Impaired Type I IFN-Induced Jak/STAT Signaling in FA-C Cells and Abnormal CD4⁺ Th Cell Subsets in Fancc^−/− Mice¹

Fanconi anemia (FA)³ is an autosomal recessive disease characterized by cellular hypersensitivity to DNA cross-linking agents, various congenital abnormalities, progressive bone marrow failure, and a high incidence of malignancy. The most common morphologic features of FA include skin anomalies; short stature; abnormal limb, kidney, and gastrointestinal development; and male hypogonadism (1). The leading causes of mortality are bone marrow failure and both epithelial and hemopoietic malignancies (2, 3). Although the extraordinary incidence of certain cancers in FA remains unexplained, the frequency of head and neck and vulvar cancers suggests that the human papillomavirus (HPV) may play a role in certain FA epithelial malignancies, and recent studies confirm this notion (4, 5, 6). These findings are compatible with the view that FA may be characterized by subtle immune defects: a view supported by mounting preliminary studies and FA patient case reports that suggest subtle defects of lymphoid subsets, especially those involved in cell-mediated immunity (7, 8, 9, 10, 11, 12). In addition, FA patients undergoing transplantation suffer higher rates of aspergillus infection, severe graft-vs-host disease, and high levels of graft rejection (13, 14). However, to date no experimental evidence has been developed to support the notion that immune dysfunction is a component of the FA phenotype or that impaired immune surveillance may be an underlying cause of the high incidence of certain malignancies in FA.

Eight of the FA genes have been identified and include: Fanca (15), Fancc (16), BRCA2 (Fancd1) (17), Fancd2 (18), Fance (19), Fancf (20), Fancg (xrcc9) (21), and Fancl (22, 23). All of the proteins encoded by these genes except BRCA2 and FANCD2 interact in a nuclear core complex. The core complex is required for monoubiquitination of FANCD2 and is believed to facilitate repair responses to DNA damage (24). However, based upon studies on FANCC, there is emerging evidence that individual FA proteins function in other pathways independent of the nuclear complex (3). Of particular relevance to immune function is the role of FANCC in facilitating activation of STAT1 in response to IFN-γ (25, 26). FANCC interacts with STAT1 following stimulation with IFN-γ and is required for proper docking of STAT1 at the IFN-γ receptor α-chain (IFN-γRα). IFN-γ-induced activation of Jak1, Jak2, and the IFN-γRα occurs normally in FA-C cells (25).

Type I and II IFNs communicate with one another, and cells derived from IFN-αR1 null mice (IFN-αR1^−/−) have impaired IFN-γ-induced responses (27). Given the cross talk that exists between type I and II IFNs, and that FA-C cells have defective IFN-γ (type II IFN) signaling through STAT1, we quantified early responses to type I IFNs in FA-C cells. We find that phosphorylation of STAT1, STAT3, and STAT5 is reduced in response to type I IFNs. However, these STAT phosphorylation defects are specific to certain ligands, because phosphorylation of STAT3 occurred normally in FA-C cells exposed to LIF or IL-6. Further analysis of the IFN-αβ signaling pathway revealed that phosphorylation of the Janus kinases, tyrosine kinase (Tyk)2 and Jak1, are reduced in FA-C cells in response to IFN-α and IFN-β, revealing a defect upstream of STAT signaling.

