NINJ1 impairs the anti-inflammatory function of hUC-MSCs with synergistic IFN-γ and TNF-α stimulation

Abstract

Purpose

To investigate the regulatory role of nerve injury-induced protein 1 (NINJ1) in the anti-inflammatory function of human umbilical cord mesenchymal stem cells (hUC-MSCs) co-stimulated by interferon-gamma (IFN-γ) and tumor necrosis factor-alpha (TNF-α).

Methods

hUC-MSCs were expanded in vitro using standard protocols, with stem cell characteristics confirmed by flow cytometry and multilineage differentiation assays. The immunomodulatory properties and cellular activity of cytokine-co-pretreated hUC-MSCs were systematically evaluated via quantitative reverse transcription RT-qPCR, lymphocyte proliferation suppression assays, and Cell Counting Kit-8 viability tests. Transcriptome sequencing, Western blotting and small interfering RNA interference were integrated to analyze the regulatory mechanisms of NINJ1 expression. Functional roles of NINJ1 in pretreated hUC-MSCs were elucidated through gene silencing combined with lactate dehydrogenase release assays, Annexin V/Propidium Iodide apoptosis analysis, macrophage co-culture models, and cytokine Enzyme-Linked Immunosorbent Assay. Therapeutic efficacy was validated in a cecal ligation and puncture-induced septic mouse model: 80 mice were randomly allocated into 4 experimental groups (n=20/group): sham group (laparotomy without cecal ligation); phosphate-buffered saline-treated group (cecal ligation and puncture (CLP) + 0.1 mL phosphate-buffered saline); hUC-MSCs (small interfering RNA (siRNA)-interferon-gamma and tumor necrosis factor-alpha co-stimulation (IT))-treated group (CLP + hUC-MSCs transfected with scrambled siRNA); and hUC-MSCs (siNINJ1-IT)-treated group (CLP + hUC-MSCs with NINJ1-targeting siRNA).

Results

hUC-MSCs demonstrated compliance with International Society for Cellular Therapy criteria, confirming their stem cell identity. IFN-γ/TNF-α co-pretreatment enhanced the immunosuppressive capacity of hUC-MSCs, accompanied by the reduction of cellular viability, while concurrently upregulating pro-inflammatory cytokines such as interleukin-6 and interleukin-1β. This co-stimulation significantly elevated NINJ1 expression in hUC-MSCs, whereas genetic silencing of NINJ1 effectively suppressed pro-inflammatory cytokine production and attenuated damage-associated molecular patterns release through inhibition of programmed plasma membrane rupture. Furthermore, the NINJ1 interference potentiated the ability of cytokine-pretreated hUC-MSCs to suppress LPS-induced pro-inflammatory responses in RAW264.7 macrophages. In cecal ligation and puncture-induced sepsis model, NINJ1-silenced hUC-MSCs exhibited enhanced therapeutic efficacy, manifested by reduced systemic inflammation and multi-organ damage.

Conclusion

Our findings shed new light on the immunomodulatory functions of cytokine-primed MSCs, offering groundbreaking insights for developing MSC-based therapies against inflammatory diseases via interfering the expression of NINJ1.

1. Introduction

Mesenchymal stem cells (MSCs), mesoderm-derived adult stem cells, have been clinically applied in treating immune-related disorders such as graft-versus-host disease, systemic lupus erythematosus, rheumatoid arthritis, and inflammatory bowel disease with demonstrated therapeutic efficacy.^1,2 Since MSCs have exhibited not only tissue regenerative potential but also potent immunomodulatory properties, which are mediated through both direct cell-to-cell interactions with immune components and paracrine secretion of immunoregulatory factors. Specifically, MSCs modulate the activity of lymphocytes, macrophages, dendritic cells, NK cells, and neutrophils via expression and secretion of indoleamine 2,3-dioxygenase (IDO), programmed death-ligand 1 (PD-L1), interleukin (IL)-10, tumor necrosis factor-alpha (TNF)-stimulated gene 6 protein (TSG-6), and prostaglandin E2,^3,4 thereby mitigating localized or systemic inflammatory responses. Studies reveal that MSCs exhibit limited immunomodulatory activity under homeostatic conditions. For example, inflammatory cytokines such as interferon-gamma (IFN-γ), TNF-α, IL-17, and IL-1β prime MSCs to acquire enhanced immunoregulatory and immunosuppressive functions.5, 6, 7, 8, 9, 10, 11, 12, 13, 14 However, this cytokine priming induces functional heterogeneity in MSCs, concurrently amplifying both anti-inflammatory and pro-inflammatory responses while compromising cell viability and therapeutic consistency.

