Oxidant hypersensitivity of Fanconi anemia type C-deficient cells is dependent on a redox-regulated apoptotic pathway.

Fanconi anemia is a genetic disorder characterized by bone marrow failure. Significant evidence supports enhanced apoptosis of hematopoietic stem/progenitor cells as a critical factor in the pathogenesis of bone marrow failure in Fanconi anemia. However, the molecular mechanism(s) responsible for the apoptotic phenotype are incompletely understood. Here, we tested whether alterations in the activation of a redox-dependent pathway may participate in the pro-apoptotic phenotype of primary Fancc -/- cells in response to oxidative stress. Our data indicate that Fancc -/- cells are highly sensitive to oxidant stimuli and undergo enhanced oxidant-mediated apoptosis compared with wild type controls. In addition, antioxidants preferentially enhanced the survival of Fancc -/- cells. Because oxidative stress activates the redox-dependent ASK1 pathway, we assessed whether Fancc -/- cells exhibited increased oxidant-induced ASK1 activation. Our results revealed ASK1 hyperactivation in H2O2-treated Fancc -/- cells. Furthermore, using small interfering RNAs to decrease ASK1 expression and a dominant negative ASK1 mutant to inhibit ASK1 kinase activity, we determined that H2O2-induced apoptosis was ASK1-dependent. Collectively, these data argue that the predisposition of Fancc -/- hematopoietic stem/progenitor cells to apoptosis is mediated in part through altered redox regulation and ASK1 hyperactivation.

Fanconi anemia (FA) 1 is a heterogeneous bone marrow (BM) failure syndrome with cellular abnormalities that include chromosomal instability, increased apoptosis, and cell cycle control defects (1)(2)(3)(4)(5). The diversity of clinical presentation in children with FA is related, in part, to the existence of multiple comple-mentation types, with eight FA complementation group cDNAs being identified thus far (FANCA, FANCC, FANCD1/BRCA2, FANCD2, FANCE, FANCF, FANCG, and FANCL) (6 -14). Despite some clinical variability between individuals with specific FA gene mutations (15,16), the major cause of mortality in all FA complementation types is BM failure (1,2,5). These studies suggest that apoptotic loss of hematopoietic stem/progenitor cells has a key pathogenetic role in this disorder. Thus, understanding molecular mechanisms involved in the predisposition of FA cells to apoptosis is of critical importance to improve current treatment approaches for children with FA.
Participation of FANCC in redox metabolism has been proposed previously and is supported by functional interactions with cytochrome P-450 reductase (CPR) (31) and glutathione S-transferase P1 (GSTP1) (32). Evidence of oxygen sensitivity was first provided by Joenje et al. (34), who demonstrated that the chromosomal instability of primary FA cells could be reduced if grown at lowered oxygen tension (33). Several conflicting studies have been reported since this original description, although most analyses were conducted in immortalized cell lines or in cells with unidentified FA complementation type and/or mutation (34 -36). Now with the availability of murine models, the issue of whether FA cells have altered redox regulation is being readdressed in primary cells. Using mice deficient in the murine FANCC homologue (Fancc), Hadjur et al. (37) showed that mice mutant at both the Fancc and superoxide dismutase 1 (SOD1) loci exhibit severe defects in hematopoiesis, including histological evidence of BM hypoplasia, an observation not detected in singly mutant mice. Although these data provide strong genetic evidence that Fancc Ϫ/Ϫ cells are hypersensitive to endogenously generated oxidants, it is unknown whether the molecular mechanism responsible for this hypersensitive response is due to altered redox signaling.
