Genomic insights into inherited bone marrow failure syndromes: a single-center clinical study

INTRODUCTION

Inherited bone marrow failure syndromes (IBMFS) are a group of rare genetic disorders due to germline pathogenic variants associated with hematopoiesis^[1-3], telomere biology^[4,5], ribonucleic acid (RNA) maturation and processing^[6], ribosome biogenesis^[7,8], and deoxyribonucleic acid (DNA) repair and cellular maintenance pathways^[9]. Overall, the estimated incidence of IBMFS is approximately 65 per million live births^[10]. The phenotypic heterogeneity in IBMFS and variability in disease penetrance results in significant overlap in clinical and laboratory findings in various IBMFS and other genetic disorders^[11-15]. Hematological abnormalities, such as uni-lineage or multi-lineage cytopenias, are the predominant clinical features across all IBMFS and usually lead to the patient seeking medical attention. As some of these IBMFS have an increased risk of developing myelodysplastic syndrome, acute myeloid and lymphoid leukemias, as well as other solid tumors^[16-20], they may be diagnosed later in life by their primary hematology oncologist. Extra-hematopoietic features are diagnosed earlier in life when patients present with multiple congenital anomalies, dysmorphism, developmental delays, skeletal dysplasia or short stature^[14], while adults may present with recurrent infections, idiopathic pulmonary fibrosis or even cryptogenic liver cirrhosis^[12]. Although most IBMFS demonstrate classical Mendelian inheritance consistent with autosomal recessive, autosomal dominant, and X-linked patterns, a positive family history may be absent in up to 40% of patients that harbor a germline pathogenic variant in genes associated with IBMFS^[21].

Despite all these complexities in the clinical setting, the rapid development of next-generation sequencing (NGS) has accelerated gene discovery in the setting of IBMFS^[22-27]. The advent of massively parallel sequencing has not only revolutionized biomedical research but has now become more accessible in the clinical sector due to rapid turnaround time at a more affordable cost^[28]. This has led to various consortiums and professional bodies recommending NGS as an early diagnostic test in patients evaluated for IBMFS^[29-31]. The classification criteria of hematological malignancies have evolved to incorporate NGS testing for both tumor profiling using somatic panels as well as defining the co-existence of any germline variants that predispose to the development of malignancies, which includes genes associated with various IBMFS^[32-35].

While the progress in genomic diagnoses has been exponential, it is a double-edged sword. The current NGS technique is unable to differentiate between clonal hematopoiesis (CH) caused by somatic mutations within the blood and germline variants seen in IBFMS^[36,37]. The risk of a false positive result due to the presence of CH and that of a false negative result due to a rare phenomenon known as somatic revertant genetic rescue^[38] are both problematic. This understanding has resulted in the recommendation of NGS testing on cultured skin fibroblasts obtained through a skin punch biopsy as the optimal sample source for determining the presence of germline variants^[39]. Additional challenges include the different NGS methodologies used in different cohorts with variable numbers of genes evaluated and the interpretation of variants of unknown significance (VUS) or genes of uncertain significance (GUS)^[36].

The clinical importance of an early and accurate diagnosis of IBMFS cannot be over-emphasized. Early identification of patients with IBMFS avoids the use of immunosuppressive therapy (IST), which is not only unhelpful in this group of patients but has the risk of causing harm through cytopenia-related complications and delay in appropriate management^[12]. Mortality and morbidity have been shown to improve in patients with IBMFS who undergo hematopoietic stem cell transplantation (HSCT)^[40] through the selection of an appropriate donor, conditioning regimen^[41], and post-HSCT surveillance^[42-47]. The genomic classification of IBMFS not only impacts the patient’s own health and reproductive choices but also influences the health of their immediate relatives. In this study, we want to highlight the novelty of our approach, where the use of exome sequencing in younger patients revealed dual diagnoses in a subset of these patients.

