Bcl-2 family regulation of neuronal cell fate plays an important role in normal mammalian nervous system development and in determining the pathological consequences of neurotoxic stimuli. However, a detailed understanding of how the Bcl-2 family accomplishes this regulatory task in the nervous system has defied simplistic explanation. Following the identification of Bcl-2's potent antiapoptotic action and the discovery that Bcl-2 heterodimerized with a proapoptotic homologue termed Bax, Korsmeyer and colleagues proposed that the intracellular ratio of Bcl-2 to Bax acted as a rheostat that determined the cellular response to death stimuli. This hypothesis has in general stood the test of time; however, the number of interacting proapoptotic and antiapoptotic proteins is exceedingly large and includes both Bcl-2 family members and nonmembers. Cell-specific, stimulus-dependent, and developmentally regulated expression of Bcl-2 family members combined with functional redundancy and overlapping death pathways continue to challenge cell death investigators and the development of Bcl-2 family-related neurotherapeutics.

Introduction

Following the discovery of Bcl-2 as a novel prosurvival oncogene in B-cell lymphoma,1 seminal advances were made by Korsmeyer and others in describing its pivotal role in regulating apoptotic cell death. For instance, studies investigating protein–protein interactions determined that at baseline, that is prior to receipt of a death stimulus, the antiapoptotic protein Bcl-2 heterodimerizes intracellularly with the proapoptotic protein Bax.2 When overexpressed, Bcl-2 forms homodimers concomitant with an inhibition of stimulus-induced apoptosis. Conversely, overexpression of Bax causes its homodimerization and an increased sensitivity to apoptotic stimuli. These findings led in part to the initial Bcl-2/Bax ‘Rheostat’ model of cell death3 (Figure 1). Since its inception, this simple but elegant model has grown increasingly complex, in part through the discovery of additional pro- and antiapoptotic Bcl-2-related molecules.

Figure 1
figure 1

Rheostat model of Bcl-2/Bax dimerization, adapted from original published version.3 In response to a death stimulus, the life or death of a cell is dictated in part by the ratio of Bax to Bcl-2, such that the predominant homodimerization of Bax promotes apoptotic death while the predominant homodimerization of Bcl-2 inhibits apoptotic death

Full size image

Bcl-2 family members are classified into three distinct subgroups based on their composition of Bcl-2 homology (BH) domains and by function (antiapoptotic versus proapoptotic). The presence of one to four BH domains thus defines membership in the Bcl-2 family. In general, BH1–3 domains are considered important for protein interactions within the Bcl-2 family. Homodimerization, heterodimerization or oligomerization of Bcl-2 family members occurs via interactions among BH domains, and have been shown to critically regulate apoptosis induction.

Many studies have shown that Bcl-2 family members regulate central nervous system (CNS) programmed cell death and may play a role in CNS injury and neurodegenerative disease. Thus, the Bcl-2 family remains a major focus for the development of pharmacological therapeutics to treat brain injury or neurodegeneration.4 In this review, we summarize some of the currently recognized complexities of Bcl-2 family regulation of normal and pathologic CNS death.

Antiapoptotic Subgroup: It all Began with Bcl-2

The antiapoptotic Bcl-2 family subgroup includes Bcl-2, Bcl-XL, Bcl-w and myeloid cell leukemia-1 (Mcl-1). Bcl-2, Bcl-XL and Bcl-w possess all four BH domains; BH1 and BH4 are considered important in maintaining their antiapoptotic activity.4 Bcl-2 is expressed in the nervous system and its targeted deletion induces the profound loss of peripheral motor neurons, sensory neurons and sympathetic neurons.5, 6 CNS neurons are largely spared in Bcl-2 deficiency, due most likely to relatively low basal expression of Bcl-2 in adult brain.6 In many neuronal populations, the more critical antiapoptotic Bcl-2 family member appears to be Bcl-XL, which was identified through cloning as a Bcl-2-independent regulator of cell death.7 Bcl-XL is expressed at very low levels in neural precursor cells but is upregulated upon their migration from the ventricular zone and differentiation, and remains high in mature, postmitotic CNS neurons.8 Bcl-XL-deficient mice exhibit massive apoptotic death of immature neurons throughout the developing nervous system and die around embryonic day 13.5 from hematopoietic cell failure.9 Numerous studies have revealed an important role for Bcl-XL in promoting neuron survival in both the developing and adult CNS.

