Copy Number Variants Underlying Inherited Bone Marrow Failure Syndromes

Background. Inherited bone marrow failure syndromes (IBMFSs) comprise a genetically heterogeneous group of diseases with hematopoietic failure and varying degrees of physical malformations. The diagnosis of an IBMFS and categorizing the specific syndrome critically impact on clinical care; however, these are commonly challenging and rely on genetic testing. Since over 80 genes have been associated with IBMFSs and might be affected by different types of DNA aberrations, the best strategy to establish a diagnosis in a timely and cost effective manner is unknown. The aims of this study were to evaluate the role of genome-wide copy number variant (CNV) analysis in unraveling causal genetic alterations in IBMFS patients with unknown genotype and determine whether correlation exists between large CNVs and more severe phenotype.

Methods. Patients from the Canadian Inherited Marrow Failure Registry (CIMFR) who were genetically investigated were included in this analysis. Genetic and clinical data were extracted and analyzed. Mann-Whitney test and Fisher's exact test were used to assess statistical significance.

Results. Among 328 patients from the CIMFR who underwent molecular investigation, a causal genotype was identified in 185 cases (56.4%). 69 patients had genome-wide CNV analysis by SNP/CGH arrays, among which ten (14.5%) had positive results. In four out of ten cases who were genotyped by SNP/CGH array, genome-wide CNV analysis was critical for establishing the diagnosis. Among 308 patients who were tested for nucleotide-level mutations by either targeted gene analysis or next generation sequencing panels, casual mutations were found in 169 (54.9%). Three patients had compound heterozygosity for a CNV and nucleotide-level mutation.

To determine whether large deletions are correlated with more severe phenotype we included nine additional patients with causal CNVs whose genotype was identified by MLPA (n=1), targeted FISH (n=1), DNA-qPCR analysis (n=1), Southern blotting (n=1) or metaphase cytogenetics (n=5). The causal CNVs among patients in our cohort ranged from 0.02 to 145.5 Mb in size. The most common disease associated with causal CNVs was Diamond-Blackfan anemia (four patients).

Patients with CNVs tended to have significantly more non-hematological organ system involvement (p=0.03), developmental delay (mean=56% vs. 28%, p=0.03) and short stature (mean=67% vs. 40%, p=0.04) than patients with nucleotide-level mutations. The difference remained significant when we compared all patients with mutations that are predicted to result in truncation or lack of protein from the respective allele (large CNV, nonsense, and indel/ frameshift) to patients with mutations that are predicted to be hypomorphic or affect function (splicing, indel/ inframe and missense). There was no correlation between CNVs and the severity of the hematological disease.

Conclusions. Most patients with IBMFSs have nucleotide-level mutations. However, a significant proportion of patients without such mutations have large CNVs that are not efficiently detected by current nucleotide-level testing methods. Therefore, genome-wide CNV analysis should be considered in IBMFS cases, where nucleotide-level sequencing does not reveal the causal mutation. Patients with IBMFSs and large CNVs had more non-hematological organ system involvement, a higher prevalence of developmental delay and short stature. This might be related to an additional impact of the CNVs on other genes close to the affected IBMFS gene or the severe damaging effect of the CNVs.