Differential regulation of retinoic acid metabolism in Fanconi anemia

Abstract

Fanconi anemia (FA) is a genetic disease characterized by congenital abnormalities and increased risk for bone marrow failure and cancer. FA is caused by mutation of 23 genes, the protein products of which function in the maintenance of genome stability. An important role for the FA proteins in the repair of DNA interstrand crosslinks (ICLs) has been established. While the endogenous sources of ICLs relevant to the pathophysiology of FA have yet to be determined, a role for the FA proteins in the detoxification of reactive metabolic aldehydes has been established. To discover new metabolic pathways linked to FA, we performed RNA-seq analysis on a previously established isogenic FA-D2 patient cell model. Multiple retinoic acid metabolism and signaling genes were differentially expressed in the FA-D2 patient cells, including ALDH1A1 and RDH10, two key enzymes in the retinoic acid pathway. Increased levels of ALDH1A1 and RDH10 were confirmed by immunoblotting. FA-D2 (FANCD2^−/−) patient cells displayed increased aldehyde dehydrogenase activity compared to the FANCD2-complemented cells. Following exposure to retinaldehyde, FA-D2 (FANCD2^−/−) patient cells exhibited increased DNA damage and checkpoint activation. Our findings provide an example of a novel link between retinoic acid metabolism and FA and suggest that retinaldehyde may be an additional reactive metabolic aldehyde relevant to the pathophysiology of FA.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-50045-9.

Introduction

Fanconi anemia (FA) is a rare genetic disease, characterized by congenital abnormalities, increased risk for bone marrow failure and both hematologic and nonhematologic cancers, accelerated aging, and premature mortality¹. Most FA patients harbor compound heterozygous germline mutations in one of twenty-three genes. Consequently, age at diagnosis and clinical manifestations can vary widely. Long-term therapeutic options for FA patients remain underdeveloped. A greater understanding of the molecular etiology of FA is critical for the discovery of new therapeutic options for FA patients.

At the cellular level, the proteins encoded by the FA genes function collectively in the maintenance of genome stability. Exemplifying this function, FA patient cells frequently exhibit spontaneous structural and numerical chromosome aberrations^2,3. In vitro studies have established a critical role for the FA proteins in the repair of DNA interstrand crosslinks (ICLs), a complex process that encompasses elements of the canonical DNA repair pathways nucleotide excision repair (NER), translesion DNA synthesis (TLS), and homologous recombination (HR)⁴. However, the endogenous sources of chromosome instability, including ICLs, in the physiological setting remain to be determined. Several studies over the past decade indicate that aldehydes generated during metabolic processes pose an acute genotoxic threat in the setting of FA. For example, mice with combined inactivation of Fancd2 and Aldh2 (which oxidizes acetaldehyde to acetate) exhibit developmental defects and increased risk for leukemia and/or aplastic anemia^5,6. Similarly, combined inactivation of the FA pathway and Adh5 (major formaldehyde detoxifying enzyme) leads to synthetic lethality^7,8. These findings have led to the hypothesis that the FA pathway functions in a two-tier protection system to mitigate the cytotoxic and genotoxic effects of endogenous metabolic aldehydes. In the first tier, substrate-specific aldehyde dehydrogenases oxidize reactive aldehydes to their corresponding nontoxic carboxylic acids. In the second tier, aldehydes that escape oxidation and react with DNA are repaired in an FA-pathway dependent manner. Inactivation of both tiers leads to severe cellular and physiological consequences^9,10. It remains to be determined if this system extends to other metabolic aldehydes, in addition to acetaldehyde and formaldehyde.

Retinoid metabolism is the cellular process whereby vitamin A (retinol) from dietary sources is converted to all-trans retinoic acid (atRA), a signaling molecule critical for patterning, morphogenesis, and organogenesis during development^11,12. During atRA generation, retinol must first be oxidized to retinaldehyde, a reversible reaction carried out by retinol dehydrogenase enzymes including RDH10. Retinaldehyde is subsequently oxidized to atRA by the ALDH1A1, ALDH1A2, and ALDH1A3 paralogous aldehyde dehydrogenases. As retinaldehyde is a potential genotoxin, vitamin A levels must be precisely regulated during pregnancy, as excessive or insufficient levels can lead to congenital abnormalities and/or embryonic lethality¹³. Nuclear atRA functions primarily as a regulator of gene expression. atRA binds to heterodimers of two classes of nuclear receptors, the retinoic acid receptors (RARs) and the retinoid X receptors (RXRs). RAR-RXR bind to retinoic acid response elements (RAREs) in gene promoters and, upon atRA binding, recruit transcriptional co-activators and histone acetyltransferases (HATs) to activate gene expression. In the absence of atRA, the RAR-RXR nuclear receptors promote gene silencing via corepressor/histone deacetylase (HDAC) complex recruitment¹⁴.

In this study, we describe a novel link between FA and retinoic acid metabolism. RNA-seq analysis of non-transformed skin fibroblasts isolated from a 1-year-old FA-D2 (FANCD2^−/−) patient and the same cells complemented with wild-type FANCD2 revealed differential expression of multiple genes linked to retinoic acid metabolism and signaling. For example, we observed markedly increased expression of the ALDH1A1 and RDH10 genes, which encode for retinaldehyde and retinol dehydrogenases, respectively, in the FA-D2 patient cells. FA-D2 patient cells displayed increased aldehyde dehydrogenase activity compared to FANCD2 complemented cells. Upon exposure to retinaldehyde, FA-D2 patient cells exhibited increased DNA double-strand breaks, replication arrest, and checkpoint activation indicative of a defect in the repair of retinaldehyde-induced DNA damage. Our findings describe a novel case of altered retinoic acid metabolism in an FA-D2 patient and suggest that retinaldehyde may represent an additional reactive metabolic aldehyde linked to the pathophysiology of FA.