The terminal stage of lytic cell death involves cellular swelling and plasma membrane rupture (PMR), leading to leakage of proteins, nucleic acids, metabolites, and other cellular components into surrounding tissues. These released substances act as damage-associated molecular patterns (DAMPs) or danger signals, which are recognized by cell surface receptors such as Toll-like receptors, P2Y receptors, and scavenger receptors. This recognition triggers transcriptional programs that induce inflammation, recruit and activate immune cells, thereby amplifying inflammatory responses.^15,16 PMR was regarded as a passive lytic event caused by osmotic imbalance prior to 2021. Afterwards, Kayagaki et al.¹⁷ demonstrated that PMR is not a passive process but rather an actively regulated post-mortem lytic event mediated by nerve injury-induced protein 1 (NINJ1).

NINJ1 not only participates in PMR but also plays a pivotal role in myeloid cell-mediated inflammatory responses, particularly in macrophages and monocytes. Studies by the Ifergan and his team revealed that NINJ1-neutralizing antibodies significantly reduce disease activity in autoimmune encephalomyelitis and inhibit the infiltration of macrophages, dendritic cells, and antigen-presenting cells into the central nervous system.¹⁸ Further research by Lee et al.¹⁹ demonstrated that NINJ1 drives neutrophil migration to ischemic brain tissue, while suppressing NINJ1 expression via small interfering RNA or neutralizing antibodies effectively alleviates neuroinflammation. Current evidence has preliminarily elucidated the molecular mechanisms underlying NINJ1-mediated inflammation regulation: it modulates macrophage migration and adhesion to orchestrate inflammatory progression.^20,21 Jennewein et al.²² discovered that inhibiting NINJ1 with neutralizing peptides or RNA interference markedly attenuates LPS-induced inflammatory responses in macrophages and endothelial cells, an effect linked to the regulation of p38 mitogen-activated protein kinase phosphorylation and the activator protein 1 signaling pathway. In murine sepsis models, upregulated NINJ1 expression exacerbates systemic inflammatory responses and multi-organ damage by promoting leukocyte migration and activating Toll-like receptor 4 signaling.

The functional role of NINJ1 in MSCs, particularly its regulatory effects on inflammatory responses, remains poorly characterized. To address this knowledge gap, this study identifies NINJ1 as a potential regulatory target and investigates its mechanistic contributions to the anti-inflammatory efficacy of human umbilical cord (hUC)-MSCs under synergistic priming with pro-inflammatory cytokines IFN-γ and TNF-α. These findings aim to establish novel theoretical foundations for enhancing the immunomodulatory capacity of MSCs in inflammatory pathologies.

2. Methods

2.1. Materials and reagents

Mesenchymal stem cell serum-free medium (Chongqing Fumei Stem Cell Biotech), human umbilical cord MSC trilineage differentiation kit (Cyagen Biosciences), anti-human CD29, CD34, CD44, CD45, CD73, CD90, CD105 antibodies (BD Biosciences), phosphate-buffered saline (PBS) (BI Israel), recombinant human IFN-γ and TNF-α (PeproTech), mouse TNF-α/IL-1β/IL-6 Enzyme-Linked Immunosorbent Assay (ELISA) kits (Elabscience), Lipofectamine RNAiMAX (Invitrogen), Opti-MEM (Gibco), anti-NINJ1 antibody (HuaAn Biotech, #HA500367), anti-β-actin antibody (Bioss, #bs-0061R), Cell Counting Kit-8 (CCK-8) (Vazyme), lactate dehydrogenase (LDH) Cytotoxicity Assay Kit (Beyotime), carboxyfluorescein diacetate succinimide ester (CFSE) proliferation kit (Beyotime), fetal bovine serum (FBS) (Gibco), and trypsin-ethylenediaminetetraacetic acid (EDTA) (Gibco) have been used in this study.

2.2. Culture and characterization of hUC-MSCs

2.2.1. Cell culture

hUC-MSCs (Chongqing Fumei Biotechnology Co., Ltd., China) were isolated from Wharton's jelly and expanded under standard culture conditions as described in the literature.²³ At passage 3 (P3), cells exhibited a homogeneous spindle-shaped morphology with characteristic vortex-like spatial organization. Upon reaching ≥80% confluence, adherent cells were dissociated using 0.125% trypsin-EDTA, resuspended in cryoprotective medium (10% dimethyl sulfoxide, 90% FBS in α-MEM), and cryopreserved in liquid nitrogen vapor phase at a density of 1 × 10⁶–2 × 10⁶ cells/mL per vial. For experimental use, cryopreserved cells were rapidly thawed at 37 °C, washed twice with PBS to remove residual dimethyl sulfoxide, and subsequently cultured in serum-free medium under 5% CO₂ at 37 °C. Cells were passaged at 1:2–1:3 ratios. All functional assays were performed using cells between (P5–P6), ensuring genomic stability and compliance with International Society for Cellular Therapy (ISCT) criteria for MSC characterization.

