Apoptosis is a regulated process of cell death that plays important roles in homeostasis, human development, and other vital processes, such as immune responses (1, 2). Apoptosis is characterized by two regulatory pathways, extrinsic and intrinsic ( Figure 1 ). These pathways are interconnected with a common element of caspase activation, which leads to the breakdown of cellular components, followed by cell death.

The extrinsic pathway is initiated when receptors on the cell surface, death receptors, are bound by ligands. Some examples of death receptors include TNFR, FAS (CD95), and DR3/WSL (3). Binding of ligands leads to activation of adapter proteins intracellularly and the formation of signaling complexes that trigger caspases.

The intrinsic or mitochondrial pathway is activated by cellular stress. This is the pathway that the Bcl-2 family of proteins regulates through interactions with mitochondrial outer membrane permeabilization (MOMP) (4). BAX and BAK, two executioner proteins of the Bcl-2 family, interact directly with the outer membrane of mitochondria and induce MOMP, which results in release of proteins such as cytochrome c from the intermembrane space of mitochondria into the cytosol (5, 6). The release of cytochrome c then leads to the formation of an apoptosome protein complex which triggers caspase activation. Defects in the intrinsic regulatory elements of apoptosis are thought to be involved in the development and progression of malignancies. The first pro-survival protein of the Bcl-2 family, Bcl-2, was identified in the early 1990s and it was noted that overexpression of Bcl-2 contributed to lymphomagenesis in follicular lymphoma, especially in cases with chromosome 14 and 18 translocations. Bcl-2 inhibited BAX and BAK through direct protein-protein interactions (7).

Members of the Bcl-2 protein family are grouped together due to sharing one or more of BH domains (BH1-4), despite some members being pro-apoptotic and others anti-apoptotic. Pro-apoptotic members, proteins sharing BH3 domains, include BIM, BID, NOXA, PUMA, BAD, HRK, BMF, and BIK (8). Anti-apoptotic members include Bcl-2, Bcl-xL, Bcl-w, Mcl-1, and A1/BFL-1 (9). Pro-apoptotic and anti-apoptotic members of the Bcl-2 protein family interact through hydrophobic interactions of BH domains, with antagonism of the anti-apoptotic proteins by the pro-apoptotic proteins. The net effect is promotion or demotion of apoptosis. There are several models for the interactions among pro-apoptotic and anti-apoptotic Bcl-2 proteins, which are beyond the scope of this review (6).

Evasion of apoptosis promotes tumor initiation followed by progression (10). There has been increasing evidence that all malignancies possess some degree of apoptotic resistance, but mechanisms of this resistance may differ (11). Overexpression of anti-apoptotic proteins, including Bcl-2, has been noted in cancers, such as follicular lymphoma and has been tied to increased radioresistance and chemotherapy resistance (12). Downregulation or inhibition of pro-apoptotic members of the Bcl-2 family has been noted in many tumor types as well. For instance, in Burkitt lymphomas, pro-apoptotic BIM and PUMA are silenced through promoter hypermethylation (13). The premise of Bcl-2 inhibitors evolved out of a desire to inhibit anti-apoptotic members of the Bcl-2 family.

One of the first identified Bcl-2 inhibitors was gossypol, a natural phenol derived from cotton plants. In addition to inhibiting pro-survival Bcl-2 proteins, it can activate pro-apoptotic members of the Bcl-2 family, such as NOXA and PUMA (14). Gossypol has been implicated in other signaling pathways as well, including NF-kB regulation and MDM2-VEGF interactions (15). Though very broad-acting, gossypol has a significant adverse effects profile, including hypokalemia and infertility, and so utilization has been limited (16).

More specific agents for Bcl-2 inhibition, BH3 mimetics, have been developed with better understanding of Bcl-2 protein family structure and BH domains. The first of these agents, ABT-737, was developed by Abbott Laboratories based on analysis of the BAD protein BH3 region via NMR (17). ABT-737 targeted several members of the pro-survival Bcl-2 family (Bcl-2, Bcl-xl, and Bcl-w proteins) with high affinity. It demonstrated promise in animal models causing tumor regression, but is not orally bioavailable, leading to the development of navitoclax (18).

