Phenotypic and functional characterization of natural killer cells in rheumatoid arthritis-regulation with interleukin-15

Rheumatoid arthritis (RA) is an autoimmune disease characterized by synovial inflammation and joint destruction. Previous studies have shown that natural killer (NK) cells may play an important role in the pathogenesis of RA. Interleukin (IL)-15, a pro-inflammatory cytokine which induces proliferation and differentiation of NK cells, is overexpressed in RA. In this present study, we examine various NKRs and adhesion molecule expression on NK cells from RA patients and their response to IL-15 stimulation. We also sought to study cytokine-induced memory-like (CIML) NK cells in RA patients. We established that 1. RA patients had higher NK cell percentages in peripheral blood and their serum IL-15 levels were higher compared to healthy volunteers; 2. NK cells from RA patients showed lower NKp46 expression and an impaired CD69 response to IL-15; 3. NK cells from RA patients showed higher CD158b and CD158e expression but lower CD62L expression; 4. exogenous IL-15 up-regulated CD69, CD158b, CD158e but down-regulated NKp46 and CD62L expression in RA; 5. As to CIML NK cells, restimulation - induced NK cytotoxicity and IFN-γ production was impaired in RA patients, 6. Reduced NKp46, perforin, and granzyme B expression on NK cells was found in RA patients with bone deformity and erosion, 7. RA disease activity (DAS28) showed inverse correlation with the percentages of CD56+CD3− NK cells, and NKp46 and perforin expression on NK cells, respectively. Taken together, our study demonstrated differential expression of various NK receptors in RA patients. NKp46, CD158e, and perforin expression on NK cells may serve as markers of RA severity.

Rheumatoid arthritis (RA) is a chronic autoimmune disease characterized by the inflammation of synovial membrane and destruction of joints 1,2 . NK cells, defined by expression of CD56 and lack of CD3, are traditionally regarded as the member of innate immunity providing anti-tumor and anti-viral defense 3,4 . NK cells may also play a pathogenic role in autoimmune diseases, as several reports have shown accumulation of NK cells in inflammatory joints of RA and other arthritic diseases 5,6 .
Interleukin (IL)-15, a member of γ-chain cytokine, is a pro-inflammatory cytokine essential in NK cell proliferation and differentiation. Higher IL-15 levels are detected in serum and synovium of RA patients 7,8 . Cytokine-induced memory-like (CIML) NK cells refers to NK cells,when pre-activated with IL-12, IL-15, and IL-18, may exhibit enhanced function when re-activated following a resting period cultured with low dose IL-15 9 . CIML NK cells have enhanced anti-leukemic and anti-tumor capacity over conventional NK cells 10,11 . However, the role of CIML NK cells in autoimmunity remains to be determined.
In this present study, we perform a phenotypic and functional characterization of NK cells in RA patients and their response to IL-15 stimulation. We also sought to examine CIML NK cells in RA patients compared with Results characteristics of RA patients and healthy volunteers. The clinical characteristics of RA patients and healthy volunteers are shown in Table 1. RA patients were predominantly female (M:F = 9: 23), and age ranged from 22 to 84 years. Patients had higher CD56 + CD3 − NK cell percentages in the peripheral blood compared with healthy controls (14.8 ± 1.5% vs. 7.7 ± 0.8%, p = 0.001).
Enhanced CD94/NKG2A expression of NK (CD3 − /CD56 + ) cells were observed in pre-stimulated MNCs compared to controls (no pre-activation) from healthy volunteers (30.3 ± 5.0% vs. 22.6 ± 4.0%, p = 0.026).  We next examined the NK receptor expression in RA patients with advanced disease including bone erosion and deformity. As shown in Table 2, NKp46 expression on NK cells from RA with deformity and erosion was lower than those from the controls. (30.0 ± 9.7% vs. 71.2 ± 4.1%, p = 0.003).

