Abstract

T-large granular lymphocyte leukemia (T-LGLL) is a clonal lymphoproliferative disorder of cytotoxic T-cells (CTL) that is associated with cytopenias, predominantly neutropenia and reticulocytopenic anemia. From a scientific point of view, T-LGLL provides a natural model to study the dynamics of CTL responses; the heterogeneity of the disorder allows for examining the diversity of CTL responses in both autoimmune disorders and putatively chronic reactive conditions. A proportion of patients may have an extreme reactive process that mimics an indolent neoplastic lymphoproliferation. NGS and deep T-cell repertoire (TCR) sequencing provide insight into the clonal dynamics at work in T-LGLL patients.

A large proportion of T-LGLL patients present with a bona-fide low-grade leukemia; this notion is supported by the discovery of recurrent somatic STAT3 mutations in some patients. STAT3 clonal burden represents an excellent marker that can be serially monitored along with clinical milestones to ultimately gain a more comprehensive understanding of disease etiology and natural history.

We collected a cohort of 183 LGLL patients and screened them via deep NGS for mutation status of STAT3. In 36% of patients, 4 distinct somatic mutations (Y640F, N647I, D661V, D661Y) were identified in the SH2 domain of STAT3. In patients with wildtype STAT3, no somatic mutation was implicated in clonal expansion except for a small minority with STAT5 mutations present.

We performed a longitudinal analysis of 20 representative STAT3-mutated T-LGLL patients with up to 10-year follow-up and an average of 7 analyzed blood samples per case. All serial samples were deep-sequenced to detect and determine the VAF of the known STAT3 mutations. Overall, STAT3 mutation VAF had a significant, inverse relationship to both hemoglobin and absolute neutrophil count (ANC) (both p<=0.001). In 7/11 cases harboring the Y640F mutation, chemotherapy led to remission accompanied by a decrease in VAF; 3 were asymptomatic and received no treatment. In patients with D661V or D661Y, 6/9 achieved remission with treatment. Only 1/3 cases with N647I entered remission.

This longitudinal cohort can be sub-categorized into distinct patterns of clonal dynamics: 1) emerging STAT3 mutation in 20% of patients with a decrease in ANC as VAF of STAT3 clones expand; 2) an opposite trend in 40% of patients where VAF decreased due to therapeutic manipulations; 3) stable VAF in 20% of patients with little change over time; 4) codominant or dominant/secondary STAT3 mutations with distinct subclonal burden in 20% of patients.

We performed deep TCR NGS on a representative subset of 9 patients to explore how STAT3 mutations correlated with T-cell clonal expansions. The data were processed by an extensive bioanalytic pipeline to quantify the relative abundance of each CDR3 rearrangement within a patient's TCR. Our cohort had an average of over 53,000 CDR3 templates per sample and was compared with 587 healthy controls.

Our results demonstrate multiple patterns of clonal dynamics over the course of T-LGLL. Within each case, the immunodominant clones in serial samples were identified and correlated with STAT3 VAF burden over time. When patients were in remission, both STAT3 VAF and clonality were typically low. Interestingly, functional remission occurred in 2 cases despite increases in both clonality and STAT3 VAF. In 5/9 cases, the T-LGLL process involved 1 STAT3 mutation and 1 corresponding pathogenic clonotype displaying similar dynamics over time. In patients with 2 mutations, multiple high-frequency clonotypes were observed. Most significantly, comparison of STAT3 VAF and the dominant clonotype(s) revealed that STAT3 mutation can arise within a pre-existing clonal expansion that may harbor 2 branching mutations in extreme cases.

Identification of CDR3 rearrangement sequences allowed for analysis of the distribution of clonotypes among patients and controls. The pathogenic clonotypes found in T-LGLL patients were detected in a high proportion of controls but at extremely low frequencies. This suggests that these potentially autoimmune clones exist in normal individuals but are effectively suppressed. No pathogenic clonotypes were shared among disease patients. In sum, analysis of clonal dynamics suggests that STAT3 mutations can occur in the context of pre-existing oligoclonal responses and involve otherwise low-frequency clonal specificities.

Disclosures

Sekeres:Celgene: Membership on an entity's Board of Directors or advisory committees; Millenium/Takeda: Membership on an entity's Board of Directors or advisory committees. Carraway:Celgene: Research Funding, Speakers Bureau; Baxalta: Speakers Bureau; Incyte: Membership on an entity's Board of Directors or advisory committees; Amgen: Membership on an entity's Board of Directors or advisory committees; Novartis: Membership on an entity's Board of Directors or advisory committees. Mustjoki:Novartis: Honoraria, Research Funding; Pfizer: Honoraria, Research Funding; Bristol-Myers Squibb: Honoraria, Research Funding; Ariad: Research Funding.

Author notes

*

Asterisk with author names denotes non-ASH members.