2.2.2. Phenotypic identification

P5 hUC-MSCs were incubated with phycoerythrin-labeled monoclonal antibodies targeting CD29, CD44, CD73, CD90, CD105 (positive markers), CD34, and CD45 (negative markers) at a concentration of 1 μL per 10⁵ cells for 30 min at 37 °C under light-protected conditions. Parallel isotype-matched immunoglobulin controls were included to validate antibody specificity. Following 3 washes with ice-cold PBS, cellular suspensions were subjected to surface marker quantification using a NovoCyte™ flow cytometer (ACEA Biosciences (Hangzhou) Co., Ltd.).

2.2.3. Trilineage differentiation

Osteogenic differentiation: P5 hUC-MSCs were seeded at a density of 2 × 10⁵ cells/well into 6-well plates pre-coated with 0.1% gelatin for 1 h at room temperature. When cells reached 70% confluence, the basal medium was replaced with osteogenic induction medium. The medium was replenished every 72 h for 21–28 days. Cells were fixed with 0.4% paraformaldehyde (PFA) and stained with Alizarin Red solution at room temperature for 10 min.

Adipogenic differentiation: P5 hUC-MSCs were seeded at 2 × 10⁵ cells/well into 0.1% gelatin-coated 6-well plates. After achieving complete confluence, the medium was replaced with adipogenic induction medium A for 72 h, followed by a 24-h incubation in adipogenic medium B. This cyclic protocol was repeated for 14–21 days until mature lipid vacuoles were microscopically confirmed. Cells were fixed with 0.4% PFA and stained with Oil Red O solution at room temperature for 30 min.

Chondrogenic differentiation: P5 hUC-MSCs (3 × 10⁵ cells/tube) were transferred into 15 mL centrifuge tubes and washed twice with chondrogenic induction premix solution. The cell pellet was resuspended in 0.5 mL complete chondrogenic induction medium and centrifuged at 150×g for 5 min at 20 °C. The tube caps were loosened, and the pellets were cultured vertically in an incubator. After 24–48 h, the tubes were gently tapped to detach the chondrospheres, which were then maintained in suspension. The medium was refreshed every 3 days until chondrospheres reached 1.5–2.0 mm in diameter. Samples were fixed with 0.4% PFA, embedded, sectioned, and stained with Alcian Blue for 1 h.

2.3. RNA interference

P5 hUC-MSCs were seeded into 6-well plates and cultured until reaching 80% confluency. In 2 sterile 1.5 mL DNase/RNase-free microcentrifuge tubes, the following mixtures were prepared.

• •Tube 1: 100 μL Opti-MEM medium +5 μL small interfering RNA (siRNA) (10 nM) (Table 1) targeting the gene of interest.
• •Tube 2: 100 μL Opti-MEM medium +4 μL RNA Mix.

Table 1.

siRNA sequences.

siRNA name	Sequence (5′–3′)	Conc.	Catalog No.
hNINJ1-#1	CUGGUGUUCAUCAUCGUGGUA	10 nM	1943120379
hNINJ1-#2	GCUGGACAUCGCGCUGCUGAU	10 nM	1943120381
hNINJ1-#3	CCUUGUCAAGUACGACCUUAA	10 nM	1943120377

siRNA: small interfering RNA; Conc.: concentration.

The mixtures in both tubes were thoroughly vortexed. Subsequently, the RNA Mix mixture (Tube 2) was added to the siRNA mixture (Tube 1), gently mixed, and incubated at room temperature for 15 min to allow complex formation. The RNA interference complexes were then added dropwise to the corresponding wells, mixed gently, and the cells were cultured for 48 h prior to further processing. siRNAs targeting the specified genes were designed and synthesized by Sangon Biotech (Shanghai) Co., Ltd.

