Recurrent Missense Mutations in the STAT3 Gene in LGL Leukemia Provide Insights to Pathogenetic Mechanisms and Suggest Potential Diagnostic and Therapeutic Applications

T-cell large granular lymphocyte (LGL) leukemia is an uncommon lymphoproliferative disorder characterized in most cases by expansion of mature, clonal CD3+CD8+ cytotoxic T lymphocytes (CTLs). The pathogenesis of LGL-leukemia is unknown, and leukemic cells closely resemble normal terminally differentiated effector memory CTLs. While resistance to apoptotic pathways (Fas/Fas ligand, sphingolipid) and activation of survival signaling pathways (Ras) have been implicated in LGL leukemia, the underlying genetic defects have not yet been elucidated. We aimed to identify somatic mutations in LGL leukemia by whole exome sequencing of leukemic and matched healthy control cells.

Our index patient is a 70 year-old male with untreated CD8+ LGL leukemia diagnosed in 2009 with a clonal rearrangement in the T-cell receptor (TCR) delta and gamma gene. He has been asymptomatic with grade 2 neutropenia and an absolute lymphocyte count of 4–15 ×10⁹/L. The patient had one large predominant T-cell clone: 94% of CD8+ cells consisted of a single Vβ16 clone, as assessed by flow cytometry. No clonal expansions were observed in the CD4+ fraction. DNA was extracted from FACS-sorted CD8+ (leukemic) and CD4+ (control) cells and sequenced by exome capture using an Agilent SureSelect All exon 50 MB capture kit and the Illumina GAII sequencing platform. Candidate somatic mutations were identified with a bioinformatics pipeline consisting of BWA for sequence alignment, Samtools for alignment filtering and Varscan for somatic mutation calling. Mutations were manually reviewed in IGV for alignment artifacts and validated by capillary sequencing.

DNA samples from 8 additional untreated LGL-leukemia patients were used for further screening of confirmed somatic mutations by capillary sequencing. From six of these patients DNA was extracted from CD8 sorted cells and from two patients from whole blood.

Whole exome sequencing of CD8+ leukemic DNA from the index patient identified a missense mutation in the STAT3 gene (D661V), which was subsequently confirmed by capillary sequencing. As STAT3 signaling has been associated with LGL leukemia pathogenesis previously, we next designed primers for the secondary screening of the six exomes of STAT3 SH2 region from the remaining patients. Another recurrent somatic missense mutation (STAT3 Y640F) was identified in two additional patients. Thus, three out of nine LGL patients (33%) showed evidence of mutations in the STAT3 SH2 region. Both missense mutations found (D661V and Y640F) were located in the area of the SH2 domain known to mediate STAT3 protein dimerization and activation. The Y640F mutation alters a conserved tyrosine residue leading to a hyperactivating STAT protein (Scarzello et al. Mol Biol Cell, 2007) and was recently found in a human inflammatory hepatocellular adenoma causing cytokine-independent tyrosine phosphorylation and activation as well as cytokine-dependent hyperactivation of STAT3 (Pitali et al., J Exp Med, 2011). The D661V mutation has not been described previously.

Our data imply for the first time that STAT3 is a common mutational target in LGL leukemia, revealing insights to the molecular pathogenesis of this rare disease. Known structural and functional data on STAT biology imply that the mutations are leading to STAT3 hyperactivation and could also confer ligand-independent signaling. While confirmatory data from a larger series of patients are necessary, our results pinpoint STAT3 mutations and aberrations in the STAT3 pathway as key pathogenetic events in true clonal LGL leukemia. Detection of STAT3 mutations could therefore be applied in the diagnostic assessment, disease stratification and therapeutic monitoring of LGL patients.

Koskela:Novartis: Honoraria. Kuittinen:Roche: Consultancy. Porkka:Novartis: Honoraria; Bristol-Myers Squibb: Honoraria. Mustjoki:Novartis: Honoraria; Bristol-Myers Squibb: Honoraria.