Mutant IL-2-expressing mesenchymal stromal cells promote regulatory T cells proliferation and activation in collagen induced arthritis mice

Abstract

Objectives

Research indicates that low doses of interleukin-2 (IL-2) can effectively mitigate Rheumatoid arthritis (RA) symptoms by promoting Treg cells, while high doses may enhance immune responses and exacerbate the disease. Consequently, this study employed mutated IL-2 to minimize its impact on CD8⁺ T and NK cell activation while preserving its influence on Treg cells.

Methods

We used a previously published mutation sites to construct the murine IL-2 mutants by overlap PCR. Then we assessed its impact on the proliferation and functionality of Treg cells by flow cytometry and PCR. The synergistic effects of mutated IL-2 and MSC on collagen-induced arthritis (CIA) in mice were evaluated through the infusion of lentiviral-transduced umbilical cord-derived mesenchymal stromal cell (UC-MSC) for CIA treatment and through pathological section staining to assess inflammatory joint injury, cartilage destruction, and osteoclast infiltration.

Results

Mutant IL-2 demonstrated targeted enhancement of both the proportion and proliferative activity of Treg cells with a diminished capacity to stimulate the proliferation of CD8⁺ T cells and NK cells relative to wild-type IL-2. Moreover, MSC-mutant IL-2 significantly augmented the proportion of Treg cells compared to either MSC or mutant IL-2 in isolation. Treatment with MSC-mutant IL-2 infusion in CIA mice ameliorated arthritis symptoms and reduced inflammatory infiltration and cartilage damage in their joints.

Conclusion

Mutant IL-2 enhances Treg function and proliferation while exerting reduced effects on CD8⁺ T and NK cell activation. MSC expressing mutant IL-2 demonstrates therapeutic benefits in CIA by increasing the proportion of Treg cells and reducing the proportion of CD8⁺ T cells.

Introduction

Rheumatoid arthritis (RA) is a chronic autoimmune disorder characterized by its deleterious effects on bone and cartilage within joints, leading to swelling, pain, and functional impairment[1]. Clinically, treatment involves the use of disease-modifying anti-rheumatic drugs (DMARDs), NSAIDs, glucocorticoids, and biologics; however, their prolonged use may result in a spectrum of adverse effects[2]. Regulatory T cells (Treg), a specialized subset of CD4⁺CD25⁺FOXP3⁺ T cells, play a pivotal role in suppressing autoimmune effector cells, thereby preventing autoimmune diseases and maintaining immune homeostasis[3, 4]. A reduction in the proportion of Treg cells and their dysfunction are critical factors in the development and progression of rheumatoid arthritis[5, 6]. Collagen-induced arthritis (CIA) mice, a widely recognized mouse model for rheumatoid arthritis, served as the experimental subjects in this study[7].

Mesenchymal stromal cells (MSC) possess the ability to suppress memory effector T cells[8], promote Treg cells proliferation and activation[9], secrete anti-inflammatory cytokines, and migrate to sites of inflammation rendering MSC a promising therapeutic candidate for RA[10]. Numerous preclinical and clinical studies have corroborated the efficacy and safety of MSC in treating RA[11]. MSC-derived extracellular vesicles have also been shown promising effects in the treatment of rheumatoid arthritis[12]. Previous studies have confirmed the efficacy and safety of human-derived MSC in the treatment of mouse models of arthritis[13, 14].

The primary function of IL-2 is to promote the proliferation and activation of T cells, thereby exerting an immune response against pathogens and tumor cells. Additionally, IL-2 is crucial for the activation of natural killer cells and B lymphocytes[15]. High dose Interleukin-2 (IL-2) has been clinically demonstrated to be effective in the treatment of cancer (particularly for metastatic renal cell carcinoma and metastatic melanoma) [16, 17]. However, long-term use of high doses of IL-2 can cause vascular leakage syndrome, immune imbalance, and other serious adverse effects[18]. On the other hand, clinical studies have corroborated the efficacy of low-dose IL-2 in treating autoimmune diseases, including systemic lupus erythematosus and rheumatoid arthritis[19, 20]. IL-2, a pleiotropic cytokine produced post-antigen activation, is pivotal in the immune response, promoting the activation and proliferation of various immune cells[21]. The IL-2 receptors vary among immune cells, with high affinity trimeric receptors composed of IL-2Rα (CD25), IL-2Rβ (CD122), and IL-2Rγ (CD132) predominantly found on Treg cells and activated effector T cells[15, 22, 23], and medium affinity dimeric receptors comprised of IL-2Rβ (CD122) and IL-2Rγ (CD132) are primarily expressed on memory T cells, helper T cells, resting effector T cells, resting memory T cells and NK cells[24–27]. Treg cells, due to their expression of trimeric receptors, are more responsive than other immune cells to low-dose IL-2, thereby enabling low-dose IL-2 to act as an immunosuppressant through the promotion of Treg cells[28]. However, the dose-dependency and short half-life of IL-2 therapy necessitate frequent administration and raise concerns about the adverse effects of long-term use, thereby limiting its clinical application[29].

As the ability to receive IL-2 signals is determined by the affinity for distinct IL-2-receptor complexes on different cell subsets, selective targeting of regulatory T cells or cytotoxic lymphocytes can be achieved through the construction of IL-2 mutants or fusion proteins that reduce affinity for different IL-2 receptor subunits, offering a promising therapeutic strategy for cancer and autoimmune disease treatments[30–34]. Sun et al. showed that mutant IL-2, characterized by a diminished affinity for IL-2Rα, led to decreased Treg cell infiltration, thereby enhancing the efficacy of tumor treatment[35]. Similarly, Khoryati, Liliane, et al. demonstrated that murine IL-2 mutants N103R V106D, possessing a reduced affinity for IL-2β, increased reliance on IL-2Rα. Despite a weaker activation relative to wild-type IL-2, their heightened selectivity for Treg cells contributed to improved treatment outcomes in NOD mice[30].