2.4. hUC-MSCs-mediated peripheral blood mononuclear cell (PBMC) proliferation suppression

2.4.1. Isolation and culture of PBMCs

Peripheral venous blood (20 mL) was collected in sodium heparin-coated vacutainers (BD Biosciences), and immediately transferred to 50 mL conical tubes. After initial centrifugation at 1000×g for 10 min, the plasma fraction was aseptically aspirated and subjected to heat inactivation at 56 °C for 30 min in a water bath, followed by secondary centrifugation at 2000×g for 10 min to remove precipitated fibrinogen aggregates. The remaining blood cells was resuspended in 30 mL physiological saline and carefully layered over Ficoll-Paque™ Plus density gradient medium (GE Healthcare, 1.077 g/mL) at a 2:1 blood-to-separation medium ratio. Density gradient centrifugation was performed at 800×g for 20 min (acceleration rate 3, deceleration rate 1) under ambient conditions. The PBMCs-enriched interphase layer was harvested using sterile transfer pipettes, diluted 1:3 with physiological saline, and washed twice at 300×g for 5 min with brake disengaged to minimize platelet contamination.

Purified PBMCs were resuspended in serum-free lymphocyte medium (Chongqing Fumei Biotechnology Co., Ltd., China) supplemented with 10% autologous heat-inactivated plasma, 200 IU/mL recombinant human IL-2 (PeproTech), and CD3/CD28 T-cell activator reagent (Thermo Fisher). Cultures were maintained at 37 °C in a humidified 5% CO₂ incubator, with 50% medium replacement performed every 48–72 h based on metabolic acidification rates monitored by phenol red indicator.

2.4.2. Co-culture of PBMCs and hUC-MSCs

PBMCs were labeled with CFSE for proliferation tracking. Briefly, PBMCs were adjusted to a density of 5 × 10⁶–10 × 10⁶ cells/mL in protein-free PBS and incubated with 1 μM CFSE for 10 min at 25 °C under light-protected conditions. To terminate labeling, a 5-fold volume of pre-cooled complete medium, followed by 5 min incubation at 4 °C. Cells were subsequently washed thrice with complete medium (300×g, 8 min, 20 °C) to eliminate unbound dye.

For co-culture experiments, hUC-MSCs monolayers in 6-well plates (pre-cultured to 80% confluence) were gently rinsed twice with PBS. CFSE-labeled PBMCs were added at a 2:1 PBMC-to-hUC-MSCs ratio (1 × 10⁶ PBMCs/well) in serum-free lymphocyte medium. Control wells containing PBMCs alone were included to assess baseline proliferation. Following 72-h co-culture under 5% CO₂ at 37 °C, cells were harvested using PBS and analyzed on a NovoCyte™ flow cytometer equipped with 488 nm excitation. Proliferation indices were calculated based on sequential halving of fluorescence intensity in divided daughter cells.

2.5. hUC-MSCs and RAW264.7 Co-culture

RAW 264.7 murine macrophages were resuscitated from cryopreservation and maintained in high-glucose Dulbecco's Modified Eagle Medium (Corning) supplemented with 10% heat-inactivated FBS under standard culture conditions (37 °C, 5% CO₂, 95% humidity). Medium changes occurred every 72 h until ready for passage. Following expansion, cells were seeded into 6-well tissue culture plates at a density of 5 × 10⁴ cells/cm² and incubated until attaining 80% confluence. To synchronize cellular metabolic states, the growth medium was replaced with serum-free high-glucose Dulbecco's Modified Eagle Medium for 12 h prior to co-culture initiation.

Preconditioned hUC-MSCs (P5) were introduced into the macrophage cultures at a 1:1 effector-to-target ratio (2 × 10⁵ cells/well). The co-culture system was subsequently stimulated with 200 ng/mL LPS and 20 ng/mL recombinant murine IFN-γ for 24 h to induce M1 macrophage polarization.

2.6. LDH release assay

P5 hUC-MSCs were seeded into 96-well flat-bottom plates at a density of 5 × 10³ cells/well in 100 μL complete growth medium. Following experimental interventions, 10 μL of Lysis Buffer was added to maximum LDH activity control wells to achieve complete cellular disruption. Plates were gently agitated for 30 s and incubated at 37 °C for 60 min to stabilize enzymatic activity. Subsequent to incubation, plates were centrifuged at 400×g for 5 min, a volume of 120 μL supernatant from each well was carefully transferred to fresh 96-well assay plates using multichannel pipettes, avoiding cellular debris. LDH detection reagent (60 μL/well, prepared per manufacturer's protocol) was added and mixed via orbital shaking (300 rpm, 30 s). After 30-min incubation at 25 °C protected from light, absorbance was measured at 490 nm with a reference wavelength of 620 nm microplate reader.

