The Efficacy of Mesenchymal Stem Cell-Based Therapy for Sjögren’s Syndrome: A Systematic Review and Meta-Analysis of Preclinical Studies

Abstract

Background

Sjögren’s syndrome (SjS) is a chronic autoimmune disease primarily characterized by xerostomia, often accompanied by xerophthalmia, cutaneous dryness, arthralgia, and stiffness. Mesenchymal stem cell (MSC)-based therapies have shown promising immunomodulatory potential in autoimmune diseases, yet their efficacy in SjS remains uncertain. Before advancing to clinical application, a systematic evaluation of preclinical evidence is essential to clarify their therapeutic impact and experimental consistency.

Methods

This systematic review and meta-analysis were registered in PROSPERO (CRD42023471348) and conducted in accordance with PRISMA guidelines. Comprehensive searches were performed in PubMed, Embase, Scopus, Cochrane Library, and Web of Science up to 1 May 2025. Random-effects meta-analyses were conducted to assess changes in salivary flow rate (SFR), inflammatory infiltration, and cytokine expression.

Results

A total of 25 studies met the inclusion criteria. Pooled analyses demonstrated that MSCs, MSC-derived exosomes, and MSC-conditioned medium significantly improved stimulated SFR (standardized mean difference = 3.19, 95% confidence interval 2.50-3.88, P < .001) and reduced inflammatory infiltration (standardized mean difference = −2.04, 95% confidence interval −2.66 to −1.43, P < .001). MSC-based therapies decreased serum levels of proinflammatory cytokines (interleukin-6 [IL-6] and interferon-gamma) and increased IL-10. Exploration of heterogeneity indicated that MSC type and dosage influenced SFR outcomes, MSC intervention type affected inflammatory infiltration and serum IL-6 levels, and MSC dosage influenced serum IL-10 levels.

Conclusion

MSC-based therapies show potential for treating SjS in animal models, improving salivary secretion, reducing inflammation, and modulating immune cytokines. Substantial heterogeneity across studies highlights the need for standardized protocols regarding MSC source, dosage, and treatment duration before clinical translation.

Graphical abstract

MSCs derived from multiple tissue sources and their derivatives, including exosomes and conditioned medium, were administered to a Sjögren’s syndrome mouse model. MSC-based interventions significantly improved salivary flow rate (SFR), alleviated salivary gland inflammatory infiltration, and modulated systemic cytokine levels.

Introduction

Sjögren’s syndrome (SjS) is a systemic autoimmune disorder that primarily affects the exocrine glands, especially the salivary and lacrimal glands, leading to severe dryness of the mouth and eyes.¹ The disease is thought to result from a combination of genetic, hormonal, infectious, and environmental factors. Its hallmark feature is immune-mediated destruction of glandular epithelial cells, largely driven by autoreactive B and T lymphocytes targeting antigens such as Ro/SSA and La/SSB.² SjS is classified as either primary, occurring independently, or secondary, when associated with other connective tissue diseases such as rheumatoid arthritis or systemic lupus erythematosus.³ Beyond glandular dysfunction, SjS may cause systemic complications including glomerulonephritis, renal tubular acidosis, pulmonary fibrosis, and neurological involvement.4, 5, 6 Current treatments for SjS are mainly symptomatic, focusing on relieving dryness and suppressing immune activity. Common therapies include glucocorticoids, hydroxychloroquine, methotrexate, azathioprine, cyclosporine, and intravenous immunoglobulin.⁷ However, these treatments often provide only limited relief and are associated with significant adverse effects, underscoring the urgent need for more effective and targeted therapies.

In recent years, mesenchymal stem cell (MSC)-based therapies have emerged as a promising alternative for autoimmune diseases due to their immunomodulatory and regenerative properties.^8,9 MSC therapy is a novel therapeutic approach that utilizes the unique properties of stem cells, including self-renewal and differentiation, to regenerate damaged cells and tissues in the human body or replace these cells with new, healthy and fully functional cells by delivering exogenous cells into a patient.¹⁰ Preclinical studies have demonstrated that MSC transplantation can ameliorate SjS-like symptoms in animal models.¹¹ Parallel to whole-cell therapies, MSC-derived exosomes (MSC-Exos) have received increasing interest as a novel cell-free strategy for overcoming limitations associated with direct cell transplantation. These nanosized extracellular vesicles (30-150 nm) carry proteins, lipids, and nucleic acids essential for intercellular communication.^12,13 MSC-Exos exhibit significant immunomodulatory properties,¹⁴ promote tissue regeneration,¹⁵ and targeted delivery capabilities,¹⁶ while their low immunogenicity enhances their therapeutic potential for SjS.¹⁷ Similarly, as a cell-free approach, MSC-conditioned medium contains a complex cocktail of bioactive molecules, including cytokines, growth factors, extracellular vesicles, and microRNAs.^18,19 These components exert multifaceted biological effects, such as modulating inflammatory responses,²⁰ promoting angiogenesis,²¹ inhibiting cellular apoptosis,²² thereby restoring salivary gland function in SjS.

A systematic review by Chihaby et al²³ highlighted the beneficial effects of MSC transplantation on sicca symptoms in SjS but was limited by its database, which included studies only until 21 July 2021, and lacked a meta-analysis. Another meta-analysis conducted by Joachim Hansen compared the alterations in unstimulated salivary flow rate (SFR) following treatment with adipose-derived or bone marrow-derived MSC in radiotherapy and SjS patients.²⁴ However, it excluded other types of MSC. Given the unanswered questions regarding the optimal MSC therapy for SjS, analysing the overall therapeutic effects of MSC in SjS animal models under different experimental conditions is essential for future clinical translation. Moreover, despite promising preclinical results, substantial heterogeneity exists among studies concerning MSC transplantation protocols, and no consensus has been reached on which interventional parameters (eg, MSC source, route of administration, dose, and follow-up time) offer the greatest therapeutic benefit.

