FANCD2 functions as a critical factor downstream of MiTF to maintain the proliferation and survival of melanoma cells

Proteins involved in genetic stability maintenance and safeguarding DNA replication act not only against cancer initiation but could also play a major role in sustaining cancer progression. Here, we report that the FANC pathway is highly expressed in metastatic melanoma harboring the oncogenic microphthalmia-associated transcription factor (MiTF). We show that MiTF downregulation in melanoma cells lowers the expression of several FANC genes and proteins. Moreover, we observe that, similarly to the consequence of MiTF downregulation, FANC pathway silencing alters proliferation, migration and senescence of human melanoma cells. We demonstrate that the FANC pathway acts downstream MiTF and establish the existence of an epistatic relationship between MiTF and the FANC pathway. Our findings point to a central role of the FANC pathway in cellular and chromosomal resistance to both DNA damage and targeted therapies in melanoma cells. Thus, the FANC pathway is a promising new therapeutic target in melanoma treatment.

In spite of recent clinical breakthroughs, cutaneous metastatic melanoma remains the most deadly form of skin cancer 1 . Extensive analysis previously reported that the mRNAs coding proteins involved in double-strand break repair and in the rescue of stalled DNA replication forks, including several members of the FANC pathway, are over-expressed in metastatic melanomas and are associated with poor patient survival 2,3 . Indeed, proteins and pathways involved in DNA damage response (DDR) act not only as a tumour barrier but can also be required for tumour cell survival, proliferation and invasion 4,5 . Thus, DDR proteins represent a promising pharmacological target for improving therapeutic responses in melanoma.
Both the absence of melanoma in FA patients and the observed over-expression of mRNAs encoding FANC proteins in metastatic melanoma implicitly suggest that the FANC pathway may be required to support melanoma development. Thus, we decided to decipher if and how the FANC pathway participates in melanomagenesis. We show here that a fully proficient FANC pathway is a key determinant for melanomagenesis ensuring the high proliferation and migration capabilities that characterize melanomas overexpressing the transcription factor microphthalmia (MiTF), which is amplified and/or mutated in a subset of patients with malignant melanoma where it acts as an oncogene [12][13][14][15] .

Results
FANC pathway silencing alters proliferation, migration, senescence and intracellular redox status of human melanoma cells. To comprehensively evaluate the role of the FANC pathway in melanoma behaviour, we silenced in several human melanoma cells either FANCA, the most frequently mutated FANC gene in FA syndrome, or FANCD2, the monoubiquitination and subnuclear relocalisation of which constitute a readout for the activation of the pathway. FANCA or FANCD2 silencing induced a significant decrease in the number of proliferating cells (Fig. 1a) and in their colony formation ability ( Supplementary Fig. 1a). Moreover, FANCA or FANCD2 knockdown dramatically reduced the migration ability of the melanoma cells (Fig. 1b) and suppressed their capacity to sustain anchorage-independent cell growth ( Supplementary Fig. 1b), which are indicators of tumorigenicity and invasiveness.
FANC proteins silencing resulted in increased cell size and granularity ( Supplementary Fig. 1c,d), which are associated with cellular senescence. Specifically, by measuring senescence-associated β -galactosidase activity (SA-β Gal) via fluorescence-activated cell sorting (FACS) analysis or blinded individual microscopic scoring, we consistently demonstrated that FANCD2 silencing specifically induced a significant increase in SA-β Gal activity, in the expression of p53, p21 and p27, as well as intracellular reactive oxygen species (ROS) accumulation, which are known inducers and/or markers of the senescence process (Fig. 1c,d,e,f and Supplementary Fig. 1e,f,g) [16][17][18] . To rule out non-specific effect we assessed the effect of a second siRNA to FANCA (FANCA#2) and FANCD2 (FANCD2#2). As previously observed, these siRNA, which efficiency is shown (Supplementary Figure 1h), reduced cell proliferation as illustrated by a reduced number of cells (Supplementary Figure 1i) and increased expression of cell cycle inhibitors and promoted senescence phenotypes such as increase in SA-β Gal activity and increase in 53BP1 foci ( Supplementary Fig. 1j,k). Thus, FANC pathway silencing in melanoma cells resulted in growth arrest, migration defect and