Clinical Severity Correlates with in Vitro Residual Function of FANCB Missense Variants

Fanconi anemia (FA) is the most common inherited bone marrow failure syndrome, characterized by congenital anomalies, bone marrow failure and cancer predisposition. FA patients with mutations in FANCB exhibit X-linked inheritance pattern. FANCB patients have been frequently associated with a severe phenotype, resembling VACTERL association with hydrocephalus, although heterogeneity in these patients still exists and is not well understood. To determine the underlying pathogenesis of clinical heterogeneity, we investigated phenotype-genotype correlations in 19 patients from 16 families with FANCB pathogenic variants. We also investigated functional properties of five FANCB missense variants to determine their possible contribution to clinical manifestations.

Patients were accrued at the Rockefeller University Hospital in the International Fanconi Anemia Registry. Among the 16 families with FANCB pathogenic variants, three carry nonsense variants, five have indels, two with large deletions, one has an intragenic duplication, two with canonical splice site variants that should result in aberrant splicing, and five display missense variants. One patient carried both an indel and a missense variant. Patients with variants leading to FANCB truncations (including nonsense variants, indels, large deletions/duplications, and splice variants) tend to show shorter survival and earlier hematologic onset, than patients with missense variants. However, the differences did not reach statistical significance due to small sample size.

Functional properties of five missense variants were examined by overexpression of variant cDNA (HA-tagged) in FANCB-null fibroblasts, followed by assessment of FANCD2 ubiquitination, mitomycin C (MMC) sensitivity, FANCD2 foci formation and HA-FANCB localization. We found that expression of p.W479G and p.L676P FANCB variants leads to near-normal FANCD2 ubiquitination upon MMC exposure, and MMC sensitivity. Conversely, p.L43S-expressing cells behaved like FANCB-null fibroblasts. Expression of p.L329P and p.G750V resulted in low levels of FANCD2 ubiquitination upon MMC exposure and intermediate MMC sensitivity. Expression of p.L43S resulted in smaller-sized and lowest quantity of FANCD2 foci with predominant cytoplasmic localization of HA-FANCB, while other variants showed intermediate phenotype. We quantified the nuclear expression of HA-FANCB, which confirmed the lowest relative nuclear expression of p.L43S variant, suggesting a defect in nuclear localization necessary for the interstrand crosslink repair. We also assessed splicing defects in the patient primary cells by reverse transcription PCR. Majority of the transcripts in the patient carrying c.1435T>G (predicted to express p.W479G) variant showed aberrant splicing, exon 7 skipping, which should result in an out-of-frame truncated protein. Although the expression level of c.1435T>G encoding the missense variant p.W479G is lower than the splicing variant, its functional characteristics were near-normal, thus reflecting relatively milder phenotype in this patient. Overall patient survival correlated with residual function of their FANCB variants as assessed by in vitro assays described above.

In summary, we report clinical course and mutation types of 19 patients with FANCB pathogenic variants. We also report that FANCB missense variants can have varying degree of residual function, which correlates with patient survival. Our data suggest that variant-level analysis may be a useful prognostic marker of patient outcome. In vitro assays evaluating the residual function of proteins with missense variants may help to predict patient outcome.