Expanding the phenotypic and genetic spectrum of radioulnar synostosis associated hematological disease

Congenital radioulnar synostosis (RUS) is a rare developmental abnormality involving fusion of the bones of the forearms (radius and ulna) preventing normal supination of the affected forearm and has been recognized in orthopaedic literature since the mid-1800s. It is apparent that there is a subgroup of patients with RUS who also present with hematological abnormalities ranging from thrombocytopenia to myelodysplastic syndrome (MDS). Here, we report on a series of seven families with RUS and diverse hematological defects ranging from single cytopenias to global bone marrow failure (BMF); some individuals had severe BMF in childhood whilst others only a modest single cytopenia well into adulthood. We identify three different germline variants (two novel) in the MDS1 and EVI1 complex locus (MECOM) in five of the families. The other two families remain uncharacterized. We have further expanded the range of genotypes and phenotypes associated with RUS and hematological disease.

RUS can occur with other abnormalities in the skeleton, heart, urinary tract, as well as aneuploid syndromes.1 RUS also occurs as part of the inherited BMF spectrum of diseases and is known as radioulnar synostosis with amegakaryocytic thrombocytopenia (RUSAT). It is usually characterized by thrombocytopenia which progresses to pancytopenia coupled with a congenital fusion of the radius and ulna. Heterozygous germline variants in two genes have been associated with RUSAT: HOXA11 (Homeobox A11)2 (RUSAT1, MIM#605432) and MECOM (MDS1 and EVI1 complex locus)3 (RUSAT2, MIM#616738). HOXA11 is involved in upper limb development in early embryogenesis as well as normal hematopoiesis and leukemogensis. MECOM is essential for the proliferation of hematopoietic stem cells and is expressed in developing limbs.4

We identified a cohort of nine cases from seven families with RUS and variable hematological disease. The clinical characteristics of these patients (including one we have previously described5) are detailed in Table 1. Radiographs of the radius and ulna fusions from families 4–6 (Figure 1), show all cases have a very similar radiological appearance. Using Sanger sequencing, we screened the coding exons of HOXA11 and MECOM in the index cases of these families. No coding variants were identified in HOXA11. However, we did identify coding variants in MECOM (NM_001105078) in seven individuals from five families. Families 1–3 shared the same variant (c.2248C>T; p.Arg750Trp: Figure 2A) as reported by Niihori et al.3 Novel variants were identified in families 4 and 5. These were c.2278G>C; p.Pro760Ala and c.2272G>A; p.Glu758Lys, respectively (Figure 2B). Based on genetic studies and clinical presentation, an autosomal dominant inheritance pattern was observed in families 1 and 5. From the information available it appears that the variants in families 2–4 were de novo which is in agreement with the initial report.3 After direct Sanger sequencing, families 6 and 7 (Figure 2C) remain uncharacterized despite having a similar presentation, suggesting further genes may be involved in this disorder. There are three reports in literature75 of patients presenting with pancytopenia and RUS with no indication of the underlying genetic cause being given, as well as a more unusual case of RUS-associated B-cell acute lymphoblastic leukemia8 where the BCR-ABL gene fusion is unlikely to explain the RUS (Online Supplementary Table S1). It is unlikely that a large germline MECOM deletion has been missed in families 6 and 7 as reports describe two cases where deletions encompassing some or all of MECOM result in thrombocytopenia but not the associated RUS.109 Further studies are needed in these two families to determine the underlying basis of the disease.

Table 1.Characteristics of cases in this study.

Figure 1.X-rays of the forearms showing proximal fusion of the radius and ulna. (A) Radiographs of both arms from the index case of family 4. (B) Affected mother of family 5 (generation III). (C) Index case of family 6 in which neither HOXA11 nor MECOM variants were detected. Arrows highlight proximal fusion of the radius and ulna.

Figure 2.Family trees of all cases and MECOM variants identified in this study. (A) Families 1–3 share the same heterozygous variant in MECOM. (B) Family trees and sequencing traces for families 4 and 5. (C) Families 6 and 7 do not have variants in either MECOM or HOXA11. Unless indicated by a genotype, a DNA sample was unavailable for sequencing. Representative sequence traces of both the wild type and the variant are shown. RUS: radioulnar synostosis; AA: aplastic anemia; MDS: myelodysplasia; BMF: bone marrow failure; M/T: macrocytosis and thrombocytopenia; RUS*: radioulnar synostosis is presumed from the clinical description; NAD: no abnormality detected; arrow indicates the index case. (D) Protein sequence from different species showing the conservation of the variants identified in this study (red arrows). The underlined residues highlight part of zinc fingers 8 and 9. H.sapiens (human) NP_001098548.2, M.musculus (mouse) XP_011247963.1, D.rerio (zebra fish) XP_017206701.1, X.laevis (African clawed frog) NP_001089139.1, D.melanogaster (fruit fly) NP_001260545.1, S.cervisiae (yeast) KZV08884.1. E - Schematic representation of MECOM encompassing both the MDS1 and the EVI1 components of the complex. All germline variants identified to date are mapped on to this diagram. Red arrows/text - variants identified in patients with RUS and hematological defects. Blue arrows/text – variants in patients with hematological defects but no RUS. (6) – total number of different families reported for this variant. Protein numbering is relative to the EVI1 component of the complex only (NP_001098548.2). NLS: nuclear localisation signal; CtBP: C-terminal binding protein domain.

