Peripheral blood mononuclear cells (PBMCs) were isolated by Ficoll (GE17-1440-03, Sigma-Aldrich) gradient centrifugation from buffy coats from healthy donors in compliance with institutional review board protocols (CME2019/042 and CME2016/629). CD4⁺ T cells were isolated using RosetteSep™ Human CD4⁺ T Cell Enrichment Cocktail (15062, Stemcell Technologies) according to manufacturer’s protocol. CD25⁺ T cells were isolated from PBMCs or CD4⁺ T cells using the Human CD25 MicroBeads II kit (130-097-044, Miltenyi Biotecs) according to manufacturer’s protocol and subsequently Tregs were sorted from these cells as propidium iodide (PI)^-CD4⁺CD25⁺CD127^- on a FACS Aria II (BD Biosciences).

After isolation, Tregs were cultured for six days in 24-well plates at 250.000 cells/well in 1 mL X-vivo (BE02-060F, Lonza) + 5% heat-inactivated fetal bovine serum (FBS) (S1400, Biowest) with 10 µg/mL plate-bound anti-CD3 (555329, BD Biosciences), 1 µg/mL soluble anti-CD28 (555725, BD Biosciences) and 300 U/mL IL-2 (11147528001, Sigma-Aldrich) or 1500 IU/mL Proleukin^® (Novartis). For short term expansion experiments, Tregs were cultured for 24 hours in above-mentioned conditions and underwent subsequent nucleofection without prior re-plating to 6-well plates.

Methylation at the Treg-Specific Demethylated Region (TSDR) was studied in 7 day-in vitro expanded CD4⁺CD25^-CD127⁺ T conventional cells (Tconv) and CD4⁺CD25⁺CD127^- Tregs from matched donors. Genomic DNA was extracted from frozen samples using QIAamp DNA blood mini kit (51104, Qiagen) according to manufacturer’s protocol. Methylation analysis was performed by EpigenDx (Hopkinton, USA) by pyrosequencing of bisulfite-converted DNA. Nine representative CpG residues in the TSDR were analyzed using ADS783-FS2 assay for human FOXP3.

24 hours prior to nucleofection, cells were cultured in 6-well plates at a density of 250.000 cells/mL in 2 mL X-vivo + 5% FBS and 100 U/mL IL-2 (11147528001, Sigma-Aldrich) or 500 IU/mL Proleukin^® (Novartis). For transfection, cells were collected, centrifuged at 90g for 10 minutes at room temperature, and 1 million Tregs were resuspended in 20 µl P3 Primary Cell 4D-Nucleofector X Kit S (V4XP-3032, Lonza). In PCR tubes, 20 pmol Cas9 nuclease (9212-0.25MG, Aldevron) was mixed with 100 pmol sgRNA (Synthego, Table S1 ) and incubated at 37°C for a minimum of 10 minutes before adding to the cells. For multiplexing, RNP complexes for each sgRNA were generated separately and equal amounts of each sgRNA was added. The cell/RNP mixture was transferred to Nucleofection cuvette strips (4D-Nucleofector X Kit S, Lonza) and cells were electroporated using the 4D-Nucleofector Core Unit (AAF-1002B, Lonza) and X Unit (AAF-1002X, Lonza) with program EO115. After transfection, 80 µl medium at room temperature (X-vivo + 5% FBS + 100 U/mL IL-2 or 500 IU/mL Proleukin^®) was added to the wells of the cuvette strip. Cells were collected and plated in 1 mL pre-warmed medium in 24-well plates and incubated at 37°C until read-out. For re-stimulation, cells were activated 2 hours to 4 days after nucleofection with anti-CD3 plate bound mAb (1 – 10 µg/mL) and 1 µg/mL soluble anti-CD28 (555725, BD Biosciences) in the presence of IL-2.

gRNAs targeting B2M, CD4 and IL2RA were described before ( Table S1 ). gRNAs targeting IL6RA were designed using Nucleotide (NCBI) and CRISPOR (http://crispor.tefor.net/) and tested for their in vitro targeting efficiency.