To address these gaps, we conducted a comprehensive meta-analysis of preclinical studies to: (1) quantify their effects on salivary gland function and inflammation, (2) explore interventional characteristics associated with enhanced efficacy, and (3) provide evidence-based guidance for the future clinical translation of MSC-based therapies in SjS.

Methods

Guidelines and protocol registration

This review was registered in PROSPERO (CRD42023471348) and conducted according to the registered protocol without major deviations. As some databases did not yield the required literature, the sources for literature retrieval were adjusted from the originally planned PubMed, Embase, Web of Science, Cochrane Library, and ClinicalTrials.gov to Embase, Web of Science, Scopus, Cochrane Library, and PubMed databases. The study followed the Preferred Reporting Items for Systematic Reviews and Meta-Analyses (PRISMA 2020) guidelines and flow diagram.²⁵ Notably, this review focused exclusively on the in vivo components of the included studies.

Search strategy

A comprehensive literature search, limited to English-language studies, was conducted across Embase, Web of Science, Scopus, Cochrane Library, and PubMed databases up to 1 May 2025. The search strategy employed the following terms: ('mesenchymal stem cell' OR 'mesenchymal stem cells' OR 'stem cell, mesenchymal' OR 'stem cells, mesenchymal') AND ('Sjögren syndrome' OR 'Sjögren’s syndrome' OR 'syndrome, Sjögren' OR 'sicca syndrome' OR 'syndrome, sicca of primary Sjögren syndrome' OR 'Sjögren’s syndrome').

Eligibility criteria

The inclusion criteria were as follows: (1) the study must be an animal experiment conducted in mice, (2) the study must focus on SjS, (3) the intervention must involve MSC, MSC-derived extracellular vesicles (exosomes), or MSC-associated conditions (intravenous or intraperitoneal injection of MSC, exosomes, or combinations such as MSC with exosomes); cotreatment studies (combining stem cells and exosomes) were also eligible, provided they were interventional in vivo studies with a comparator control, and (4) the primary outcome must assess the efficacy of the intervention.

Exclusion criteria were as follows: (1) reviews, case reports, letters, clinical studies, systematic reviews, and meta-analyses, (2) studies in which MSC, MSC-Exos, or MSC-associated conditions were not used in animal models of SjS or were conducted in humans, (3) studies in which the therapeutic regimen included other agents with unknown effects, and (4) studies that were not published in peer-reviewed journals.

Study selection

Study selection was processed through two stages, with two reviewers (X.Y. Yang and Y.Q. Yu) independently conducting screening. The first stage involved reviewing titles and abstracts, while the second stage entailed assessing the full text of articles based on predefined exclusion and inclusion criteria. Disagreements between reviewers were resolved through discussion. An animal filter was applied to identify preclinical studies, and grey literature, preprints, and conference abstracts were not included. Trial registries were not searched. Reference lists of included articles were manually screened to identify any additional eligible studies, and the search was rerun prior to final analysis to ensure completeness. The PRISMA 2020 checklist is available in Supplementary Data 1. After duplicate removal, two reviewers independently screened all records. Irrelevant studies were excluded, and potentially eligible studies were retrieved for full-text assessment. Final inclusion was determined by consensus, with disagreements resolved through discussion or adjudication by a third investigator.

Data collection

Data extraction was performed independently by two reviewer pairs (X.Y. Yang and Y.Q. Yu; C. Shi and Z.Y. Wu). Data were gathered from various sources, including texts, tables, figures, Supplementary Materials, and referenced methods. Two investigators independently extracted and documented the data listed below from the eligible studies: general information (first author, publication year); experimental methods (number of mice per group for individual comparisons; type of mice; sources and types of MSC; MSC dose; MSC delivery route; follow-up duration; study endpoints; sialogogue; sialogogue dosage; and stimulation time).

Data on the mean and standard deviation (SD) for each parameter in the MSC-treatment group and control group were extracted independently by both investigators. If data were presented only graphically, the mean and SD were estimated using a ‘digital ruler’ (Graph Digitizer 2.26; GetData; available at: https://getdata-graph-digitizer.com/index_php.html). In cases where SD was not reported, it was calculated by multiplying the standard error (SE) by the square root of the group size. Studies were excluded from the meta-analysis if sample size and treatment dose information were not provided. Any disagreements between the two investigators were resolved through mutual review of the data in the original publications.

Outcome measures

The efficacy of MSC treatment was evaluated based on the following outcome measures: SFR; infiltrating area of lymphocyte (% of total area, FOCUS score, Histological score, area ratio of lymphocyte infiltration, number of lymphocytic foci); interleukin-6 (IL-6); IL-10 and interferon-gamma (IFN-γ). Any disagreements between the investigators were resolved through discussion with a third reviewer.

Quality and risk of bias assessment

Risk of bias was evaluated by two independent reviewers (X.Y. Yang and Y.Q. Yu) using the SYRCLE (Systematic Review Centre for Laboratory Animal Experimentation) risk of bias tool.²⁶ Disagreements between reviewers were resolved through discussion. The SYRCLE tool features 10 different parameters: (1) sequence generation (selection bias); (2) baseline characteristics (selection bias); (3) allocation concealment (selection bias); (4) random housing (performance bias); (5) blinding (performance bias); (6) random outcome assessment (detection bias); (7) blinding (detection bias); (8) incomplete outcome data (attrition bias); (9) selective outcome reporting (reporting bias); (10) other sources of bias (bias). All disagreements were solved by the authors (C. Shi and J.H. Hu). Each item was answered ‘Yes’, ‘Unclear’, or ‘No’, representing low, unclear, or high-risk, respectively.

