Brca1 deficiency causes bone marrow failure and spontaneous hematologic malignancies in mice

Fanconi anemia (FA) is an inherited bone marrow failure (BMF) syndrome characterized by hypersensitivity to DNA crosslinking agents, congenital anomalies, BMF, and an increased risk of developing leukemia and solid tumors.¹  FA is caused by mutations in one of 17 genes that make up the FA DNA repair pathway.^2-4 

BRCA1 binds directly to several FA proteins, all of which are essential for normal hematopoiesis, but the effects of BRCA1 deficiency on hematopoiesis are unknown. Several lines of evidence suggest that BRCA1 may also be an important regulator of hematopoiesis. First, BRCA1 is highly expressed in hematopoietic tissues,^5,6  whereas its expression is lost in myeloid leukemias.^7,8  Second, overexpression of Brca1 in the bone marrow (BM) disrupts stem cell quiescence and differentiation.⁹  Finally, the first person ever reported to have biallelic BRCA1 mutations experienced unusually severe myelosuppression after exposure to the DNA crosslinking agent carboplatin,¹⁰  which provides direct evidence that BRCA1 is important in human hematopoiesis.

To test whether BRCA1 has a critical role in hematopoiesis, we created a conditional Mx1-Cre–induced Brca1 deficiency mouse model. We demonstrated that mice with homozygous deficiency of Brca1 in the BM have FA-like cytopenias, DNA crosslinking agent hypersensitivity, and susceptibility to BMF and hematologic malignancies (HMs). Furthermore, Brca1^−/− cells have a hematopoietic precursor functional defect characterized by a reduced capacity to form hematopoietic colonies in vitro and to reconstitute hematopoiesis in irradiated recipient mice, supporting a critical role for Brca1 in normal BM function.

All mice were housed in a pathogen-free barrier facility accredited by the Association for Assessment and Accreditation of Laboratory Animal Care and maintained under a protocol approved by the Institutional Animal Care and Use Committee. Mice with a floxed Brca1 allele¹¹  were bred with Mx1-Cre⁺ mice¹²  to generate Brca1^+/+, Brca1^+/−, and Brca1^−/− mice after injections of polyinosinic:polycytidylic acid (supplemental Figure 1A-C available on the Blood Web site). Targeted Brca1 deletion was confirmed by polymerase chain reaction and real-time polymerase chain reaction (supplemental Figure 1C-D). Mice were monitored daily for the development of tumors, and they had monthly tail vein bleeds to measure complete blood counts. Full necropsy was performed on any mouse with changes in appearance that indicated illness.¹³  Histopathology, flow cytometry, and spectral karyotyping were performed on isolated tissues. Detailed methods are included in the supplemental Data.

As early as 1 month of age, Brca1^−/− mice demonstrated macrocytic anemia (hemoglobin [Hb] levels, 10.3 ± 3.3 vs 12.5 ± 3 g/dL; P = .042 [Figure 1A]; mean corpuscular volume 66.3 ± 5.0 vs 55.4 ± 4 fL; P < 1 × 10⁻⁵ [Figure 1B]) and had lower mean total white blood cell counts compared with Brca1^+/+ mice (supplemental Figure 2).

Four (30%) of 13 Brca1^−/− mice developed a steady decline in their Hb levels, eventually causing lethargy and resulting in euthanasia (supplemental Figure 3A). Necropsy confirmed severe anemia (mean Hb, 6.95 ± 0.6 g/dL), and peripheral blood (PB) smears showed polychromasia with irregular red blood cell morphology that was not present in Brca1^+/+ mice (Figure 1C-D). BM sections were also hypocellular compared with age-matched Brca1^+/+ mice (Figure 1E-F and supplemental Figures 8 and 9), consistent with spontaneous BMF. Spectral karyotyping of whole BM from 2 BMF Brca1^−/− mice (Figure 1G-H and supplemental Figure 3B-C) revealed multiple cytogenetic abnormalities, resulting from 18 and 16 double-strand breaks detected in the 20 cells examined in each mouse, indicative of genomic instability.

By days 140 to 190, 6 additional Brca1^−/− mice (6 [50%] of 13) developed HMs, including 4 lymphomas, 1 acute myeloid leukemia, and one erythroleukemia (Figure 2A-B).^14,15 Trp53 sequencing identified DNA binding domain mutations in 2 of the lymphomas (Figure 2B). Histologic review of tissues from these mice demonstrated malignant cells circulating in the PB, disrupting hematopoietic organ architecture, and infiltrating nonhematopoietic organs (Figure 2C-J and supplemental Figure 9). Splenic cells isolated from the mouse with erythroleukemia were transplanted into 2 sublethally irradiated recipient mice, both of which developed erythroleukemia. Spectral karyotyping of malignant cells from the secondary transplant revealed multiple chromosomal translocations, a result of double-strand breaks (Figure 2K-L), indicative of genomic instability in a malignancy in this model as well. None of the Brca1^+/+ or Brca1^+/− mice developed a malignancy during 12 months of follow-up. These results demonstrate that lack of Brca1 within the BM leads to PB cytopenias, macrocytosis, spontaneous BMF, and susceptibility to HMs reminiscent of the hematopoietic defects observed in human FA.^16,17 