2.7. Animal experiments

Specific Pathogen-Free-grade male C57BL/6J mice (6–8 weeks old, 20–25 g) were procured from the Experimental Animal Center of Army Medical Center and maintained under controlled conditions (22 ± 1 °C, 55% humidity, 12 h light/dark cycle) with ad libitum access to sterilized feed and water. All experimental protocols were approved by the Animal Ethics Committee of Army Medical University (Approval No.: AMUWEC20237429) in compliance with ARRIVE guidelines.

Eighty mice were randomly allocated into 4 experimental groups (n=20/group): Sham group (laparotomy without cecal ligation); PBS-treated group (cecal ligation and puncture (CLP) + 0.1 mL phosphate-buffered saline); hUC-MSCs (siRNA-IT)-treated group (CLP + hUC-MSCs transfected with scrambled siRNA); and hUC-MSCs (siNINJ1-IT)-treated group (CLP + hUC-MSCs with NINJ1-targeting siRNA). The polymicrobial sepsis model was established via standardized CLP as previously described.²⁴ Briefly, mice were anesthetized via intraperitoneal administration of 1% sodium Pentobarbital (40 mg/kg) and maintained on a 37 °C warming pad. Following midline laparotomy, the mid-distal cecum was ligated 1 cm from the ileocecal junction using 3.0 silk suture, then perforated twice with a 21-gauge needle to ensure luminal content extrusion. Postoperatively, all mice received subcutaneous pre-warmed normal saline (50 mL/kg) for resuscitation. At 3 h post-surgery, therapeutic interventions were administered via tail vein injection: sham-operated and PBS control groups received 0.1 mL sterile phosphate-buffered saline, while treatment groups received 0.1 mL hUC-MSCs suspension containing 5 × 10⁵ viable cells in PBS supplemented with 1% human serum albumin. Survival analysis was conducted at 12-h intervals for 5 days.

An additional cohort of 40 male C57BL/6J mice (6–8 weeks old, 20–25 g) were randomly allocated into four experimental groups (n = 10/group). Following identical modeling and therapeutic protocols, blood samples were collected via Retro-orbital bleeding 24 h post-intervention. Serum was isolated through centrifugation (3000×g, 15 min, 4 °C) and stored at −80 °C until analysis. Inflammatory cytokines (TNF-α, IL-1β, IL-6) were quantified using high-sensitivity ELISA kits following manufacturer protocols. Aliquots of serum were submitted to the institutional clinical laboratory for quantification of alanine aminotransferase, aspartate aminotransferase, creatinine, and urea using standardized enzymatic assays.

2.8. Statistical analysis

Data were analyzed using SPSS 26.0 statistical software. For comparisons between 2 groups, a two-tailed Student's t-test was performed. For comparisons among three or more groups, one-way ANOVA was first used to assess overall differences. If significant overall differences were found (p < 0.05), Tukey's post-hoc test was conducted to determine pairwise group differences. A p < 0.05 was considered statistically significant. Data are presented as mean ± standard deviation to describe the variability of the raw data or mean ± standard error of the mean to indicate the precision of the mean estimate. All experiments were performed with at least 3 biological replicates (n ≥ 3).

3. Results

3.1. Culture and characterization of hUC-MSCs

hUC-MSCs were expanded under standard culture conditions and systematically characterized for phenotypic and functional properties according to ISCT guidelines. Phase-contrast microscopy revealed typical spindle-shaped morphology with whirlpool-like growth patterns at 85%–95% confluency (Fig. 1A). The trilineage differentiation assay confirmed that the hUC-MSCs retained their multipotent capacity to differentiate into osteogenic, adipogenic, and chondrogenic lineages under standardized induction protocols (Fig. 1B). Flow cytometric analysis demonstrated ≥95% positivity for CD29, CD44, CD73, CD90 and CD105, while maintaining ≤2% expression of hematopoietic markers CD34 and CD45 (Fig. 1C). These findings confirm that the isolated hUC-MSCs strictly adhere to the ISCT minimal criteria for MSCs, including plastic adherence, multipotent differentiation capacity and definitive surface marker profile.

Fig. 1.

Characterization of hUC-MSCs. (A) Morphological features of hUC-MSCs. (B) Trilineage differentiation capacity (left to right: adipogenic, osteogenic, and chondrogenic differentiation). (C) Flow cytometric analysis of surface markers. hUC-MSCs: human umbilical cord mesenchymal stem cells.