senescence, which are phenotypes that we previously observed following the silencing of the oncogenic microphthalmia-associated transcription factor (MiTF) 19 . We therefore hypothesised a potential interplay between the MiTF and the FANC pathway. Accordingly, the simultaneous silencing of FANCA or FANCD2 with MiTF resulted in the same effect following the silencing of MiTF alone in terms of ROS production and senescence (Fig. 1e,f). FANC pathway and/or MiTF silencing generate genetic instability and hypersensitivity to MMC-induced DNA damage. Similar to the consequences of FANC pathway loss-of-function, MiTF-silenced cells showed sustained recruitment of γ H2AX and 53BP1 foci, increased frequency of chromosome aberrations, anaphase abnormalities (i.e., anaphase bridges and chromosome loss) and post-mitotic micronuclei, indicating that MiTF-depletion results in DNA breakage as possible consequence of an altered rescue of stalled replication forks ( Fig. 2a,b,c,d and Supplementary Fig. 2a,b). Specifically, less than 10% of exponentially growing melanoma cells displayed more than five 53BP1 foci per nucleus, a percentage that exceeded 60% when MiTF, FANCA or FANCD2 were silenced individually ( Fig. 2b and Supplementary Fig. 2). Moreover, MiTF-depleted cells presented a cellular and chromosomal hypersensitivity to ICL-inducing agents similar to that reported in FANCA-or FANCD2-depleted melanoma cells (Fig. 2e,f, Supplementary Fig. 2c,d,e and data not shown). The silencing of both MiTF and FANCA or FANCD2 did not show additive or synergistic effects in terms of micronuclei induction or cellular sensitivity to ICLs (Fig. 2d,e). A more elevated level of cell death, a more pronounced growth arrest or both could be responsible for the slightly reduced level of micronucleated cells observed 96 hours post-transfection in the MITF-FANCA double-depleted cells compared to the MITF-depleted cells.
Finally, by flow cytometry we quantified cell cycle distribution in BrdU and/or propidium iodide stained cell populations ( Supplementary Fig. 2f,g), demonstrating that, in melanoma cells, FANC and MiTF depletion with specific siRNAs triggers a similar G0/G1 phase accumulation together with a corresponding reduction in the frequency of S-phase cells, as evaluated 72 hours post-siRNA transfection.
Collectively, our data support the hypothesis that MiTF and the FANC proteins operate within one pathway to ensure DNA integrity, high proliferation and survival in melanoma.

MiTF is involved in FANC genes expression.
In line with a crosstalk between MiTF and the FANC pathway, MiTF silencing in melanoma cells triggered a strong reduction in the mRNA level of 4 analysed FANC genes: FANCA, FANCD2, FANCC and FANCG ( Fig. 3a and Supplementary Fig. 3a). Using recently described MiTF chromatin immunoprecipitation sequencing (ChIP-seq) data set, we investigated the MiTF binding profile in the genomic region of FANCD2 and FANCA ( Supplementary Fig. 3b). We identified several significantly enriched MiTF-binding sites also containing E-box motifs, a known binding motif of MiTF 20,21 . Moreover, increased transcription correlates with the presence of H3K27ac modification. These data, together with reduced FANCD2 and FANCA mRNA level observed following MiTF-depletion, indicate a direct role for MiTF in the transcriptional activation of FANCD2 and FANCA through binding in enhancer and promoter regions. We have indeed observed a modest activation of the transcriptional activity of the FANCD2 promoter by MiTF, consistently with the fact that MiTF binding sites are mainly localized in intronic regions ( Supplementary Fig 3c). Contrastinlgy, we found no MITF binding in the genomic region of FANCC and FANCG, indicating an indirect regulation of these genes by MITF.
Likewise, immunoblot analysis demonstrated a robust and significant time-dependent decrease in the expression of both FANCA and FANCD2 following MiTF silencing (Fig. 3b,c, and Supplementary Fig. 3d). To rule out non-specific effects a second MiTF siRNA was used and showed similar inhibition of FANCA and FANCD2 expression ( Supplementary Fig. 3e).
MMC exposure induced FANCD2 monoubiquitination and relocalisation to subnuclear foci where it largely colocalised with γ H2AX foci (Fig. 3d,e,f). In MMC-treated MiTF-silenced cells, despite the reduced levels of FANCA and FANCD2, the latter underwent relatively normal monoubiquitination, but its monoubiquitinated form is able to aggregate into subnuclear foci in cyclin A-positive cells (i.e., cells into the S/G2 phase) (Fig. 3d,e,f). Obviously, as might be expected, the quantity of FANCD2 foci per cell is significantly reduced in MiTF-silenced cells (Fig. 3f).