Cells derived from Tyk2^−/− mice have reduced type I IFN-induced STAT phosphorylation, similar to what is found in Fancc^−/− cells. In addition, differentiation of Th cell subsets is impaired in Tyk2^−/− mice (26, 27). We sought to determine whether other phenotypic abnormalities observed in Tyk2^−/− mice were also present in Fancc^−/− mice. Taking into account that certain STAT proteins and the FANCC-interacting protein repressor of GATA/Franconi anemia zinc finger/testis zinc finger protein are required for proper differentiation of Th cell subsets (28), and that FA patients are reported to have reduced numbers of CD4⁺ cells (12), we hypothesized that these specific signaling defects in FA-C cells may lead to a relative Th subset deficiency. To test this hypothesis, we removed splenocytes from mice nullizygous (Fancc^−/−) or heterozygous (Fancc^+/−) at the Fancc locus, isolated CD4⁺ cells, and quantified IFN-γ production from activated CD4⁺ cells. Fancc^−/− CD4⁺ cells released lower levels of the Th1 cytokine, IFN-γ, and have fewer IFN-γ-secreting CD4⁺ cells. Moreover, IFN-γ production was reduced in Fancc^−/− splenocytes placed under in vitro conditions favoring induction of the Th1 subset. Taken together, these data suggest that differentiation of Fancc^−/− CD4⁺ cells into the Th1 subset is impaired in Fancc^−/− mice.

EBV-transformed human lymphoblasts were maintained in RPMI 1640 medium (Invitrogen Life Technologies, Grand Island, NY) supplemented with 15% heat-inactivated FCS, and grown in a humidified 5% CO₂-containing atmosphere at 37°C. The lymphoblast lines, JY (normal), HSC536N (FANCC mutant), HSC536/FANCC (corrected), and HSC536/neo (mutant, transduced with vector only), were described previously (29). Murine embryonic fibroblasts (MEFs) were established from Fancc knockout and wild-type mice and transformed, as previously described (26, 30, 31).

B lymphoblasts were plated 1 day before treatment. Cells were incubated with 1 ng/ml human rIFN-γ (R&D Systems, Minneapolis, MN), 50 U/ml human rIL-6 (R&D Systems), 1000 U/ml human rIFN-α2β (Schering-Plough, Kentworth, NJ), or 1000 U/ml human rIFN-β (Biogen Idec, Cambridge, MA) for 15 min at 37°C, 5% CO₂. MEFs were serum starved for 24 h and treated with the indicated amount of murine rLIF (R&D Systems) for 15 min at 37°C, 5% CO₂.

Total cell lysates were made using radioimmunoprecipitation assay buffer (26). SDS-PAGE and Western analysis were performed, as previously described (26). Blots were incubated with primary Ab: total STAT1 (Santa Cruz Biotechnology, Santa Cruz, CA) (1/2000), STAT2 (Santa Cruz Biotechnology) (1/1000), total STAT3 (Santa Cruz Biotechnology) (1/1000), total STAT5 (Santa Cruz Biotechnology) (1/500), Tyk2 (BD Transduction Laboratories, Lexington, KY) (1/1000), phospho-specific STAT1 (Cell Signaling Technology, Beverly, MA) (1:500), phospho-specific STAT3 (Cell Signaling Technology) (1/1000), phospho-specific STAT5 (Cell Signaling Technology) (1/1000), phospho-specific Jak1 (1:250) (Cell Signaling Technology), and phospho-specific Tyk2 (Cell Signaling Technology) (1/1000) at the indicated dilutions for 20 h at 4°C, washed with TBST, then incubated with the appropriate secondary Ab for 1 h at room temperature, and washed again with TBST. Ab-reactive proteins were detected using ECL (Amersham, Arlington Heights, IL).

