Mutational Landscape of SH2B3-Mutated Hematologic Malignancies

Background:

SH2B3 acts as a negative regulator of cytokine signaling, cell proliferation and is involved in myelopoiesis. Mutations are known to occur both somatically as well as germline. Thus far, effects of these mutations have been mostly studied in the context of myeloproliferative neoplasms (MPN) and acute lymphoblastic leukemia (ALL). However, mutations have also been identified in other myeloid neoplasms, including MDS/MPN and MDS. In the present study, we sought to identify and characterize the mutational spectrum of SH2B3-mutated hematological malignancies across different entities.

Methods:

Our database was mined for cases with variants in SH2B3 detected by Next Generation Sequencing. Out of 28,563 cases with confirmed hematologic neoplasms, and information for variants in SH2B3, we identified 621 cases that showed a variant in the SH2B3 gene. In order to characterize the mutational landscape of these SH2B3-mutated cases, further 52 genes which are frequently associated with myeloid neoplasms were analyzed.

Results:

Variants in SH2B3 were identified in cases from all entities (n total = 621, grouped into 10 subgroups: 5 acute leukemia, 15 ALL, 101 AML, 4 CML, 67 CMML, 188 MDS, 41 MDS/MPN, 153 MPN, 40 cases lymphatic neoplasms other than ALL and 7 other). The incidence of mutated SH2B3 was calculated from a smaller representative subcohort and was 3.9% in total (216/5568). It was highest in patients with CMML (8%, 27/338), MDS/MPN (5.4%, 11/204) and CML (5.6%, 4/71) and lowest in lymphatic entities (2.8%, 13/470). No difference in age, sex, Hb levels or bone marrow blasts was detected for any entity comparing cases with mutated (mut) and wildtype (WT) SH2B3. ALL cases with mutated SH2B3 however showed an increase in white blood cell counts (mut: 122,355 /µL; WT: 46,833/µL; p<0.001). Intriguingly, CMML cases with mutated SH2B3 showed a correlation with WHO subgroups of CMML, with a higher frequency of mut cases being classified CMML-1 compared to WT (mut: 67%; WT: 49%, p<0.05).

In total, 669 variants in SH2B3 were identified in 621 patients, with 1-7 variants per patient. The most frequent variations in SH2B3 across all entities were missense (n=452, 67%), followed by frame-shift (n=137, 20%), and nonsense (n=48, 7%) variants. Splice site (n=24, 3%) and in-frame variants (n=8, 1%) were less abundant. The distribution of types of variations was similar across most entities, with CMML cases showing higher frequencies of frame-shift (32%) and nonsense variants (14%) compared to other entities (p<0.05). The median variant allele frequency (VAF) of variants in SH2B3 was 42% (range: 1.3-100%). Overall, 42% (n=281) of the variants were classified as pathogenic/likely pathogenic (Median VAF: 34%), 10% (n=70) were characterized as functionally uncertain but likely somatic based on VAF (VAF: 7%) and 47% (n=318) as variants of unknown significance (VUS) (VAF: 44%). Of note, pathogenic variants were detected in all entities, other than the four CML cases. ALL and CMML cases showed the highest frequency of pathogenic variants (65% and 53% respectively). Interestingly, variants were distributed across the whole gene and did not cluster in the functional domains, irrespective whether the variant was classified as VUS or pathogenic.

On average patients with SH2B3 variants showed 3.9 co-mutations (range 1-14), with patients with CMML and AML harboring the most changes (on average 5.8 and 4.6 mutations, respectively). In 75/621 cases a variant in SH2B3 was the sole variation of all examined genes. Overall the 5 most frequently mutated genes other than SH2B3 were TET2 (n=332), JAK2 (n=157), ASXL1 (n=151), SRSF2 (n=111) and DNMT3A (n=84). In all myeloid entities, TET2 mutations occurred most often, whereas in ALL patients, NOTCH1 was the most frequently mutated gene.

Conclusion:

Here we show a comprehensive characterization of SH2B3 mutations in patients with diverse hematologic malignancies. This study revealed that pathogenic mutations in SH2B3 occur in almost all hematologic neoplasm. Given a high prevalence of pathogenic variants and an association with lower risk subclasses in CMML, this study suggests a role of SH2B3 in this specific entity that needs further investigation.

Shoumariyeh:Blueprint: Honoraria. Haferlach:Abbvie: Consultancy, Honoraria.