Statistical analysis

A meta-analysis was performed to compare MSC therapy vs control for three outcomes: SFR, inflammatory infiltration, and inflammatory cytokines. Analyses were conducted using RevMan 5.3 and STATA 12.0. Continuous outcomes were pooled using the inverse-variance method under a random-effects model, with results expressed as standardized mean differences (SMDs) and 95% confidence intervals (CIs) to account for expected heterogeneity in measurement techniques and protocols. Statistical heterogeneity was assessed using the I² statistic, with values above 75% indicating substantial heterogeneity.²⁷ When I² exceeded 50%, a 95% prediction interval (PI) was calculated to illustrate the expected range of effects in future studies. The PI was computed as the pooled effect estimate ± t_{k − 2, 0.975} × √(SE² + τ²), incorporating both within-study variance and between-study heterogeneity (τ²), as recommended by PRISMA. Exploratory meta-regression was performed to examine potential sources of heterogeneity, including MSC type, dose, and follow-up duration; these analyses were considered hypothesis-generating due to the limited number of studies. A leave-one-out sensitivity analysis was conducted to evaluate the influence of individual studies on the pooled results. Publication bias was assessed using Egger’s and Begg’s tests in STATA 12.0. Additionally, trim-and-fill analysis was conducted to estimate the number of missing studies and recalculate the effect estimate based on assumed and unobserved studies that would rectify the asymmetry of the funnel plot.²⁸ In cases where publication bias was detected, we employed the trim-and-fill method to make necessary adjustments. A two-tailed P value <.05 was considered statistically significant.

Results

Search results

A comprehensive search across multiple databases identified preclinical studies evaluating the efficacy of MSC-based therapies for SjS in murine models. The study selection process is outlined in Figure 1. An overview of the key characteristics of the included studies is provided in Table, while a more detailed description of the experimental conditions and outcome measures are presented in Supplementary Data 2. Detailed search strategies for each database are presented in Supplementary Data 3.

Fig. 1.

PRISMA flow diagram of study selection.

Table.

Summary of included preclinical studies evaluating MSC-based therapies.

Study	Y	Follow-up	MSC type	Dose	Endpoints
Khalili29	2010	52 wk	BM-MSC	1 × 107 cells	SFR; inflammatory infiltration score
Xu43	2012	18 wk	BM-MSC	1 × 105 cells	SFR; inflammatory infiltration score; TGF-β; IFN-γ; IL-10; IL-6
Khalili52	2012	45 wk	BM-MSC	1 × 107 cells	SFR; TNF-α
Huang44	2016	3 wk	UC-MSC	5 × 105 cells	IL-6; TNF-α; IFN-γ
Aluri45	2017	4 wk	BM-MSC	1 × 106 cells	inflammatory infiltration score
Ruan30	2017	2 wk	BM-MSC	5 × 105 cells	SFR; inflammatory infiltration score; IL-6; IFN-γ; IL-10; TGF-β
Hai46	2018	3 wk	iPSC-MSC (Exos) or BM-MSC (Exos)	1 × 106 cells or 30 μg Exos	inflammatory infiltration score
Shi31	2018	5 wk	UC-MSC	1 × 106 cells	SFR
Yao50	2019	1 wk	UC-MSC	1 × 106 cells	SFR; inflammatory infiltration score; TGF-β
Abughanam33	2019	16 wk	BM-MSC or BM-MSC-Exos	100 µL CM (derived from 2 × 106 cells)	SFR; inflammatory infiltration score; IL-10
Hu40	2020	1 wk	BM-MSC	5 × 105 cells	SFR; inflammatory infiltration score; TNF-α
Qi34	2020	4 wk	UC-MSC	1 × 106 cells	SFR
Kim47	2021	2 wk	iPSC Exos (P5 or P15)	Exos derived from 2.5 × 107 cells	inflammatory infiltration score; TGF-β
Li36	2021	6 wk	LG-MSCs isolated from patients	2 × 105 cells	SFR; inflammatory infiltration score; TGF-β; IL-6; TNF-α
Ogata48	2021	2 wk	DPSCs-CM or BMMSC-CM	500 μL CM (derived from 5 × 105 cells)	SFR; inflammatory infiltration score; IFN-γ; IL-10; IL-6
Liu37	2021	12 wk	UC-MSC	1 × 106 cells	SFR; inflammatory infiltration score; IL-6; IL-10; TNF-α
Matsumura51	2021	2 wk	DPSCs-CM or BMMSC-CM	500 μL CM (derived from 5 × 105 cells)	SFR; inflammatory infiltration score; TGF-β
Xing35	2022	16 wk	LG-MSC or LG-MSC-Exos	1 × 106 cells or 50 μg Exos	SFR; inflammatory infiltration score
Sun53	2022	2 wk	UC-MSC	5 × 105 cells	SFR; inflammatory infiltration score
Yang38	2022	4 wk	SHED or BM-MSC	2 × 106 cells	SFR; IL-10
Cong39	2022	12 wk	UC-MSC	5 × 105 cells	SFR
Zhao49	2023	2 wk	iPSC Exos	Exos derived from 1.5 × 1010 cells	inflammatory infiltration score
Hu32	2023	10 wk	DPSCs-Exos	25 mg/kg	SFR
Chu41	2024	7 wk	SHED-Exos	50 μg	SFR; inflammatory infiltration score
Xie42	2025	8 wk	LG-MSC-Exos	100 μg	SFR; inflammatory infiltration score; IL-6; IL-10; TGF-β

Different dosing units were used depending on the type of intervention.

BM-MSC, bone marrow-derived mesenchymal stem cells; CM, conditioned medium; DPSC, dental pulp stem cells; EVs, extracellular vesicles; Exos, exosomes; IFN-γ, interferon gamma; IL-10, interleukin-10; IL-6, interleukin-6; iPSC, induced pluripotent stem cells; LG-MSC, labial gland-derived mesenchymal stem cells; SFR, salivary flow rate; SHED, stem cells from human exfoliated deciduous teeth; TGF-β, transforming growth factor beta; TNF-α, tumour necrosis factor alpha; UC-MSC, umbilical cord-derived mesenchymal stem cells.