Discussion
In the present study, we sought to examine the expression of various NK cell receptors from RA patients and determined their response to exogenous IL-15. We found the percentages of NK cells are increased in PBMC from RA patients compared to controls. Higher serum IL- 15     NKp46, a natural cytotoxicity receptor, mediates NK cell-induced cell lysis and provide an innate defense against intracellular pathogens, NKp46 also serve as a key factor in the pathogenesis of type 1 diabetes mellitus by recognition of pancreatic β cells ligands [15][16][17] . We observed NK cells from RA patients expressed lower NKp46 than those from controls, and IL-15 decreased NKp46 expression of NK cells from RA patients, similar to that observed in controls. NKp46 expression level on RA NK cells was negatively correlated with RA disease activity.
CD69, a triggering C-type lectin receptor that plays a crucial role in inducing NK-mediated cytotoxicity, involved in NK cell proliferation and served as a signal transmitting receptor in NK cells 18,19 . We found there was no difference between expression of CD69 on NK cells from RA patients and controls. However, we found that expression of CD69 on RA NK cells showed impaired response to IL-15. Guo et al. also demonstrated that CD69 expression of NK cells could be up-regulated in response to IL-15-superagonist stimulation 20 .
CD158b and CD158e are Ig-like inhibitory receptors on the surface of NK cells 21,22 . CD158b recognizes the group HLA-C alleles (HLA-Cw1,3,7 and 8), presenting Ser-77 and Asn-80 26, and CD158e corresponds to HLA Bw4 allele. Both of them down regulate NK cytotoxicity via preventing cell lysis 23,24 . We found higher CD158b expression level on NK cells from RA patients compared to controls, in discrepancy with Pridgeon et al. who showed that CD158b expression on NK cells from controls and RA patients were almost identical 25 . IL-15 up-regulated CD158b on RA NK cells and control NK cells to a similar extent.
CD62L is an adhesion molecule that may play a crucial role in mediating the NK cell recruitment to inflamed synovium in RA. We showed that CD62L expression was lower on RA NK cells compared to controls, and exogenous IL-15 further down-regulated the CD62L expression. Yang et al. showed that the shedding of CD62L from surface of anti-tumor T cells and acquisition of lytic activity, suggests CD62L may play a key role in T cell effector functions and anti-tumor activity 26 . CIML NK cells have been shown to have enhanced antitumor effect compared to conventional NK cells in adoptive immunotherapy 27 . However, the role of CIML NK cells in the pathogenesis of autoimmunity remains unclear. We are the first to study CIML NK cells in RA patients and found deficient cytotoxicity and IFN-γ production in CIML NK cells of RA patients. The CD69 and NKG2A expression of RA CIML NK cells was also decreased compared to healthy volunteers. Human CD94/NKG2A is an inhibitory receptor that recognizes HLA-E and is expressed by NK cells 28 . The clinical significance of this finding remains to be determined. IFN-γ is considered protective in RA as clinical symptoms of RA patients were relieved after administration of IFN-γ 29 . Our finding that deficient IFN-γ production of CIML NK cells may contribute in part to RA pathogenesis.
Granzyme B is an apoptosis inducer of chondrocytes with natural killer cell-like cytotoxicity in RA 30 . Perforin is a 70 kDa glycoprotein which is responsible for pore formation on the cell membrane of target cells induced by NK cells 31 . RA patients with bone erosion or deformity showed decreased granzyme B, perforin and NKp46 expression compared to those without bone erosion and deformity. We hypothesized that peripheral blood perforin and granzyme B expressing NK cells with great cytotoxic potential may migrate to inflamed joints in RA patients with bone erosion or deformity.
Taken together, we have characterized various NK receptor expression in RA patients and their response to IL-15, aberrant CIML NK cells in RA, and the correlation of NK receptor expression with RA disease severity, Further studies will be need to explore the pathogenic role of NK cells in RA patients.

Methods
Study subject. Study subjects include 32 RA patients and 20 healthy volunteers, the clinical characteristics of patients are summarized in Table 1. They were recruited from the Out-patient Clinic,Division of Rheumatology, Department of Internal Medicine at Chang Gung Memorial Hospital, Linkou, Taiwan. All the RA patients were diagnosed by the same physician according to the revised criteria of ACR/EULAR for RA 32 . Disease severity of patients were evaluated by the disease activity score, DAS28, according to the method of Prevoo et al. 33 , a multifaceted index composed of the number of swollen joints and number of tender joints out of 28 joints, erythrocyte sedimentation rate, and global assessment of health. Patients were separated into two groups according to whether bone deformity or erosions of fingers and toes on chest X-Ray diagnosed by radiologists or not. Peripheral blood samples were obtained from healthy volunteers with the approval of Chang Gung Medical Foundation Institutional Review Board (IRB: 201700445B0C501) and the informed consents had been obtained from all of the donors. cell culture. Peripheral blood was collected in sterile tubes containing heparin (20 units/ml of blood) and was processed within 24 hours of collection. Mononuclear cells (MNCs) were then separated from heparinized blood using Ficoll-Hypaque density gradient centrifugation. MNCs (1 × 10 6 /ml) were incubated for 18 hours in complete RPMI-1640 (with 10% Fetal calf serum) in the present or absence of rhIL-15 (10 ng/ml, Preprotech, Rocky Hill, USA) for subsequent analysis. flow cytometric analysis of nK cell markers. Cell staining was performed as previously described 34 .
Cells were washed in cold PBS with 2% FCS and 0.1% sodium azide and following stained with fluorescein isothiocyanate (FITC)-, phycoerythrin (PE)-conjugated anti-human monoclonal antibodies which were including anti-CD3/CD16 + CD56 (APC/FITC), NKp46, CD69, CD62L, CD158b, and CD158e (PE) from Becton-Dickinson for flow cytometric analysis. The fluorescent staining was analyzed on a Canto II (BD Biosciences) flow cytometer. Electronic gates were set to enable analysis of the fluorescence of the viable cell population according to FSC/SSC histograms following anti-CD3/ CD16 + CD56 stimulation. The percentage of cells stained with each monoclonal antibody was determined by comparing each histogram with one from control cells stained with FITC-or PE-labeled isotype control monoclonal antibodies.