Results

Characterization of a novel FA-D2 patient model

The FANCD2 protein is well known to play a key role in DNA interstrand crosslink (ICL) repair and in the cellular DNA replication stress response^4,15. To discover additional cellular pathways linked to FA, and FANCD2, we sought to perform global transcriptional profiling with a non-transformed FA-D2 cell model. FA-D2 (FANCD2^−/−) and FANCD2-complemented skin fibroblasts isolated from a 1-year-old patient with a severe clinical phenotype (Supplementary Table 1) were immortalized with human telomerase and characterized for functional complementation. Compound heterozygous mutations in the FANCD2 gene were confirmed by Sanger sequencing; a G > A transition at nucleotide 2444 in exon 26 [c.2444G > A (exon 26)] leading to the missense change R815Q and a G > A transition at the first nucleotide of intron 28 [c.2715 + 1G > A (IVS28 + 1G > A)], leading to a frameshift and premature truncation E906Lfs*4 (Fig. 1A). No FANCD2 expression was observed in the FA-D2 (FANCD2^−/−) cells, while robust expression - comparable to that observed in HeLa and U2OS cells - was observed in the FANCD2-complemented cells (Figs. 1B and 3B). Compared to FANCD2-complemented cells, FA-D2 (FANCD2^−/−) cells displayed increased chromosome breaks and decreased cell viability following exposure to the DNA interstrand crosslinking agent mitomycin C (MMC) (Figs. 1C and Supplementary Figure S1A). Consistent with previous studies with non-transformed cells, and in contrast to that observed in transformed lines^16,17, we observed low levels of FANCD2 monoubiquitination following MMC treatment and treatment with the DNA polymerase inhibitor aphidicolin (APH) in the FANCD2-complemented cells (Fig. 1B and Supplementary Figure S1B). FA-D2 (FANCD2^−/−) cells exhibited increased MMC- and APH-inducible CHK1 S345 phosphorylation, and increased levels of APH-inducible MRE11 and RPA2, consistent with increased replication stress in these cells compared to FANCD2-complemented cells (Fig. 1B). In a G2/M accumulation assay, FA-D2 (FANCD2^−/−) cells demonstrated a 2-fold increase in G2/M phase cells compared to FANCD2-complemented cells following MMC treatment, consistent with a block in cell cycle progression in the absence of FANCD2 (Fig. 1D). Together, our analyses confirm functional complementation of these non-transformed FA-D2 patient fibroblasts and establish these cells as an excellent isogenic cell system to perform differential transcription expression analysis.

Fig. 1.

Characterization of a novel FA-D2 patient model. (A) Sanger sequencing confirmation of compound heterozygous mutations in the FANCD2 gene in RA2645 (FA-D2) patient fibroblasts; a G > A transition at the first nucleotide of intron 28 [c.2715 + 1G > A (IVS28 + 1G > A)], leading to a frameshift and premature truncation E906Lfs*4 and a G > A transition at nucleotide 2444 in exon 26 [c.2444G > A (exon 26)] leading to the missense change R815Q. (B) FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells were incubated in the absence and presence of 200 nM MMC or 1 µM aphidicolin (APH) for 24 h and whole cell lysates were analyzed by immunoblotted using anti-FANCD2, anti-MRE11, anti-RPA2, and anti-CHK1 pS345 antibodies. For immunoblots in this figure and all others, white space between panels signifies different parts of the same gel or panels from different gels. (C) FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells were incubated in the absence and presence of 150 nM mitomycin C (MMC) for 48 h and metaphase chromosomes analyzed for chromosome breaks and radial formations. (D) FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells were incubated in the absence or presence of MMC (20 nM and 40 nM) MMC for 24 h, cells were fixed and stained with propidium iodide and analyzed by flow cytometry.

Fig. 2.

RNA-seq analysis reveals several differentially expressed gene between FA-D2 (FANCD2^−/−) and FANCD2-complemented cells. (A) Principal component analysis (PCA) of quadruplicate biological samples were projected onto a 2D plane spanned by their first two principal components. The percentage of the total variance per direction is shown. The x-axis depicts the most variance and the y-axis depicts the second most. (B) RNA-seq heatmap of differentially expressed genes between FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells. Up-regulated genes are shown in red and down-regulated genes are shown in blue. (C) Volcano plot of differentially expressed genes between FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells. Genes with an adjusted P value less than 0.05 and a log2 fold change greater than 1 are shown in red while genes with an adjusted P value less than 0.05 and a log2 fold change less than − 1 are shown in green. (D) RNA-seq expression data of FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells were further analyzed using gene set enrichment analysis (GSEA) (v4.2.1) platform sourced from the Broad Institute at MIT, for angiogenesis, neurogenesis, and cellular response to retinoic acid gene sets.

RNA-seq analysis reveals several differentially expressed gene sets in FA-D2 (FANCD2^−/−) cells

We performed bulk RNA-seq analysis on our non-transformed FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells, with quadruplicate biological replicates sequenced at a read depth of 60 M reads per sample. Principal component analysis (PCA) revealed clustering by genotype with 85% of the variance attributed to gene expression differences between FA-D2 (FANCD2^−/−) cells and FANCD2-complemented cells (Fig. 2A). A clustered heatmap illustrates the differential gene expression profiles between patient and complemented FA-D2 (FANCD2^−/−) cells (Fig. 2B). Our analysis uncovered 1,458 significantly differentially expressed genes (adjusted P value (Padj) < 0.05 and a log2 fold change greater than 1) between FA-D2 (FANCD2^−/−) and FANCD2-complemented cells (Fig. 2C). Significantly differentially expressed genes were clustered by gene ontology and the enrichment of gene ontology terms was tested using Fisher’s exact test. Gene ontology terms significantly enriched with a Padj value < 0.05 in the differentially expressed gene sets included extracellular matrix organization (GO:0030198), positive regulation of gene expression (GO:0010628), angiogenesis (GO:0001525), nervous system development (GO:0007399), and the cellular response to retinoic acid (GO:0071300). In addition, gene set enrichment analysis (GSEA) demonstrated enrichment for genes with defined roles in angiogenesis, neurogenesis, and the cellular response to retinoic acid, in FANCD2-complemented cells compared to FA-D2 (FANCD2^−/−) cells (Fig. 2D). The cellular response to retinoic acid/retinoic acid metabolism was of particular interest given recent established connections between the FA pathway and the detoxification of reactive metabolic aldehydes^9,10.

Fig. 3.

Differential regulation of retinoic acid metabolism and signaling in FA-D2 (FANCD2^−/−) patient cells. (A) Morpheus heatmap analysis demonstrating differential expression of retinoic acid metabolism and signaling genes in FA-D2 (FANCD2^−/−) cells relative to FANCD2-complemented FA-D2 cells. Genes with increased expression in FA-D2 (FANCD2^−/−) cells relative to FANCD2-complemented FA-D2 cells are depicted in red and genes with decreased expression are depicted in blue. (B) Immunoblotting validation of increased expression of ALDH1A1 and RDH10 in FA-D2 (FANCD2^−/−) cells relative to FANCD2-complemented FA-D2 cells. Antibody-protein specificity was confirmed by depleting U2OS 3xFLAG-FANCD2 cells of ALDH1A1 (α-ALDH1A1 panel) or RDH10 (α-RDH10 panel) using pooled siRNAs, referred to as siTarget. Whole-cell lysates were prepared from the indicated cells and immunoblotted using the indicated antibodies. LE, Long exposure; SE, Short exposure. (C) Increased ALDH1A1 activity in FA-D2 (FANCD2^−/−) cells as demonstrated using the isoform-selective fluorescent ALDH1A1 probe, AlDeSense, in comparison to the control nonreactive Ctrl-AlDeSense probe. (D) ALDEFLUOR flow cytometry assay demonstrating increased aldehyde dehydrogenase (ALDH) activity in FA-D2 (FANCD2^−/−) cells compared to FANCD2-complemented FA-D2 cells. ALDH-negative HeLa cells and ALDH-positive A549 cells were used as negative and positive controls, respectively. Cells were incubated in the absence or presence of the ALDH inhibitor N, N-diethylaminobenzaldehyde (DEAB) prior to analysis.

