Exploring MHC class II I-A^b blockade as a potential treatment for Sjögren’s disease in the mouse model

Abstract

Sjögren’s disease (SjD) is a chronic autoimmune disorder predominantly affecting females, characterized by exocrine gland dysfunction. This study investigates the therapeutic potential of 2-chloro-1-(4-hydroxy-phenyl)-ethanone (CHPE) and metformin in the C57BL/6.NOD-Aec1Aec2 mouse model, which closely mirrors human SjD. Molecular docking identified CHPE and metformin as high-affinity binders to the MHC class II I-A^b antigen-binding groove, suggesting their ability to inhibit antigen presentation and modulate immune responses. In-vitro assays confirmed their effectiveness in reducing T cell activation. In-vivo studies demonstrated that both preventative and therapeutic regimens of CHPE and metformin significantly reduced lymphocytic infiltration in the lacrimal glands, with metformin showing a more pronounced effect in females. Salivary gland infiltration was less responsive, though some reduction in focal scores was observed in male mice treated preventatively with CHPE. Both drugs altered the composition of lymphocytic infiltrates, particularly by reducing B cell populations, with notable sex-specific differences in response to treatment. CHPE and metformin also reduced anti-nuclear antibody levels, with CHPE showing stronger effects in females. Additionally, both drugs improved saliva and tear secretion, with metformin being more effective in the preventative regimen, especially in females. T cell receptor transductant assays revealed that CHPE and metformin exert their therapeutic effects through antigen-specific pathways, inhibiting T cell responses to SjD-associated autoantigens. Overall, this study provides compelling evidence that CHPE and metformin can modulate immune responses and improve gland function, with effectiveness varying by sex and age. These findings support the potential of these compounds as personalized treatments for SjD tailored to individual patient characteristics.

Introduction

Sjögren’s disease (SjD) is an autoimmune chronic inflammatory disease that is characterized by exocrine dysfunction^1–4; it disproportionately affects females.^5–9 The pathogenesis of SjD involves abnormal innate and adaptive immune responses, where an overproduction of T helper 17 (Th17) and T follicular helper (Tfh) cells persist in an unbalanced frequency with their regulatory counterparts, regulatory T cells (Tregs).¹⁰ Interaction of cognate T and B cells induces a dysregulated B cell response and the formation of ectopic germinal center (GC)-like structures in SjD.^11,12 This leads to an increase in autoreactive B cell activation and autoantibody production. Unfortunately, there are very few therapies available, and those that are available are broadly effective, with many side effects (ie, corticosteroids, biologics, and disease-modifying antirheumatic drugs, DMARDs).¹³

There are several mouse models of SjD, but they mainly present comorbidly with another autoimmune disorder. Here, we use the C57BL/6.NOD-Aec1Aec2 mouse model, which exhibits an SjD phenotype alone, has the major histocompatibility complex (MHC) Class II I-A^b. SjD presents in three phases in this model. Phase 1 (0 to 8 wk) is characterized by acinar epithelial cell death and delayed salivary gland (SG) morphogenesis. Phase 2 (8 to 16 wk), where interferon (IFN)-stimulated genes become activated, coincides with migrating macrophages and dendritic cells (DCs), followed by CD4⁺T and B220⁺B lymphocytes and the emergence of autoantibodies. Finally, in Phase 3 (16 wk onward), there is an overt clinical disease with a progressive measurable loss of exocrine gland function. Previously, small molecules have been utilized to ameliorate type I diabetes^14,15 and the SjD phenotype in the non-obese diabetic (NOD/ShiLtJ) mouse,¹⁶ which has MHC Class II I-A^g.7 C57BL/6.NOD-Aec1Aec2 mice develop SjD symptoms temporally and phenotypically similar to NOD/ShiLtJ mice but without diabetes, making them an ideal candidate to study spontaneous SjD^14,17,18 by screening small molecules for their druggability/capacity to sit in pockets of this MHC Class II I-A^b.

2-Chloro-1-(4-hydroxy-phenyl)-ethanone (CHPE) is a reagent used in the preparation of hydroxypyrimidine derivatives and their HDAC (histone deacetylases) inhibitory activity.¹⁹ Metformin, conversely, is an anti-diabetic drug that reduces hepatic gluconeogenesis by AMP-activated protein kinase (AMPK)-dependent and AMPK-independent mechanisms.^20–22 CHPE and metformin were analyzed to observe systemic effects and changes in the salivary glands' Th17 and Th1 cell populations. In addition, in this study, we examine the impact the drugs have on helper T cell activation of B cells, accumulation of anti-nuclear antibodies, and lymphocytic infiltration of the exocrine glands. Finally, this study aims to ascertain if sex and/or the age of drug administration affects the efficacy of these small molecules as either a prophylactic or therapeutic intervention in B6.NOD-Aec1Aec2 mice, both drugs were studied in male and female, old and young cohorts of mice; together, these findings offer insights into how the aberrant activity of innate and adaptive immune cells mediate the pathogenesis of SjD and indicate the possibility of personalized medicine-based on the patient’s age, sex, disease activity, and genes.

Materials and methods

Mice

SjD-susceptible (SjD^S) C57BL/6 J.NOD-Aec1Aec2 and C57BL/6 J (B6) control mice were housed under specific pathogen-free conditions in the animal facilities of the University of Florida Animal Care Services. The breeding and use of animals described herein were approved by and conducted under the direction of the University of Florida Institutional Animal Care and Use Committee. All methods were performed per the relevant guidelines and regulations. The development of C57BL/6.NOD-Aec1/2 mouse and its SjD-like disease phenotype are described previously.^23,24 Briefly, the SjD^S mouse was developed by introducing two genetic regions, one on chromosome 1 (designated Aec2) and the second on chromosome 3 (designated Aec1) derived from the NOD/LtJ mouse into the B6 mouse. All animals were maintained on a 12-h light-dark schedule, and food and acidified water were provided ad libitum. At times indicated in the study, mice were euthanized by cervical dislocation after deep anesthetization with isoflurane, and their organs and tissues were freshly harvested for analyses.

