Coordinated Priming of NKG2D Pathway by IL-15 Enhanced Functional Properties of Cytotoxic CD4⁺CD28⁻ T Cells Expanded in Systemic Lupus Erythematosus

A total of 149 patients with SLE were enrolled from the Department of Rheumatology and Immunology, Shenzhen People’s Hospital, including 80 patients with nephritis (LN group) and 74 patients without nephritis (NLN group). In total, 39 healthy individuals (age and gender-matched) were recruited as controls (Supplementary Table 1). Patients with other autoimmune diseases, infectious diseases, or over 70 years old were excluded from this study. Blood samples were obtained from all patients and healthy controls. All patients fulfilled the Systemic Lupus International Collaborating Clinics (SLICC) classification criteria for SLE [35]. The Systemic Lupus Erythematosus Disease Activity Index 2000 (SLEDAI-2 K) score for patients is calculated using the SLEDAI-2 K score calculator according to both clinical and laboratory parameters (https://​qxmd.​com/​calculate/​calculator_​335/​sledai-2k) [36]. The SDI [37] was also assessed. The diagnosis of LN was confirmed by renal biopsy as per the International Society of Nephrology and Renal Pathology Society (ISN/RPS) criteria [38]. Demographic, clinical, and laboratory data were collected from hospital records or by questionnaire and reviewed by experienced physicians.

Blood samples were collected in EDTA anticoagulant tube, and 100 µl blood samples were transferred to a new tube for flow cytometry analysis. Peripheral blood mononuclear cells (PBMCs) were isolated using density gradient centrifugation. The frequency of CD4⁺CD28⁻ T cells and their phenotypical characterization were determined by staining blood with antibodies, followed by erythrocyte lysis with Lyse/Fix buffer (BD Biosciences). The percentage of circulating CD4⁺CD28⁻ T cells was calculated as the percentage of CD4⁺ T cells. For intracellular staining, including staining for cytotoxic molecules (granzyme B (GZMB), granzyme A (GZMA), perforin (PRF1)), transcription factor (T-bet, Runx3) and cytokines (interferon-gamma (IFN-γ), tumor necrosis factor-alpha (TNF-α)), PBMCs or cells from whole blood after surface marker staining and erythrocyte lysis, was fixed, permeabilized and stained with antibodies using a Fixation/Permeabilization Solution Kit (BD Biosciences) according to the manufacturer’s instructions. For assessing the phosphorylation state of the signal transducer and activator of transcription 5 (STAT5), CD4⁺CD28⁻ T cells were isolated and then cultured alone or treated with IL-15. Phosphorylated STAT5, Akt, and extracellular signal-regulated kinase (ERK) were assessed via the BD Phosflow method using STAT5 antibody, Akt antibody, ERK1/2 antibody, BD Phosflow Lyse/Fix Buffer, and BD Perm Buffer III (all BD Biosciences) as per the manufacturer’s instructions.

Antibodies targeting the following molecules were used in this study: CD3 (clone: HIT3a; PerCP/Cy5.5 BioLegend, clone: UCHT1; Brilliant Violet421 BioLegend), CD4 (clone: A161A1; APC-Cy7, PE/Cyanine7 BioLegend), CD8 (clone: SK1; APC-Cy7 Biolegend and BD Biosciences), CD28 (clone: CD28.2; FITC, PE-Cy7 BioLegend), granzyme B (clone: QA18A28; PE Biolegend, clone: QA16A02; Alexa Fluor 647 Biolegend), granzyme A (clone: CB9; PE Biolegend), perforin (clone: dG9; PE/Cyanine7 BioLegend), CX3CR1 (clone: 2A9-1; APC eBioscience), NKG2D (clone: 1D11; APC BD Biosciences), T-bet (clone: 4B10; PE/Cyanine7 BioLegend), Runx3 (clone: R3-5G4; PE BD Biosciences), TNF-α (clone: MAb11; APC BioLegend), IFN-γ (clone: 4S.B3; PE BioLegend and BD Biosciences, PE/Cyanine7 BD Biosciences), STAT5 (pY694; PE BD Biosciences), AKT (pS473; Alexa Fluor 647 BD Biosciences), ERK1/2 (pT202/pY204; PE/Cyanine7 BD Biosciences). Data were acquired using FCAS Canto II (BD Biosciences) and analyzed with FlowJo software.

An intracellular cytokine assay was performed to assess cytokine levels. Briefly, PBMCs were stimulated with a combination of phorbol 12-myristate 12-acetate (PMA) and ionomycin (both from Sigma-Aldrich) for 4 h. Three hours prior to analysis, Brefeldin A (Sigma-Aldrich) was added. PBMCs were stimulated in a standard culture medium without PMA and ionomycin (Ion) as a control. To analyze IFN-γ and TNF-α producing cells, the cell surface of PBMCs was stained with antibodies (anti-human CD3, anti-human CD4, anti-human CD28). Next, cells were fixed, permeabilized, and stained with antibodies (anti-human IFN-γ and anti-human TNF-α).

