Apoptosis-inducing factor (AIF), which is confined to mitochondria of normal healthy cells, was initially described by Kroemer and coworkers (Susin et al. 1999) as the first caspase-independent death effector. Under conditions of cell death induction, AIF is released from mitochondria and translocates to the nucleus where it contributes to chromatin condensation and DNA fragmentation, two features that are classically associated with apoptosis (Hangen et al. 2010a). Since its initial discovery, the structure of the AIF protein has been resolved and the AIF gene has been subjected to genetic manipulations in mice, flies, nematodes, and yeast, revealing the phylogenetically conserved contribution of AIF to cell death, as well as its crucial role in physiological cell survival, proliferation, and differentiation (Hangen et al. 2010a). The recent discovery of disease-aggregating mutations of AIF reflects the relevance of its mitochondrial activity to human health (Sevrioukova 2016).
AIF Protein Synthesis and Regulation
The primary AIF transcript is subjected to tissue-specific alternative splicing (Hangen et al. 2010a) (Fig. 2). The alternative usage of exon 2 (2a or 2b) allows for the production of two splice variants (AIF1 and AIF2). Exon 2a is included in the most abundant AIF isoform (AIF1). The two isoforms exhibit a limited difference, which is confined to the C-terminal part of their MLS (Hangen et al. 2010b). While AIF1 is expressed in almost all tissues, AIF2 mRNA expression is restricted to brain and retina. The expression of AIF2 is modulated during brain development and in vitro-induced neuronal differentiation (Hangen et al. 2010b). Several additional AIF isoforms have been described (Hangen et al. 2010a; Sevrioukova 2011). AIFsh is a short variant that is produced from an alternative transcript whose transcriptional start site is located within intron 9 of AIFM1. This variant lacks the N-terminal MLS and the enzymatic domain but retains the C-terminal domain (which harbors the proapoptotic segment). Transfection-enforced expression of AIFsh results in its accumulation in the nucleus and triggers apoptosis (Hangen et al. 2010a). Another short form of AIF (AIFsh2) results from the alternative usage of exon 9b, which contains a stop codon. This isoform maintains the conserved mitochondrial localization and redox function but lacks the C-terminal proapoptotic domain (Hangen et al. 2010a). A third short form of AIF (AIFsh3) lacks the mitochondrial localization signal but otherwise resembles AIFsh2 (Hangen et al. 2010a). Quantitative profiling of mRNA expression and proteomics will be required to assess the precise distribution of each isoform in various tissues.
AIF’s Involvement in Cell Death
Upon mitochondrial outer membrane permeabilization (MOMP) – a feature of most, if not all, apoptotic pathways – AIF is released from mitochondria and translocates to the nucleus, where it mediates chromatin condensation and DNA degradation (Hangen et al. 2010a). The mitochondrio-nuclear translocation of AIF has been observed during developmental cell death, in cells that die in response to genotoxic agents, in the context of exitotoxicity induced by glutamate or other NMDA receptor agonists, in hypoxia–ischemia followed by reperfusion, in neurodegeneration, and in pathogen exposure. As a result, it has been hypothesized that AIF’s lethal activity could regulate a wide range of cell death paradigms (Hangen et al. 2010a). Nonetheless, the genetic deletion or downregulation of murine Aifm1 revealed that AIF was not a general death effector and that its contribution to cell death depended on the cell type and/or the apoptotic insult (Hangen et al. 2010a). Today, AIF’s lethal activity is considered to be required for the programmed death of neurons and photoreceptor cells provoked by excitotoxins, hypoxia–ischemia, hypoglycemia, or withdrawal of trophic support, as in the case of retinal detachment (Hangen et al. 2010a). The neuro-specific death activity of AIF was first suspected from in vitro observations, which implicated AIF in poly(ADP-ribose) polymerase 1 (PARP-1)-dependent neuronal death and then confirmed through the characterization of the mutant Harlequin (Hq) mice, which carry a hypomorphic Aifm1 mutation that provokes an 80% reduction in the expression level of AIF compared to wild type animals (Hangen et al. 2010a). For example, in vivo excitotoxic studies using kainic acid-induced seizures revealed that the brains of Hq mice developed less hippocampal damage than wild type animals. In addition, compared to wild type animal, the brain of Hq mouse is more resistant to ischemia-induced damage. Likewise, the prevention of AIF’s nucleo-mitochondrial translocation has a neuroprotective effect (Hangen et al. 2010a).
