Article

• The EMBO Journal (2005) 24, 2815 - 2826
• doi:10.1038/sj.emboj.7600746

Published online: 7 July 2005

• Subject Categories:

  • Differentiation and Death | Molecular Biology of Disease

AIF suppresses chemical stress-induced apoptosis and maintains the transformed state of tumor cells

Alexander Urbano¹, Umayal Lakshmanan¹, Poh Heok Choo¹, Jair Chau Kwan¹, Poh Yong Ng¹, Ke Guo², Saravanakumar Dhakshinamoorthy¹ and Alan Porter¹

1. Cell Death and Human Disease Group, Institute of Molecular and Cell Biology, Proteos, Singapore, Republic of Singapore
2. Histology Unit, Institute of Molecular and Cell Biology, Proteos, Singapore, Republic of Singapore

Correspondence to:

Alan Porter, Cell Death and Human Disease Group, Institute of Molecular and Cell Biology, 61 Biopolis Drive, Proteos, Singapore 138673, Republic of Singapore. Tel.: +65 6586 9675; Fax: +65 6779 1117; E-mail: mcbagp@imcb.a-star.edu.sg

Received 7 February 2005; Accepted 16 June 2005

• Keywords:

  • AIF,
  • apoptosis,
  • colon cancer,
  • NADH oxidase,
  • tumorigenesis

Introduction

A single X-linked mammalian gene aif encodes a 57-kDa mitochondrial flavoenzyme, apoptosis-inducing factor, homologues of which are present in human, mouse, Caenorhabditis elegans and probably yeast (Susin et al, 1999; Joza et al, 2001; Lipton and Bossy-Wetzel, 2002; Wissing et al, 2004) apoptosis-inducing factor (AIF) is released from mitochondria to the cytoplasm of many mammalian cells induced to undergo apoptosis, where it combines with cyclophilin A to form an active DNase, translocates to the nucleus and contributes to nuclear DNA fragmentation and chromatinolysis (Loeffler et al, 2001; Cande et al, 2004a). This death pathway is conserved in C. elegans, where the AIF homologue WAH-1 functionally associates with the mitochondrial DNA endonuclease CPS-6/Endo G. Inhibition of WAH-1 delayed the normal progression of apoptosis during development (Wang et al, 2002). Knockout of the aif gene in mouse embryonic stem (ES) cells is lethal, being associated with a defect in sculpting of the early embryo and reduced susceptibility of the cells to serum withdrawal (Joza et al, 2001). Thus, AIF/WAH-1 appears to play crucial roles in programmed cell death during early animal development.

The crystal structures of mouse and human AIF revealed two important regions: a domain required for DNA fragmentation (a putative DNA-binding site) and a domain with homology to a bacterial NADH-dependent ferredoxin reductase that is structurally similar to the eukaryotic glutathione reductase family of enzymes (Mate et al, 2002; Ye et al, 2002). In functional in vitro assays, AIF acts as an NADH oxidase, accepting electrons from NADH, and transferring them to molecular oxygen to form the superoxide (O₂^-) free radical, which subsequently undergoes dismutation to H₂O₂ (Miramar et al, 2001). The NADH oxidase and DNA fragmenting activities of AIF are completely independent, suggesting that the NADH oxidase activity of AIF is not required for apoptosis (Ye et al, 2002).

The first evidence that AIF has a protective function in some cell types came from studies on the Harlequin mutant mouse, which has an 80% reduction in AIF, but appears normal except for premature neurodegeneration and increased peroxide sensitivity in specific subsets of neurons in adult animals (Klein et al, 2002). The precise mechanism of action of AIF in free radical scavenging and counteracting oxidative stress in this mouse model remains to be established. More recently, it has been shown that a redox-active domain of AIF and reduced glutathione are required for the inhibition of cytoplasmic stress granule formation under conditions of chemical stress, suggesting that AIF is involved in an adaptive response (Cande et al, 2004b). Interestingly, AIF deficiency compromises oxidative phosphorylation (OXPHOS) by inhibiting respiratory chain complex I in vitro and in vivo, revealing a 'life' function for AIF (Vahsen et al, 2004). However, the mechanistic connections between the redox activity of AIF, OXPHOS and cell survival remain unclear.

Here, we investigate the role of the NADH oxidase activity of AIF by knocking out the gene in different human colon cancer cell lines, and analyzing the phenotype of AIF knockout cells re-expressing either AIF or mutant AIF deficient in NADH oxidase activity.

Generation of AIF-deficient carcinoma cell lines

We generated viable AIF-null human colon carcinoma cell lines by modifying the single X-linked locus of the human aif gene using bipartite promoter-less targeting vectors (Jallepalli et al, 2001) (Figure 1A). The human colon cancer cell lines HCT116 and DLD-1 are near diploid with a single X chromosome and have been used extensively for gene targeting (Sedivy et al, 1999). aif^-/y knockout clones were identified by PCR analysis of genomic DNA using two unique sets of primers. One primer set detected only the correct integration of the targeting vector reporter gene (hygromycin, hyg) together with the deletion of the first exon, while the other set detected only the original configuration of the human aif gene (Figure 1B and C). Lack of aif expression in two independent clones derived from both cell lines was further confirmed by RT–PCR of the exon 1–exon 2 boundaries of aif mRNA (Figure 1D). Western blot analysis demonstrated AIF protein in the isogenic control cell lines and the complete absence of AIF in all knockout cell lines (Figure 1E).

Figure 1.

Generation of AIF knockout cell lines by homologous recombination. (A) Schematic representation of bipartite promoter-less targeting vector. Hyg, hygromycin; p-A, poly(A). (B) Two different primer pairs for genomic DNA PCR assay for correct gene targeting. (C) Genomic DNA from HCT116 and DLD-1 cells was analyzed by PCR using the primers described in panel B. As a control, the wild-type locus of AIF was amplified together with the diagnostic PCR for integration of the promoter-less Hyg-pA cassette. hvt and dvt refer to vector controls for HCT116 and DLD-1 cells, respectively. hKO and dKO refer to HCT116 and DLD-1 knockout cells, respectively. (D) aif mRNA-specific RT–PCR analysis of exon 1–exon 2 boundary confirming lack of aif expression in HCT116 and DLD-1 clones. (E) Western blot analysis with human AIF antibodies to confirm the absence of AIF in knockout cells. Hsp60, a mitochondrion-specific protein, was used as a loading control.

View full figure (107 KB)

To extend our studies to other tumor types, we generated stable AIF knockdown cell lines via siRNA in colon cancer cell line SW480, breast cancer line MCF-7 and lung cancer cell line A549. Two independent siRNA knockdown clones for each cell line showed 50–80% reduction in AIF protein (Supplementary Figure S1A).