Thus, MiTF silencing appears to alter the expression and capacity of the FANC pathway.
FANC pathway is necessary to maintain genetic stability but dispensable to downregulate ROS levels downstream MiTF depletion. To  FANCD2 in MiTF-depleted cells significantly reduced the occurrence of spontaneous DNA damage, as indicated by γ H2AX and 53BP1 foci formation (Fig. 4a,b) and foster melanoma cell proliferation in response to MMC exposure (Fig. 4c). On the contrary, neither FANCA nor FANCD2 forced expression is able to modify the siMiTF-associated increase in both β -Gal activity (Fig. 4d,e) and ROS intracellular levels, as determined by FACS analysis of DCFDA-stained cells (Fig. 4f). Since FANCD2 forced expression significantly reduced the increase in molecular markers associated to DNA damage response and growth arrest that are strongly expressed in siMiTF-depleted cells (Fig. 4g), we argue that FANCcore complex-mediated FANCD2 monoubiquitination appears to be required for an optimal DNA damage response in order to ensure genomic stability and cell proliferation but it is not sufficient per se for protecting cells against siMiTF-induced increased ROS content. Thus, in MiTF depleted cells, senescence program follows also other pathways than the DNA-damage induced.

FANC pathway participates to the poor outcome and vemurafenib resistance in melanoma.
Collectively, our data suggest that the FANC pathway may participate in cancer progression and resistance to therapeutics and could be responsible for the poor outcome associated with melanoma. We further assessed the importance of a proficient FANC pathway for tumor growth in vivo. We injected subcutaneously nude mice with 1 × 10 6 B16-F10 cells 72 hours after their transfection with siRNA targeting Fancd2 or Mitf, which efficiency is shown by Q-RT-PCR analysis ( Supplementary Fig. 4a). As previously observed for human melanoma cell lines, 3 days after transfection both MiTF and Fancd2 siRNA reduced their corresponding  Fig. 4b). Ten days later, mice were sacrificed, tumor mass isolated and weighted to estimate tumor growth capability. As reported in Fig. 5a, FANCD2 depletion strongly affects tumor development in vivo, clearly supporting the important role of a proficient FANC pathway in tumor progression and growth. Moreover, an analysis of datasets from patients with stage III and IV metastatic melanoma 22 revealed that high FANCD2 expression significantly correlated with poorer prognosis (Fig. 5b, supplementary Fig. 4c). BRAF V600E mutation, which leads to MAPK/ERK signalling pathway activation and hyperproliferation, is the most common causative mutation in melanoma 23 . Vemurafenib, a BRAF V600E inhibitor, has been recently shown to trigger transient melanoma regression; patients rapidly developed resistance and subsequently relapsed 24 . We wondered if FANCD2 could influence vemurafenib response in BRAF V600E melanoma cells. Immunoblot analysis showed that vemurafenib efficiently reduced ERK activation, decreased FANCD2 expression, as well as caused the disappearance of the zymogenic form of caspase 3, which occurred in association with cell growth arrest and death (Fig. 5c). The decrease in FANCD2 following vemurafenib treatment was validated by analysing a publicly available dataset (Supplementary Fig. 4d). More importantly, FANCD2 silencing improved and FANCD2 overexpression abolished the cellular effects of vemurafenib, demonstrating that FANCD2 has a protective function against the pharmacological effect of the vemurafenib (Fig. 5d and Supplementary Fig. 4e,f).
Altogether, our results demonstrate that FANCD2 participates in the cellular resistance of melanoma cells even to non-DNA damage therapeutics. Consequently, since the FANC pathway appears to oppose the efficiency of targeted therapies, specific inhibitors of FANCD2 activity/activation or inducers of its degradation could be useful therapeutic agents for increasing the efficiency of current melanoma treatments.

