The understanding of the BCL2 oncogene and the BCL-2 protein family have motivated the current advancements in cancer therapeutics that target different members of apoptotic regulatory networks.
Every day, the average adult loses approximately 60 billion cells due to apoptosis.1 Although apoptosis facilitates the cellular turnover that is required to maintain homeostasis and plays important roles in development, misregulated apoptotic processes can contribute to disease. Indeed, a disruption of the balance between cell death and proliferation is a feature of many cancers. Apoptosis is controlled by a tightly regulated process that is conserved across metazoans and involves proteins related to the B-cell lymphoma 2 (BCL-2) family. Characterization of the BCL2 oncogene and the BCL-2 protein family led to the current appreciation of the significant role that apoptosis plays in many types of malignancies. This understanding has motivated the current advancements in cancer therapeutics that target different members of apoptotic regulatory networks.
The BCL2 Gene Family
The gene encoding BCL-2 was cloned from the breakpoint of the common chromosomal translocation that is observed in patients with follicular lymphoma (FL) t(14;18).2 The suspected oncogenic properties of this gene were confirmed when BCL-2 overexpression was shown to confer a survival advantage to cells, somehow preventing them from undergoing programmed cell death.3 The ability of BCL-2 to block cell death signaling was a novel oncogenic feature; at the time, all the known oncogenes caused cancer by interfering with cell proliferation mechanisms. The BCL-2 family members have since been implicated in both proapoptotic and antiapoptotic roles and are established master-regulators of apoptosis.
Discovery and Characterization
Since the initial characterization of BCL-2 more than 3 decades ago, at least 30 other BCL-2 family members have been identified. The defining feature of the BCL-2 family members is the presence of ≥1 BCL-2 homology (BH) domains. The BCL-2 family members fall into 3 subcategories, based on function and structural features: (1) the prosurvival members, (2) the proapoptotic members, and (3) the BH3—only members (Figure).4
Known helical and transmembrane regions are indicated, as well as BH domains 1-4.
Figure reprinted with open access from Adams JM, Cory S. Cell Death Differ. 2018;25(1):27-36.
In addition to BCL-2, the human prosurvival family members include BCL-XL, BCL-W, MCL-1, and BFL1.5 These proteins each harbor 4 BCL-2 homology domains (BH1-BH4). Together, these homology domains form the tertiary structure that interacts directly with the proapoptotic family members. BCL-2 and the other prosurvival proteins are localized on the mitochondrial outer membrane and promote cell survival by binding to and sequestering the proapoptotic family members.
The proapoptotic members also contain multiple BH domains and have a structure that is similar to that of the prosurvival members until they are activated, upon which they undergo conformational change.6 The proapoptotic family members, also referred to as pore-formers for their role in mitochondrial outer membrane permeabilization (MOMP), include BAX, BAK, and BOK. BAX was discovered by an immunoprecipitation experiment searching for proteins that interact with BCL-2 and, upon characterization, was the first indication that some BCL-2 family members perform a proapoptotic role.7 Two years later, BAK was discovered via a polymerase chain reaction screen searching for genes encoding BH-domain—containing proteins.8 BAX and BAK interact with different members of the prosurvival group to contribute to apoptosis regulation. Cells missing either protein are largely unaffected, but the disruption of both proteins interferes with apoptosis, as mice with both genes knocked out maintain cells that would otherwise undergo apoptosis during development.9 The function of BOK remains unclear.10
BH3-only family members are proapoptotic proteins that include BIM, PUMA, BAD, BMF, BID, NOXA, HRK, and BIK. These members are more distantly related to each other; the similar feature they share is the BH3 domain.11 This subcategory is further divided into activator and sensitizer proteins; activators can directly bind to and activate BAX/BAK, whereas sensitizers bind to and inhibit antiapoptotic proteins. These proteins are natively unfolded and adopt structures upon interaction with binding partners, except BID and probably BIK.12
Decades of research have led to a detailed understanding of the mechanisms of BCL-2 family-mediated apoptosis. Studies have used a range of systems, from structural biology experiments to clinical samples, to gain insight into the network of interactions among BCL-2 family members. Together, these data reveal the key role BCL-2—related proteins play in cellular turnover in healthy and disease states.