3.2. Heterogeneity and adverse effects of hUC-MSCs primed with IFN-γ and TNF-α co-stimulation

hUC-MSCs were primed in vitro with IFN-γ (20 ng/mL) and TNF-α (20 ng/mL) for 24 h. RT-PCR analysis revealed that combined IFN-γ/TNF-α stimulation significantly upregulated transcription levels of anti-inflammatory genes (IDO, PD-L1, and IL-10) (Fig. 2A–C) alongside elevated pro-inflammatory genes (IL-6 and IL-1β) (Fig. 2D and E), indicating heterogeneous inflammatory regulation in primed hUC-MSCs. CCK-8 assays demonstrated reduced cell viability in IFN-γ/TNF-α-primed hUC-MSCs (Fig. 2F). However, co-culture experiments with PBMCs showed enhanced suppression of lymphocyte proliferation by primed hUC-MSCs compared to untreated controls (Fig. 2G and H). These findings suggest that while IFN-γ/TNF-α co-stimulation amplifies the global anti-inflammatory capacity of hUC-MSCs, concurrent proinflammatory activation and viability impairment may partially counteract therapeutic efficacy. Targeted mitigation of these adverse effects—such as pathway-specific inhibitors or metabolic preconditioning—could refine the anti-inflammatory potential of primed hUC-MSCs for clinical applications.

Fig. 2.

Immunoregulatory heterogeneity and adverse effects of hUC-MSCs primed with IFN-γ and TNF-α co-stimulation. (A–C) Transcriptional levels of anti-inflammatory genes (IDO, PD-L1, IL-10) in primed hUC-MSCs (n=3). (D–E) Transcriptional levels of proinflammatory genes (IL-6, IL-1β) in primed hUC-MSCs (n=3). (F) Cell viability assessed by CCK-8 assay after 24 h stimulation with IFN-γ/TNF-α alone or in combination (n=7). (G) CFSE dilution profiles showing PBMC proliferation after 3-day co-culture with hUC-MSCs (n=3). (H) PBMC proliferation suppression rate (n=3).^# IDO: indoleamine 2,3-dioxygenase; mRNA: messenger RNA; IFN-g: interferon-gamma; TNF-a: tumor necrosis factor-alpha; PDL1: programmed death-ligand 1; IL: interleukin; PBMC: peripheral blood mononuclear cell; hUC-MSC: human umbilical cord mesenchymal stem cell; IT: interferon-gamma and tumor necrosis factoralpha co-stimulation; CFSE: Carboxyfluorescein diacetate succinimide ester.

^# An unpaired, 2-tailed Student t-test was used to analyze the data. The data are presented as the mean ± standard deviation.

ns p ≥ 0.05, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

3.3. NINJ1 expression dynamics in hUC-MSCs under IFN-γ/TNF-α stimulation and transcriptomic profiling post-knockdown

In hUC-MSCs, IFN-γ stimulation alone failed to induce transcriptional or translational upregulation of NINJ1, whereas TNF-α monotherapy moderately enhanced its expression (Fig. 3A and B). Notably, co-stimulation with IFN-γ (20 ng/mL) and TNF-α (20 ng/mL) synergistically amplified both messenger RNA and protein levels of NINJ1 (Fig. 3A and B), suggesting its pivotal role in regulating the immunomodulatory reprogramming of hUC-MSCs under cytokine co-stimulation.

Fig. 3.

NINJ1 expression in hUC-MSCs under IFN-γ/TNF-α stimulation and transcriptomic profiling post-NINJ1 knockdown. (A) Transcriptional expression of NINJ1 in hUC-MSCs stimulated with IFN-γ, TNF-α alone, or in combination (n=3). (B) Protein expression of NINJ1 under corresponding stimulation conditions (n=3). (C) Transcriptional validation of NINJ1 knockdown via siRNA (n=3). (D) Protein-level validation of NINJ1 knockdown (n=3). (E) Volcano plot of differentially expressed genes after NINJ1 silencing and analysis of inflammation-associated downregulated following NINJ1 knockdown. (F) Enriched pathways of upregulated genes post-NINJ1 knockdown. (G) Enriched pathways of downregulated genes.^# NINJ1: nerve injury-induced protein 1; mRNA: messenger RNA; IFN-g: interferon-gamma; TNF-a: tumor necrosis factor-alpha; siRNA: small interfering RNA; IT: interferon-gamma and tumor necrosis factor-alpha co-stimulation.