Discussion
We previously showed through high throuput screening that MITF directly regulates a set of genes involved in DNA replication and repair 25 . However, it is not known whether these genes have a role in melanoma. We report here that the FANC pathway plays an unexpected yet key role in the proliferation and survival of melanoma cells downstream of MiTF (Fig. 5e). We identified MiTF as a critical regulator of the FANC pathway in melanoma and demonstrated that MiTF-silenced cells display the primary characteristics of FA cells, i.e., the cellular and chromosomal hypersensitivity to ICL-inducing agents. Our observations strongly support that thanks to its ability to cope with DNA lesions and replication stress the FANC pathway appears to be important for maintaining the high proliferative potential of melanoma cells as well as their high resistance to therapeutics.
It is noteworthy that silencing of FANCD2 in melanoma cells recapitulates all the key cellular phenotypes of FA cells while FANCA depletion fails to induce a significant increase in the intracellular level of ROS (Fig. 1f). This apparent discrepancy could be due to the fact that FANCA depletion by siRNA is, in general, not so strong as that obtained for FANCD2. Alternatively, the effect of FANCA deficiency in melanoma cells could have a less severe effect in term of senescence and ROS production than FANCD2 depletion. Indeed, all the analyzed markers of senescence (Fig. 1d) are less activated following FANCA depletion than following FANCD2 or MiTF downregulation.
Nevertheless, even if FANCD2 depletion resulted per se in an increasing intracellular level of ROS, MiTF downregulates ROS level in a FANC-pathway independent manner (Fig. 4f). Thus, as depicted in Fig. 5e, FANC pathway, contributes to DNA damage response downstream of MiTF, avoiding the DNA damage-dependent senescence programme to take place, whereas MITF controls both ROS levels as well as DNA damage response, in a FANC pathway-independent and -dependent manner, respectively.
It interesting to note that our data indicate that the FANC pathway influences also melanoma cells migration and invasiveness. Probably, both migration and invasiveness are independent from the DDR activity of the pathway, but, more realistically, linked to the recently reported FANCA loss-of-function-associated defect in CXCR5 neddylation, which decreases the membrane localisation of this receptor involved in cell motility, invasion and (c) Immunoblot of control or FANCD2-suppressed WM9 melanoma cells that were untreated or exposed to vemurafenib (1 μ M for 48 h) with the indicated antibodies. (d) Proliferation of WM9 or FANCD2silenced WM9 melanoma cells that were untreated or exposed to vemurafenib (1 μ M for 48 h). The statistical analysis was performed using Student's T test, ** indicates p < 0.01 and *** indicates p < 0.001. (e) Schematic model summarising the observations made in this manuscript. FANC pathway, contributes to DNA damage response downstream of MiTF, avoiding the DNA damage-dependent senescence programme to take place, whereas MITF controls both ROS levels as well as DNA damage response, in a FANC pathway-independent and -dependent manner, respectively.
Scientific RepoRts | 6:36539 | DOI: 10.1038/srep36539 homing 26 . Clearly, our data indicate that the FANC proteins have a central role in melanoma development, favouring growth, motility, invasiveness and DNA damage resistance of melanoma cells.
Although sustained MITF inhibition triggers the hallmarks of cellular senescence 27 , MiTF expression must transiently decrease to enable melanoma cell migration and invasion but high MiTF level must be recovered to allow the growth of the metastases 13,28 . This oscillatory regulation of MiTF expression is a critical determinant of melanoma metastasis development and could account for melanoma heterogeneity, plasticity and drug resistance. As FANCD2 expression was higher in melanoma cells with a proliferative signature compared to melanoma cells with an invasive signature (http://www.jurmo.ch/hopp/hopp_mpse.php), FANCD2, within the MiTF/FANC axis might follow the same oscillatory model. Hence, the low MiTF phenotype-associated reduction in FANC expression during the metastatic process could contribute to increased genomic instability and the selection of cells that could seed and survive during the migration process. The cells would subsequently need to re-express MiTF and the FANC pathway once they reach their final destination, thereby contributing to the "genetic stability" and proliferation of the cells at the metastatic site.