Redox signaling has a critical role in controlling multiple complex cellular processes including apoptosis, proliferation, senescence, and differentiation (38 -44). This highly conserved regulatory process involves maintenance of the intracellular environment in an overall reduced state. Cellular oxidative stress results in the oxidation of key cysteine residues on redoxsensitive proteins, a post-translational modification that affects intracellular signaling pathways in a fashion similar to phosphorylation. A notable example of redox apoptotic signaling involves the serine-threonine kinase apoptosis signal-regulating kinase 1 (ASK1). In the normal reducing environment of a cell, ASK1 activity is inhibited via binding to thioredoxin, glutaredoxin, and glutathione S-transferases (45)(46)(47)(48)(49)(50)(51). After direct or indirect oxidant stress (i.e. H 2 O 2 , TNF-␣, glucose/serum deprivation, and heat shock), these proteins are oxidized forming intramolecular disulfide bonds, which result in a conformational change and dissociation from ASK1. Unbound ASK1 is then available to autophosphorylate and subsequently phosphorylate downstream kinases, initiating an apoptotic program.
To extend our understanding of oxidant hypersensitivity in FA, we investigated whether primary Fancc Ϫ/Ϫ cells exhibit alterations in the redox-dependent ASK1 apoptotic pathway. Our data demonstrate that Fancc Ϫ/Ϫ cells exhibit ASK1 hyperactivation after H 2 O 2 treatment. In addition, we show that enhanced H 2 O 2 -induced apoptosis in Fancc Ϫ/Ϫ cells is ASK1dependent and that pretreatment with antioxidants preferentially protects Fancc Ϫ/Ϫ cells from apoptosis induced by H 2 O 2 as compared with controls. Collectively, these data argue that the predisposition of primary Fancc Ϫ/Ϫ cells to oxidant-induced apoptosis is mediated through hyperactivation of a redoxdependent apoptotic signaling pathway.

EXPERIMENTAL PROCEDURES
Mice-Fancc ϩ/Ϫ mice in a C57Bl/6J genetic background were bred to generate Fancc Ϫ/Ϫ and wild type (WT) mice for hematopoietic progenitor assays and timed embryos for murine embryo fibroblast (MEF) cell lines as previously described (52,53). All of the studies were approved by the Indiana University Laboratory Animal Research Center.
Hematopoietic Progenitor Assays-WT and Fancc Ϫ/Ϫ BM low density mononuclear and ckitϩlinϪ cells were prepared as previously described (54,55). Cells from WT and Fancc Ϫ/Ϫ mice were resuspended in Iscove's modified Dulbecco's medium (Invitrogen) supplemented with 20% fetal calf serum (Biowhittaker, Walkersville, MD) and then exposed to H 2 O 2 (Sigma) for 1 h. After oxidant treatment, the cells were washed and plated in clonogenic assays as described previously (55). For hyperoxia exposure, low density mononuclear cells were placed in an airtight chamber before infusing with a gas mixture containing 50% O 2 , 5% CO 2 , and 45% N 2 (Praxair, Indianapolis, IN). The chamber was then incubated for 4 or 16 h at 37°C before the cells were harvested for clonogenic assays. An O 2 analyzer was used to measure the O 2 concentration before and after each incubation period (50 ϩ 3%) to ensure an airtight culture system. Control cultures were incubated at 21% O 2 for 4 or 16 h.
MEF Survival Assays-MEFs were maintained as previously described (53). All of the studies were conducted in at least two or three different MEF cell lines/genotype, and only MEFs that were less than passage 5 were utilized. To assess H 2 O 2 sensitivity, WT and Fancc Ϫ/Ϫ MEFs were cultured with H 2 O 2 for 24 h before assessing viability by trypan blue exclusion. In some experiments, MEFs were pretreated with 20 M selenomethionine (SeMet; Sigma) overnight or 4 mM Nacetylcysteine (NAC; Sigma) for 1 h prior to culturing with H 2 O 2 . To evaluate apoptosis, MEFs were treated with 100 M H 2 O 2 for 4 -6 h and analyzed by the terminal deoxynucleotidyltransferase-mediated nickend labeling (TUNEL) assay as previously described (55,56).