PBMCs were isolated from 10 ml fresh peripheral blood samples by Ficoll-Paque (GE Healthcare, USA) density gradient centrifugation. To detect surface mobilization of lysosomal associated membrane protein-1 (CD107a), a marker of degranulation associated with cytolytic function, PBMCs were seeded in 96-well plates and cultured alone or stimulated with anti-CD3 antibody (Clone: OKT3, Biolegend) or treated with 50 ng/ml IL-15 for 4 days. Anti-human CD107a PE was added to and incubated with the culture for 4 h to provoke degranulation before analysis. Three hours before analysis, Brefeldin A was added. Flow cytometry was performed by labeling cells with CD3, CD4, CD28, and CD107a expression was quantified by flow cytometry.

After 24 h of treatment with anti-CD3 antibody and extensive washing, CD4⁺CD28⁻ T cells were co-cultured with human renal glomerular endothelial cells (HRGEC) at a ratio of 1:1 in RPMI 1640 supplemented with 100 mg/ml streptomycin, 100 μg/ml penicillin, and 10% fetal bovine serum (FBS) for 48 h. HREGC were cultured alone or co-cultured with untreated CD4⁺CD28⁻ T cells for 48 h as a control. Cells from the above co-cultures were washed to remove CD4⁺CD28⁻ T cells and stained the remaining adherent HREGC with annexin V-FITC and propidium iodide (PI) and evaluated for apoptosis by flow cytometry conducted with FCAS Canto II (BD Bioscience).

Tissue samples were obtained from patients with LN. These tissue samples were fixed in formalin and paraffin-embedded prior to sectioning. After deparaffinization, the tissue sections were placed in a plastic container filled with Tris–EDTA buffer for antigen retrieval. All sections were incubated with the primary antibodies to GZMB (496B, eBioscience; dilution 1:200) and CD4 (2H4A2, Protein Tech; dilution 1:400) in humidified boxes overnight at 4 °C after blocking with 0.01 M phosphate buffer saline (PBS) containing 3% bovine serum albumin (BSA) for 30 min. Next, the sections were washed and incubated with secondary antibodies (Cy3 conjugated AffiniPure Goat Anti-Rat IgG (GB21302, Servicebio) and FITC conjugated Goat Anti-Mouse IgG (GB22301, Servicebio). The slides were mounted using ProLong Gold antifade reagent with 4′, 6-diamidino-2-phenylindole (DAPI) (G1012, Servicebio) for nuclear staining. Images were acquired with a fluorescence microscope (Leica, Germany).

Plasma was isolated from fresh EDTA-treated blood samples from SLE patients and healthy controls and stored frozen. Quantitative analysis of IL-5 was performed using enzyme-linked immunosorbent assay (ELISA) (R&D Systems) according to the manufacturer’s instructions.

In order to investigate whether CD4⁺ cytotoxic T cells are present in SLE patients, we analyze the frequency of CD4⁺ cytotoxic T cells in blood samples from SLE patients using flow cytometry. As shown in Fig. 1A, D, SLE patients exhibited significantly higher occupancy of CD4⁺CD28⁻ T cells in the circulation than healthy controls. In addition, SLE patients diagnosed with nephritis showed much higher occupancy than those without nephritis. We found that these cells abundantly expressed cytolytic granule Granzyme B (GZMB; 93%), suggesting that these CD4⁺CD28⁻ T cells function as cytotoxic cells (Fig. 1B, D). The percentage of these cytotoxic CD4⁺CD28⁻ T cells positively correlates with SDI (r = 0.36, P < 0.001), whereas no correlation was found between the percentage of CD4⁺CD28⁻ T cells and SLEDAI (systemic lupus erythematosus Disease Activity Index) (Fig. 1C and Supplementary Fig. 1). We also divided the patients into two different groups, one group without lupus nephritis (NLN) and another group with lupus nephritis (LN) and found that the percentage of CD4⁺CD28⁻ T cells was also positively correlated with SDI after grouping (NLN (r = 0.32, P = 0.006) and LN (r = 0.36, P = 0.003). Furthermore, multicolor immunofluorescence studies showed that in patients with lupus nephritis, CD4⁺GZMB⁺ T cells infiltrated into kidney tissues (Supplementary Fig. 2). We also found that a portion of CD4⁺CD28⁻ T cells expressed cytotoxic molecular Granzyme A (GZMA; 20.5%) and Perforin (RPF1; 19.4%) (Fig. 2E). The GZMB/PRF1 and GZMB/GZMA staining profiles showed that the PRF1-expressing and GZMA-expressing CD4⁺CD28⁻ T cells were predominantly GZMB positive, suggesting that cytotoxic granules in CD4⁺CD28⁻ T cells were co-expressed (Supplementary Fig. 3). In addition to the observed increase in granules molecules, the expression levels of transcription factors Runx3 and T-bet were increased in CD4⁺CD28⁻ T cells (Fig. 1G). The cooperation of these two factors is thought to activate the cytotoxic program and contribute to the reprogramming of CD4 Th cells into cytolytic effectors [39]. Furthermore, we found that the chemokine receptor CX3CR1 and NK cell-associated activating receptor NKG2D were also significantly higher in samples from SLE patients compared with those from healthy controls (Fig. 1H). Importantly, these cell subsets showed an effector-memory phenotype evidenced by dual loss of CCR7, CD45RA (TEM), together with re-expressed CD45RA (TEMRA) (Fig. 1F). Together, these results indicated that CD4⁺CD28⁻ T cells from SLE patients are characterized by both an effector-memory state as well as high cytotoxic potential and involved in SLE-related tissue damages.