Ten gRNAs targeting IL6RA were designed and tested in HEK293T cells for their in vitro effectiveness of creating indels as described before (35). Briefly, HEK293T cells were transfected using jetOptimus buffer (Polyplus, #117-07) with 300 ng Cas9 plasmid (pU6-(BbsI)_Cbh-Cas9-T2A-mCherry; Addgene plasmid #64324) and 150 ng OOF plasmid (pBS SK mCherryROSAegfp; Addgene plasmid #54322) according to manufacturer’s protocol and incubated at 37°C for 48 hours before flow cytometry read-out. gRNAs were considered working when at least 33% of the transfected cells were GFP⁺.

Cells were stained with LIVE/DEAD^® Fixable Red Dead Cell Stain Kit (L34972, Thermo Fisher), LIVE/DEAD^® Fixable Near-IR Dead Cell Stain Kit (L34976, Thermo Fisher) or Propidium Iodide Staining Solution (PI, 556463, BD Biosciences) according to the manufacturer’s instructions for assessment of viability. Cell surface staining was performed in MACS buffer [PBS (17-516F, Lonza) + 0.5% BSA (A2153-100G, Sigma-Aldrich) + 2 mM EDTA (15575-038, Invitrogen)] by incubating fluorochrome-conjugated antibodies for 20 minutes at 4°C. Afterwards, cells were fixed and permeabilized using eBioscience™ FOXP3/Transcription Factor Staining Buffer set (00-5523-00, Invitrogen) according to manufacturer’s protocol. Intracellular staining was performed in Perm buffer by incubating antibodies for 30 minutes at 4°C. For intracellular cytokine staining, cells were stimulated with 50ng/ml phorbol12-myristate13-acetate (PMA) and 250ng/ml Ionomycin (Sigma) in the presence of GolgiPlug (BD) for 5 hours. Flow cytometry analyses were performed using LSRFortessa X-20 (BD Biosciences) and FlowJo™ (BD Biosciences). Antibodies used were CD4 – APC-Cy7 (557871, BD Biosciences), CD8 – APC (17-0088-73, eBioscience), CD25 – PE-Cy7 (557741, BD Biosciences), CD127 – PerCP-Cy5.5 (351322, Biolegend), B2M – FITC (316304, Biolegend), CD126 – PE (352804, Biolegend), FOXP3 – PE (320108, Biolegend), FOXP3 – AF700 (56-4776-41, eBioscience), Helios Alexa Fluor 488 (563950, BD Biosciences), TIGIT Brilliant Violet 605 (372712, Biolegend), CD39 FITC (328206, Biolegend), CTLA4 PE (555853, BD Biosciences), IL-2 APC (17-7029-82, eBioscience), IFNγ FITC (11-7319-82, eBioscience), IL-10 PE (559330, BD Biosciences), IL-17A PerCP-Cy5.5 (45-7179-42, eBioscience).

Treg ability to suppress T cell proliferation was assessed as previously described (36, 37). In brief, allogeneic PBMCs (100.000 cells/well) were stained with CellTrace™ CFSE Cell Proliferation Kit (C34554, Thermo Fisher) at 1 µM and cultured with Tregs (in different Treg : PBMC ratios: 1:2, 1:4 and 1:8) in 96-well U-bottom plates in X-vivo + 5% FBS. Cells were stimulated using Treg Suppression Inspector beads (130-092-909, Miltenyi Biotec) on a 1:1 bead to cell ratio and cultured for 4 days before FACS analysis. After incubation, cells were stained for flow cytometry analysis as mentioned above and acquired on a BD LSRFortessa. Briefly, cells were stained with LIVE/DEAD^® Fixable Red Dead Cell Stain Kit to exclude dead cells. Cells were then extracellularly stained for CD4 and CD8 expression, fixed and permeabilized using eBioscience™ FOXP3/Transcription Factor Staining Buffer set and lastly stained intracellularly for FOXP3 expression. Treg suppression capacity was assessed based on PBMC proliferation extracted from positive CellTrace™ CFSE staining on both CD4 and CD8 positive populations, allowing a clear exclusion of Tregs from PBMC cells.