Following cytokine treatment, cells were washed twice with ice-cold PBS and lysed with digitonin lysis buffer (1% digitonin, 50 mM Tris-HCl, pH 8.0, 150 mM NaCl, 50 mM NaF, 5 mM sodium pyrophosphate) supplemented with the following freshly made protease inhibitors: 1% aprotinin, 1 μg/ml leupeptin, 1 mM PMSF, and 2 mM sodium orthovanadate. After rocking at 4°C for 1 h, cell lysates were cleared by centrifugation at 13,200 rpm for 20 min at 4°C, and protein concentrations were determined by the Bradford method using a protein microassay reagent (Bio-Rad, Hercules, CA). Whole cell extracts (WCE) (1 mg of total protein) were precleared with 50 μl of 50% protein A-Sepharose suspension (Pharmacia Biotech, Uppsala, Sweden) for 1 h at 4°C. After separation of the protein A-Sepharose from the lysate by centrifugation at 1000 rpm for 1 min at 4°C, the lysates were incubated with either anti-phosphorylated tyrosine Ab 4G10 (a gift from B. Druker, Oregon Health and Science University) (12 μg/ml) or anti-IFN-αR1 (Research Diagnostics, Flanders, NJ) for 3–12 h at 4°C with constant agitation. Immunocomplexes were then bound to protein A-Sepharose beads (50 μl of 50% slurry) during a 2-h incubation at 4°C. The immunoprecipitates were recovered by centrifuging at 1000 rpm for 1 min at 4°C. The immunocomplexes were washed three times with 0.1% digitonin wash buffer (same as digitonin lysis buffer, except that the concentration of digitonin was 0.1%). Immunoprecipitates were then run on SDS-PAGE, electroblotted, and probed as standard immunoblots.

Fancc^−/− mice were previously described (30). All mice used were from a C57BL6/129SvJ mixed background. Mice were euthanized by cervical dislocation. Spleens were removed, minced, and passed through 50 mesh screens and an 18 G needle to make single cell suspensions. Splenocytes were mixed with CD4⁺ microbeads (Miltenyi Biotec, Auburn, CA), washed, and passed through MACS magnetic columns (Miltenyi Biotec), as per manufacturer’s instructions. Flow through was collected as CD4-depleted cells. The column was removed from the magnetic field, and the CD4⁺ fraction was collected. Isolated CD4⁺ cells were incubated on anti-CD3-coated plates (BD Discovery Labware, Bedford, MA) for 16–20 h.

Mouse IFN-γ secretion assay kits (Miltenyi Biotec) were used according to the manufacturer’s instructions. Cells were kept on ice and washed with cold wash buffer (PBS containing 2.5% BSA and 2 mM EDTA). Cells were incubated with cytokine catch reagent (CD45 Ab conjugated to anti-IFN-γ) for 45 min at 37°C and then washed twice with cold wash buffer. Cells were incubated for 10 min with an IFN-γ-specific detection Ab conjugated to PE. Cells were then stained with an Ab specific for CD4 conjugated to FITC. Cells were washed twice with cold buffer. Immediately before flow cytometric analysis, cells were stained with propidium iodide.

Quantitative sandwich enzyme immunoassay kits for murine IFN-γ and murine IL-4 (mIL-4) were obtained from R&D Systems and were run as per manufacturer’s recommendations. Briefly, standards, controls, and samples were pipetted into IFN-γ- or IL-4-precoated microplate wells. Any unbound substrate was washed away, and an enzyme-linked polyclonal Ab specific for mouse IFN-γ or IL-4, respectively, was added to the wells. Wells were washed to remove any unbound Ab-enzyme reagent. The reaction was stopped by addition of a stop solution, and color intensity was measured on a microplate reader. IFN-γ and IL-4 levels were then determined using the standard curve.

Mice were euthanized by cervical dislocation. Spleens were removed, and single cell suspensions were made. Splenocytes were incubated on plate-bound anti-CD3 plates in RPMI 1640 containing 15% FCS alone or under conditions known to induce production and survival of either Th1 or Th2 subsets. Th1 conditions included 10 μg/ml anti-IL-4 (R&D Systems) and 10 ng/ml rmIL-12 (R&D Systems) added immediately, and 100 U/ml rmIL-2 added the following day. Th2 conditions included anti-IFN-γ (R&D Systems) (10 μg/ml) added immediately, and 10 ng/ml rmIL-4 and 100 U/ml rmIL-2 added the following day. After 3 days, medium and cytokines were replaced and cells were cultured for 3 more days. Cells were washed, and an equal number of cells was replated on anti-CD3-bound plates in RPMI 1640 with 15% FCS in the absence of any additional cytokines for 20 h.

