Fancb deficiency impairs hematopoietic stem cell function

Fanconi anemia (FA) is a genetic disorder characterized by bone marrow failure, variable congenital malformations and a predisposition to malignancies. FANCB (also known as FAAP95), is the only X-linked FA gene discovered thus far. In the present study, we investigated hematopoiesis in adult Fancb deficient (Fancb−/y) mice and found that Fancb−/y mice have decreased hematopoietic stem cell (HSC) quiescence accompanied by reduced progenitor activity in vitro and reduced repopulating capacity in vivo. Like other FA mouse models previously reported, the hematopoietic system of Fancb−/y mice is hypersensitive to DNA cross-linking agent mitomycin C (MMC), which induces bone marrow failure in Fancb−/y mice. Furthermore, Fancb−/y BM exhibits slower recovery kinetics and less tolerance to myelotoxic stress induced by 5-fluorouracil than wild-type littermates. RNA-seq analysis reveals altered expression of genes involved in HSC function and cell cycle regulation in Fancb−/y HSC and progenitor cells. Thus, this Fancb−/y mouse model provides a novel approach for studying the critical role of the FA pathway not only in germ cell development but also in the maintenance of HSC function.


Figure 1. Reduced HSC/P cells in Fancb −/y mice. (A)
Fancb deficiency reduces the HSC/P pool in Fancb −/y mice. Whole bone marrow cells (WBMCs) isolated from 8-week-old Fancb −/y mice or their male Wild-type (WT) littermates were subjected to flow cytometric analysis for LSK (Lin − Sca1 + c-kit + ) and SLAM (Lin − Sca1 + c-kit + CD150 + CD48 − ) staining. Representative plots (Upper) and quantification (Lower) are shown. Results are means ± standard deviation (SD) of three independent experiments (n = 9 per group). (B) Fancb mutation increases HSC cycling. Cells described in (A) were gated for the SLAM population and analyzed for the cell cycle using Hochest 33342/Ki67 staining. Representative plots (left) and quantification (right) are shown.
Results are means ± standard deviation (SD) of three independent experiments (n = 9 per group). (C) Fancb mutation decreases HSC quiescence. BM cells from WT or Fancb −/y mice were gated for SLAM population and analyzed for BrdU incorporation. Representative plots (left) and quantification (right) are shown. Results are means ± standard deviation (SD) of three independent experiments (n = 9 per group). (D) Fancb deficiency does not increase cell apoptosis. Cells described in (A) were gated for SLAM population and analyzed for apoptosis by Annexin V and 7AAD. Representative plots (left) and quantification (right) are shown. Results are means ± standard deviation (SD) of three independent experiments (n = 9 per group). littermates. Thus, there is no indication of anemia in these mutant animals under steady state, which is consistent with other FA mouse models 17-19,24,26 . Reduced HSC/P frequencies in Fancb −/y mice. Analysis of different cell compartments in the bone marrow (BM) of the Fancb −/y mice showed that although the total BM cellularity of Fancb −/y mice was comparable to that of WT littermates (Fig. 1A, Lower), Fancb deficiency caused a significant reduction in the frequencies of HSC/P cells (Lin − Sca1 + c-kit + ; LSK) and, importantly, this was also seen in the phenotypic HSC (Lin − Sca1 + c-kit + CD150 + CD48 − ; Signaling lymphocyte activation molecule; SLAM) 27 compartment (Fig. 1A). Thus, the Fancb −/y mice have a reduced HSC/P pool compared to WT controls at the steady state.

Absolute and differential WBC counts Characterization of red blood cells Plts
Decreased HSC quiescence in Fancb −/y mice. Since quiescence is an important feature of HSC homeostasis 28 , we next analyzed the cell cycle profile of Fancb −/y HSCs. Hochest 33342/Ki67 staining showed that there was a significant decrease in the number of quiescent (G 0 ) and an increase in the number of cycling (S/G 2 /M) SLAM cells in Fancb −/y mice compared to WT control animals (Fig. 1B). We also performed an in vivo BrdU incorporation assay to determine the proliferative status of HSCs in the BM. In line with the cell cycle data, the percentage of SLAM cells in S phase was significantly higher (33.98% ± 3.922 in Fancb −/y verse 22.65% ± 1.491 in WT, p = 0.0356) in Fancb −/y mice than WT animals (Fig. 1C). However, mutation of Fancb did not affect the viability of SLAM cells at the steady state, as analyzed by Annexin V/7AAD staining (Fig. 1D). These results suggest that the Fancb protein may play a role in HSC homeostasis, probably by maintaining quiescence.