Similar to Bcl-XL, Bcl-w was cloned by homology screening to Bcl-2 and is expressed in both the developing and adult rodent nervous systems, with highest levels of Bcl-w expressed in adult brain.10, 11 A comparison of Bcl-w and Bcl-XL in sensory neuron survival suggests that unlike the importance of Bcl-XL in programmed neuron death, Bcl-w plays a greater role in maintaining survival at later ages.12 This may explain in part why Bcl-w-deficient mice do not exhibit any obvious developmental CNS abnormalities.11 Mcl-1 was isolated originally from differentiating myeloid cells.13 Unlike Bcl-2 and Bcl-XL, Mcl-1 lacks the ‘antiapoptotic’ BH4 domain of Bcl-2 and Bcl-XL yet still possesses potent antiapoptotic activity that may be related to its interaction with proapoptotic Bcl-2 family members.14 The expression of Mcl-1 increases upon trophic factor stimulation and may subserve an antiapoptotic function.14 Conversely, the caspase-specific cleavage of Mcl-1 effectively converts the C-terminal fragment of Mcl-1 into a potent prodeath molecule.14 Mcl-1 is also effectively degraded by the proteasome.14 Although early immunohistochemical analysis detected expression of Mcl-1 in human sympathetic neurons, little if any expression was detected in human brain and spinal cord.15 However, abundant expression of Mcl-1 was shown recently in neural precursor cells, suggesting a possible role for Mcl-1 in regulating the survival of this early CNS population.16

Proapoptotic Multidomain Subgroup: Presenting the Rheostat Model

The proapoptotic, multidomain subgroup of the Bcl-2 family consists of Bax, Bcl-2-homologous antagonist/killer (Bak), Bcl-2-related ovarian killer (Bok) and Bcl-XS. Each of these proteins share BH1–3 domains except for Bcl-XS, which possesses only BH3 in addition to BH4. This subgroup is defined functionally by its ability to induce cytochrome c release and subsequent apoptotic death, possibly by formation of the mitochondrial permeability transition pore.17

The isolation of Bax as a proapoptotic molecule that heterodimerized with Bcl-22 provided inspiration for the initial rheostat model (Figure 1). Bax was shown subsequently to heterodimerize with Bcl-XL, which also inhibited its proapoptotic function.18 The interaction between Bax and Bcl-XL clearly regulates immature neuron apoptosis during development since the absence of Bax completely abrogates the increased neuronal apoptosis observed in the Bcl-XL-deficient nervous system.19 Apoptotic stimuli induce the dissociation of Bax from Bcl-XL and/or Bcl-2, which allows for its mitochondrial association and induction of the intrinsic apoptotic pathway. Bax-deficient mice exhibit a marked decrease in programmed neuron death and resultant increased numbers of postmitotic neurons in areas such as the brainstem, cerebellum, dorsal root ganglia, hippocampus and spinal cord.20 However, Bax-deficient mice also show increased numbers of atrophic neurons with atrophic axons, suggesting that prevention of neuron death does not guarantee functional recovery.21

The identification of Bak was reported concurrently by several different laboratories.4 Bak is structurally very similar to Bax, which suggests functional overlap. In some systems, both Bak and Bax are required to promote apoptotic cell death. For instance, combined Bax and Bak deficiency prevented DNA damage-induced death of neuronal precursor cells, whereas their single deficiency did not.22 In many neuronal cell populations, however, Bak does not seem to play a critical role in the induction of apoptosis. For example, the release of cytochrome c was prevented in cultures of peripheral neurons deficient in Bax but not Bak.23 Together, these findings suggest that while similar in structure, the relative contribution of Bax and Bak to apoptosis is often stimulus and cell-type specific. Bcl-XS, formed by alternative splicing of the bcl-X gene,7 is expressed at low levels in the mammalian nervous system8 and thus is thought to contribute minimally to CNS-regulated apoptosis. Finally, the role of Bok in regulating apoptotic neuron death remains to be determined.