Decreased superoxide (O₂^-) and reactive oxygen species in AIF-deficient tumor cell lines

AIF has been reported to exhibit in vitro NADH oxidase activity that generates O₂^- and subsequently H₂O₂ by dismutation of O₂^- (Miramar et al, 2001), and therefore the lack of AIF might be expected to lower reactive oxygen species (ROS) levels. On the other hand, AIF has been reported to behave as an antioxidant protein that suppresses oxidative stress in neurons (Klein et al, 2002). AIF is required for the proper assembly or function of mitochondrial complex I (Vahsen et al, 2004), and its loss might affect ROS levels as a result of altered leakage of O₂^- from a defective complex I. To explore these possibilities in carcinoma cells, we first assessed intracellular O₂^- rigorously by employing three different methods to overcome limitations associated with measuring the short-lived O₂^- radical with one method only. Using assays for aconitase (which is inhibited by O₂^-), independent clones of both AIF knockout cell lines exhibited higher aconitase activity than the parental and vector control cell lines, reflecting 20–40% less intracellular O₂^- (Figure 2A). Similar results were obtained when O₂^- production was assessed with lucigenin (Figure 2B) and NBT reduction (data not shown). Similar decreases in intracellular O₂^- were also observed in the three different AIF knockdown cell types (Supplementary Figure S1B). Consistent with the O₂^- data, we demonstrated corresponding decreases in ROS in both AIF knockout colon cancer cell lines (Figure 2C) and in AIF knockdown colon and breast tumor cell lines (Supplementary Figure S1C). Thus in human carcinoma cells, AIF contributes significantly to intracellular O₂^-, either directly through its NADH oxidase activity or indirectly through leakage of O₂^- from AIF-mediated mitochondrial complex I function (or both). Either way, the generated O₂^- would undergo dismutation to H₂O₂ in most cell lines, accounting for the elevated ROS.

Figure 2.

Reduction in O₂^- and ROS in AIF knockout cell lines. Relative O₂^- levels were measured using aconitase (A) and lucigenin (B) assays in AIF knockout (hKO, dKO) clones derived from HCT116 and DLD-1 cell lines, and are meanss.d. of three independent experiments (note that increased aconitase activity but decreased lucigenin activity both reflect reduction in O₂^-). (C) Flow cytometry analysis of relative intracellular ROS levels detected by measuring oxidized dichlorodihydrofluorescein diacetate (H₂DCFH-DA) in two independent knockout clones compared to the wild-type HCT116 and DLD-1 cells. Values represent % of AIF knockout cells having decreased ROS compared to wild-type cells. Graphs are representative of three independent measurements.

View full figure (125 KB)

AIF-deficient cell lines show increased sensitivity to DNA-damaging agents and oxidative stress

Surprisingly, all AIF knockout clones derived from HCT116 and DLD-1 showed a three- to five-fold increased sensitivity to the DNA-damaging agents campthothecin and etoposide, as well as the oxidative stress agents H₂O₂ and t-butylhydroperoxide, as compared to the AIF-expressing cell lines (Figure 3A and B). Likewise, aif siRNA knockdown colon, lung and breast carcinoma cell lines were markedly more sensitive to etoposide and t-butylhydroperoxide than the parental cell lines (Supplementary Figure S2A and B). Evaluation of apoptosis using caspase-3 assays and annexin-V or morphological staining using Sytox-Hoechst dyes confirmed the greatly increased sensitization of both AIF knockout colon carcinoma cell lines to drug and oxidative stress-induced apoptosis (Figure 4A–C). In contrast, the knockout cells showed no alterations in cell death sensitivity with TRAIL, anti-Fas or serum starvation (Figure 3A and B), or with staurosporine or actinomycin D (data not shown).

Figure 3.

Increased sensitivity of AIF knockout cell lines to cancer drugs and oxidative stress. (A) Cytotoxicity assays for HCT116 wild-type and vector control (hvt) cells, and counterpart AIF knockout cell lines (hKO1, hKO2) after exposure to 60 M campthothecin, 60 M etoposide, 500 M hydrogen peroxide, 25 M t-butylhydroperoxide, 25 ng/ml TRAIL, 0.5 g/ml anti-FAS or 4 days serum starvation. (B) Cytotoxicity assays for DLD-1 wild-type and vector control (dvt) cells, and counterpart AIF knockout clones (dKO1, dKO2) after exposure to 5 M campthothecin, 2.5 M etoposide, 250 M hydrogen peroxide, 12.5 M t-butylhydroperoxide, 25 ng/ml TRAIL, 0.5 g/ml anti-FAS or 4 days serum starvation. (C) Cytotoxicity assays for HCT116 cells pre-exposed to EUK-8 (20 M) and NAC (5 mM) overnight prior to treatment with 60 M etoposide or 25 M t-butylhydroperoxide. In panels A–C, surviving cells were detected by incubation with MTS/PES (Promega Cell Titer Proliferation Assay) and expressed as a % of untreated cells and are meanss.d. of three independent experiments.

View full figure (151 KB)

Figure 4.

Increased sensitivity of aif^-/y HCT116 cells to apoptosis. (A) Flow cytometry analysis of HCT116 wild-type cells, vector control (hvt) cells and two AIF knockout clones stained with Annexin-V (PE) after 16 h exposure to 100 M t-butylhydroperoxide or 50 M etoposide. (B) Caspase-3 activity in HCT116 and two counterpart AIF knockout clones after 16 h exposure to 100 M t-butylhydroperoxide or 50 M etoposide. (C) Fluorescence microscopy of HCT116 cells and two counterpart AIF knockout clones stained with Sytox/Hoechst dye after exposure to 50 M etoposide for 48 h.

View full figure (183 KB)

To further investigate the link between decreased O₂^-/ROS and increased sensitivity to DNA-damaging agents and oxidative stress, we exposed parental HCT116 cells to the general antioxidant N-acetylcysteine (NAC) and the MnSOD mimetic (O₂^- scavenger) EUK-8, which decreased intracellular ROS by 23% (5%) and O₂^- by 30% (6%), respectively. This is comparable to the reduction in O₂^-/ROS observed in AIF-deficient cells (Figure 2A–C). Exposure of HCT116 cells pretreated with NAC or EUK-8 to etoposide or t-butylhydroperoxide resulted in increases in sensitivity in the 1.5- to 3-fold range (Figure 3C), which reiterates the effects of curbing AIF expression, and establishes a closer link between a pro-oxidant state and resistance to apoptosis in these tumor cells.

Dramatic loss of tumorigenicity of AIF-deficient carcinoma cell lines

To assess tumorigenicity, we grew the AIF knockout cells in anchorage-independent conditions using the standard soft agar assay. Remarkably, the four independent AIF knockout clones derived from HCT116 and DLD-1 cells showed a substantial 5- to 20-fold decrease in colony growth in soft agar compared to the AIF-expressing cells (Figure 5A and B). Comparable large decreases in soft agar colony formation were observed with AIF siRNA knockdown clones from the colon, lung and breast cancer cell lines (Supplementary Figure S2C). However, the antioxidant NAC and the O₂^- scavenger EUK-8 both failed to curb growth of HCT116 or DLD-1 cells in soft agar at concentrations that cause a 30% or greater reduction in ROS (data not shown). Subsequently, we injected the HCT116 and DLD-1 wild-type and AIF knockout cells subcutaneously into nude mice to determine the in vivo tumorigenicity of the AIF knockout cell lines. The two independent clones from both knockout cell lines exhibited dramatic reduction or absence of tumor growth in vivo compared to the wild-type and vector control cell lines, even when the AIF-expressing and knockout cells were injected on opposite flanks of the same animal (Figure 5C). Measurements of tumor weight and volume showed an overall 6- to 10-fold decrease in tumor sizes for the AIF knockout cells (Figure 5D). These data demonstrate a dramatic loss of tumorigenicity of AIF-deficient carcinoma cells that does not appear to depend on an overall reduction in cellular ROS.