Navitoclax has been evaluated in phase I and II clinical trials and has clinical activity in small cell lung cancer and acute lymphoblastic leukemia but causes significant dose-dependent thrombocytopenia due to inhibition of Bcl-xL (19). Obatoclax has also been developed which targets all members of the pro-survival Bcl-2 family, but experiences issues with solubility and toxicity (20, 21).

ABT-199 or venetoclax was subsequently developed as a more selective Bcl-2-specific agent. Despite venetoclax lacking the pleiotropic inhibition capabilities of other members of the anti-apoptotic Bcl-2 group, venetoclax has demonstrated promise as the first-in-class Bcl-2 inhibitor approved by the FDA, first as a breakthrough therapy for relapsed CLL/SLL in 2015 (22). Since then, its approved uses have expanded to all CLL/SLL patients and to AML patients ≥ 75 years old (or with significant comorbidities) as part of a combination regimen with azacitidine, decitabine, or low-dose cytarabine (23). Venetoclax has minimal toxicity to platelets compared to navitoclax or obatoclax.

Chemoimmunotherapy with regimens including fludarabine, cyclophosphamide, rituximab, chlorambucil and/or obinutuzumab have traditionally been mainstays in the treatment of CLL (24). In recent years, PI3K inhibitors and BTK inhibitors have been introduced as targeted agents centered on inhibition of oncogenic signaling pathways. The development of venetoclax has continued this trend. CLL is notable for high levels of Bcl-2 expression. Notably in the most common 13q deletion CLL, this has been attributed to a loss of tumor suppressor microRNA (miR-15 and miR-16) expression. miR-15 and miR-16 directly target and inhibit Bcl-2 expression (25).

The first clinical trial for venetoclax was a phase I dose-escalation trial with 116 patients diagnosed with refractory CLL or small lymphocytic leukemia SLL (13). In this cohort, the overall response rate (ORR) was 79% with 20% achieving complete remission (CR), and 5% with no minimal residual disease (MRD). Response rates were 71-79% even among subgroups with poor prognostic features, such as resistance to fludarabine, 17p deletions, or unmutated IGHV. A manageable safety profile was noted as well. Though three of 56 patients experienced tumor lysis syndrome (TLS) in the dose-escalation cohort, none were noted in the expansion cohort after adjustments were made to the dosing schedule with a ramp-up from 20 mg to 200 mg. A subsequent phase II, single arm, multicenter study (M13-982) in 107 patients with 17p deletion demonstrated an ORR 79.4% and CR 8% (26). The most common adverse events were grade 3-4 neutropenia (40%), infection (20%), anemia (18%), and thrombocytopenia (15%).

Though venetoclax monotherapy induced significant response in patients with relapsed CLL, it was notable in these trials that CR occurred only for 8-20% of the cohort with median progression free survival (PFS) at 25 months, which suggested limited efficacy of venetoclax monotherapy. This observation has subsequently led to the development and study of various combination regimens. Current NCCN guidelines recommend venetoclax-obinutuzumab as a preferred first line regimen for CLL with and without del(17p)/TP53 mutation. Additionally, ibrutinib-venetoclax is also recommended first-line, followed by venetoclax-rituximab as preferred second/third line regimens. Several phase III trials summarized below have supported these recommendations ( Table 1 ).

CLL14, a multicenter, open-label, randomized phase III trial examined venetoclax-obinutuzumab vs chlorambucil-obinutuzumab in patients with untreated CLL (27). The venetoclax-obinutuzumab arm had significantly longer PFS than the chlorambucil-obinutuzumab arm (HR 0.31, 0.22-0.44, p <0.0001) and did not reach median PFS as opposed to 35.6 months PFS in the chlorambucil-obinutuzumab arm. Grade 3-4 neutropenia was common in both arms.

