Single-cell profiling identifies a CD8bright CD244bright Natural Killer cell subset that reflects disease activity in HLA-A29-positive birdshot chorioretinopathy.

doi:10.21203/rs.3.rs-2272009/v1

Download PDF

Article

Single-cell profiling identifies a CD8^bright CD244^bright Natural Killer cell subset that reflects disease activity in HLA-A29-positive birdshot chorioretinopathy.

https://doi.org/10.21203/rs.3.rs-2272009/v1

This work is licensed under a CC BY 4.0 License

Journal Publication

published 31 Jul, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

MHC-I-opathies are inflammatory conditions strongly associated with HLA class I genes. The striking association with HLA class I suggests involvement of T cells, whereas natural killer (NK) cell involvement remains largely unstudied. Here we show that HLA-A29-positive birdshot chorioretinopathy patients have a skewed NK cell pool containing expanded CD16-positive NK cells which produce more proinflammatory cytokines. These NK cells contain populations that express the HLA class I restricted antigen CD8A which display gene signatures indicative of high cytotoxic activity (IGFBP7, MYOM2, and LINC00996), and signaling through NK cell receptor CD244 (SH2D2A and SH2D1B). Long-term monitoring of a cohort of birdshot chorioretinopathy patients with active disease identified a population of CD8bright CD244bright NK cells, which rapidly declined to normal levels upon clinical remission following successful treatment. Collectively, these studies implicate CD8bright CD244bright NK cells in the sight-threatening MHC-I-opathy, birdshot chorioretinopathy.

Biological sciences/Immunology/Innate immune cells/Innate lymphoid cells

Biological sciences/Immunology/Autoimmunity

Biological sciences/Immunology/Neuroimmunology

Non-infectious uveitis (NIU) is a clinically and prognostically heterogeneous group of ocular inflammatory diseases and a major cause of severe visual handicap¹. Birdshot chorioretinopathy (Birdshot Uveitis or BCR-UV) is a relatively rare form of NIU that has clinically distinct features in the form of retinal and choroidal inflammatory lesions visible on examination. BCR-UV can lead to progressive deterioration of visual function^{2, 3, 4} due to persistent inflammation in the retina and choroid⁵. Consequently, patients often require systemic immunomodulatory therapy to control ocular inflammation^{3, 6}. The disease mechanisms driving BCR-UV remain to be elucidated, but scientific advances have shed light on the immunopathology of this clinically well-defined ophthalmological condition⁷.

One of the most striking molecular features of BCR-UV is that it’s strongly associated with the presence of HLA-A29 allele such that all patients carry the HLA-A29 allele (most often the common HLA-A*29:02)^{7, 8}. Genome-wide genetic studies have revealed that the susceptibility to BCR-UV also maps to other factors of the MHC-I pathway^{9, 10, 11}. This incriminates CD8 + T cells and Natural Killer (NK) cells in the pathogenesis of BCR-UV⁷. Indeed, CD8 + T cells have been shown to infiltrate eye tissues of patients and secrete cytokines considered central to its immunopathology^{12, 13, 14, 15}.

The role of NK cells in BCR-UV has remained significantly underexplored. Classically, blood NK cells are subdivided into two well-established overarching populations; the CD56^dim CD16 + NK cells (~ 90% of circulating NK cells) also known as NK1 cells which are considered to be cytotoxic and produce greater amounts of pro-inflammatory cytokines and the CD56^bright (~ 10% of circulating NK cells) - or NK2 cells - that are considered to be more immunoregulatory¹⁶. Of interest, CD56^bright NK cells have been shown to be lower in patients with NIU, including BCR-UV patients with active uveitis¹⁷ and successful immunosuppressive treatment of NIU is accompanied by the recovery of CD56^bright NK cell levels¹⁸. However, beyond this phenotypic bifurcation, NK cell population is substantially more diverse and thus far single-cell RNA sequencing (scRNAseq) has uncovered at least 10 transcriptionally distinct clusters in peripheral blood^{19, 20, 21}. This includes ‘inflammatory’ CD56^dim CD16 + NK subsets that was defined by high levels of cytokine and interferon response genes^{20, 21} and “adaptive-like” NK populations that expand during infection²².

Transcriptomically distinct NK cell subsets are considered to exhibit differential effector functions mediated by an ensemble of surface immunoregulatory molecules²³, in particular Killer cell immunoglobulin-like receptors (KIR), IgG Fc receptors (e.g., CD16), and integrins (e.g., CD47)^{24, 25, 26}. Perturbations in the composition of the NK cells have been reported in other MHC-I associated conditions, such as HLA-B27-positive ankylosing spondylitis²⁷ and were shown to be predictive for clinical outcome in autoimmune diseases, such as Multiple Sclerosis²⁸. Collectively, these observations suggest that deep phenotyping of the blood NK cell compartment could provide better understanding of disease biology and may hold clinically relevant information to the clinical course of BCR-UV.

Here, we have taken a multi-omics approach to phenotype the NK cell compartment at single-cell resolution of patients with BCR-UV and report on the expansion of a CD56^dim CD16 + subset of NK cells which are CD8^bright and CD244^bright, and whose reduction is correlated with clinical improvement after systemic immunosuppressive therapy.