The lethal activity of AIF is also determined by the nucleic acid binding potential of AIF. Crystal structures of AIF and mutagenesis experiments revealed the existence of positively charged amino acids that are scattered at the surface of the molecule and that are required for the interaction with DNA or RNA and for the induction of nuclear apoptosis by overexpressed AIF. Recombinant AIF provokes DNA condensation through direct, sequence-independent interactions with single or double stranded DNA (Hangen et al. 2010a). The nondegradative ubiquitination of AIF by XIAP has a negative effect on the capacity of AIF to bind DNA and induce chromatin condensation (Fig. 3) (Lewis et al. 2011). It was recently reported that the resistance of ovarian cancer cells to cisplatin could be overcome by the treatment with a natural food phytochemical hirsutenone that inhibits XIAP/AIF interaction (Farrand et al. 2014). HSP70 was also revealed as a negative regulator of AIF apoptotic function (Fig. 3) (Hangen et al. 2010a). Moreover, AIF interacts with several other proteins, in particular cyclophilin A (Artus et al. 2010; Hangen et al. 2010a) and histone H2AX (Artus et al. 2010) that participate to the activation of the lethal action of nuclear AIF (Fig. 3). The RNA binding activity of AIF that was revealed in vitro also requires further investigation (Hangen et al. 2010a).
AIF Involvement in Cell Survival, Proliferation, and Differentiation
AIF is a bifunctional flavoprotein. Indeed, in addition to its lethal function within the nucleus of dying cell, AIF plays a vital role in healthy cells, due to its impact on mitochondrial bioenergetics (Hangen et al. 2010a). Initially, the revelation of a significant homology between the internal, nonapoptotic, segment of AIF and bacterial NADH-oxidases hinted towards the possibility that AIF could fulfill a nonapoptotic enzymatic function in the mitochondrion of healthy cells (Hangen et al. 2010a). Later, the phenotypic characterization of Harlequin (Hq) mice, a model of late-onset neurodegeneration, was instrumental in highlighting the vital nonapoptotic activity of AIF and its impact on cell survival, proliferation, and differentiation (Hangen et al. 2010a). Hq mice, which harbor a retroviral insertion in the first intron of Aifm1 that causes an 80% reduction in AIF expression, were first noted for growth retardation, fur loss, and the development of ataxia and blindness due to an age-related, oxidative stress-associated progressive loss of terminally differentiated cerebellar and retinal neuron (Hangen et al. 2010a). The impact of AIF hypomorphic Hq mutation on cell metabolism and survival seems to be cell-specific as, in addition to the loss of cerebellar, retinal neurons, and skeletal muscle cells (Hangen et al. 2010a; Armand et al. 2011; Sevrioukova 2011), Hq mice exhibit a T-cell lineage (but not B lineage) block that can be overcome either by antioxidant treatment or by the expression of a recombinant AIF protein that carries a wild type NADH-oxidase domain (Banerjee et al. 2012). It was also observed that hypoxia/ischemia condition is particularly harmful to the hearts of Hq mice and that the administration of the synthetic superoxide dismutase and catalase mimetic EUK8 can reduce cardiac oxidative stress and ameliorate the survival of Hq mice subjected to experimental constriction of the aorta (Hangen et al. 2010a).