DNA repair process has been shown to mediate resistance of cells to DNA-damage-induced cell death 29 . Whether they mediate the resistance of metastatic melanoma to chemo-and radio-therapy remained to be demonstrated. Here we show that reduction in FANCD2 improves the effect of vemurafenib. Vemurafenib has been shown to cause a slowdown in replication rate in melanoma cells. FANC pathway is required to cope with replication stressess preventing the accumulation of replication-associated DNA double strands breaks. Since vemurafenib exposure decreases FANC protein expression (Fig. 5c and Supplementary S4d), it is likely that vemurafenib treated cells have a reduced ability to cope with DNA damage and stalled replication forks. It is why the forced expression of FANCD2 allows the cells to better sustain proliferation: thanks to FANCD2 they are able to better manage replication difficulties. Thus, our observations support the idea that inhibition of DNA repair processes may have a therapeutic value in metastatic melanoma. In this context, some pharmacological inhibitors of FANCD2 have been recently identified. Although not really specific for FANCD2, their utilization as therapeutic agents could open new opportunities to increase the efficiency of melanomas treatments.
pCDNA3 (empty vector), pFLAG-FANCD2 and pFLAG-FANCA were a gift of A. Constantinou's lab (Montpellier). Briefly, 1 μ g of plasmid was transfected to the cells using 6 μ L JetPEI (Polyplus) in 150 mM NaCl for the time indicated in the figure legends. The medium was changed 6 h after transfection. For siRNA and plasmid cotransfection, the siRNA were transfected 8 h after seeding; the plasmids were transfected 16 h later.
Treatment. As indicated in the figure legends, cells were treated with mitomycin C (MMC) 48 h after transfection at the concentration of 1 μ g/mL for 1 h or 100 ng/mL for 24 h. After several washes, they were then incubated in fresh medium prior to collection for Western blot or fixation for immunofluorescence.
For micronuclei and anaphase bridges formation, cells were treated with 2 μ g/mL of cytochalasin B (Sigma) with or without 50 ng/mL of MMC for 24 h. RT-qPCR. qPCR was performed as previously reported 30 . The primer sequences for each cDNA were designed using qPrimerDepot (http://primerdepot.nci.nih.gov) and are available upon request. After incubation for 15 min at room temperature, the protein concentrations were determined using the Bradford assay (Bio-Rad), and the samples were combined with 4X Laemmli buffer containing β -mercaptoethanol and denatured via boiling. The proteins (25 μ g) were separated using SDS-PAGE, transferred onto a nitrocellulose membrane, and then exposed to the appropriate antibodies. The proteins were visualised using an enhanced chemiluminescence system. All western blot quantifications were performed using densitometry measures and the ImageJ software. Immunofluorescence. Cells grown on glass coverslips were fixed in 4% formaldehyde supplemented with 0.1% Triton X-100 for 15 min at room temperature prior to permeabilisation in 0.5% Triton X-100 for 5 min. After blocking with 3% BSA in PBS containing 0.05% Tween 20, the cells were stained for 1 h in blocking solution with antibodies. Primary antibody detection was achieved via incubation with anti-rabbit or anti-mouse Alexa Fluor 594-or 488-conjugated secondary antibodies (Invitrogen) for 45 min at room temperature. The slides were mounted in DAKO mounting medium supplemented with DAPI (Sigma) and examined at a magnification of 63× via fluorescence microscopy (Zeiss Axio Observer Z1). The images were captured with an ORCA-ER camera (Hamamatsu). The microscope and camera parameters were adjusted for each series of experiments to avoid signal saturation. Image processing and analysis were performed using ImageJ software.

Chromosomal analysis (conventional karyotyping). 501 mel cells were transfected with control or
MiTF-specific siRNA; 96 h later, these cells were exposed to DEB (0.01 μ g/ml) for 48 h or mitomycin C (0.01 μ g/ml) for 24 h. Briefly, the cells were arrested at metaphase by adding colchicine (1 μ g/ml) to the growing culture for 6 h. The arrested cells were then collected via trypsinization with immediate neutralisation with cell culture medium followed by 5 min of centrifugation at 400 g. We incubated the cell pellet suspension in hypotonic buffer, potassium chloride (5.59 g/l in double distilled water), and sodium citrate (9.0 g/l in double-distilled water) at 1:1 (vol/vol) for 10 min at 37 °C. The cells were then pelleted at 400 g for 5 min to remove the hypotonic buffer and fixed with an ice-cold solution (methanol/acetic acid, 3:1, vol/vol). The fixative was changed once, and then the cells were fixed overnight at 4 °C. To spread the metaphase of the fixed cells onto slides, we dropped the cell suspension from a height onto a chilled, precleaned SuperFrost Plus microscopic slide (Gerhard Menzel, Braunschweig, Germany) that was slightly sloped on a freezer block. Then, the slides were breathed on to enhance spreading and were mounted after drying. The metaphase chromosomes of the metaphase-arrested cells were identified and captured using an automated cytogenetic scanner workstation (MetaSystems) for analysis. Twenty metaphase spreads of distinctly separated chromosomes were analysed for chromosomal abnormality.