Increased frequency of CD16 + NK cells in Birdshot uveitis patients

To investigate the relationship between NK cell-mediated inflammation and BCR-UV, we first sought evidence that circulating NK cells were specifically perturbed in BCR-UV. To this end, we quantified the major lineages of immune cells (10-marker panel, 6 lineages) (Supplementary Fig. 1A) in peripheral blood using flow cytometry in a cohort of 18 BCR-UV patients, 80 healthy controls, and 121 non-infectious uveitis (NIU) patients other than BCR-UV (Fig. 1A). Global comparison of major lineages in all NIU patients versus healthy controls revealed a significant increase in frequency of blood NK cells (Fig. 1B), but not T cells, B cells, monocytes, or dendritic cells (Supplementary Fig. 1B-D). Flow cytometry analysis revealed that blood NK cell frequency appeared to be increased in several uveitis subtypes, but this increase was most significant for BCR-UV (P < 0.0001) (Fig. 1C). This became more evident after quantification of the two major NK populations (i.e., NK1, and NK2) that can be distinguished by their expression of surface CD56 and CD16 (Fig. 1D). We detected a significant increase of CD56^dim CD16+ [NK1] cells and a concomitant decrease of the CD56^bright CD16⁻ [NK2] cells only in BCR-UV patients or when considering all NIU patients collectively, but not individually in any of the other types of NIU (Supplementary Fig. 1E-F). This skew in NK1/NK2 balance also remained evident after strict comparison to 15 age-matched healthy controls (mean age ± SD = 62.2 ± 8.8) (Fig. 1E).

Importantly, NK cells of BCR-UV patients showed enhanced responsiveness to restimulation by production of significantly higher tumor necrosis factor-ɑ (TNF-ɑ, P = 0.007) and interferon-γ (IFN-γ, P = 0.002) (Fig. 1F), indicating that the altered NK1/NK2 balance results in a more pro-inflammatory NK repertoire. Collectively, these data show an imbalance in NK1/NK2 cells in peripheral blood of patients with BCR-UV and a skew towards a more proinflammatory phenotype.

PBMC scRNA-seq identifies altered NK repertoire in Birdshot Chorioretinopathy

To allow characterization of the changes in peripheral blood NK cells in BCR-UV in an unbiased manner, we used single-cell RNA-sequencing (scRNAseq) of peripheral blood mononuclear cells (~ 300K cells) of 24 BCR-UV patients and healthy controls (Fig. 2A and Supplementary Fig. 2A). Unsupervised clustering followed by uniform manifold approximation and projection (UMAP) and automated cell type annotation, identified an NK cell population (9,619 cells of cluster C4, Fig. 2B and Supplementary Fig. 2B) with an altered NK cluster structure in two-dimensional UMAP space in BCR-UV patients compared to healthy controls (Supplementary Fig. 2C). NK-specific GZMB (granzyme B), KLRD1, GNLY, PRF1 (perforin), NKG7 and SH2D1B (CD244 signaling) were among the most differentially upregulated genes in BCR-UV (Supplementary Fig. 2D). We extracted the NK cells from the PBMC scRNAseq data for further analysis. Unsupervised clustering of the NK cell population revealed a high level of transcriptomic heterogeneity and the existence of 12 distinct clusters ranging from 57 cells (cluster 11) to 2,093 cells (cluster 0) in each cluster (Fig. 2C and Supplementary Fig. 3A). Gene expression levels of characteristic NK lineage surface markers revealed that these clusters expressed different levels of transcripts encoding NK activating receptors CD244 and CD8A, as well as NK inhibitory receptors KIR3DL1, KLRD1 and B3GAT1 (Fig. 2D) which is compatible with an altered composition of functional NK subsets. At a false discovery rate of 5%, cluster 1 was significantly decreased (948 cells in HC vs 275 cells in BCR-UV) while clusters 2 (351 vs 862 cells in HC vs BCR-UV), 6 (208 vs 457 cells in HC vs BCR-UV) and 10 (25 vs 235 cells in HC vs BCR-UV) were significantly increased in frequency in BCR-UV patients compared to healthy controls (Fig. 2E). Clusters 1, 2, 6 and 10 were uniquely represented by the expression of MYOM2, SH2D1B, IGFBP7 and LINC00996 genes, respectively (Supplementary Fig. 3B).

Cluster 1 also showed high expression of DUSP1, FOS, JUN, and CD69 highly reminiscent of the gene expression profile of CD56^bright CD16⁻ NK cells [i.e., NK2]²⁰, which corroborates our findings by major lineage flow cytometry (Fig. 2F). In contrast, the increased clusters 2, 6 and 10 expressed lower levels of NCAM1 (CD56) (Fig. 2D) but high levels of FCGR3A (CD16) and the surface co-receptor encoding CD8A (CD8 Antigen, Alpha Polypeptide) (Fig. 2D), which suggests that clusters 2, 6, 10 are subpopulations of the bulk population of CD8 + NK1 cells.

We further observed that cluster 10 showed high CD244 (Fig. 2D, F) and clusters 2 and 10 displayed enrichment of CD244 binding Src homology 2 (SH2) domain-encoding genes SH2D2A and SH2D1B, which control signal transduction through the surface receptor CD244²⁹ (Fig. 2F). This implicates active CD244 signaling in these NK clusters. Other highly expressed activation-associated genes in these sub-clusters include TNFRSF18 (also known as GITR) and ISG15 (Interferon-Stimulated Protein, 15 kDa) (Fig. 2F, Supplementary Fig. 3C, D). In summary, these results indicate expansion of activated CD8 + NK1 subpopulation characterized by high levels of CD244-signaling molecules in the blood of patients with BCR-UV.