Dysregulation of retinoic acid metabolism in non-transformed FA-D2 (FANCD2^−/−) patient cells

RNA-seq analysis uncovered differential expression of multiple genes linked to retinoic acid metabolism and/or retinoic acid signaling in FA-D2 (FANCD2^−/−) cells compared to FANCD2-complemented FA-D2 cells (Fig. 3A and Supplementary Table 2). For example, we observed a negative 7.2 log2 fold change of the ALDH1A1 gene, indicating markedly increased expression in FA-D2 (FANCD2^−/−) cells compared to FANCD2-complemented cells. Increased expression of ALDH1A1 and the differential expression of retinoic acid-responsive genes PAX9 and SOX11^18,19 were confirmed by qPCR (Supplementary Figure S1C). ALDH1A1 was the most highly differentially expressed ALDH gene out of 15 of the 19 human ALDH genes expressed in these cells (Supplementary Table 3). ALDH1A1 encodes for an aldehyde dehydrogenase that irreversibly oxidizes retinaldehyde to retinoic acid. We also observed negative 1.4 and 1.1 log2 fold changes for RDH10 and CRABP2, respectively (Fig. 3A and Supplementary Table 2). RDH10 encodes for retinol dehydrogenase and converts retinol to retinaldehyde in a reversible reaction. CRABP2 encodes for cellular retinoic acid binding protein II and transports cytoplasmic retinoic acid to the nucleus^11,12. A very minor difference in ALDH2 gene expression was observed (-0.6 log2 fold change), while we detected a -1.1 log2 fold change for the ADH5 gene (Supplementary Tables 3 and 4). We used immunoblotting to confirm the increased expression of ALDH1A1, RDH10, and CRABP2 in the FA-D2 (FANCD2^−/−) cells using positive and negative controls (Fig. 3B and Supplementary Figures S1D and E). A549, a lung epithelial-like carcinoma line, was used as a positive control for ALDH1A1 expression as this line is known to express high levels of this protein (Human Protein Atlas) (Fig. 3B)^20,21. Increased ALDH1A1 expression in FA-D2 (FANCD2^−/−) cells, and not in FANCD2-complemented cells, was also confirmed using the isoform-selective fluorescent ALDH1A1 probe, AlDeSense (Fig. 3C and Supplementary Figure S1F)²². Cellular fractionation experiments revealed that ALDH1A1, CRABP2, and RDH10 were primarily expressed in the soluble fraction, with no change in subcellular localization following exposure of the cells to exogenous retinaldehyde (Supplementary Figure S1G). Low levels of RDH10 were detected in the chromatin fraction of FA-D2 (FANCD2^−/−) cells (Supplementary Figure S1G).

We next analyzed ALDH1A1 and ALDH2 expression in FA-A (FANCA^−/−), FA-F (FANCF^−/−), and FA-D2 (FANCD2^−/−) patient lymphoblasts and their FANCA-, FANCF-, or FANCD2-complemented counterparts, respectively (Supplementary Figure S1E). ALDH1A1 expression was not observed in any of the lymphoblasts tested (Supplementary Figure S1H). We did, however, observe increased expression of ALDH2 in FA-A, FA-F, and FA-D2 lymphoblasts, compared to their complemented counterparts (Supplementary Figure S1H). We also examined Aldh1a1 and Rdh10 expression in wild-type, Fancd2^+/+/Aldh2^−/−, Fancd2^−/−/Aldh2^+/+, and Fancd2^−/−/Aldh2^−/− non-transformed mouse ear fibroblasts. In contrast to wild-type cells, we observed an Aldh1a1 immunoreactive band - albeit at a lower molecular weight than expected - and increased levels of Rdh10 in Fancd2^+/+/Aldh2^−/−, Fancd2^−/−/Aldh2^+/+, and Fancd2^−/−/Aldh2^−/−cells (Supplementary Figure S1I). We also examined Aldh1a1 and Rdh10 protein expression in Fancd2^+/+, Fancd2^+/−, and Fancd2^−/− mouse embryonic fibroblasts (MEFs) and observed increased levels of Aldh1a1 in Fancd2^−/− MEFs compared to Fancd2^+/+ and Fancd2^+/− MEFs (Supplementary Figure S1J). No appreciable differences in Rdh10 expression levels between these cells were observed (Supplementary Figure S1J). Taken together, these results from several cell systems establish that loss of FANCD2/Fancd2 is associated with altered retinoic acid metabolism.

To determine if increased ALDH1A1 expression results in higher aldehyde dehydrogenase (ALDH) activity in the non-transformed FA-D2 (FANCD2^−/−) cells, we performed an ALDEFLUOR assay with FA-D2 (FANCD2^−/−) and FANCD2-complemented cells. Cells were incubated with the ALDH substrate BODIPY-aminoacetaldehyde (BAAA), which is converted into the fluorescent molecule BODIPY-aminoacetate (BAA) via ALDH activity. This reaction was performed in the absence and presence of the ALDH inhibitor N, N-diethylaminobenzaldehyde (DEAB). FA-D2 (FANCD2^−/−) cells showed an approximate 2-fold increase in ALDH activity compared to FANCD2-complemented cells, indicating that increased ALDH1A1 expression is associated with higher ALDH activity in these cells (Fig. 3D and Supplementary Figure S1K). HeLa and A549 cells were included as negative and positive controls for ALDH activity, respectively (Fig. 3D).