Figure 5.

Dramatic loss of in vitro anchorage-independent growth and in vivo tumorigenicity of aif^-/y colon cancer cells. (A, B) Representative soft agar assays in duplicate (14 days incubation) for HCT116 cells (A, left panels) and DLD-1 cells (B, left panels) and their respective counterpart AIF knockout (KO) cells. Histograms (A and B, right panels) show numbers of colonies formed and are meanss.d. of three independent experiments. (C) HCT116 and DLD-1 cells and their counterpart AIF knockout (KO) cells were injected as xenografts in nude mice, which were killed after 3 weeks (HCT116) and 5 weeks (DLD-1). (D) Histograms show data points representing average tumor weights (left panel) and volumes (right panel) from 15 mice (except dvt=10 mice)s.d.

View full figure (272 KB)

Deficiency in mitochondrial complex I in AIF knockout cells

As previously reported for ES and HeLa cells (Vahsen et al, 2004), the absence of AIF in HCT116 and DLD-1 cells resulted in partial reduction (HCT116) or loss (DLD-1) of several complex I proteins (Figure 6A). In agreement and as measured by three independent assays, mitochondrial complex I activity (an indirect indicator of AIF function) was 40–60% lower in AIF knockout HCT116 and DLD-1 cells (Figure 6B and Supplementary Figure S3). In contrast, citrate synthase activity, which assesses mitochondrial redox status independently of OXPHOS, was unchanged in independent clones representing both AIF knockout cell lines (Figure 6C). These results indicate serious defects in complex I in AIF-null colon cancer cell lines.

Figure 6.

Mitochondrial complex I is compromised in HCT116 and DLD-1 knockout (KO) cells. (A) Representative protein analysis showing decrease in yields of three mitochondrial complex I subunits, which were performed using 5 g of mitochondrial proteins. The loading control is mitochondrial Hsp60. (B) Decrease in NADH-coenzyme Q₁ oxidoreductase activity in AIF knockout cell (KO) cells. Graph shows meanss.d. of three independent experiments. (C) Citrate synthase activity in isolated mitochondria from AIF KO cells represented as meanss.d. of three independent experiments.

View full figure (116 KB)

NADH oxidase activity of AIF is required to restore tumorigenicity and complex I activity in AIF knockout cells

Exhaustive attempts to generate stable knockout cell lines re-expressing AIF were unsuccessful. We attributed this to the pro-apoptotic property of AIF, since we found that overexpressed AIF was not localized only to mitochondria, but distributed overwhelmingly in the cytoplasm, where it can readily translocate to the nucleus and initiate DNA degradation (data not shown). A similar phenomenon was reported in HeLa cells expressing exogenous AIF in which mitochondrial AIF leaked spontaneously into the cytoplasm on long-term culture (Loeffler et al, 2001). To circumvent the apoptotic function of AIF and test the requirement for its NADH oxidase activity in tumorigenicity, we stably expressed two different classes of mutant AIF-GFP molecules in the AIF knockout cell line hKO1 derived from HCT116. One class had mutations in one or other of the (DNA-binding) domains known to be required for AIF-mediated apoptosis (mDNA1 and mDNA2; Figure 7A). The other class denoted 'double mutants' had the DNA-binding mutations as well as mutations in two amino acids implicated in NADH binding (mDNA1 plus NADH mut; Figure 7A), which compromised the NADH oxidase function of AIF (Supplementary Figure S4) (Mate et al, 2002).

Figure 7.

NADH oxidase activity of AIF mutants expressed in aif^-/y HCT116 (hKO1) cells is required for restoration of anchorage-independent growth and mitochondrial complex I activity. (A) Locations of AIF DNA-binding and NADH oxidase mutations. Non-apoptotic DNA-binding mutants mDNA1 and mDNA2 and NADH oxidase-deficient mutant (NADH mut) all have two amino-acid substitutions (arrowed). (B) Western blot analysis of mitochondrial AIF-GFP expression levels in aif^-/y HCT116 (hKO1) cells detected using anti-GFP antibody. DNA muts are three independent clones containing AIF mDNA1 and mDNA2 mutations, NADH muts are three independent clones transfected with aif harboring NADH-binding and mDNA1 DNA-binding mutations. hKO1-GFP represents aif^-/y HCT116 (hKO1) cells transfected with EGFP vector. (C) Colony formation in soft agar assays for hKO1 cells transfected with AIF-GFP mutants as above. Histogram shows numbers of colonies formed and are meanss.d. of three independent experiments. (D) Increased expression of mitochondrial complex I proteins in hKO1 cells transfected with DNA-binding mutants but not NADH-binding mutants. (E) NADH-coenzyme Q₁ activity in parental HCT116 cells and in hKO1 cells transfected with AIF-GFP mutants or GFP alone. Histogram shows enzyme activities normalized against amounts of total proteins and are meanss.d. of three independent experiments.

View full figure (115 KB)

Independently isolated clones of all these various mutants gave similar levels of mitochondrial AIF-GFP expression in aif^-/y HCT116 cells (Figure 7B). The nonapoptotic DNA-binding mutants ('DNA muts') were able to restore anchorage-independent growth in soft agar comparable to parental HCT116 cells, which is in sharp contrast to the GFP control or the double mutants ('NADH muts') harboring amino-acid substitutions in the NADH-binding domain as well as in the DNA-binding domain (Figure 7C). Mitochondrial complex I proteins as well as complex I activity (measured by two methods) were re-established in aif^-/y HCT116 cells stably transfected with the DNA-binding mutants, but not with the double mutants deficient in NADH binding or with GFP alone (Figure 7D and E, and Supplementary Figure S5A). Similar to the independent clones, recovery of colony formation in soft agar was observed with pools of aif^-/y HCT116 cells stably re-expressing DNA-binding mutants, but not the double mutants defective in NADH binding (Supplementary Figure S5B and C). These data altogether show that NADH oxidase activity of AIF is necessary for its tumorigenic properties and important for mitochondrial complex I activity in HCT116 cells.