High-dimensional cytometry reveals accumulation of a CD8^bright CD244^bright subset of CD16+ NK cells in the circulation of Birdshot Chorioretinopathy patients

We wished to validate our scRNAseq findings using a 12-marker panel (Supplementary Fig. 4A) flow cytometric phenotyping of the blood NK cell repertoire. Unbiased cell clustering considering the surface marker phenotypes by FlowSOM discerned 12 NK cell clusters (Fig. 3A). As expected, the majority of the clusters were CD56^dim CD16 + NK1 (cluster 0–2, 5–11) and a minor population was CD56^bright and CD16⁻ (NK2, cluster 3, 4) (Supplementary Fig. 4B). These 12 flow cytometry clusters broadly intersected with the NK clusters detected by the scRNAseq analysis: scRNAseq clusters 1 and 3 (Fig. 2C) are represented by the flow cytometry clusters 3 and 4 (Fig. 3A), all of which are CD56^bright population. The remaining 10 clusters in both scRNAseq and flow cytometry analyses are CD56^dim and CD16 + populations. Differential cluster abundance analysis revealed that clusters 4 and 5 were significantly reduced in the BCR-UV, while cluster 0 was significantly increased (Fig. 3B, C). The NK2 cluster 4 was further defined by high expression of CD336 and CD94 whereas the expanded NK1 cluster 0 was defined by high co-expression of CD8 and CD244, in line with our scRNA-seq data. We further found that cluster 0 expressed surface markers CD314 (NKG2D) and CD337 (NKp30), but not CD57 (Fig. 3D, E and Supplementary Fig. 4E). Principal component analysis supported that cluster 0 was the most distinguished cluster by high co-expression of CD8a and CD244 (Fig. 3E).

We determined by in vitro culture that CD244 expression but not CD8 was upregulated in NK cells by restimulation with IL-15 and IL-18, indicating that CD8a + CD244 + cells may represent activated CD8 + NK cells (Fig. 3F). To validate these findings, we manually gated the NK cell population based on the relative expression of CD8a and CD244 and divided the NK cells into three categories: CD8a^bright/CD244^bright (hi), CD8a/CD244 ^medium (mid) and CD8a/CD244 ^low (lo) expressing cells. Quantitative analysis demonstrated that the CD8a^bright/CD244^bright population was significantly more abundant in BCR-UV compared to that of healthy controls (Student’s t-test, P < 0.01) (Fig. 3G, H and Supplementary Fig. 4C, D). Together, these data show that CD8a^bright/CD244^bright NK1 cells are enriched in the blood of BCR-UV patients.

CD8a^bright CD244^bright cytotoxic NK cell frequency correlates with disease activity

Finally, we were interested to determine the dynamics of the newly identified NK cell subset in BCR-UV patients over the course of the disease. We had the opportunity to analyze samples taken prior to commencing therapy, during, and upon achieving clinical quiescence (according to the SUN criteria³⁰) following treatment with systemic immunomodulatory therapy (1-year follow-up) (Fig. 4A, Supplementary Fig. 5A). We assessed the expression of CD8a and CD244 in the CD56^dimCD16 + NK1 cell population in this cohort by flow cytometry. The mean fluorescent intensity (MFI) of surface expression for both CD8a and CD244 was elevated in patients with active BCR-UV and decreased over the course of treatment (Fig. 4B and Supplementary Fig. 5B). Similarly, the CD8a and CD244 double positive cells within CD56^dimCD16 + NK population were significantly increased in patients with active BCR-UV and gradually decreased with one year of treatment and normalized to the frequency observed in healthy controls (P < 0.05) (Fig. 4C, D). In conclusion, these results show that CD8a^bright/CD244^bright NK1 cells are expanded during active uveitis in BCR-UV patients but decrease upon successful systemic immunomodulatory treatment and clinical remission, compatible with the interpretation that CD8a^bright/CD244^bright NK1 cells are a pro-inflammatory NK subset that are likely to be involved in the underlying disease mechanism.

In this study, we conducted deep molecular phenotyping of peripheral blood NK cells and identified altered changes in NK1/NK2 subsets in peripheral blood of BCR-UV patients. We found expansion of a CD8a^bright CD244^bright NK1 subset in BCR-UV patients that decreased upon successful treatment with systemic immunomodulatory therapy. Our findings also corroborate previous reports on the decreased CD56^bright NK cells (NK2) in BCR-UV¹⁷ and the normalization of CD56^bright NK cell abundance upon immunosuppressive treatment of non-infectious uveitis¹⁸. In other MHC-I associated conditions, such as HLA-B51-associated Behcet’s disease (BD) and HLA-B27-associated ankylosing spondylitis (AS), NK1/NK2 changes have been reported. In BD, total NK cells are increased in frequency in blood^{31, 32} and produce increased IFN-gamma and TNF-alpha^{33, 34}, which we also demonstrate in BCR-UV. In AS, the number of CD56^dim CD16 + subset of NK cells (NK1) in the peripheral blood is increased,^{35, 36, 37} which is in line with our finding of elevated CD16 + CD56^dim NK cells in BCR-UV. Here, we add to these previous observations that the decline in CD56^bright regulatory NK cells in BCR-UV is accompanied by a concomitant expansion of CD56^dim CD16 + NK cells, that also express the alpha-chain of the CD8 co-receptor for HLA class I^{38, 39}, advancing the understanding of NK cell dynamics in ocular inflammatory disease. The subset of CD8 + NK cells (also known as “NK8⁺” cells²⁸) make up approximately half of the blood NK cells⁴⁰ and are found in both the CD56^dimCD16 + and CD56^brightCD16 + NK subsets (Supplementary Fig. 6A). Note that in contrast to CD8 + T cells, which express the alpha and beta chain of CD8, NK8⁺ cells only express the alpha chain of CD8 and mouse NK cells do not express CD8 at all^{38, 39}.