Increased retinaldehyde genotoxicity in FA patient cells

Aldehydes are strong electrophiles, and the terminal carbonyl group can form adducts with various cellular targets including DNA. The mechanisms of retinaldehyde genotoxicity have not been clearly elucidated. To learn more about the mechanisms of retinaldehyde genotoxicity, we exposed HeLa and U2OS cells to the type II topoisomerase inhibitor etoposide (VP-16) and all-trans-retinaldehyde (atRALD) and analyzed protein biomarkers of DNA damage, DNA repair, and cell cycle checkpoint activation. As expected, VP-16 exposure led to a marked increase in H2AX S139 (γH2AX) and CHK1 S345 phosphorylation, indicative of DNA double-strand break formation and activation of the intra-S phase checkpoint, respectively (Figs. 4A and B and Supplementary Figs. 2 A and 2B). At later time points following exposure, atRALD treatment also resulted in increased levels of γH2AX (Figs. 4A and B and Supplementary Figs. 2 A and 2B). atRALD exposure also resulted in increased CHK1 S345 and CHK2 T68 phosphorylation in both HeLa and U2OS cells (Figs. 4A and B and Supplementary Figs. 2 A and 2B). As ALDH1A1 exhibits specificity for atRALD and 9-cis-retinaldehyde (9cRALD), we also exposed HeLa and U2OS cells to both isomers and analyzed the same protein biomarkers of DNA damage, DNA repair, and cell cycle checkpoint activation. Exposure to both atRALD and 9cRALD resulted in DNA damage and checkpoint activation to varying degrees. Under the conditions tested, however, we observed a greater increase in levels of γH2AX following atRALD treatment, compared to 9cRALD (Supplementary Figure S2C and D). We next compared the response of the non-transformed FA-D2 (FANCD2^−/−) and FANCD2-complemented cells to atRALD. We observed increased levels of γH2AX, CHK1 pS345, and CHK2 T68 phosphorylation in atRALD-treated FA-D2 (FANCD2^−/−) cells compared to the FANCD2-complemented cells (Fig. 4C and Supplementary Figures S2E). We also observed increased γH2AX nuclear foci formation in FA-D2 (FANCD2^−/−) cells compared to the FANCD2-complemented cells following exposure to atRALD and VP-16 (Fig. 5A and Supplementary Figure S2F). Cell cycle analysis revealed a higher percentage of FA-D2 (FANCD2^−/−) cells in S-phase both with and without prior exposure to atRALD, compared to FANCD2-complemented cells (Fig. 5B and Supplementary Figures S2G and H). We also observed increased levels of atRALD-induced mitotic aberrations, including micronuclei, in FA-D2 (FANCD2^−/−) cells compared to FANCD2-complemented cells (Fig. 5C and Supplementary Figure S2I). These results indicate that FA-D2 (FANCD2^−/−) cells are compromised in their ability to repair atRALD-induced DNA damage. Similar findings were observed for EBV-immortalized FA-A (FANCA^−/−) cells, whereby atRALD exposure led to increased γH2AX and CHK2 T68 phosphorylation in FA-A (FANCA^−/−) cells compared to FANCA-complemented cells (Supplementary Figure S2J-L). We did not however observe increased sensitivity to cell killing by retinaldehyde in the transformed FA-D2 (FANCD2^−/−) cell line, PD20, compared to the same cells complemented with wild-type FANCD2 (Supplementary Figure S2M). Taken together, these results suggest an important role for the FA pathway in the repair of atRALD-induced DNA damage, similar to that previously observed for acetaldehyde and formaldehyde.

Fig. 4.

Increased retinaldehyde-induced activation of the DNA damage response in FA-D2 (FANCD2^−/−) patient cells. HeLa (A) and U2OS (B) cells were exposed to either 25 µM all-trans-retinaldehyde (atRALD) or 25 µM etoposide (VP-16) for 2 h, released into drug-free media, and whole-cell lysates were prepared at 0, 3, 6, and 24 h following the exposure period. Whole-cell lysates were immunoblotted with the indicated antibodies. (C) Non-transformed FA-D2 (FANCD2^−/−) and FANCD2-complemented cells were exposed to 25 µM atRALD for 2 h, released into atRALD-free media, and whole-cell lysates were prepared at 0, 3, and 6 h following the exposure period. Whole-cell lysates were immunoblotted with the indicated antibodies.

Fig. 5.

Increased retinaldehyde-induced genotoxicity in FA-D2 (FANCD2^−/−) patient cells. (A) Non-transformed FA-D2 (FANCD2^−/−) and FANCD2-complemented cells were exposed to 25 µM all-trans-retinaldehyde (atRALD) or 25 µM etoposide (VP-16) for 2 h and released into drug-free media. Immunofluorescence microscopy (IF) for γH2AX was performed at the indicated time points following release. Error bars represent the standard errors of the means from two independent experiments. At least 600 nuclei were scored per biological replicate. (B) FA-D2 (FANCD2^−/−) and FANCD2-complemented cells were treated as in (A) and EdU S-phase IF labeling was performed. EdU labeling demonstrates a higher percentage of FA-D2 (FANCD2^−/−) cells in S-phase both with and without prior exposure to atRALD, compared to FANCD2-complemented cells. Error bars represent the standard errors of the means from two independent experiments. (C) FA-D2 (FANCD2^−/−) and FANCD2-complemented cells were treated as in (A), nuclei were labeled with DAPI and micronuclei scored by IF. Error bars represent the standard errors of the means from two independent experiments. ***, P < 0.001; **, P < 0.01; *, P < 0.05.

Attenuated apoptosis in non-transformed FA-D2 (FANCD2^−/−) cells

Gene set enrichment analysis (GSEA) of our RNA-seq data also revealed differential expression of apoptosis genes, suggesting downregulation of apoptosis in FA-D2 (FANCD2^−/−) cells compared to the FANCD2-complemented cells (Fig. 6A). To experimentally validate differential apoptosis gene expression, we analyzed cleavage of the poly(ADP)ribose polymerase enzyme PARP1 by immunofluorescence microscopy (IF). We observed an approximate 2-fold reduction in PARP1-positive nuclei following staurosporine (STS) exposure in FA-D2 (FANCD2^−/−) cells compared to cells complemented with wild-type FANCD2 (Fig. 6B and Supplementary Figure S3). Similarly, using immunoblotting, we observed reduced STS-induced PARP, caspase 3 (CASP3), and caspase 7 (CASP7) cleavage in FA-D2 (FANCD2^−/−) cells compared to FANCD2-complemented cells (Fig. 6C). These results demonstrate that apoptosis is functionally attenuated in the non-transformed FA-D2 (FANCD2^−/−) patient cells.

Fig. 6.

Attenuated apoptosis in non-transformed FA-D2 (FANCD2^−/−) patient cells. (A) RNA-seq gene set enrichment analysis demonstrates increased expression of apoptosis genes in FANCD2-complemented cells compared to FA-D2 (FANCD2^−/−) cells. (B) Increased PARP1 nuclear staining in FANCD2-complemented cells compared to FA-D2 (FANCD2^−/−) cells. Cells were incubated in the absence or presence of 1 µM staurosporine (STS) for 3 h, released into STS-free media, and the numbers of PARP1-stained nuclei scored. Error bars represent the standard deviations from at least 5 technical replicate measurements from a single experiment. At least 500 nuclei were scored per sample. This experiment was repeated three times with similar findings. ***, P < 0.001. (C) Decreased STS-induced apoptosis in FA-D2 (FANCD2^−/−) cells. FA-D2 (FANCD2^−/−) and FANCD2-complemented cells were incubated in the absence or presence of 1 µM STS for 3 h, released into STS-free media, whole-cell lysates prepared at 0, 3, and 6 h, and immunoblotted with the indicated antibodies.