^# An unpaired, 2-tailed Student t-test was used to analyze the data. The data are presented as the mean ± standard deviation.

ns p ≥ 0.05, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To elucidate NINJ1's functional contributions, we transfected hUC-MSCs with NINJ1-specific siRNA (Fig. 3C and D) and performed transcriptome sequencing post IFN-γ/TNF-α co-stimulation. RNA-seq analysis identified 1363 differentially expressed genes (DEGs): 718 upregulated and 645 downregulated (Fig. 3E). Gene Ontology enrichment analysis of inflammation-related DEGs revealed marked suppression of proinflammatory mediators, including IL-1β and IL-6 (Fig. 3E). Upregulated genes were predominantly associated with extracellular matrix remodeling, particularly collagen biosynthesis (Fig. 3F), while downregulated genes clustered in chemotaxis and inflammatory pathways (Fig. 3G). These findings position NINJ1 as a molecular rheostat governing hUC-MSCs' immunoregulatory plasticity under IFN-γ/TNF-α co-stimulation.

3.4. Functional alterations in hUC-MSCs following NINJ1 knockdown

Our study revealed that in hUC-MSCs without IFN-γ/TNF-α co-stimulation, NINJ1 knockdown exerted minimal effects on the expression of immunomodulatory genes (PD-L1, IL-10) or proinflammatory mediators (IL-6, IL-1β), except for a modest reduction in IDO levels (Fig. 4A–E). This observation aligns with the basal low expression of NINJ1 in steady-state hUC-MSCs. Strikingly, under IFN-γ/TNF-α co-stimulation, NINJ1 silencing significantly attenuated the upregulation of proinflammatory genes (IL-6, IL-1β, Fig. 4A and B), corroborating transcriptomic sequencing data, while leaving PD-L1 and IDO unaffected (Fig. 4C–E). These findings position NINJ1 as a selective mediator of proinflammatory activation in cytokine-primed hUC-MSCs, suggesting targeted NINJ1 inhibition could refine their immunosuppressive profile.

Fig. 4.

Functional alterations of hUC-MSCs after NINJ1 interference. (A–B) Transcription levels of pro-inflammatory-related genes (n=3). (C–E) Transcription levels of anti-inflammatory-related genes (n=3); (F) Cell viability detected by CCK-8 assay (n=7). (G) LDH release assay results (n=3). (H–I) Apoptosis detection results (n=3).^# IL: interleukin; mRNA: messenger RNA; siRNA: small interfering RNA; NINJ1: nerve injury-induced protein 1; IT: interferon-gamma and tumor necrosis factor-alpha co-stimulation; PD-L1: programmed death-ligand 1; IDO: indoleamine 2,3-dioxygenase; IFNg: interferon-gamma; TNF-a: tumor necrosis factor-alpha; LDH: lactate dehydrogenase; PI: Propidium Iodide.

^# An unpaired, 2-tailed Student t-test was used to analyze the data. The data are presented as the mean ± standard deviation.

ns p ≥ 0.05, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

To investigate NINJ1's role in cytokine-induced viability loss, we observed that NINJ1 knockdown did not alter baseline cell survival (Fig. 4F), excluding its involvement in terminal cell death regulation. However, LDH release assays—a marker of membrane integrity disruption—demonstrated that IFN-γ/TNF-α co-stimulation dose-dependently exacerbated spontaneous cell rupture (Fig. 4G). NINJ1 silencing not only suppressed baseline LDH release but also abolished the cytokine-driven escalation of membrane permeabilization. Annexin V/PI apoptosis assays further revealed that NINJ1 depletion reduced both apoptotic initiation (Annexin V + population) and progression (Annexin V+/PI + cells) under cytokine challenge (Fig. 4H and I).

Collectively, NINJ1 acts as a dual-functional regulator in cytokine-primed hUC-MSCs, selectively suppressing proinflammatory IL-6/IL-1β pathways while preserving immunosuppressive IDO/PD-L1 axes through transcriptional reprogramming, and concurrently maintaining membrane integrity to attenuate apoptosis-associated rupture, thereby reducing DAMPs release and secondary inflammation. This synergistic mechanism—neutralizing intrinsic proinflammatory signals while minimizing DAMPs-driven immunogenicity—provides a strategic framework for enhancing the anti-inflammatory efficacy of hUC-MSCs-based therapies in inflammatory disorders.

3.5. Knockdown of NINJ1 enhances the inflammatory suppressive function of hUC-MSCs under IFN-γ and TNF-α co-stimulation in both in vitro and in vivo settings

In vitro studies, we established a co-culture system of hUC-MSCs and RAW264.7 macrophages to investigate NINJ1-mediated immunomodulatory effects under IFN-γ and TNF-α co-stimulation. Results demonstrated that combined IFN-γ and TNF-α stimulation enhanced the inhibitory effects of hUC-MSCs on IL-6, TNF-α, and IL-1β secretion in RAW264.7 cells. Notably, NINJ1 knockdown further suppressed IL-6 and TNF-α production (Fig. 5A–C).

Fig. 5.