NK8⁺ cells are considered to be functionally distinct and produce more IFN-gamma and TNF-alpha compared to CD8– NK cells⁴¹. Accordingly, our data show that the NK8⁺-enriched population in patients secretes more IFN-gamma and TNF-alpha compared to healthy controls. In addition, the activation marker CD69 was increased in patients (Supplementary Fig. 6B) and was induced upon treatment with IL-15 or IL-18 (Fig. 3F). This is significant because cross-linking of CD8 on NK cells induces increased expression of the activation marker CD69 at the cell surface⁴². Of interest, elevated frequency of peripheral blood NK8⁺ cells with homing marker CXCR3 is associated with an increased risk for Type 1 diabetes, a T-cell mediated autoimmune condition ⁴³. The most expanded NK8⁺ subpopulation identified by scRNAseq (cluster 10) in BCR patients was also characterized by high CXCR3 expression (Supplementary Fig. 3C). Whether this NK8⁺ subset (i.e., cluster 10) may directly contribute to eye inflammation in BCR-UV, remains to be determined. Alternatively, the NK8⁺ skewing may be a reflection of diminished regulatory capacity in the NK cell compartment. NK cells are required to suppress CD8 + T cell autoimmunity⁴⁴ which is attributed to the negative immunoregulation of activated T cells by CD56^bright NK cells⁴⁵. As shown in this study, CD56^bright NK cells are diminished in circulation of NIU patients, including BCR-UV patients. This makes it tempting to speculate that proinflammatory skewing of the NK cell population diminishes control of autoreactive CD8 + T cell immunity directed towards the eye in BCR-UV.

Detailed immunoprofiling by scRNAseq and flow cytometry revealed that expanded NK8⁺ cells in BCR-UV co-express high levels of other functional receptors, including CD244, which we demonstrate is a marker for activated NK cells. CD244 [or SLAMF protein 2B4] is an immunoregulatory surface receptor expressed by NK cells^{46, 47}. Interaction of CD244 on NK cells with it’s ligand CD48 expressed by other cells causes the intracellular domain of CD244 to recruit proteins EAT2 and TSAd encoded by SH2D1B and SH2D2A, which were also associated with BCR-UV. The concomitant expression of TNFRSF18 (Tumor Necrosis Factor Receptor Superfamily, Member 18) and ISG15 (Interferon-Stimulated Protein, 15 kDa) levels in the novel NK subset supports the ‘inflammatory’ phenotype of this NK subset in BCR-UV patients. This is also supported by flow cytometry analysis that revealed expression of the activating receptor NKG2D (CD314) and the cytotoxicity receptor (NCR) NKp30 (CD337)⁴⁸ by the CD8 + CD244 + NK cells in BCR-UV.

We showed that the CD244^bright NK8⁺ cells correlate with disease activity during longitudinal monitoring of patients treated with systemic immunomodulatory therapy. The NK8⁺ cell frequency has previously been shown to correlate with clinical parameters in several conditions, including HIV-1 and multiple sclerosis (MS)^{28, 41}. More specifically, in MS, the NK8⁺ cell frequency was predictive of the relapse rate in a longitudinal cohort²⁸. It would be interesting to use flow cytometry analysis of this cell subset to see if its abundance can be used to predict treatment outcome or clinical course in advance (measured at diagnosis).

In conclusion, using complementary immunophenotyping platforms, we identified an expanded CD8a^bright CD244^bright population of circulating NK cells in BCR-UV whose abundance reflects inflammatory disease activity. Better understanding of the molecular underpinnings of BCR-UV and its relation to clinical outcome may pave the way towards implementation of more effective personalized therapeutic approaches in ocular inflammatory diseases.

Patients

This study was conducted in compliance with the Declaration of Helsinki and ethical principles regarding human experimentation. All samples were obtained under a National Institutes of Health (NIH) Institutional Review Board (IRB) approved protocol (Uveitis/Intraocular Inflammatory Disease Biobank (iBank); NCT02656381). Informed consent was obtained from all enrolled participants. We recruited 139 adult patients (Supplementary Table 1) with non-infectious uveitis (NIU), including 18 patients with Birdshot Chorioretinopathy uveitis (BCR-UV) at the National Eye institute (NEI) outpatient clinic. In total, 80 healthy donors- majority (iBank IDs) of whom were screened for no personal or family history of autoimmune diseases, were recruited and served as unaffected healthy controls. NIU was classified and graded in accordance with the SUN classification³⁰. All patients with BCR-UV were HLA-A29-positive confirmed by HLA typing. A retrospective review of patient charts, fluorescein angiography (FA), indocyanine green angiography (ICGA), and electroretinography (ERG) was performed in order to determine disease activity, clinical course and response to systemic immunomodulatory treatment. Patient demographics are summarized in Supplementary Table 1.