The study utilized two groups of mice: the preventative group, consisting of male and female mice aged 15 to 21 wk, and the therapeutic group, consisting of male and female mice aged 29 to 33 wk. Both groups were treated with either CHPE or metformin for 14 wk. The dosing regimen was 20 mg/kg of CHPE administered twice a week and 20 mg/kg of metformin administered 5 d a week. Age- and sex-matched PBS-injected mice served as control groups for each cohort. Metformin is commonly used in animal models for various diseases, with doses ranging from 20 to 205 mg/kg body weight.^21,25–27 In contrast, CHPE is seldom used and has not been previously tested in mouse models. The C57BL/6.NOD-Aec1/2 strain is highly sensitive to drugs, especially in older mice at higher doses. Consequently, we initiated a survival dose experiment. We selected 5 to 10 mice aged 29 to 33 wk and treated them with either metformin (5 d a week) or CHPE (twice a week) at doses of 20, 50, and 100 mg/kg body weight. The mice were monitored for three weeks, with daily assessments for Bright, Alert, Responsive (BAR) status and survival. These assessments indicated that 20 mg/kg was the optimal dose for metformin and CHPE.

Molecular modeling and docking

Crystal structures of I-Ab complexed to the 3K peptide, Protein Data Bank code 1ES0, were used as the basis for molecular docking. The site preparation for docking included removing all water molecules and protonation of I-Ab residues with SYBYL (Tripos), which was done previously.^14,15 The SPHGEN program was used to define the sites selected for molecular docking. The methodology included generating a grid of points that reflected the shape of the selected site and filtering the results through CLUSTER. The final step included energy scoring for van der Waals forces of attraction and contact scoring using the algorithm. Atomic coordinates for ∼140,000 small molecules in the National Cancer Institute Developmental Therapeutics Program (NCI/DTP repository) were positioned in the selected structural pocket in 1,000 different orientations and scored based on predicted polar (H bond) and nonpolar (van der Waals) interactions. The best orientation and scores (contact and electrostatic) were calculated. PYMOL was used to generate molecular graphic images.

In-vitro screening of small molecules

Small molecules for screening were obtained from the NIH’s Developmental Therapeutics Program (DTP) repository in vialed sets, dissolved in DMSO, and diluted in PBS for a final concentration of 0.1% DMSO in each well. Each small molecule was screened at a concentration of 100 μM as previously determined to have the optimal inhibitory effect.^14,15 The B3K506 TCR transductant was cultured with the 3K peptide (ASFEAQKAKANKAVDKA), with or without small molecules, in the presence of murine C57BL/6.NOD-Aec1/2 splenic cells. Where indicated, plates were coated (4 °C, overnight) with 50 µl of 10 µg/ml anti-CD3 (BD Pharmingen, Franklin Lakes, New Jersey). After washing the plate, 5 µg/ml anti-CD28 (BD Pharmingen, Franklin Lakes, NJ) was added prior to other well contents. Single-cell suspension of spleens was achieved, followed by red blood cell lysis (5 ml cold lysis buffer [0.802% NH4Cl, 0.084% NaHCO3, and 0.037% EDTA] for 13.5 min), and then 2 × 10⁵ splenoctyes were cultured with Ro60 peptide (LFTFIQFKKDLKESM, 100 µg/ml) with CHPE or Metformin for 48 hours in a 96-well, U-bottomed plate. The supernatant was collected to measure IL-2 secreted by the TCR hybridoma and splenocytes using an ELISA (R&D DuoSet, R&D Systems, Minneapolis, Minnesota) per the manufacturer's instructions.

Measurement of saliva flow rate

To measure stimulated flow rates of saliva, individual mice were weighed and given an intraperitoneal (IP) injection of 100 μl of a mixture containing isoproterenol (0.2 mg/1 ml of PBS) and pilocarpine (0.05 mg/1 ml of PBS). Saliva was collected for 10 min from the oral cavity of individual mice using a micropipette starting 1 min after injection of isoproterenol and pilocarpine. The volume of each saliva sample was measured.

Measurement of tear flow

To measure stimulated tear flow, individual mice were weighed and given an IP injection of pilocarpine (0.05 mg/1 ml of PBS). A phenol red thread is inserted next to the conjunctiva of the mouse after one minute post-injection, which indicates a color change based on the number of tears produced. The length of the thread stained was recorded in millimeters (mm).

Histological examination of the salivary and lacrimal glands

Salivary and lacrimal glands, including cervical draining lymph nodes of C57BL/6.NOD-Aec1Aec2 were fixed in 10% phosphate-buffered formalin for 48 h. Fixed tissues were embedded in paraffin and sectioned at a thickness of 5 μm. The prepared tissue sections were stained with hematoxylin and eosin (H&E) dye (Histology Tech Services, Gainesville, Florida). Stained sections were observed at 200× magnification using a Nikon Eclipse Ti-E inverted microscope. A blinded examiner examined a single histological section per gland per mouse to detect and determine leukocytic infiltrations in salivary glands. Lymphocytic infiltrations were defined as aggregates of >50 leukocytes.

Evaluation of anti-nuclear antibody staining patterns

Examination of anti-nuclear antibody staining patterns was performed as previously described. In brief, sera were diluted 1:40 in PBS and incubated on permeabilized HEp2 ANA slides (ImmunoConcepts, Sacramento, California) prior to incubation with FITC secondary antibody sections were observed at 200× magnification by using Nikon Eclipse Ti-E inverted microscope.

