Fanconi anemia with sun-sensitivity caused by a Xeroderma pigmentosum-associated missense mutation in XPF

The diagnosis of FA in a 51-year-old woman (at the time of this report, patient 3104) was made when she presented at the age of seven with thrombocytopenia and bilateral radial ray abnormalities including right thumb absence and severe left hypoplasia (Fig. 1a), and increased sensitivity to DNA interstrand-crosslinking agents of her cells (data no longer available). Other features of FA included mild facial anomalies (Fig. 1a−c), short stature, microcephaly and skin pigmentation changes. The patient is the younger of two sisters of an unrelated White British couple, both over 70 and well, with an unremarkable past medical and family history. The patient is married and worked for a long time in administrative and secretarial positions. During her adult life borderline thrombocytopenia and macrocytosis persisted. Further details with respect to clinical manifestations and management are provided in Table 1. Since she was referred to our service in her fifth decade, her hypo-cellular bone marrow remained without features of dysplasia or clonal cytogenetic aberrations (current platelet count 117 × 10⁹/L, MCV 109 fl). Notably, she reported to burn very easily in the sun, and avoided sun exposure all her life. She developed so far neither bone marrow failure nor malignancy.

Cyclosporine A was used to establish Epstein-Barr virus (EBV)-transformed lymphoblastoid cell lines (LCL) [11]. Peripheral blood lymphoblasts were grown in RPMI 1640 medium supplemented with GlutaMAX (Gibco) and 15% fetal bovine serum (FBS). Primary fibroblasts of the patient concerned were cultured in Amniopan medium (PAN Biotech), while SV40 large T antigen-immortalized fibroblasts were maintained in MEM with GlutaMAX and 10% FBS. For retrovirally transduced fibroblasts 10% Tet System Approved FBS (Clontech Laboratories) was used. All cultures were maintained in incubators with 5% CO₂. Cells were exposed to 40 ng/μl mitomycin C (MMC) for 16 h prior to immunofluorescence analysis of RAD51 foci, immunoblotting or cell fractionation studies. Cell lines from disease and normal controls were maintained as previously reported [9].

Cytogenetic assays were performed on whole blood cultures to assess spontaneous and MMC-induced breakage rates. MMC at indicated final concentrations was added to blood samples as a G₀ pulse before culture initiation. After 1 h the cells were washed and transferred to fresh complete RPMI 1640 medium. Lymphocytes were stimulated with phytohemagglutinin (PHA). The cultures were incubated at 37 °C and harvested after 72 h. Chromosome preparations were made by the air drying method [12]. Solid (Giemsa)-stained metaphases were examined for chromatid and chromosome type damage, and the results were compared to age and sex-matched normal controls.

PHA-stimulated lymphocytes or primary fibroblasts were cultured untreated for 72 h or continuously exposed to MMC. Mono- or bivariate (BrdU-Hoechst 33,258/Ethidium bromide) cell cycle analysis was performed on a triple-laser-equipped flow cytometer (LSRII, BD Biosciences). Data were analyzed using the MPLUS AV software package (Phoenix Flow Systems) [13].

A modified salting-out technique [14] or the GeneJET Genomic DNA Purification Kit (Thermo Fisher) were used for isolation of genomic DNA (gDNA). For total RNA isolation we employed the High Pure RNA Isolation Kit (Roche) while transcription into cDNA was performed by SuperScript Reverse Transcriptase (Invitrogen).

For relative quantification of mRNA expression levels, allele-specific primers were designed (WT for: 5’-GACGCAGAGCTAACCTTTGTTC-3′; c.1765C > T MUT for: 5′- GACGCAGAGCTAACCTTTGTTT-3′; WT/MUT rev: 5′- GTTCCTCAGTTGAACCTCCGTA-3′). For the detection of exon 5 skipping due to c.793-2A > G primer sequences were as follows: WT for: 5′- TCTGGAATCTCTGAGAGCAACG-3′, WT rev: 5′- AACATCGAGGTGCTGGAGTC-3′, MUT for: 5′- ATAACCCATCGCTTGAAGTGGA-3′, MUT rev: 5′- CAAGAA\ACAGCCAACCTTGTCA-3′. The PCR reaction was performed with HOT FIREPol® EvaGreen® qPCR Mix Plus (Solis BioDyne) on an ABI ViiaTM 7 System (Applied BiosystemsTM). Evaluation of melting curves and amplification plots, and relative quantification (RQ) was done with the QuantStudioTM Real-Time PCR Software v1.2.4 (Applied BiosystemsTM) using the ΔΔCt method. Each sample was analyzed in technical triplets.

