APAF-1 Gene Structure and Regulation

The APAF1 gene, encoding the APAF-1 protein, spans about 55 kb of genomic region mapping on chromosomal band 12q22, between the polymorphic markers D12S296 and D12S346. Several allelic variants of the APAF1 gene have been described. Some of these alleles (such as the E777K, N782T, C450W, and Q465R variants) have been shown to segregate with major depression (MDD) in families where a significant linkage had been found previously between MDD and markers at 12q22. The intron–exon structure of the APAF1 gene comprises 26 introns and 27 exons. The APAF-1 mRNA is ubiquitously expressed in human adult and fetal tissues and yields a 130 kDa cytoplasmatic protein. The APAF1 gene is one of the transcriptional targets of  p53 in DNA damage-induced apoptosis. A genomic region upstream of the APAF-1 transcription start site (at −604 to −570 relative to the transcriptional start site) contains two consensus palindromic sequences defined as p53-responsive elements. Expression of the gene is promoted also by UVC irradiation which enhances translation of APAF1 by a cap-independent mechanism facilitated by internal ribosome entry (IRES) elements located in the 5′-UTR of the gene.

APAF-1 Protein Structure

The APAF-1 protein belongs to the superfamily of AAA⁺ (AAA⁺) proteins. AAA⁺ proteins form large ring-shaped complexes acting as energy-dependent unfoldases of macromolecules. Most AAA⁺ proteins have a single ATPase domain containing a canonical phosphate-binding (P-loop) domain. In the APAF-1 protein, the AAA⁺ ATPase domain is located between an N-terminal caspase recruitment domain (CARD) and a C-terminal domain containing several WD-40 repeats. The overall structure of the APAF-1 protein is as follows: (i) an N-terminal CED-3-like domain (aminoacids 1–89) named CARD that binds to the CARD domain of procaspase-9; (ii) a CED-4 homologous domain (aminoacids 94–412) containing a P-loop sequence that binds dATP/ATP and a putative Mg²⁺-binding site; and (iii) a C-terminal regulatory domain (aminoacids 412–1,194) containing 12 WD-40 repeats involved in the regulation of APAF-1. APAF-1 has different splice isoforms. These include APAF-1 L, APAF-1XL, APAF-1 M, and APAF-1XS. These alternative APAF-1 forms differ in the number of WD-40 repeats (12 or 13) and/or for the presence of additional sequences inserted between the CARD and the CED-4 homologous domains.

Assembly and Structure of the Apoptosome

The inactive form of APAF-1 is thought to form a compact monomer containing bound dATP. According to current models, when cytochrome-c binds to the C-terminal regulatory region of APAF-1, it promotes dATP hydrolysis (dATP → dADP). Subsequently, nucleotide exchange (dATP for dADP) takes place leading to an “active monomer” conformation that is poised to apoptosome assembly. Analysis of the apoptosome structure first at 27 Å and at 12.8 Å resolution, by electron cryomicroscopy, has confirmed that the heptameric complex is a wheel-like structure with seven spokes radiating from a central hub. The N-terminal CARD domains of seven APAF-1 molecules contribute to build the central hub of the apoptosome (the “CARD ring”). The CARD ring represents the active center of the apoptosome where interaction with procaspase-9 takes place. At this level, procaspase-9 molecules may bind to the apoptosome, by CARD–CARD interactions with the hub domain. Activation of procaspase-9 molecules is then thought to occur by an intermediate step requiring the formation of caspase-9 dimers. The C-terminal regulatory regions of each of the seven APAF-1 molecules in the apoptosome contribute an arm and a Y shape domain ending with two lobes. Each lobe, made of six or seven WD-40 repeats (depending on the APAF-1 isoform) folds as a β propeller and a cytochrome-c molecule binds between the two β propellers.

Positive and Negative Regulation of Apoptosome Function

The apoptosome is subjected to positive and negative regulatory interactions with several molecules. In general, activation or inhibition of APAF-1 function may be achieved by mechanisms that interfere with (i) APAF-1 oligomerization; (ii) APAF-1 interaction with caspase-9 or cytochrome-c; and (iii) caspase-9 activation. Examples of positive regulators are (i) NAC, a CARD-containing protein that associates with APAF-1 and promotes the activation of procaspase-9 by the apoptosome; (ii) Nucling, a protein that binds the apoptosome promoting its translocation to the nucleus and the activation of apoptosome-associated caspase-9; and (iii) PHAPI, a protein that promotes caspase-9 association with the apoptosome. Examples of negative regulators of the apoptosome include (i) IAP proteins, as  XIAP, that can associate with the apoptosome and inhibit caspase-9; (ii) heat shock proteins (HSP) that can bind to cytochrome-c (as Hsp27) thus preventing its association with APAF-1, or that can bind to APAF-1 (as Hsp70 and  Hsp90) and prevent caspase-9 activation; (iii) posttranslational modifications of apoptosome components, as exemplified by phosphorylation of caspase-9 at serine 196 (by AKT), at threonine 125 (by ERK), or at serine 144 (by PKC) that prevent caspase-9 activation or recruitment to the apoptosome; and (iv) a caspase-9 splice variant (named Casp-9γ), retaining only the CARD domain but lacking the catalytic domain, that may compete with functional caspase-9 for binding to the apoptosome.

