Comparison of somatic STAT3 and STAT5b gene mutations between Felty's syndrome and T-cell large granular lymphocytic leukemia with rheumatoid arthritis

Vadim R. Gorodetskiy (  gorodetskiyblood@mail.ru ) V.A.Nasonova Research Institute of Rheumatology https://orcid.org/0000-0001-8428-1281 Yulia V. Sidorova National Research Center for Hematology Natalia A. Kupryshina N.N.Blokhin Russian Cancer Research Center Vladimir I. Vasilyev Diagnostic center of the MEDSI clinic Natalya A. Probatova N.N. Blokhin Russian Cancer Research Center Natalya V. Ryzhikova National Research Center for Hematology Andrey B. Sudarikov National Research Center for Hematology


Introduction
Currently, Felty's syndrome (FS) is considered an uncommon subset of seropositive rheumatoid arthritis (RA). [1] Persistent unexplained neutropenia (absolute neutrophil counts below 1.5-2.0 × 10 9 /L) is a mandatory criterion for suggesting FS in a patient with RA. Although splenomegaly was one of the triads originally described by Felty, later studies showed that spleen size does not correlate with neutropenia and that patients with RA and neutropenia are like patients with the full triad. [2][3][4] At present, splenomegaly is not an absolute diagnostic requirement for diagnosing FS. [5,6] FS is a clinical diagnosis and there is no speci c single diagnostic test to con rm or exclude it; therefore, FS is essentially a diagnosis of exclusion. The pathogenesis of FS is unknown.
T-LGLL with RA resembles FS in many aspects and some researchers hypothesize that RA-associated T-LGLL and FS are part of a single disease process with a common pathogenic mechanism. [5,18,19] However, the prevalence of STAT3 and STAT5b mutation status in FS is unknown. In this study, we strati ed 81 patients with RA and unexplained neutropenia into 2 groups based on the presence and absence of T-cell clonality (RA-associated T-LGLL and FS), and then examined STAT3 and STAT5b gene mutations in both groups. We also present the clinical and laboratory characteristics of 25 patients with FS and discuss their differential diagnosis with RA-associated T-LGLL.

Patients And Methods
We diagnosed FS if all three criteria were met: (i) RA diagnosed according to the 2010 American College of Rheumatology/European League Against Rheumatism criteria [20]; (ii) neutropenia (absolute neutrophil counts < 1.5 × 10 9 /L) without an alternative explanation such as drug-induced suppression; (iii) absence of T-cell clonality.
Peripheral blood (PB) smears for LGL counting were re-examined in 18 cases. Bone marrow aspiration with differentiated cell counts was performed in 14 cases and in 11 of these, a bone marrow biopsy was also performed. The collected clinical data included patient age, sex, presence of splenomegaly, RA duration, titers of rheumatoid factor (RF), antibodies against cyclic citrullinated peptides (anti-CCP), antibodies against mutated citrullinated vimentin (anti-MCV), maximal Disease Activity Score derivative for 28 joints calculated using C-reactive protein (max.DAS28-CRP) during follow-up, erosive arthritis, and associated autoimmune diseases.
Evaluation of STAT3 and STAT5b gene mutations and T-cell clonality STAT3 and STAT5b gene mutations and T-cell clonality were examined using genomic DNA extracted from blood (17 patients), blood and bone marrow (5 patients), bone marrow (1 patient), blood and spleen (1 patient), and blood, bone marrow and spleen tissue (1 patient) samples.
Evaluation of T-cell clonality was based on the rearrangements of the T-cell receptor (TCR) gamma (Vγ-Jγ) and TCR beta (Vβ-Jβ, Dβ-Jβ) genes. T-cell clonality assays were performed according to the BIOMED-2 standardized protocol. [21] Polymerase chain reaction (PCR) was carried out using an automated DNA Engine thermocycler (BioRad, Hercules, USA), and fragments were detected using an ABI PRISM 3130 Genetic Analyzer (Applied Biosystems, Foster City, CA); the data were analyzed using GeneMapper software version 4.0 (Applied Biosystems, Foster City, CA).
Allele-speci c (AS) TaqMan Real-Time PCR assays were employed to determine the somatic point mutations in STAT3 (p.Y640F; p.N647I; p.D661V; p.D661Y; p.D661H; p.D661N) and STAT5b (p.N642H) genes. DNA (200-400 ng) was added to 25 µL of the reaction mixture containing 10 pmol of WT (wild type)-speci c or MT (mutated type)-speci c forward primer, 10 pmol of common reverse primer, and 7.5 pmol of the uorescent probe. AS-PCR was then performed in triplicate (3 WT + 3 MT) using a StepOne Real-Time PCR System (Applied Biosystems, USA). PCR conditions included preliminary denaturation at 95 °C for 5 min, followed by 45 cycles at 95 °C for 30 s, 62 °C for 30 s, and 72 °C for 30 s.
A mixture of DNA from healthy donors was used as a negative control. Samples with mutations con rmed by Sanger sequencing were used as positive controls. The primer and probe sequences are shown in Supplement 1. As a comparison group, we tested the STAT3 and STAT5b mutational status in 56 patients with T-LGLL in the context of RA.