Migration assay. Cell migration was assessed using a Boyden chamber with 8-μ m pore filter inserts for 24-well plates (BD Bioscience). Human 501 mel cells melanoma cells were seedd on the upper trans-well chamber; 7% DMEM was placed in the lower chamber. After 24 h, the cells that had adhered to the underside of the filters were fixed with 4% PFA and stained with 0.4% crystal violet. Three random fields were counted at x20 magnification. The results represent the average of samples from three independent experiments. Proliferation assay. Cells were plated at 1 × 10 5 per well of a six-well plate. The cells were transfected 24 h after seeding and grown for 48 h prior to MMC treatment. The cells were incubated for 72 h in medium supplemented with or without MMC and counted with trypan blue using a Neubauer haemocytometer.
Agarose colony-forming assays. The cells (5.10 3 ) were transfected with 50 nM siRNA for 48 h, suspended in 1 ml of 7% FBS DMEM medium containing 0.3% agarose and plated on a 0.6% agarose base in 6-well plates. The cells were then placed in a 37 °C and 5% CO 2 incubator. The colonies of cells allowed to grow for 14 days were stained with 0.04% crystal violet-2% ethanol in PBS for 30 min. The colony formation assay was performed in duplicate.
Senescence-associated β-galactosidase assay. Fluorescence-based assay. Senescence-associated β -galactosidase (SA-β -Gal) activity was assessed with a fluorescence-based method for analysis via flow-cytometry, as previously described 31 . In brief, the cells were incubated for 1 h with bafilomycin A1 (Sigma) to prevent lysosomal acidification and were subsequently incubated with 5-dodecanoylaminofluorescein di-β -D-galactopyranoside (C12FDG, Invitrogen), a fluorogenic substrate for β -galactosidase. Fluorescence was detected using an Accuri C6 Flow Cytometer (BD Biosciences). Relative SA-β -Gal activity was estimated using the Median Fluorescence Intensity (MFI) of the population. Non-labelled samples were used to determine the autofluorescence.
Chromogenic assay. The senescence β -galactosidase staining kit (Cell Signaling Technology) was used to histochemically detect β -galactosidase activity. SA-β -Gal activity results in a cytoplasmic blue staining that can be visualised using light microscopy. The percentage of the means and SDs were derived from counting 100 cells after 72 h and 96 h.
Cell cycle analysis. Cells were stained with propidium iodide (40 μ g/ml) containing ribonuclease A (10 μ g/ml) and were analyzed using a fluorescence activated cell sorter (MACSQuant Analyzer) and MACSQuantify software.
5-Bromo-2-deoxyuridine (BrdU, BD Bioscience) was added to the culture media at a final concentration of 10 μ M for 10 min. Pelleted cells were resuspended in 30 mM HCl/0.5 mg/mL pepsin. BrdU was immunodetected with a mouse anti-BrdU antibody (DAKO, clone Bu20a) and a fluorescein-conjugated donkey anti-mouse antibody (Life Technologies) and cells were stained with PI. Flow cytometry analyses were performed using a Fluorescence was detected using an Accuri C6 Flow Cytometer (BD Biosciences).
Scientific RepoRts | 6:36539 | DOI: 10.1038/srep36539 Tumour assay in vivo. Animal experiments were carried out at the Animal Facilities of the Gustave Roussy Institute under the conditions established by the French ministry (87-848) and European Community (Directive 86/609/CCE) on the protection of animals used for scientific purposes. The project was officially approved by the Animal Experimentation Ethics Committee of the Gustave Roussy Institute (IGR) and registered under the N° 26 by the IGR Department of Research and conducted in accordance with French Laws and regulations. The animals were maintained in a temperature-controlled facility (22 °C) on a 12-h light/dark cycle and were given free access to standard laboratory chow. Equal amount of B16 melanoma cells (0.5.10 6 cell per condition) transfected with control siRNA or siRNA targeting FANCD2 or MiTF for 72 h were injected subcutaneously in nude mice. Ten days later, mice were sacrificed and tumour isolated and weighted.