FANCC is required for proper IFN-γ-induced STAT1 activation. Given that cross talk occurs between type I (IFN-αβ) and type II (IFN-γ) IFNs (27), we quantified levels of phosphorylated STAT proteins in response to type I IFNs. Cells were treated with 1000 U/ml IFN-α, 1000 U/ml IFN-β, or 1 ng/ml IFN-γ for 15 min. WCE were used for immunoblot analysis. Blots were probed with Abs specific for the tyrosine-phosphorylated forms of STAT1, STAT3, and STAT5. Phosphorylation of STAT1 is decreased in response to IFN-α, IFN-β, and IFN-γ in FANCC mutant B lymphoblasts, HSC536 and HSC536/neo compared with isogenic cells corrected for the defect by retroviral transduction of the FANCC cDNA, HSC536N/FANCC, and normal B lymphoblasts, JY. Levels of total STAT1 are slightly increased in HSC536 and HSC536/neo compared with corrected and normal cells (but phosphorylation levels are reduced) (Fig. 1,A). Phosphorylation of STAT3 was also decreased in response to IFN-α, IFN-β, and IFN-γ in the FA-C cells, HSC536 and HSC536/neo, compared with wild-type (WT) and corrected cells, JY and HSC536/FANCC. Levels of total cellular STAT3 were equivalent between mutant and corrected cells (Fig. 1,B). Phosphorylation of STAT5 in response to IFN-α was also reduced in the FANCC mutant cells, HSC536 and HSC536/neo, compared with the corrected cell line, HSC536/FANCC. No differences in total cellular STAT5 were detected (Fig. 1 C). These STAT phosphorylation defects were confirmed in another FA-C cell line, PD149 (data not shown). Thus, tyrosine phosphorylation of multiple STAT proteins is impaired in FA-C cells.

Phosphorylation of multiple STAT molecules is decreased in FA-C cells in response to type I and II IFNs. We therefore questioned whether decreased STAT activation is a global defect in FA-C cells. We measured phosphorylation levels of STAT3 in response to other cytokines including IL-6 and LIF using immunoblot analysis and an Ab specific for the phosphorylated form of STAT3. We found no differential phosphorylation of STAT3 between the FA-C cells, HSC536 and HSC536/neo, and the corrected isogenic cell line, HSC536/FANCC, in response to IL-6 (Fig. 2,A). Moreover, phosphorylation of LIF-induced STAT3 was equivalent between SV40-transformed murine embryonic cell lines derived from a Fancc^−/− mouse (MEF61) and a WT mouse (MEF11.1; Fig. 3 B). These studies indicate that abnormal STAT phosphorylation in FA-C cells is not a global defect and is unique to specific extracellular signals.

EBV-transformed B lymphoblasts were treated with 1000 U/ml IFN-α for 15 min. Subsequently, WCE were made and cells were immunoprecipitated with an Ab specific for phosphorylated tyrosines. These extracts were run on SDS-PAGE, electroblotted, and probed with an Ab specific for STAT2. These studies consistently revealed no differential IFN-α-induced phosphorylation between the FANCC mutant cells HSC536N and HSC536/neo and the corrected cell HSC536N/FANCC (Fig. 3). However, we consistently found low-level constitutive activation of STAT2 in mutant cells compared with normal cells (Fig. 3). Therefore, STAT2 is constitutively active in FA-C cells at low levels; however, this does not result in abnormal amounts of inducible activated STAT2 in FA-C cells.

