Intersection between STAT3, Rheumatoid Arthritis, Large Granular Lymphocyte Leukemia, and Therapeutics

Large granular lymphocyte (LGL) leukemia, is a rare chronic lymphoproliferative malignancy that frequently co-occurs with rheumatoid arthritis (RA). LGL leukemia is characterized by clonal expansion of cytotoxic CD8+ T-cells, cytopenias such as neutropenia and anemia, as well as autoimmune manifestations. Treatment is initiated upon symptomatic disease presentation, and unfortunately, there are few effective therapeutic options in LGL leukemia. First-line therapies include general immunosuppressive agents like methotrexate (MTX) or cyclosporine. Over half of LGL leukemia patients harbor somatic activating mutations in signal transducer and activator of transcription 3 (STAT3), and nearly all patients have increased levels of STAT3 activation. Additionally, as high as 36% of patients may suffer from concomitant RA. Through ddPCR analysis of the most common STAT3 alterations, Y640F and D661Y, and any mutations in the S614-G618 region, we sought to better understand the relationship between RA, STAT3, and LGL leukemia. We discovered that 53% (52/98) of RA patients without LGL leukemia exhibit activating STAT3 mutations in their CD8+ T cells compared to only 11% (1/9) healthy donor controls (Chapter 2). The majority of these mutations were low variant allele frequency (VAF) (0.008-0.05%). However, 3/98 patients assayed displayed STAT3 mutations at a high enough frequency to be indicative a clonal expansion (VAF >5%). In RA patients, these mutations also correlated to clinical disease parameters like anti-cyclic citrullinated peptide (CCP) positivity.

In LGL leukemia patients, we also sought to harness the capabilities of ddPCR to identify low frequency Y640F and D661Y STAT3 mutations (Chapter 3). Although past studies observed that approximately 50% of LGL leukemia patients exhibit STAT3 mutations, we utilized a 183-person cohort and divided based on RA status to determine that 61% of LGL leukemia patients without RA harbor STAT3 mutations, and 83% of LGL leukemia patients with RA have mutated STAT3. The newly detected mutation VAFs included a wide range in size from a low 0.007% to 25.83%. Additionally, these LGL leukemia studies were performed in DNA isolated from PBMCs rather than CD8+ T-cell DNA as was utilized in Chapter 2, so it is possible that with CD8+ T-cell isolation, these rates could be even higher. Y640F STAT3 mutation occurred more frequently than D661Y, and interestingly, we noted that 30% of patients exhibit double mutations in both Y640F and D661Y that likely occur in different clones. We show that these low VAF STAT3 mutations can fluctuate over time in conjunction with clinical features and in response to treatment. We also demonstrate that these mutations strongly correlate with clinical features such as neutropenia. Interestingly, neutrophil cell death has been implicated in production of citrullinated proteins, which can lead to anti-CCP positivity like we observed in RA patients with STAT3 mutations in Chapter 2. Together, these discoveries may help to inform treatment decisions in the future, as patients with STAT3 mutations may respond more effectively to therapeutics that decrease STAT3 signaling.

One therapeutic utilized in LGL leukemia is the anti-folate and first-line RA therapy, MTX. Of particular interest, patients with STAT3 mutations tend to respond better to MTX. MTX has previously been shown to impact STAT3 signaling in certain cell types, but its effects on LGL cells in vitro have never been characterized despite frequent therapeutic use for LGL leukemia. We, therefore, sought to better understand MTX in an LGL leukemia context (Chapter 4). We also expanded our studies to include a similar, but more potent analog never-before used in LGL leukemia, pralatrexate (PDX). Another therapeutic shown to impact JAK/STAT signaling is vitamin D. Vitamin D treatment induces vitamin D receptor (VDR), a transcription factor that has been implicated in the control of MTX’s mode of cellular entry, reduced folate carrier (RFC). We hypothesized that combination of the anti-folates and vitamin D may allow more anti-folates into cells and lead to beneficial effects on viability and signaling. We observed strong cell killing effects with single agent MTX and PDX therapy. Combinatorial effects with vitamin D were dependent on cell culture conditions, with vitamin D exhibiting cell killing as a single agent and in combination with MTX or PDX following cytokine deprivation of the system. Combination of vitamin D with anti-folates demonstrated strong and sustained induction of VDR, and promising decreases in STAT3 and STAT1 activation.

Overall, we have identified STAT3 Y640F and D661Y mutations as frequent alterations in RA, LGL leukemia, and RA-associated LGL leukemia. STAT3 mutation associates with clinical features in all three disease manifestations. These observations, combined with frequent disease and symptomatic overlap, suggest there may be a common etiology or disease spectrum containing RA and LGL leukemia. Future studies are required to expand on the relationship between RA and LGL leukemia. There is room to investigate the impact of low VAF mutations on LGL biology and leukemic pathogenesis, as well as to explore STAT3’s potential as a biomarker for therapeutic targeting. Vitamin D and MTX combination may be a promising therapeutic approach for patients regardless of STAT3 status. Additional studies interrogating signaling changes and potential mechanisms of these agents in LGL leukemia will provide insight into how the disease functions and how better to treat it.