NADH oxidase activity of AIF is required for foci formation in NIH3T3 cells

To further investigate whether AIF has transforming properties, we overexpressed the various AIF-GFP mutants in the premalignant mouse cell line NIH3T3, and selected stable cell lines that synthesized similar levels of AIF-GFP protein (Figure 8A). NIH3T3 cells transfected with the non-apoptotic DNA-binding mutants mDNA1 and mDNA2 formed obvious foci, which were 5- to 10-fold higher in number compared with cells transfected with either GFP alone or the double mutants with compromised NADH oxidase activity (Figure 8B and C). NIH3T3 cells transfected with parental AIF coupled to GFP did not form significantly more foci than GFP-expressing cells with background levels of foci (Figure 8C). Fluorescence microscopy demonstrated 100% colocalization of AIF-GFP mutants (but not GFP alone) with the foci (Figure 8D). These experiments (and those involving reintroduction of aif into AIF knockout HCT116 cells; Figure 7) suggest that AIF has latent transforming properties revealed either by artificial mutagenesis or inherent suppression of its pro-apoptotic function in tumor cells. These latent transforming properties are a reflection of the involvement of AIF in the maintenance of tumorigenicity of established cancer cells. Moreover, the NADH oxidase activity of AIF is required for foci formation in NIH3T3 cells.

Figure 8.

Stable transformation of NIH3T3 cells by AIF-GFP DNA-binding mutants with NADH oxidase activity. (A) Expression of AIF-GFP mutants stably transformed in NIH3T3 cells detected by Western blot using anti-GFP antibody. mDNA1 and mDNA2, non-apoptotic mutants; mDNA1N and mDNA2N, nonapoptotic plus NADH oxidase mutants. (B) Representative plates showing foci formation in NIH3T3 cells stably transfected with AIF-GFP mutants. (C) Histogram shows numbers of foci formed in panel B. Values are meanss.d. of three independent experiments. (D) Phase-contrast and fluorescence microscopy of GFP and AIF-GFP mutants showing colocalization of AIF-GFP expression and foci only in AIF-GFP-transfected NIH3T3 cells.

View full figure (289 KB)

AIF contributes to pro-oxidant state of carcinoma cells, and NADH oxidase activity of AIF is important for mitochondrial complex I activity

Dying neurons of the AIF-deficient Harlequin mutant mouse exhibited increased oxidative stress, and it was proposed AIF acts as a free radical scavenger (Klein et al, 2002). No differences in oxidative damage or ROS levels were found in wild-type and aif^-/y mouse ES cells, and it was concluded that AIF does not behave as an antioxidant protein in these cells (Vahsen et al, 2004). In contrast, we showed that knockout or knockdown of the aif gene in tumor cell lines of various tissue origins generally resulted in a marked reduction in both O₂^- and ROS levels. AIF contributes to the activity of mitochondrial complex I and consequently to energy production via OXPHOS (Vahsen et al, 2004). In agreement, we found that complex I activity was much reduced in AIF knockout cells, but was restored with stable transfection of aif cDNA encoding mutant AIF lacking pro-apoptotic function. Complex I activity was not, however, restored with NADH-binding mutants of AIF, indicating that the NADH oxidase activity of AIF is important for complex I function. Thus, AIF normally contributes a significant amount of O₂^- in various carcinoma cell types, which could be derived directly from the NADH oxidase activity of AIF (Miramar et al, 2001) or indirectly as a result of AIF-dependent complex I activity causing leakage of O₂^- from the respiratory chain (or both) (Vahsen et al, 2004).

AIF suppresses apoptosis of carcinoma cell lines—role of a pro-oxidant state and complex I

We showed that AIF-deficient cell lines (representing colon, breast and lung tumors) were all sensitized to apoptosis induced by peroxides and some cancer drugs. We found no evidence for an essential cell death promoting function for AIF in carcinoma cells; on the contrary, AIF appears to suppress cell death in the carcinoma cell types we tested. Although much evidence exists for an evolutionarily conserved pro-cell death function for AIF in development (Joza et al, 2001; Wang et al, 2002), and for a proapoptotic signaling role for AIF in nuclear chromatinolysis (Cande et al, 2004a), it is likely that in some scenarios, the survival function of AIF is dominant over its death-inducing activity. Consistent with these conclusions and our results, only certain neurons in the AIF-deficient Harlequin mouse were more prone to H₂O₂-induced cell death (Klein et al, 2002). Thus, the precise involvement of AIF in regulating the life–death balance may be rather specific for the cell type and death inducer.

How might a reduction in intracellular ROS (especially O₂^-) be reconciled with the increased susceptibility of AIF-deficient tumor cells to stress-induced apoptosis that we demonstrated? Physiological levels of some ROS, particularly O₂^- and H₂O₂, are variously required for cell adhesion, immune functions and certain growth factor-dependent survival pathways (Droge, 2002; Reth, 2002; Finkel, 2003). Many tumor cells exist in a pro-oxidant state that is not necessarily detrimental and does not equate with lethal oxidative stress (Cerutti, 1985; Szatrowski and Nathan, 1991; Oberley, 2002; Go et al, 2004). Thus, AIF-negative transformed cells may be more vulnerable to stress if, for example, lower O₂^- (and consequently lower H₂O₂) lessens survival signaling mediated by Ras, NF-KB or AP-1 (Irani et al, 1997; Gupta et al, 1999; Finkel, 2003). It is known that modest reductions in ROS (O₂^- in particular) can promote apoptosis of diverse tumor cell types (Clement and Stamenkovic, 1996; Lin et al, 1999; Pervaiz et al, 1999; Vaquero et al, 2004). NAD(P)H oxidases are usually the source of protective ROS in these reports, which is in line with our findings (Pervaiz et al, 2001; Clement et al, 2003; Vaquero et al, 2004). In our study, the antioxidant NAC (which reduced ROS) and a SOD mimetic (which reduced O₂^-) both strongly enhanced drug- or peroxide-induced apoptosis of AIF-expressing cells.

Deficiency in mitochondrial respiratory chain complex I (confirmed in the present study) and consequent dependence on the less efficient anaerobic ATP generation (Vahsen et al, 2004) provide another not necessarily mutually exclusive explanation for the raised sensitivity of AIF-deficient carcinoma cells to chemical stress-mediated cell death. OXPHOS deficiencies in animals can lead to cell death, particularly in energy demanding cells (Klein et al, 2002; Atorino et al, 2003; Vahsen et al, 2004). Therefore, tumor cells, which are heavily reliant on mitochondrial ATP production, may be more vulnerable to drug and oxidative stress when AIF is absent. Moreover, cells lacking complex I are less able to effect mitochondrial detoxification of H₂O₂ (Atorino et al, 2003; Zoccarato et al, 2004), which likely contributes to the increased peroxide sensitivity of AIF-deficient cells in our study.

In sum, our findings suggest that AIF helps to counteract chemical stress by mechanisms dependent on a pro-oxidant state and probably also a properly functioning mitochondrial complex I.