Blood sample processing

Blood samples from patients and healthy controls are collected through venipuncture and all the samples were processed within 4 h of blood collection. Fresh whole-blood samples were directly used for flow cytometry analysis. PBMCs were purified by standardized density gradient isolation (Ficoll-Paque) and stored in liquid nitrogen until further use.

Flow cytometry

Three mL of whole blood was incubated with 30 mL 1x RBC lysis buffer (BioLegend #420391) at room temperature for 15 minutes, centrifuged at 400 x g for 5 minutes and resuspended in 3 ml ACK lysis buffer (Lonza #BP10-548E) for 4 minutes at room temperature. Cell suspension was washed in a 30 ml FACS buffer (FB: 1x PBS w/o calcium and magnesium chloride + 2% FBS + 2 mM EDTA + 0.01% NaN₃). In total, 1-3x10⁶ cells were stained for 45 minutes following resuspension in 100 uL FACS buffer with Fc block (Human TruStain FcX™, Biolegend # 422302) with the following antibodies (Supplementary Table 2): Alexa Fluor® 488 anti-human CD3 Antibody (clone OKT3), Alexa Fluor® 700 CD16 (clone 3G8), APC anti-human CD57 Antibody (clone HNK-1), APC-Fire 750 CD14 (clone 63D3), APC-Fire 750 CD8 (clone SK1), Biotin anti-human CD19 Antibody (clone HIB19), Biotin anti-human CD20 Antibody (clone 2H7), Biotin anti-human CD3 Antibody (clone SK7), Biotin anti-human CD56 Antibody (clone HCD56), BV 421™ anti-human CD158e1 Antibody (clone DX9), BV 421™ anti-human CD19 Antibody (clone HIB19), BV 510™ anti-human CD20 Antibody (clone 2H7), BV 510™ anti-human CD56 Antibody (clone HCD56), BV 650™ anti-human CD314 Antibody (clone 1D11), BV 650™ anti-human HLA-DR Antibody (clone L243), FITC anti-human CD4 antibody (clone SK3), FITC anti-human CD94 antibody (clone DX22), PE anti-human CD336 Antibody (clone P44-8), PE/Cy7 anti-human CD337 Antibody (clone AF29-4D12), PE/Dazzle™ 594 anti-human CD244 Antibody (clone C1.7), PerCP/Cy5.5 anti-human CD3 Antibody (clone SK7). BV 605 Streptavidin and Pe/Cy7 streptavidin were used for secondary staining. The Pe/Cy7 Lin cocktail in monocyte/DC panel includes anti-human CD3, CD19, CD20 and CD56 Antibodies. Live/dead staining was carried out using Fixable Viability Dye eFluor® 455UV concomitantly with surface staining. Cells were fixed with Fixation Buffer (Biolegend #420801) and resuspended in 300 uL FACS buffer for acquisition on BD LSR Fortessa 1–3 days after staining.

Data were analyzed using FlowJo v10. For ex vivo restimulation, NK cells from a healthy donor were cultured with recombinant human IL-15 and IL-18 using five serial dilutions as indicated of ED₅₀ concentrations. Cell-surface expression of CD69, CD244 and CD8 was measured using a BD LSRFortessa and analyzed by FlowJo V.10.

Intracellular cytokine staining

For intracellular cytokine staining, cryopreserved PBMCs were thawed into warm RPMI/10% FBS, washed once in cold PBS and divided cells in two equal proportions. Protein Transport Inhibitor Cocktail (eBioscience #00-4980-03, 1/500 vol/vol) was added to each proportion. Cells were incubated with cell stimulation cocktail (eBioscience # 00-4970-03, 1/500 vol/vol) for 4 hours at 37 degrees or kept under similar conditions without the stimulation cocktail. Cells were washed with cold PBS and resuspended in 100 uL FBS with Fc block (Human TruStain FcX™, Biolegend # 422302) followed by surface markers staining for 45 min. Cells were then washed once in cold PBS and incubated in fix/perm buffer (eBioscience FOXP3/Transcription Factor Staining Buffer, invitrogen #00-5523-00) for 20 min and then incubation with the following intracellular antibodies for another 30 min: BV510 anti-human Granzyme B (clone GB11), APC/Cyanine7 anti-human TNF-α (clone MAb11), PE/Dazzle 594 anti-human IFN-gamma (clone 4S.B3) and PE anti-human Perforin (clone B-D48).

FlowSOM analysis

Live NK cells (Aqua L/D-CD3-CD20-CD56+) were gated from 18 BCR-UV patients sampled at baseline (i.e., at disease onset or relapse) and 10 healthy controls. NK cell fraction (range 5,011 to 42,358 NK cells) was down-sampled to 5,000 NK cells per donor using FlowJo DownsampleV3 plugin, concatenated and subjected to t-distributed stochastic neighbor embedding (t-SNE, iteration 1000, perplexity 30). The FlowSOM plugin in FlowJo was used to cluster cells into 12 meta-clusters (following 12 clusters identified in scRNAseq).

