We previously investigated the consequences of Rnls deletion on immune cell infiltration and activation by transplantation of the syngeneic mouse beta cell line NIT-1 into Non-obese diabetic (NOD) mice, a model of T1D. Single-cell RNA sequencing (scRNAseq) of graft-infiltrating CD45⁺ cells revealed that Rnls mutant (Rnls^mut) beta cells broadly influence immune cell activation and metabolism towards anti-inflammatory oncostatin M and Tgfβ1 expression, accompanied by enriched expression of genes important for glycolysis. In addition to elevated numbers of CD4⁺ T cells and tolerogenic PD-L1⁺ APCs, beta cell grafts deficient for Rnls showed markedly reduced frequency of NK cells. Because we previously demonstrated that PD-L1 blockade abrogates the advantage of Rnls^mut beta cells to survive autoimmunity, and NK cells are known to modulate APC maturation and activation, we now investigated the role of NK cells in protection of Rnls^mut beta cells in more detail (10). The gene expression profile of both graft-infiltrating NK cells and closely related type 1 innate lymphoid cells (ILC1) demonstrated markedly increased expression of genes involved in pathways of cellular activation and inflammation (interferon responses, allograft rejection, etc.) when derived from WT beta cell grafts compared to Rnls^mut ( Figure 1A and Supplementary Figures 1A–C ). Whereas ILC1 showed downregulated gene expression of inflammatory cytokines interferon gamma (Ifng) and tumor necrosis factor alpha (Tnf), NK cells upregulated gene expression of immune-regulatory Tgfβ1 when infiltrated in Rnls^mut beta cell grafts ( Supplementary Figures 1A, B ). Interestingly, the frequency of NK cells in Rnls^mut beta cell grafts was reduced not only in late stages (10), but also at earlier time points preceding the detection of graft weight difference between control and Rnls^mut ( Supplementary Figures 1D, E ; see Supplementary Figure 2 for gating strategy). Of note, the ILC1 frequency was not changed between the grafts and only represented a vast minority of infiltrating immune cells (≤1%) compared to NK cells (up to 14% according to scRNAseq analysis) (10), indicating that ILC1 likely does not play a major role in mediating the protective immune-regulation caused by Rnls^mut beta cells. These observations strengthened our hypothesis that NK cells might be crucial for regulating autoimmunity within Rnls^mut grafts. To investigate the importance of NK cells in shaping autoimmunity in vivo, we depleted NKp46⁺ innate lymphoid cells from autoreactive splenocytes before adoptive transfer into NIT-1 beta cell graft-bearing mice ( Figure 1B , Supplementary Figures 1F–H ). The isogenic WT and Rnls^mut beta cells were transplanted subcutaneously (s.c.) on opposite flanks of the same recipient mice following intravenous (i.v.) injection of diabetogenic splenocytes similar to described before (9, 10). Strikingly, NKp46-depleted autoreactive splenocytes lost the ability to protect Rnls^mut cells from autoimmunity in both immunodeficient recipients, NOD-Prkdc^scid (NOD.scid) mice and NOD.scid gamma (NSG) mice completely deficient for functional T, B, and NK cells ( Figure 1C ). Whereas the accumulation of CD4⁺ T cells was independent of innate lymphoid cells, the reduced frequency of mature MHCII^high+ APCs and downregulation of MHCII expression level on APCs in Rnls^mut grafts completely depended on the presence of graft-infiltrating NKp46⁺ immune cells ( Figures 1D, E , Supplementary Figure 1I ). In the absence of innate lymphoid cells within Rnls^mut beta cell grafts, MHCII^high+ APCs lost the ability to upregulate the immune checkpoint PD-L1, thus forfeiting the survival advantage over WT beta cells as shown before ( Figures 1G, H ) (10). In summary, these observations clearly demonstrated that graft-infiltrating NKp46⁺ innate lymphoid cells are crucial for the induction of a tolerogenic APC phenotype important to reduce autoimmune destruction of Rnls^mut beta cells.