AIF maintains the transformed state via its NADH oxidase activity—role of complex I

We found that AIF knockout clones from two distinct colon cancer cell lines, differing in p53 and oncogenic H-ras status, displayed very similar properties characteristic of the loss or reversal of the transformed state. These are lack of anchorage-independent growth in soft agar and extremely poor development of tumors in athymic nude mice. The failure of anchorage-independent growth of AIF-deficient cells was confirmed by siRNA knockdown in SW480 colon carcinoma cells and representative lung and breast carcinoma cells. Anchorage-independent growth was only restored in AIF knockout HCT116 cells by AIF mutants with intact NADH-binding domains. Thus, AIF maintains the tumorigenicity of various carcinoma cell types regardless of p53 function or oncogenic H-ras, and the NADH oxidase activity of AIF is essential for maintaining the transformed state of HCT116 colon carcinoma cells.

NIH3T3 cells transformed with the non-apoptotic AIF mutant formed foci in culture, whereas NIH3T3 cells transformed with AIF harboring deleterious mutations in both the pro-apoptotic and NADH-binding domains did not, which provides additional evidence that the NADH oxidase activity of AIF functions in cancer. Overall, the ability of transfected DNA-binding mutants of AIF (but not unmodified AIF) to stimulate foci formation in nonmalignant NIH3T3 cells and generate soft agar colonies in AIF knockout HCT116 cells indicates a latent transforming effect of AIF that is a reflection of its important role in maintaining the transformed state.

How does the NADH oxidase activity of AIF contribute to the transformed phenotype? There is ample evidence that a pro-oxidant state is associated with the development and maintenance of the transformed phenotype of many tumors, but the source of ROS is unknown (Cerutti, 1985; Szatrowski and Nathan, 1991; Dreher and Junod, 1996; Irani et al, 1997; Ha et al, 2000; Oberley, 2002). One possible source is the superoxide-generating Nox1 NAD(P)H oxidase, a Ras-inducible gene that is functionally required for oncogenic Ras-mediated transformation of fibroblasts and maintenance of their malignant phenotype (Suh et al, 1999; Lambeth, 2004; Mitsushita et al, 2004). Dismutation of O₂^- to H₂O₂ was proposed to mediate the increased cell growth and transformation caused by nox1 overexpression (Arnold et al, 2001).

In our study, although reduction in ROS correlated with loss of tumorigenicity, both O₂^- and ROS scavengers failed to block tumorigenicity at concentrations that significantly reduced cellular ROS, which argues against a prominent role for ROS in maintaining the transformed state of colon cancer cells. Alternatively, complex I function might be linked to tumorigenicity, which is supported by our data showing that the NADH oxidase activity of AIF is required both for full complex I activity and for the transformed state. Interestingly, complex I has been linked to carcinogenesis, since the classical specific complex I inhibitor, rotenone, is known to potently block the formation of spontaneous or chemically induced carcinomas of the tongue, liver and colon in rodents (Cunningham et al, 1995; Yoshitani et al, 2001; Tanaka et al, 2002, and references therein).

Together, our data and these considerations suggest that AIF NADH oxidase-dependent complex I function may be important for tumorigenicity in carcinoma cells rather than ROS per se.

Inhibitors of AIF as potential cancer drugs?

Targeting mitochondrial AIF in cancer is attractive. First, the enhanced rates of glucose uptake and glycolysis of transformed cells enable small molecule inhibitors to accumulate at higher concentrations in the mitochondria of cancer cells (Don and Hogg, 2004). Second, blocking the NADH oxidase activity of AIF should leave its extramitochondrial pro-apoptotic function intact (Ye et al, 2002). Third, whereas in our study an incomplete reduction in AIF expression inhibited anchorage-independent growth, an 80% reduction in AIF in the Harlequin mouse had no adverse effects, except in a subset of neurons and in the retina in the adult animal (Klein et al, 2002). Therefore, specific inhibitors of the NADH oxidase function of mitochondrial AIF that do not cross the blood–brain barrier should be considered as a novel approach to the therapy of colorectal cancers and other carcinomas. We speculate that inhibitors of the NADH oxidase activity of AIF would be differentially toxic to tumor cells and both sensitize colon cancer cells to cancer drugs and block or reverse their tumorigenicity regardless of their p53 or H-ras status.

Cell culture and materials

HCT116 and DLD-1 colon carcinoma cells were maintained in McCoy's 5A Medium (Sigma-Aldrich), SW480 colon carcinoma cells and NIH3T3 cells in DMEM (Sigma-Aldrich) and A549 colon carcinoma cells and MCF7 breast carcinoma cells in RPMI medium. All media were supplemented with 10% fetal bovine serum (FBS; HyClone, Logan, UT) and penicillin (100 U/ml) and streptomycin (100 g/ml) (Sigma-Aldrich).

Inactivation of the human aif gene by homologous recombination

Gene targeting in diploid male cell lines HCT116 and DLD-1 via homologous recombination has been described previously (Sedivy et al, 1999). We utilized this strategy to modify the single X-linked locus of the human aif gene by adapting the design for promoter-trap vectors (Sedivy and Dutriaux, 1999) in construction of the targeting vector. Briefly, PCR amplifications of a 0.6 kb region upstream of exon 1 and 5.2 kb region downstream of exon 1 from the human aif locus of HCT116 genomic DNA were cloned into the pGEM-T vector (Promega, Madison, WI). Exon 1 was replaced by the hyg gene (with a polyadenylation signal) exactly at the start codon of the human aif gene (Figure 1). To further reduce background selection of hygromycin-resistant clones due to nonhomologous recombination, the bipartite method was utilized (Jallepalli et al, 2001). The conventional promoter-trap vector described above was digested with Mlu1, SacII and Pst1 (New England Biolabs, Beverly, MA) to generate vectors with overlapping regions of the truncated hyg gene. The vectors were gel-purified and transfected at 1 g/10⁶ HCT116 or DLD-1 cells via electroporation using the Bio-Rad Gene Pulser system (250 V, 960 F). Transfected cells were plated at 10³ cells/well in 96-well plates and selected in McCoy's 5A medium (Sigma-Aldrich) and 0.3 mg/ml hygromycin (Clontech) for 2–3 weeks. The 96-well plates were screened for individual clones, which were isolated and analyzed by PCR analysis of genomic DNA using the following primers: PgAFA: TTG GGT GGG AAA GAG CCG GCA ACT GGT AAA AG; Hyg-Rev: AGA TTC TTC GCC CTC CGA GAG CTG CAT CAG GT. Candidate aif^-/y clones were confirmed by additional PCR analysis for the loss of the 950 bp wild-type junctions with the following primers: primer A: GAG TCT GCG TAA TGT GCG TGT GAA GAG AGA CTG G; primer B: TTG CCT GGA ATG GGT CAG TCA CCT GGG AG. To further validate gene targeting of the human aif gene, we analyzed clones using RT–PCR by preparing oligo-dT primed cDNA from clones and analyzed with the following mRNA specific PCR primers for AIF and GAPDH: AFRTA: CGG TCG CCG AAA TGT TCC GGT GTG GAG; AFRTB: ACG CGG CCT TTT TCT GTT TCT GTT CTG G; GPDA: CCA ATA TGA TTC CAC CCA TGG C; GPDB: TTC TCC ATG GTG GTG AAG AC. Clones negative for aif mRNA were analyzed for protein expression using Western blotting.

