Non-coding autoimmune risk variant defines role for ICOS in T peripheral helper cell development

Human samples

This research complies with all relevant ethical regulations, as approved by the Institutional Review Board at Mass General Brigham. Written informed consent was obtained from all participants, except as noted below. Blood samples from healthy subjects were recruited from genotyped volunteer donors within the Mass General Brigham Biobank through the Recruitment Core of the Joint Biology Consortium (JBC, www.jbcwebportal.org). 24 subjects (8 A/A, 8 A/C, and 8 C/C genotypes at rs117701653) were enrolled for low-input RNA sequencing. 46 subjects (17 A/A, 18 A/C, 11 C/C), including 7 individuals (2 A/A, 3 A/C, 2 C/C) included also among the first cohort of 24 subjects, were recruited for CD4⁺ memory T cell immunophenotyping and differentiation assays.

Patients with RA fulfilled the ACR/EULAR 2010 Rheumatoid Arthritis classification criteria⁵⁴. Synovial fluid samples were obtained as excess material from patients undergoing clinically indicated diagnostic or therapeutic arthrocentesis as directed by the treating rheumatologist. Blood samples from healthy controls were obtained from blood bank leukoreduction collars from anonymous platelet donors, and as de-identified discard samples, were obtained without written consent and were not employed for genomic studies. Synovial fluid samples from RA patients and blood samples from healthy controls were used to estimate the correlation between ICOS expression and Tph proportion in memory CD4⁺ T cells.

FREP and mass spectrometry

FREP followed the previously described protocol²³. The bait DNA fragment (rs117701653-C/5Biosg), competitor DNA fragment (rs117701653-C), and irrelevant DNA sequence were used; sequences are listed in Supplementary Table 3. Mass spectrometry was performed using a Thermo Scientific Q Exactive HF Orbitrap LC-MS/Ms system.

Electrophoretic mobility shift assay (EMSA)

EMSA was performed using the LightShift Chemiluminescent EMSA kit (Thermo Scientific, 20148) according to manufacturer’s instructions. A probe made of the 31-bp sequences centered on SNP rs117701653 was made by annealing two biotinylated oligonucleotides. Nuclear proteins were extracted from Jurkat T cells using NE-PER Nuclear and Cytoplasmic Extraction Reagents (Thermo Scientific, 78835) per manufacturer’s instructions. For gel supershift, the indicated antibody was added before or after an additional 30 min incubation with DNA probe and nuclear protein extract (Supplementary Fig. 1).

Chromatin Immunoprecipitation (ChIP) qPCR

ChIP-IT PBMC kit (Catalog no. 53042, Active Motif) was used according to the manufacturer’s instructions. In brief, human CD4 + T cells or CRISPR-Cas9-edited Jurkat cell clones were cross-linked for 15 min using 1% formaldehyde. Cross-linked cells were lysed, and chromatin was sheared by using a QSONICA Q125 (42% amplitude, pulse 30 seconds on/off for 5 minutes of “on” time per rounds, total four rounds of sonication). 20 μg of chromatin was incubated with either 10 μg of anti-SMCHD1 (Abcam, ab179456) or 10 μg of rabbit IgG (NOVUS, NB810-56910) to control for non-specific binding. Quantitative PCR was used to measure relative fold enrichment of SMCHD1 binding at SNP rs117701653 or HS17 promoter as a positive control⁵⁵. ChIP-qPCR primers are listed in Supplementary Table 3.

HDR-editing and SMCHD1 deletion by CRISPR-Cas9

CRISPR-mediated homology-directed repair (HDR) was applied for the generation of HDR-edited Jurkat cell lines using sgRNA targeting three bases upstream from rs117701653 and asymmetrical single-stranded DNA donors²⁹. 2 × 10⁵ Jurkat cells were nucleofected with 20 picomole of sgRNA-Cas9 complex and 100 picomole of DNA donor template using program CL-120 of Amaxa^TM 4D-Nucleofector and SE cell line kit S (Lonza, V4XC-1032). The edited single-cell clones were sorted into 96-well plate by BD Aria II sorter and expanded for two months. Using a GeneArt^TM Genomic Cleavage Detection Kit (ThermoFisher Scientific, A24372) following manufacturer’s instructions, modified DNA was isolated from the survived clones (904/1,440) and used in a PCR reaction. PCR products were analyzed by Sanger sequencing to identify A/A wild-type clones (n = 38), A/C edited clones (n = 3), C/C edited clones (n = 3).