Our strategy involves expressing murine mutant IL-2, which has a high affinity for Treg cells, through MSC, while concurrently leveraging the inflammatory tropism of MSC to induce a selective Treg cell-promoting effect[36], thereby exerting anti-inflammatory effects in CIA mice. It is further anticipated that this treatment modality will be extended and applied to clinical therapy to effectively relieve rheumatoid arthritis and prolong remission duration, minimizing the need for frequent medication.

Material and methods

Prediction of protein structure and physicochemical properties

The tertiary structure of wild-type murine IL-2 was sourced from the Research Collaboratory for Structural Bioinformatics Protein Data Bank (RCSB PDB, http://www.rcsb.org), and the structure of mutant IL-2 N103R V106D was predicted using the online tool AlphaFold (https://cosmic-cryoem.org/tools/alphafold2) [37], A comparison of these two protein structures was conducted using PyMOL (PyMOL Molecular Graphics System, v2.5.7), and the physicochemical properties of the proteins were predicted using Expasy (https://web.expasy.org/protparam/).

Construction of lentiviral vectors

293 T cells were purchased from Wuhan Pricella Biotech Co., Ltd. Cultured 293 T cells were seeded into 10-cm dishes. Upon reaching 80% confluency, target plasmids (pSLenti-EGFP-Puro- IL-2-WPRE, pSLenti-EGFP-Puro-IL-2 N103R V106D-WPRE, control plasmid pSLenti-EGFP- Puro-WPRE) and packaging plasmids (PMD2G, PsPAX2) were introduced into serum-free DMEM medium, all plasmids were obtained from GenScript Biotech Co., Ltd. EZ-Trans transduction reagent was mixed with diluted plasmid DNA to form EZ-Trans-DNA complexes, which were then added to the cell culture dishes. After 6–8 h, the medium was changed; after 48 h, the supernatant was collected, then centrifuged and filtered using a 0.45-µm micro-filter. PEG-8000 was added and mixed by inverting 3–5 times, with a 30-min interval between each inversion. The mixture was left overnight and then centrifuged at high speed, the supernatant was discarded, and the lentiviral precipitates were resuspended in PBS. The suspension was flash-frozen in liquid nitrogen before storage in an ultra-low-temperature freezer.

Overlap PCR

Primers were designed according to the mutation sites identified by Khoryati et al.[30]. Using the IL-2 plasmid as a template, primers IL-2-F(ACACCTTTAATTGGTCAACACGA) and IL-2-M1-R (AGTATGTCTGATCGTGGT), IL-2-M2-F(ACCACGATCAGACATACT) and IL-2-R (CCTGCTACGTTCTCTACCTCT) were used to amplify mutant fragments M1 and M2 by PCR using Taq DNA polymerase respectively. The IL-2 N103R V106D mutant plasmid was obtained by amplifying with primers IL-2-F and IL-2-R. The obtained mutant plasmid IL-2 N103R V106D was sequenced by GenScript Biotech Co., Ltd to verify the correctness of the mutation.

RNA extraction and real-time quantitative PCR

Tissues or cells were lysed using RNA isolater Total RNA Extraction Reagent (Vazyme), and the resulting supernatant was collected. Protein was denatured using chloroform and then centrifuged at 12,000 g. The supernatant, containing total RNA, was removed, and total RNA was precipitated using isopropanol, washed with ethanol, and dissolved in RNase-free water. The RNA was reverse-transcribed into cDNA (HiScript II 1 st Strand cDNA Synthesis Kit, Vazyme). cDNA was used as a template for PCR amplification (ChamQ Universal SYBR qPCR Master Mix, Vazyme), following standard procedures. Data were recorded from the amplification curve generated by the PCR machine (Takara-Standard Co., Ltd.), using GAPDH as an internal control. Finally, the ratio between the target gene and GAPDH was calculated and subjected to statistical analysis.

Western blot

Protein extraction involved lysing splenocytes using RIPA lysis buffer (5 mM Tris–HCl, pH 7.6 150 mM NaCl, 1% NP-40, 1% sodium deoxycholate, 0.1% SDS and 1% protease inhibitor), with subsequent ice incubation for 30 min and heating at 100 °C for 15 min in a metal bath. This was followed by protein electrophoresis, transfer, and blocking. Subsequently, the PVDF membrane were incubated overnight with β-acting rabbit monoclonal antibody (CST:#8457, 1:5000), pSTAT5 (phospho Y694) rabbit monoclonal antibody (Abcam:ab32364, 1:1000) and STAT5 rabbit monoclonal antibody (Abcam:ab230670, 1:1000), and then for one hour with the goat anti-rabbit IgG Fc (HRP) antibody (Abcam:ab97200, 1:10,000). After the PVDF membrane was washed with TBST (137 mM NaCl, 2.7 mM KCl, 25 mM Tris and 0.05% Tween-20), the Enhanced Chemiluminescence developer reagent (Vazyme) was applied to the membrane, and the membrane was scanned using a western imager (Syngene Co., Ltd.) at the appropriate exposure time to capture the image of protein bands.. Expression levels of pSTAT5 and STAT5 proteins in each group were quantified by gray value analysis using ImageJ software(v1.8.0.) with β-acting as an internal control.