Examination of salivary gland T helper cells, cytotoxic T cells, and B cells by flow cytometry

Salivary glands were digested in a digest buffer (1 mg/ml DNase (Sigma-Aldrich, St Louis, Missouri, USA) and 1 mg/ml Collagenase Type 4 (Worthington, Lakewood, New Jersey, USA) in RPMI (Lonza, Allendale, New Jersey, USA) complete media containing 10% FBS, 2 mM L-glutamine, 0.05 mM β-mercaptoethanol) and placed in a MACS C tube (Miltenyi Biotec, San Diego, California, USA) for desiccation on GentleMACS V1.02 for a pulse of 38 s, twice. After a 10-min incubation at 37 °C, the digest buffer was removed and placed into 4 °C RPMI complete media, and the digest buffer was replaced with another digest buffer to repeat the process twice. Single-cell suspensions were centrifuged (2,500 rpm, 10 min, 4 °C) and resuspended in PBS for filtration through a 70-μm sterile cell strainer (Fisher, Pittsburgh, Pennsylvania, USA). After a wash with PBS, cells were resuspended again in PBS for lymphocyte isolation with Lympholyte-M cell separation media (Cedar Lane, Burlington, Ontario, Canada) per the manufacturer's instructions. Single-cell suspensions were stained for surface markers using the following stains: PE for CD3 (BD Bioscience catalog no. 553064), PerCP-Cy5.5 for CD4 (Biolegend catalog no. 116011), FITC for CD8 (BD Bioscience catalog no. 553031), APC for IFN-Gamma (BD Bioscience catalog #552772), BV421 for IL-17(Biolegend catalog no. 505810), PE Cy7 for B220 (Biolegend catalog no. 506926), AmCyan as a live dead marker (ThermoFisher catalog no. 2031176). The samples were analyzed using BD Fortessa Flow Cytometer (BD Biosciences, San Jose, California, USA), and analysis was performed using FlowJo VX software (FlowJo, Ashland, Oregon, USA).

Third generation lentiviral transduction

TCR sequences were identified via 10X Genomics Single Cell 5' expression with V(D)J enrichment, and full-length TCR (paired alpha/beta chains) (method below) were custom-made by GenScript (Piscataway, New Jersey) as an insert in pcDNA3.1. After receiving the TCR-containing plasmids, they were transformed into NEB DH5alpha (New England Biolabs, Ipswich, Massachusetts) per the manufacturer’s instructions. Colonies were picked after growth on 100 ug/ml ampicillin agar plates for growth in LB broth with 100 µg/ml ampicillin. After plasmid purification with the QIAprep Spin Miniprep kit (Hilden, Germany), the plasmid was sent to Genewiz (South Plainfield, New Jersey) for sequence confirmation. Meanwhile, the following bacterial stabs were received from Addgene: pMDLg/pRRE (12251), pCMV-VSV-G (8454), pRSV-Rev (12253), and pFUGW-H1 empty vector (25870). Colonies were picked after growth on 100 ug/mL ampicillin agar plates for growth in LB broth with 100 ug/mL ampicillin. After plasmid purification with QIAprep Spin Miniprep kit (Hilden, Germany), double digest with NotI and AgeI was performed for all vectors except pRSV-Rev, which was digested only with EcoRI, and run on an agarose gel for band determination. After confirmation, colonies were expanded and purified using a QIAgen Plasmid Maxi kit (Hilden, Germany); plasmids were sent to sequencing for confirmation with Genewiz. For TCR insert into pFUGW-H1, the plasmid was cut with BamHI (New England Biolabs, Ipswich, Massachusetts), then ligated with T4 Ligase (New England Biolabs, Ipswich, Massachusetts) per manufacturer’s instructions. Due to a single digest, rSAP (New England Biolabs, Ipswich, Massachusetts) was utilized to link the digest and ligation steps. After growth and purification above (with BamHI cutting for band confirmation), H1-TCR plasmid was sent to Genewiz for sequence confirmation.

HEK293T cells were grown in a T75 flask with cDMEM + Ab (Gibco, Waltham, Massachusetts) to 80% confluency (5% CO2, 37 °C). Media was replaced with antibiotic-deficient DMEM, while 37.5 µl Lipofectamine 3000 (ThermoFisher, Waltham, Massachusetts) in 1.25 ml OPTI-MEM (Gibco, Waltham, Massachusetts) was incubated for 5 mins at room temperature. H1-TCR (27.5 ug) and helper plasmids (VSVg, pMDLg/pRRE, and pRSV-Rev; 9.15 µg each) were mixed in 1.25 ml OPTI-MEM and incubated for 5 mins at room temperature. After both OPTI-MEM mixtures were combined with 55 uL P3000 and incubated for 15 min at room temperature, the transfection mixture was added dropwise to the flask. Likewise, 2.5 ml 10X caffeine was added to the flask. T75 was then incubated at 37 °C for 3 d. The supernatant was collected in 50 ml conical tube, and centrifuged (500 × g, 10 min, 4 °C). The supernatant was then filtered through a 0.45 um cellulose acetate filter (ThermoScientific, Waltham, Massachusetts) to remove cell debris before centrifugation in a Beckman (Brea, CA) XL-90 ultracentrifuge (55k rpm, 1.5 h, 4 °C). The supernatant was removed, and pellets were washed with 25 uM HEPES in PBS (Hyclone, Logan, Utah). After overnight incubation at 4 °C, the lentivirus was aliquoted and frozen for 2 h at −80°C. 2.5 × 10⁵ Jurkat cells/well were plated with 0 µl (no virus control), 0.1 µl, 0.5 µl, 1 µl, 5 µl, or 10 µl lentivirus concentrate on a 24-well plate. For the APC-peptide presentation, see the section below. Plates were incubated (5% CO₂, 37 °C, 2 d), then supplemented with additional media. After incubating for an additional day, Cytek Aurora (Cytek Biosciences, Fremont, California) flow cytometry was performed on cells to determine the presence of TCRs via GFP fluorescence.