DATA AVAILABILITY STATEMENT.

Discussion

In this manuscript, we describe a previously undiscovered molecular link between FA and retinoic acid metabolism and signaling. Bulk RNA-seq analysis of non-transformed FA-D2 patient cells and the same cells complemented with wild-type FANCD2 uncovered differential expression of multiple retinoic acid metabolism and signaling genes. For example, we observed markedly increased cytoplasmic expression of the ALDH1A1 and RDH10 genes in FA-D2 (FANCD2^−/−) patient cells compared to FANCD2-complemented cells, findings confirmed by immunoblotting. Consistent with increased expression of ALDH1A1, FA-D2 (FANCD2^−/−) patient cells demonstrated increased aldehyde dehydrogenase activity as determined by the ALDEFLUOR™ assay. We did not detect ALDH1A1 expression in EBV-immortalized FA-A, FA-F, or FA-D2 patient lymphoblasts or SV40-transformed FA-A and FA-D2 fibroblasts. To the best of our knowledge, all these cells were derived from FA patients greater than 7 years of age.

We also observed increased Aldh1a1 expression in Fancd2^−/− mouse embryonic fibroblasts and increased Rdh10 expression in non-transformed Fancd2^−/− mouse ear fibroblasts. Taken together, these results suggest that dysregulated retinoic acid metabolism is a feature of non-transformed FA models and restricted to younger FA patients, consistent with the critical role of retinoic acid metabolism and signaling during early development^14,23. FA patients frequently present with heterogenous multisystemic congenital defects, including short stature, microcephaly, microphthalmia, and upper and lower limb defects. The FA-D2 (FANCD2^−/−) cells described in this study were isolated from a 1-year-old patient with a particularly severe clinical phenotype, characterized by multiple congenital defects, in addition to hypersensitivity to DEB-induced chromosome breakage during diagnosis. Our findings suggest that dysregulated retinoic acid metabolism and signaling during development could play a significant role in the etiology of FA patient congenital abnormalities.

While important roles for the FA proteins in the repair of DNA interstrand crosslinks (ICLs) generated by mitomycin C and cisplatin have been clearly defined in vitro, the endogenous sources of genome instability in the physiological setting remain poorly understood. Over the past decade, several studies have established a critical role for certain FA proteins in a 2-tier system designed to mitigate the cytotoxic and genotoxic effects of reactive metabolic aldehydes, including acetaldehyde and formaldehyde^9,10. For example, Aldh2^−/− Fancd2^−/− mice exhibit congenital defects and increased risk of leukemia, while aged Aldh2^−/− Fancd2^−/− mice that escape leukemia develop aplastic anemia^5,6. Similarly, Aldh2 gene inactivation or inhibition is synthetic lethal with loss of the homologous recombination and FA proteins Brca1, Brca2, and Rad51 ²⁴. A synthetic lethal interaction between the alcohol dehydrogenase 5 gene ADH5, which restricts the cellular accumulation of formaldehyde, and the FA pathway has also been established⁷. Here we demonstrate upregulated tier-1 activity in FA-D2 cells, suggesting that modulation of tier-1 activity may be a cellular strategy to mitigate genotoxic stress upon disruption of the FA pathway. Our study also indicates that the FA pathway may also play an important role in the mitigation of retinaldehyde genotoxicity. Cellular exposure to retinaldehyde resulted in the induction of phosphorylation of histone variant H2AX, a marker of DNA double-strand breaks (DSBs), albeit to a lesser extent than the topoisomerase type II inhibitor etoposide. In contrast to etoposide, which led to a marked increase in CHK1 S345 phosphorylation, indicative of activation of the intra-S-phase checkpoint, retinaldehyde exposure primarily induced phosphorylation of CHK2 T68, which is associated with activation of the G2-M checkpoint. We observed increased phosphorylation of H2AX and CHK2 T68 in both FA-A (FANCA^−/−) and FA-D2 (FANCD2^−/−) cells treated with retinaldehyde compared to FANCA- or FANCD2-complemented cells, respectively. Furthermore, following exposure to retinaldehyde, we observed increased γH2AX nuclear foci formation in FA-D2 (FANCD2^−/−) cells compared to FANCD2-complemented cells. Combined, these results indicate that the FA pathway plays an important role in the repair of retinaldehyde-induced DNA damage. Our results also suggest that certain cell lineages that are particularly dependent on retinoic acid signaling, e.g., neural stem and progenitor cells and hematopoietic stem cells, may be acutely reliant on an intact FA DNA repair pathway.

ALDH1A1 is one of 19 evolutionarily conserved human aldehyde dehydrogenase (ALDH) enzymes that catalyze the NAD(P)⁺-dependent irreversible oxidation of aldehydes to their corresponding carboxylic acids²⁵. The importance of the ALDH gene family is underscored by their association with increased risk for numerous pathologies including Sjögren-Larsson syndrome, type II hyperprolinemia, alcohol-related diseases, cancer, and neurodegeneration^25,26. The ALDH1A subfamily, comprised of ALDH1A1, ALDH1A2, and ALDH1A3, all exhibit specificity for retinaldehyde, with ALDH1A1 exhibiting a preference for 9-cis-retinaldehyde while ALDH1A3 exhibits a preference for all-trans-retinaldehyde^27,28. In our studies, we observed higher levels of phosphorylated H2AX following exposure to all-trans-retinaldehyde compared to 9-cis-retinaldehyde. Immunohistology studies have detected increased ALDH1A1 levels in a wide spectrum of cancers, with increased ALDH1A1 levels frequently associated with a poorer prognosis^26,29. The mechanisms by which ALDH1A1 promotes cancer cell stemness and poorer outcomes in cancer patients remain to be clearly elucidated. While we observed evidence for increased DNA damage and genotoxicity in retinaldehyde-treated FA cells, we did not observe differential sensitivity to retinaldehyde cytotoxicity between patient and complemented FA lines. We speculate that FA-D2 (FANCD2^−/−) cells will exhibit an increased mutation frequency and an increased frequency of structural variants following exposure to retinaldehyde, compared to FANCD2-complemented cells. Ongoing experiments are testing this hypothesis. We also speculate that increased levels of ALDH1A1 lead to increased oxidation of retinaldehyde to the biologically active molecule retinoic acid, which plays a key role in the transcriptional regulation of a wide variety of biological processes including proliferation, differentiation, and apoptosis^12,23. Indeed, all-trans-retinoic acid is used to rescue arrested differentiation in PML-RARA acute promyelocytic leukemia (APL)³⁰. Consistent with this prediction, we observed differential expression of numerous retinoic acid-regulated genes in non-transformed FA-D2 (FANCD2^−/−) patient cells, e.g., RARA, EGR1, and TGM2, and several members of the HOX, PAX, and SOX gene families. In our study, gene set enrichment analysis revealed differential expression of apoptosis genes between FA-D2 (FANCD2^−/−) cells and the FANCD2-complemented cells. Attenuated apoptosis in FA-D2 (FANCD2^−/−) cells was functionally validated using immunofluorescence microscopy for cleaved-PARP1 and immunoblotting for cleaved-PARP1 and cleaved caspase 3 and caspase 7. Taken together, our findings indicate that altered retinoic acid metabolism and signaling is likely to modulate several pathways and endpoints in non-transformed FA-D2 (FANCD2^−/−) cells.