Statistical analysis
Fisher's two-tailed test was used to compare the frequency of STAT3 mutations between the groups of patients with FS and RA-associated T-LGLL.

Results
We strati ed 81 patients with RA and neutropenia according to the presence and absence of T-cell clonality, respectively, into 2 groups: RA-associated T-LGLL (56 patients) and FS (25 patients Seven of the 25 (28%) patients also had a concomitant autoimmune condition: 6 patients had Sjogren syndrome (SS) and one patient had SS and autoimmune thyroiditis. Nineteen of the 25 (76%) patients were females. The median age at FS diagnosis was 54 years (range, 30-79 years). The median duration of RA prior to FS diagnosis was 7 years (range, 0-35 years). The median maximum RA activity over the observation period, estimated by DAS28, was 4.9 (range, 2.1-7.01). Erosive arthritis at FS diagnosis was detected in 16 of 22 (73%) of patients. RF was positive in 23 of 25 (92%) patients; however, 2 of these 23 RF-positive patients had low RF levels (less than 2 normal ranges). The median RF level was 332.7 IU/mL (range, 9.5-12900.0 IU/mL). Anti-CCP was positive in 23 of 24 (96%) patients. Anti-MCV was positive in all 11 patients examined. In all 4 RF-negative and weakly positive patients, the anti-CCP and anti-MCV titers were highly positive. The median neutrophil count was 0.496 × 10 9 /L with a range of 0.052-1224.0 × 10 9 /L.  LGLs (× 10 9 /L), median (range) 0.413 (0.117-1.036) A neutrophil count < 0.5 × 10 9 /L was observed in 13 of 25 (52%) patients. None of the patients with FS in our cohort had lymphocytosis. The absolute number of lymphocytes in PB ranged from 0.42 to 2.32 × 10 9 /L with a median of 1.092.
The results of ow cytometric immunophenotyping of lymphocytes performed in 19 patients are summarized in Table 2. A weakening or absence of CD5 pan T-cell antigen expression on cytotoxic (CD3+/CD8+) T-lymphocytes was the most commonly identi ed phenotypic abnormality in 13 of 18 (72%) cases. CD16, CD56, and CD57 expression on cytotoxic T-lymphocytes was found in 1 of 15 (7%), 1 of 16 (6%), and 10 of 18 (56%) cases, respectively. PB, peripheral blood; BM, bone marrow; -, express less than 10% of CD3 + CD8 + lymphocytes; +, express more than 10% of CD3 + CD8 + lymphocytes; pos., positive; ND, not done Normal or hypercellular bone marrow resulting from myeloid hyperplasia was revealed in all cases. There were no signs of myelodysplasia. The number of lymphocytes in the bone marrow was not elevated and comprised 3.8-18.4% of nucleated cells. In 11 of 14 patients bone marrow aspirate differential counts showed a signi cant reduction in segmented neutrophils, and in 1 patient, they showed a reduction in band and segmented neutrophils. However, in 2 patients, the bone marrow aspirate differential count was normal despite detection of PB neutrophils. A bone marrow immunohistochemical study in 2 of 6 cases showed interstitial clusters and/or linear arrays of intravascular CD8+/granzyme B + lymphocytes.

