questions to be resolved include defining the mechanisms underlying synergism between various agents and venetoclax, and identifying those regimens likely to be the most successful in the clinic. It has become clear that mechanisms other than MCL1 down-regulation or disabling may contribute to interactions between venetoclax and other targeted agents. For example, while initial preclinical studies of venetoclax + DNMT1 inhibitors suggested MCL1 down-regulation as a primary mechanism underlying synergism [20,21,23], subsequent studies identified perturbations in OXPHOS and ROS as potential contributors. BCLXL and dysregulation of energy metabolism represent other important mechanisms of resistance as discussed previously. Additional therapeutic strategies designed to improve venetoclax activity e.g., up-regulation of pro-apoptotic proteins (e.g., BIM by HDAC inhibitors or NOXA by DNA-damaging agents), modulation of cell dependence upon anti-apoptotic proteins, and disruption of autophagy remain to be explored both pre-clinically and in the clinical arena. Whether one of these strategies will prove to be optimal in enhancing venetoclax activity is unknown at this point, but answering this question will depend upon further pre-clinical and clinical investigation. In this regard, one intriguing and potentially promising possibility involves the use of multiple strategies simultaneously.

With respect to early-phase clinical trials, multiple factors need to be considered in selecting the most promising concepts to pursue. Clearly, evidence of activity in pre-clinical studies, both in vitro and in vivo, in specific malignancies will clearly be a prerequisite for further development. It will also be important to demonstrate a molecular basis for therapeutic selectivity, as parallel increases in venetoclax lethality directed at normal host tissues will not improve the therapeutic index. In this context, venetoclax has been associated with certain toxicities e.g.,cytopenias and tumor lysis syndrome, and it would be important to determine that combination therapy would not exacerbate these events. It should also be recognized that even when early phase trial results appear promising, it is possible that clinical efficacy may not be confirmed in phase 3 trials, as suggested by NCT03069352 involving the combination of low dose cytarabine with venetoclax in newly diagnosed AML. Finally, for all such trials, it will be critical to perform correlative pharmacodynamic studies to determine if the mechanisms identified in pre-clinical studies can be recapitulated in patients receiving agents in vivo. Such information will be important in determining whether a regimen fails due to the inability to achieve pharmacodynamic endpoints or inadequate pharmacokinetics.

As in the case of essentially all therapeutic strategies, the availability of a biomarker capable of predicting responses to venetoclax-containing regimens would be extremely valuable. Currently, no definitive markers exist for either venetoclax alone or venetoclax-containing regimens. However, it is tempting to speculate that cells in which BCL2 is over-expressed (e.g., by gene amplification) might be more likely to respond to such a regimen. The premise is that cells that do not express high levels of BCL2 are unlikely to be addicted to this protein and less likely to respond to venetoclax alone or in combination. Prospective studies will be necessary to confirm or refute this notion. In addition, BH3-profiling has been proposed as a way in which responses to venetoclax can be predicted. Parallel studies involving venetoclax-containing regimens should be performed to determine if this approach is valid in this setting. A caveat is that changes in the molecular wiring of cells by the agents themselves may make predictions of long-term responses to such regimens challenging.

Finally, while potentially synergistic venetoclax doublet regimens remain the subject of great interest in the treatment of hematologic malignancies, a focus on rationally designed triplet regimens is attracting considerable attention in these disorders. For example, as noted earlier, regimens combining venetoclax and DNMT1 inhibitors are now approved in AML. Furthermore, the NEDD8 inhibitor pevonedistat has shown single-agent activity in AML [102], and initial evidence of promising activity in combination with 5-azacytidine [103]. In light of pre-clinical evidence of venetoclax/pevonedistat synergism [104], the venetoclax/5-azacytidine/pevonedistat triplet regimen, if tolerable, could display significant anti-leukemic activity. A recently initiated trial will test this possibility (NCT03862157, NCT04172844, NCT04266795). While the identification of mechanisms responsible for complex interactions between such 3-agent regimens will be challenging, pursuing this strategy appears very well justified. In this context, it will be particularly interesting to pursue three-agent strategies in which each of the 3 doublet regimens displays therapeutic synergism.

In summary, it is clear that venetoclax has significantly improved the landscape of treatment options for patients with hematologic malignancies who are not eligible for conventional/intensive therapies or who have relapsed/refractory disease. Based upon emerging pre-clinical as well as clinical developments, it is also very likely that identifying newer rational combination approaches involving this novel agent will lead to further significant therapeutic advances in the years to come. Finally, strategies involving the use of three or more rationally combined agents and/or the selection of patients with specific molecular or pharmacodynamic profiles could further enhance the promise of venetoclax in hematologic malignancies.

Article Highlights.

• BCL2 is one of the anti-apoptotic proteins and an important determinant of cell survival in various hematological malignancies.

• Venetoclax is an oral, highly selective BCL2 inhibitor that was developed to overcome on-target-related adverse effects of navitoclax.

• While venetoclax showed promising efficacy in vitro and in vivo, in early phase clinical trials, resistance to single agent treatment occurred either de novo or was acquired during treatment.

• MCL1 upregulation is one of the major mechanisms of venetoclax resistance.

• Rational combination approaches with other novel agents to overcome resistance based on mechanistic considerations will be important to improve clinical outcomes.