Antigen presentation assay using Ro and La peptides

Single-cell suspension of splenocytes was achieved by mashing the spleen through a 70-µm strainer, followed by RBC depletion with ammonium chloride. Cells were stained with 3 uL PE anti-CD3, and the manufacturer’s protocol for anti-PE microbeads was followed (Miltenyi Biotec, Germany); labeled cells were run on a MACS MS column for isolation of antigen-presenting cells. APCs were treated with Mitomycin C (20 min, 37 °C), then plated on a 24-well plate at a density of 2.5 × 10⁵ APCs per well. Peptides were added at 100 µg/ml concentrations: Ro52 ANPWLILSEDRRQVR, Ro60 LFTFIQFKKDLKESM, and La KEALKKIIEDQQESL. As a negative control for peptide specificity, we selected an unrelated peptide, corresponds to ribonucleotide reductase subunit 1 from human alphaherpesvirus 1, positions 709 to 721 (NVTWTLFDRDTSM). The peptide estimated to bind H2-IA^b with low affinity. CHPE and metformin were both added for a well concentration of 100 µM. Cells were incubated as indicated previously.

TCR-pMHC docking

Peptides were docked onto MHC for presentation to a unique SjD^S mouse TCR via TCRmodel,²⁸ and an image was generated with docking in RCSB PDB Mol* 3D Viewer.²⁹ The previously identified sequence utilized here was: Vα 14-3, CDR3α CAATNTGYQNFYF, Jα 49, TRAC, Vβ 13-1, CDR3β CASSANSGNTLYF, Jβ 1-3, and TRBC sequences with a GSG linker and P2A between the alpha and beta sequences.

10× genomics single-cell sample processing, cDNA library preparation, and TCR analysis

Samples were prepared using the Chromium Next GEM Single Cell V(D)J Kit v1.1, mouse (10× Genomics, Pleasanton, California) following the manufacturer's instructions. In brief, sorted single cells of the salivary glands for each sample were resuspended in RPMI containing 10% FBS to a final concentration of 700 to 1200 cells/µl. A total of 8 samples were loaded onto a Chromium Next GEM Chip G, analyzed by the Chromium Controller (10× Genomics, Pleasanton, California) for Gel Beads-in-emulsion (GEMs) generation and reverse transcription. The generated cDNA was purified with SPRIselect (Beckman Coulter Inc, Indianapolis, Indiana) and used for 5′ gene expression with V(D)J library construction. The cDNAs and libraries were examined for quality control using D5000 ScreenTape (Agilent Technologies, Waldbronn, Germany), and Qubit (Thermo Fisher Scientific, Waltham, Massachusetts) was used for quantification. To achieve 5,000 reads per cell for the V(D)J library, the libraries were sequenced using the Illumina NovaSeq6000 system (Illumina, San Diego, California). scRepertoire (v2.0.4) was used to analyze TCR V(D)J data from the 10× Cell Ranger. Analysis of clones between samples and changes in dynamics was done using the clonalCompare() function in the same package.

Statistical analysis

All presented data are the mean ± standard error of the mean (SEM). Statistical significance was tested using GraphPad Prism 9 (Prism, Irvine, California). All results are discerned to the 95% confidence interval, and P > 0.05 was considered statistically significant.

Results

Molecular docking of MHC class II I-A^b and in vitro screening of drug-like small molecules

We employed the molecular docking process to assess the efficacy of compounds identified through an in silico rational approach for binding to specific pockets of the MHC class II I-A^b in inhibiting antigen presentation and deactivating T cell function. In total, 139,735 small molecules from the NCI/DTP repository were analyzed for their potential to bind distinct structural pockets within the I-A^b antigen-binding groove. Identifying the antigen-binding clefts of I-A^b was facilitated by pinpointing sites of essential contact between the compounds and MHC. Each compound was oriented in 1,000 different positions using the DOCK program algorithm for the docking process. The scoring of these compounds was based on their cumulative attractive and repulsive polar and nonpolar interactions, which helped in estimating the free energy (ΔG) of binding. Focus was directed toward individual pockets at positions 3, 6, 9, and 11 for predicted compound binding affinity. Notably, buformin and metformin were positioned in pocket 9. In contrast, CHPE was located in pocket 6 of I-A^b (Fig. 1A). The top 40 compounds were further evaluated for their capacity to mitigate T cell responses within the T cell hybridoma B3K506, employing I-A^b splenocytes as antigen-presenting cells. The assay measured secreted IL-2 levels to determine T cell proliferation; a decrease in IL-2 levels indicates a significant inhibition of the 3K peptide presentation to I-A^b. This led to the identification of seven molecules that significantly reduced IL-2 secretion by T cells (Fig. 1B) with predicted binding to the antigen-binding cleft of I-Ab at specific residues that formed intermolecular contacts with peptides at position 6 (Fig. 1C). Lastly, to determine whether the high-affinity small molecules target I-A^b in an antigen-specific manner, we stimulate I-A^b splenocytes with anti-CD3/CD28 and SjD-associated Ro60 peptide. Both CHPE and metformin lowered the activation of T cells by decreasing IL-2 secretion compared to anti-CD3/CD28 stimulation. When Ro60 was added to the simulation, CHPE and metformin exhibited a more significant decrease of IL-2 than anti-CD3/CD28 stimulation alone. The data suggest that CHPE and metformin could lower the function of T cells via a non-MHCII stimulation like anti-CD3/CD28; however, the impact was significantly higher using MHCII stimulation with SjD-specific antigen. Based on their potent inhibitory effect and well-defined toxicity in animal models, metformin and CHPE, were chosen for further analysis in this study.

Figure 1.