Study characteristics

The included 25 experimental studies were conducted in nonobese diabetic mice.29, 30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43, 44, 45, 46, 47, 48, 49, 50, 51, 52, 53 The publication years of the included studies spanned from 2010 to 2025. The sample size per study ranged from 10 to 30 animals, with follow-up durations varying from 1 to 52 weeks. The MSCs utilized were derived from diverse sources, including bone marrow MSC (BM-MSC), umbilical cord MSC, dental pulp stem cell (DPSC), stem cells from human exfoliated deciduous teeth, labial glands MSC (LG-MSC/LGMSC), and induced pluripotent stem cells (iPSC-MSC). The administered cell doses ranged from 1 × 10⁵ to 1 × 10⁷ cells per animal. Some studies employed stem cell-derived conditioned media or exosomes as the intervention. The outcome measures assessed were SFR, histopathological inflammatory infiltration scores, and the levels of various inflammatory cytokines (IL-6, IL-10, IFN-γ, TNF-α, TGF-β). The route of administration for all interventions was systemic, predominantly via intravenous or intraperitoneal injection.

Quality assessment

The risk of bias in the included animal studies was evaluated using SYRCLE’s Risk of Bias Tool, with the comprehensive findings summarized in Figure 2. Overall, the studies demonstrated acceptable methodological quality, with most domains rated as ‘low risk’ or ‘unclear risk’.

Fig. 2.

Risk of bias assessment for included studies (SYRCLE’s ROB Tool). (A) Risk of bias summary of the included preclinical studies based on SYRCLE’s Risk of Bias tool. (B) Aggregate risk of bias graph for all included preclinical studies based on SYRCLE’s Risk of Bias tool. The assessed domains include: random sequence generation, baseline characteristics, allocation concealment, random housing, blinding, random outcome assessment, blinding of outcome assessment, incomplete outcome data, selective reporting, and other biases. The judgements for each domain are categorized as low (green), unclear (yellow), or high (red) risk of bias.

Eighteen studies (72.0%) mentioned random assignment of animals to experimental groups, but only five (20.0%) provided sufficient details on their randomization methods, leading to a high proportion of unclear risk in this domain. Twenty-three studies (92.0%) adequately reported baseline characteristics, suggesting a low risk of selection bias in this aspect. However, allocation concealment was unclear in all studies (100%), and no studies confirmed random housing of animals, which are common limitations in preclinical research. Blinding of personnel was poorly implemented: 7 studies (28.0%) were judged as high-risk of performance bias, while the remaining 18 (72.0%) provided insufficient details to assess this domain. Blinding of outcome assessors was adequately reported in six studies (24.0%), reducing the risk of detection bias in these cases. For selective reporting, 18 studies (72.0%) showed consistency between their methods and reported outcomes, though none had a publicly available study protocol for verification. Attrition bias was judged as low risk in 12 studies (48.0%), while the remaining 13 (52.0%) were rated as unclear risk due to insufficient information on missing data.

Assessment of SFR

Sixteen studies30, 31, 32, 33, 34, 35, 36, 37, 38, 39, 40, 41, 42, 43^,48,51 were included in the meta-analysis evaluating the effect of MSC therapy on SFR. The methods of collecting saliva (sialogogue, dosage, and stimulation time) included in the article are provided in Supplementary Data 2. In order to eliminate potential errors caused by different units, we converted the SFR units in each included study to a uniform unit of μL/10 minutes. The results showed that the MSC treatment group had significantly higher SFR levels compared to the non-MSC treatment group in SjS (SFR: SMD = 3.19, 95% CI 2.50-3.88; P < .001; Figure 3). The heterogeneity index was high (I² = 72%), indicating there may be substantial variability in MSC types, follow-up duration, MSC dose, sialogogue, dosage, and stimulation time among the studies. The 95% PI ranged from 1.38 to 5.00 (Supplementary Data 4), indicating that the effect of MSC therapy on SFR in a future, similar study is predicted to be positive, though its magnitude may vary considerably. Despite beneficial point estimates for MSC, exosomes, and conditioned media, no significant subgroup differences were found (P = .36).

Fig. 3.

Forest plot of SFR. (A) MSCs, (B) MSC-derived exosomes, and (C) MSC-conditioned medium were analysed separately. Each square represents the effect size of an individual study, with the size proportional to its weight, and the diamond indicates the pooled SMD. I² value represents the statistical heterogeneity among studies. BM-MSC, bone marrow-derived mesenchymal stem cells; CM, conditioned medium; Exos, exosomes; LG-MSC, labial gland-derived mesenchymal stem cells; SFR, salivary flow rate; SHED, stem cells from human exfoliated deciduous teeth; UC-MSC, umbilical cord-derived mesenchymal stem cells.

Assessment of inflammatory infiltration

In addition to SFR, we also pooled data from 18 experimental studies^29,30,33,35, 36, 37^,40, 41, 42, 43^,45, 46, 47, 48, 49, 50, 51^,53 assessing inflammatory infiltration, a crucial pathological hallmark of SjS. For the inflammatory infiltration group, subgroup analyses were performed using five distinct assessment units: percentage of total area, FOCUS score, histological score, area ratio of lymphocyte infiltration, and number of lymphocytic foci. Overall, the results indicated that the MSC-treated group showed significantly lower levels of inflammatory infiltration compared to the control group (Inflammatory infiltration: SMD = −2.04, 95% CI −2.66 to −1.43; P < .001; Figure 4), with substantial heterogeneity observed across studies (I² = 74%). The 95% PI (−3.18 to −0.90; Supplementary Data 4) suggests that while the effect size may vary considerably among studies, the direction of the treatment effect is consistently favourable. The heterogeneity index was low in the FOCUS, histological score, and area ratio of lymphocyte infiltration groups (I² = 0%, 25%, and 0%), but high in the % of total area and number of lymphocytic foci groups (I² = 85% and 82%; Figure 4). Notably, the test for subgroup differences was statistically significant (P = .005), suggesting that the magnitude of the treatment effect varied according to the specific histological assessment method employed.