Another multicenter open-label, randomized phase III trial, GLOW, examined ibrutinib-venetoclax vs chlorambucil-obinutuzumab in patients ≥65 or those younger with cumulative illness rating scale (CIRS) > 6 (28). PFS was longer for the ibrutinib-venetoclax arm (median PFS not reached) than the chlorambucil-obinutuzumab arm (median PFS 21.0 months, 95% CI 16.6-24.7, HR 0.216, 0.131-0.357, p < 0.001). Overall improvement was consistent between patients ≥65 or those younger with CIRS > 6. Additionally, the proportion of patients with undetectable minimal residual disease (uMRD) from 3 to 12 months after treatment was greater in the ibrutinib-venetoclax arm. Side-effect profiles were consistent with expectations.

Finally, the MURANO trial examined venetoclax-rituximab vs bendamustine-rituximab in patients with relapsed/refractory CLL (29). Two-year rates of PFS were significantly higher in the venetoclax-rituximab arm (84.9% vs 36.3%, HR 0.17, 0.11-0.125, p < 0.001). This was seen across clinical and biologic subgroups. Though rates of grade 3-4 neutropenia were higher in venetoclax-rituximab (57.7% vs 38.8%), grades 3-4 febrile neutropenia (3.6% vs 9.6%), infection, and infestation (17.5% vs 21.8%) were lower.

The current first-line treatment for newly diagnosed AML is induction with cytarabine and anthracycline in what is known as a “7 + 3” regimen, followed by consolidation with 2-4 cycles of intermediate dose cytarabine or hematopoietic stem cell transplant (32). With this therapy, 60-80% of patients ≤60 years old achieve CR and only 40-60% of patients >60 years old achieve CR (33). The presence of small populations of leukemic stems cells (LSC) resistant to therapy is thought to contribute to relapse (34). This is a particularly significant issue in elderly patients and patients with poor functional status that cannot tolerate standard induction. LSCs are thought to be cell cycle quiescent compared to hematopoietic stem cells and overexpress Bcl-2, a rationale for evaluation of Bcl-2 inhibitors, such as venetoclax, as potential AML therapeutics.

The earliest studies to evaluate venetoclax monotherapy in AML was a phase II single arm trial in relapsed/refractory AML in 2016 which achieved ORR 19% (35). Venetoclax demonstrated pharmacologic effect by International Working Group Criteria in these patients with an acceptable safety profile. This study has been a foundation for later trials examining venetoclax combination regimens of venetoclax with hypomethylating agents (HMAs) or low-dose cytarabine (LDAC) for newly diagnosed AML in adults ≥75 years, or adults with comorbidities that prevent standard induction (36).

In initial phase Ib study of venetoclax-decitabine or azacitidine in patients ≥ 65 years old with untreated AML the regimen demonstrated an overall CR/CRi rate of 61% in three arms: venetoclax-decitabine, venetoclax-azacitidine, and a substudy group of venetoclax-decitabine with posaconazole (37). A follow-up phase II study narrowed down the dosing of venetoclax to 400 mg daily (38). Notably in this study, a subgroup analysis of patients with intermediate and poor risk AMLs demonstrated response with CR/CRi of 74% and 60% respectively and patients with TP53, IDH1/2, and FLT3 mutations had response with CR/CRi 47%, 71%, and 72%. A subsequent phase III study demonstrated that venetoclax-azacitidine was superior to azacitidine alone in patients ineligible for standard induction (median OS 14.7 months vs 9.6 months, HR 0.66, 95% CI 0.52–0.85, p = 0.001) (30). CR/CRi was also higher comparing venetoclax-azacitidine at 66.4% vs 28.3% (p = 0.001) in azacitidine alone. It was notable that the incidence of febrile neutropenia was higher in the venetoclax-azacitidine group (42% vs 19%).

In parallel, low dose cytarabine (LDAC) has been evaluated in combination with venetoclax in a phase Ib/II trial with adults 60 years or older or ineligible for intensive therapy with previously untreated AML (39). Forty-nine percent of these patients had secondary AML, 29% with prior HMA treatment, and 32% with poor cytogenic features. Fifty-four percent achieved CR/CRi with median OS 10.1 months and median duration of response (DOR) of 8.1 months. In the group of patients without prior HMA exposure, CR/CRi was documented in 62% and median OS was 13.5 months and DOR 14.8 months. This study showed venetoclax and LDAC combination produced a durable remission. A follow-up randomized placebo-controlled phase III trial in newly diagnosed patients ineligible for intensive induction did not meet primary endpoint, but analysis after an additional six-month follow-up showed improved survival for the venetoclax-LDAC arm (HR 0.70, 0.50-0.99, p = 0.04) (31). Key adverse events in the venetoclax-LDAC arm included febrile neutropenia (32%), neutropenia (46%), and thrombocytopenia (45%). These randomized trials ultimately led to FDA approval of venetoclax in combinations with hypomethylating agents or LDAC in older AML patients unfit for induction chemotherapy in 2018.