Structure-based selection of small molecules predicted to bind MHC II I-A^b and validation by in vitro screening. (A) The crystal structure of I-A^b is shown in panel A, red for pockets 6, green for the I-A^b α-chain, and violet for the I-Ab β-chain (PDB code 1MUJ). (B) 139,735 small molecules in the NCI DTP 2007 plated set were screened by high-throughput molecular docking to identify candidates predicted to bind the antigen-binding cleft of I-A^b. The top 40 scoring compounds (based on estimated ΔG values) were assayed for effects on IL-2 secretion. Buformin, metformin, and CHPE are shown as spheres in central panels. (C) Buformin, metformin, and CHPE are shown sticks in the right panels. Buformin, metformin, and CHPE were predicted to bind the antigen-binding cleft of I-A^b at residues (indicated in red) that formed intermolecular contacts with peptides at position 6 (P6) in 1MUJ. (D) I-A^b splenoctyes of C57BL/6.NOD-Aec1/2 were stimulated with anti-CD3/CD28 (Stim) or without anti-CD3/CD28 (No Stim). Ro60 peptide was added to the culture at 100 µg/ml. CHPE or Metformin were used at 100 μM final concentration to evaluate their inhibitory effect. After 48 h, the supernatant was collected to measure IL-2 secretion using an ELISA (R&D DuoSet, R&D Systems, Minneapolis, MN) per the manufacturer's instructions. Statistical analysis was done using 1-way ANOVA with ****P < 0.0001. The experiment was performed in triplicate.

CHPE and metformin reduced lymphocytic infiltration restricted to the lacrimal glands, not the salivary glands

In the B6.NOD-Aec1/2 animal model, which closely mimics SjD with autoimmune exocrinopathy marked by salivary and lacrimal gland infiltration and the development of autoantibodies by 12 to 16 wk and secretory dysfunction by 24 w of age, we sought to assess the efficacy of the candidate drugs CHPE and metformin in inhibiting I-Ab antigen presentation. This intervention is conceived to prevent the autoimmune process if administered before the full SjD phenotypes emerge or to alleviate the disease symptoms after their onset. To achieve this, we employed 2 mouse cohorts: a preventative group (male and female mice aged 15 to 21 wk old) and a therapeutic group (male and female mice aged 29 to 33 wk). Both groups were treated with CHPE or metformin for 14 wk, at a dose of 20 mgs/kg twice a week (CHPE) and 20 mgs/kg 5 d a week (metformin), with age- and sex-matched PBS-injected mice as a control group for each cohort. Evaluating the lacrimal glands of the preventative group indicated that metformin improved the focal scores drastically in both males and females. In contrast, only females showed improvement when treated with CHPE (Fig. 2A). In the therapeutic group, both sexes responded to either treatment, with females accounting for the significantly lower number of the focal score (Fig. 2B). Analysis of the salivary gland pathology indicated that only male mice receiving CHPE during the preventative phase exhibited a reduced focal score, with no significant changes observed in other treated groups compared to the PBS-controlled mice, whether in the preventative or therapeutic regimen (Fig. 2C, D). In conclusion, CHPE and metformin demonstrated a limited effect on the focal scores of the salivary glands, except for male mice treated preventatively with CHPE. Conversely, both drugs significantly improved the lacrimal gland focal scores in preventative and therapeutic regimens. Mice treated with metformin exhibited a complete resolution of lymphocytic infiltration when treated early, and only female mice showed significant improvement in focal scores under treatment.

Figure 2.

Reduced lymphocytic infiltration in the lacrimal glands, not the salivary glands, by CHPE and metformin. Representative H&E images of the lacrimal glands of preventative (A) and therapeutic (B), as well as salivary glands of preventative (C) and therapeutic (D) mice, where treatment is indicated under each image. The focal scores were determined as described in the methods. Welch's 1-tailed t-tests were performed where *P < 0.05, **P < 0.01, and NS: not significant.

CHPE and metformin modulated the composition of lymphocytic infiltrates in the salivary gland

As discussed, CHPE and metformin had a limited impact on the focal scores of the salivary glands. However, we sought to evaluate whether the drugs may affect the lymphocyte composition, specifically the pathogenic cell populations in the glands: B cells, CD4⁺ T, CD8⁺ T, Th1 (CD4⁺IFN-y⁺), and Th17 (CD4⁺IL-17⁺) cells. CHPE and metformin administration significantly decreased B cells in the preventative and therapeutic regimens (Fig. 3A, F, Fig. S1). There was an increase of CD4⁺ T cells by metformin administration (Fig. 3B, Fig. S1) and no statistically significant changes to CD8⁺ T, Th1, or Th17 cells in the preventative regimen (Fig. 3C, D, E, Fig. S1) nor in the therapeutic regimen (Fig. 3G, H, I, J, Fig. S1). Further analysis focused on the effects of these drugs on male and female mice separately. In the preventative regimen, administration of both drugs led to decreased B cells in males, whereas in females, only metformin showed a trend toward reduction compared to the PBS-treated group (Fig. 3K, Fig. S1). There was no alteration in CD4⁺ T cell counts in males, but an increased number was observed in females treated with both drugs; CHPE alone showed an increase in CD8⁺ T cell population in males (Fig. 3L, M, Fig. S1). Neither drug affected Th1 cells, and only CHPE decreased Th17 cells in females compared to the PBS-treated group (Fig. 3N, O, Fig. S1). In the therapeutic regimen, CHPE and metformin exhibited limited effects on the cell populations of male and female mice. Results showed that both drugs significantly reduced B cell counts in both treatment regimens with a more pronounced effect on male mice and reduced B cells in females only with metformin in the preventative regimen. Neither drug affected Th1 cells, but CHPE reduced Th17 cells in females when administered preventatively.

Figure 3.