PBMC single cell RNA sequencing

PBMCs from a total of 24 BCR-UV patients and healthy controls (Supplementary Table 3) were thawed quickly at 37°C and resuspended in RPMI media supplemented with 10% FBS. Approximately, 1,000–1,200 viable cells per microliter were loaded for capture onto the Chromium System using the v2 single-cell reagent kit (10X Genomics). Following capture and lysis, scRNA-Seq libraries were prepared from ~ 16k cells using a 10X Genomics Chromium device and Chromium Single Cell 3’ Reagent Kit v3.1 according to manufacturer’s protocol. GEX (transcriptome) libraries were sequenced on a NovaSeq 6000 DNA sequencer (Illumina, Inc.) using V1.5 chemistry, generating ~ 160M GEX reads per sample and ~ 40M ADT reads per sample. The raw data was processed using RTA 3.4.4. Sequencing information is shown in Supplementary Table 4.

Preprocessing of scRNAseq data

The raw fastq data were processed using cellranger v6.0.0⁴⁹ with genome assembly GRCh38 (hg38), 3' and 5' assay chemistry SC3Pv3, and an expected cell count of 10,000 per sample (Supplementary Table 5). Further analysis of the single cell data was done using the Seurat v4.0.5⁵⁰ package in the R v4.1.0 environment⁵¹. Cells with fewer than 300 genes and number of transcript counts less than 1,000 and more than 12,000, more than 8% mitochondrial and 40% ribosomal fraction were excluded from the dataset. Doublets were detected and removed using the doubletFinder v3.0 package⁵² in R, set with an expected doublet level of 7%. Data were normalized using the “LogNormalize” method with the scaling factor set at 10,000. The variables, sample batch, percent mitochondria, percent ribosomes, transcript counts and gene counts, were regressed out using the ScaleData function.

Dimension reduction, clustering and visualization

Cells were clustered using Principal Component Analysis (PCA) using the RunPCA function. The first 20 PCs identified (by Elbow method) were used in the ‘FindNeighbors’ (based on k-nearest neighbor (KNN) graphs) and ‘FindCluster’ (Louvain algorithm) functions in Seurat. RunTSNE and RunUMAP functions were used with “pca” as the reduction method, to visualize the data. The FindAllMarkers function was used to identify cluster specific markers. Cell type annotation was generated manually combining automatic annotation results from ScType⁵³, with “Immune System” set as the tissue type, and SCSA⁵⁴, with the whole database as reference. Outcomes from ScType and SCSA were used to curate cluster annotation and identify NK cells by plotting expression of lineage-specific genes. Differences in proportions of scRNAseq data between the groups were assessed using the “scProportionTest” package⁵⁵, with number of permutations set at 10,000.

Data availability

All raw data and workflow are available as an open-source resource, with documentation.

BCR-UV: birdshot chorioretinopathy uveitis; NIU: non-infectious uveitis; KIR: killer cell immunoglobulin-like receptors; scRNAseq: single-cell RNA sequencing; MFI: mean fluorescent intensity; SUN: standardization of uveitis nomenclature.

FUNDING

The work is funded by the NEI intramural research program (project number 1ZIAEY000556-04) and made possible by the Lasker Clinical Research Grant to HNS, and the Vision Grant from Prevention of Blindness Society of Metropolitan Washington to PRN.

ACKNOWLEDGEMENT

We acknowledge the NIAMS Flow Core (James Simone, Kevin Tinsley, Jeffrey Lay), NEI Flow Core (Rafael Villasmil, Jyothi Priya Mandala) and Jenny McDowell at NISC for help with this project.