Retroviral Constructs and Transduction-PG13 retroviral packaging cells containing the FANCC mutants (FANCC-E251A and FANCC-del322G) in the pLXSN backbone were generously provided by Dr. Grover C. Bagby, Jr. (Oregon Health Sciences, Portland, OR) (29). Retroviral supernatants were harvested and utilized to transduce GPϩE86 retroviral packaging cells as previously described (52) to pseudotype viral particles with an ecotropic envelope. The MFG-FAC retrovirus encoding the FANCC cDNA was used as a control, which previously was shown to correct mitomycin C sensitivity of Fancc Ϫ/Ϫ cells to WT levels (52). The dominant negative ASK1 cDNA (ASK1-K709M) (57) was generously provided by Dr. Hidenori Ichijo (University of Tokyo, Tokyo, Japan) in a pcDNA plasmid. The ASK1-K709M cDNA was subcloned into the NotI site of the bicistronic retroviral plasmid MIEG3 (58), which is 5Ј to an internal ribosomal entry siteenhanced green fluorescent protein (EGFP) cassette. A GPϩE86 packaging cell line was developed for MIEG3 and MIEG3-ASK1-K709M as previously described (52). Retroviral supernatants were collected, filtered, and stored at Ϫ80°C until utilized for transduction of MEFs. Early passage MEFs (P0 -P1) were transduced with retroviral supernatants four times over 2 consecutive days in the presence of polybrene as previously described (52,53).
ASK1 in Vitro Kinase Assay-ASK1 kinase activity was determined by depriving MEFs of serum for 4 h followed by treatment with 100 M H 2 O 2 for 5 min. MEFs were then washed twice with cold phosphatebuffered saline containing 1 mM sodium orthovanadate and lysed in nonionic lysis buffer. The protein concentrations were determined using the BCA assay (Pierce). The ASK1 immunoprecipitations were conducted using protein A Sepharose beads (Amersham Biosciences) and anti-ASK1 antibody (Cell Signaling, Beverly, MA). Immunobeads were subjected to an in vitro kinase reaction using either myelin basic protein (Sigma) or MKK4 (Upstate Biotechnologies, Inc.) as substrates for ASK1. The kinase mixtures contained 20 mM MgCl 2 , 0.1 M sodium orthovanadate, 1 M dithiothreitol, 30 mM ␤ glycerol phosphate, 5 mM EGTA, 20 mM MOPS, 1 M ATP, and 10 g of substrate/sample before adding 2.5 Ci of [␥-32 P]ATP/sample. The kinase reaction buffer was added to each sample and incubated at 30°C for 30 min. The reactions were terminated by the addition of sample buffer. The protein samples were separated on a 12% SDS-PAGE gel (Invitrogen), transferred to a nitrocellulose membrane, and subjected to autoradiography.
Western Blotting-Equivalent amounts of protein (200 -500 g) were separated on a 12% SDS-PAGE gel and transferred onto a nitrocellulose membrane. For immunodetection of FANCC mutants, a primary rabbit anti-FANCC antibody, previously generated by our laboratory (52), and a secondary anti-rabbit horseradish peroxidase antibody (Amersham Biosciences) were used as described (52) before visualizing by chemiluminescence (Amersham Biosciences). To document equal protein loading, the membrane was stripped and reprobed with ␤-actin (Sigma). For immunodetection of ASK1, rabbit anti-ASK1 antibody (Cell Signaling) was used at a 1:200 dilution before incubating with the secondary antibody anti-rabbit horseradish peroxidase (1:2000 dilution).
Small Interfering RNA Transfection Protocol- Table I lists the RNA sequences used for these studies. The ASK1 siRNA sequence targeted nucleotides 570 -590 of the ASK1 mRNA and was designed according to the manufacturer's recommendations (Dharmicon, Lafayette, CO). Either sense or scrambled oligonucleotides were used as a control for every transfection experiment. WT and Fancc Ϫ/Ϫ MEFs were cultured in a 6-well tissue culture dish to 30 -50% confluency. The RNA oligonucleotides were diluted in Opti-MEM (Invitrogen) to obtain a 250 nM solution per Dharmicon's recommendations. Oligofectamine transfections were conducted exactly per the manufacturer's recommendations (Invitrogen). Following the transfection, 500 l of Dulbecco's modified Eagle's medium (Invitrogen) containing 30% fetal calf serum was added without removing the transfection mixture. The cells were incubated for 72 h at 37°C before harvesting for H 2 O 2 cytotoxicity assays and ASK1 Western blotting. Four independent transfections were conducted with similar results.