Proapoptotic, BH3 Domain-Only Subgroup: From Rheostat to Real Complex

The proapoptotic BH3 domain-only subgroup includes Bcl-2-associated death protein (Bad), Bid, Bcl-2-interacting mediator of cell death (Bim), Bcl-2 Nineteen kilodalton interacting protein (BNip), death protein 5/harakiri (DP5/Hrk), Noxa and p53-upregulated modulator of apoptosis (Puma). Although this subgroup is diverse, regulation of cell death by the entire group relies critically upon the BH3 domain and the presence of Bax and/or Bak.24 Both prosurvival and prodeath signaling molecules regulate apoptosis through the coordinated expression and/or function of proapoptotic, BH3 domain-only proteins. BH3 domain-only proteins are thought to induce death either by interacting with and thus preventing the antiapoptotic activities of Bcl-2 and/or Bcl-XL25, 26, 27, 28, 29 or by increasing their interaction with proapoptotic Bax.30 Conversely, this interaction has been proposed to sequester BH3 domain-only proteins from Bax,24 thus preventing apoptosis.

Bad was isolated originally by its heterodimeric binding of both Bcl-2 and Bcl-XL25 and is described as a latent death factor whose proapoptotic function requires dephosphorylation, cleavage or splicing.31, 32 Bad-deficient mice develop normally and display no obvious neuropathological abnormalities.23 Bad is expressed in the adult nervous system, but at levels lower than during neuronal development.33 Bad is also proposed to maintain cell survival and to regulate glycolytic metabolism.31, 32, 34 Bid was isolated originally by its heterodimerization with Bcl-2 or Bax.26 Bid-deficient mice develop normally and show no neuronal abnormalities.35, 36 Bid is unique among Bcl-2 family members in that it acts as a cleavage substrate for caspase 8, thus being the only member of the Bcl-2 family that links directly the extrinsic or death receptor-mediated apoptotic pathway with mitochondrial release of cytochrome c.23 The interaction of Bid with Bax is thought to permit the Bcl-XL-inhibitable formation of a large mitochondrial pore to allow the release of mitochondrial proteins into the cytosol.37 Bim, identified originally as a Bcl-2-binding protein exists as three major splice variants, BimEL, BimL and BimS.23 Recent studies have indicated isoform-specific binding of Bim to Bax.29 In addition to its stress-induced upregulation, the phosphorylation of Bim also regulates apoptosis.23 Only BimEL and BimL are detected in postmitotic CNS neurons.38 A subset of Bim-deficient mice die around embryonic day 9.5, but those surviving show no obvious neuropathological abnormalities.39 Bim deficiency delays but does not prevent the death of cerebellar granule neurons induced by trophic factor withdrawal or low potassium.23