Discussion
Historically, LGLL could be readily recognized by reviewing a PB smear. An LGL count of more than 2 × 10 9 /L (normal LGL count in PB: 0.2-0.4 × 10 9 /L) lasting > 6 months, was considered a criterion for determining this disease. [12,22] For T-LGLL, the current diagnostic requirements have lowered this threshold to > 0.4 or 0.5 × 10 9 /L provided that a clonal T-LGL population is found with an appropriate clinical context. [23][24][25] Recent studies have shown that 49% of patients with T-LGLL have no absolute lymphocytosis and 36% of patients have blood LGLs < 1 × 10 9 /L. [9] As the clinical manifestations of RAassociated T-LGLL are often identical to those in which one would suspect an FS, it may be di cult to differentiate RA-associated T-LGLL with a low LGL count (0.4-2.0 × 10 9 /L) from FS. Moreover, expansion of LGLs can be detected in patients with FS. [26][27][28] RA-associated T-LGLL and FS can be distinguished by T-cell clonality determined by assessing the TCR gene rearrangements present in T-LGLL but not in FS. [1,5,29] However, there is considerable discussion regarding the signi cance of dominant T-cell clones as a hallmark of T-cell malignancy because small populations of clonally expanded T-LGLs are revealed in healthy individuals and in an exuberant reactive response. [30][31][32][33][34] Considering that the difference between RA-associated T-LGLL and FS often depends on a single test with well-known gray areas in interpretation and limitations, [35] it is necessary look for additional distinctions. We can use mutations in STAT3 and STAT5b genes as molecular markers for T-LGLL diagnostics, [36] but their prevalence in FS and their diagnostic value for differential diagnosis between FS and RA-associated T-LGLL are unclear. In this study, we did not detect STAT3 mutations in any of the 24 cases with FS, as opposed to 22 of 56 patients with RA-associated T-LGLL. Further, no STAT5b mutation was detected in any FS or RAassociated T-LGLL patient in our cohort.
Savola et al. examined STAT3 and STAT5b mutations in 14 patients with RA and neutropenia. [37] Similar to our patient cohort, they did not nd any STAT5b mutations. However, in contrast to our results, they identi ed STAT3 mutations in 6 of 14 (43%) patients. We believe that difference between outcomes obtained by Savola et al. and our study can be attributed to different methods of assessing T-cell clonality and the patient selection criteria. We tested T-cell clonality based on the rearrangement of gamma and beta chain-encoding genes by a PCR-based assay, whereas Savola et al. studied the clonality of T cells by ow cytometry using a Vβ kit, covering only 70% of the Vβ T-cell repertoire. In contrast to Savola et al., we did not include patients with T-cellular clonality in the FS group.
Female prevalence, age at FS diagnosis, and duration of RA prior to FS diagnosis in our series were comparable to the results found in literature. [4,5,38] Overall, in our patient cohort, RA was of moderate activity, even though RA is typically severe in patients with FS. All our patients were seropositive: RF+/anti-CCP+/anti-MCV + or RF−/anti-CCP+/anti-MCV+. Splenomegaly ranging from massive to detectable only based on abdominal imaging modalities, was detected in 83% of patients. SS was diagnosed in our study in 28% of patients, which is signi cantly less than in the FS patient cohorts reported by other authors: 48% (Sienknecht et al.), 69% (Barnes et al.), and 53% (Campion et al.). [2,3,39] In our study, low count expansion of LGLs (0.4-2.0 × 10 9 /L) in PB was detected in 56% of cases, but did not exceed 2.0 × 10 9 /L and the bone marrow aspirate differential count showed no increase in lymphocytes. Flow cytometric immunophenotyping studies play an important role in the diagnosis of T- LGLLs. The expression of CD57 and CD16 antigens, one or both of which are detected in the vast majority of T-LGLL cases, was found on cytotoxic T-lymphocytes in only 56% and 7% of our cases, respectively. In contrast, aberrant expression of CD5 was the most common nding in our patient group. This abnormality is frequently associated with T-LGLL, but is also found in T-cell reactive expansion. [41,42] Bone marrow involvement is present in at least 75% of T-LGLL cases, although it is often subtle and di cult to detect. Speci c criteria have been proposed for the diagnosis of T-LGLL in bone marrow sections using immunohistochemistry. [42; 43] However, as reported by Burks et al., there are probably no distinctive features in bone marrow biopsies that would separate T-LGLL from FS. [5] In 2 patients with FS in our cohort, immunohistochemical studies also revealed that bone marrow in ltration by cytotoxic Tlymphocytes was indistinguishable from T-LGLL lesions.
The pathogenesis of neutropenia in FS has not yet been fully studied and seems to be multifactorial. In 12 of 14 cases in our study, the bone marrow aspirate differential count t into the expected consequence of peripheral destruction/sequestration of neutrophils. Although the role of splenic sequestration/destruction in neutropenia pathogenesis is not supported by all studies, [5] splenectomy produces a long-term hematologic response in 80% of patients with FS. [44] We also observed persistent recovery of neutrophil levels after splenectomy in 2 of our patients with FS and massive splenomegaly.