Statistical Analyses-For all data shown, a Student's t test was conducted to evaluate for differences between treatment groups, and a p value Յ 0.05 was considered significant.   1A). In addition, Fancc Ϫ/Ϫ progenitors exposed to 50% O 2 for 4 or 16 h exhibited a marked reduction in colony formation as compared with WT control cultures (Fig. 1B). BM low density mononuclear cells are a heterogeneous cell population that includes a significant proportion of differentiated cells compared with the relatively low frequency of clonogenic progenitor cells (0.01-0.5%). Given our previous observations that Fancc Ϫ/Ϫ progenitors are exquisitely sensitive to multiple inhibitory cytokines such as interferon-␥ and TNF-␣ (55), together with the knowledge that inflammatory cells such as lymphocytes and granulocytes are major sources of secreted inhibitory cytokines, it was crucial to eliminate these differentiated cells from our culture system. To test whether the observed oxidant hypersensitivity was due to an intrinsic abnormality in Fancc Ϫ/Ϫ progenitor cells and not secondary to accessory cells present in BM low density mononuclear cell populations, WT and Fancc Ϫ/Ϫ ckitϩlinϪ cells were purified by fluorescence cytometry, treated with 100 M H 2 O 2 , and plated in colony assays. This phenotypically defined cell population enriches for immature hematopoietic stem and progenitor cells and excludes differentiated progeny cells. Similar to prior studies with low density mononuclear cells, Fancc Ϫ/Ϫ ckitϩlinϪ cells were hypersensitive to H 2 O 2 (Fig. 1C), supporting an intrinsic hematopoietic progenitor cell defect. Because of the difficulty in obtaining sufficient numbers of primary hematopoietic progenitor cells, we established WT and Fancc Ϫ/Ϫ MEFs to utilize as a cellular model system for evaluation of alterations in oxidant responsiveness in Fancc Ϫ/Ϫ cells. Initial studies determined that Fancc Ϫ/Ϫ MEFs exhibited an enhanced sensitivity to H 2 O 2 ( Fig. 2A), comparable with that observed in Fancc Ϫ/Ϫ progenitors. TUNEL assays confirmed that the H 2 O 2 hypersensitivity in Fancc Ϫ/Ϫ MEFs was due to enhanced apoptosis (Fig. 2B). In addition, we observed that Fancc Ϫ/Ϫ MEFs exhibit a slight increase in apoptosis when grown in basal conditions, similar to previous data in cultured Fancc Ϫ/Ϫ hematopoietic cells (52). To test whether antioxidants protect Fancc Ϫ/Ϫ MEFs from oxidant exposure, MEFs were pretreated with either SeMet or NAC before culturing with H 2 O 2 . Consistent with an altered redox state, Fancc Ϫ/Ϫ MEFs pretreated with SeMet or NAC were protected from H 2 O 2 treatment compared with control Fancc Ϫ/Ϫ MEFs (Fig. 2C). Importantly, survival of Fancc Ϫ/Ϫ MEFs was restored to WT levels after SeMet and NAC pretreatment. Furthermore, SeMet pretreatment of Fancc Ϫ/Ϫ MEFs reduced apoptosis in basal and H 2 O 2 -treated conditions (Fig. 2D).

Fancc
Correction least two independent functions, separable by FANCC mutant cDNAs (29). One recognized function is to maintain normal alkylating agent sensitivity, which is associated with restoration of nuclear FA protein complex formation (corrected by FANCC-E251A mutant). The second identified function is to sustain normal PKR-mediated apoptotic signaling via an interaction with HSP70 (corrected by FANCC-del322G mutant, an FA patient-derived mutation resulting in deletion of amino acids 1-54). To determine whether oxidant hypersensitivity in Fancc Ϫ/Ϫ MEFs segregates with a known FANCC function, we transduced Fancc Ϫ/Ϫ MEFs with a retrovirus encoding either FANCC, FANCC-E251A, or FANCC-del322G. Mock infected Fancc Ϫ/Ϫ MEFs were utilized as a control. FANCC expression in transduced MEFs was confirmed by Western blotting (Fig. 3A) before evaluating H 2 O 2 sensitivity (Fig. 3B)  exhibited a significant decrease in ASK1 expression (Fig. 5A, a  representative experiment). Interestingly, Fancc Ϫ/Ϫ MEFs transfected with the ASK1 siRNA oligomer were completely protected against H 2 O 2 treatment compared with untreated controls (Fig. 5B), whereas WT MEFs transfected with the ASK1 siRNA oligomer did not exhibit a significant increase in survival.