BNips were isolated as proteins interacting with the adenovirus E1B 19 kDa protein, an antiapoptotic homologue to Bcl-2 known to limit host–viral defense.40 Currently, there are four human BNips: BNip1-3 and Nip3-like protein X (Nix). Nix was cloned subsequent to BNip1-3 and shares more than 50% homology with BNip3.40 The binding of BNips with Bcl-2 is BH3 independent40 and both BNip1 and BNip2 rely on their BH3 domain for proapoptotic activity, yet deletion of BH3 from BNip3 does not decrease apoptosis.40 BNip1 and BNip2 are localized to the endoplasmic reticulum and nuclear membrane whereas Nix and BNip3 are localized to the mitochondria and nucleus.40, 41 The C-terminal transmembrane domain of BNip3 is considered integral for its induction of mitochondrial-dependent apoptosis, but deletion of this region induces the non-mitochondrial expression of BNip3 in addition to apoptosis, suggesting that its mitochondrial localization is not critical for the induction of cell death.40 In an unstressed state, BNip3 associates to the mitochondrial surface, but upon cell stress inserts into the outer mitochondrial membrane, which may contribute to the formation of the mitochondrial permeability transition pore and induction of the intrinsic apoptotic pathway.40 BNip2 is developmentally regulated late in the embryonic rat brain, which suggests a possible role in programmed cell death.40 The differential display of genes stimulated in sympathetic neurons from trophic factor deprivation resulted in the isolation of DP5.23 Hrk, the human homologue of DP5, was identified as a Bcl-2-binding protein.23 The expression of DP5 in mouse brain peaks postnatally between P7 and P14 and is lower in adult.42 DP5-deficient mice develop normally and DP5-deficient motorneurons and superior cervical ganglia exhibit enhanced protection from axotomy-induced death and NGF withdrawal, respectively.23 Together with other reports,23 the stress-induced upregulation of DP5 may play an important role in the death of differentiated neurons.

Noxa was identified through cloning of a mouse genomic library and is expressed constitutively, albeit in small amounts in mouse brain.27 Noxa-deficient mice show no nervous system abnormalities during development or as adults.43 Noxa is upregulated in many cell types including neural precursor cells and neurons from stimuli that induce its p53-dependent transcription and resultant apoptosis, whereas Noxa deficiency prevents p53-dependent apoptosis.27, 43, 44, 45, 46 Puma was identified via cloning in simultaneous reports.28, 47 Puma-deficient mice appear normal throughout development and do not have any obvious neuronal abnormalities.23 Similar to Noxa, Puma is upregulated in response to p53-dependent apoptotic stimuli28 and Puma deficiency prevents p53-dependent apoptosis.23 Interestingly, Puma is also upregulated in response to ER stress in the absence of p53,48 suggesting that Puma-induced neuron death can occur through both p53-dependent and -independent signaling pathways thus warranting further investigation of the role of Puma in CNS models of neuronal cell death.

Bcl-2 Family Interacting Proteins: the Plot Thickens Even Further

An emerging group of Bcl-2 family interacting proteins includes Bcl-2-associated athanogenes (BAGs) and Bcl-2 interacting protein (Beclin) that lack known homology with the Bcl-2 family yet interact with them to regulate stimulus-specific cell death.

The BAG protein family consists of BAG1-5, all with at least one copy of the approximately 50 amino acids, evolutionarily conserved C-terminal BAG-domain.49 BAG5 is unique among BAGs in that it is predicted to possess multiple BAG domains.49 BAG1, the first identified member of this family, was isolated through cloning of genes encoding proteins that bound Bcl-2.49 The combined overexpression of BAG1 and Bcl-2 enhanced protection against apoptotic insults compared to overexpression of either alone, suggesting that the interaction between BAG1 and Bcl-2 is an important apoptosis regulator. In addition, the overexpression of Bcl-2 causes a redistribution of BAG1 to a punctuate organellar location, which suggests its mitochondrial localization49 and potential regulation of the intrinsic apoptotic pathway.

In addition to their interactions with Bcl-2, BAG proteins bind with high affinity to the ATPase domain of Hsp70 through interaction of their conserved BAG domains.49 The BAG1–Hsp70 interaction prevents neuron death,50 which may suggest the formation of a multimember complex containing BAG1, Hsp70 and Bcl-2 that has yet to be described. BAG1 also stimulates neuronal differentiation50, 51 independent of its interaction with Hsp70.50 BAG1 is expressed in the developing mouse nervous system and exhibits development-specific intracellular staining patterns such that its localization in neural precursor cells is predominantly nuclear and is primarily cytosolic in differentiated neurons.51 Currently, it is not known whether Bcl-2 plays a role in BAG1-mediated neuronal differentiation.