Fancb −/y HSC/P cells show reduced CFU and repopulating ability. Increased cycling in Fancb −/y
HSCs may lead to HSC/P exhaustion. To test this notion, we first compared Fancb −/y and WT progenitor activity using the colony formation unit (CFU) assay. Similar to Fanca −/− and Fancc −/− mice 29-31 , LSK cells derived from Fancb −/y mice produced significantly fewer colony formation units than WT LSK cells when plated in methylcellulose supplemented with hematopoietic cytokines ( Fig. 2A). However, there appeared to be no bias in lineage differentiation in Fancb −/y HSC/P cells ( Fig. 2A, Right). Significantly, Fancb −/y LSK cells showed a marked decrease in serial replating activity compared to WT LSK cells ( Fig. 2A, Left), indicative of replicative exhaustion in the absence of stromal support. We next determined the hematopoietic repopulating ability of Fancb −/y HSCs by performing competitive bone marrow transplantation (BMT). We transplanted whole bone marrow cells (WBMCs) from Fancb −/y mice or their male WT littermates (CD45.2 + ), along with equal number of WBMCs from congenic BoyJ mice (CD45.1 + ), into lethally irradiated BoyJ recipients. Flow cytometric analysis demonstrated a reduced donor-derived chimera (CD45.2 + ) in the PB of the recipients transplanted with Fancb −/y cells compared to those transplanted with WT cells (Fig. 2B), which is again consistent with previous reports of impaired repopulating ability of HSCs derived from other FA mouse models [29][30][31] . However, Fancb mutation did not alter lineage differentiation, as we observed similar percentages of donor-derived cells stained positive for CD3ε , B220 and Gr1/Mac1 in the recipients transplanted with either Fancb −/y or WT cells (Fig. 2C). Furthermore, serial BMT transplantation confirmed a long-term repopulation defect of Fancb −/y HSCs (Fig. 2D). In corroboration with the cell cycle data from native mice (Fig. 1B,C), we found a significantly decreased proportion of quiescent cells in donor-derived Fancb −/y LSK cells as compared to WT donor cells (Fig. 2E). Reduced short-and long-term engraftment suggests that Fancb deficiency may affect both stem and progenitor cells, perhaps by disrupting a competitive advantage at the early stage of engraftment. To test this notion, we performed a whole-bone-marrow (WBM) homing assay 32 , and found that Fancb deficiency decreases homing efficiency (albeit statistically not significant) of stem/progenitor-enriched (both Lin − and Lin − c-Kit + ) cells in the BM of the recipients 16 hours post-transplant (Fig. 2F). Taken together, these results indicate a crucial role of FANCB in maintaining HSC/P function.
Fancb −/y HSC/P cells are hypersensitive to MMC. BM failure is common in FA patients but does not occur spontaneously in FA mouse models. It has been shown that the DNA cross-linker mitomycin C (MMC) can induce BM failure in Fancc −/− mice 18 . We thus examined whether MMC can also induce BM failure in Fancb −/y mice. Because the hematologic phenotype of Fancb −/y mice is similar to other FA mouse models and that FANCA mutation is most predominant among FA patients 5 , we included the Fanca −/− mice for comparison. Eight weeks after chronic exposure to a low dose of MMC (0.3 mg/kg) 18 , Fancb −/y mice showed severe pancytopenia, as evidenced by significantly decreased RBC, Hb and PLT (Fig. 3A), and BM cellularity (Fig. 3B). In addition, MMC induced marked reduction in LSK and SLAM cells in the BM of Fancb −/y mice, as compared to WT controls (Fig. 3C). Similar to the Fanca −/− mice, all MMC-treated Fancb −/y mice died within three weeks due to BM failure (Fig. 3D). These results indicate that the hematopoietic system of Fancb −/y mice is hypersensitive to the DNA cross-linking agent MMC. This is consistent with the shared MMC sensitivity of both mouse and human FA cells to MMC in vitro [5][6][7][8]18 .