It is important to note that despite this progress for older AML patients, TP53 mutation status continues to be a challenge in treatment. Studies in the in vivo mouse models and structural biology findings have supported the conclusion that Bcl-2-induced apoptosis is regulated by p53 (40, 41). It has been shown that loss of or mutant TP53 increases the threshold of BAX/BAK activation resulting in resistance to venetoclax (42, 43). These findings have been consistent with clinical trial results. In a subset analysis of the patients with poor-risk cytogenetics and TP53 mutations in the phase III trial by DiNardo et al., it was noted that remission rates were improved for patients with poor-risk cytogenetics and TP53 mutations, but duration of remission and OS were not similar between venetoclax-azacitidine vs azacitidine cohorts (30, 44). The limitations of venetoclax in patients with the TP53 mutation has been demonstrated in studies with different HMAs as well; for instance, outcomes were noted to be significantly worse in patients with TP53 mutant status compared to TP53 wild-type (ORR 66% vs 89%, p = 0.002) on decitabine-venetoclax in a phase II study of 118 patients with new AML (45). These results collectively show that TP53 mutation status warrants consideration of alternative agents.

Given the overall impact of venetoclax in older AML patients and pre-clinical studies indicating synergy with chemotherapy, combinations of venetoclax as part of standard induction are currently being explored too. A phase Ib/II trial examining venetoclax combined with FLAG-Ida induction in 68 patients with newly diagnosed or refractory AML demonstrated deep remissions and high rates of transition to autologous stem cell transplant (ASCT) (56%) (46). A follow-up report of 45 patients with extended follow-up demonstrated high response rates with MRD-negative CR rates >90%, along with estimated 24-month OS 76% (95% CI, 62%–94%). Other studies have demonstrated support for venetoclax with cladribine, idarubicin, and cytarabine (CLIA) and venetoclax with cytarabine and idarubicin (47, 48). Venetoclax combined with 7 + 3 induction chemotherapy is also being explored in a large, randomized study of the HOVON-AALG intergroup (NCT04628026).

In myeloma, the t(11;14) translocation found in 15-20% of myeloma patients has been associated with Bcl-2 overexpression and higher Bcl-2 to Mcl-1 ratio (49). Sensitivity to venetoclax has been shown in pre-clinical data with certain myeloma cell lines as well and further attributed to B cell patterns of gene expression (50). Additionally, it has been shown that other myeloma therapies modulate Bcl-2 expression. Dexamethasone promotes a Bcl-2 dependent state, and bortezomib and carfilzomib are thought to reduce Mcl-1-mediated venetoclax resistance (44, 51).

A phase III study, BELLINI, examined bortezomib and dexamethasone with venetoclax vs with placebo in 291 patients with relapsed or refractory MM (52). Seventy-nine percent of these patients had high Bcl-2 expression. Median PFS was higher at 22.4 months vs 11.5 months (HR 0.63, 0.44-0.90, p=0.01). However, increased mortality was seen in the venetoclax group with eight cases (4%) treatment-emergent fatal infections noted and none in the placebo group.

A phase I/II, multi-center, dose escalation trial (NCT03314181) examining venetoclax-daratumumab-dexamethasone in patients with t(11,14) relapsed/refractory MM has recently finished enrollment for part three as of May 31, 2022 with preliminary data demonstrating durable responses: 24-month PFS was 94% (95% CI, 63.2-99.1%) and 83% (95% CI, 27.3-97.5%) in the Ven400Dd and Ven800Dd arms (53).