After showing that NKp46⁺ innate lymphoid cells are important to mediate protective immune-regulation leading to prolonged graft survival of Rnls^mut beta cells, we subsequently investigated if Rnls^mut beta cells directly regulate the activation of innate lymphoid cells. Hence, we purified splenic NKp46⁺ cells (further declared as NK cells as ILC1 only represents approximately 5%–10% of all NKp46⁺ cells in the spleen) (14) by negative selection to collect “untouched” cells for functional in vitro co-culture experiments. Activation of NK cells was achieved by supplementation of interleukin-2 (IL2), well-known to enhance the proliferation and effector function of NK cells (15). Co-culture with Rnls^mut NIT-1 beta cells significantly reduced NK cell activation indicated by impaired expression of activation marker CD44 and CD69 compared to WT, but did not completely abolished NK cell activation ( Figures 2A, B ). However, the surface expression of degranulation marker CD107a (also known as LAMP1) on NK cells, a molecule that predicts cytotoxic activity (16), was strikingly abolished when co-cultured with Rnls^mut NIT-1 beta cells ( Figures 2C, D ). The ability of NK cells to proliferate following repeated IL2 stimulations was also completely prevented when co-cultured with Rnls^mut NIT-1 beta cells. This effect was solely dependent on cell–cell contact as separation of NK cells and Rnls^mut beta cells in trans-well abrogated the anti-proliferative function of Rnls^mut beta cells ( Figures 2E, F ). The cell–cell contact dependency on NK cell regulation by Rnls^mut NIT-1 beta cells also became evident for regulation of activation marker CD69 ( Figure 2G ). As described above, differential gene expression analysis of graft-infiltrated NK cells showed enriched Tgfβ1 when derived from Rnls^mut NIT-1 beta cell grafts ( Supplementary Figure 1A ). To investigate if Tgfβ1 is also upregulated on protein level, we stained NK cells for latency-associated peptide (LAP) representing membrane-bound Tgfβ1. Indeed, NK cells co-cultured with Rnls^mut NIT-1 beta cells demonstrated elevated Tgfβ1 cell surface expression compared to WT NIT-1 beta cells. In contrast to the modulation in NK proliferation, Rnls^mut NIT-1 beta cells drive Tgfβ1 expression independent of cell–cell contact ( Figures 2H, I, L ). The fact that Rnls^mut NIT-1 beta cells significantly upregulated Tgfβ1 expression and secretion in NK cells by soluble factors was demonstrated by two different methods, in a trans-well assay ( Figures 2H–L ) and by supplementation of conditioned media derived from Rnls^mut NIT-1 beta cells in comparison to supernatant of WT cells ( Figures 2K, L ). However, the increased intensity of Tgfβ1 expression on Tgfβ1⁺ cells mediated by Rnls^mut NIT-1 beta cells was dependent on cell–cell contact ( Figure 2J ). Next to Tgfβ1, the pro-inflammatory cytokine Ifng is also known to facilitate the upregulation of PD-L1 (11). However, the secretion of Ifng was not significantly affected when NK cells were co-cultured with Rnls^mut NIT-1 cells compared to WT ( Figure 2M ). In summary, Rnls^mut NIT-1 beta cells directly shape NK cell activity in multiple modes, dependent and independent of cell–cell contact. The presence of NIT-1 beta cells deficient for Rnls causes NK cells to fail to proliferate as well as to lose cytotoxic activity in response to IL2, but at the same time elevate the expression level of anti-inflammatory Tgfβ1.

We demonstrated that Rnls^mut NIT-1 beta cells are potent modulators of NK cell activation ( Figures 1 , 2 ). To explore how Rnls-deficient beta cells might influence the activity of NK cells, we re-analyzed bulk RNA sequencing comparing differential gene expression data of WT and Rnls^mut NIT-1 cells (10). Rnls^mut NIT-1 beta cells showed an enrichment of genes involved in immune-regulatory interactions between a lymphoid and a non-lymphoid cell ( Supplementary Figure 1J ), indicating that upregulated expression of inhibitory cell surface molecules may regulate cell contact-dependent NK cell stimulation. Next, we selected all candidates from the top 1,500 most significant upregulated genes in Rnls^mut NIT-1 cells that are known to be involved in modulation of immune cell activity according to the GSEA-MSigDB data base (https://www.gsea-msigdb.org). In addition to the elevated expression of CD44 and Itga4 important for cell–cell adhesion processes (17, 18), Rnls^mut NIT-1 beta cells also upregulate key immune checkpoint surface molecules such as Ceacam1 (also known as CD66a), CD47 (also known as integrin associated protein), and CD200 (also known as OX-2), which are all eminent potent inhibitors of NK cell activation ( Figure 3A ) ^19-22. To validate if elevated mRNA expression correspond to higher protein expression, we performed cell surface staining for these three key inhibitory NK cell ligands analyzed by flow cytometry. Rnls^mut NIT-1 beta cells showed significant upregulated surface expression of CD47 for both percentage of cells and mean fluorescence intensity (MFI, Figures 3B–D ). The surface expression of Ceacam1 was also significantly upregulated in terms of Ceacam1^high+-expressing cells, whereas Ceacam1 MFI showed a tendency of elevated expression on Rnls^mut NIT-1 beta cells ( Figures 3E–G ). Although the percentage of cells positive for cell surface CD200 was not dramatically changed between WT and Rnls^mut, the expression level of CD200 (MFI) on Rnls^mut NIT-1 beta cells was significantly upregulated ( Figures 3H–J ). These observations indicate that Rnls^mut NIT-1 beta cells could shape NK cell activation at least partially by the upregulation of multiple inhibitory NK cell ligands on their cell surface.