Fig. 4.

Forest plots of inflammatory infiltration. Subgroup analyses were performed for (A) percentage of total area, (B) focus score, (C) histological score, (D) area ratio of lymphocyte infiltration, and (E) number of lymphocytic foci. Each square represents the SMD of an individual study, with the size proportional to study weight, and the diamond represents the pooled SMD. I² value represents the statistical heterogeneity among studies. BM-MSC, bone marrow-derived mesenchymal stem cells; CM, conditioned medium; Exos, exosomes; LG-MSC, labial gland-derived mesenchymal stem cells; SHED, stem cells from human exfoliated deciduous teeth; UC-MSC, umbilical cord-derived mesenchymal stem cells.

Assessment of inflammatory cytokines

Regarding inflammatory cytokines, due to the limited number of included studies, only IFN-γ, IL-10, and IL-6 were analysed.

IFN-γ

Four studies^30,43,44,48 assessed the effect of MSC treatment on IFN-γ levels. Pooled analysis demonstrated a significant reduction in IFN-γ levels following MSC treatment (SMD = −2.20, 95% CI −3.51 to −0.89; P = .001; Figure 5). However, substantial heterogeneity was observed across studies (I² = 74%), indicating considerable between-study variability. Subgroup analysis revealed particularly pronounced effects in salivary gland tissue (SMD = −4.11, 95% CI −6.14 to −2.08; P < .001; Figure 5B) compared to serum measurements (SMD = −0.97, 95% CI −1.66 to −0.27; P = .006; Figure 5A), with significant subgroup differences (P = .004). Furthermore, the 95% PI for the overall effect ranged from −4.09 to −0.31 (Supplementary Data 4), suggesting that despite the observed heterogeneity, future studies are still likely to demonstrate a reduction in IFN-γ levels following MSC treatment.

Fig. 5.

Forest plots of inflammatory cytokine levels. (A and B) Interferon-γ (IFN-γ) levels in serum and salivary gland tissues; (C-E) interleukin-10 (IL-10) levels in serum, salivary gland, and ELISA assays; (F and G) interleukin-6 (IL-6) levels in serum and salivary gland tissues. Each square represents the SMD of an individual study, with its size proportional to study weight, and the diamond denotes the pooled SMD with 95% confidence interval (CI). I² value represents the statistical heterogeneity among studies. BM-MSC, bone marrow-derived mesenchymal stem cells; CM, conditioned medium; Exos, exosomes; LG-MSC, labial gland-derived mesenchymal stem cells; SHED, stem cells from human exfoliated deciduous teeth; UC-MSC, umbilical cord-derived mesenchymal stem cells.

IL-10

Seven studies^{30,33,37,38,42,43,48} were included to evaluate the effects of MSC on IL-10. The overall analysis indicated a significant increase in IL-10 levels (SMD = 2.44, 95% CI 0.90-3.98; P = .002; Figure 5). However, this effect was primarily driven by serum measurements (SMD = 3.37, 95% CI 1.20-5.55; P = .002; Figure 5C), whereas IL-10 levels in salivary gland tissue showed no significant change (SMD = 0.13, 95% CI −2.80 to 3.06; P = .93; Figure 5D).

IL-6

Seven studies^{30,36,37,42,43,44,48} were included to assess the effect of MSC treatment on IL-6 levels. Analysis of IL-6 levels showed a consistent reduction across studies (SMD = −2.25, 95% CI −3.29 to −1.22; P < .001; Figure 5). Both serum (SMD = −1.92, 95% CI −3.18 to −0.65; P = .003; Figure 5F) and salivary gland (SMD = −3.10, 95% CI −5.35 to −0.85; P = .007; Figure 5G) subgroups demonstrated significant improvements, with no statistical difference between measurement sites (P = .37).

Sources of heterogeneity

SFR

To explore potential sources of heterogeneity, subgroup analyses and meta-regression based on the SFR outcome were conducted according to MSC type, dose, and follow-up duration only, as the limited number of studies precluded meaningful analyses of other variables, such as sialogogue use, dosage regimen, and stimulation time.

MSC source

Subgroup analyses were conducted to explore potential sources of heterogeneity in the effect of MSC-based therapies on SFR (Figure 6). When stratified by the source of MSCs (Figure 6A), the pooled analysis demonstrated a statistically significant improvement in SFR across all sources, including BM-MSC (SMD = 2.50, 95% CI 1.56-3.44, P < .001), LGMSC (SMD = 4.78, 95% CI 0.12-9.43, P = .04) and umbilical cord MSC (SMD = 2.27, 95% CI 0.93-3.60, P < .001). However, the test for subgroup differences did not provide strong evidence that MSC source moderated the treatment effect (P = .60).

Fig. 6.

Subgroup analysis. Subgroup analyses were performed according to (A) MSC source for SFR, (B) exosome source for SFR, (C) MSC dose for SFR, (D) conditioned medium source for SFR, (E) MSC intervention type for inflammatory infiltration area percentage, (F) MSC dose for serum IL-10, and (G) MSC intervention type for serum IL-6. The square represents the SMD for each study, with the size proportional to study weight, and the diamond represents the pooled SMD with 95% CI. Horizontal lines indicate 95% CI. The I² statistic reflects heterogeneity among studies. BM-MSC, bone marrow-derived MSCs; CM, conditioned medium; Exos, exosomes; LG-MSC, labial gland-derived MSCs; SFR, salivary flow rate; SHED, stem cells from human exfoliated deciduous teeth; UC-MSC, umbilical cord-derived MSCs.