Mice

DBA/1 J and C57BL/6 mice, aged 8–10 weeks, were acquired from Nanjing Junke Biotechnology Co. These mice were housed under specific pathogen-free conditions and subjected to a 12-h light/12-h dark cycle (lights on at 7:00 AM). All mice were placed under deep anaesthesia with an intravenous injection of 40 mg/kg of sodium pentobarbital followed by cervical dislocation, C57BL/6 mice were utilized in both in vivo and ex vivo experiments to evaluate mutant IL-2 targeting for Treg cell promotion. DBA/1 J mice were employed in collagen-induced arthritis models to assess the efficacy of mutant IL-2-expressing MSC in arthritis treatment. All experimental protocols involving these mice were approved by the Ethics Committee of Drum Tower Hospital, Nanjing Medical University.

CIA induction and joint symptom evaluation

On Day 0, 100 µL of complete Freund's adjuvant and bovine type II collagen(2 mg/mL, Chondrex, Inc.) were homogenized on ice at high speed for 1 min per cycle, allowed to cool for 2 min, repeated 10 times, and subsequently injected into the tail-root area of each DBA/1 J mouse, and on Day 21, Incomplete Freund's Adjuvant and bovine collagen were similarly homogenized and injected. Arthritis scoring commenced on Day 22 and was independently conducted by a blinded professional unaware of the subgroups. Arthritis scoring was based on the following criteria: 0—normal; 1—erythema of the foot only, without significant swelling; 2—erythema with mild swelling of the foot and ankle; 3—erythema with significant swelling involving the toes and ankle joints; 4—severe erythema, deformity, rupture, and bleeding. The sum of the four paw scores was recorded as the arthritis score.

Cell culture

Umbilical cord-derived mesenchymal stromal cells (UC-MSC) were purchased from Jiangsu Ruiyuan Biotechnology Co., Ltd., and cultured in DMEM/F12 complete medium (comprising 90% DMEM/F12 basal medium (Bio-Channel Biotechnology Co., Ltd.), 10% fetal bovine serum (FBS), and 1% Penicillin–Streptomycin Solution) in a cell incubator at 37 °C and 5% CO2, and were passaged using trypsin–EDTA digestion upon reaching 80–90% confluency. Subsequent in vivo and in vitro experiments used UC-MSC with a PDL of 9.38.

In vitro co-culture

MSC were treated with mitomycin to inhibit their division at a concentration of 10 μg/mL and incubated at 37 °C for 2 h. A density of 10⁵ cells per well was inoculated into 24-well cell culture plates overnight, and then removed the medium and washed them twice with PBS. Splenocytes were inoculated into 24-well plates at a ratio of 1:5, with 5 × 10⁵ splenocytes per well, and 2 μg/mL anti-CD3 functional monoclonal antibody and 1 μg/mL anti-CD28 functional monoclonal antibody were added to activate splenic immune cells. After a 5-day co-culture period, cells were collected in suspension in culture plates. For co-culture with supernatants, splenocytes were inoculated at 5 × 10⁵ per well in 24-well plates pre-coated with 2 μg/mL anti-CD3 functional monoclonal antibody and 1 μg/mL anti-CD28 functional monoclonal antibody, and supernatants of MSC or 293 T were added according to the relevant experimental requirements after 24 h of activation.

Lentivirus and MSC injections

The mice were exposed to an infrared lamp for 5–10 min to fully dilate their tail veins, and subsequently secured in a mouse restrainer. Pre-prepared lentivirus or lentivirally transduced MSC was then slowly injected into the lateral tail vein using an insulin syringe, and any bleeding was halted with an alcohol-soaked cotton ball. Each mouse received a lentivirus dose of 10^8 UI or 5 × 10^5 MSC. MSC infusion was scheduled for arthritic mice on days 28, 35, 42, and 49 post-collagen induction.

Lentiviral transduction of MSC

Lentivirus (MOI = 80) was added to well-growing MSC with a cell confluency of 60–70%, and the medium was replaced 12 h post-transduction, followed by a further 48-h incubation, after which GFP fluorescence was observed. Subsequently, lentivirus-transduced cells were screened using 5 μg/ml puromycin, and the medium was refreshed 24 h post-screening..

Histopathology

Following euthanasia, the right hind limb joints of the mice were excised and immersed in 4% paraformaldehyde (PFA), decalcified in 10% EDTA for 30 days, and subsequently embedded in paraffin. The joints were then assessed for inflammatory infiltration, cartilage destruction, and osteoclastogenesis using hematoxylin and eosin (H&E), Safranin O/Fast Green staining, and TRAP staining (Wuhan Servicebio Technology Co., Ltd.). Histological scoring of H&E-stained joint sections followed established scoring systems for synovial inflammation and cartilage destruction.[38].

Flow cytometry

Single-cell suspensions were prepared from cultured cells, mouse spleens, and lymph nodes. Dead cells were excluded using eBioscience Fixable Viability Dye eFluor 506. Mouse cell surface antigens were stained with anti-mouse CD4 (FITC, RM4-5, Biolgend), CD25 (APC, PC61, Biolgend), CD8 (PE, S18018E, Biolgend), and NK1.1 (Alexa Fluor 700, S17016D, Biolgend) antibodies in Flow Cytometry Staining Buffer (Thermo Fisher) at 4 °C for 25 min. For nuclear cytokine staining, cells were fixed and permeabilized post-surface staining with eBioscience™ Foxp3/Transcription Factor Staining Buffer Set (Thermo Fisher), followed by staining with anti-mouse Foxp3 (Brilliant Violet 421, MF-14, Biolgend) and Ki-67 (PE/Cyanine7, 11F6, Biolgend) antibodies at 25℃ for 40 min. Phosphorylation of STAT5 was detected by flow cytometry using Phosflow Perm™ Buffer III (BD Biosciences) after permeabilization and fixation, followed by staining with anti-mouse pSTAT5 (Tyr694) (PE, A17016B, Biolgend) antibody at 4℃ for 40 min. Data acquisition was performed using a BD LSR Fortessa (BD Biosciences), and analysis was conducted with FlowJo software (Tree Star, version 11.0).