To investigate the functional attributes of CD4⁺CD28⁻ T cells expanded in SLE patients, we measured their inflammatory cytokine production and cytolytic function. As shown in Fig. 2A, B, after stimulation with PMA/ionomycin, both CD4⁺CD28⁻ T cells and CD4⁺CD28⁺ T cells in SLE patients produced IFN-γ and TNF-α. The frequencies of IFN-γ and TNF-α producing CD4⁺CD28⁻ T cells were much higher than that in CD4⁺CD28⁺ T cells, suggesting that CD4⁺CD28⁻ T cells from SLE patients have the potential signature of proinflammatory cytokine. CD107a has been described as a marker of activation-induced degranulation and is a prerequisite for cytolysis. We next assessed the cytotoxicity of CD4⁺CD28⁻ T cells from SLE patients by quantifying CD107a expressed at the cell surface. We found that activation of CD4⁺CD28⁻ T cells with anti-CD3 antibody increased the expression levels of CD107a (Fig. 2C). In addition, treatment of CD4⁺CD28⁻ T cells with anti-CD3 antibody-induced apoptosis in co-cultured HRGECs, and cd107a expression on the surface of co-cultured CD4⁺CD28⁻ T cells was upregulated (Fig. 2D and Supplementary Fig. 4). Together, these results suggested that cytotoxic CD4⁺CD28⁻ T cells expanded in SLE are functionally active and are capable of producing proinflammatory cytokines.

While we successfully validated and characterized CD4⁺ cytotoxic T cells in our study, evidence for their existence remains circumferential, with little direct evidence of how these cells expand or function in the pathogenesis of SLE. Previous studies have shown that the cytokine IL-15 is overproduced in some inflammatory and autoimmune diseases. Chronic exposure to IL-15 may promote the expansion of CD4⁺CD28⁻ T cells in vivo [40‐42]. Here, we investigated whether IL-15 is involved in the expansion and regulates the function of cytotoxic CD4⁺CD28⁻ T cells in SLE. Our results showed that IL-15 was increased in the plasma of SLE patients (Supplementary Fig. 5). Furthermore, incubation in IL-15 significantly increased the frequency of CD4⁺CD28⁻ T cells in SLE patients (Fig. 3A, B). Next, we investigated the effects of IL-15 on the proliferation of CD4⁺CD28⁻ T cells in SLE patients by determining the expression of Ki67, a well-known proliferation marker for cell proliferation evaluation. The results showed that Ki67-expressed CD4⁺CD28⁻ T cells and the expression level (based on mean fluorescence intensity (MFI)) of Ki67 on CD4⁺CD28⁻ T cells were both significantly increased after stimulation with IL-15. In contrast, minor changes in Ki67-expressed cells and Ki67 expression level were observed in the CD4⁺CD28⁺ T cells followed by IL-15 treatment (Fig. 3C and Supplementary Fig. 6A). To identify the active status of CD4⁺CD28⁻ T cells, we measured the expression level of the activation marker CD69. As shown in Fig. 3D and Supplementary Fig. 6B, treatment with IL-15 significantly increased the proportion of CD69-expressing CD4⁺CD28⁻ T cells and the CD69 expression levels (MFI) on CD4⁺CD28⁻ T cells. We observed only minor differences in the CD69 expression of the CD4⁺CD28⁺ T cells counterparts. In addition, we assessed the effects of IL-15 treatment on the expression of granzyme B in CD4⁺CD28⁻ T cells from SLE patients. As shown in Fig. 3E and Supplementary Fig. 6C, IL-15 did increase the expression levels (MFI) of granzyme B in CD4⁺CD28⁻ T cells, although it did not change the percentage of CD4⁺CD28⁻ T cells expressing granzyme B. Next, we investigated whether the release of cytotoxic molecules was affected by IL-15 using a degranulation assay based on the CD107a expression. Our results showed that IL-15 significantly upregulated the degranulation of CD4⁺CD28⁻ T cells. Moreover, IL-15 significantly upregulated the anti-CD3-induced degranulation of CD4⁺CD28⁻ T cells (Fig. 3F and Supplementary Fig. 6D). We also found that IL-15 induces the production of disease-related proinflammatory cytokines in CD4⁺CD28⁻ T cells. The percentage of cells that produce TNF-α and IFN-γ significantly increased in both CD4⁺CD28⁻ and CD4⁺CD28⁺ T cells treated with IL-15 (Fig. 3G, H and Supplementary Fig. 6E), and the increase was higher in CD4⁺CD28⁻ T cells compared to CD4⁺CD28⁺ T cells. Together, these results suggest that following IL-15 treatment, CD4⁺CD28⁻ T cells develop a higher proliferative capacity and higher activity status than their CD4⁺CD28⁺ T cells counterparts.

