Abstract

Background

Large granular lymphocytic (LGL) leukemia is a clonal disease of mature cytotoxic T- or natural killer (NK)-cells, which was recently characterized by gain-of-function somatic STAT3 mutations in 40-70% of patients. Most of the T-LGL leukemia patients harbor one major Vbeta clone corresponding even up to 90% of total CD8+ T-cell population. Interestingly though, in a small proportion of T-LGL leukemia patients we have detected multiple mutations in the STAT3 gene suggesting the presence of subclones. Here, we aimed to study the clonal architecture and mutation spectrum of expanded lymphocytes with deep sequencing method and to follow the clones during immunosuppressive treatment.

Methods

DNA samples from 228 LGL leukemia patients were available for STAT3 mutation analysis. Additional flow cytometry-based sorting was done from 12 STAT3 mutation-positive patients, 6 of which had multiple STAT3 mutations in peripheral blood mononuclear cells (PBMNCs). First, frozen live PBMNCs were sorted into CD4+, CD8+ T-cell, and NK-cell fractions using antibodies for CD3, CD4, CD8, and CD16/56. Then CD8+ T-cell population was further sorted into clonal/non-clonal cells based on the flow cytometry analysis of T-cell receptor beta chain expression (Vbeta). STAT3 exon 21 was sequenced using Illumina Miseq platform with coverage aim over 10,000. The data was analyzed using an in-house bioinformatics pipeline: mutations were considered to be true if variant allele frequency (VAF) was over 1%, and false-positives were filtered out by comparing the VAF with calculated error rate of the amplicon.

Results

In total, 12/228 patients had multiple STAT3 mutations (16% of all STAT3 mutation positive cases). In all studied patients with multiple mutations (Table 1, 4 cases presented), the VAF was 30-50% in the purified major Vbeta clone suggesting that the whole population belonged to the same clone harboring a heterozygous STAT3 mutation (the clone size can be estimated to be twice the VAF). In addition to the major clone, STAT3 mutations were also discovered in smaller Vbeta expansions and in some cases in the non-clonal CD8+ population (Table 1). Interestingly, one patient diagnosed with T-LGL leukemia did not have STAT3 mutations in the major Vbeta expansion (65% of CD8+ cells) but harbored a D661Y mutation with 38% VAF in the NK-cell fraction (Patient 5 in Table 1).

The follow-up samples during the treatment were available from 4 patients. In patients 1 and 2 (Table 1), the size of the clone was unchanged during 32 and 37 months follow-up despite of the treatment with methotrexate and cyclophosphamide. In patient 3 (Table 1), the Vb7.1+ clone carrying Y640F mutation decreased from 10% to 3% of CD8+ cells, whereas the Vb5.1+ clone (D661Y) was unchanged during the methotrexate treatment. In patient 5 complete remission was achieved with cyclophosphamide treatment and that was accompanied with the disappearance of D661Y-mutated NK-cells.

Table 1

Vbeta results and variant allele frequencies (VAFs) of STAT3-mutations. The size of the mutated clone is approximately 2xVAF in the case of heterozygous mutation.

Pat.
No 
Vbeta (% of CD8+) PB MNC CD3+CD8+Vb+ CD3+CD8+Vbneg CD3negCD16/56+ 
Vb13.6 (96%) D661Y 38%, D661V 1% D661Y 47% D661V 17% N/A 
Vb3 (90%) Y640F 42%, I659L 1% Y640F 49% I659L 20% N/A 
Vb7.1 (13%) Vb5.1 (3%) Y640F 1%, D661Y 2%, D661V 2% Vb7.1: Y640F 35%
Vb5.1: D661Y 33% 
Vb7.1neg: D661Y 6%, D661V 3%, N647Y 5%
Vb5.1neg: Y640F 1%, D661V 4%, N647Y 2% 
No mutations 
Vb16 (50%) Vb13.6 (11%) D661V 6% Vb16: D661V 41%
Vb13.6: D661Y 11% 
Vb16neg: D661Y 4%
Vb13.6neg: D661V 26% 
N/A 
Vb20 (65%) D661Y 6% No mutations No mutations D661Y 38% 
Pat.
No 
Vbeta (% of CD8+) PB MNC CD3+CD8+Vb+ CD3+CD8+Vbneg CD3negCD16/56+ 
Vb13.6 (96%) D661Y 38%, D661V 1% D661Y 47% D661V 17% N/A 
Vb3 (90%) Y640F 42%, I659L 1% Y640F 49% I659L 20% N/A 
Vb7.1 (13%) Vb5.1 (3%) Y640F 1%, D661Y 2%, D661V 2% Vb7.1: Y640F 35%
Vb5.1: D661Y 33% 
Vb7.1neg: D661Y 6%, D661V 3%, N647Y 5%
Vb5.1neg: Y640F 1%, D661V 4%, N647Y 2% 
No mutations 
Vb16 (50%) Vb13.6 (11%) D661V 6% Vb16: D661V 41%
Vb13.6: D661Y 11% 
Vb16neg: D661Y 4%
Vb13.6neg: D661V 26% 
N/A 
Vb20 (65%) D661Y 6% No mutations No mutations D661Y 38% 
Discussion

Our preliminary results provide evidence that the STAT3 mutations are not only restricted to the significantly expanded lymphocyte clone in LGL leukemia patients, but they can also be found in smaller subclones mimicking the situation in acute leukemia. The actual cause of the mutations is unknown, but the results suggest the presence of a strong initial immune activation, which predisposes existing lymphocyte clones to somatic mutagenesis during cell proliferation. Considering the effects of treatment on STAT3-mutated clones, the only complete remission seen was connected to the disappearance of the mutated clone, which warrants STAT3-inhibitor trials in the future.

Disclosures:

Porkka:BMS: Consultancy, Research Funding, Speakers Bureau; Novartis: Consultancy, Research Funding, Speakers Bureau. Maciejewski:NIH: Research Funding; Aplastic anemia&MDS International Foundation: Research Funding. Mustjoki:Novartis: Consultancy, Speakers Bureau; BMS: Consultancy, Speakers Bureau.

Author notes

*

Asterisk with author names denotes non-ASH members.