Fancb −/y mice are hypersensitive to 5-FU. To further examine the impact of the loss of FANCB on stressed hematopoiesis, we injected Fancb −/y mice and their male WT littermates, along with the Fanca −/− mice as a comparative FA model, with the myeloid-ablating agent Fluorouracil (5-FU) and monitored BM recovery over a period of 30 days 28 . It is known that administration of 5-FU induces hyper-proliferation and exhaustion of HSCs 33 . We found a similar drop of the white blood cell (WBC) count at the first one week after 5-FU injection in both WT and Fancb −/y mice (Fig. 4A). However, WBC recovery in WT mice started as early as 10 days after 5-FU treatment; whereas the recovery of WBC counts in Fancb −/y mice persistently lagged as compared to WT mice (Fig. 4A). Of note, Fanca −/− mice showed similar lagging recovery pattern (Fig. 4A). Concomitantly, a significantly higher fraction of SLAM cells from Fancb −/y mice were found in the S/G 2 /M phase of cell cycle (Fig. 4B).

Fancb deficiency alters expression of genes involved in stem cell function and cell-cycle regulation.
To explore the molecular mechanisms underlying the compromised HSC function observed in Fancb −/y mice, we performed RNA sequencing (RNA-seq) analysis using LSK cells isolated from Fancb −/y mice and their male WT littermates. Sequencing data were aligned using Tophat and the mm9 version of the mouse genome. Using GeneSpring GX analysis, we found 2753 unique differentially expressed genes (Fig. 5A), of which 973 were up-regulated and 1780 were down-regulated in Fancb −/y LSK cells as compared to WT cells (Moderate T-test, FC ≥ 2.0).
To further characterize the differentially expressed genes in Fancb −/y HSPCs, we performed pathway analysis and found that the affected pathways in Fancb −/y HSPCs included Pluripotency (86 transcripts), cell cycle regulation (85 transcripts), Wnt signaling pathway (54 transcripts), and the Delta-Notch pathway (79 transcripts). Furthermore, the pathway analysis highlighted the role of Fancb in DNA replication (41 transcripts) and oxidative damage (17 transcripts) (Fig. 5B). Further classification based on functional annotation revealed that several differentially expressed genes in Fancb −/y LSK cells are involved cell cycle control (Cdc25c, Ccnb1, Chek1, Ccne1,  Mcm4, Mcm2), consistent with the increase in cycling HSCs observed in Fancb −/y mice. Another group of genes with altered expression was those involved in HSC function (Wnt10a, Wnt16, Wnt3a, Fzd1, Fzd5, Fzd8, Prkcb) (Fig. 5C). We validated the RNA-seq results by quantitative RT-PCR (qPCR) for selected genes. Consistent with our phenotypic findings (Figs 1 and 2), many important cell cycle control genes, such as such as Cdc25c, Ccnb1,  (Fig. 5D, Upper) and Wnt/pluripotency genes, such as Wnt5a, Fzd8, Prkcb, Wnt3a, Wnt6 and Fzd5, are deregulated in Fancb −/y LSK cells (Fig. 5D, Lower). Thus, these transcriptional changes may link dysregulation of multiple pathways to impaired HSC function in Fancb −/y . Finally, we performed rescue experiments to address whether the transcriptional dysregulation of the cell-cycle related genes in Fancb −/y LSK cells was associated with the observed increase in HSPC cycling. We chose Cdc25c and Ccnb1, both of which are critical regulators of cell entry into mitosis [34][35][36] , for further study. We first acutely lowered the expression of Cdc25c and Ccnb1 in LSK cells, using lentiviral shRNA. We found that reducing the level of either Cdc25c or Ccnb1 mRNA in Fancb −/y LSK cells to near WT level (Fig. 5E, Right) significantly increased (albeit not completely rescued) quiescence of the Fancb −/y LSK cells (Fig. 5E, Left), indicating that dysregulation of these two cell-cycle regulators were at least partially responsible for the decreased quiescence of Fancb -/ HSCs. To  examine if knock-down of Cdc25c and Ccnb1 also rescued Fancb −/y HSC self-renewal, we performed in vitro serial replating assays using the shRNA-transduced LSK cells. The downregulation of Cdc25c or Ccnb1 in Fancb −/y LSK cells significantly increased colony numbers in first plating and, to a greater extent, in second and third platings (Fig. 5F). Thus, genetic correction of Cdc25c and Ccnb1 expression improves Fancb −/y HSC self-renewal capability.