Changes in infiltrating immune cells in the salivary glands with treatment by CHPE or metformin. Flow cytometric analysis of B220 B cells (A), (F), (K), and (P); CD4 T cells (B), (G), (L), and (Q); CD8 T cells (C), (H), (M), and (R); Th1 cells (D), (I), (N), and (S); and Th17 cells (E), (J), (O), and (T). The analysis here includes preventative group with combined females and males (A–E): PBS: n = 4, CHPE: n = 4, and metformin: n = 5; Therapeutic group (F–J): PBS: n = 4, CHPE: n = 5, and metformin: n = 6; Preventative group with separated males and females (K–O): PBS: n = 4, CHPE male: n = 2, CHPE female: n = 2, metformin male: n = 3, and metformin female: n = 2; and therapeutic group (P–T): PBS: n = 4, CHPE male: n = 4, CHPE female: n = 1, metformin male: n = 4, and metformin female: n = 2. Welch's one-tailed t-tests were performed where *P < 0.05, **P < 0.01, on groups where n ≥ 3.

Reduction of anti-nuclear antibodies in CHPE- and metformin-treated mice

SjD pathogenesis is characterized by high circulating serum anti-nuclear antibody (ANA) levels. As presented in Table 1, the PBS controls displayed a 100% positive staining pattern in the preventative regimen. In CHPE or metformin treatment, the staining shifted to 50% and 33% negative, respectively, wherein the majority (75%) of the females changed to a normal phenotype with CHPE treatment. However, this sexually dimorphic response was not noted in the metformin group, as both the male and female groups were equal at 33% positive. In the therapeutic regimen, the PBS controls were 75% positive staining. The therapeutic administration of CHPE and metformin drastically reduced ANA positivity, in which 1 male in the CHPE group was positive, and none of the females were. Males treated with metformin were 50% negative, whereas the females were 75% negative. The data indicate that both drugs reduced the levels of autoantibodies and were more effective in the therapeutic regimen, with CHPE eliciting a more drastic change in the ANA profile of females.

Table 1.

Improvement in autoantibody profile with CHPE and metformin treatment.

	Preventative												Therapeutic
	Total				Males				Females				Total				Males				Females
	Negative		Positive		Negative		Positive		Negative		Positive		Negative		Positive		Negative		Positive		Negative		Positive
	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%	n	%
PBS	0	0	4	100									1	25	3	75
CHPE	3	50	3	50	0	0	2	100	3	75	1	25	6	86	1	14	3	75	1	25	3	100	0	0
Met	2	33	4	67	2	33	1	67	2	33	1	67	5	62.5	3	37.5	2	50	2	50	3	75	1	25

Constituent antinuclear antibody (ANA) staining was reported for preventative and therapeutic regimens of CHPE and Metformin (Met) compared to PBS-treated mice for the groups (Total) as a whole and broken down by sex. Percentage (%) is given as well as the individual values (n).

Recovering of saliva and tear levels in a sex- and age-dependent manner

One of the major clinical signs of SjD is the salivary and lacrimal glands' loss of saliva and tear secretion. As shown in Fig. 4A, the administration of CHPE and metformin in the preventative regimen increased saliva secretion, and the effect of metformin was more pronounced at 9 wk post-treatment. The increase was more drastic in females than males by metformin (Fig. 4C). In the therapeutic regimen, CHPE and metformin treatment improved saliva flow, especially with CHPE at 5 and 9 wk post-treatment (Fig. 4C). Female mice showed the largest increase in saliva flow when treated with CHPE (Fig. 4D). To access the tear flow rates, we collected tears at endpoint to minimize stress on the mice. As presented in Fig. 5A, CHPE, and metformin showed significant improvement in tear flow compared to PBS-treated mice in the preventative regimen. Male and female mice had similar positive effects with both drugs (Fig. 5B). Similarly, tear flow was increased by CHPE and metformin in comparison to PBS-treated mice in the therapeutic regimen (Fig. 5C). When the responses were separated by sex, all sex groups had a significant improvement in tear flow with the most pronounced change exhibited by males treated with metformin (Fig. 5D). These results indicate that metformin used in the preventative regimen showed the most improvement in saliva flow, specifically in females. Both CHPE and metformin improved saliva flow in the therapeutic regimen with a more pronounced effect by females treated with CHPE. Lastly, both drugs effectively improved lacrimal gland function in the preventative and therapeutic regimens.

Figure 4.

Improvement of salivary gland secretion with CHPE or metformin. Saliva was collected as described in the materials and methods section. Saliva flow rate (SFR) indicated in both preventative (A) and therapeutic (B) groups in combined females and males, where (A) Preventative group (PBS: n = 4, CHPE, n = 6, and metformin, n = 6). (B) Therapeutic group (PBS n = 5, CHPE n = 7, and metformin n = 8). SFR is also displayed, showing the difference between CHPE and metformin by sex in the preventative (C) and therapeutic (D) groups. (C) Preventative group (PBS: n = 4, CHPE male: n = 2, CHPE female: n = 4, and metformin male: n = 3, metformin female: n = 3). (D) Therapeutic group (PBS: n = 5, CHPE male: n = 4, CHPE female: n = 3, metformin male: n = 4, and metformin female: n = 4). Two-way ANOVA was performed where *P < 0.05, **P < 0.01 and ***P < 0.001.

Figure 5.

Improvement of lacrimal gland secretion with CHPE or metformin. Normalized tear flow rates calculated as the TFR against baseline per mouse are given in the preventative with combined females and males (A) and separated males and females (B). Normalized TFR is presented in the therapeutic regimen with combined females and males (C) and separated males and females (D). Welch's 1-tailed t-tests were performed where *P < 0.05, **P < 0.01, and ***P < 0.0001.