BAG3, also known as Bis, is similar to BAG1 such that its antiapoptotic activity increases synergistically in the presence of Bcl-2.52 BAG3 is expressed in many neuron populations throughout the adult rat CNS including cerebellar Purkinje cells and in ventral spinal cord motor neurons, and is also found in astrocytes localized to the rostral migratory stream.53 To date, a neuron-specific, antiapoptotic effect of BAG3 has not been documented although BAG3 is selectively upregulated in mouse astrocytes following kainic acid-induced seizures.54

Beclin was isolated upon investigation of adult mouse proteins that interacted with Bcl-2.55 The interaction of Bcl-2 with Beclin is important in protecting against CNS infection of Sindbis virus, since the expression of viral constructs of Beclin that lacked Bcl-2-binding capacity induced significantly greater neuronal apoptosis and resultant animal death.55 Beclin, also known as ATG6 in mammals, forms an integral component of the class III phosphoinositol 3-kinase (PI3-K) complex, which is important in autophagosome formation.56 Beclin-null mice are embryonic lethal, suggesting an important role for Beclin in embryogenesis, but a specific role for Beclin in CNS development is not known at this time.57 The overexpression of Bcl-2 inhibits Beclin-dependent autophagy, and Beclin mutants that do not bind Bcl-2 increase autophagic death. Present reports of Beclin-mediated alterations in the CNS remain limited, although Beclin increases in neurons and astrocytes following mouse head injury.58 Future studies will no doubt focus on the interaction of Beclin with Bcl-2 family members in regulating apoptotic and autophagic death in disease and injury models of the nervous system.

The Bcl-2 Family and Neurodegeneration

While it is clear that Bcl-2 family members play a pivotal role in nervous system development, many studies have also shown that neurodegeneration induced by chronic disease or acute traumatic events is associated with alterations in expression of Bcl-2 family members. Many reports describe decreased levels of antiapoptotic Bcl-2 proteins and/or increased levels of proapoptotic proteins in affected areas. Conversely, other reports show increased expression of antiapoptotic genes or proteins, which may reflect a compensatory response of the surviving population of cells. Caution must be used in interpreting results of such studies, since altered expression of Bcl-2 family members does not necessarily imply a causal relationship between the Bcl-2 family and neurodegeneration. Regardless of their etiological role in neurodegenerative disease, therapeutics designed to increase the antiapoptotic balance of Bcl-2 family members remains a popular treatment goal. For instance, several potential neuroprotective agents increase the expression of Bcl-2 via upstream activation of the PI3-K/Akt pathway.4

Alzheimer's disease (AD), the most prevalent dementia-associated disease in humans is characterized by an extensive loss of cortical neurons which has led to speculate a role for apoptosis in this devastating disease. Indeed, AD brain shows altered expression of Bcl-2 family members, but conflicting reports of apoptotic morphology and caspase activation have generated significant controversy on the role of apoptosis in AD pathogenesis.59 The upregulation of pro- (Bak, Bad) and anti- (Bcl-2 and Bcl-XL) apoptotic molecules has been reported in AD brain.4 Bcl-2 and Bcl-XL expression may increase presumably as a chronic, antiapoptotic stress response of compromised neurons. However, in one report the AD-induced increase in Bcl-2 was found to be specific for reactive astrocytes,60 which highlights the need for careful immunohistochemical analysis of Bcl-2 family members in the early stages of AD, to differentiate possible etiologic from compensatory alterations in Bcl-2 family members.

Bcl-2 family members are also implicated in the amyloid beta (Aβ)-induced death of cultured neurons. Treatment with Aβ increases the expression of Bax and DP5 and decreases the expression of Bcl-2 and Bcl-w.61, 62, 63 In addition, the overexpression of Bcl-2, Bcl-XL or Bcl-w and the targeted deletion of Bax prevent Aβ-induced death.63, 64, 65 Future studies are necessary to validate the relative contribution of Bcl-2 family members for Aβ-induced neurotoxicity in vivo.