Generation and transfection of AIF mutants

The pEGFP-C1 vector (Clontech) alone and vector containing the human aif coding region in-frame with gfp cDNA were routinely transfected into 2 10⁶ cells using Fugene 6 (Roche) transfection reagent. Stable expressing cells were selected using 1 mg/ml G418 (geneticin; Invitrogen). Mutagenesis of aif was performed using the transformer site-directed mutagenesis kit (Clontech). Mutations introduced were as follows: mDNA1: K255A, R265; mDNA2: K510A, K518A; mDNA1N: K255A, R265A, T263A, V300A; mDNA2N: K510A, K518A, T263A, V300A (Mate et al, 2002; Ye et al, 2002).

Western blot

Western blotting was performed with 10 g protein as described (Towbin et al, 1979) using anti-AIF polyclonal antibody (N-1, 1:2000) (QED Biosciences, San Diego, CA) or anti-GFP polyclonal antibody (Clontech) and horseradish peroxidase-conjugated goat anti-rabbit IgG (New England Biolabs) as secondary probe. The transfer conditions were 10 mM NaHCO₃, 3 mM NaCO₃ and 20% methanol pH 9.9 for AIF, and 192 mM glycine, 20% methanol and 25 mM Tris pH 8.3 for complex I proteins and GFP. All blots were washed with phosphate-buffered saline and 0.5% Tween 20 (Sigma-Aldrich) and developed using enhanced chemiluminescence (Amersham Biosciences).

Aconitase activity and lucigenin assays for O₂^-

O₂^- measurement using the aconitase assay was as described (Gardner et al, 1995). A total of 10⁶ cells were lysed using M-PER (Pierce, Rockford, IL) solution and protein concentration measured using Bio-Rad protein assay solution. A 30 g portion of proteins was incubated with 50 mM Tris–HCl pH 7.4, 0.6 mM MnCl₂, 20 mM NADP⁺, 250 mM sodium citrate pH 7.0 and 1 U isocitrate dehydrogenase (Sigma-Aldrich) in a 200 l reaction volume. Aconitase activity was quantified by measuring the rate of NADPH formation during citrate isomerization to isocitrate at 340 nm using a TECAN microplate reader (Maennedorf, Switzerland) and expressed as mU/mg protein, where mU is nmol/min computed from the NADPH molar extinction coefficient of 6.22 mM^-1 cm^-1. Lucigenin-based chemiluminescence measurement of intracellular O₂^- has been described (Clement and Stamenkovic, 1996). A total of 2 10⁶ cells were lysed in 400 l ATP-releasing solution (Sigma-Aldrich) and pelleted at 12 000 g. Together with the supernatant, 100 l of 850 M lucigenin (Sigma-Aldrich) was immediately added and chemiluminescence was measured for 20 s using a Turner TD-20e Luminometer. Data are expressed as relative light units/s and normalized against total protein content.

Flow cytometric analysis of intracellular ROS

Intracellular ROS levels were determined by staining with H₂DCFH-DA (Molecular Probes, Oregon), which is oxidized to dichlorofluorescein (DCF) by ROS. Cells were exposed to 5 M H₂DCFH-DA at room temperature for 1 h, analyzed by flow cytometry using a Becton Dickinson FACScan machine and analyzed using the CellQuest program.

Cytotoxicity and apoptosis assays

A total of 5 10³ cells were plated in 96-well plates and exposed to different concentrations of drugs: etoposide, campthothecin, staurosporine, doxorubicin and actinomycin D (Sigma-Aldrich); oxidative stress agents t-butylhydroperoxide and hydrogen peroxide (Sigma-Aldrich); and apoptosis inducers TRAIL (Biomol, Philadelphia) and anti-FAS (Roche). Cells were incubated for up to 4 days at 37°C in 5% CO₂. Survival was measured by processing cells with 20 l of CellTiter 96 AQ_ueous One Solution Cell Proliferation Assay ((3-(4,5-dimethylthiazol-2-yl)-5-(3-carboxymethoxyphenyl)-2-(4-sulfophenyl)-2H-tetrazolium (MTS); Promega, Madison, WI) for 2 h and measurement of color reaction absorbance at 485 nm using a TECAN microplate reader, which was expressed as a fraction of untreated cells. Apoptosis was measured using Annexin-V staining, Sytox-Hoechst stain and caspase-3 activity assay as previously described (Hentze et al, 2003; Li et al, 2004). For Annexin-V staining, 5 l of Annexin-V (PE) (Pharmingen) was incubated with 10⁵ pretreated cells for 1 h and analyzed by flow cytometry using a Becton Dickinson FACScan machine and using the CellQuest program.

Measurement of mitochondrial complex I and citrate synthase activities

The NADH-coenzyme Q₁ oxidoreductase activity of complex I was measured using mitochondria isolated using the Mitochondrial Isolation Kit (Pierce, Rockland, IL). Assays were performed as described (Estornell et al, 1993). Mitochondria were freeze-thawed three times and incubated immediately with 50 mM Tris–HCl, pH 8.0, 50 mM KCl, 1 mM EDTA, 2 mM KCN, 150 M NADH and 100 M coenzyme Q₁. All reagents were obtained from Sigma-Aldrich. A time course for NADH oxidation was measured at 340 nM using a TECAN microplate reader at 37°C and calculated using an extinction coefficient of 5.5 mM^-1 cm^-1. Citrate synthase activity was measured using the Citrate Synthase Assay Kit (Sigma-Aldrich) and performed according to the manufacturer's instructions. Protein concentration measurements were performed using the Bio-Rad Protein Assay Solution.

Colony formation in soft agar

Cells were trypsinized and resuspended in McCoy's 5A medium supplemented with 10% fetal bovine serum, 100 U/ml penicillin, 100 g/ml streptomycin and 0.3% noble agar (Difco). A total of 10³ cells/well were plated in six-well plates containing solidified medium with 1% noble agar. Plates were incubated at 37°C in 5% CO₂ and fed with fresh media every 5 days. After 14 days, cells were incubated with 1 mg/ml 3-[4,5-dimethylthiazol-2-yl]-2,5-diphenyltetrazolium bromide (MTT; Sigma-Aldrich) for 1 h.

Tumorigenesis in athymic mice

Cells (5 10⁶) were subcutaneously injected into the flanks of 4- to 5-week-old athymic nude mice (Animal Resource Centre, Australia) and incubated for 3 weeks. Tumors were excised, weighed and dimensions were measured using calipers. Tumor volume (cm³) was determined using the standard formula a² b/2, where a is the width and b is the length of the horizontal tumor perimeter. A total of 10–15 athymic nude mice were used for each cell line in two experiments, and the data were represented by the mean values. All animal experiments were performed strictly in accordance with Institute of Molecular and Cell Biology and A^*Star regulations.

Foci formation

To demonstrate the transforming properties of human AIF, NIH3T3 cells were transfected with cDNA encoding AIF-GFP or various mutants, and the cells were sorted as described above. Cells were continuously cultured for 3–6 weeks and monitored for foci formation. Plates were analyzed after 6 weeks of incubation by staining with crystal violet (0.3% crystal violet, 30% methanol), and only dense colonies more than 2 mm in diameter were scored as foci.