One explanation for the ligand-specific defects of STAT molecules in FA-C cells is that FANCC might preferentially influence upstream intracellular signaling molecules with specificity for the same ligands. IFN-αβ activates STAT molecules, in part, through the Janus kinases, Tyk2 and Jak1 (32). Using immunoblot analysis and an Ab specific for tyrosine-phosphorylated Tyk2, we determined that phosphorylation of Tyk2 in response to IFN-α (Fig. 4,A, upper panel) and IFN-β (data not shown) is consistently reduced in FA-C mutant B lymphoblasts (HSC536 and HSC536/neo) compared with normal cells (HSC536/FANCC), revealing for the first time a defect in activation of a Jak/STAT pathway upstream of STAT proteins in FA-C cells. Levels of total cellular Tyk2 were equivalent between mutant and corrected cells (Fig. 4,A, lower panel). Reduced type I IFN-induced Tyk2 phosphorylation was confirmed in a MEF cell line derived from a Fancc^−/− mouse compared with a WT cell line and in another FA-C EBV-transformed B lymphoblast cell line, PD149, compared with FANCC-corrected PD149 cells (data not shown). In addition to reduced Tyk2 phosphorylation, IFN-α-induced phosphorylation of Jak1 is also reduced in response to IFN-α, but not IFN-γ in FA-C cells (Fig. 4 B), once again demonstrating ligand-specific activation defects of the same molecule.

We next asked whether IFN-αR1 is expressed properly in FA-C cells. We found no difference in expression of IFN-αR1 in FANCC mutant and normal cells (Fig. 5,A). We next asked whether FANCC is required for Tyk2 to associate with the IFN-αR1. Cells were treated with nothing or 1000 U/ml IFN-α for 10 min. WCE were made and used for immunoprecipitation with an Ab specific for IFN-αR1. Immunoprecipitates were electrophoresed on SDS-PAGE transferred to nitrocellulose and probed using an Ab specific for Tyk2. Tyk2 associates with IFN-αR1 constitutively, an association unaffected by IFN-α treatment in all cells tested, including the FANCC mutant B lymphoblasts, HSC536N (Fig. 5,B, upper panel). When the same blot is stripped and reprobed with STAT2, equivalent inducible association of STAT2 with IFN-αR1 is detected between mutant and corrected cells (Fig. 5 B, lower panel). These studies suggest the altered Tyk2 phosphorylation is not due to altered binding to the receptor.

Because Tyk2 has an important role in differentiation and maintenance of Th subsets (28, 32, 33), we isolated CD4⁺ cells from Fancc^−/− mice and quantified cytokine production. Splenocytes were removed from Fancc^+/− or Fancc^−/− mice. CD4⁺ cells were isolated and placed in anti-CD3-coated plates for 20 h. ELISA demonstrated that IFN-γ production is reduced from Fancc^−/− CD4⁺ cells compared with Fancc^+/− cells (Fig. 6,A), suggesting that Fancc^−/− mice have abnormal levels of Th type I subsets. Overall CD4⁺ cell numbers were roughly equivalent (data not shown). Moreover, IFN-γ-secreting CD4⁺ cells were measured using flow cytometric analysis and the Miltenyi Biotec murine IFN-γ detection assay. Fewer CD4⁺ IFN-γ-secreting cells were detected from the Fancc^−/− splenocytes compared with Fancc^+/− splenocytes (percentage of CD4⁺ cells that are IFN-γ secreting) (Fig. 6 B). These studies indicate that Fancc^−/− spleens have reduced levels of differentiated Th subsets.

Fancc^−/− CD4⁺ splenocytes produce reduced amounts of the classic Th1 cytokines, IFN-γ (Fig. 6). We next asked whether Fancc^−/− CD4⁺ can be differentiated into polarized subsets in vitro. Fancc^+/− and Fancc^−/− mice were euthanized by cervical dislocation, spleens were removed, and a single cell suspension was prepared. Any animal displaying signs of infection or an abnormal spleen size was removed from the study. Splenocytes were activated with murine plate-bound anti-CD3 (BD Bioscences, San Jose, CA) and placed under conditions conducive for Th1 or Th2 differentiation (see Materials and Methods). Cells were then washed and incubated with plate-bound anti-CD3 in the absence of any additional cytokines for 20 h before harvest. Supernatants were removed from cells, and ELISAs were used to determine levels of IL-4 and IFN-γ production. Analysis of IFN-γ production in supernatants by ELISA consistently shows that Fancc^−/− splenocytes placed under Th1 conditions produce less IFN-γ than Fancc^+/− splenocytes. Moreover, unlike Fancc^+/− splenocytes that produce significantly higher levels of IFN-γ when placed under Th1 conditions compared with Th2 conditions, no difference in IFN-γ was detected in Fancc^−/− splenocytes placed under Th1 or Th2 conditions (Fig. 7,A), suggesting that Fancc^−/− splenocytes were unable to polarize into the Th1 subset. No consistent differences in IL-4 production were detected in cells placed under our Th2-polarizing conditions (Fig. 7 B). Taken together, these data indicate that CD4⁺ cells from Fancc^−/− mice fail to properly differentiate into Th1 subsets.