Transduction with different retroviral vectors containing XPF/ERCC4/FANCQ cDNA constructs or mock was performed according to standard protocols [15, 16]. Transduced immortalized fibroblasts were analyzed for their sensitivity. Aliquots of 20,000 cells per well were seeded in 6-well culture plates and grown at the indicated concentrations of MMC. After eight days cell viability was determined by image cytometry using VitaBright-48 and propidium iodide staining on a NucleoCounter NC-250 instrument (ChemoMetec A/S).

LCLs were grown in T25 cell culture flasks at concentrations of 0–1000 nM MMC for eight days. Cell viability was analyzed by propidium iodide exclusion and assayed by flow cytometry.

Nuclear RAD51 focus formation was analyzed in fibroblasts grown on glass chamber slides (Sarstedt). Cells were washed with PBS and subsequently fixed in 4% (vol/vol) paraformaldehyde in PBS for 15 min at room temperature. Thereafter they were exposed to ice-cold 100% methanol and kept on ice for 30 min. Blocking of non-specific antibody binding sites was accomplished with 20% (vol/vol) FBS in PBS for 30 min at room temperature. Rabbit anti-RAD51 (1:800; ab63801, Abcam) served as primary antibody, to be detected by Alexa 594-conjugated anti-rabbit secondary antibody (1:2000; A11037, Molecular Probes/Life Technologies). ProLong Gold Antifade Mountant with DAPI (Thermo Fisher) was used for counter-staining and as mounting medium. Foci-positive cells (>5 foci/nucleus) were scored by fluorescence microscopy (Axiovert 40C, Zeiss).

Aliquots of 40 μg whole protein extracts from cultured cells were loaded on NuPAGE Novex 7% Tris-Acetate protein gels (Invitrogen). Electrophoresis was carried out over night at 70 V with constant cooling. Proteins were transferred using a dry blotting system (iBlot 2, Life Technologies). FANCD2 isoform immunodetection was carried out using primary anti-FANCD2 antibody (1:500; sc-20,022, Santa Cruz) and secondary anti-mouse IgG-HRP (sc-2005, Santa Cruz) diluted 1:3000. FANCJ (1:1000; NB100–416, Novus) or vinculin (1:10,000; ab129002, Abcam) served as loading controls. For visualization of ERCC4/XPF/FANCQ whole protein extracts were separated on NuPAGE 4–12% Bis-Tris gels (Invitrogen). Anti-XPF (1:200; ab17798, Abcam) was used for immunodetection while Anti-Tubulin (1:1000; ab44928, Abcam) was employed as loading control. Luminescence was generated by chemical reaction with a standard ECL reagent (Millipore Corporation).

The NE-PER Nuclear and Cytoplasmic Extraction Kit (Thermo Fisher) was used for extraction of cytoplasmic proteins while nuclear and chromatin-associated proteins were isolated by the Subcellular Protein Fractionation Kit (Thermo Fisher). SDS-PAGE was run on NuPAGE 4–12% Bis-Tris gels (Invitrogen). For XPF, YY1, tubulin and histone H3 immunodetection we employed Anti-XPF (1:200; ab17798, Abcam), Anti-Tubulin (1:1000; ab44928, Abcam), Anti-YY1 (1:5000; ab199998, Abcam) and Anti-Histone H3 (1:800; ab1791, Abcam) antibodies. Secondary antibodies were as above.