Tregs were isolated from buffy coats as mentioned before and incubated in X-vivo + 5% FBS the presence of 1 µg/mL of plate-bound anti-CD3 (555329, BD Biosciences), 1 µg/mL of soluble anti-CD28 (555725, BD Biosciences) and 25 U/mL IL-2 (11147528001, Sigma-Aldrich) in 96-well U-bottom plates at 5x10⁴ cells per well for 24 hours. Where indicated, Tregs were further incubated in the presence of 25 ng/mL IL-6 (206-IL-010, R&D Systems).

50.000 cells/well were plated in V-bottom 96-well plates and incubated for 2 hours at 37°C. After incubation, 100 U/ml IL-2 (11147528001, Sigma-Aldrich) or 50 ng/ml IL-6 (206-IL-010, R&D Systems) was added to each well and cells were incubated at 37°C for 15 minutes. Subsequently, cells were fixed using BD Cytofix™ Fixation Buffer (554655, BD Biosciences) according to manufacturer’s protocol (10 minutes at 37°C). Next, cells were washed in PBS + 0.5% BSA and permeabilized using BD Perm III buffer (558050, BD Biosciences) according to manufacturer’s instructions (30 minutes on ice). Then cells were washed twice in PBS + 0.5% BSA and stained in PBS + 0.5% BSA with pSTAT5 – Pacific Blue (560311, BD Biosciences) or pSTAT3 – FITC (651019, Biolegend) and analyzed using LSRFortessa X-20 (BD Biosciences) and FlowJo™ (BD Biosciences).

Genomic DNA was extracted from Tregs with QIAamp DNA blood mini kit (51104, Qiagen) according to manufacturer’s protocol. Specific gene fragments were amplified using HotStarTaq Master Mix Kit (203443, Qiagen) according to manufacturer’s protocol and gene specific primers listed in Table S2 . In a thermal cycler, the following PCR program was used: 15 minutes at 95°C, 40 cycles of 30 seconds at 95°C, 30 seconds at 65°C and 1 minute at 72°C, followed by 10 minutes at 72°C. PCR products were loaded on a 1% agarose gel and extracted using Nucleospin PCR and Gel Clean-up kit (740609.250, Macherey-Nagel). Sanger sequencing was performed at LGC Genomics GmbH (Berlin, Germany) and the data were analyzed using the ICE analysis tool (Synthego).

Statistical analyses were performed using GraphPad Prism 8.2 software (GraphPad Software). Error bars represent mean ± SD. Results were compared using two-tailed unpaired and paired t tests and one-way ANOVA if the data were normally distributed. Wilcoxon and Kruskal-Wallis tests were used as non-parametric tests. Normality was assessed using Shapiro-Wilk tests. For all experiments, significance was defined as *p ≤ 0.05, **p ≤ 0.01, ***p ≤ 0.001 and ****p ≤ 0. 0001.

In order to generate sufficient cell numbers for functional experiments with highly pure Tregs, we optimized an in vitro expansion protocol for human Tregs, maintaining high viability and functionality. Tregs were isolated from peripheral blood mononuclear cells (PBMC) and stimulated for six days with plate-bound anti-CD3 and soluble anti-CD28 mAbs in the presence of IL-2. Since it has been observed that high cell densities negatively affect transfection efficiency (38), cells were rested for 24 hours prior to transfection in 6-well plates in the presence of IL-2 but without further TCR stimulation. A schematic representation of this protocol is shown in Figure 1A . Tregs were isolated from PBMCs using CD25 microbeads and subsequently FACS-sorted as CD4⁺CD25⁺CD127^- cells ( Figure 1B and Figure S1 ). Expression of Treg-associated markers FOXP3, Helios, TIGIT, CD39 and CTLA4, assessed after sorting, confirms the high purity of isolated Tregs ( Figure 1B and Figure S2 ). Tregs were then expanded in vitro for 6 days and were rested for one more day before transfection ( Figure 1C ). Importantly, expanded Tregs maintained high cell viability and expression of Treg-associated markers after expansion ( Figures 1D, E and Figure S2 ). Moreover, in vitro expanded Tregs preserved their suppressive capacity ( Figure 1F ) and a demethylated TSDR profile ( Figures 1G, H ). Overall, these data indicate that our protocol efficiently induces proliferation of highly pure ex vivo isolated Tregs with sufficient proliferation rates in a short timeframe and without loss of functionality, FOXP3 expression or Treg stability. Importantly, the protocol did not lead to expansion or outgrowth of contaminating effector cells as demonstrated by the phenotype and low TSDR methylation in the expanded Treg product.