For generation of SMCHD1 deleted A/A and C/C Jurkat clones, 2 × 10⁵ cells were nucleofected with 20 picomole of sgRNA-Cas9 complex that targets exon 8 of the SMCHD1 gene using program CL-120 of Amaxa^TM 4D-Nucleofector and SE cell line kit S. Afterward, the cells were cultured for 4-5 days and analyzed for SMCHD1 deletion by western blot. For generation of SMCHD1 deleted A/A and C/C differentiated Tph cells, memory CD4 + T cells from 6 A/A and 6 C/C healthy donors were preactivated with anti-CD3/CD28 beads for 2 days. Thereafter, 1 × 10⁶ cells were nucleofected with either non-targeting or SMCHD1 targeting sgRNA using a 40 picomole of sgRNA-Cas9 complex. The nucleofection was performed using program EO-115 of Amaxa^TM 4D-Nucleofector and P3 primary cell kit S. The nucleofected cells were then differentiated with anti-CD3/CD28 beads, TGF-β and anti-ICOS stimulation for 3 days. Deleted SMCHD1 was confirmed by western blotting. IL-21 and ICOS expression in SMCHD1-deleted Tph cells were compared to control cells using flow cytometry. The sequences of sgRNAs and DNA donor template are listed in Supplementary Table 3.

Protein and RNA quantifications in Jurkat clone cells

Protein levels in resting and stimulated Jurkat clones were measured by western blotting. A/A and C/C clones were stimulated with either plate-bound anti-CD3 (Biolegend, 317302)/CD28 (Biolegend, 302943) or anti-CD3/ICOS (Invitrogen, 16-9948-82) antibodies over different time periods. Whole-cell lysate were obtained using Cell Lysis Buffer (Cellsignal, 9803). Protein extracts were subjected to western blotting using anti-SMCHD1 (ABCAM, ab179456), anti-\(\alpha\)-Tubulin (Cellsignal, 2144 S), anti-ICOS (ABCAM, ab175401), anti-RAPH1 (Cellsignal, 91138 T), anti-CD28 (Cellsignal, 38774 S), anti-AKT (Cellsignal, 9272 S), Ser473 Phospho-AKT (Cellsignal, 4060 S), JNK (Cellsignal, 9252 S), Thr183/Tyr185 Phospho-JNK (Cellsignal, 4668 S) antibodies. Details of the antibodies are available in Supplementary Table 4.

For measurement of RNA level, total RNA was isolated from resting Jurkat clones using the RNeasy Micro kit (Qiagen, 74004) and used for cDNA library construction using oligo-dT primer and reverse transcriptase (Agilent, 600559). Quantitative PCR was performed with SYBR green fluorescent dye (Agilent, 600882) using the real-time qPCR detection system (ThermoFisher Scientific, QuantStudio3). We used specific primers from PrimerBank (https://pga.mgh.harvard.edu/primerbank/index.html) for human ACTB, SMCHD1, ICOS, CD28, and RAPH1. Relative mRNA levels were estimated using the comparative Ct method, \(\Delta \Delta\)Ct method⁵⁶. All qPCR primers are listed in Supplementary Table 3.

PBMC isolation and immunophenotyping

Whole blood was collected from 46 healthy human subjects based on genotype at rs117701653. Peripheral blood mononuclear cells (PBMCs) were isolated by a Ficoll gradient (GE Healthcare, 17-1440-02) and cryopreserved in 10 % DMSO in fetal bovine serum. Immunophenotyping was carried out by flow cytometry on thawed PBMCs. We characterized major populations of human CD4⁺ T cells as follows: naïve T cells (CD3⁺CD4⁺CD45RA⁺), memory T cells (CD3⁺CD4⁺ CD45RA⁻), and effector memory subsets: Th1 (CD3⁺CD4⁺CD45RA⁻ CC6⁻CXCR3⁺CCR4⁻), Th2 (CD3⁺CD4⁺CD45RA⁻CCR6⁻CXCR3⁻CCR4⁺), Th17 (CD3⁺CD4⁺CD45RA⁻CCR6⁺CXCR3⁻CCR4⁺), memory Treg (CD3⁺CD4⁺CD45RA⁻CD25^highCD127⁻FOXP3⁺), Tfh (CD3⁺CD4⁺CD45RA⁻CXCR5⁺PD-1^high), and Tph (CD3⁺CD4⁺CD45RA⁻CXCR5^–PD-1^high) (Supplementary Fig. 5). Fluorescent-conjugated and isotype control antibodies used in flow cytometry are listed in Supplementary Table 4.

For gating Tph and Tfh cells, we considered cells with a higher signal level than the negative controls as positive for CXCR5 and PD-1 expression, and cells with similar signal to the control as negative. To distinguish cells with high and intermediate PD-1 among the positive cells in resting T cells, we established a threshold of 4,100 on a biexponential scale (Fig. 3C). For the differentiated Tph cells, we used a threshold of 9,700 on a biexponential scale to clearly distinguish cells with PD-1 high expression after 3 days of differentiation (Fig. 5A).