Cell pellets were resuspended in 200 μl lysis buffer (20 mM Tris-HCl pH 7.4, 10 mM KCl, 140 mM NaCl, 1 mM EDTA, 0.2 mM EGTA, 0,5% Triton X-100, 5% Glycerol, 0.1 mM PMSF, 0.5 mM DTT, Halt Protease/Phosphatase Inhibitor Cocktail 1:50) and lysed by sonication. After centrifugation (14,000×g for 10 min at 4 °C) and lysate pre-clearing, 500 μl cell lysate with 1 mg/ml cell total protein was incubated with the primary antibody (ERCC1: ab2356, Abcam, 1:150; XPF: ab76948, Abcam, 4 μg/mg total protein in the lysate) at 4 °C overnight. 50 μl of pre-washed Protein G Dynabeads (Thermo Scientific) was added. After incubation for 2 h at 4 °C, the beads were washed five times using 500 μl washing buffer (20 mM Tris-HCl pH 7.4, 10 mM KCl, 140 mM NaCl, 1 mM EDTA, 0.2 mM EGTA, 0.1% Triton X-100, 5% Glycerol, 0.1 mM PMSF, 0.5 mM DTT, Halt Protease/Phosphatase Inhibitor Cocktail 1:100). Beads were taken up in 20 μl elution buffer, 4× LDS sample buffer and 10× reducing agent. Following incubation for 10 min at 70 °C beads were collected and the supernatant was loaded on a 4–12% Bis-Tris gel. Immunodetection was carried out using anti-XPF (1:200; ab17798, Abcam) or anti-ERCC1 (1:100; sc-17,809, Santa Cruz) primary antibodies. YY1 (1:5000; ab199998, Abcam) served as loading control.

Aliquots of 5000 fibroblasts per well were seeded in 6-well plates. After two days cells were washed with PBS and irradiated at the indicated doses in quadruplicate (0 J/m²) or triplicate (others) using an UV-C germicidal lamp (254 nm; Philips). An UVX Digital Radiometer (Serial No. E27846, UVP) was used to quantify the exact UV dose. Control cells were UV irradiated simultaneously to serve as an internal control. After another five days, before the non-irradiated cultures reached confluency, cells were pulse-labeled with [methyl-³H]-thymidine (40–60 Ci/mmol; 5 μCi/ml; Amersham Biosciences)-containing medium for 3 h, washed with PBS and chased for 15 min in medium without ³H–thymidine. Finally cells were lysed in 0.25 M NaOH. Each lysate was counted with 7.5 ml Hionic Fluor scintillation fluid in a liquid scintillation counter (Packard) for 10 min. The results obtained from irradiated plates were expressed as percentages of non-irradiated plates (set 100%) and plotted [17].

Unscheduled DNA synthesis (UDS) and recovery of RNA synthesis (RRS) experiments after UV-C irradiation were done with mixed cell populations [7]. C5RO wildtype (WT) fibroblasts were incubated and preloaded with polystyrene beads of 2 μm diameter for three days, then mixed with the patient (3104 or 1333) fibroblasts and seeded on coverslips. Two days later adherent cells were washed with PBS and UV-C irradiated at 16 J/m². Thereafter they were incubated for 3 h in medium containing 0.1 μM 5-ethynyl-2′-deoxyuridine (EdU; Invitrogen). Afterwards they were washed with PBS, incubated with medium without EdU for 15 min, washed with PBS, fixed with 3.7% formaldehyde in PBS containing 0.5% Triton for 15 min and washed two times with PBS again. The cells were incubated for 30 min with fluorescent dye-coupling buffer containing 10 mM CuSO₄ and Alexa Fluor 594 azide (Qlick-iTTM; Invitrogen). After washing with PBS, cells were mounted in Vectashield Antifade Mounting Medium with DAPI (Vector Laboratories). For RRS studies cells were treated identically as for UDS experiments, but 16 h after irradiation incubated with culture medium containing 0.1 μM 5-ethynyl-uridine (EU) for 2 h.

Visual light distinguished WT cells that contained beads in their cytoplasm from patient cells that did not. On micrographs DAPI is the blue signal that stains nuclei. Red is the UDS (or RRS) signal; it is much brighter in WT than in patient cells. UDS and RRS levels are expressed as the average fluorescence intensity in the nucleus of patient vs. WT cells (set at 100%).

An unpaired two-tailed Student’s t test was used to compare the results of qPCR analyses. A p value less than 0.05 was considered significant; *p < 0.05, **p < 0.01, ***p < 0.001.