Amyotrophic lateral sclerosis (ALS) is a fatal disease characterized by the progressive death of upper and lower motor neurons and eventual loss of motor function. It is well accepted that a percentage of dying motor neurons in ALS exhibit apoptotic morphology.66 Postmortem ALS spinal cord indicates a shift towards proapoptotic members of the Bcl-2 family, such as increased expression of Bax66 and DP5/Hrk67 and decreased expression of Bcl-2.66 However, other reports indicate no changes in levels of these Bcl-2 family members.66 Analysis of symptomatic, transgenic mice with mutant superoxide dismutase (SOD1), the most common genetic mutation in familial ALS, also show decreased expression of Bcl-2 and Bcl-XL and increased expression of Bax and Bad.66 Truncated Bid is also evidenced in mutant SOD1 mice,68 which suggests participation of the extrinsic apoptotic pathway in ALS. While apoptosis is suspected to play a role in ALS, more studies are needed to determine its relative contribution to the overall neuropathology.

Huntington's disease (HD) induces the chronic, progressive neurodegeneration of striatal and cortical neurons and is caused by a mutation generating expanded CAG repeats in the Huntington gene. A recent study indicates increased levels of Bax and Bcl-2 in HD brain specific to the caudate nucleus, which may be interpreted as an antiapoptotic response in stressed but viable neurons.69 However, a relative lack of cleaved caspase-3 immunoreactivity in the same samples precludes a net prodeath role for Bcl-2 family members in HD brain.69 Animal models of HD show more convincing evidence of a proapoptotic tone of Bcl-2 family members in affected neuronal populations. Mouse models of HD, including the R6/2 transgenic mouse, show increases in mitochondrial Bim and Bax and a decrease in phosphorylated Bad.70 Levels of Bcl-2 or Bcl-XL are unchanged in these mice, but the overexpression of Bcl-2 does slow disease progression.70 Treatment with 3-nitroprussic acid, a chemical inhibitor of mitochondrial succinate dehydrogenase used to model HD, induces a selective striatal lesion when administered in vivo that increases the ratio of Bax/Bcl-2 and Bax/Bcl-XL,69, 71 thus shifting their balance to a proapoptotic state.

Parkinson's disease (PD) is a chronic motor disease caused by the degeneration of dopamine-containing neurons of the substantia nigra that project to the striatum. In animal models of PD, treatment with 6-hydroxydopamine and 1-methyl, 4-phenyl,1,2,3,6-tetrahydropyridine (MPTP) increases expression of Bax72, 73 and Puma,74 and decreases expression of Bcl-2.72 Furthermore, Bax deficiency decreases MPTP-induced neurodegeneration72 and Puma deficiency prevents 6-OHDA-induced neuron death.74 Bcl-2 and Bcl-XL are reportedly increased in postmortem PD brain samples,75 which may represent an antiapoptotic response of surviving neurons, yet other studies of PD have not detected changes in Bcl-2 family members.75 Thus, involvement of Bcl-2 family members in PD pathogenesis remains speculative.

In contrast to the neuroprotective effects of BAG1, BAG5 enhances the degeneration of dopamine-containing neurons in an in vivo model of PD and inhibits the chaperone activity of Hsp70.76 In addition, BAG5 directly interacts with and inhibits the E3 ubiquitin ligase activity of parkin,76 whose gene is mutated in autosomal recessive forms of PD. Furthermore, attenuation of BAG5 expression decreases the accumulation of parkin within Lewey Body-containing aggregates in an animal model of PD.76 The attenuated function of wild-type parkin in the presence of BAG5 may implicate this protein interaction in sporadic PD, but a role of Bcl-2 family members in this context is unclear at present.