Type I (IFN-α₁, -α₂, -ω, -τ, and -β) and type II (IFN-γ) IFNs are involved in a broad range of mammalian functions, including antiviral responses, inflammation, hemopoiesis, cell proliferation, and differentiation (34, 35). Although they are structurally unrelated and signal through different receptors, type I and II IFNs share many overlapping biological functions (36). Recent findings that type I and II IFNs affect signaling of one another (27, 37, 38, 39) have clarified the mechanism of overlap and suggested to us that FANCC, known to be required for optimal IFN-γ signaling, might influence signaling pathways ignited by the type I receptor.

FANCC associates with STAT1 following IFN-γ stimulation and facilitates docking of STAT1 on IFN-γR1 (25). Therefore, loss of a functional FANCC results in reduced tyrosine phosphorylation of STAT1 (25). Given that proper IFN-γ signaling requires an intact type I IFN receptor (27), we tested proximal elements of type I IFN signaling in FA-C cells. Phosphorylation of STAT1, STAT3, and STAT5 is significantly reduced in response to type I IFNs (Fig. 1). However, defective STAT phosphorylation is not global because, although STAT3 phosphorylation is decreased in FA-C cells in response to IFN-α, IFN-β, and IFN-γ (Fig. 1), phosphorylation of STAT3 occurs normally in response to LIF and IL-6 (Fig. 2). Focusing on the defective activation of STATs 1, 3, and 5, we reasoned that ligand specificity might result from differential influences of specific ligands on specific Jak molecules. Therefore, we quantified activation of Jak1 and Tyk2 in response to type I IFNs. These studies revealed the first activation defects in Jak molecules, finding that both Jak1 and Tyk2 tyrosine phosphorylation is reduced in FANCC mutant cells (Fig. 4). Previous studies demonstrated and we once again confirm that activation of Jak1 in response to IFN-γ occurs normally (25) (Fig. 5 B). Therefore, like STATs, defective activation of Jak1 in FA-C cells is ligand dependent. The precise mechanism for impaired type I IFN signaling in FA-C cells is unclear. Tyk2 may lie upstream in the process, given that cross talk occurs between type I and II IFN signaling pathways and IFN-γ signaling requires an intact IFN-α receptor (35). Moreover, like FA-C cells, cells derived from mice nullizygous at the Tyk2 locus (Tyk2^−/−) have reduced (but not absent) responses to type I and II IFNs, as demonstrated by reduced phosphorylation of STAT proteins (32). However (28), like FA-C cells, responses to IL-6 and LIF, two cytokines previously implicated in Tyk2 signaling by in vitro studies, were unaffected in Tyk2^−/− mouse cells. Given that STAT1 and FANCC coimmunoprecipitate in response to IFN-γ, one provocative model is that a functional interaction between FANCC and Tyk2 is required in type I IFN signaling. However, we were unable to detect a FANCC:Tyk2 interaction (data not shown). Therefore, FANCC most likely acts indirectly through one of the many type I IFN regulatory pathways. In light of the cochaperone-like role played by FANCC in modulating protein kinase R activation in an heat shock protein 70-dependent fashion (40, 41), it may be the case that FANCC serves a chaperone-like function with STAT molecules, permitting proper association of specific molecules to discrete receptors. Clarifying the mechanism of abnormal type I IFN signaling in FA-C cells will be the target of future studies.

