ANKRD26 Coding Variants Presenting with Giant Platelets and a Predisposition to Myeloid Neoplasia

Mutations (MT) in the 5' untranslated region (UTR) of ANKRD26 (A26) are implicated in ANKRD26- related thrombocytopenia (A26-RT), an autosomal dominant disorder of mild to moderate thrombocytopenia (TP) often presenting in adulthood, although severe and pediatric cases are reported. Erythrocyte and leukocyte counts are normal to increased, with unremarkable morphology. Platelet (plt) size is usually normal, as with ETV6- and RUNX1-mutated TP. Together, A26, ETV6, and RUNX1 germline (GL) MT comprise a separate 2016 WHO category of myeloid neoplasms (MN) with GL predisposition and preexisting platelet disorders. Normal plt size separates A26-RT from other familial TPs with giant plts. No consistent morphological plt aberrations have been reported. Bleeding history is absent or mild, and while TP is not life threating, 8-10% of patients (pts) develop a MN, including a 30-fold increased risk for AML relative to the general population.

A26 is an inner membrane adaptor protein with 2 major domains: ankyrin repeats (ANKR) and coil-coil (C-C), both of which interact with signaling and cytoskeletal proteins. A26-RT MT are almost exclusively in the 5'UTR, however, rare A26 coding variants (A26-CV) are reported to segregate with familial TP. Still, some studies on A26-RT limit sequencing to the 5'UTR. As such, A26-CV are not well represented or described.

We performed whole-exome sequencing (WES) on 195 pts with MN seen at Cleveland Clinic between 2004 and 2012, and downloaded WES .bam files from online sources, totaling 653 MN cases. Using a standard pipeline for discovery of rare GL variants (<1% population frequency), we identified 22 pts with A26-CV (6 MDS, 3 MDS/MPN, 13 AML) with 13 unique variants (2 benign/likely benign, 1 variant of unknown significance, 10 not found) and a median (med) VAF of 47%. Half of the variants were within either ANKR or C-C domains. The occurrence of these variants in our cohort was 3% vs.1% of the healthy population (p=0.0001).

Complete clinical description was available for 8 A26-CV pts, all diagnosed with MDS or MDS/MPN at med age of 64. Two pts had antecedent bruising 1 year prior to diagnosis (dx). No prior bleeding was noted. Three pts had prior TP, two of whom had bicytopenia and pancytopenia. First degree family history (FH) was positive for cancer in 6 pts (75%), including 2 pts with FH of hematologic neoplasms. One pt had a 2^nd-degree relative with a non-malignant hematologic condition requiring transfusions.

At dx, cytogenetics were normal and complex in 2 pts each (25%). Deletions in chromosomes 5, 17, 20, and Y were observed in 1 pt each, and monosomy 7 in 2 pts. On bone marrow aspirate, 6 pts (75%) had dysmegakaryopoiesis, found in interstitial patterns and clusters. Mild to moderate dyserythropoiesis was observed in 5 pts (63%), and 5 pts had dysgranulopoiesis. The med blast percentage was 1.5% (range 0-8%), with 5 pts having hypercellular marrow, med 65% (range 40-95%). On corresponding CBC, pts were anemic (med hemoglobin of 9.5 g/dL), with the majority (n=5) showing signs of erythroid dysplasia, and also thrombocytopenic (med plt of 99 k/μL), with giant plts observed in 3 pts. Two pts had both giant plts and abnormal erythroid morphology. Hypo and monolobate, hyposegmented, and hypogranular forms were observed in the 3 lineages, as well as detached nuclear lobes, budding, segmentation, and nuclear-cytoplasmic dyssynchrony. On next generation sequencing, co-occurring SRSF2 and ASXL1 MT were observed in 3 and 2 pts, respectively.

In sum, we have identified 22 A26-CV in MN, suggesting a role in predisposition as with A26-RT. We have seen in our cohort that A26-CV pts present differently from those with A26-RT. They have a variable past medical history and limited FH of TP, are anemic, with multilineage dysplasia observed not just in bone marrow, but also on peripheral blood smear, especially in megakaryocytes and erythrocytes. This is not surprising, as A26 is expressed in both lineages. The presence of giant plts is noteworthy. The mechanism for hypomorphic A26-CV may differ from that of the A26 5' UTR, which increase A26 levels in late-stage megakaryopoiesis by abrogating RUNX1/FLI1 binding, leading to aberrant proplatelet formation. Given the plt size and presence of nuclear phenotypes, altered interactions with signaling and cytoskeletal proteins could be involved, and may represent a novel A26 phenotype. Further investigation and association of A26-CV with MN ontogeny is under way.