We thank Michael Choi for his help with some experiments. zzAGP is an adjunct staff member of The Department of Surgery, National University of Singapore. This study was supported by funds from A^*STAR to the Institute of Molecular and Cell Biology, Singapore.

Arnold RS, Shi J, Murad E, Whalen AM, Sun CQ, Polavarapu R, Parthasarathy S, Petros JA, Lambeth JD (2001) Hydrogen peroxide mediates the cell growth and transformation caused by the mitogenic oxidase Nox1. Proc Natl Acad Sci USA 98: 5550–5555 | Article | PubMed | ChemPort |

Atorino L, Silvestri L, Koppen M, Cassina L, Ballabio A, Marconi R, Langer T, Casari G (2003) Loss of m-AAA protease in mitochondria causes complex I deficiency and increased sensitivity to oxidative stress in hereditary spastic paraplegia. J Cell Biol 163: 777–787 | Article | PubMed | ISI | ChemPort |

Cande C, Vahsen N, Kouranti I, Schmitt E, Daugas E, Spahr C, Luban J, Kroemer RT, Giordanetto F, Garrido C, Penninger JM, Kroemer G (2004a) AIF and cyclophilin A cooperate in apoptosis-associated chromatinolysis. Oncogene 23: 1514–1521 | Article | PubMed | ISI | ChemPort |

Cande C, Vahsen N, Metivier D, Tourriere H, Chebli K, Garrido C, Tazi J, Kroemer G (2004b) Regulation of cytoplasmic stress granules by apoptosis-inducing factor. J Cell Sci 117: 4461–4468 | Article | PubMed | ISI | ChemPort |

Cerutti PA (1985) Prooxidant states and tumor promotion. Science 227: 375–381 | Article | PubMed | ISI | ChemPort |

Clement MV, Hirpara JL, Pervaiz S (2003) Decrease in intracellular superoxide sensitizes Bcl-2-overexpressing tumor cells to receptor and drug-induced apoptosis independent of the mitochondria. Cell Death Differ 10: 1273–1285 | Article | PubMed | ISI | ChemPort |

Clement MV, Stamenkovic I (1996) Superoxide anion is a natural inhibitor of FAS-mediated cell death. EMBO J 15: 216–225 | PubMed | ISI | ChemPort |

Cunningham ML, Soliman MS, Badr MZ, Matthews HB (1995) Rotenone, an anticarcinogen, inhibits cellular proliferation but not peroxisome proliferation in mouse liver. Cancer Lett 95: 93–97 | Article | PubMed | ChemPort |

Don AS, Hogg PJ (2004) Mitochondria as cancer drug targets. Trends Mol Med 10: 372–378 | Article | PubMed | ChemPort |

Dreher D, Junod AF (1996) Role of oxygen free radicals in cancer development. Eur J Cancer 32A: 30–38 | Article

Droge W. (2002) Free radicals in the physiological control of cell function. Physiol Rev 82: 47–95 | PubMed | ISI | ChemPort |

Estornell E, Fato R, Pallotti F, Lenaz G (1993) Assay conditions for the mitochondrial NADH:coenzyme Q oxidoreductase. FEBS Lett 332: 127–131 | Article | PubMed | ChemPort |

Finkel T. (2003) Oxidant signals and oxidative stress. Curr Opin Cell Biol 15: 247–254 | Article | PubMed | ISI | ChemPort |

Gardner PR, Raineri I, Epstein LB, White CW (1995) Superoxide radical and iron modulate aconitase activity in mammalian cells. J Biol Chem 270: 13399–13405 | Article | PubMed | ISI | ChemPort |

Go YM, Gipp JJ, Mulcahy RT, Jones DP (2004) H₂O₂-dependent activation of GCLC-ARE4 reporter occurs by mitogen-activated protein kinase pathways without oxidation of cellular glutathione or thioredoxin-1. J Biol Chem 279: 5837–5845 | Article | PubMed | ChemPort |

Gupta A, Rosenberger SF, Bowden GT (1999) Increased ROS levels contribute to elevated transcription factor and MAP kinase activities in malignantly progressed mouse keratinocyte cell lines. Carcinogenesis 20: 2063–2073 | Article | PubMed | ISI | ChemPort |

Ha HC, Thiagalingam A, Nelkin BD, Casero Jr RA (2000) Reactive oxygen species are critical for the growth and differentiation of medullary thyroid carcinoma cells. Clin Cancer Res 6: 3783–3787 | PubMed | ChemPort |

Hentze H, Lin XY, Choi MS, Porter AG (2003) Critical role for cathepsin B in mediating caspase-1-dependent interleukin-18 maturation and caspase-1-independent necrosis triggered by the microbial toxin nigericin. Cell Death Differ 10: 956–968 | Article | PubMed | ISI | ChemPort |

Irani K, Xia Y, Zweier JL, Sollott SJ, Der CJ, Fearon ER, Sundaresan M, Finkel T, Goldschmidt-Clermont PJ (1997) Mitogenic signaling mediated by oxidants in Ras-transformed fibroblasts. Science 275: 1649–1652 | Article | PubMed | ISI | ChemPort |

Jallepalli PV, Waizenegger IC, Bunz F, Langer S, Speicher MR, Peters JM, Kinzler KW, Vogelstein B, Lengauer C (2001) Securin is required for chromosomal stability in human cells. Cell 105: 445–457 | Article | PubMed | ISI | ChemPort |

Joza N, Susin SA, Daugas E, Stanford WL, Cho SK, Li CYJ, Sasaki T, Elia AJ, Cheng H-YM, Ravagnan L, Ferri KF, Zamzami N, Wakeham A, Hakem R, Yoshida H, Kong Y-Y, Zuñiga-Pflücker JC, Kroemer G, Penninger JM (2001) Essential role of the mitochondrial apoptosis-inducing factor in programmed cell death. Nature 410: 549–554 | Article | PubMed | ISI | ChemPort |

Klein JA, Longo-Guess CM, Rossmann MP, Seburn KL, Hurd RE, Frankel WN, Bronson RT, Ackerman SL (2002) The harlequin mouse mutation downregulates apoptosis-inducing factor. Nature 419: 367–374 | Article | PubMed | ISI | ChemPort |

Lambeth JD (2004) NOX enzymes and the biology of reactive oxygen. Nat Rev Immunol 4: 181–189 | Article | PubMed | ISI | ChemPort |

Li L, Feng Z, Porter AG (2004) JNK-dependent phosphorylation of c-Jun on serine 63 mediates nitric oxide-induced apoptosis of neuroblastoma cells. J Biol Chem 279: 4058–4065 | PubMed | ChemPort |

Lin KI, Pasinelli P, Brown RH, Hardwick JM, Ratan RR (1999) Decreased intracellular superoxide levels activate Sindbis virus-induced apoptosis. J Biol Chem 274: 13650–13655 | Article | PubMed | ISI | ChemPort |