Sequence-based analysis initially indicated that BCL-2 is an integral membrane protein; the carboxy-terminus (C-terminus) of the protein embeds into a membrane while the functional portion is exposed to the cytoplasm.13 Cell-based experiments subsequently revealed that BCL-2 is a component of the mitochondrial outer membrane, the endoplasmic reticulum, and the nuclear envelope.14
Initial experiments with BCL-2—expressing cells revealed the prosurvival activity of the protein. Transduced interleukin-3 (IL-3)–dependent cells that express BCL-2 do not proliferate or die when cultured without IL-3, remaining instead in a G0 phase, with the cells resuming proliferation if IL-3 is added to the media.15 Although the oncogenic implications of this prosurvival activity were apparent, cells that overexpress only BCL-2 did not develop tumors when injected into mice.15 Cells expressing both BCL-2 and MYC, however, do induce lymphoma in mouse models.15,16
Additionally, BCL-2—expressing transgenic mice accumulate B lymphocytes, and when these cells are removed from the animals and cultured, they exhibit a 14-day survival rate that is 500-fold greater than those of normal littermates.17 Further indicating the role played by BCL-2 in preventing apoptosis, BCL2 mice exhibit elevated levels of cell death, particularly in lymphoid organs. This phenotype is rescued in mice that are also unable to express BIM.18,19 In addition, some BAX and BAK mice are able to survive until adulthood but exhibit morphologic defects in areas that require cell death for their development.9
BCL-2 overexpression blocks the apoptosis triggered by a variety of cellular stresses and can protect cells from apoptosis triggered by anticancer agents.3,20 These studies highlight an important feature of numerous cytotoxic anticancer therapies: Many execute their antitumor effects by triggering apoptosis. Cancer cells that are protected from apoptosis are thus able to evade the effects of such therapies.
Collectively, these experiments emphasize the antiapoptotic activity of BCL-2 and the central role played by BCL-2 family members in regulating apoptosis. Over time, research has revealed a complex web of interactions among the BCL-2 family members that controls the balance between cellular life and death.
Regulation of Mitochondrial Apoptosis
Much of the early insight into the function of the BCL-2 family protein was derived from experiments performed in the nematode Caenorhabditis elegans, in which the details of programmed cell death mechanisms were more clearly understood. C elegans experiments revealed that BCL-2 expression mimicked the phenotype of a gain-of-function mutation of the CED-9 gene.21 CED-9 was shown to inhibit CED-4, a caspase-inhibiting protein that is a homologue of the human apoptotic protease-activating factor 1 (APAF1). This connected BCL-2 to the caspase cascade, but a key difference between the CED-9— and BCL-2–mediated cell death pathways is the involvement of the mitochondria.
The knowledge of mitochondrial localization of BCL-2 paved the way for investigators to discover the connection between BCL-2 family signaling and mitochondrial-mediated apoptosis.22 Eventually, it was established that BCL-2 interacts with BAX to control pore formation in the mitochondrial outer membrane, ultimately leading to activation of the caspase cascade and subsequent cell death.3 The balance between cellular life and death was initially thought to be controlled by BCL-2 concentrations, but it is becoming more evident that the local concentration in the mitochondrial membrane and the binding kinetics of the individual family members add another layer of precision and complexity to this regulation.23
Two modes have been identified via which BCL-2 family members negatively regulate apoptosis. In inhibition mode 1, prosurvival proteins bind to the BH3 domain of proapoptotic BH3-only proteins.24 This sequesters the proapoptotic proteins and prevents further apoptosis signaling.25 In inhibition mode 2, prosurvival proteins bind to a proapoptotic effector protein, BAX or BAK, blocking downstream signaling.24 These modes are executed simultaneously.