Venetoclax has been assessed in various subtypes of NHL, but activity was only modest. A phase I study of venetoclax in patients with relapsed/refractory NHL (mantle cell lymphoma (MCL), follicular lymphoma (FL), diffuse large B-cell lymphoma (DLBCL), Waldenström macroglobulinemia (WM), marginal zone lymphoma) showed that venetoclax was well-tolerated. Anemia, neutropenia, and thrombocytopenia were the most recorded adverse events. The response rates for MCL, FL, and DLBCL respectively were 75%, 38%, and 18%. The median PFS was estimated at 14, 11, and 1 month respectively. Of the subtypes of NHL studied, the response in MCL was notable, but durable response on monotherapy was poor with CR 21%. Subsequently, a phase II study of ibrutinib and venetoclax in 24 patients with primarily relapsed/refractory MCL (23/24) demonstrated an impressive 71% CR overall (54).

The focus of this review has been on Bcl-2 inhibitors and the first-in-class approved agent venetoclax, but the family of Bcl-2 proteins is broad with other members demonstrating potential as therapeutic targets, such as Bcl-xL and Mcl-1.

It has been shown that cutaneous t-cell lymphomas (CTCL) express high levels of Bcl-xL (55). Furthermore, several AML subsets such as erythroleukemia and megakaryoblastic leukemia, overexpress and are functionally dependent on Bcl-xL for survival (56). The proteolysis-targeting chimera (PROTAC) DT2216 was developed to target Bcl-xL with the design taking advantage of engagement with VHL E3 ligase that is not expressed in platelets. This directs the PROTAC to malignant cells, addressing thrombocytopenia concerns noted with previous inhibitors of Bcl-xL, such as navitoclax (57). DT2216 has been shown in preclinical mouse models to selectively kill various Bcl-xL-dependent lymphoma and leukemia cells, supporting further clinical testing.

Mcl-1 has been implicated as an apoptotic gatekeeper in TCL, myeloma, and AML. Mcl-1 is expressed in every differentiation step of B-cells and in most MM cells (58). Increased Mcl-1 is also one of the key mechanisms of Bcl-2 inhibitor resistance. It has been shown that AML blasts in recurrent disease have higher levels of Mcl-1 and that Mcl-1 inhibition reverses Bcl-2 inhibitor resistance, potentiating venetoclax activity (58, 59).

Clinical trials of Mcl-1 inhibitors have been limited so far despite potential in preclinical studies, such as AZD5991 (55). Preliminary results from 26 patients with relapsed MM in a phase I trial for an intravenous Mcl-1 inhibitor (AMG-176) demonstrated anti-tumor effect (CR, partial CR) in only 3 patients (60). Concerns have additionally been raised about cardiac toxicity of other Mcl-l-inhibitor candidates, such as AMG-397 (61).

With Bcl-2 and Mcl-1 inhibitors targeting the intrinsic pathway of apoptosis, another approach being considered is targeting the extrinsic pathway of apoptosis (62). The IAP family of proteins, including XIAP and cIAP, inhibit apoptosis through the extrinsic pathway and overexpression has been noted to cause poor outcomes in AML. AEG35156, an antisense oligonucleotide inhibitor of XIAP-induced apoptosis of malignant cells in vitro, but in subsequent clinical trials has failed to meet goals (63, 64). Other forms of IAP inhibition, such as cIAP inhibitors and survivin inhibitors have suffered similar fates, where trials in humans have not yielded adequate results (65, 66).

Further upstream in the extrinsic pathway, TRAIL and other ligands with associated death receptors, such as Fas, TNF-R, TRAIL-R, can trigger death in malignant cells. Various recombinant TRAIL and TRAIL-R agonists have been developed in this context for cancer treatments. Though with good safety profiles, initial candidates have had limited effects. This was attributed to acquired TRAIL-resistance and evidence that oligomerization and anchoring of antibodies may be necessary (67). A first-in-human trial of Abbv-621, a TRAIL receptor agonist fusion protein, in 17 patients with refractory/relapsed AML and DLBCL demonstrated antitumor activity in one patient who was able to achieve CR. Hepatoxicity, increased ALT and AST, was the most common treatment related adverse event, noted in 34.8% of patients (68, 69).

JW: Conceptualization, Writing – original draft, Writing – review & editing. MK: Conceptualization, Writing – original draft, Writing – review & editing.