Enzyme-linked immunosorbent assay

As per the manufacturer's instructions for the IL-2 ELISA kit (BioLegend), MSC supernatant, mouse plasma, medium control and gradient-diluted IL-2 standards were added to pre-coated plates and incubated for 2 h. These plates were then incubated with detection antibody and horseradish peroxidase (HRP) for 1 h. Subsequently, the plates were treated with chromogen solution, incubated for 30 min in the dark, followed by the addition of the stop solution, and finally read by an enzyme-linked immunosorbent assay (ELISA) reader. A standard curve was plotted based on the concentration of the standards, and the IL-2 concentration in the samples was determined from their absorbance.

Statistical analysis

All data are expressed as Mean ± SD, and Student's t-test, One-way ANOVA (with Tukey's range test as the post hoc test) and Two-way ANOVA (with Bonferroni’s correction) were used for comparisons between two and more groups when the data conformed to a normal distribution, while data that did not conform to a normal distribution were analysed using Kruskal–Wallis analyses (with Dunn's post-hoc test). Analysis was conducted using GraphPad Prism software (version 9.0, San Diego, CA). P-values less than 0.05 were considered statistically significant, denoted as follows: *P < 0.05, **P < 0.01, ***P < 0.001, ****P < 0.0001;'ns'indicates not significant.

Results

Construction of mutant IL-2 N103R V106D

To construct the desired mutants, the wild-type murine IL-2 plasmid underwent site-directed mutagenesis through overlap extension PCR (Fig. 1A)[39]. According to the findings of Khoryati et al.[30], the asparagine at position 103 and valine at position 106 of wild-type IL-2 were substituted with arginine and alanine, respectively (Fig. 1B). The corresponding lentiviruses were generated using a triple plasmid system[40]. IL-2 secretion was confirmed in both wild-type and mutant IL-2 N103R V106D lentivirus-transduced 293 T cells through ELISA assays of cell supernatants (Fig. 1C). Predictive analyses of physicochemical properties indicate that mutations at these two sites do not significantly alter the spatial structure, estimated half-life, extinction coefficient, instability index, aliphatic index, or grand average of hydropathicity of IL-2. (Table 1, Fig. 1D).

Fig. 1.

Construction and verification of the IL-2 mutein N103R V106D. A Construction of the IL-2 mutein N103R V106D. B Sequencing results of wild-type IL-2 and IL-2 N103R V106D. C IL-2 concentration in cell supernatants post-lentiviral transduction of 293 T cells (n = 3). D Comparative analysis of the structure of IL-2 N103R V106D (predicted by AlphaFold) and wild-type IL-2. ****, P < 0.0001

Table 1.

Physical and chemical property prediction of IL-2 wild type and IL-2 N103R V106D

	Estimated half-life	Extinction coefficients	Instability index	Aliphatic index	Grand average of hydropathicity
IL-2 Wild Type	4.4 h	8605	70.16	81.81	−0.574
IL-2 N103R V106D	4.4 h	8605	68.06	79.87	−0.633

IL-2 N103R V106D can target and activate mouse Treg cells

We assessed the effect of expressing mutant IL-2 on Treg cells in B6 mice by injecting recombinant lentivirus into the tail vein of the mice, B6 mice transdued with empty lentivirus were used as the control and plasma IL-2 concentrations were measured 5 days later. Transduction of B6 mice with either wild-type or mutant IL-2 resulted in increased serum IL-2 levels in the mice (Fig. 2A), and mutant IL-2 N103R V106D lentiviral transduction in mice significantly increased the levels of splenic Treg cells compared to the control group and the wild type IL-2 treated group, whereas no significant difference was seen in the proportion of Treg cells in wild-type IL-2 compared to the control (Fig. 2B, C).

Fig. 2.

Targeted Activation of Treg Cells by IL-2 N103R V106D. A Plasma IL-2 levels increased in IL-2 lentiviral- transduced C57BL/6 mice (ANOVA with Tukey’s post-hoc. n = 5). B-C Representative flow cytometry plots (B) and statistical analysis graphs (C) showing the proportion of splenic Treg cells in C57BL/6 mice transduced with mutant IL-2 N103R V106D was elevated (ANOVA with Tukey’s post-hoc. n = 5). D-E Representative flow cytometry plots (D) and statistical analysis graphs (E) showing the proportion of Ki-67⁺ Treg was elevated in the spleens of mutant IL-2 N103R V106D-treated C57BL/6 mice compared with controls, whereas the proportions of Ki-67⁺ CD8⁺ T cells and NK cells were lower than those in the wild-type IL-2-treated group (ANOVA with Tukey’s post-hoc. n = 3). F The proportions of Treg and NK cells in the spleens of mutant IL-2 N103R V106D-treated C57BL/6 mice were elevated compared to controls, whereas the proportion of CD8⁺ T cells was lower than that of the wild-type IL-2 group (ANOVA with Tukey’s post-hoc. n = 3). SP, spleen; LN, lymphatic node; PBMC, peripheral blood mononuclear cell; *, P < 0.05, **, P < 0.01, ***, P < 0.001; ns, not significant

To further assess the effect of IL-2 N103R V106D on the proliferation of different immune cells, 200μL cell supernatants with an IL-2 concentration of 15 μg/ml from wild-type and mutant IL-2 N103R V106D lentivirus-transduced 293 T were injected intraperitoneally into B6 mice and the proportion of splenic Treg cells, CD8⁺ T cells, NK cells and the proportion of Ki67⁺ cell were detected after 3 days. Ki-67⁺ Treg cell levels in IL-2 N103R V106D-treated B6 mice were comparable to those in wild-type IL-2-treated mice, however, the proportion of Ki-67⁺ NK cells and CD8⁺ T cells was significantly lower than in the wild-type IL-2 group, a statistically significant difference (Fig. 2D, E). Conversely, the proportion of CD8⁺ T cells in IL-2 N103R V106D-treated B6 mice was significantly lower compared to the wild-type IL-2 group, while Treg cell levels remained similar in both (Fig. 2F). This indicates that IL-2 N103R V106D selectively targets Treg cell proliferation and activation, exerting minimal effects on other immune cells.