To further test whether H 2 O 2 -induced apoptosis in Fancc Ϫ/Ϫ MEFs was dependent on ASK1 kinase activity, we constructed a bicistronic retroviral vector that encodes a catalytically inactive, dominant negative ASK1 cDNA and EGFP (MIEG3-ASK1-K709M), allowing for selection of transduced cells (EGFPϩ). WT and Fancc Ϫ/Ϫ MEFs were transduced with either MIEG3-ASK1-K709M or vector control and then evaluated for H 2 O 2 -induced apoptosis. Fancc Ϫ/Ϫ MEFs trans-duced with MIEG3-ASK1-K709M exhibited significantly less apoptosis after H 2 O 2 treatment compared with Fancc Ϫ/Ϫ MEFs transduced with vector alone (Fig. 5C). Consistent with previously published data for ASK1 Ϫ/Ϫ MEFs, inhibiting ASK1 kinase activity in WT MEFs also resulted in reduced apoptosis after H 2 O 2 treatment, although the inhibition was more dramatic in Fancc Ϫ/Ϫ MEFs. Collectively, these data show that the predisposition of Fancc Ϫ/Ϫ MEFs to H 2 O 2 -mediated apoptosis is ASK1-dependent and identify ASK1 as a critical mediator of oxidant-induced apoptosis in Fancc Ϫ/Ϫ cells. DISCUSSION The first demonstration of oxygen hypersensitivity in FA cells was made over 20 years ago (33); however, since that initial report little progress has been made to understand the molecular mechanisms involved. In fact, contradictory data have led to significant debate (60,61). Our data in primary Fancc Ϫ/Ϫ hematopoietic progenitors and MEFs clearly demonstrate hypersensitivity to oxidative agents. Furthermore, these studies are the first to establish an intrinsic defect in the oxidant responsiveness of Fancc Ϫ/Ϫ progenitors, independent of differentiated hematopoietic cell populations and the BM microenvironment. Our data compliment previous studies showing that Fancc Ϫ/Ϫ mice devoid of superoxide dismutase 1 expression develop a hypoplastic BM (37) linking increased endogenous oxidant stress with marrow failure in Fancc Ϫ/Ϫ mice. Together these observations suggest that when Fancc Ϫ/Ϫ hematopoietic cells encounter an increase in either endogenous or exogenous oxidant stress, enhanced apoptotic cell loss may occur, contributing to the pathogenesis of BM failure in FA.
Numerous reports in immortalized cell lines suggest that the loss of FA protein function may result in a pro-oxidant cellular environment (31-33, 60, 62-71). Furthermore, protein-protein interaction studies provide support for the concept that FANCC may directly participate in redox metabolism by interacting with two cytoplasmic binding proteins, GSTP1 (32) and CPR (31). These studies demonstrated that FANCC regulates the activity of both GSTP1 and CPR, with the loss of FANCC predicting increased GSTP1 oxidation (decreased activity) and increased CPR activity. In a reduced conformation GSTP1 inhibits stress-activated apoptotic signaling (50,72,73), suggesting that the loss of FANCC may result in altered redox-dependent stress signaling. In addition, CPR transfers electrons to cytochromes and molecular oxygen (74 -77); hence increased CPR activity in FANCC deficient cells may subsequently result in a pro-oxidant cellular environment by generating increased reactive oxygen species. We reasoned that if an altered redox environment was responsible for the oxidant hypersensitivity in Fancc Ϫ/Ϫ MEFs, pretreatment with antioxidants that provide additional cellular reducing equivalents would preferentially protect Fancc Ϫ/Ϫ cells compared with WT controls. Indeed, our data showed that SeMet or NAC pretreatment not only preferentially protected Fancc Ϫ/Ϫ MEFs from H 2 O 2induced apoptosis but improved viability to WT levels, supporting a potential role for FANCC in redox metabolism.