RNA-Sequencing and data processing

CD3⁺CD4⁺ T cells from PBMCs of healthy subjects from the Mass General Brigham Biobank genotyped for rs117701653 (A/A = 8, A/C = 8, C/C = 8) were isolated by negative selection using EasySep human CD4⁺ T cell Isolation kit (STEMCELL, 17952). From the isolated CD3⁺CD4⁺ T cells of each individual, RNA was isolated using RNeasy Micro kit (Qiagen, 74004) and eluted in 14 μl of water. 10 ng samples of RNA were transferred into wells of a 96-well plate, and RNA-seq libraries were prepared at Broad Technology Labs at the Broad Institute of MIT and Harvard (Cambridge, Massachusetts, USA) using the Illumina SmartSeq2 platform. Samples were sequenced on a NextSeq 500 generating a median of 5.6 million 38 bp paired-end reads per sample.

Raw data were processed using release 3.9 of the nextflow nf-core “rnaseq” pipeline^57,58. The pipeline was executed on the BCH HPC Clusters Enkefalos 2, using singularity⁵⁹ containers to ensure optimal reproducibility. Briefly, we performed adapter and quality trimming using Trim Galore (version 0.6.7) and subsequently aligned reads to the GRCh38 reference genome using STAR⁶⁰ (version 2.7.10). We quantified transcript expression with Salmon⁶¹ (version 1.5.2) and aggregated transcript abundances to gene-level measurements with bioconductor-tximeta⁶² (version 1.8.0).

Targeted eQTL analysis

We used QTLtools^63,64,65 (version 1.3.1-12-gba66d62ef4) to perform a targeted cis-eQTL mapping analysis for protein coding genes that were expressed above 0 log₂(tpm+1) in at least 8 samples and have a transcription start site (TSS) within a 1MB window of SNP rs117701653, namely WDR12, NBEAL1, CYP20A1, ABI2, RAPH1, CD28, CTLA4, ICOS, and PARD3B. We corrected expression levels for age and sex and rank-normal transformed residuals via QTLtools’ “–normal” option. We computed nominal P-values for the association between genotypes and expression levels of the selected genes using a linear model implemented in QTLtools.

Memory and naive CD4⁺ T cell isolation and differentiation

Memory and naive CD4⁺ T cells from PBMCs of healthy subjects were isolated by negative selection using the EasySep human memory CD4⁺ T cell Isolation kit (STEMCELL, 19157) and the human naive CD4 + T cell isolation kit II (Miltenyi, 130-094-131). T cells were resuspended in RPMI (supplemented with 10 % fetal bovine serum and 100 units/mL penicillin/streptomycin) at 0.25 \(\times\) 10⁶ cells/mL and stimulated with anti-CD3/CD28 beads (Invitrogen, Dynabeads human T-activator CD3/CD28) at a ratio of 5:1 (cell:bead), to which the indicated combination of 2 ng/mL TGF-β1 recombinant protein (R&D system, 7754BH005/CF) and 2 μg/ml anti-ICOS antibody (Invitrogen, 16-9948-82) were added for varying periods (3, 6, 10, 14, and 18 days). For the prolonged expansion, cells were stimulated again with anti-CD3/CD28 beads together with TGF-β, ICOS stimulation by anti-ICOS, or both at 6, 10, and 14 days of differentiation.

Cells were harvested at indicated time points for intracellular cytokine staining. Cells were re-stimulated with anti-CD3/CD28 beads at a ratio 5:1 (cell:bead) for 24 h and with both phorbol 12-myristate 13-acetate and ionomycin (Both 1:500, Biolegend 423301) for the last 5 hours. Brefeldin (1:1000, BD Bioscience 555029) and monensin (1:1500, BD Bioscience 554724) were added for the last 5 hours. Cells were washed twice in cold PBS, incubated for 30 minutes with Fixable Viability Dye (Invitrogen, 65-0863-14), washed in 1% FBS/PBS, and then incubated in cell surface antibodies with anti-CD3, CD4, PD-1, CXCR5 for 20 minutes. Cells were then washed again in 1% FBS/PBS, and fixed and permeabilized using Transcription Factor Buffer Set (eBioscience, 00-5523-00). Permeabilized cells were incubated in intracellular antibodies with anti-IL-21 and CXCL13 for 1 hour. Flow cytometry analysis was performed on a BD Fortessa analyzer. Antibodies used in flow cytometry are listed in Supplementary Table 4.

Mass cytometry data

AMP mass cytometry data followed the previously described gating method³⁴. Briefly, frequency of Tph cells (CD3⁺CD4⁺CD45RO⁺CXCR5⁻PD-1^high), Tfh cells (CD3⁺CD4⁺CD45RO⁺CXCR5⁺PD-1^high), plasmablasts (CD45⁺CD19 + CD20⁻CD38^highCD27⁺) were quantified by manual gating, with uniform gates applied to all samples. FlowJo 10.8.0 was used for determination of cell population frequencies.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.