IL-2 N103R V106D Targeted activation of IL-2 downstream of the PSTAT5 signaling pathway

The JAK-STAT5 signaling pathway, recognized as a crucial pathway downstream of the IL-2 receptor, effectively responds to IL-2R activation through STAT5 phosphorylation. We assessed the phosphorylation levels of pSTAT5 in various cells using flow cytometry and western blot. 20 min Stimulation of B6 mice splenocytes with wild-type IL-2 significantly increased STAT5 phosphorylation [41], whereas mutant IL-2 N103R V106D stimulation led to reduced STAT5 phosphorylation levels (Fig. 3A,B). This disparity in phosphorylation levels primarily arises from wild-type IL-2 activating a broader spectrum of cells, including Treg and conventional T cells, T cells were stimulated with either wild-type IL-2 or mutant IL-2 N103R V106D. Wild-type IL-2 induced STAT5 phosphorylation in both conventional T cells and regulatory T cells. In contrast, stimulation with the mutant IL-2 N103R V106D resulted in reduced STAT5 phosphorylation in conventional T cells compared to wild-type IL-2 (Fig. 3C). A 3-day co-culture of wild-type IL-2 with splenocytes from B6 mice increased the proportions of Treg, CD8⁺ T cells, and NK cells. In contrast, the mutant IL-2 N103R V106D only increased the proportion of Treg cells (Figs. 3D to F). Wild-type IL-2 significantly promoted the proliferation of Treg, CD8⁺ T cells, and NK cells, whereas the mutant IL-2 N103R V106D enhanced the proliferation of only Treg and NK cells. This effect was blocked by STAT5 inhibitors (Figs. 3G to I). Owing to its decreased affinity for the IL-2 receptor subunit β, IL-2 N103R V106D targets the activation of the IL-2 trimeric receptor on Treg cells and CD56 bright NK cells[42], thereby activating downstream STAT5 phosphorylation and promoting their proliferation.

Fig. 3.

Targeted Activation of STAT5 Phosphorylation in Treg Cells by IL-2 N103R V106D to Promote Proliferation. A-B (A) Wild-type IL-2 substantially increases levels of phosphorylated STAT5, while mutant IL-2 has only a modest effect on pSTAT5 levels (ANOVA with Tukey’s post-hoc. n = 3). B Neither wild-type IL-2 nor mutant IL-2 change total STAT5 levels. C Both wild-type IL-2 and mutant IL-2 N103R V106D stimulation elevated the phosphorylation level of STAT5 in Treg cells, whereas mutant IL-2 N103R V106D did not elevate the phosphorylation level of STAT5 in non-Treg cells. D-F The proportion of splenocyte Treg cells co-cultured with supernatants from both wild-type IL-2 and mutant IL-2N103R V106D-transduced 293 T cells was elevated (D), whereas the proportions of CD8⁺ T (E) cells and NK cells (F) were high in the wild-type IL-2 group only (ANOVA with Tukey’s post-hoc. n = 5). G-I Splenocytes cultured in the supernatant of mutant IL-2 N103R V106D-transduced 293 T cells have a reduced Ki-67 MFI in Treg cells after STAT5 inhibitor treatment (G), while their Ki-67 MFI in CD8⁺ T cells (H) and NK cells (I) is lower than that of the wild-type IL-2 group (ANOVA with Tukey’s post-hoc. n = 5). *, P < 0. 05, **, P < 0. 01; ***, P < 0. 001; ****, P < 0. 0001; ns, not significant

N103D and V106D mutant IL-2-expressing MSC can target and promote Treg cells proliferation and function

We transduced MSC with constructed wild-type IL-2 and mutant IL-2 N103R V106D lentiviruses to obtaina higher concentration of IL-2 secretion. Green fluorescence protein (GFP) validated the transduction efficiency of MSC after 8 h of transduction(Fig. 4A). The baseline concentration of IL-2 secreted by MSC alone is 50 pg/ml. Following lentiviral transduction with both wild-type and mutant IL-2, the IL-2 secretion level in MSC increased by more than 100-fold. (Fig. 4B). Previous research has confirmed that transduced MSC retained their inherent osteogenic, chondrogenic, and lipogenic capabilities[43, 44].

Fig. 4.