Rothova, A. et al. Causes and frequency of blindness in patients with intraocular inflammatory disease. Br J Ophthalmol 80, 332–336 (1996).
Howe, L.J. et al. Choroidal abnormalities in birdshot chorioretinopathy: an indocyanine green angiography study. Eye (Lond) 11 (Pt 4), 554–559 (1997).
Kiss, S. et al. Birdshot retinochoroidopathy. Int Ophthalmol Clin 46, 39–55 (2006).
Shah, K.H. et al. Birdshot chorioretinopathy. Surv Ophthalmol 50, 519–541 (2005).
Testi, I. et al. Peripheral Capillary Non-perfusion in Birdshot Retinochoroiditis: A Novel Finding on Ultra-widefield Fluorescein Angiography. Ocul Immunol Inflamm 28, 1192–1195 (2020).
You, C. et al. Long-term outcomes of systemic corticosteroid-sparing immunomodulatory therapy for Birdshot Retinochoroidopathy. Ocul Immunol Inflamm 28, 966–974 (2020).
Kuiper, J.J.W. & Venema, W.J. HLA-A29 and Birdshot Uveitis: Further Down the Rabbit Hole. Front Immunol 11, 599558 (2020).
Kuiper, J. et al. The immunopathogenesis of birdshot chorioretinopathy; a bird of many feathers. Prog Retin Eye Res 44, 99–110 (2015).
Kuiper, J.J. et al. A genome-wide association study identifies a functional ERAP2 haplotype associated with birdshot chorioretinopathy. Hum Mol Genet 23, 6081–6087 (2014).
Kuiper, J.J.W. et al. Functionally distinct ERAP1 and ERAP2 are a hallmark of HLA-A29-(Birdshot) Uveitis. Hum Mol Genet 27, 4333–4343 (2018).
Gelfman, S. et al. ERAP1, ERAP2, and Two Copies of HLA-Aw19 Alleles Increase the Risk for Birdshot Chorioretinopathy in HLA-A29 Carriers. Invest Ophthalmol Vis Sci 62, 3 (2021).
Pulido, J.S. et al. Histological findings of birdshot chorioretinopathy in an eye with ciliochoroidal melanoma. Eye (Lond) 26, 862–865 (2012).
Kuiper, J.J. et al. Detection of choroid- and retina-antigen reactive CD8(+) and CD4(+) T lymphocytes in the vitreous fluid of patients with birdshot chorioretinopathy. Hum Immunol 75, 570–577 (2014).
Kuiper, J.J. et al. Intraocular interleukin-17 and proinflammatory cytokines in HLA-A29-associated birdshot chorioretinopathy. Am J Ophthalmol 152, 177–182.e171 (2011).
Dagur, P.K. et al. Secretion of interleukin-17 by CD8 + T cells expressing CD146 (MCAM). Clin Immunol 152, 36–47 (2014).
Nielsen, N. et al. Cytotoxicity of CD56(bright) NK cells towards autologous activated CD4 + T cells is mediated through NKG2D, LFA-1 and TRAIL and dampened via CD94/NKG2A. PLoS One 7, e31959 (2012).
Verhagen, F.H. et al. High-Dimensional Profiling Reveals Heterogeneity of the Th17 Subset and Its Association With Systemic Immunomodulatory Treatment in Non-infectious Uveitis. Front Immunol 9, 2519 (2018).
Li, Z. et al. Cutting edge: in vivo blockade of human IL-2 receptor induces expansion of CD56(bright) regulatory NK cells in patients with active uveitis. J Immunol 174, 5187–5191 (2005).
Brownlie, D. et al. Expansions of adaptive-like NK cells with a tissue-resident phenotype in human lung and blood. Proc Natl Acad Sci U S A 118 (2021).
Smith, S.L. et al. Diversity of peripheral blood human NK cells identified by single-cell RNA sequencing. Blood Adv 4, 1388–1406 (2020).
Yang, C. et al. Heterogeneity of human bone marrow and blood natural killer cells defined by single-cell transcriptome. Nat Commun 10, 3931 (2019).
Béziat, V. et al. NK cell responses to cytomegalovirus infection lead to stable imprints in the human KIR repertoire and involve activating KIRs. Blood 121, 2678–2688 (2013).
Sivori, S. et al. TLR/NCR/KIR: Which One to Use and When? Front Immunol 5, 105 (2014).
Lanier, L.L. Up on the tightrope: natural killer cell activation and inhibition. Nat Immunol 9, 495–502 (2008).
Nath, P.R. et al. CD47 Expression in Natural Killer Cells Regulates Homeostasis and Modulates Immune Response to Lymphocytic Choriomeningitis Virus. Front Immunol 9, 2985 (2018).
Nath, P.R. et al. Natural Killer Cell Recruitment and Activation Are Regulated by CD47 Expression in the Tumor Microenvironment. Cancer Immunol Res 7, 1547–1561 (2019).
Ren, C. et al. Single-cell RNA-seq reveals altered NK cell subsets and reduced levels of cytotoxic molecules in patients with ankylosing spondylitis. J Cell Mol Med 26, 1071–1082 (2022).
McKinney, E.F. et al. A CD8(+) NK cell transcriptomic signature associated with clinical outcome in relapsing remitting multiple sclerosis. Nat Commun 12, 635 (2021).
Agresta, L. et al. The Emerging Role of CD244 Signaling in Immune Cells of the Tumor Microenvironment. Front Immunol 9, 2809 (2018).
Classification Criteria for Birdshot Chorioretinitis. Am J Ophthalmol 228, 65–71 (2021).
Kaneko, F. et al. Natural killer cell numbers and function in peripheral lymphoid cells in Behcet's disease. Br J Dermatol 113, 313–318 (1985).
Suzuki, Y. et al. Increased peripheral blood gamma delta + T cells and natural killer cells in Behçet's disease. J Rheumatol 19, 588–592 (1992).
Sakly, K. et al. Phenotypic abnormalities of peripheral blood mononuclear cells in patients with Behçet's disease and association with HLA-B51 expression. Immunol Invest 43, 463–478 (2014).
Hasan, M.S. et al. Circulating NK cells and their subsets in Behçet's disease. Clin Exp Immunol 188, 311–322 (2017).
Szántó, S. et al. Intracytoplasmic cytokine expression and T cell subset distribution in the peripheral blood of patients with ankylosing spondylitis. J Rheumatol 35, 2372–2375 (2008).
Azuz-Lieberman, N. et al. The involvement of NK cells in ankylosing spondylitis. Int Immunol 17, 837–845 (2005).
Mousavi, T. et al. Phenotypic study of natural killer cell subsets in ankylosing spondylitis patients. Iran J Allergy Asthma Immunol 8, 193–198 (2009).
Perussia, B. et al. Peripheral NK cell phenotypes: multiple changing of faces of an adapting, developing cell. Mol Immunol 42, 385–395 (2005).
Gibbings, D. & Befus, A.D. CD4 and CD8: an inside-out coreceptor model for innate immune cells. J Leukoc Biol 86, 251–259 (2009).
Vokurková, D. et al. CD8 + natural killer cells have a potential of a sensitive and reliable biodosimetric marker in vitro. Physiol Res 55, 689–698 (2006).
Ahmad, F. et al. High frequencies of polyfunctional CD8 + NK cells in chronic HIV-1 infection are associated with slower disease progression. J Virol 88, 12397–12408 (2014).
Addison, E.G. et al. Ligation of CD8alpha on human natural killer cells prevents activation-induced apoptosis and enhances cytolytic activity. Immunology 116, 354–361 (2005).
Barcenilla, H. et al. Mass Cytometry Identifies Distinct Subsets of Regulatory T Cells and Natural Killer Cells Associated With High Risk for Type 1 Diabetes. Front Immunol 10, 982 (2019).
Lang, P.A. et al. NK Cells Regulate CD8(+) T Cell Mediated Autoimmunity. Front Cell Infect Microbiol 10, 36 (2020).
Bielekova, B. et al. Regulatory CD56(bright) natural killer cells mediate immunomodulatory effects of IL-2Ralpha-targeted therapy (daclizumab) in multiple sclerosis. Proc Natl Acad Sci U S A 103, 5941–5946 (2006).
Garni-Wagner et al. A novel function-associated molecule related to non-MHC-restricted cytotoxicity mediated by activated natural killer cells and T cells. J Immunol 151, 60–70 (1993).
Valiante, N.M. & Trinchieri, G. Identification of a novel signal transduction surface molecule on human cytotoxic lymphocytes. J Exp Med 178, 1397–1406 (1993).
Memmer, S. et al. The Stalk Domain of NKp30 Contributes to Ligand Binding and Signaling of a Preassembled NKp30-CD3ζ Complex. J Biol Chem 291, 25427–25438 (2016).
Zheng, G.X. et al. Massively parallel digital transcriptional profiling of single cells. Nat Commun 8, 14049 (2017).
Hao, Y. et al. Integrated analysis of multimodal single-cell data. Cell 184, 3573–3587.e3529 (2021).
R 2022. R v4.1.0, https://www.r-project.org/.
McGinnis, C.S. et al. DoubletFinder: Doublet Detection in Single-Cell RNA Sequencing Data Using Artificial Nearest Neighbors. Cell Syst 8, 329–337.e324 (2019).
Ianevski, A. et al. Fully-automated and ultra-fast cell-type identification using specific marker combinations from single-cell transcriptomic data. Nat Commun 13, 1246 (2022).
Cao, Y. et al. SCSA: A Cell Type Annotation Tool for Single-Cell RNA-seq Data. Front Genet 11, 490 (2020).
Policastro, M. https://github.com/rpolicastro/scProportionTest (2020).