METHODS

RESULTS

A total of 16 individuals from 15 families were included in the analysis for this study. Patient demographics are summarized in Table 1. Our results identified eight individuals with a telomere biology disorder (TBD) (50%). Interestingly, four patients (25%) had dual diagnoses. Two of these patients had multiple congenital anomalies and global developmental delays. An exome sequencing was performed and revealed two rare genetic disorders in each individual, which is summarized in Table 2. Two individuals from the same family (individuals 15 and 16) were diagnosed with hereditary breast and ovarian cancer syndrome (HBOC) and TBD due to a family history of breast cancer and a personal history of B cell acute lymphoblastic leukemia in individual 16. The telomere length measurement was normal in individual 16 using the flow-FISH methodology (summarized in Table 2).

Nine patients had genomic DNA extracted from cultured skin fibroblasts and only one patient had repeated culture failure (11%). This patient had germline genetic testing on peripheral blood during remission from her underlying diagnosis of acute myeloid leukemia.

Only one patient had a negative genetic test result. Two patients were found to have copy number variants (CNV). Variants in 22 genes were detected among those with an abnormal test result. Recurrent mutations were seen in the TERC, TERT, and RTEL genes and diagnostic for TBD. A total of 26 variants were identified in this patient cohort. Missense variants accounted for the majority of variants, with 20 identified (77%). The second most frequent variants found were those constituting frameshift or insertion/deletions (indels) variants, with 6 identified (23%). Lastly, 2 non-coding variants were identified (7%). Overall, the diagnostic yield was approximately 50% (8/16 patients). A summary of the diagnostic yield and number of variants detected per gene can be found in Figure 1. Eight patients had at least one VUS detected on clinical-grade genetic testing (50%), with one individual having 5 VUS reported. One patient (individual 8) was given the diagnosis of TBD due to short telomeres (1%ile) with myelodysplastic syndrome and pulmonary fibrosis and remains genetically uncharacterized. Another patient (individual 2) received a clinical diagnosis of DBA and remains genomically uncharacterized as well. Individual 12 was found to have a novel TERC variant (n.116C>G) and reclassified from a VUS to a likely pathogenic variant (LPV) based on the American College of Medical Genetics and Genomics and the Association for Molecular Pathology (ACMG/AMP) guidelines for variant interpretation^[48], given his clinical presentation of hypoplastic myelodysplastic syndrome with short telomere length of less than the first percentile using the flow-FISH methodology (PS3). These results are summarized in Table 3.

Figure 1. Number of genetic variants and diagnostic yield.

We found that patients with shortened telomeres had delta telomere lengths for lymphocytes and granulocytes less than -2.00. We also summarized these results in eight patients within our cohort and the relevant genes associated with TBD in Figure 2.

Figure 2. Delta telomere lengths and Telomere Biology Disorder related genes.

Clinical outcomes were significant for two deaths during this period, leading to a mortality rate of 12.5%. One was due to respiratory failure because of pulmonary fibrosis secondary to an underlying diagnosis of TBD. Another was due to relapsed acute myeloid leukemia. A full summary of the clinical phenotypes and referral indications are in the Supplementary Table 1.

DISCUSSION

Overall, the diagnostic yield of clinical-grade genomic testing in a selected population of patients referred for suspected IBMFS is approximately 50%. Although one of the limitations of this study is the small number of patients (n = 16), it is interesting to note that approximately 25% of patients may have dual diagnoses that were not apparent prior to genetic testing. We also identified patients at an earlier age through increased access to exome sequencing in patients presenting with multiple congenital anomalies, global developmental delays, and intellectual disability based on a recent ACMG guideline published in 2021^[51]. We also acknowledge the potential referral bias in this cohort as these patients were referred by other healthcare providers based on the presence of cytopenias concerning an underlying IBMFS, personal and/or family history of a hematological malignancy, solid tumor or extra-hematopoietic organ dysfunction associated with IBMFS, personal history of pre-existing cytopenia(s), and abnormal genetic test results in genes related to IBMFS. The lack of a standardized NGS panel with a fixed number of genes tested reflects this heterogeneity in this patient population and highlights the complexities in clinical care when patients receive genetic testing through other healthcare providers.