When required, proapoptotic signaling is initiated via a variety of transcriptional and posttranslational modifications of BH3-only proteins. These family members are thought to serve as sensors that detect cues indicative of cellular stress and trigger an apoptotic response when necessary. The relatively large number of BH3-only proteins facilitates precise control over the responses to cellular assault. Understanding the specialized functions of the BH3-only proteins can provide additional avenues for therapeutic development. For example, it has been shown that loss of BIM decreases the toxicity of such cytotoxic agents as paclitaxel.26
BH3-only proteins induce MOMP and subsequent cell death through both direct and indirect means. Sensitizer BH3-only proteins bind to and inhibit the prosurvival family members by blocking their BH3-binding groove, thus indirectly promoting MOMP by releasing BAX and BAK from inhibitory interactions.23,27 Activator BH3-only proteins (BIM, tBID, PUMA) bind directly to BAX/BAK, recruiting them to the mitochondrial outer membrane and triggering conformational changes that result in the formation of homodimers that subsequently oligomerize to form pores.23 The mechanistic details of BAX/BAK pore formation itself remain unclear.
Once signaling has begun to favor the formation of BAX and BAK oligomers, the cell is committed to death. Because of pore formation, apoptogenic proteins, particularly cytochrome c, are released into the cytoplasm. Cytochrome c subsequently activates APAF1, which leads to caspase 9 activation and initiation of the caspase cascade. This ultimately results in cell death.28 This apoptosis activation mechanism has been termed the intrinsic pathway. This contrasts with the extrinsic pathway, which involves signaling via the tumor necrosis factor-α and Fas receptors.
Molecular Mechanisms of BCL-2—Mediated Apoptotic Signaling
The BCL-2 family members form an elegant signaling network that controls the balance between cellular life and death. The molecular and kinetic details of these interactions are key to understanding how this balance is maintained and are areas of active research. Structural biology experiments investigating BH domains with and without binding partners have shed light on the important structural features that mediate the interactions among different family members. Kinetic studies are also revealing the importance of protein folding and the role played by the mitochondrial membrane in this signaling network. Together, these pieces of data are improving our understanding of apoptotic mechanisms in both healthy and disease states.
Nuclear magnetic resonance spectroscopy and x-ray crystallography experiments indicate that BCL-2 family members are globular proteins that are mostly α-helical. BCL-2 and its closest homologues, BCL-XL and BCL-W, share a conserved structure that consists of 5 α-helices surrounding a hydrophobic core of 2 additional helices. The transmembrane domain resides at the C-terminus of the protein. The BH3-binding groove is formed by amino acids from BH1, BH2, and BH3. Helices within these domains form a groove into which the BH3 domains from other BCL-2 family members dock. This groove—helix interaction is the molecular basis for BCL-2–mediated regulation of apoptosis.29
The antiapoptotic BAX, BAK, and BOK also share a similar α-helical core structure with a C-terminal transmembrane domain. Interestingly, the BAX C-terminal tail can bind to its hydrophobic groove, but the binding runs in the opposite orientation of BH3 ligands.30 Competitive interactions at these groove sites are hypothesized to play a key role in controlling the activation of apoptosis and thus the determination of cell fate.
Over the past 3 decades, it has become well established that the BCL-2 proteins regulate cell death by interacting physically with each other. It is becoming increasingly apparent, however, that the local concentration, the binding kinetics, and the membrane interactions are key components that contribute to the execution of these interactions. Early experiments performed with peptides and truncated versions of the BCL-2 family members provided only pieces of the story. Experiments with full-length proteins are critical to obtaining the entire picture of the regulatory interactions that determine the life-to-death switch.23
BH3-only proteins are the most distantly related of the BCL-2 family members, united by their shared BH3 domain. Unlike the globular BCL-2 and BAX-like family members, the BH3-only proteins appear to be intrinsically disordered proteins that fold upon interaction with binding partners.12 These proteins are the initial sensors for cellular stress and can be stimulated by a variety of posttranslational modifications and protein—protein interactions that control this folding. The BH3 domain fold itself harbors a surface-exposed hydrophobic patch that mediates protein–protein interactions with the hydrophobic grooves of other BCL-2 family members.31
This molecular insight into the intrinsic regulation of apoptosis provides key details on the complex networks that control cell fate in the presence of cytotoxic insults. As kinetic and molecular descriptions of these interactions improve, efforts to develop more effective BCL-2—targeting antitumor therapeutics will become more powerful.