There is NO Competing Interest.

SupplementTable1DonorDemograpy.xlsx
Supplementary Dataset 1
SupplementTable2Flowantibodies.xlsx
Supplementary Dataset 2
SupplementTable310xGenomics.xlsx
Supplementary Dataset 3
SupplementTable4scRNAseqSampleDemograpy.xlsx
Supplementary Dataset 4
SupplementTable5scRNAseqLibraryinfo.xlsx
Supplementary Dataset 5
Supplementarydata.docx

Download PDF

Journal Publication

published 31 Jul, 2024

Read the published version in Nature Communications →

Version 1

posted

You are reading this latest preprint version

Single-cell profiling identifies a CD8^bright CD244^bright Natural Killer cell subset that reflects disease activity in HLA-A29-positive birdshot chorioretinopathy.

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Increased frequency of CD16 + NK cells in Birdshot uveitis patients

PBMC scRNA-seq identifies altered NK repertoire in Birdshot Chorioretinopathy

High-dimensional cytometry reveals accumulation of a CD8^bright CD244^bright subset of CD16+ NK cells in the circulation of Birdshot Chorioretinopathy patients

CD8a^bright CD244^bright cytotoxic NK cell frequency correlates with disease activity

Discussion

Methods

Patients

Blood sample processing

Flow cytometry

Intracellular cytokine staining

FlowSOM analysis

PBMC single cell RNA sequencing

Preprocessing of scRNAseq data

Dimension reduction, clustering and visualization

Data availability

Abbreviations

Declarations

FUNDING

ACKNOWLEDGEMENT

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1

Single-cell profiling identifies a CD8bright CD244bright Natural Killer cell subset that reflects disease activity in HLA-A29-positive birdshot chorioretinopathy.

Status:

Journal Publication

Version 1

Abstract

Figures

Introduction

Results

Increased frequency of CD16 + NK cells in Birdshot uveitis patients

PBMC scRNA-seq identifies altered NK repertoire in Birdshot Chorioretinopathy

High-dimensional cytometry reveals accumulation of a CD8bright CD244bright subset of CD16+ NK cells in the circulation of Birdshot Chorioretinopathy patients

CD8abright CD244bright cytotoxic NK cell frequency correlates with disease activity

Discussion

Methods

Patients

Blood sample processing

Flow cytometry

Intracellular cytokine staining

FlowSOM analysis

PBMC single cell RNA sequencing

Preprocessing of scRNAseq data

Dimension reduction, clustering and visualization

Data availability

Abbreviations

Declarations

FUNDING

ACKNOWLEDGEMENT

References

Additional Declarations

Supplementary Files

Status:

Journal Publication

Version 1

Single-cell profiling identifies a CD8^bright CD244^bright Natural Killer cell subset that reflects disease activity in HLA-A29-positive birdshot chorioretinopathy.

High-dimensional cytometry reveals accumulation of a CD8^bright CD244^bright subset of CD16+ NK cells in the circulation of Birdshot Chorioretinopathy patients

CD8a^bright CD244^bright cytotoxic NK cell frequency correlates with disease activity