In a first attempt to investigate why Rnls^mut cells upregulate the expression of NK inhibitory molecules, we wondered if the changes in Rnls^mut beta cell metabolism towards increased glucose metabolism (10) could influence the cell surface expression of CD47. As CD47 expression has been described to be similarly regulated like the immune checkpoint PD-L1 (19), and PD-L1 is positively regulated by glycolysis (20, 21), we treated NIT-1 cells with 2-deoxy-d-glucose (2DG) to inhibit glucose metabolism. Strikingly, 2DG treatment for 48 h dramatically downregulated the surface expression of CD47 on both alive WT and Rnls^mut NIT-1 cells ( Supplementary Figures 3A–D ). Whereas treatment with low concentration of 2DG (1 mM) decreased the percentage of CD47-expressing Rnls^mut cells to about the level of untreated WT cells, small amounts of 2DG had no effect on the percentage of CD47 expression on WT cells ( Supplementary Figures 3A, C ). This observation indicates that enhanced glucose metabolism in Rnls^mut beta cells might be responsible for the upregulation of CD47 surface expression.

Rnls^mut NIT-1 beta cells strongly modulate NK cell activity towards a regulatory phenotype. To make sure that our observations are not restricted to mouse-derived NIT-1 beta cells, we also analyzed human iPSC-derived beta-like cells (SCBCs) lacking the RNLS gene (9) for NK regulatory characteristics ( Figure 4A ). Intracellular staining for beta cell marker demonstrated that SCs differentiate into beta-like cells regardless of genotype, as previously described ( Supplementary Figures 4A, B ) (9). Rnls^mut NIT-1 beta cells showed elevated surface expression of NK inhibitory molecules CD47, Ceacam1, and CD200 ( Figure 3 ). In line with Rnls^mut NIT-1 beta cells, RNLS^mut SCBCs also demonstrated significantly upregulated expression of NK inhibitory ligands CD47 and CD66a/c/e, whereas, in contrast to NIT-1 beta cells, CD200 surface expression was not elevated on RNLS^mut SCBCs ( Figures 4B, C , Supplementary Figures 4C–H ). We co-cultured WT and RNLS^mut with allogenic peripheral blood mononuclear cells (PBMCs) stimulated with IL2 for 3 days ( Figure 4A ). Although surface expression of the NK cell activation marker was only modestly changed in this highly stimulatory allogenic setting ( Figure 4D and data not shown), RNLS^mut SCBCs significantly upregulated the immune-regulatory cytokine TGFβ1 on human CD3^-CD56⁺ NK cells ( Figures 4E, F ). These findings clearly show that human beta cells lacking RNLS acquire key characteristics to induce a regulatory NK cell phenotype. Because we previously demonstrated that Tgfβ1 expression is elevated on other NOD-derived immune cells such as CD3⁺ CD4⁺ T cells when infiltrating in Rnls^mut beta cell grafts (10), we stained human CD3⁺ T cells for membrane-bound TGFβ1 as well. Indeed, PBMC-derived CD3⁺ T cells showed elevated surface expression of LAP/TGFβ1 when co-cultured with RNLS^mut SCBCs compared to WT SCBCs ( Figure 4G ). Finally, we tested if RNLS^mut SCBCs had higher resistance to immune-mediated destruction by IL2-stimulated allogenic PBMCs. The total number of alive CD45 negative SCBCs as well as the ratio of alive/dead SCBCs were significantly higher for RNLS^mut SCBC compared to WT indicating that the protective immune-regulatory effects of RNLS-deficient beta cells also translate, at least to some extent, to the human system ( Figures 4H, I ). Our results highlight that knocking out a single gene in beta cells can drive multifaceted modulation on the immunogenicity of beta cells to prevent autoimmunity.