Some of the challenges encountered in the clinical setting included difficulty in obtaining a skin punch biopsy for patients who presented with acute myeloid leukemia (AML) and myelodysplastic syndrome (MDS) at diagnosis. Given the presence of somatic variants and the possibility of clonal hematopoiesis in peripheral blood, it is recommended that germline genetic testing be performed on non-hematopoietic tissue, such as cultured skin fibroblast^[52]. We overcame this challenge by performing the skin punch biopsy during the repeat bone marrow biopsy after induction treatment. For patients who have undergone a prior HSCT, skin punch biopsy was performed by the clinical geneticist in the outpatient clinic. In an ideal situation, this procedure could be performed when the diagnostic bone marrow biopsy is scheduled, especially in a young patient with a suspected germline predisposition to myeloid neoplasms. Other tissue sources such as eyebrow hair follicles, fingernail clippings, and buccal swabs have been evaluated as possible alternative sources of germline DNA in patients with MDS^[53] or AML^[54]. A recent practice resource by the National Society of Genetic Counselors has highlighted the caveats and potential pitfalls such as the risk of clonal hematopoiesis of indeterminate potential (CHIP) when using other sample sources like blood, saliva, or buccal swabs for germline genetic testing in patients with hematological malignancies^[55].

Another challenge encountered is the inability to send ancillary tests that would help clarify any VUS that is reported from genomic testing. One of the reasons that these tests were not able to be sent (e.g., telomere length or chromosomal breakage analysis) was leukopenia during active disease or because of active chemotherapy. Culture failure of skin fibroblasts was a rare event that occurred only in one patient (11%).

In addition to the challenges mentioned above, the VUS rate is approximately 50%, with eight patients having at least one VUS detected in their genomic test. Functional studies are required to clarify the pathogenicity of these VUS but are not always available, and the development of in vitro assays requires significant investment in terms of time, personnel, and funding. Novel methods to improve in silico algorithms that predict pathogenicity may include saturation mutagenesis-reinforced functional assays (SMuRF)^[56]. Innovative methods for telomere length measurement, which can be performed on non-hematopoietic tissues, including in patients post-HSCT, would be highly desirable and useful in the clinical setting, and some research groups have attempted to address this challenge using long-read sequencing technology^[57-59]. The use of multi-omics^[60,61] such as transcriptomics^[62], metabolomics^[63], and proteomics^[64], as well as novel disease models^[65] using induced pluripotent stem cells (iPSCs) or organoids^[66,67], have been used to clarify VUS in patients with undiagnosed diseases and should be used in the analysis of the remaining 50% of these IBMFS patients that remain genomically uncharacterized.

For patients with dual diagnoses, further studies have been planned to utilize multi-omics and novel cellular-based disease models to identify the pathomechanisms and interacting pathways in these patients with two rare disorders.

Finally, some patients were diagnosed genomically but did not exhibit the typical clinical phenotype associated with bone marrow failure given their young age. Early diagnosis in younger individuals is important to better understand these rare disorders and to direct screening and surveillance plans. Continued long-term follow-up is very important to understand the natural history of patients with these underlying genetic variants associated with IBMFS and their risk of solid and hematological malignancies.

CONCLUSION

Early diagnosis of an underlying IBMFS is important to reduce mortality and morbidity. While there is increased awareness among providers of the multi-organ involvement in IBMFS and the risk of malignancies, access to genetic evaluation and counseling may be limited or delayed in certain circumstances. We highlighted our institutional experience within the Division of Medical Genetics and Genomics and the challenges involved in the diagnosis and management of these patients who are suspected of having an underlying IBMFS. In this selected group of patients, we identified a genomic diagnosis in half of the patients while the remaining patients were genomically uncharacterized. Ancillary tests such as erythrocyte adenosine deaminase, telomere lengths, and chromosomal breakage analysis are very useful in clarifying variants of uncertain significance and should be performed early during the clinical evaluation of these patients. Although overall mortality is improving with better supportive care, curative options are still limited. Continued research is required in this field to better understand disease-causing coding and non-coding variants in patients with IBMFS - clarifying the pathogenicity of variants of uncertain significance, consequences of gene interactions in patients with dual diagnoses, and the underlying mechanisms of disease in this group of patients. Natural history studies, in combination with translational research, are required to improve the mortality and morbidity in these patients.