A previous study established an important role for the FA genes BRCA1/FANCS and PALB2/FANCN in the transcriptional regulation of retinoic acid-regulated genes³¹. Depletion of BRCA1/FANCS and PALB2/FANCN resulted in decreased expression of the HOXA1 and HOXA2 genes following treatment with retinoic acid³¹. BRCA1/FANCS and PALB2/FANCN were shown to be recruited to the promoters of retinoic acid-regulated genes indicating that these proteins may play a direct role in the transcriptional regulation of retinoic acid-responsive genes³¹. Our studies establish an important link between retinoic acid metabolism - including retinaldehyde detoxification and retinoic acid-mediated transcriptional regulation - and the FA pathway. Considering the critical role of retinoic acid during embryogenesis, altered retinoic acid metabolism may contribute to many aspects of the pathophysiology of FA including the heterogeneous congenital abnormalities and increased risk for bone marrow failure observed among FA patients. Nervous system defects have become increasingly prevalent among FA patients^32–34. While retinoic acid is a well-established and critical signaling molecule during embryonal neuronal patterning and differentiation, retinoic acid also plays a key role in adult axon outgrowth and nerve regeneration and in the maintenance of the differentiated state of adult neurons^35,36. Further studies using developmentally appropriate and/or tissue-specific models are necessary to further define the role of retinoic acid metabolism in the pathophysiology of FA.

Materials and methods

Cell culture and generation of cell lines

A549, HeLa, Hepa1c1c7 and U2OS cells were obtained from American Type Culture Collection (ATCC). U2OS 3xFLAG-FANCD2 cells were a generous gift from the Takata lab at Kyoto University. Mouse ear fibroblasts were a generous gift from the Fanconi Anemia Research Fund. Mouse embryonic fibroblasts were a generous gift from the Andreassen lab at Cincinnati Children’s Hospital. The non-transformed FA-D2 (FANCD2^−/−) patient samples described in this manuscript originated from a male patient diagnosed at the Rockefeller University International Fanconi Anemia Registry (IFAR) (Supplementary Table 1). Skin fibroblasts (RA2645) were prepared at 1 year of age. These cells harbor a maternally inherited missense FANCD2 hypomorphic mutation leading to an R815Q change and paternally inherited 2715 + 1G > A frameshift that leads to a truncated protein. Skin fibroblasts RA2645 were complemented with wild-type FANCD2 and functionally characterized by Helmut Hanenberg at the University of Duisberg-Essen and Detlev Schlinder at the University of Wuerzburg. FA-D2 (FANCD2^−/−) and FANCD2-complemented cells were immortalized with human telomerase and grown in DMEM complete medium supplemented with 1 µg/ml puromycin. HSC72 (FA-A) and FANCA-complemented, EUFA121 (FA-F) and FANCF-complemented, and PD20 (FA-D2) and FANCD2-complemented patient-derived EBV-immortalized lymphoblasts were grown in Roswell Park Memorial Institute 1640 medium (RPMI 1640) supplemented with 20% (vol/vol) fetal bovine serum (FBS), 2 mM L-glutamine, and 50 µM penicillin-streptomycin. PD20 (FA-D2) and FANCD2-complemented patient-derived SV40-transformed fibroblasts were grown in DMEM medium supplemented with 15% (vol/vol) fetal bovine serum (FBS), 2 mM L-glutamine, and 50 µM penicillin-streptomycin. HSC72 (FA-A) lymphoblasts were derived from a 10-year-old female patient harboring a homozygous in-frame deletion of exons 18–28 in the FANCA gene. EUFA121 (FA-F) lymphoblasts were derived from a male patient of unknown age with compound heterozygous mutations in the FANCF gene (c.16 C > T, c.351_397del)³⁷. PD20 cells were derived from a 7-year-old male patient with compound heterozygous mutations in the FANCD2 gene (c.376 A > G, c.3707G > A)³⁸. Mouse ear fibroblasts, A549 lung carcinoma, HeLa cervical carcinoma, Hepa1c1c7 mouse liver hepatoma, and U2OS osteosarcoma cells were grown in Dulbecco’s modified Eagle’s medium (DMEM) supplemented with 10% (vol/vol) FBS, L-glutamine, and penicillin-streptomycin.

Immunoblotting

For immunoblotting analysis cell pellets were washed in phosphate-buffered saline (PBS) and lysed in 2% (wt/vol) SDS, 50 mM Tris-HCl, and 10 mM EDTA followed by sonication for 10s at 10% amplitude using a Sonic Dismembrator model 500 (Fisher Scientific). Proteins were resolved on NuPAGE 3–8% (wt/vol) Tris-acetate or 4–12% (wt/vol) Bis-Tris gels (Invitrogen) and transferred to polyvinylidene difluoride (PVDF) membranes. The following antibodies were used: rabbit polyclonal antisera against AKR1C3 (11194-1-AP; Proteintech), ALDH2 (18818 S; Cell Signaling), ATR pS428 (2853 S; Cell Signaling), CRABP2 (10225-1-AP; Proteintech), FANCD2 (NB100-182; Novus Biologicals), MRE11 (PC388; Calbiochem), PARP (9542; Cell Signaling), and RDH10 (14644-1-AP; Proteintech), or rabbit monoclonal antisera against ALDH1A1 (54135; Cell Signaling), ATM pS1981 (5883; Cell Signaling), ATR (13934; Cell Signaling), Caspase-3 (14220; Cell Signaling), Caspase-7 (12827; Cell Signaling), CHK1 pS345 (2348T; Cell Signaling), CHK2 pT68 (2197; Cell Signaling), Cleaved Caspase-3 (D175) (9664; Cell Signaling), Cleaved Caspase-7 (D198) (8438; Cell Signaling), Cleaved PARP (D214) (5625; Cell Signaling), and H2AX pS139 (γH2AX) (9718; Cell Signaling), or mouse monoclonal antisera against alpha-tubulin (MS-581-PO; Neomarkers), ATM (11G12) (92356; Cell Signaling), CHK1 (2G1D5) (2360; Cell Signaling), p53 (DO-1) (F0909; Santa Cruz Biotechnology), p53 pS15 (9286; Cell Signaling), and RPA2 (NA18; Calbiochem). Protein band intensities were quantified using ImageJ (National Institutes of Health, USA). Immunoblot images were imported into ImageJ, converted to 8-bit grayscale format, and inverted to obtain positive peak profiles. Bands were selected using a uniform rectangular selection area for all lanes, and densitometric analysis was performed using the Gel Analysis function. Peak areas corresponding to individual bands were measured using the Wand tool. Background signal was determined from an empty region of the membrane or gel and subtracted from each band intensity value. Where indicated, band intensities were normalized to the corresponding loading control and expressed as relative protein levels.