Enhancement of Treg Cell Function and Proportion by MSC-IL-2 N103R V106D. A Validation of MSC transduction efficiency by GFP fluorescence. B Concentration of IL-2 in supernatants of lentiviral-transduced mesenchymal stromal cells is elevated. C-D Representative flow cytometry plots (C) and statistical analysis graphs (D) showing the proportion of Treg cells in splenocytes co-cultured with mutant IL-2N103R V106D-transduced MSC was higher compared to those co-cultured with wild-type IL-2-transduced MSC and conventional MSC(ANOVA with Tukey’s post-hoc. n = 5). E Treg proportions were elevated in splenocytes co-cultured with high concentrations of mutant IL-2 N103R V106D-transduced MSC supernatants, whereas wild-type IL-2 increased the proportion of CD8⁺ T cells (ANOVA with Tukey’ s post-hoc. n = 5). F-G The proportion (F) or proliferative capacity (G) of CD8⁺ T cells and NK cells in splenocytes co-cultured with MSC-IL-2 N103R V106D supernatant was lower than that of MSC-IL-2. whereas the proportion of Treg cells was elevated compared to MSC, MSC-IL-2 and 293 T-IL-2 N103R V106D supernatants (ANOVA with Tukey’s post-hoc. n = 5). H-I Compared to MSC-IL-2 WT, MSC-IL-2 N103R V106D co-cultured splenocytes had elevated levels of expression of Treg-related genes (H) and reduced levels of expression of effector T-cell-related genes (I) (ANOVA with Tukey’s post-hoc. n = 5). *, P < 0. 05, **, P < 0. 01; ***, P < 0. 001; ****, P < 0. 0001; ns, not significant

Upon 3 days co-culturing with B6 mice splenocytes, the MSC transduced with mutant IL-2 N103R V106D lentivirus significantly increased Treg cell content (Fig. 4C, D). Splenocytes were incubated with supernatants from MSC transduced with wild-type and mutant IL-2 at IL-2 concentrations of 1.5 μg/ml or 15 μg/ml for 3 days. At both high and low concentrations, MSC-IL-2 N103R V106D supernatant elevated Treg cell content when co-cultured with splenocytes, in contrast to MSC-IL-2 wild-type, which increased the proportion of CD8⁺ T cells (Fig. 4E). Compared to MSC-IL-2, the MSC-IL-2 N103R V106D supernatant had an attenuated ability to increase the proportion or proliferative capacity of CD8⁺ T cells and NK cells. Meanwhile, MSC-IL-2 N103R V106D supernatant more effectively promoted a higher proportion of Treg cells compared to MSC or 293 T-IL-2 N103R V106D supernatant (Fig. 4F, G). Furthermore, MSC-IL-2 N103R V106D significantly increased the expression of Treg-related functional genes of B6 mice splenocytes, including Foxp3, TNFR2, CD25, and others. (Fig. 4H). Conversely, wild-type IL-2 lentivirus-transduced MSC significantly elevated the expression of cytotoxic lymphocyte-associated genes in B6 mice splenocytes, such as perforin and granzyme compared to the conventionally cultured splenocytes, a difference that was statistically significant (Fig. 4I).

MSC-IL-2 N103R V106D is effective in treating rheumatoid arthritis by elevating Treg cell levels in CIA mice

To investigate the efficacy of MSC-IL-2 N103R V106D in collegan-induced arthritis, CIA mice were treated via tail vein injection with lentivirus-transduced MSC on days 28, 35, 42, and 49 post-collagen induction. Joint symptom scores, along with histopathological sections of the ankle joints, were assessed every 2 days. Mice were executed on day 56 post-collagen induction. Compared to MSC or IL-2-overexpressing MSC treated CIA mice, lentiviral transduction of MSC expressing the N103D and V106D mutant IL-2 injection significantly alleviated joint erythema and arthritis scores following injection into CIA mouse models via the tail vein (Fig. 5A, B). Histological analysis of H&E-stained ankle joint sections revealed reduced inflammatory cell infiltration in the MSC–IL-2 N103R V106D treatment group. Furthermore, Safranin O–Fast Green and TRAP staining indicated reduced cartilage destruction and decreased osteoclastic activity in the MSC–IL-2 N103R V106D treatment group. (Fig. 5C), and H&E hematological scores were lower in MSC-IL-2 N103R V106D-treated mice compared to other groups (Fig. 5D).

Fig. 5.

Alleviation of arthritis severity and scores by MSC-IL-2 N103R V106D. In this study, the CIA group received PBS treatment, while various MSC groups received four infusions of 5 × 10.⁵ cells per mouse via the tail vein. A MSC-IL-2 N103R V106D infusion treatment reduces arthritis score (Two-way ANOVA with Bonferroni’s. n = 5). B Representative joint images from different groups. C Representative images of ankle joints stained with hematoxylin and eosin (H&E), safranin O/fast green, and TRAP in different treatment groups. D MSC-IL-2 N103R V106D infusion therapy reduces H&E-stained sections histological score (Kruskal–Wallis analyses with Dunn’s post-hoc. n = 5). * P, < 0. 05

Flow cytometry confirmed significantly elevated spleen Treg cell levels in CIA mice injected with MSC-IL-2 N103R V106D, with a statistically difference (Fig. 6A, B), and it did not increase spleen effector T cell levels, unlike IL-2-overexpressing MSC (Fig. 6C, D). In joint tissues, both MSC-IL-2 N103R V106D and MSC-IL-2 treatments increased STAT5 phosphorylation downstream of the IL-2 receptor in CD4⁺ T cells (Fig. 6E, F), whereas the proportion of CD8⁺ T cells was lower in mice treated with MSC-IL-2 N103R V106D compared to those treated with MSC-IL-2 wild-type (Fig. 6G, H). Additionally, functional genes related to effector cells, including granzyme, perforin, and T-bet, which were elevated in IL-2-overexpressing MSC treated CIA mice were not elevated in MSC-IL-2 N103R V106D-treated CIA mice compared to the control group, (Fig. 6I). This confirms the therapeutic potential of lentiviral transduction of N103D and V106D mutant IL-2-expressing MSC in treating rheumatoid arthritis, by selectively targeting and enhancing Treg cell proliferation and function without affecting effector T cells and NK cells.

Fig. 6.