Stroke or cerebral ischemia is difficult to treat in humans due to the narrow time between the ischemic event and irreversible brain injury. In rodent models, postischemic alterations in expression of Bcl-2 family members, including increased expression of Bax and reduced expression of Bcl-2 and Bcl-w are found both in the ischemic core and the penumbra of the infarct.15 Conversely, Bcl-2 and Bcl-w expression have been shown to increase in areas of ischemic insult,11, 15 suggesting a compensatory survival response in sublethally injured neurons. Expression of Bim,77 BNip341 and Puma48 also increase in ischemic brain, suggesting a role for BH3 domain-only proteins in ischemia-induced neuron apoptosis. In addition, the overexpression of BAG1 and Bcl-2, and deficiencies in Bax and Bid have been shown to attenuate ischemia-induced damage,15, 78, 79 which suggests potentially effective treatment strategies that target the Bcl-2 family.

Conclusions

Numerous advances have been made over the last 15 years that have helped clarify the role of the Bcl-2 family in regulating cell fate in the CNS. The simple rheostat model of Bcl-2/Bax regulation of cell death formulated in the early 1990s provided a solid foundation from which a more complete understanding of the complex molecular pathways regulating neuronal death has emerged. We now appreciate that life and death decisions in the nervous system are controlled by a myriad of interacting Bcl-2 and non-Bcl-2 family members. In many cases, the interactions of Bcl-2 family members and resultant antiapoptotic versus proapoptotic tone are regulated in part by their stimulus-specific expression and post-translational modifications (Figure 2). The challenge remains to define the critical proteins and molecular interactions regulating neuronal cell fate and to use this information to further our understanding of nervous system development and to generate novel therapeutics for a variety of human neurodegenerative conditions.

Figure 2
figure 2

Current model shows how alterations in Bcl-2 family members and Bcl-2-binding partners leads to stimulus-specific induction of apoptosis, inhibition of apoptosis or inhibition of autophagy in the CNS. The initial stimulus may arise from a lack of trophic support as occurs in neurodevelopment or aging, a chronic neurodegenerative state such as Alzheimer's, Huntington's and Parkinson's disease and ALS, or acute nerve injury that may occur, for example, from cerebral ischemia. As indicated in the text, the relative contribution of each Bcl-2 family member is cell-type and stimulus-type specific. Stimulus-specific alterations in BH3-domain members of the Bcl-2 family, including cleavage of Bid, phosphorylation of Bim, dephosphorylation of Bad and increased expression of Bad, Bim, BNip3, DP5, Noxa or Puma are proposed to alter the interaction of Bax with Bcl-2 or Bcl-x and induce Bax-dependent apoptosis. This induction of Bax-dependent apoptosis may occur either by (1) the increased association of BH3 domain only proteins with antiapoptotic Bcl-2 family members, thus freeing Bax; (2) the direct interaction of Bax with BH3 domain-only proteins; or (3) a decrease in the baseline interaction of BH3 domain-only members with antiapoptotic Bcl-2 family members. Direct stimulus-specific alterations in antiapoptotic and multidomain proapoptotic Bcl-2 family members have also been reported in several neurodegenerative conditions, and the resultant changes in their levels may shift the balance between a proapoptotic versus antiapoptotic state. The association of Bcl-2 with nonfamily members BAG1/3 and Beclin has also been shown to prevent apoptosis

Full size image

Dedication

Stan Korsmeyer and his laboratory contributed enormously to our current understanding of Bcl-2 family regulation of apoptosis. In addition, Stan had an almost unique ability to attract non-‘apoptologists’ into the field of cell death research and to turn short-term scientific collaborations into life-long friendships. Prior to his illness, Stan had agreed to give the 2004 Pritchett Endowed Lecture in Pathology at the University of Alabama in Birmingham. Subsequently, his cancer was discovered and despite undergoing intensive therapy, Stan insisted upon delivering the lecture entitled ‘Gateway to Apoptosis’ in June 2004. After his brilliant presentation, I expressed to Stan my gratitude and told him that I felt somewhat selfish for taking up his time during this difficult period in his life. In an e-mail that I will always treasure and reflecting Stan's spirit and determination he replied ‘The trip was just what the Dr. ordered.’ The biomedical research community has lost one of its superstars; for those of us lucky enough to know Stan personally, we have lost a true friend.