Limitations
We are aware of some limitations concerning our study design. Due to the retrospective design of the study, some of the data were incomplete. Additionally, we used of allele-speci c TaqMan real-time PCR rather than Sanger sequencing to detect somatic point mutations in STAT3 and STAT5b genes. A set of primers for most common mutations in STAT3 and STAT5b genes was developed. Even though this

Declarations Acknowledgments and a liations
We would like to thank Editage (www.editage.com) for English language editing.
Authors' contributions V.R.G. substantially contributed to the conception of the work, acquisition, analysis, and interpretation of data; made a draft of the work; approved the nal version of the work for publication; and agreed to be responsible for all aspects of the work. Y.V.S., N.A.K., V.I.V., N.A.P., N.V.R., and A.B.S. substantially contributed to the acquisition, analysis, and interpretation of data for the work; critically revised the work for important intellectual content; approved the nal version of the work for publication; and agreed to be responsible for all aspects of the work.
Funding: This study did not receive any funding support.

Availability of data and materials
All data generated or analyzed during this study are included in this article.
Ethics approval and consent to participate approach provides much higher sensitivity compared to Sanger sequencing, some rare mutations, not covered by the developed primers, could not be identi ed.

Conclusion
In conclusion, we did not detect STAT3 and STAT5b gene mutations in any of the 24 FS cases in this study, whereas STAT3 gene mutations were found in 22 of 56 patients (39%) with T-LGLL in RAassociated T-LGLL (p < 0.001). Although further data are required, our results suggest that identifying STAT3 mutations in patients with clinical pattern resembling FS may be a counterargument to make this diagnosis and that careful patient evaluation is required to exclude RA-associated T-LGLL.
Abbreviations FS: Felty's syndrome; RA:rheumatoid arthritis; T-LGLL:T-cell large granular lymphocytic leukemia; LGLs:large granular lymphocytes; STAT:signal transducer and activator of transcription; RF:rheumatoid factor; anti-CCP:antibodies against cyclic citrullinated peptides; anti-MCV:antibodies against mutated citrullinated vimentin; DAS28:Disease Activity Score derivative for 28 joints; CD:cluster of differentiation; TCR:T-cell receptor; PCR:polymerase chain reaction; AS:allele-speci c; WT:wild type; MT:mutated type; PB:peripheral blood; BM:bone marrow; ND:not done The study was approved by the V.A. Nasonova Research Institute of Rheumatology Ethics Committee (protocol #17 on the 20-02-2020) and was conducted in accordance with the Declaration of Helsinki of 1975, as revised in 2008. Written informed consent was given by all participants included in the study.

Consent for publication
Not applicable.