NOD (nonobese diabetic), NOD.scid (NOD.Cg-Prkdc^scid/J), and NSG (NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ) mice were purchased from The Jackson Laboratory. Animals were housed in pathogen-free facilities at the Joslin Diabetes Center and all experimental procedures were approved and performed in accordance with institutional guidelines and regulations (IACUC protocol number 2013–03).

NIT-1 cells were obtained from ATCC (cat # CRL-2055). Cells were maintained in DMEM, high glucose, and pyruvate (cat # 11–995-073; Gibco), supplemented with 2 mM L-glutamine (cat # 25–030-081; Gibco),10% FCS (cat # 10–082-147; Gibco), 50 μM 2-mercaptoethanol (cat # 60–24-2; Sigma-Aldrich), and penicillin/streptomycin (cat # 15140122; Thermo Fisher Scientific) in a 37°C incubator with 5% CO₂. Rnls^mut NIT-1 cells were generated previously (10). Human induced pluripotent stem cells (iPSCs) obtained from Local donor 3 with T1D were carefully maintained and differentiated following robust protocols tailored for beta cell differentiation (31). Cultured in planar form using mTeSR medium (cat# 85850; Stem Cell Technologies), iPSCs subjected to daily media changes to sustain optimal growth conditions. Employing a 2-D differentiation protocol (32), iPSCs were seeded at stage 0 onto culture plates at a density of 0.8 × 105 cells/cm², supplemented with 10 μM Y-27632 (cat # DNSK-KI-15–02; DNSK International) to enhance cell survival and adherence. Quality control measures were rigorously implemented at each differentiation stage, utilizing immunofluorescence (IF) or flow cytometry analyses with specific differentiation marker to assess differentiation efficiency and cellular identity (data not shown). For effective cell dissociation, TrypLE Express (cat # 12604013; Gibco) treatment was administered at 37°C for 10 min to disrupt cell–cell adhesions, facilitating the generation of single-cell suspensions. Following dissociation, cells were promptly fixed and stained as per established procedures (33) to visualize cellular marker indicative of beta cell differentiation. Quality control at stage 6 involved the use of antibodies targeting C-peptide and NKX6.1 analyzed by flow cytometry (see Table 1 ). Ethical approval for all human cell experiments was meticulously obtained from the Harvard University Institutional Review Board (IRB) and the Embryonic Stem Cell Research Oversight (ESCRO) committees, underscoring our commitment to ethical research practices and compliance with regulatory guidelines. RNLS^mut iPSCs from an individual with T1D (Local donor 3) were obtained, generated, and gifted by the Melton Lab (Harvard University) as described before (9).

Bulk RNAseq of WT and Rnls^mut NIT-1 cells were performed previously (10). In brief, triplicates of 2 × 106 NIT-1 cells were used for RNA isolation using Zymo Quick-RNA miniprep plus kit (cat # R1058; Zymo Research), following the manufacturer’s protocol. RNA libraries were prepared, and subsequently quality control tested by Novogen Corporation Inc. Sequencing was performed on an Illumina NovaSeq 6000 sequencing system for a >20 million read data output (Novogen Corporation Inc.). For enrichment analysis of immune system-related genes (according to GSEA-MSigDB, https://www.gsea-msigdb.org) the 1,500 most significant upregulated genes in Rnls mutant vs. WT control are shown ( Figure 3A ).