Lipton SA, Bossy-Wetzel E (2002) Dueling activities of AIF in cell death versus survival: DNA binding and redox activity. Cell 111: 147–150 | Article | PubMed | ISI | ChemPort |

Loeffler M, Daugas E, Susin SA, Zamzami N, Metivier D, Nieminen AL, Brothers G, Penninger JM, Kroemer G (2001) Dominant cell death induction by extramitochondrially targeted apoptosis-inducing factor. FASEB J 15: 758–767 | Article | PubMed | ISI | ChemPort |

Mate MJ, Ortiz-Lombardia M, Boitel B, Haouz A, Tello D, Susin SA, Penninger J, Kroemer G, Alzari PM (2002) The crystal structure of the mouse apoptosis-inducing factor AIF. Nat Struct Biol 9: 442–446 | Article | PubMed | ISI | ChemPort |

Miramar MD, Costantini P, Ravagnan L, Saraiva LM, Haouzi D, Brothers G, Penninger JM, Peleato ML, Kroemer G, Susin SA (2001) NADH oxidase activity of mitochondrial apoptosis-inducing factor. J Biol Chem 276: 16391–16398 | Article | PubMed | ISI | ChemPort |

Mitsushita J, Lambeth JD, Kamata T (2004) The superoxide-generating oxidase Nox1 is functionally required for Ras oncogene transformation. Cancer Res 64: 3580–3585 | Article | PubMed | ISI | ChemPort |

Oberley TD (2002) Oxidative damage and cancer. Am J Pathol 160: 403–408 | PubMed | ChemPort |

Pervaiz S, Cao J, Chao OS, Chin YY, Clement MV (2001) Activation of the RacGTPase inhibits apoptosis in human tumor cells. Oncogene 20: 6263–6268 | Article | PubMed | ISI | ChemPort |

Pervaiz S, Ramalingam JK, Hirpara JL, Clement MV (1999) Superoxide anion inhibits drug-induced tumor cell death. FEBS Lett 459: 343–348 | Article | PubMed | ISI | ChemPort |

Reth M (2002) Hydrogen peroxide as second messenger in lymphocyte activation. Nat Immunol 3: 1129–1134 | Article | PubMed | ISI | ChemPort |

Sedivy JM, Dutriaux A (1999) Gene targeting and somatic cell genetics—a rebirth or a coming of age? Trends Genet 15: 88–90 | Article | PubMed | ISI | ChemPort |

Sedivy JM, Vogelstein B, Liber HL, Hendrickson EA, Rosmarin A (1999) Gene targeting in human cells without isogenic DNA. Science 283: 9 | Article

Suh YA, Arnold RS, Lassegue B, Shi J, Xu X, Sorescu D, Chung AB, Griendling KK, Lambeth JD (1999) Cell transformation by the superoxide-generating oxidase Mox1. Nature 401: 79–82 | Article | PubMed | ISI | ChemPort |

Susin SA, Lorenzo HK, Zamzami N, Marzo I, Snow BE, Brothers GM, Mangion J, Jacotot E, Costantini P, Loeffler M, Larochette N, Goodlett DR, Aebersold R, Siderovski DP, Penninger JM, Kroemer G (1999) Molecular characterization of mitochondrial apoptosis-inducing factor. Nature 397: 441–446 | Article | PubMed | ISI | ChemPort |

Szatrowski TP, Nathan CF (1991) Production of large amounts of hydrogen peroxide by human tumor cells. Cancer Res 51: 794–798 | PubMed | ISI | ChemPort |

Tanaka T, Kohno H, Sakata K, Yamada Y, Hirose Y, Sugie S, Mori H (2002) Modifying effects of dietary capsicain and rotenone on 4-nitroquinolene 1-oxide-induced rat tongue carcinogenesis. Carcinogenesis 23: 1361–1367 | Article | PubMed | ChemPort |

Towbin H, Staehelin T, Gordon J (1979) Electrophoretic transfer of proteins from polyacrylamide gels to nitrocellulose sheets: procedure and some applications. Proc Natl Acad Sci USA 76: 4350–4354 | Article | PubMed | ChemPort |

Vahsen N, Candé C, Brière J-J, Bénit P, Joza N, Larochette N, Mastroberardino PG, Pequignot MO, Casares N, Lazar V, Feraud O, Debili N, Wissing S, Engelhardt S, Madeo F, Piacentini M, Penninger JM, Schägger H, Rustin P, Kroemer G (2004) AIF deficiency compromises oxidative phosphorylation. EMBO J 23: 4679–4689 | Article | PubMed | ISI | ChemPort |

Vaquero EC, Edderkaoui M, Pandol SJ, Gukovsky I, Gukovskaya AS (2004) Reactive oxygen species produced by NAD(P)H oxidase inhibit apoptosis in pancreatic cancer cells. J Biol Chem 279: 34643–34654 | Article | PubMed | ISI | ChemPort |

Wang X, Yang C, Chai J, Shi Y, Xue D (2002) Mechanisms of AIF-mediated apoptotic DNA degradation in Caenorhabditis elegans. Science 298: 1587–1592 | Article | PubMed | ISI | ChemPort |

Wissing S, Ludovico P, Herker E, Buttner S, Engelhardt SM, Decker T, Link A, Proksch A, Rodrigues F, Corte-Real M, Frohlich KU, Manns J, Cande C, Sigrist SJ, Kroemer G, Madeo F (2004) An AIF orthologue regulates apoptosis in yeast. J Cell Biol 166: 969–974 | Article | PubMed | ISI | ChemPort |

Ye H, Cande C, Stephanou NC, Jiang S, Gurbuxani S, Larochette N, Daugas E, Garrido C, Kroemer G, Wu H (2002) DNA binding is required for the apoptogenic action of apoptosis inducing factor. Nat Struct Biol 9: 680–684 | Article | PubMed | ISI | ChemPort |

Yoshitani S, Tanaka S, Kohno H, Takashima S (2001) Chemoprevention of azoxymethane-induced rat colon carcinogenesis by dietary capsaicin and rotenone. Int J Oncol 19: 929–939 | PubMed | ChemPort |

Zoccarato F, Cavallini L, Alexandre A (2004) Respiration-dependent removal of exogenous H₂O₂ in brain mitochondria: inhibition by Ca²⁺. J Biol Chem 279: 4166–4174 | Article | PubMed | ISI | ChemPort |

Main navigation

• Search EMBO Journal

  Search

• Advanced search

• The EMBO Journal

  • Home
  • Aims and scope
  • Current issue
  • Advance online publication
  • Web focuses
  • Archive:
  • Browse by issue
  • Browse by subject
  • Browse by category
  • Free online sample issue
  • Press releases

• Authors and Referees

  • Editorial process
  • Guide for authors
  • Submit an article
  • Guide for referees
  • Editors and Editorial Board
  • Contact editorial office

• Customer Service

  • Subscribe
  • Order sample copy
  • Purchase articles
  • Reprints and permissions
  • Contact NPG
  • Advertising