Discussion
FA patients suffer BM failure, presumably resulting from depletion of HSCs. Here, we have shown that mice mutated in the Fancb gene exhibit defective HSC maintenance. Several observations support this conclusion: 1) Fancb −/y mice have a reduced HSC pool size at steady state; 2) Fancb mutation compromises repopulating capacity of HSCs due, at least partially, to increased HSC cycling and premature stem cell exhaustion; 3) Fancb −/y HSPCs are hyper-sensitive to genotoxic and myelotoxic stresses; 4) Fancb mutation deregulates the expression of genes involved in stem cell function and cell cycle control pathways. This novel mouse model adds to the utility of FA mice for understanding the in vivo functions of the FA proteins.
Under steady state, FA mice, including Fanca −/− , Fancc −/− and Fancg −/− mice, fail to recapitulate the anemia phenotype of FA patients [17][18][19]24,26 . Consistently, our attempt to generate anemic mice by mutational disruption of the Fancb gene has not resulted in a model characteristic of the human disease. In fact, Fancb −/y mice show only minor hematological parameters characteristic of FA patients (Table 1). Although the mechanism behind this is still lacking, possible causes are suggested by the scientific literature. For example, a recent study suggested that mice harbor longer telomere lengths than patients with FA 37 , which might protect hematopoietic cells from senescence. Alternatively, the mild phenotype in mice could be due to fewer spontaneous DNA lesions or lower levels of endogenous DNA-damaging agents, such as formaldehyde 38 or malondialdehyde 39 .
Although spontaneous BM failure has not been observed in FA mice, a defect in HSC function has been observed in FA mice using competitive repopulation and serial BM transplant studies 19 , [29][30][31][40][41][42][43] . Indeed, we found that loss of Fancb protein compromises HSC repopulation ability and increases proportion of cycling cells in the SLAM compartment, which eventually leads to premature stem cell exhaustion (Fig. 2). Moreover, Fancb −/y HSCs in lethally-irradiated recipients are less quiescent and are rapidly exhausted upon 5-FU challenge. This correlated with a significantly increased mortality of Fancb −/y mice (Fig. 4). These results are also consistent with the cell cycle-related HSC defects in Fancc −/− mice reported by Li et al. 31 . It is in this context that we identify FANCB as a critical regulator that prevents HSCs from myelotoxic stress-induced cycling and excessive proliferation.
It is well-accepted that DNA repair is essential for the maintenance of hematopoietic function. In fact, mice defective in different mechanisms of genome maintenance, including homologous recombination (HR), non-homologous end joining (NHEJ), nucleotide base excision (NER), mismatch repair (MMR), DNA interstrand crosslink repair (ICL), as well as telomere maintenance, have hematopoietic stem/progenitor cell defects [44][45][46][47][48][49] . Our data on the loss of HSC function in Fancb −/y mice further supports the importance of genome stability in maintaining the long-term repopulating capacity of HSCs. The impaired regenerative potential of the Fancb −/y bone marrow may be due to a reduced HSC reserve, functionally defective HSCs, or a combination of the two. In support of a possible role for FANCB in the pathways regulating the balance between stem cell quiescence and cell-cycle entry, our RNA-seq data identified significant difference in gene expression of several particular genes in Fancb deficient HSPCs, including those related to cell cycle control (Cdc25c, Ccnb1, Chek1, Ccne1, Mcm4, Mcm2) and HSC function (Wnt10a, Wnt16, Wnt3a, Fzd1, Fzd5, Fzd8, Prkcb) (Fig. 5). Furthermore, our rescue experiments show that knocking down Cdc25c or Ccnb1 mRNA in Fancb −/y LSK cells to near WT level significantly increased quiescence and self-renewal capability of the Fancb −/y HSCs cells.
In summary, the present study demonstrated that inactivation of Fancb in mice induces premature HSC exhaustion and identify excessive HSC cycling as one of the underlying mechanisms for the defect. These findings reveal functional interaction between the FA DNA repair pathway and cell-cycle progression in HSC maintenance. To our knowledge, this is the first study that describes X-linked hematopoietic defects associated with FA.