To investigate the etiology of these hematopoietic defects, we performed flow cytometry on single-cell suspensions from BM and spleen to assess for differences in specific hematopoietic cell populations. In the spleens of BMF mice, we found an accumulation of early red blood cell precursors at the basophilic erythroblast (CD71⁺, Ter119⁺) stage relative to Brca1^+/+ mice (P = .003) (Figure 1I-J) that was not present in their BM (supplemental Figure 3D), suggesting a niche effect. We then assessed for differences in cell populations in the BM of 2-month-old Brca1^−/− mice before the expected age of BMF or HM development vs age-matched Brca1^+/− or Brca1^+/+ mice. These experiments demonstrated an expansion of Lineage⁻ Sca-1⁺ c-Kit⁺ cells, a population enriched for hematopoietic stem and progenitor cells (HSPCs) (supplemental Figure 5), but no differences in the numbers of long-term hematopoietic stem cells, short-term hematopoietic stem cells, multipotent progenitors, or other committed progenitors (supplemental Figures 5-7).

Because most other single-gene FA mouse models demonstrated reduced HSPC numbers and function,^18,19  we tested the hematopoietic progenitor function of Brca1^−/− BM cells. By using methylcellulose colony-forming assays, we found that Brca1^−/− BM cells formed fewer hematopoietic colonies in vitro (44.4 ± 31.9) than Brca1^+/+ cells (200.3 ± 30.5; P = .004). Similarly, BM cells from 2-month-old Brca1^−/− mice were unable to reconstitute hematopoiesis in lethally irradiated congenic mice (supplemental Figure 4). These data demonstrate that although Lineage⁻ Sca-1⁺ c-Kit⁺ cell numbers are expanded, HSPCs in our model have decreased function as seen in other FA models.

Finally, to test Brca1^−/− BM cells for hypersensitivity to DNA crosslinking agents, a classic hallmark of FA, we performed colony-forming assays in the presence of increasing concentrations of mitomycin C. Brca1^−/− BM cells demonstrated increased sensitivity compared with Brca1^+/+ cells at all mitomycin C concentrations tested (Figure 1K). Taken together, these data demonstrate that Brca1 is critical for normal hematopoiesis and maintenance of genomic stability in the BM, and its deficiency results in an FA-like hematopoietic phenotype in mice.

Until recently, BRCA1 was not considered a true member of the FA pathway because the effects of homozygous deficiency of this gene in people were not known.³  Recently, 2 patients with biallelic BRCA1 mutations, each with 1 hypomorphic and 1 null allele, were reported. The FA-like phenotype in these patients included developmental anomalies and increased sensitivity of lymphocytes to chromosomal breakage with exposure to diepoxybutane,^4,10  leading to the designation of BRCA1 as the seventeenth FA gene, FANCS. However, neither of these individuals has the characteristic FA BMF phenotype, leaving the effects of BRCA1 deficiency on BM function unknown.

Our data provide direct evidence of an FA-like hematopoietic defect in Brca1-deficient BM, strengthening the evidence that BRCA1 is a bona fide FA-like gene with a critical role in hematopoietic function. Furthermore, these mice provide a model system in which to determine the effects of Brca1 haploinsufficiency, as seen in humans, on BM function as well as ready access to a renewable source of Brca1-deficient tissue for characterization of the complex functions of Brca1. Finally, because nearly all other single FA gene mouse models have not recapitulated the BM phenotype of FA,^18,20  further investigation of our model may provide novel insights into FA pathogenesis.

The authors thank Drs Amittha Wickrema and Ryan Duggan for their helpful suggestions.

This work was funded by National Institutes of Health grants No. P01 CA40046 (M.M.L.B.) and K12CA139160 (J.E.C.) from the National Cancer Institute, K08HL129088 (J.E.C.) from the National Heart, Lung, and Blood Institute, and by the Cancer Research Foundation (M.M.L.B., L.A.G., J.E.C.) and Cancer Center Support grant No. P30 CA14599 from The University of Chicago Medicine Comprehensive Cancer Center.

Contribution: A.V., M.M.L.B., L.A.G., and J.E.C. designed the research; A.V., S.A., R.M., J.L., G.R., A.A., M.W., R.D., and E.M.D. performed the research; A.V., S.A., R.M., G.R., B.N., E.M.D., M.M.L.B., J.W.V., L.A.G., and J.E.C. analyzed data; A.V. and J.E.C. wrote the manuscript; and all authors edited and approved the final manuscript.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Jane E. Churpek, The University of Chicago, 5841 S. Maryland Ave, MC2115, Chicago, IL 60637; e-mail: jchurpek@bsd.uchicago.edu.