As AIF is indispensable for cell survival during embryogenesis, all attempts to create AIF null mice by homologous recombination were unsuccessful and consequently only conditional genetic deletion of AIFm1 was an option for the determination of the in vivo role of AIF in cell survival and organogenesis (Hangen et al. 2010a). For instance, the specific deletion of AIF in the prospective midbrain and cerebellum revealed that AIF is necessary for cell-type specific neurogenesis in the developing brain (Ishimura et al. 2008). A major defect in cortical development and reduced neuronal survival was also observed when AIFm1 was specifically lost in the telencephalon (Cheung et al. 2006). The conditional deletion of AIFm1 in muscle and liver had an important impact on whole-body metabolism and revealed that compared to control littermates, muscle- and liver-specific AIF mutant mice were resistant to diet-induced obesity and diabetes (Pospisilik et al. 2007). It was also reported that with aging, mutant mice with muscle-specific loss of AIF develop severe skeletal muscle atrophy and a dilated cardiomyopathy before becoming lethargic around the age of 5 months (Joza et al. 2005). Recently, the tissue-specific knockout of AIF, in T and B cells, confirmed that the metabolism of T (but not B) lymphocytes depends on the pro-survival function of AIF linked to the maintenance of mitochondrial function (Milasta et al. 2016).
In addition to its involvement in mitochondrial import process in the IMS, AIF could harbor additional functional segments that regulate the structure and/or stability of the inner mitochondrial membrane. This is suggested by the fact that the loss of AIF in the telencephalon entails a degenerative phenotype accompanied by fragmentation of the mitochondrial network and aberrant cristae in cortical neurons (Cheung et al. 2006). The potential inner membrane-stabilizing activity of AIF may reside in its transmembrane segment. This possibility is supported by the description of a brain-specific isoform of AIF (AIF2) that is produced through the alternative usage of exon 2 (Fig. 2) and differs from the ubiquitously expressed AIF (AIF1) only within its transmembrane region (Hangen et al. 2010b). Both AIF1 and AIF2 localize to the same mitochondrial subcompartment and are similar in their capacity to regulate the stability of complex I subunits, yet differ in their membrane anchorage capacity and in their effects on mitochondrial morphology (Hangen et al. 2010b).
The Implication of AIF in Disease
Among the four mutations (AIFΔR201, G308E, G338E, G262S), which have been found associated with severe early onset encephalomyopathies, AIFΔR201 and G308E are the most studied at the structure/function level. AIFΔR201, which consists in the deletion of three base pairs coding for arginine residue 201 of the precursor AIF protein, was described by Ghezzi et al. (2010) while exploring the genetic basis for the X-linked mitochondriopathy manifested in two male infant patients born from monozygotic twin sisters and unrelated fathers. The expression of AIFΔR201 negatively affected OXPHOS. The biochemical examination of fibroblasts from both patients revealed a major defect in CIII and CIV that was partially corrected by the overexpression of recombinant wild type AIF or by continuous culture of mutant fibroblasts in the presence of riboflavin, the precursor of FAD (Ghezzi et al. 2010). Moreover, muscle biopsies from both patients revealed a severe loss of mitochondrial DNA that could be responsible for the combined multicomplex (CI, CIII, and CIV) dysfunction (Ghezzi et al. 2010). Molecular modeling of the mutant AIF, as well as in vitro experiments indicated that the pathogenic AIF protein was unstable and exhibited altered folding and redox properties compared to the nonmutated version of AIF (Ghezzi et al. 2010; Sevrioukova 2016). Moreover, it was revealed that the deletion of arginine 201 enhanced the DNA-binding capacity of the mutant AIF and rendered cells more sensitive to death stimuli (Ghezzi et al. 2010).