Treatments. 5 into lethally irradiated BoyJ recipients. Hematopoietic reconstitution in recipient mice by donor (CD45.2 + ) cells at 8 and 16 weeks post transplantation was determined by staining for CD45.1-PE and CD45.2-FITC markers followed by flow cytometry analysis with a FACSCanto I (BD Biosciences, San Jose, CA). For non-competitive bone marrow transplantation (BMT), 1,000 LSK cells from WT mice or 2,000 LSK cells from Fancb −/y mice were transplanted to lethally irradiated BoyJ recipient to establish similar chimera. For 2 nd BM transplantation, 3 million BM cells from primary recipients were injected to lethally irradiated BoyJ recipients.
Mouse competitive homing assay. Mouse competitive homing experiments were performed as described 32 .
In brief, BM cells from Fancb −/y or their male WT littermates (CD45.2 + ) were labeled with DiO dye at a density of 2 × 10 6 cells per ml at 37 °C for 30 min. In parallel, BM cells from BoyJ mice (CD45.1 + ) were labelled with DiD dye (1:200) at 37 °C for 30 min. DiO and DiD were purchased from Vybrant Multicolor Cell-Labelling Kit (Molecular Probes, V-22889). After labelling, dyes were washed off. The DiO-labelled CD45.2 and DiD-labelled CD45.1 cells were mixed at a 1:1 ratio and transplanted into lethally irradiated (11 Gy one day before transplantation) CD45.1 recipients (2.5 × 10 6 from each donor). Sixteen hours after transplant, the recipients were euthanized and the BM was analyzed by flow cytometry for both DiO/DiD and surface lineage markers (Gr1, Mac1, B220, CD3, Ter119, from Ebioscience) and c-Kit (2B8, BD Biosciences). The ratio between the percentages of DiO + (donor) and DiD + (competitor) cells within different cell populations was quantified. For the BrdU incorporation assay, Bromodeoxyuridine (BrdU, 150 μ l of 10 mg/ml) were intraperitoneally (i.p.) injected to subjected mice followed by BM cells isolation 14 hours later. BrdU incorporated cells (S phase) were analyzed with the APC BrdU Flow Kit (BD Biosciences, San Jose, CA), following the manufacturer's instructions. Briefly, cells were surface stained then fixed and permeabilized using BD Cytofix/Cytoperm Buffer. After 1 hour incubation with DNase at 37 °C, cells were stained with APC-conjugated anti-BrdU monoclonal antibody. 7-aminoactinomycin (7-AAD) was added to each sample right before Flow Cytometry analysis (BD Biosciences, San Jose, CA).
For cell sorting, lineage negative cells were enriched using lineage depletion columns (StemCell Technologies, Vancouver, BC, Canada) according to the manufacturer's instructions. The LSK (Lin − c-Kit + Sca-1 + ) population was acquired by using the FACSAria II sorter (BD Biosciences, San Jose, CA).
Colony-forming unit assay. LDBMCs isolated from Fancb −/y mice or their male WT littermates were plated in a 35-mm tissue culture dish in 4 mL of semisolid medium containing 3 mL of MethoCult M3134 (Stem Cell Technologies, Vancouver, BC, Canada) and the following growth factors: 100 ng/ml SCF, 10 ng/ml IL-3, 100 ng/ml GM-CSF, and 4 units/mL erythropoietin (Peprotech, Burlington, NC). On day 7 after plating, erythroid and myeloid colonies were enumerated. For serial plating, cells from primary or secondary CFU assays were pooled and re-plated to evaluate secondary or tertiary CFUs, respectively. Hematopoietic clonal growth results were expressed as means (of triplicate plates) ± SD of three experiments. RNA sequencing. For RNAseq analysis, RNA was extracted from LSK cells isolated from Fancb −/y mice or their male WT littermates by Trizol-type method. The RNA quality and quantity assessment, single-end sequencing and alignment of reads on mouse genome (mm9 version) were carried out in the Cincinnati Children's Hospital Medical Center Affymetrix Core using standard procedures using Illumina HiSeq2000. Bam files provided by the Core were analyzed using Genespring GX v12 (Agilent Technologies, Santa Clara, CA). Quantification of mRNA expression was done using the RefSeq database. Gene expression and fold change (FC) were evaluated between WT and Fancb −/y samples and genes with a significant difference (moderate-T test, p ≤ 0.05) and a FC ≥ 2.0 were selected to run the pathway analysis module of GeneSpring GX v12 using the curated WikiPathway database(http:// www.wikipathways.org).
Statistics. Data were analyzed statistically using a Student's t test. Statistical measures stated in the text were based on the p values. p values < 0.05 were considered statistically significant.