Because of the Jak/STAT signaling similarities between Fancc^−/− and Tyk^−/− cells, we reasoned that other phenotypic abnormalities may be overlapping. Splenic Th1 differentiation is impaired in Tyk2^−/− mice (28, 32). Moreover, certain STAT proteins are involved in differentiation and maintenance of Th cell subsets (28, 32, 42), and two FANCC-interacting proteins, heat shock protein 70 and repressor of GATA (the murine form of Fanconi anemia zinc finger), are also important in Th1 differentiation and control of Th cytokine expression, respectively (41, 43, 44, 45). Therefore, we quantified cytokine secretion in CD4⁺ T cells from Fancc^−/− mice. Fancc^−/− mice have reduced numbers of CD4⁺ IFN-γ-secreting cells and produce less IFN-γ. Because the numbers of CD4⁺ cells in Fancc^−/− and Fancc^+/− spleens are equivalent (data not shown), these studies suggest that differentiation of CD4⁺ cell subsets is impaired in Fancc^−/− mice. Supporting this, IFN-γ secretion from in vitro differentiation of Fancc^−/− splenocytes into Th1 cells is reduced (Fig. 7). We emphasize that these studies were conducted in murine cells, and that significant differences exist between mice and humans in differentiation of Th1 cells (46). In humans, type I IFNs can also induce Th1 differentiation in addition to the classical IL-12 induction pathway. Fancc^−/− mice do not mimic the severity of the FA phenotype. These mice show no overt physical anomalies, do not exhibit overt spontaneous bone marrow failure, and do not exhibit an increased risk of cancer. The Th1 defect in FA may be less pronounced in mice than in patients, given that IFN-α signaling does not occur properly in human lymphocytes (Figs. 1 and 4). In fact, preliminary evidence supporting this notion finds that FA patients have reduced CD4⁺ cell populations (12). Finally, Th cell differentiation requires chromatin remodeling for proper differentiation (47). Given the presumed role of the FA genes in transcription-coupled DNA repair (17, 24, 48), it is entirely possible that chromatin at the cytokine locus is not remodeled properly in FA cells, contributing to impaired Th differentiation, linking defective cytokine signaling and DNA repair/chromatin rearrangement in FA cells (49).

A proper balance between Th1 and Th2 is required for hemopoietic progenitor cell homeostasis and tumor surveillance (50, 51, 52, 53, 54, 55, 56, 57). The Th1:Th2 ratio is particularly important in the progression and outcome of neoplastic disease (58, 59, 60), and cytokines secreted by CD4⁺ cells are known to have potent antitumor and antimetastatic activities (61, 62, 63, 64). Defective Th development in FA cells may not only affect immune function, but may also be reflected in the increased risk of cancer in the FA population, particularly in solid tumors. FA patients are at a significantly increased risk of squamous carcinoma of the head and neck, and esophageal, liver, and gynecological cancers compared with the general population (2, 4, 65, 66, 67), cancers commonly associated with oncogenic strains of the HPV (68, 69, 70, 71). In fact, there are emerging reports that squamous cell cancers in FA patients are HPV positive (4, 5, 6). Given that Th1 responses are involved in protecting cells against HPV and in eradicating HPV-positive tumors (72, 73), we speculate that FA Th defects may contribute to these cancers. Evidence that lymphocytes and STAT1/IFN-γ signaling both contribute to the immunogenic phenotype of tumors (74) supports this hypothesis, emphasizing the need for studies designed to define the role of immune surveillance/cancer immunoediting in the context of FA malignancies. Finally, the discovery of defective lymphocyte subpopulations in the Fancc^−/− mouse also underscores the importance of focusing on both lymphoid and myeloid lineages in any study seeking to clarify bone marrow failure, carcinogenesis, and leukemogenesis in FA patients.

We thank Michael Forte and Qishen Pang for advice and support, and Hanqian Carlson, Winifred Keeble, and Dylan Zodrow for technical support.