Further subgroup analyses were conducted according to the source of MSC-derived products. For exosome-based therapies (Figure 6B), exosomes derived from LGMSCs were associated with a statistically significant improvement in SFR (SMD = 6.17, 95% CI 3.71-8.63, P < .001), whereas the pooled effect for DPSC-derived exosomes was not statistically significant (SMD = 4.40, 95% CI −2.79 to 11.58, P = .23). The test for subgroup differences was not significant (P = .65), suggesting that the exosome source did not significantly moderate the treatment effect. In contrast, subgroup analysis of CM (Figure 6D) demonstrated a significantly greater effect for DPSC-derived CM (SMD = 5.29, 95% CI 3.28-7.30, P < .001) compared with BM-MSC-derived CM (SMD = 2.65, 95% CI 1.43-3.88, P < .001). The difference between these subgroups was statistically significant (P = .03), suggesting that the source of conditioned media may represent an important moderating factor.

MSC dosage

Subgroup analysis based on MSC dose (Figure 6C) demonstrated statistically significant improvements in SFR across all dose categories. Specifically, MSC doses <10⁶ were associated with a significant effect (SMD = 3.69, 95% CI 2.41-4.98, P < .001), as were doses ≥10⁶ (SMD = 16.33, 95% CI 9.66-23.00, P < .001). Studies administering an MSC dose of 10⁶ similarly showed a significant pooled effect (SMD = 1.47, 95% CI 0.66-2.27, P < .001). Notably, the test for subgroup differences was statistically significant (P < .001), suggesting that MSC dose may contribute to heterogeneity in treatment effects.

Follow-up duration

When stratified by follow-up duration (Supplementary Data 5), SFR improvement remained statistically significant in both subgroups with follow-up durations of less than 12 weeks (SMD = 3.31, 95% CI 2.03-4.58, P < .001) and 12 weeks or longer (SMD = 2.41, 95% CI 1.32-3.49, P < .001). The test for subgroup differences did not provide strong evidence that follow-up duration moderated the overall treatment effect (P = .29).

Meta-regression

Univariable meta-regression was performed to further assess the potential moderating effects of MSC type, MSC dose, and follow-up duration on the overall SFR effect size (Supplementary Data 5). The results indicated that both MSC type (coefficient = 2.13, 95% CI 0.49-3.76, P = .017) and MSC dose (coefficient = −4.31, 95% CI −7.10 to −1.53, P = .007) significantly moderated the treatment effect, whereas follow-up duration did not show a significant influence (coefficient = 0.54, 95% CI −1.64 to 2.72, P = .588). The joint test for all covariates was statistically significant (P = .026), and the model explained 72.2% of the between-study variance (adjusted R² = 72.17%), suggesting that MSC type and dose account for a substantial portion of the observed heterogeneity.

Inflammatory infiltration

Due to the limited number of studies using the number of lymphocytic foci to assess inflammatory infiltration, we performed a subgroup analysis based on the percentage of total glandular area to explore potential sources of heterogeneity (Figure 6E). Regarding inflammatory infiltration, the study by Aluri et al⁴⁵ was excluded to reduce methodological heterogeneity. The subgroup analysis showed that when inflammatory infiltration was quantified as a percentage of total glandular area, whole MSC therapy significantly reduced inflammatory infiltration compared with MSC-Exos (SMD = −3.02 vs −0.29). Notably, the test for subgroup differences was statistically significant (P < .001), suggesting that the type of intervention may contribute to heterogeneity in inflammatory infiltration meta-analysis.

IL-10

Dose–response subgroup analysis of MSC treatment revealed a significant dose-dependent effect on serum IL-10 levels (Figure 6F). High-dose MSC administration (MSC dose >10⁶) resulted in a pronounced increase in IL-10 expression (SMD = 8.11, 95% CI 5.17-11.05; P < .001), with no detectable heterogeneity (I² = 0%). Similarly, moderate-dose MSC treatment (10⁵ < MSC dose ≤10⁶) significantly elevated serum IL-10 levels (SMD = 1.97, 95% CI 1.13-2.80; P < .001; I² = 0%). In contrast, low-dose MSC administration (MSC dose ≤10⁵) did not produce a statistically significant change in IL-10 levels (SMD = 0.70, 95% CI −0.21 to 1.61; P = .13). The test for subgroup differences was statistically significant (P < .001), indicating that MSC dose was a major contributor to the observed heterogeneity and that higher MSC doses were associated with greater IL-10 upregulation.

IL-6

Subtype analysis based on intervention type revealed distinct response patterns: while MSC administration significantly reduced serum IL-6 levels (SMD = −1.27, 95% CI −2.14 to −0.39; P = .005; Figure 6G) with moderate heterogeneity (I² = 49%), MSC-Exos demonstrated a substantially more pronounced reduction (SMD = −10.02, 95% CI −15.08 to −4.96; P < .001; Figure 6G) with no heterogeneity (I² = 0%). Notably, the test for subgroup differences was highly significant (P < .001), indicating that the type of intervention substantially moderated the treatment effect, with exosomes showing superior efficacy in reducing serum IL-6 levels compared to whole MSC administration.

Sensitivity analysis

A sensitivity analysis was conducted for the primary outcome SFR (MSCs and Exos) to assess the stability of the results by sequentially omitting each study due to significant heterogeneity. As shown in Supplementary Data 5, omission of any single study did not significantly alter the overall effect size, with the estimated values remaining within the 95% CI range. This indicates that no individual study had a disproportionate impact on the overall meta-analysis results, suggesting a high level of stability and reliability of the findings.