ALDEFLUOR assay

Cells were counted and adjusted to a concentration of 5 × 10⁵ cells/mL with ALDEFLUOR assay (01700; Stem Cell Technologies) buffer. Activated ALDEFLUOR reagent was added to cell suspensions and mixed well. Subsequently, half of the cell suspension was transferred to tubes containing the aldehyde dehydrogenase inhibitor ALDEFLUOR N, N-diethylaminobenzaldehyde (DEAB). Experimental and control samples were incubated at 37^°C for 45 min. Cells were washed with ALDEFLUOR buffer once and resuspended in ALDEFLUOR buffer for FACS analysis. Samples were analyzed on a BD Biosciences-LSRII flow cytometer.

AlDeSense assay

Cells were plated and incubated at 37^°C/5% CO₂ overnight. The following day the growth media was removed, and cells were incubated with Opti-MEM supplemented with either control AlDeSense AM or AlDeSense AM probes²². Cells were incubated at room temperature with AlDeSense probes for 30 min and then imaged with a Zeiss Axio Observer A.1 inverted fluorescent microscope with ZEN lite image acquisition software.

RNA sequencing

Quadruplicate biological replicates of non-transformed FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells were harvested and washed with ice-cold PBS. Cell pellets were flash frozen and sent for sequencing with GENEWIZ/Azenta Life Sciences. RNA library prep was completed via polyA selection and HiSeq sequencing. RNA samples were quantified using Qubit 2.0 Fluorometer (Life Technologies, Carlsbad, CA) and RNA integrity was checked using Agilent TapeStation 4200 (Agilent Technologies, Palo Alto, CA). Sequencing libraries were prepared using NEBNext Ultra II RNA library prep kit using the manufacturer’s instructions (NEB, Ipswich, MA). Sequencing libraries were validated on the Agilent TapeStation and quantified using Qubit 2.0 Fluorometer as well as via quantitative PCR (KAPA Biosystems, Wilmington, MA). Sequencing libraries were clustered on a single lane of a flowcell and loaded onto an Illumina HiSeq 4000 instrument. Samples were sequenced using a 2 × 150 bp paired-end configuration at 60 million reads per sample.

RNA-seq analysis

Reads were trimmed to remove adapter sequences and bases of poor quality using Trimmomatic v.0.36. Trimmed reads were then mapped to the hg19 ENSEMBL reference genome using STAR aligner v.2.5.2b. FeatureCounts subread package v.1.5.2 was run on BAM files generated from the alignment, to calculate reads uniquely mapping to genes in the genome. FeatureCounts output files were used for downstream differential expression analysis using DESeq2. The RNA-seq data generated in this study have been deposited to the NCBI Gene Expression Omnibus under accession number GSE313989.

RNA extraction and qPCR

Cells were grown in 10 cm² dishes, pelleted by centrifugation and resuspended in ice cold PBS. RNA was isolated using the PureLink RNA Mini Kit (Invitrogen) with on-column DNase treatment and quantified using a Nanodrop spectrophotometer. RNA was converted into single-stranded cDNA using the High-Capacity Reverse Transcription Kit (ThermoFisher Scientific). One µg of RNA was used for cDNA synthesis. Synthesis was performed in triplicate and subsequently pooled. Quantification was conducted using TaqMan Gene Expression probes (ThermoFisher Scientific) for GAPDH, ALDH1A1, PAX9, and SOX11, with TaqMan Gene Expression Master Mix (ThermoFisher Scientific) in a Roche 480 Lightcycler. CP values for each gene were averaged and normalized to the housekeeping gene GAPDH. Relative transcript abundance was then calculated using the formula 1.8^{(Normalized CP Value)}. Reactions were performed in triplicate with 2 µl of cDNA template per well.

MTT assay

PD20 (FA-D2) and FANCD2-complemented patient-derived SV40-transformed fibroblasts were seeded in 96-well dishes at a concentration of 10 × 10³ cells/well and allowed to attach overnight. The following day cells were treated with varying concentrations of retinaldehyde for 72 h. After the incubation period, cell viability was determined using the MTT assay (ThermoFisher Scientific) per the manufacturer’s instructions. Viability was calculated by determining the difference in absorbances at 490 nm and 750 nm (A₄₉₀ - A₇₅₀). Relative viability was determined using control wells without retinaldehyde exposure. All conditions were performed in triplicate.

Immunofluorescence microscopy

For immunofluorescence microscopy (IF) analysis, cells were seeded in 8-well tissue culture slides (Millipore) and treated with retinaldehyde for 2 h, followed by a recovery period of 0, 3, or 6 h, or with staurosporine for 3 h only. For combined FANCD2 and γH2AX IF, cells were fixed in 4% w/v paraformaldehyde and 2% w/v sucrose at 4 °C followed by permeabilization in 0.3% v/v Triton X-100 in PBS. For cleaved PARP IF, cells were fixed in ice-cold 100% methanol. Fixed cells were blocked for 30 min in antibody dilution buffer (5% v/v goat serum, 0.1% v/v NP40 in PBS) followed by incubation with anti-FANCD2 (NB100-182, Novus Biologicals) and anti-H2AX pS139 (JBW301, Millipore) or cleaved PARP (D214) (5625; Cell Signaling) antibodies for 1 h. Cells were washed three times in PBS and incubated in goat anti-mouse Texas Red (T-862, ThermoFisher Scientific) and goat anti-rabbit Fluorescein (F-2765, ThermoFisher Scientific) secondary antibody for 1 h at room temperature. The slides were counterstained and mounted in vectashield plus 4’6-diamidine-2-phenylindole dihydrochloride (DAPI). Nuclear foci were analyzed using a Zeiss Axioimager.A2 upright epifluorescence microscope with ZEN lite image acquisition software.

Micronucleus assay

For micronuclei analysis, cells were seeded on 8-well tissue culture slides (Millipore) overnight. On the following day, cells were treated with retinaldehyde for 2 h, followed by a recovery period of 0, 6, or 24 h. Cells were fixed in fixing buffer (4% w/v paraformaldehyde, 2% w/v sucrose in PBS, pH 7.4) at 4 °C for 10 min. Permeabilization was performed using 0.3% v/v Triton-X in PBS for 10 min at room temperature. Cells were washed in PBS and stained with 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI, Vector Laboratories). At least 300 nuclei were scored per group.