Next, expanded Tregs were used for RNP-based genome editing. To assess feasibility and efficacy of the KO procedure in Tregs, we chose to KO beta-2-microglobulin (B2M) and CD4 genes. B2M is a component of the major histocompatibility (MHC) class I molecules and is present on all nucleated cells except for red blood cells (39). CD4 is a membrane-bound glycoprotein that is expressed on helper T cells (40). B2M-KO reached efficiencies up to over 90% depending on the donor and on average about 60% at protein level at the time points analyzed ( Figures 2A, B ). This is well in line with the representative 67% efficiency observed by Inference of CRISPR Edits (ICE) analysis ( Figure S3 ). Briefly, ICE uses Sanger sequencing data for analysis of CRISPR KO efficiencies on DNA level (41). Efficacy for CD4-KO reached about 40% for CD4 expression at protein level ( Figures 2A, C ). Treg KO efficiencies were similar for B2M and slightly lower for CD4 compared to genome editing in total CD4⁺ T cells ( Figures 2B, C and Figure S4 ). B2M-KO did not affect viability compared to mock condition ( Figure S5A ). Moreover, FOXP3 expression was stable in both mock and KO cells, indicating that the gene editing protocol does not compromise cell integrity ( Figure S5B ). Furthermore, re-stimulation after nucleofection did not affect viability, FOXP3 expression nor KO efficiency ( Figures S6A–D ). Altogether, these data indicate that B2M and CD4 can be efficiently knocked-out in Tregs without compromising FOXP3 expression or viability, and KO phenotype is conserved upon re-stimulation with diverse TCR stimulating conditions.

Since in vitro cell expansion does not recapitulate the physiology of ex vivo Tregs, certain studies may require gene editing of non-expanded Tregs. Therefore, we investigated whether Tregs could be gene edited under short-term stimulation prior to RNP nucleofection. Using these conditions we achieved B2M-KO efficiencies of, on average, 45% ( Figures 2D, E ), with no effect on viability ( Figure S7A ) or FOXP3 expression ( Figure S7B ) compared to mock controls. Thus, our data demonstrates that human Tregs can be efficiently genome edited, also by using a short-term protocol.

We next tested the applicability and efficiency of our gene editing protocol for multiplexing, knocking-out multiple genes at once. When targeting B2M and IL2RA (encoding for the α-chain of the IL2 receptor; CD25) in parallel, we were able to generate about 20-30% double KO by co-transfection of both sgRNAs simultaneously ( Figures 3A, B ). Multiplexing slightly affected viability ( Figure 3C ) and did not affect FOXP3 expression ( Figure 3D ), assessed four days after nucleofection. KO efficiencies were 49.8% and 25.4% of the single KOs of B2M and IL2RA respectively, when measured at protein level ( Figure S8 ). Overall, these data show that multiplexing in human Tregs can be performed in an efficient way with limited effects on cell viability and cell integrity, offering possibilities to target combinations of genes to study Treg function and potentially improve Treg-based immunotherapy.