Improvement of the SjD phenotypes was mediated in an antigen-specific manner

The data demonstrate that CHPE and metformin improved the SjD signs in the animal model when treated preventatively and therapeutically. These drugs were determined via in-silico to have a high affinity for MHC class II I-A^b and were shown to downregulate T cell response (Fig. 1). Furthermore, mice treated with the drugs exhibited improved signs of SjD with changes in infiltrating immune cells in the glands. We sought to determine whether these immunological and clinical effects were attributed to the specific antigens associated with SjD. To address this issue, we first develop a TCR transductant that expresses a unique TCR to SjD. We employed scRNA-seq using sorted salivary gland immune cells from normal B6 and B6 to obtain the unique TCR. NOD-Aec1Aec2 mice were performed with V(D)J enrichment. Analyzing TCR repertoires, we identified two unique TCRs with paired alpha/beta chains only present in B6.NOD-Aec1/2 mice and shared between male and female mice (Fig. 6A). Using TCRmodel docking algorithm, we found that Vα14-3/Vβ13-1 exhibited high affinity for SjD-associated autoantigens, Ro52, Ro60, and La (Fig. 6B). To further confirm the antigen-specificity of this TCR, we generated the TCR transductant by expressing the paired Vα14-3/Vβ13-1 chains using lentivirus (Fig. 6C). As presented in Fig. 6D, Ro52, Ro60, and La peptides induced the proliferation of Jurkat cells with TCR, whereas the negative control peptide had no effect on cell proliferation, indicating that the receptor recognized and responded to the autoantigens. To evaluate the inhibitory effect of CHPE and metformin, we stimulated the TCR transductant with Ro52, Ro60, and La peptides and subsequently subjected the cells to CHPE or metformin. Both drugs exhibited significant inhibition on the TCR transductant. The result demonstrates that CHPE and metformin inhibit T cell response in an antigen-specific manner.

Figure 6.

MHC blockade by small molecules prevents autoantigen presentation to the SjD pathogenic TCR. (A) Comparison of TCR clones in each experimental group of mice. The proportion of the single top expanded clone in each mouse group is shown as a stacked bar graph, with clones shared between SjD-susceptible (SjD^s) female and male mice, which are not shared with control B6 mice in both sexes, highlighted in red and green. The CDR3 sequences are listed. (B) Theoretical peptide presentation of SjD peptides to unique SjD TCR. TCR-pMHC modeling was performed on TCRmodel, and the image was generated with docking in RCSB PDB Mol* 3D Viewer. (C) The MHC is displayed in green, the TCRalpha in yellow, the TCRbeta in orange, and the peptide in pink. (D) Reduction of responsive pathogenic TCR to peptides of common SjD autoantigens by flow cytometry. The normalized count is reported where the cell counts of GFP⁺ lentivirus-transduced Jurkat cells plated with mitomycin-treated APCs and indicated peptides above the background control of lentivirus with Jurkat and mitomycin-treated APCs but without the presence of peptides is given, where pep indicates negative control peptide. Here, black indicates no treatment, and red indicates CHPE and metformin. The statistical differences between no treatment and CHPE or metformin were determined using a 1-way ANOVA with an uncorrected Fisher's LSD test at all threshold events. Error bars indicate 95% class interval confidence, ****P < 0.0001. The experiment was performed in triplicate.

Discussion

The study underscored applying a structural-based approach to selectively target major histocompatibility complex molecules, offering an immunosuppressive strategy for treating SjD. Previous studies have explored the role of small molecules in inhibiting MHC class II antigen presentation. This study employed an algorithm to identify small molecules predicted to interact with the peptide-binding groove of MHC class II I-Ab. The study examined the effects of metformin and CHPE, as I-Ab binders, on alleviating the clinical signs in a mouse model of SjD. The study revealed that both drugs improved the lacrimal gland focal scores in preventative and therapeutic regimens, reduced the levels of autoantibodies, and were more effective in the therapeutic regimen. Females responded better with CHPE in changing their ANA profile. Metformin and CHPE normalized saliva and tear secretory function. Using a TCR transductant, exposure to both of these small molecules led to autoantigen-specific inhibition of T cell response in the mice. The study demonstrated that a structure-based approach can identify critical structural features at the peptide/MHC interface and discover molecules that modulate TCR signaling in response to specific peptides. This methodology, combining structure-guided in silico screening with targeted intervention at specific MHC pockets, offered significant potential as a preventative and therapeutic measure for modulating T cell response in SjD.

CD4⁺ T cells are causative agents of gland inflammation and B cell activation, specifically Th17 cells, which play a prominent role in exocrine gland infiltration and germinal center activation.^30–34 Once within the affected glands, CD4⁺ T cells become activated upon encountering their specific antigens and possibly differentiate into effector T cell subsets, which release key cytokines, such as IFN-γ, IL-4, and IL-17^.35 CD8⁺ T cells were identified as a significantly lower proportion of the T cell population compared to CD4⁺ T cells in affected glands. The frequency of CD8⁺ T cells can vary among patients and is influenced by several factors, including disease severity, stage, and individual immune response variability. Their function in SjD has not been thoroughly investigated; however, recent data suggest their important role in the pathogenesis (70), especially when considered coupled with blood granzyme K⁺ (GZMK⁺) and/or CXC chemokine receptors (CXCRs). Patients were found to have CD8⁺ T cells with a tissue-resident memory phenotype, while in mice, infiltrating CD8⁺ T cells showed a significant elevation of IFN-γ production in the salivary and lacrimal glands. Notably, the depletion of CD8⁺ T cells fully protected mice against the pathologic manifestations of SjD, even after the onset of the disease (71). Therefore, targeting T cells or T cell co-stimulation provides a feasible therapy. Abatacept is an IgG-Fc and cytotoxic T lymphocyte antigen 4 (CTLA-4) fusion that modulates CD28-mediated T cell co-stimulation between antigen-presenting cells and T cells. Early and active primary SjD patients showed decreased ESSDAI, ESSPRI, RF, and IgG levels. However, salivary and lacrimal gland function remained unchanged, with a reduction in germinal centers per mm² of glandular tissue but not the focus score.³⁶