Elevation of Treg cell proportion by MSC-IL-2 N103R V106D in CIA mice without affecting CD8 + T cells and related genes. A-D Representative flow cytometry plots (A, C) and statistical analysis graphs (B, D) showing the proportion of Treg cells in splenocytes was elevated in the MSC-IL-2 N103R V106D-treated group of mice while the proportion of CD8 + T cells decreased compared to the MSC-IL-2-treated group (ANOVA with Tukey’s post-hoc. n = 5). E-D Representative flow cytometry plots and statistical analysis graphs showing Both MSC-IL-2 N103R V106D and MSC-IL-2 enhanced STAT5 phosphorylation downstream of the IL-2 receptor in mice joint CD4⁺ T cells. E, F However, the proportion of joint CD8⁺ T cells was reduced in mice treated with the mutant MSC-IL-2 N103R V106D compared to those treated with wild-type MSC-IL-2. G, H (ANOVA with Tukey’s post-hoc. n = 5). I The expression levels of genes associated with effector T cells were lower in the splenocytes of mice treated with MSC-IL-2 N103R V106D compared to those treated with MSC-IL-2 (ANOVA with Tukey’s post-hoc. n = 5). *, P < 0. 05, **, P < 0. 01

Discussion

RA, a prevalent immune system disease primarily affecting small and medium-sized joints, has its pathogenesis significantly influenced by the scarcity of Treg cells, a critical immunoregulatory cell type [5, 6]. While Treg cell transplantation has shown efficacy in treating RA [45], this approach may encounter challenges such as low specificity and potential for non-specific rejection. IL-2, a critical immunomodulatory factor, has garnered sustained attention for its role in treating autoimmune diseases and cancer. Thirty years ago, IL-2 became the first immunotherapy drug approved by the US Food and Drug Administration for treating metastatic renal cell carcinoma and metastatic melanoma[46, 47]. Subsequently, the role of IL-2 in regulating Treg cells has been further explored, leading to its use in treating various autoimmune diseases[19, 20, 48]. Research indicates that IL-2 expression levels are elevated in patients with rheumatoid arthritis (RA) and positively correlate with osteoarthritis. This elevation may be due to aberrant T-cell activation and a compensatory increase in IL-2 release by the immune system to regulate its response[49, 50]. Given IL-2 being able to activate various immune cells, ensuring its specific targeting in disease treatment is crucial. Additionally, IL-2, being a small molecule, is endocytosed and degraded after binding to the IL-2 receptor on the cell surface, resulting in a short half-life (< 15 min)[51]. Overcoming the short half-life of IL-2, prolonging its in vivo action, and reducing the frequency of dosing are key research objectives.

Low-dose IL-2 therapy in vivo could restore the imbalance between autoimmune response and self-tolerance toward self-tolerance via promoting Treg cell expansion and inhibiting follicular helper T (Tfh) and IL-17-producing helper T (Th17) cell differentiation and compared to standard therapy alone, low-dose IL-2 therapy, when combined with the respective standard treatments for systemic lupus erythematosus and rheumatoid arthritis, demonstrates improved efficacy and safety. [48, 52–54]. However, due to the short half-life of low-dose IL-2, prolonged and frequent administration of IL-2 does not result in a stable long-term increase in Treg cells and potentially causes off-target effects[34, 55]. Enhancing the targeting and half-life of IL-2 through engineering technology to achieve better efficacy is a topical area of current research, The four main types of IL-2 engineered proteins include IL-2 muteins(IgG–(IL-2N88D)2), PEGylated IL-2(NKTR-358), IL-2-anti-IL-2 immune complexes(IL-2–JES6) and IL-2-CD25 fusion proteins[30, 34, 56–58]. These engineered proteins preferentially stimulate the IL-2 trimeric receptor expressed on Treg cells by relatively decreasing the capacity of IL-2 to interact with CD122 (IL-2Rβ) and delaying IL-2 receptor-mediated clearing of IL-2 in vivo compared to wild-type IL-2, thereby prolonging half-life. These engineered IL-2 have achieved relatively satisfactory results in clinical trials or animal experiments.

In this study, a mutant IL-2 N103R V106D, characterized by a reduced affinity for IL-2Rβ and enhanced dependency on the IL-2α receptor, was explored as a potential treatment for RA.. This mutation facilitates targeted activation of Treg cells. Additionally, previous research indicates that MSC contribute to an increased proportion of Treg cells in patients with autoimmune diseases[59–61]. While both MSC infusion and low-dose IL-2 injection are effective individually for treating autoimmune diseases, previous studies suggest that their combination does not yield the desired therapeutic effect, potentially due to off-target activation of effector T cells[62]. Enhanced Treg cell targeting and more stable, sustained IL-2 expression in MSC were achieved by transducing them with the mutant IL-2 N103R V106D lentivirus, offering a promising therapeutic approach for autoimmune diseases. By selectively targeting and elevating Treg cells, without activating or increasing CD8⁺ T cells and NK cells, MSC-IL-2 N103R V106D demonstrates an immunosuppressive effect and has proven effective in treating collagen induced arthritis, providing a foundation for the clinical exploration of new treatments for rheumatoid arthritis.

The mutant IL-2 N103R V106D demonstrated improved targeting of the trimeric IL-2 receptor and Treg cells compared to wild-type IL-2, resulting in a reduced capacity to stimulate CD8⁺ T cells and NK cells. Transducing MSC with this mutant enabled more stable expression and fewer treatment cycles. However, this mutant IL-2 variant increases the selectivity for CD25 (IL-2Rα) but reduces overall activity, necessitating higher treatment doses for efficacy. Additionally, activated effector T cells and CD56 bright NK cells may also express the trimeric IL-2 receptor in vivo, leading to off-target effects, a side effect present in the treatment of autoimmune diseases with low-doses IL-2.[34]. Lentiviral transduction of MSC ensures the efficient expression and secretion of IL-2 N103R V106D. The secretion of mutant IL-2 by MSC circumvents the need for frequent injections associated with conventional IL-2 therapy and addresses the issues of dose-dependence that can lead to the non-targeted activation of pro-inflammatory cells and exacerbation of the disease[63]. Together, MSC and mutant IL-2 synergistically modulate the immune response, presenting a promising strategy for the treatment of arthritis.