Chromosome breakage studies from peripheral lymphocyte culture of the patient confirmed FA at ages 7, 33 and 49, and demonstrated increased levels of spontaneous damage, and in response to MMC (Additional file 1: Table S1, results of the analysis carried out age 7 are no longer available). In primary fibroblast cultures a spontaneous and MMC-induced G2-phase arrest was observed (Additional file 1: Figure S1A). Also peripheral blood lymphocyte cultures showed G2-phase accumulation within the FA-range, excluding mosaicism in the hematopoietic system (Additional file 1: Figure S1B). Functional analysis of FANCD2 ubiquitination in 3104 fibroblasts demonstrated both FANCD2 isoforms (Additional file 1: Figure S1C), indicating a mutation in an FA gene downstream of those encoding the FA core- and FANCD2/FANCI-complex. RAD51 foci developed normally after exposure to MMC in both 3104 fibroblasts and lymphoblasts (Additional file 1: Figure S1D), excluding the post-FANCD2 groups FA-D1, -N and -O, −R, –S, and -U. WES identified two mutations in the XPF/ERCC4/FANCQ gene, which were confirmed by Sanger sequencing. A single nucleotide substitution located two base pair upstream of exon 5 (c.793-2A > G) was detected on the maternal allele, affecting a canonical splice acceptor (Fig. 1d). cDNA sequencing verified aberrant splicing. A major splice product revealed exon 5 skipping (Fig. 1e). This event is predicted to result in premature termination of translation (p.Thr265Valfs*13). The second change in the XPF/ERCC4/FANCQ sequence is the missense mutation c.1765C > T in exon 8 (Fig. 1f) on the paternal allele, which is listed on the ExAC-Browser with a minor allele frequency of 0.0066%. It substitutes a highly conserved amino acid residue (p.Arg589Trp) in the SF2 helicase-like domain. This mutation has previously been reported in two XP-F patients (XP24BR and XP32BR) [18], one patient with XP with neurodegeneration (AS871) [5], and an individual showing an intermediate XP/CS phenotype and features of FA (XPCS1CD) [10]. Expression of both mutations in mRNA was re-confirmed by qPCR (Additional file 1: Figures S1E and F). Complementation of 3104 fibroblasts by wildtype, but not Arg589Trp-mutant XPF/ERCC4/FANCQ rescued MMC resistance, providing evidence that the mutations of XPF/ERCC4/FANCQ cause the cellular FA phenotype (Fig. 1g). Premature termination of translation due to exon 5 skipping would result in a truncated protein of 31.4 kDa. However, on XPF/ERCC4/FANCQ immunoblot analysis of 3104 fibroblasts only one low signal intensity band of approximately 100 kDa was detected (Fig. 1h), suggesting the presence of residual mutant XPF/ERCC4/FANCQ protein of normal size but reduced abundance. Dilution studies demonstrated that XPF/ERCC4/FANCQ residual protein with the Arg589Trp mutation is present at approximately a 1:15th the level of wildtype XPF/ERCC4/FANCQ protein (Additional file 1: Figure S1G). However, it is difficult to tell the individual contributions of mutant transcript and protein instability to this reduction. Fractionation studies of 3104 fibroblasts showed that residual full-length XPF/ERCC4/FANCQ protein was detectable in the nucleus (Fig. 1i), while in cells from XP-F or XFE patients the mutant XPF/ERCC4/FANCQ appears to be mislocated and not to be part of the XPF-ERCC1 complex, presumably as a consequence of XPF protein misfolding [19].

In addition to ICL repair, NER activity was also impaired in 3104 fibroblasts. Cell survival after UV-C irradiation was reduced (Fig. 2a) (LC₅₀ = 2.8 J/m²), as it was the case in fibroblasts from an XP-F patient with known UV sensitivity and a mild clinical phenotype (LC₅₀ = 2.2 J/m²) [20], and FA-Q 1333 fibroblasts as previously reported [9]. Fibroblasts from the unique XFE patient [8] showed much higher sensitivity (LC₅₀ = 1.3 J/m²). Both global genome NER (GG-NER) and transcription-coupled NER (TC-NER) were affected in cells from patient 3104. Unscheduled DNA synthesis (UDS) and recovery of RNA synthesis (RRS) in 3104 fibroblasts after UV irradiation were more than 80% decreased compared to controls (Fig. 2b-d), to similar levels as for FA1333 fibroblasts was previously reported [9] and is shown here again (Fig. 2c and d). Fibroblasts from an XP-F patient with mild clinical UV-light sensitivity (XP42RO) [7] also exhibited comparable reduction of UDS and RRS rates, whereas fibroblasts from the XPE patient (XP51RO) had even lower activity (Fig. 2c and d). In contrast, 3104 cells were more MMC-sensitive than 1333 cells (Fig. 2e).