Publication bias

Publication bias was assessed using Begg’s and Egger’s tests. Begg’s rank correlation test indicated significant publication bias (z = 4.57, P < .001; Supplementary Data 6A), which was further supported by Egger’s regression test showing a significant intercept (P < .001; Supplementary Data 6A). These results suggested potential small-study effects and asymmetry among the included studies. To further evaluate the impact of publication bias on the pooled estimates, the trim-and-fill method was applied under a random-effects model (Supplementary Data 6D). However, no studies were imputed, and the pooled effect size remained unchanged after adjustment, indicating that the observed publication bias did not materially affect the robustness of the overall meta-analytic results.

Discussion

Summary of findings

In this systematic review and meta-analysis of preclinical animal studies, we comprehensively evaluated the therapeutic efficacy of MSC-based therapies in SjS models. Our findings demonstrate that MSC treatment significantly enhances salivary secretion, modulates inflammatory cytokine profiles, and markedly reduces inflammatory cell infiltration within the salivary glands. Collectively, these results suggest that MSC-based interventions exert both functional and immunomodulatory benefits in experimental SjS, supporting their potential as a promising therapeutic strategy for SjS.

Previous studies have reported that MSC-based therapies can ameliorate xerostomia.⁵⁴ Notably, a prior meta-analysis demonstrated that MSC treatment has significant therapeutic potential in alleviating radiation-induced xerostomia following head and neck radiotherapy.⁵⁵ In contrast, our meta-analysis extends these findings by suggesting that MSC-based interventions may also confer benefits in improving xerostomia in the context of SjS. Beyond the improvement of xerostomia, accumulating evidence indicates that MSC-based therapies exert broad anti-inflammatory and immunomodulatory effects.⁵⁶ A meta-analysis in acute pancreatitis showed that MSC therapy reduced pancreatic histopathology and proinflammatory cytokines, with the cytokine-inhibitory effect correlating with MSCs dose, suggesting a dose-dependent immunoregulatory mechanism.⁵⁷ These findings are consistent with the results of our meta-analysis. Importantly, multiple meta-analyses have demonstrated the therapeutic potential of MSC-based interventions in a range of autoimmune diseases, including multiple sclerosis,⁵⁸ rheumatoid arthritis,⁵⁹ systemic sclerosis,⁶⁰ and systemic lupus erythematosus.⁶¹ However, no meta-analysis has systematically evaluated the efficacy of MSC therapy in SjS. The present study addresses this gap by quantitatively assessing the effects of MSC-based therapies on SFR, inflammatory infiltration, and modulation of inflammatory cytokines in preclinical SjS models. In addition, we performed comprehensive subgroup analyses to explore the potential sources of heterogeneity, including types of MSC-based interventions, MSC sources, administered cell doses, and follow-up duration. Our findings extend the current understanding of MSC-mediated immunotherapy in SjS and provide new insights into their potential role in the treatment of autoimmune diseases.

Notably, subgroup analyses revealed that most predefined variables, including MSC type and follow-up duration, did not significantly influence improvements in SFR. In contrast, MSC dosage emerged as a potential moderator of therapeutic efficacy. Specifically, administration of more than 1 × 10⁶ MSCs was associated with a substantially greater improvement in SFR compared with lower doses (SMD = 16.33 vs 3.69; Figure 6C). However, this high-dose subgroup comprised only two datasets, both derived from the study by Yang,³⁸ indicating that this finding should be interpreted cautiously due to limited sample size and potential study-level bias. Moreover, the MSC source of conditioned media emerged as a significant moderator: DPSC-CM demonstrated superior efficacy compared with bone marrow MSC-conditioned (P = .03; Figure 6D). Subsequent analysis demonstrated that when inflammatory infiltration was quantified as a percentage of the total glandular area, treatment with whole MSCs resulted in a greater reduction in local inflammation than MSC-Exos (SMD = −3.02 vs −0.29; Figure 6E). Interestingly, MSC-Exos showed superior efficacy compared with whole MSC therapy in reducing serum IL-6 levels (SMD = −10.02 vs −1.27; Figure 6G), suggesting that these two therapeutic modalities may exert distinct anti-inflammatory effects at the tissue and systemic levels. These findings are consistent with previous studies indicating that MSCs, due to their intrinsic multipotent differentiation and tissue-homing abilities, are particularly effective for direct local tissue repair and regeneration.⁶² In contrast, MSC-Exos, as nanoscale extracellular vesicles carrying diverse bioactive cargo, have demonstrated potent systemic immunomodulatory and anti-inflammatory effects in preclinical models of autoimmune and inflammatory diseases.^63,64

We observed that MSC-Exos showed notable effects on serum IL-6 levels. Notably, both Xie⁴² and Xing³⁵ utilized exosomes derived from LGMSC, and Xie et al further employed exosomes modified with miRNA let-7f-5p, which may have contributed to their enhanced anti-inflammatory efficacy. Although overall heterogeneity remained high in the SFR subgroup forest plots for both MSC and exosome interventions, a consistent trend was observed, indicating that LGMSC and their exosomes exhibited superior efficacy in improving SFR compared with other MSC sources. Nevertheless, the current evidence base for LGMSC in SjS remains limited, with only three studies included in the present meta-analysis. Future experimental studies are warranted to further clarify whether LGMSC and their exosomes confer distinct and reproducible therapeutic benefits in SjS.