Reviewing literature, the first disease gene to be associated with both RUS and bone marrow failure was HOXA11. The heterozygous variant c.872delA, p.Asn291ThrfsX3 was identified in six individuals from two families (Online Supplementary Table S1)2 and causes a premature truncation of HOXA11. In both families the variant was inherited from the father. Interestingly, both fathers showed skeletal abnormalities of the arm but neither had any hematological disease. All the affected children presented with thrombocytopenia from birth, with three out of the four undergoing BMT. Recently, several papers have been published identifying germline variants in MECOM in patients with various bone marrow failures either with or without the presence of RUS (Online Supplementary Tables S1 and S2).131132 From these reports, together with findings of this study, we are able to demonstrate that even with the same variant the age of onset and the severity of BMF is highly variable. We observe that the MECOM variant p.Arg750Trp is a recurrent change that can present with a variety of hematological defects in association with RUS. In five out of the six families with this variant (families 2 and 3 in this study, and cases described in references 3, 10, 12) the disease appears de novo, with presentation very early in infancy/childhood, and all cases required a BMT. The sixth case with p.Arg750Trp (index case family 1) shows dominant inheritance pattern and the hematological involvement progressed from aplastic anemia to myelodysplastic syndrome (MDS); this presentation of RUS and development of MDS in adulthood is similar to the family reported by Ripperger et al.12 even though the causative variant is different. Autosomal dominant inheritance was also seen in a second family in our cohort (family 5). The hematological abnormality in the affected cases from family 5 was variable: ranging from normal blood count, mild single cytopenia to severe pancytopenia, again demonstrating variability in severity and age of onset. Interestingly, there are reports of seven patients with germline heterozygous variants (including two deletions) affecting MECOM where there is no presence of RUS but the severe associated hematological disease required treatment with BMT.119 These patients form a distinct group as they were sporadic, de novo cases who presented with disease in infancy. The reason behind this difference in phenotype is unclear at present.

All the heterozygous MECOM variants reported in this study are highly conserved from humans to yeast (Figure 2D), cluster within 10 amino acids (aa750-760) and impact on either the highly conserved Cys2His2 zinc finger motif (zinc finger 8, aa733-755) or the adjacent linker motif (aa756-760). Preliminary data suggest levels of MECOM appear to be reduced in lymphoblastoid cell lines heterozygous for p.Glu758Lys (Online Supplementary Figure S1). Cys2His2 zinc fingers are one of the most common DNA-binding motifs found in eukaryotic transcription factors and usually occur in groups allowing for multiple tandem contacts along the DNA. EVI1 has 2 zinc finger domains; proximal domain contains 7 motifs and the distal contains 3 motifs (Figure 2E) and it has been shown that these 2 domains can have different functions.14 Studies removing the whole of the 8 zinc finger domain have shown that this can block granulopoiesis confirming this region is involved in hematopoiesis. All germline variants identified to date in MECOM are shown in Figure 2E. It is noteworthy that all variants associated with the co-presentation of RUS and hematological disease cluster in region spanning zinc fingers 8 and 9. The variants observed in patients with hematological disease without RUS (verified by x-ray)11 are more widely distributed throughout the complex, including two deletions that either remove the MDS1 part of the complex or the entire gene region.109 Further reports may help to clarify whether there is a definitive genotype-phenotype correlation for disease associated variants in MECOM.

In this study, we have genetically characterized patients from five out of seven families with RUS and hematological disease. As well as identifying novel pathogenic MECOM variants, we show that three of the families share the same previously identified MECOM variant. The differences in presentation have enabled us to expand the clinical phenotype from a severe global bone marrow failure requiring transplantation that presents very early in infancy to a syndrome that presents with variable hematological disease in adults. The disease can also be inherited in families in a dominant manner and in these cases the hematological involvement is more variable both in terms of severity and age of presentation. It is also notable that the hematological features can be very variable, even amongst affected members of the same family as exemplified by families 1 and 5. Lack of variants in MECOM and HOXA11 in two of the families, together with uncharacterized families in the literature, suggest there may be other disease genes to be identified. Also, in light of the number of BMF cases where MECOM variants can occur independently of the development of RUS, a genetic screen of MECOM should be undertaken in all new patients with BMF, particularly if there is no family history of disease. Finally, it has become clear that RUS-associated hematological abnormalities are frequently more global and variable, rather than being limited to thrombocytopenia. The description RUS-associated hematological disease (RUSHD) is therefore perhaps more appropriate than RUS-associated amegakaryocytic thrombocytopenia (RUSAT) which was initially coined for the disease features observed in patients with germline HOXA11 or MECOM variants.