Similarly, we demonstrated that local delivery of a recombinant adeno-associated virus vector serotype 2 (AAV2) expressing a chimera of mouse CTLA-4 fused with a human immunoglobulin decreased sialadenitis and improved gland function in the SjD mouse model.³⁷ Furthermore, Iscalimab is a monoclonal anti-CD40 antibody in which CD40 is expressed on B-cells and antigen-presenting cells, and it has been shown to improve ESSDAI score in patients compared to placebo.³⁸ Metformin is a widely used drug that causes a reduction in hepatic gluconeogenesis via activation of AMP-activated protein kinase (AMPK)-dependent and AMPK-independent mechanisms, which is prescribed as an anti-diabetic drug. It is currently being tested for efficacy in several diseases due to the wide-reaching effects of AMPK signaling. Recently, Kim et al. tested metformin as a treatment for SjD in NOD/ShiLtJ mice,²¹ which have a different MHC class II (IA^g7) under the premise that AMPK activation would suppress the mTOR pathway. In turn, this will suppress signal transducer and activator of transcription 3 (STAT3) production via phosphorylation of runt-related transcription factor 1 (RUNX1), a negative regulator of suppressor of cytokine signaling 3 (SOCS3), suppressing STAT3^.39 The study found that metformin reduced infiltration of the salivary glands and improved saliva flow while reducing factors canonically associated with pathogenic Th17 cells and reducing splenic germinal center B cells. CHPE is known as a compound in synthesizing hydroxypyrimidine derivatives whose biological function has yet to be characterized. We found that both drugs had differential immunological and clinical impacts at different treatment regimens. Both drugs were shown to lower the gland infiltrating B cells at the preventative and therapeutic regimens, whereas only Th17 cells were reduced in the therapeutic regimens. B cells and Th17 cells are critical in the development of SjD. Specifically, we have learned that in mouse models, different populations of pathogenic cells are involved in different stages of the disease. Th1 and Th17 cells are necessary for the initial inflammatory stage, during which neutrophils and monocytes are recruited to the glands. B cells are critical during the effector stage when Th2 cells and B cells are activated for a humoral response. Lastly, Th17 cells are required for gland destruction at the clinical stage. Therefore, modulating the changes in these populations at different stages of the disease by metformin and CHPE had differential clinical impacts, as presented in the study.

SjD exhibits the highest sexual dimorphic autoimmune disease with a 10 to 20 female-to-male ratio. Sex differences in autoimmunity are a result of either genetic or hormonal factors. A recent study⁴⁰ showed that mice with XX sex chromosome complement as compared with XY^- exhibited upregulated X-linked CD40L and activation markers, with higher levels of MMPs, TGFβ, IL13, splenic follicular B cells, and peritoneal marginal zone B cells. The mice showed accelerated spontaneous lupus. Hormonal signaling has a complex and distinctive role in immune response. Mice with estrogen receptor (ER)α deficiency in T cells failed to develop colitis due to reduced Th1 and Th17 cell responses and increased Treg differentiation.⁴¹ Androgen is considered to be protective in certain autoimmune diseases by enhancing Treg development via increasing Foxp3 expression.⁴² Androgen can shift the pathogenic to tolerogenic response (Th17 to Th2), as evidenced by the decreasing severity of experimental autoimmune encephalomyelitis.⁴³ Regarding drug efficacy, males and females may differ in specific drug pharmacokinetics and pharmacodynamics. As demonstrated in the study, male mice responded better with both drugs by decreasing salivary gland B cells when treated preventatively. Females showed the most efficacy in reducing Th1 and Th17, primarily with CHPE preventatively. In the therapeutic regimen, males decreased Th1 with CHPE, unlike females with metformin. Clinically, female mice showed greater improvement in the lacrimal focus score with both drugs, whereas only males showed lower focal scores in the salivary gland with CHPE when treated therapeutically. Lower B, Th1, and Th17 cells resulted in the most improvement of lacrimal and salivary secretory function by females treated with both drugs. The data suggested a distinctive efficacy of CHPE and metformin in both immune response and clinical outcomes of SjD. The primary limitation of this study was the small sample size when the animals were divided by sex. Although the immunological and clinical impacts of both drugs in males and females showed promise, future research with larger numbers of both sexes is necessary to substantiate the biological and statistical significance of these findings.

Inhibiting auto-antigen presentation in the MHC binding cleft of the self-antigen would prevent symptoms and, as a result, overt disease in SjD by preventing the autoimmune cascade, which reduces inflammatory T cell responses to autoantigens. The immediate result shifts the development of pathogenic T cells. This, in turn, prevents activation of B cells as helper T cells can assist in B cell function. GC-like structures and epithelial inflammation are observed in the exocrine glands of patients with SjD.^10,44 These GC-like structures in target organs are related to the production of autoantibodies and loss of glandular function.⁴⁵ As seen here, treatment with small molecules effectively impairs the helper T/B cell activation in an antigen-specific manner, as evidenced by the diminished infiltrating B cell populations and reduced anti-nuclear antibody profiles. The study supports the use of in-silico drug screening to target specific MHCs, which could provide a reliable therapy for SjD patients in the near future.

Author contributions

B.I. and C.N. conceived the study and designed the experiments. A.V., S.G., and P.G. performed the in vitro and in vivo experiments. Y.S. performed the TCR single-cell analysis. D.L. and D.O. conducted the in-silico screening and analysis. All authors reviewed and approved the manuscript.

Supplementary material

Supplementary material is available at ImmunoHorizons online.

Funding

This research was funded by the National Institutes of Health (NIH) and the National Institute of Dental and Craniofacial Research (NIDCR) (DE028544, DE028544-02S1, PI-Nguyen).

Conflicts of interest

All authors have no competing interests in the subject of the study.

Data availability

The RNA-seq data have been deposited in NCBI’s Gene Expression Omnibus (GEO) and are accessible through GEO Series accession number GSE290595.