The overall methodological quality of the included preclinical studies was generally suboptimal, with several domains demonstrating a high or unclear risk of bias. In particular, key sources of potential bias were related to performance and detection bias. Most studies did not clearly report the use of blinding procedures or random housing, resulting in predominantly unclear or high-risk ratings in these domains. Similarly, random outcome assessment and blinding of outcome assessors were insufficiently described in many studies, which may have influenced the reliability of outcome measurements. In contrast, certain methodological aspects were relatively well addressed. A high proportion of studies demonstrated a low risk of bias in baseline characteristics, selective outcome reporting, and other sources of bias, suggesting that group comparability at baseline and outcome reporting were generally adequate across studies. These findings indicate that while some fundamental design elements were appropriately considered, critical safeguards against bias during intervention implementation and outcome assessment were frequently lacking. Collectively, these results highlight the need for improved methodological rigour in future preclinical studies investigating MSC therapy for SjS. Greater adherence to established guidelines for animal research, including the implementation and transparent reporting of randomization, blinding, and housing procedures, is essential to enhance internal validity. Addressing these limitations will improve the robustness and translational value of preclinical evidence in this field.

Research gaps

Our study has several limitations. First, regarding study quality, the frequent absence of blinded outcome assessment and the relatively small sample sizes in many studies may compromise the robustness and reproducibility of the findings. Second, the number of included studies was small, and the sample sizes were limited, which may reduce the reliability of our conclusions. Finally, substantial heterogeneity was observed across most pooled outcomes. Several factors may have contributed to these findings: (1) selective reporting of positive results without corresponding negative findings; (2) restriction of included studies to English-language publications, excluding potentially relevant studies published in other languages such as Chinese; (3) the limited number of randomized controlled animal studies; (4) exclusion of relevant studies from which quantitative data could not be extracted; and (5) Considerable variability existed in MSC-related factors (eg, cell source, dose, administration route, and preparation methods) and in salivary flow assessments (eg, stimulant type/concentration, anaesthesia, stimulation duration, and timing of measurement), which may contribute to heterogeneity across studies. These methodological inconsistencies underscore the urgent need for standardized experimental protocols and highlight the necessity for larger, well-designed preclinical studies to more accurately delineate the therapeutic potential of MSCs in SjS.

Clinical translation and implications

Rigorous preclinical validation remains a critical prerequisite before routine clinical application. During the conduct of this systematic review and meta-analysis, we identified a substantial lack of standardization across preclinical studies with respect to MSC source, dosage, delivery route, and treatment regimens. Moreover, the mechanisms by which MSCs, exosomes, and conditioned media exert their therapeutic effects in vivo have not yet been fully elucidated. Consequently, further investigation is required to comprehensively evaluate the safety, efficacy, and optimal dosing strategies of MSC-based therapies prior to the initiation of large-scale clinical trials. Additional challenges that must be addressed include the heterogeneity of MSC sources and the high costs associated with large-scale manufacturing and quality control. Nonetheless, the favourable outcomes observed in animal models provide a strong rationale for advancing MSC-based therapies towards clinical translation.

If supported by well-designed clinical trials, MSC-based therapies may represent a next-generation treatment strategy for SjS. Compared with conventional immunosuppressive therapies, MSC-based interventions have the potential to offer improved safety profiles, targeted immunomodulatory effects, and enhanced clinical feasibility. By modulating immune responses and targeting underlying inflammatory processes, MSC-based therapies could shift future treatment paradigms towards regenerative and immune-regulatory approaches aimed at restoring salivary gland function and alleviating systemic inflammation.⁸ From a clinical perspective, the observed improvements in SFR and reductions in key proinflammatory cytokines in this study point to measurable and clinically relevant benefits. Subgroup analyses further underscore the importance of treatment-related heterogeneity, particularly with respect to MSC dosage and cell type, highlighting the need for protocol standardization and the identification of potential effect modifiers in future research. If these findings are validated in human studies, MSC-based therapies may emerge as a promising therapeutic option, especially for patients with SjS who exhibit inadequate responses or intolerance to conventional immunosuppressive treatments.

Conclusion

In conclusion, the present study supports the potential therapeutic value of MSC-based therapies and their derivatives, including exosomes and conditioned media, for the treatment of SjS in preclinical animal models. These findings underscore the need for further well-designed clinical studies to rigorously evaluate the efficacy of MSC-based interventions in human SjS. The pooled results demonstrate that MSC-based therapies significantly improved SFR, reduced inflammatory infiltration of the salivary glands, and effectively modulated inflammatory cytokine profiles. Specifically, MSC treatment was associated with decreased levels of the proinflammatory cytokines IFN-γ and IL-6, alongside increased levels of the anti-inflammatory cytokine IL-10. Subgroup analyses further suggest that MSC dosage and cell type may contribute substantially to the observed interstudy heterogeneity. Taken together, these findings highlight the importance of standardizing treatment protocols and identifying key effect modifiers in future research. Moreover, they support the continued investigation of MSC-based therapies in clinical settings, with particular emphasis on their broader effects on immune regulation and inflammatory pathways in SjS.

Author contributions

Xinyue Yang, Yaqiong Yu: Conceptualization, project administration, methodology, data curation, formal analysis, investigation, visualization, writing – original draft. Cheng Shi, Bo Fu: Conceptualization, data curation, writing – review and editing. Zhuoyang Wu, Jiahao Hu: Methodology, formal analysis, writing – review and editing; and Aokang Yao, Sheng Luo: Conceptualization, data curation, writing – review and editing.

Funding

This study was supported by the National Natural Science Foundation of China (81900991); China Postdoctoral Science Foundation (2019M651174); Basic Research Program of Liaoning Province in 2022 (Project No. 2022020310-JH2/1013).

Data availability

The datasets used and/or analysed during the current study are available from the corresponding author upon reasonable request.

Conflict of interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this article.

Appendix. Supplementary materials

Supplementary Data 2 Table: Detailed characteristics of the included studies.

Supplementary Data 3 Search strategy.

Supplementary Data 4 Table: Meta-analysis of the efficacy of MSCs in the therapy of Sjögren’s syndrome.

Supplementary Data 5 Subgroup analysis of SFR (follow-up duration), meta-regression analysis, and sensitivity analysis.

Supplementary Data 6 Publication bias assessment.