Biallelic Deleterious Germline SH2B3 Variants Cause a Novel Clinical Syndrome of Myeloproliferation and Multi-Organ Autoimmunity

SH2B3 is a negative regulator of cytokine receptor signalling pathways in hematopoietic tissue. Acquired deleterious SH2B3 variants resulting in increased JAK/STAT signalling are observed in myeloproliferative neoplasms (MPN) however the clinical implications and phenotype of deleterious germline variants in SH2B3 are less clear. A single consanguineous kindred has been previously described with homozygous deleterious germline SH2B3 variants presenting with hepatosplenomegaly, autoimmune hepatitis, developmental delay, autoimmune thyroiditis and acute lymphoblastic leukemia (Perez-Garcia, A., et al, Blood, 2013), however a clinically valid gene-disease pair according to formal criteria has not been met to date for SH2B3.

We identified two probands from unrelated kindred with germline homozygous deleterious SH2B3 variants. The first patient presented with splenomegaly, thrombocytosis, neutrophilia and a leukoerythroblastic blood film at 3 months old. He developed alopecia areata at 2 years and autoimmune hypothyroidism at 6 years. He was otherwise developmentally normal without any detected liver abnormalities or thrombotic episodes. Whilst his splenomegaly and neutrophil count gradually improved throughout life, his platelet count remained markedly elevated (~1000x10⁹/L at 18 years of age). Targeted NGS identified a homozygous frameshift variant in SH2B3 (NM_005475.3:c.441_468del; p.(Arg148Profs*40)). Both parents were heterozygous carriers of this variant. The second patient was a male born to consanguineous parents and noted shortly after birth to have isolated splenomegaly. He was diagnosed with autoimmune hypothyroidism and Raynaud syndrome at 12 years and was noted to have a persistent thrombocytosis ranging from 550 - 780 x10⁹/L. He developed autoimmune hepatitis and insulin dependent diabetes at 17 years and around the same time he also suffered a right-sided middle cerebral artery territory infarct with a right carotid artery thrombus. Whole genome sequencing detected a homozygous missense variant in SH2B3 (NM_005475.3:c.1204G>A; p.(Val402Met)) in the patient with both parents being heterozygous carriers.

In order to understand these variants further we modelled the human p.(Arg148Profs*40) variant from patient 1 in zebrafish. CRISPR-Cas9 gene editing was targeted to the analogous region of zebrafish sh2b3 to create assorted carboxyl-truncating mutations in F0 crispants. On-target gene-editing was verified by Sanger sequencing and NGS. F0 crispants had a significantly increased number of macrophages and CD41-expressing cells (HSCs and thrombocytes), replicating the patient phenotype. Treatment of the sh2b3 crispant fish with the JAK inhibitor ruxolitinib intercepted the myeloproliferative phenotype. The myeloproliferative phenotype was not observed in stable deletion mutants, likely reflecting genetic compensation not present in crispants. Skin-derived fibroblasts from patient 2 who carried the Val402Met homozygous variant, showed increased phosphorylation of JAK2, STAT5 and STAT3 by immunoblotting after stimulation with IL3, GH, GM-CSF and EPO compared to healthy controls.

In conclusion, these newly described kindreds showing striking phenotypic similarity to each other as well as the previously described kindred and in conjunction with our functional data provide sufficient evidence for biallelic homozygous deleterious variants in SH2B3 to be considered a valid gene-disease association with a clinical syndrome of bone marrow myeloproliferation and multi-organ autoimmune manifestations. Notably, neither of our patients had evidence of acute lymphoblastic leukemia or overt hematological malignancy and further affected individuals will need to be described in order to fully understand the breadth of the phenotype. Moreover, our data supports ruxolitinib as being a potential therapeutic option for patients with this rare novel immune dysregulatory syndrome.