NK cells were obtained from spleens of 8- to 16-week-old non-diabetic NOD/ShiLtJ mice purchased from The Jackson Laboratory (strain #: 001976; three to four spleens were used per experiment to collect sufficient amount of NK cells) using the NKp46⁺ NK Isolation Kit (cat # 130–115-818; Miltenyi Biotec). For co-culture experiments, 10⁴ NIT-1 cells were seeded in each well of a 96-well flat bottom (trans-well assays) or round-bottom (co-culture assays) plate in 100 μl of NK cell medium (DMEM, high glucose, pyruvate (cat # 11–995-073; Gibco), supplemented with 10% FCS (cat # 10–082-147; Gibco), 2 mM L-glutamine (cat # 25–030-081; Gibco), and penicillin/streptomycin (cat # 15140122; Thermo Fisher Scientific) following incubation for 3 days in a 37°C incubator with 5% CO₂. In trans-well assays, NK cells were fluorescently labeled using Tag-it Violet proliferation dye according to the manufacturers’ protocol (cat # 425101; BioLegend) and 5 × 10⁴ NK cells were added to NIT-1 cell-containing 96-well plates either in direct co-culture or on a trans-well insert (cat # 3380; Corning). Recombinant mouse (m)IL-2 (cat # 575406; BioLegend) was added to stimulate NK cells with a total of 20–100 ng/mL final concentration as indicated. For trans-well assays, 100 ng/mL mIL2 was freshly added daily to achieve stronger NK cell proliferation. To investigate NK cell proliferation, NK cells were stained with Tag-it Viole Proliferation and Cell Tracking Dye (cat # 425101; BioLegend) according to the manufacturer’s instructions. Conditioned media experiments were performed by the addition of medium derived from WT or Rnls^mut NIT-1 cells with a ratio of 1:1 (conditioned media:NK cell media). Additional conditioned medium was supplemented following 24 h (final ratio 3:1) and 48 h (final ratio 4:1) each time in combination with repetitive IL2 stimulation (100 ng/mL). Supernatants were collected following centrifugation for 5 min at 300 g and stored at −20°C. ELISA for Ifng (cat # MIF00; R&D Systems) and Tgfβ1 (cat # DB100C; R&D Systems) were performed according to the manufacturer’s instructions. In all experiments, cells were co-cultured in a 37°C incubator with 5% CO₂ for 72 h, then collected and stained as described in the “Flow cytometry” section with indicated antibodies.

PBMCs collected from allogenic donors (under protocol Joslin CHS # 2013–10) in EDTA tubes were isolated using Lymphoprep density gradient medium (cat # 07801; Stem Cell Technologies). A total of 5 × 10⁶ PBMCs were transferred into each well of a 6-well plate with fully confluent WT or RNLS^mut SCBC in 3 mL of HEPES buffered PBMC medium [RPMI-1640 with penicillin/streptomycin, L-glutamine, and HEPES (cat # ABI-Custom American BioInnovations), supplemented with 10% human AB serum (cat # H5667; Millipore Sigma) and 100 ng/mL recombinant human IL2 (cat # 589104; BioLegend)] following incubation at 37°C with 5% CO₂. Following co-culture for 72 h, cells were detached with Trypsin-EDTA (cat # 25–200-056; Gibco) and filtered using a 70-μm pore size strainer twice to obtain a single-cell solution. Collected cells were stained with indicated antibodies listed in Table 1 as described in the “Flow cytometry” section. The quantification of alive CD45^- SCBCs was performed by using Precision Count Beads (cat # 424902; BioLegend). Cells were collected and stained as described in the “Flow cytometry” section with indicated antibodies.

WT and Rnls^mut NIT-1 beta cells (10⁷ each) were injected subcutaneously (s.c.) into opposite flanks of the same immunodeficient NOD.Cg-Prkdc^scid/J (NOD; strain #: 001303) or NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ (NSG; starin #: 005557) mice purchased from The Jackson Laboratory. Four days later, autoreactive splenocytes from recently diabetic NOD mice were isolated and red blood cells were lysed using 5 mL of ACK lysing buffer (cat # A1049201; Thermo Fisher Scientific) for 4 min at RT. Lysis was stopped using 5 mL of PBS/10% FCS followed by two washing steps using PBS only. Splenocytes were filtered two times in total through a strainer with 70 μm pore size to obtain a single-cell solution. For NK cell depletion, NKp46-expressing splenocytes were removed by using the anti-NKp46 MicroBead kit (cat # 130–095-390; Miltenyi Biotec) according to the manufacturer’s instructions. Splenocytes (10⁷/mouse) were injected intravenously (i.v.) into NIT-1 cell-bearing NOD.scid/NSG recipient mice to transfer autoimmune beta cell killing for a total of 17 days. NIT-1 beta cell graft isolation was performed as described before (10). In brief, grafts were isolated and scaled on an analytical balance 17 days after splenocyte transfer. Grafts were cut into small pieces and digested using HEPES buffered RMPI 1640 (cat # R4130–10L; Sigma-Aldrich) supplemented with 1 mg/mL Collagenase D (cat # 11088858001; Sigma-Aldrich), 20 μg/mL DNase I (cat # EN0521; Thermo Fisher Scientific), 2% FCS, and 50 μg/mL lipopolysaccharide neutralizing agent Polymyxin B sulfate (cat # 1405–20-5; Sigma-Aldrich) for 45 min at 37°C while shaking by maximum speed (750 rpm) on a heating block. Digested grafts were further disaggregated and filtered through a strainer with 70 μm pore size twice to obtain a single-cell solution. For flow cytometry analysis, cells were stained with the antibodies listed in Table 1 in two separate panels (panel 1: T cells and NK cells; panel 2: myeloid cells and B cells).