Various engineered IL-2 mutants have been utilized in clinical research pertaining to oncology and autoimmune diseases. Our subsequent experimental objectives might involve conducting clinical studies on this specific mutant IL-2-modified MSC. MSC are genetically modified via lentiviral transduction to secrete mutant IL-2 N103R V106D. This process requires further standardization to ensure its safety for use in clinical trials. Moreover, the therapeutic potential of MSC-IL-2 N103R V106D in autoimmune diseases such as systemic lupus erythematosus, Sjogren's syndrome, and type 1 diabetes mellitus merits comprehensive exploration. Additionally, future investigations could focus on combining this mutant form of IL-2 with other biologics. This approach could integrate multiple mechanisms to modulate the immune response, potentially offering more extensive immunosuppression and enhanced therapeutic outcomes.

Umbilical cord MSC, as a type of stem cell with excellent self-renewal and differentiation capabilitie, can be collected more conveniently and painlessly than other MSC types, exhibit stronger proliferation capabilities, and can secrete a large number of growth factors. Clinical studies have demonstrated their efficacy in treating rheumatoid arthritis [64] and suggested that the imbalance of Th17/Treg is a crucial mechanism in the pathogenesis of RA [65]. TGF-β secreted by MSC enhances the development of Treg and synergizes with mutant IL-2 to promote Treg cells while inhibiting Th1 and Th17 proliferation. This helps maintain the Th17/Treg balance critical for RA treatment [66]. Genetic modification using viral or non-viral vectors has been explored to enhance the therapeutic potential of MSC therapy [67]. Additionally, in rheumatoid arthritis treatment, strategies such as co-culture methods, the use of growth factors, cytokines, receptor agonists, hypoxia, autophagy, and genetic modification have been investigated to boost the therapeutic efficacy of MSC [68, 69]. Summarized in the report by Sarsenova et al. Combining MSC with IL-10-producing Treg cells significantly enhances the reduction of inflammation and prevents destructive arthritis in mice [70].

An limitation of this study is the inability to purify the mutant IL-2 N103R V106D monomer, coupled with the exclusive reliance on lentivirally transduced 293 T cell supernatants. This approach might inadvertently include the effects of extraneous substances present in the cell supernatants. Furthermore, due to the limited number of Treg cells detected in the mice joints, only the activation of STAT5 downstream of the IL-2 receptor in joint CD4⁺ T cells was assessed following MSC-IL-2 N103R V106D and MSC-IL-2 treatment, and a precise assessment of Treg cell proportions at the site of joint inflammation remains to be clarified. Concurrently, to better evaluate joint inflammation and overall immune status, additional assessments of joint thickness and anti-collagen antibody levels are warranted. Moreover, further exploration is needed to understand the specific pharmacokinetics and the levels and duration of in vivo IL-2 expression. Additionally, this study did not compare the therapeutic effects of mutant IL-2 N103R V106D with those of low-dose IL-2, despite the broader range of therapeutic doses available. Ultimately, Despite acting mainly on Treg cells, The precise mechanism driving the synergistic interaction between MSC and mutant IL-2 N103R V106D in the treatment of RA requires in-depth study.

The study revealed that genetically modified MSC expressing mutant IL-2 N103R V106D effectively alleviated arthritis in CIA mice through targeted engagement of Treg cells. In vivo self-expansion of Treg cells circumvents the challenges associated with Treg cell extraction and expansion, as well as the risk of autologous rejection that can occur with in vitro allogeneic transplantation[71].The utilization of relevant animal models for experimentation further propels this therapeutic strategy, potentially paving the way for its application in clinical trials thereby pioneering new treatments for patients with rheumatoid arthritis.

Abbreviations

IL-2

interleukin-2

RA

Rheumatoid arthritis

CIA

collagen-induced arthritis

MSC

mesenchymal stromal cell

Treg

Regulatory T cells

NOD

non-obese diabetic

PFA

paraformaldehyde

DMARDs

disease-modifying anti-rheumatic drugs

FBS

fetal bovine serum

HRP

horseradish peroxidase

ELISA

enzyme-linked immunosorbent assay

H&E

hematoxylin and eosin

TRAP

Tartrate resistant acid phosphatase

Author contributions

Zhicheng Tang, Fan Yang and Jingyi Shen contributed equally to this manuscript as co-first authors. Zhicheng Tang and Jingyi Shen performed experiments, analyzed the data, and wrote the manuscript. Fan Yang revised the manuscript. Haolin Wu and Huiming Hong performed animal experiments. Yue Wang and Fanzhang Yin performed pilot experiments. Xiaojun Tang conceptualized the idea, provided technical support, and revised the manuscript. Huayong Zhang granted project funding and supervised the project. All authors approved the final manuscript.

Funding

The work was supported by the National Natural Science Foundation of China (grant numbers: 81802126, 81671608).

National Natural Science Foundation of China,81802126,81802126,81802126,81802126,81802126,81802126,81802126,81802126,81802126

Data Availability

No datasets were generated or analysed during the current study.

Declarations

Ethics approval and consent to participate

Animal experiments are carried out under the project license (No. 2020AE01061) approved by the Ethics Committee of Drum Tower Hospital, Nanjing Medical University.

Competing interests

The authors declare no competing interests.