Next, we investigated if the suppressive capacity of Tregs could be altered by knocking out IL2RA, as previously demonstrated in murine Tregs (7, 42–45). CD25 forms, together with CD122 and CD132, a fully functional IL-2 receptor that can activate the transcription factor STAT5 (45). Roth et al. described a family with monogenic immune disease caused by a heterozygous mutation in IL2RA leading to ablation of CD25 expression and decreased STAT5 phosphorylation in Tregs. Furthermore, Tregs from those patients showed decreased suppressive capacity (46). IL2RA-KO in Tregs following our protocol, reached efficiencies up to over 80% depending on the donor and on average 60% on protein level ( Figures 4A, B ). Mock, CD25⁺ and IL2RA-KO cells were FACS sorted with high purity four days after nucleofection as living cells and based on CD25 expression ( Figure 4C ). FOXP3 expression did not differ between mock, CD25⁺ and IL2RA-KO ( Figure 4D ). IL2RA-KO Tregs also maintained similar Helios and TIGIT expression as mock Tregs, whereas a downregulation in CTLA4 expression was observed in IL2RA-KO Tregs ( Figure S9A ). These data are consistent with previous findings that reported no changes in Helios expression in patients with an IL2RA null mutation (47), and CTLA4 downregulation in IL2RA-KO murine Tregs (48). Moreover, we could not detect major changes in expression of cytokines such as IL-2, IL-10, IFNγ or IL-17A in IL2RA-KO Tregs ( Figure S9B ). DNA sequencing analysis of sorted KO cells revealed a KO score of 70%, indicating that both homozygous as well as heterozygous mutations cause loss of CD25 protein expression ( Figure S10A ). Importantly, IL2RA-KO Tregs had significantly decreased STAT5 phosphorylation upon IL-2 stimulation, proving functional KO of IL2RA targeted cells ( Figure 4E ). Moreover, IL2RA-KO Tregs had significantly blunted suppressive capacity compared to control Tregs, measured as their ability to inhibit proliferation of CD4⁺ and CD8⁺ T effector cells in suppression assays ( Figure 4F ). These data highlights the importance of CD25 as a functional component in in vitro suppression assays, comparable to published data on murine Tregs (7). Overall, the obtained results demonstrate that our gene editing protocol does not compromise cell integrity and that gene-edited cells can be efficiently used for functional downstream applications.

Several clinical trials indicate that tocilizumab might be an effective therapy to ameliorate disease severity in COVID-19 patients (24–28). Since CD126 is expressed on both CD4⁺CD25^- conventional T cells as well as on Tregs (49), tocilizumab could act on both cell types and it was recently shown that tocilizumab treatment also affects the transcriptional signature of Tregs in COVID-19 patients (50). IL-6 is a well known regulator of the Th17 and Treg balance (21, 51) and murine CD126⁺ Tregs have defective suppressive function whereas CD126^- Tregs exerted superior stability in an inflammatory context in vivo (20). However, the effects of IL-6 on human CD4⁺ T cells are less established. Ferreira et al. described a subset of human Tregs that are highly suppressive in vitro and can be characterized by high IL-6 receptor expression that potentially could relate to Treg instability in the presence of IL-6-associated inflammation in vivo (52). Further, it has been demonstrated that supplementation of IL-6 to in vitro suppression assays impairs the suppressive function of human Tregs (19). However, these observations might be the result of effects of IL-6 on responder cells rather than on Tregs. In order to test the direct effects of IL-6 on human Tregs, we pre-incubated Tregs for 24 hours with IL-6 before setting up an in vitro suppression assay and observed decreased suppressive ability compared to controls ( Figure 5A ). To further investigate the function of CD126, we knocked out IL6RA in Tregs using our protocol. We reached KO efficiencies over 70% depending on the donor and on average of about 55% on protein level, as measured by FACS for CD126 expression ( Figures 5B, C ). For downstream analysis, IL6RA-KO cells were purified by FACS-sorting ( Figure 5D ). ICE analysis of sorted KO cells confirmed a high KO score of 90%, indicating that most of the KO sorted cells are homozygous for a mutation ( Figure S10B ). FOXP3 expression was not affected by IL6RA-KO ( Figure 5E ) and IL6RA-KO Tregs also maintained similar Helios and TIGIT expression ( Figure S11A ), in line with previous data showing similar Helios expression between TIGIT⁺IL-6R^high and TIGIT⁺IL-6R^low human memory Tregs (52) or between gp130^high or gp130^low human memory Tregs (gp130 being part or the IL-6R complex) (19). Moreover, IL6RA-KO Tregs did not show major differences in IL-2, IL-10, IFNγ or IL-17A expression compared to mock Tregs at different time points after RNP nucleofection ( Figures S11B, C ). IL-6 induced STAT3 signaling could activate a Th17-like phenotype in Tregs while destabilizing their function (15, 17, 21). In order to test functional consequences of IL6RA-KO we thus examined STAT3 phosphorylation in targeted Tregs. Importantly, upon IL-6 stimulation, IL6RA-KO cells had significantly lower phosphorylation of STAT3 compared to mock-transfected Tregs ( Figure 5F ). Overall, these data further demonstrate that human Treg function could be directly affected by IL-6 and that targeting of CD126 may prevent IL-6-mediated instability.