Interestingly, we and others have previously shown that Fancc Ϫ/Ϫ cells are exquisitely hypersensitive to TNF-␣ (55,78), a potent activator of ASK1 (46,59,79). Pang et al. (28,29) demonstrated that binding of FANCC to HSP70 acts as a negative regulator for PKR-mediated apoptotic signaling induced by costimulation with interferon-␥ and TNF-␣. To evaluate whether H 2 O 2 -induced apoptosis was dependent on HSP70 binding to FANCC, we transduced Fancc Ϫ/Ϫ MEFs with a retrovirus containing a FANCC construct mutated at the HSP70-binding site (FANCC-E251A). These studies demonstrated that the FANCC-E251A mutant completely corrected the sensitivity of Fancc Ϫ/Ϫ MEFs to H 2 O 2 , suggesting a HSP70/PKR independent mechanism. In contrast, the FANCC-del322G mutant did not protect against H 2 O 2 -induced apoptosis in Fancc Ϫ/Ϫ MEFs. This is a particularly intriguing observation because CPR binding to FANCC is predicted to be disrupted by this mutation (31). Given  directly controlled by the cellular redox environment. Since this original description, it is now recognized that the redox regulation of ASK1 is complex, involving the direct interaction of ASK1 with multiple redox-dependent binding partners including thioredoxin, glutathione S-transferases, and glutaredoxin (45)(46)(47)(48)(49)(50)(51). The precise physiologic role that individual negative regulators have in inhibiting ASK1 activity remains unclear. However, there is evidence that these redox-dependent proteins may act as sensors for specific cellular redox stresses (46,49,51). Collectively, these studies suggest that ASK1 functions as a major modulator of apoptotic signaling induced by multiple types of oxidant stress including H 2 O 2 , TNF-␣, glucose/serum deprivation, and heat shock (46,48,51,80). In addition, because ASK1 activity is tightly regulated by redox-dependent mechanisms (45)(46)(47)(48), these observations support the idea that Fancc Ϫ/Ϫ cells may exhibit an altered redox environment, which predisposes to ASK1-mediated apoptosis.
An intriguing possible explanation for the enhanced propensity of Fancc Ϫ/Ϫ MEFs to ASK1-mediated apoptosis may be due to disruption of GSTP1 redox control. Previous studies showed that FANCC expression maintains GSTP1 in a reduced conformation during growth factor withdrawal and subsequently protects from apoptotic cell death (32). Importantly, reduced glutathione S-transferases are critical for inhibition of ASK1 activation and consequently oxidant-induced apoptosis (45,50,51). Interestingly, Gilot et al. (50) reported that GSTP1 overexpression in primary hepatocytes protected from ASK1dependent apoptosis, demonstrating that GSTP1 regulates ASK1 activity. Future investigation of the underlying mechanism that initiates ASK1 hyperactivation in Fancc Ϫ/Ϫ cells will be important to understand the role that FANCC has in preserving survival after oxidant stress. Regardless, our data in primary Fancc Ϫ/Ϫ cells identify ASK1 as a potentially important molecular target to defend against an apoptotic fate induced by oxidant stress.
In summary, our data indicate that Fancc Ϫ/Ϫ progenitors exhibit an intrinsic defect in oxidant responsiveness. In addition, we show that the hypersensitivity of Fancc Ϫ/Ϫ cells to oxidative stress is ASK1-dependent. Furthermore, our data showing that antioxidants protect Fancc Ϫ/Ϫ cells from enhanced oxidant-induced apoptosis suggest a potential translational role for antioxidants in the prevention and/or delay of BM failure in FA. Future preclinical studies to test the potential of such an approach will be important.