The scRNAseq experiment was performed in our previous study (10). In brief, WT or Rnls mutant graft-infiltrating immune cells were FACS sorted (CD45⁺ PI^-). Individual samples were labeled using hashtag antibodies and scRNAseq was performed on pooled samples using the Chromium Next GEM Single Cell 3’ GEM, Library & Gel Bead Kit v3.1 (cat # PN-1000213; 10x Genomics) according to the manufacturer’s instructions. Samples were super-loaded with 40,000 cells per reaction. The hashtag oligo library (HTO) was generated separately as described previously (https://citeseq.files.wordpress.com/2019/02/cell_hashing_protocol_190213.pdf). Illumina NovaSeq 6000 with approximately 1.1 billion reads was used to sequence the gene expression library, while the HTO library was sequenced separately by Illumina NextSeq with approximately 130 million reads total each. For further details on QC and analysis procedure, please refer to our previous work (10). In this study, gene set enrichment analyses of graft-infiltrating NK cells and ILC1 were performed by using the molecular signatures of hallmark gene sets provided by the MSigDB database (http://www.gsea-msigdb.org/). The indicated number of differential expressed genes within the indicated immune cell population derived from WT or Rnls^mut NIT-1 beta cell grafts were used as input.

All cell surface staining procedures were performed in PBS/2 mM EDTA/10% FCS. Single-cell suspensions were pre-incubated with FcR-blocking antibody solution (cat # 156604; BioLegend) for 5 min, then stained with antibodies against indicated molecules ( Table 1 ) for 20 min on ice diluted in FcR-blocking solution. Appropriate isotype controls were purchased from BioLegend. Dead cells were excluded using propidium iodide (PI, cat # R37169; Thermo Fisher Scientific) or fixable viability dye eFluor 450 (cat # 65–0863; eBiosciences). For intracellular staining (SCBC QC), the Transcription Factor Staining Buffer Set (cat # 00–5523-00; Thermo Fisher Scientific) was used according to the manufacturer’s instruction. In brief, following surface staining of CD47, CD66/a/c/e, and CD200 performed as described above using human FcR blocking solution (cat # 422302; BioLegend), cells were washed with PBS/2 mM EDTA/10% FCS following cell fixation with Fixation/Permeabilization working solution and pulse vortexing. Cells were incubated for 90 min at 4°C and washed following 2 permeabilization steps using the provided Permeabilization Buffer. Cells were pre-incubated in 5% human FcR-blocking antibody solution for 15 min at room temperature (RT). Cells were stained with indicated antibodies against C-peptide and NKX6.1 (working dilution of 1:20) listed in Table 1 for 30 min at RT. Cells were washed twice, kept on ice and analyzed by flow cytometry. For the 2DG experiment to inhibit glucose metabolism in NIT-1 cells, 2 × 10⁵ cells of indicated genotypes were seeded into a 24-well plate for 48 h with or without addition of indicated concentrations of 2DG or sterile dH₂O as control. Cell surface staining of indicated molecules were performed as described above. Prior to analysis, all cell suspensions were filtered using a 35-µm strainer (cat # 352235; Corning) to avoid cell cluster. Data were acquired on an LSRII instrument (BD Biosciences) and analyzed with FlowJo software v.10.6.1 (FlowJo LLC). All data are shown using log-scale axes.

Statistical analyses were performed by unpaired or paired tests or by one-way ANOVA as indicated using GraphPad Prism v 9.0.2 software. All data are presented as the mean ± standard deviation (SD). p < 0.05 was considered statistically significant. Sufficient sample size was estimated without the use of a power calculation. No samples were excluded from the analyses. No randomization was used for the animal experiments. Data analysis was not blinded. All data are representative of two or more similar experiments.