EdU incorporation assay

Cells were seeded on 8-well tissue culture slides (Millipore) and treated with retinaldehyde for 2 h, followed by a recovery period of 0, 6, or 24 h. Two hours prior to the end of each recovery period, cells were labeled with 10 µM EdU labeling solution. Post recovery cells were fixed in 3.7% formaldehyde for 15 min at room temperature, followed by permeabilization in 0.5% v/v Triton X-100 in PBS. Fixed cells were then incubated in Click-iT plus reaction cocktail for 30 min at room temperature protected from light (C10637, ThermoFisher Scientific). Cells were washed with 3% BSA in PBS and then counterstained and mounted in vectashield plus DAPI.

G2/M accumulation assay

Cells were seeded in 10 cm² dishes and incubated in the absence or presence of mitomycin C (MMC) or etoposide (VP-16) for 24 h, or retinaldehyde (RALD) for 2 h, released and allowed to recover for 0, 3, 6, and 24 h. Cells were harvested, resuspended in PBS and fixed in ice-cold methanol. Samples were then washed in PBS and incubated with 50 µg/ml propidium iodide (PI) (Sigma) and 1x RNase A for 15 min at 37 °C, followed by analysis using a BD FACSVerse flow cytometer. The percentages of cells in G₁, S, and G₂/M were determined by analyzing PI histograms with Flowjo v10.7.1 software.

Crystal violet cell viability assay

Cells were seeded in 6-well dishes at a concentration of 40,000 cells/well and allowed to adhere at 37^°C/5% CO₂ overnight. Cells were then incubated with drug for 72 h at 37 °C. After treatment, the media was aspirated, cells were washed once with PBS, and a 0.5% w/v crystal violet solution was added to each well. Plates were subsequently washed in water, allowed to dry overnight, and imaged. Finally, 100% methanol was added to each well and incubated for 2–3 h at 4 °C. Methanol solution was then transferred to 96-well plates where the optical density of each well was measured at 570 nm (OD ₅₇₀). The average OD₅₇₀ of non-treated cells was set to 100% and viability of treated cells were compared to cells that weren’t exposed to drug.

siRNA

U2OS cells were plated at 700,000 cells per 10 cm² dish. The following day cells were transfected with SMARTpool siALDH1A1 (Horizon Discovery), siRDH10 (Horizon Discovery), or a control non-targeting siRNA using lipofectamine 2000 (ThermoFisher Scientific). Cells were incubated with siALDH1A1, siRDH10, or siControl for 24 h followed by a 24 h recovery period. Cells were harvested 96 h post treatment, washed with ice cold PBS, and used for downstream experiments.

Gene set enrichment analysis

Gene sets enriched in either FA-D2 (FANCD2^−/−) or FANCD2-complemented FA-D2 cells were analyzed by gene set enrichment analysis (GSEA) (v4.2.1.) software sourced by the Broad Institute (https://www.gsea-msigdb.org/gsea/index.jsp). Four gene sets including Angiogenesis (HALLMARK_ANGIOGENESIS.v7.5.1.gmt), Neurogenesis (GOBP_NEUROGENESIS.v7.5.1.gmt), Cellular Response to Retinoic Acid (GOBP_CELLULAR_RESPONSE_TO_RETINOIC_ACID.v7.5.1.gmt), and Apoptosis (HALLMARK_APOPTOSIS.v2022.1.Hs.gmt) were utilized on normalized read counts of each sequenced sample. Gene sets tested were considered to be significantly enriched in either FANCD2-complemented FA-D2 or FA-D2 (FANCD2^−/−) phenotype at a nominal P value < 0.01.

DNA extraction and mutation analysis

DNA was extracted from cells using DNA Blood and Tissue kit (Qiagen). PCR primers were designed using IDT Primer Quest to amplify regions around the mutation sites. Allele 1 Forward primer: 5’CAC GTT GTT GAG GAC AGT TCT A 3’, Reverse primer: 5’ CAG GGA TAT TGG CCT GAG ATT T 3’. Allele 2 Forward primer: 5’ CCT CTT CTA CCT CTA GGC AGT T 3’, Reverse primer: 5’ CCT TCA CGG ATA CCC ATC ATT C 3’. PCR was performed using Phusion HF Mastermix (ThermoFisher Scientific). Amplicons were cleaned with NucleoMag beads (New England Biolabs) and amplification was verified with gel electrophoresis. Sanger sequencing was performed on an ABI 3500xl instrument and raw data was analyzed using a mixed base detection protocol with a mixed base being called if the 2nd highest peak is at least 25% of the primary peak.

Cellular fractionation

Cells were harvested, centrifuged and lysed in cytoskeletal buffer (CSK) (10 mM PIPES pH 6.8, 300 mM sucrose, 100 mM NaCl, 3 mM MgCl₂, 1 mM EGTA, and 0.5% v/v Triton-X-100) for 10 min at 4℃. Samples were centrifuged again and chromatin fractions were transferred to a separate microcentrifuge tube. Remaining pellets were washed once with CSK buffer, lysed 2% SDS sample buffer and boiled for 15 min and sonicated for 10s at 10% amplitude using a Sonic Dismembrator model 500.

Metaphase chromosome analysis

FA-D2 (FANCD2^−/−) and FANCD2-complemented FA-D2 cells were plated, then the following day were left untreated or treated with MMC for 24 h. Prior to harvest, cells were treated with colcemid for 2 h. After harvesting cells were incubated with 0.075 M KCl for 18 min at 37 °C. Cells were then fixed in 3:1 methanol: glacial acetic acid. Finally, cell suspensions were dropped onto chilled slides and allowed to air-dry before staining with 2.5% Giemsa solution.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

J.B. and N.H. wrote the main manuscript; J.B., B.N., K.H., K.H., J.A., A.A., and F.D. generated and analyzed the data presented in this manuscript; S.C. and D.S. helped to analyze the data and contributed important reagants. All authors have reviewed the manuscript.

Funding

This work was supported by an American Society of Hematology Bridge Grant to N.G.H., National Institutes of Health/National Heart, Lung, and Blood Institute R01 HL149907 to N.G.H., and Rhode Island IDeA Network of Biomedical Research Excellence (RI-INBRE) grant P20GM103430 from the National Institute of General Medical Sciences (PI, B.P. Cho).

Data availability

The RNA-seq data generated in this study have been deposited to the NCBI Gene Expression Omnibus under accession number GSE313989.

Declarations

Competing interests

The authors declare no competing interests.

Ethical approval

Anonymized FA patient materials were collected under an IRB-approved protocol through the International Fanconi Anemia Registry (IFAR) at the Rockefeller University after obtaining proper informed consent. This study was approved by the University of Rhode Island Institutional Biosafety Committee.