APAF-1 Expression and Apoptosome Regulation in Cancer

Most of the chemotherapeutic drugs used in the treatment of cancer promote apoptosis by the mitochondrial pathway that leads to cytochrome-c release and apoptosome assembly. As resistance to apoptosis is one of the hallmark of cancer, alterations in APAF-1 expression and apoptosome function have been shown to be common in both solid tumors and hematological malignancies. Loss of expression of APAF1 gene was initially described in advanced  melanoma, by a mechanism involving  methylation-induced transcriptional silencing and  allelic imbalance (loss of heterozygosity). APAF1 promoter methylation and allelic imbalance have been described even in tumors other than melanoma. For example, reduced mRNA levels for APAF1 has been shown in primary acute myeloblastic leukemia cells, due to  CpG methylation in a region between +87 and +128 of the APAF1 gene. Similarly, methylation of APAF1 gene has been described in carcinomas of the bladder and in clear cell renal carcinomas. Allelic imbalance for APAF1 gene, associated with reduced APAF1 mRNA levels, has been described in colorectal carcinomas.

Defects of APAF-1 expression may be a marker of neoplastic transformation and/or tumor  progression. In human melanoma, expression of APAF-1 protein is lower in neoplastic cells than in melanocytes and decreases with increasing thickness of the primary tumor as well as in the progression from primary lesion to metastatic disease. In  nonsmall cell lung cancer patients, the subcellular localization of APAF-1 has been shown to represent a significant prognostic factor. In fact, nuclear localization of APAF-1 was associated with 5-year survival rates of 89% compared to 54% in patients with cytoplasmic localization of APAF-1 in the tumor cells. This suggests that nuclear translocation of APAF-1 may be associated with an apoptosis-prone phenotype of the neoplastic cells.

Inactivation of the APAF1 gene, first shown in human melanoma, provides evidence for a mechanism that may prevent the execution of the apoptotic program in neoplastic cells following cytotoxic stress. The initial evidence indicated that reduced/absent APAF-1 protein expression was associated with chemoresistance of melanoma cells to DNA damaging drugs that mediate apoptosis by the p53 pathway. Reduction of APAF-1 protein can be achieved in neoplastic cells even by APAF-1 sequestration in discrete subcellular domains, not only by reduced protein expression as in melanoma cells. In Burkitt lymphoma cells, APAF-1 has been shown to be associated with discrete domains of the plasma membrane, instead of being free in the cytosol. Such APAF-1 sequestration prevents apoptosome formation in the presence of cytochrome-c and is associated with resistance to etoposide in Burkitt lymphoma.

Altered regulation of apoptosome assembly and function is another antiapoptotic strategy activated in neoplastic cells. For example, the constitutively active tyrosine kinase  BCR–ABL of chronic myelogenous leukemia has been shown to inhibit interaction of caspase-9 with APAF-1. Reduced caspase-9 binding to the apoptosome, not explained by reduced levels of caspase-9 or APAF-1, has been proposed as a chemoresistance mechanism even in  ovarian cancer. In nonsmall cell lung cancer, a defect in apoptosome function has been linked to overexpression of the inhibitor of apoptosis XIAP that binds to the processed form of caspase-9, thus suppressing activation of downstream effector caspases.

Apoptosome-Dependent and -Independent Pathways of Apoptosis in Normal and Neoplastic Cells

The role of the APAF-1 pathway in apoptosis depends on the cell-context and on the specificity of the proapoptotic signal. In some experimental models, and in some human tumors, APAF-1 expression and function has been shown to be required for apoptosis in response to different proapoptotic drugs. For example, in mouse embryonic fibroblasts from mice lacking APAF-1 (APAF-1^−/− mice), susceptibility to apoptosis induced by the  proteasome inhibitor bortezomib is inhibited. In human leukemic cells, apoptosis promoted by etoposide requires caspase-10 activation, but small interfering RNA-mediated downregulation of APAF-1 prevents etoposide-mediated caspase-10 activation and inhibits apoptosis.

On the other hand, thymocytes from APAF1^−/− mice have normal susceptibility to Fas-mediated cell death, indicating that APAF-1 is dispensable for the execution of apoptosis by the extrinsic pathway (extrinsic pathway of apoptosis) (i.e.,  death receptor-induced apoptosis). In addition, thymocytes from mice expressing a mutant cytochrome-c unable to bind APAF-1 have been shown to be susceptible to apoptosis regulated by the intrinsic pathway (and induced by stimuli as etoposide, γ and UV irradiation). In these cells caspase-9 and caspase-3 could be activated after γ-irradiation, in spite of the absence of APAF-1 oligomerization, indicating the existence of apoptosome-independent, caspase activation pathways in response to cytotoxic stress.

Apoptosome-independent pathways of cell death in response to chemotherapeutic drugs exist in neoplastic cells. These pathways may promote apoptosis even when APAF-1 is not expressed, although APAF-1 expression can amplify the cellular response to some drugs. In melanoma cells, APAF-1 expression has been shown to be dispensable for caspase-9 activation and apoptosis promoted by drugs as  cisplatin, camptothecin, betulinic acid, and etoposide. In agreement with these results, analysis of APAF-1 expression in a panel of 60 cell lines used for drug screening, and including the most frequent solid tumors and leukemias, has not provided evidence for APAF-1 as a major determinant of drug sensitivity.