The Significance of BCL-2
From the initial discovery of BCL-2 in patients with FL, this family of proteins has been implicated in a variety of malignancies, including those of blood, breast, and lung tissue. In addition to their roles in the development and persistence of malignant states, BCL-2 family members also have been shown to play a role in cellular resistance to cytotoxic agents. Studies designed to evaluate the roles of BCL-2 family members in these processes are improving our understanding of the cellular mechanisms of cancer and are aiding in the development of new anticancer therapies. The section that follows will focus on involvement of the BCL-2 family members in the pathogenesis and therapeutic resistance of hematologic malignancies. BCL-2 family proteins implicated in hematologic malignancies are also summarized in the Table.2,32-39
BCL-2 in Hematologic Malignancies
The gene encoding BCL-2 was discovered within a commonly occurring translocation of chromosomes 14 and 18 in patients with FL.2 This translocation places BCL2 gene expression under control of the immunoglobulin heavy locus enhancer, which drives overexpression of the inhibitory protein and protects the lymphoma cells from apoptosis. This protection from cell death is evidenced by the accumulation of B cells in patients with non-Hodgkin lymphomas.
BCL-2 family members also have been shown to play a key role in the development and pathogenesis of chronic lymphocytic leukemia (CLL). In the early 1990s, it was observed that some patients with CLL express high levels of BCL-2 in the absence of chromosomal translocation, sequence alterations, or abnormal DNA modifications.32 This observation led to discovery of the transregulation of the BCL2 gene, as well as the roles played by the micro-RNAs (miRNAs) miR-15a and miR-16.1. These miRNAs function in healthy cells by suppressing BCL-2 expression.33 Most patients with CLL have deleted or silenced miR-15a and/or miR-16.1.33 This discovery was the first instance in which genes that encode miRNAs were shown to play a role as tumor suppressors. Ongoing efforts are underway to identify other roles for miR-15a and miR-16.1, to determine whether these miRNAs may serve as useful therapeutic targets. In addition to BCL-2, MCL-1, BAG-1, BAX, BAK, and caspase 3 are also expressed in patients with CLL and may prove to be useful avenues for the development of antitumor agents.40
BCL2 gene amplification has been observed in patients with diffuse large B-cell lymphoma (DLBCL) or small lymphocytic lymphoma (SLL).35 Recent data demonstrate that DLBCL cells become more dependent on BCL-2 signaling if the PI3K pathway is inhibited. This indicates that combination therapies targeting both pathways may be effective in patients with DLBCL.41
BCL-2 family members have also been shown to influence the development of T-cell cancers. Transgenic mice that overexpress BCL-2 have an increased incidence of peripheral T-cell lymphoma.38 Interestingly, T-cell leukemia exhibits dependency on BCL-2 family members that changes as the cells differentiate, with immature cells relying on BCL-2 activity and more mature cells relying on BCL-XL activity. These findings highlight key details on the involvement of BCL-2 family members and demonstrate that as it matures, a malignancy can alter its dependence on antiapoptotic protein activity.39
Exploitation of apoptotic pathways has also been observed in preclinical studies of patients with acute lymphoblastic leukemia (ALL). One study, which was performed in ALL cell lines, used the BCL-2 inhibitor ABT-737 to determine whether these cells exhibit high levels of dependence on BCL-2 for survival.42 This system has since been utilized in many preclinical studies, to identify useful combination therapies that are effective against BCL-2—dependent malignancies.
In addition to BCL-2, the antiapoptotic protein MCL-1 has been shown to contribute to the survival of ALL cells. In a 2013 study, a mouse model for Philadelphia chromosome B-lineage ALL was developed that provided investigators with the ability to control expression of the MCL-1 gene in transformed B cells that express BCR-ABL. These experiments revealed a strong dependence of these cells on MCL-1 expression, both for the development and persistence of BCR-ABL—positive leukemia.36
As major determinants of cell fate, the BCL-2 family members play a key role in the pathogenesis of a variety of human cancers. As more details are uncovered that implicate these family members in different malignancies, new opportunities will emerge to develop targeted therapies and treatment plans to more effectively fight cancer.