Previous studies showed that IL-15 receptor signaling is transmitted through the JAK/STAT and RAS/Raf/ERK signaling pathway [43‐45]. In order to investigate the mechanism underlying IL-15 regulated function of CD4⁺CD28⁻ T cells, we analyzed the activity of components involved in the IL-15 receptor signaling pathways. As shown in Fig. 4A, engagement of IL-15R with IL-15 induced an increase of STAT5 phosphorylation in CD4⁺CD28⁻ T cells. In contrast, phosphorylated ERK levels changed little in the same condition (Supplementary Fig. 6). Next, we block IL-15 signaling with the JAK3 selective inhibitor tofacitinib. As shown in Fig. 4B, the expansion and proliferation of CD4⁺CD28⁻ T cells were inhibited after adding tofacitinib. In addition, our results show that blocking IL-15 signaling with tofacitinib also decreased the percentage of Ki67-expressed CD4⁺CD28⁻ T cells and the expression level (MFI) of GZMB in CD4⁺CD28⁻ T cells (Fig. 4C, D). Together, these findings suggested that IL-15 regulates the expansion, proliferation, and effector function of CD4⁺CD28⁻ T cells from SLE patients through activation of the JAK3/STAT5 and NKG2D signaling pathway.

We have shown increased surface NKG2D on CD4⁺CD28⁻ T cells compared to CD4⁺CD28⁺ T cells in Fig. 1H. And previous studies showed that IL-15 increases the abundance of NKG2D and DAP10 in the NK cell surface membrane in vitro, which primes the cytotoxic responses of NK cells [46‐48]. Here, we investigate whether IL-15 cooperatively initiates the NKG2D pathway to regulate CD4⁺CD28⁻ T cell cytotoxicity. As shown in Fig. 5A, we found that the expression of surface NKG2D on CD4⁺CD28⁻ T cells was upregulated after exposure to IL-15. To evaluate whether IL-15 induces the activation of the NKG2D-mediated signaling pathway, we assessed the phosphorylation events downstream of the NKG2D pathway. As shown in Fig. 5B, we found that AKT phosphorylation levels were significantly increased in CD4⁺CD28⁻ T cells following the addition of IL-15 and pretreatment of CD4⁺CD28⁻ T cells with the PI3K inhibitor LY294002 inhibited AKT phosphorylation. However, IL-15R engagement also mediates PI3K/AKT pathway activation. Thus, to further investigate whether the phosphorylation of AKT is completely dependent on NKG2D/DAP10 signaling pathway, we blocked the NKG2D with neutralizing antibody and analyzed the phosphorylation level of AKT in CD4⁺CD28⁻ T cells treated with IL-15. We observed that treatment of CD4⁺CD28⁻ T cells with IL-15 plus NKG2D neutralizing antibody partially inhibited the phosphorylation of AKT, indicating that IL-15 combined with NKG2D activate the PI3K/AKT pathway (Fig. 5C). Furthermore, our results show that pretreatment of CD4⁺CD28⁻ T cells with the PI3K inhibitor LY294002 significantly decreased the percentage of Ki67-expressed CD4⁺CD28⁻ T cells and CD107a-expressed CD4⁺CD28⁻ T cells and the expression level (MFI) of GZMB in CD4⁺CD28⁻ T cells (Fig. 5D–F). Together, these results suggested that IL-15 stimulation induces the activation of NKG2D-mediated signaling pathway in CD4⁺CD28⁻ T cells and the consequent phosphorylation of AKT.