Linkage analysis followed by exome sequencing led to the discovery of the missense pathogenic mutation of AIF (G308E) that is associated with a severe prenatal encephalomyopathy (Berger et al. 2011). Structural and biochemical analyses of the AIF variant revealed that the G308E substitution did not affect the general conformation of the protein, but rather perturbed its capacity to bind its cofactor NAD(P)H and to catalyze redox reactions (Sevrioukova 2016). The G308E mutation reduces the capacity of AIF to bind to its mitochondrial interactor CHCHD4 (Fig. 4), shedding new light on the molecular basis for the pathogenicity of this particular mutation (Hangen et al. 2015). Reportedly, the AIFΔR201 mutation, which perturbs the correct conformation of the protein, also compromises the interaction with CHCHD4 (Meyer et al. 2015). Other AIF mutations, which are exemplified by the E493V substitution (Rinaldi et al. 2012), are associated with late onset and slowly progressing mitochondriopathies that exhibit relatively mild clinical phenotypes (Sevrioukova 2016). The pathogenic AIF mutation E493V enhances the apoptotic activity of the protein (without affecting its capacity to control the respiratory activity) and does not affect the in vitro interaction of AIF with CHCHD4 (Hangen et al. 2015), which might explain the milder phenotype of the associated disorder (Sevrioukova 2016).
Recent studies have highlighted the impact of the mitochondrial activity of AIF on the growth and aggressiveness of prostate and pancreatic tumors (Lewis et al. 2012; Scott et al. 2016). Future molecular studies are required for a better understanding of the tissue-specificity and/or the particular metabolic demands of cancer cells that benefit from the pro-survival function of AIF.
AIF that was originally recognized as an apoptosis-inducing factor acting in the nucleus of the dying cell is now emerging as an important protein participating to the regulation of bioenergetics in the mitochondrion of healthy cells. Although the molecular bases for the tissue-specificity and signal-dependency of both lethal and vital functions of AIF remain largely elusive, recent years have witnessed a major progress in the understanding of the pro-survival mitochondrial function of the flavoprotein that is crucial for the biogenesis of the respiratory chain complexes. Thus, CHCHD4 has emerged as the missing link between AIF and the biogenesis of respiratory chain complex I subunits. The number and variety of potential mammalian CHCHD4 substrates suggest that in addition to the biogenesis of specific respiratory chain protein subunits, AIF has the potential to control redox regulation, antioxidant response, lipid homeostasis, and mitochondrial ultrastructure and dynamics (Modjtahedi et al. 2016). The creation of mouse models carrying knock-in mutations in each of the functional domains of AIF, including those that are involved in its interaction with CHCHD4, or in human mitochondrial diseases will be instrumental for a better understanding of the vital and lethal roles of AIF.
- Farrand L, Kim JY, Byun S, Im-aram A, Lee J, Suh JY, et al. The diarylheptanoid hirsutenone sensitizes chemoresistant ovarian cancer cells to cisplatin via modulation of apoptosis-inducing factor and X-linked inhibitor of apoptosis. J Biol Chem. 2014;289:1723–31. doi:10.1074/jbc.M113.513879.CrossRefPubMedGoogle Scholar
- Joza N, Oudit GY, Brown D, Benit P, Kassiri Z, Vahsen N, et al. Muscle-specific loss of apoptosis-inducing factor leads to mitochondrial dysfunction, skeletal muscle atrophy, and dilated cardiomyopathy. Mol Cell Biol. 2005;25:10261–72. doi:10.1128/MCB.25.23.10261-10272.2005.PubMedCentralCrossRefPubMedGoogle Scholar
- Lewis EM, Wilkinson AS, Davis NY, Horita DA, Wilkinson JC. Nondegradative ubiquitination of apoptosis inducing factor (AIF) by X-linked inhibitor of apoptosis at a residue critical for AIF-mediated chromatin degradation. Biochemistry. 2011;50:11084–96. doi:10.1021/bi201483g.PubMedCentralCrossRefPubMedGoogle Scholar
- Lewis EM, Wilkinson AS, Jackson JS, Mehra R, Varambally S, Chinnaiyan AM, et al. The enzymatic activity of apoptosis-inducing factor supports energy metabolism benefiting the growth and invasiveness of advanced prostate cancer cells. J Biol Chem. 2012;287:43862–75. doi:10.1074/jbc.M112.407650.PubMedCentralCrossRefPubMedGoogle Scholar