p53 and BCL-2 Signaling in Cancer
The TP53 gene is mutated in half of all human cancers and executes many interactions across the cell and crosses paths with BCL-2 family signaling. Experiments performed in leukemic cell lines using a temperature-sensitive gain-of-function p53 mutant showed that p53 triggers apoptosis.43 BCL-2 expression in these cells prevents cell death, however, indicating that TP53 works upstream of BCL-2 signaling.44 The genes encoding PUMA and NOXA have since been identified as directly interacting with p53 to trigger apoptosis.44
The integrity of the TP53 gene is also related to BCL-2 family protein function and can be an important indicator of those patients who will respond to therapy. TP53 disruption is associated with poor prognosis and therapeutic resistance in a variety of cancers. Mutation or deletion of the TP53 gene (eg, by losing parts of chromosome 17) decreases the responsiveness of BCL2 family genes to cellular stress.28
In addition to supporting the persistence of unhealthy cells, disruptions in BCL-2 signaling confer resistance to therapeutics. This highlights the fact that some therapies do not kill cells outright but instead trigger apoptotic switches that misregulated BCL-2 family members can circumvent. Many cancer cells are faulty and fragile and use the BCL-2 family proteins as a shield against apoptosis. The next section will cover the progress that has been made in the development of therapeutics that remove this crutch and restore a malignant cell’s apoptotic pathways. These efforts are showing encouraging promise in the fight against a variety of cancers.
Decades after the discovery of the BCL-2 family proteins, the Letai research group developed a method called BH3 profiling. In this strategy, cells are exposed to a panel of BH3 peptides and their unique apoptotic response profiles are correlated with the cell’s dependence on antiapoptotic proteins for survival and consequent sensitivity to BCL-2 inhibition.45 Using this method, the group found that certain cancer cells are “addicted” to antiapoptotic signaling, relying on the expression of BCL-2 family members for protection from apoptosis. These cells are referred to as being “primed for death” and require BCL-2 family signaling to prevent MOMP.45 This finding demonstrated that some cancers may be particularly sensitive to therapeutics that disable antiapoptotic BCL-2 signaling.
Because of the survival advantage that BCL-2 family signaling confers to several cancers and the apparent reliance of malignant cells on BCL-2 signaling, this protein family has been an attractive target for the development of anticancer therapeutics. Inhibitors that disrupt antiapoptotic signaling and restore a cancer cell’s suicidal abilities thus have been a major focus of this effort.
BH3 mimetics are the major class of BCL-2 family protein inhibitors. These small molecules are designed to chemically mimic the BH3 surface that interacts with the antiapoptotic family members. As do their template BH3-containing proteins, these compounds bind to the hydrophobic groove of antiapoptotic proteins such as BCL-2, blocking their ability to sequester their proapoptotic binding partners. Upon this interaction, the liberated proapoptotic proteins are then able to trigger MOMP and cell death.
Data from structural biology experiments have shown that the interaction surface between BH3 domains and their prosurvival binding partners is hydrophobic and shallow and that the proteins have a high affinity for each other. These features render it challenging to develop a therapy that can separate them.28 With the addition of protein structural data that describe the molecular interactions between proapoptotic and antiapoptotic family members, and the advent of new structure-based drug discovery techniques, compounds that effectively target BH3-binding grooves began to emerge and progressed through clinical trials, with varying levels of success. In 2016, venetoclax was approved by the FDA as a second-line treatment for CLL associated with 17p deletion.46 In 2018, the indication for venetoclax was expanded to the treatment of patients with CLL or SLL, with or without 17p deletion, who have received ≥1 prior therapies.
The BCL-2 family proteins exert tight control over cell fate in response to stress. In a healthy environment, a complex network of interactions by and among the BCL-2 family members sense and communicate apoptotic triggers, initiating cell suicide when appropriate. As is true with many master regulators of cellular processes, these pathways can be harnessed by cancer cells to confer a survival advantage. Despite this, new therapeutics that mimic the BCL-2 family BH3 domain are effectively restoring apoptotic signaling in malignant cells. These BH3 mimetics are progressing through clinical trials and are showing promise for the treatment of several types of cancer. The next article reviews the practical and therapeutic implications of BCL-2 inhibition in various hematologic malignancies.