Validation of the anti-inflammatory function of hUC-MSCs after NINJ1 knockdown in vitro and in vivo. (A–C) ELISA results of inflammatory cytokine levels in the co-culture supernatant of hUC-MSCs and RAW264.7 macrophages (n=3). (D) Survival rates of mice in each treatment group after CLP modeling and therapy (n=20). (E–G) Serum inflammatory cytokine levels in mice 24 h post-CLP modeling and treatment (n=4). (H–K) Serum levels of urea, creatinine (Crea), ALT, and AST in mice 24 h post-CLP modeling and treatment (n=3).^# IL: interleukin; siRNA: small interfering RNA; IT: interferon-gamma and tumor necrosis factor-alpha co-stimulation; TNF-a: tumor necrosis factor-alpha; PBS: phosphate-buffered saline; hUC-MSC: human umbilical cord mesenchymal stem cell; NINJ1: nerve injury-induced protein 1; AST: aspartate Transaminase; ALT: alanine Transaminase; Crea: creatinine; ELISA: Enzyme-Linked Immunosorbent Assay; CLP:cecal ligation and puncture.

^# Intergroup comparisons of survival analysis data were performed using the Mantel-Cox log-rank test, while group comparisons of other continuous variables were analyzed with independent 2-tailed Student's t-test.

ns p ≥ 0.05, ∗p < 0.05, ∗∗p < 0.01, ∗∗∗p < 0.001, ∗∗∗∗p < 0.0001.

In vivo, we established a murine sepsis model via CLP to evaluate the role of NINJ1 in the anti-inflammatory efficacy of hUC-MSCs. Survival rate, a critical therapeutic endpoint in sepsis, revealed that PBS-treated mice exhibited rapid mortality, with only 10% survival by postoperative day 4. hUC-MSCs administration significantly improved survival to 40%, though NINJ1 interference did not confer additional survival benefits (Fig. 5D). At 24h post-CLP, serum IL-6, TNF-α, and IL-1β levels surged, confirming successful sepsis induction. IFN-γ/TNF-α-primed hUC-MSCs treatment markedly reduced these proinflammatory cytokines, with NINJ1 knockdown amplifying suppression (Fig. 5E–G). Liver and kidney dysfunction biomarkers—alanine aminotransferase, aspartate aminotransferase, creatinine, and urea—were significantly elevated post-CLP. IFN-γ/TNF-α-primed hUC-MSCs effectively attenuated these increases, and NINJ1 interference potentiated this protective effect (Fig. 5H–K).

Collectively, both in vitro and in vivo experiments demonstrate that knockdown of NINJ1 enhances the anti-inflammatory function of hUC-MSCs under the IFN-γ and TNF-α co-stimulation system.

4. Discussion

MSCs were first identified in bone marrow in 1968, and subsequent research has expanded their isolation and preparation to diverse tissues including adipose tissue, synovium, bone, muscle, umbilical cord, placenta, amniotic fluid, and cord blood.²⁵ Although bone marrow-derived MSCs dominated early research, subsequent studies have increasingly demonstrated the superiority of hUC-MSCs. Compared to other sources, the umbilical cord not only offers abundant availability but also contains a higher yield of MSCs, enabling large-scale isolation in a single procedure and significantly shortening preparation timelines. Furthermore, hUC-MSCs exhibit lower immunogenicity, enhanced bioactivity, superior proliferative capacity, and fewer ethical concerns compared to bone marrow-derived MSCs.^26,27 These advantages collectively underpin our rationale for selecting hUC-MSCs as the experimental model in this study.

Globally, over 1300 registered clinical trials have explored MSC-based therapies for conditions spanning musculoskeletal disorders, chronic graft-versus-host disease, respiratory pathologies, neurodegenerative diseases, autoimmune conditions, and collagen-related disorders, with demonstrated therapeutic efficacy. Despite these advances, most transplanted MSCs fail to fully realize their therapeutic potential, and numerous clinical trials report suboptimal outcomes, resulting in only a handful of MSC-based products receiving the USA Food and Drug Administration approval.²⁸ Key challenges include batch-to-batch variability in immunocompatibility, differentiation potential, and senescence control, as well as inconsistencies in quality control protocols.^29,30 To address these limitations, 4 principal strategies are being pursued: (1) genetic engineering via viral vectors or CRISPR/Cas9 to enhance homing, proliferation, and migration; (2) functional priming through hypoxia or small-molecule stimulation (e.g., cytokines) to improve survival and immunomodulatory capacity; (3) biomaterial integration using scaffolds to optimize engraftment, and (4) secretome utilization leveraging MSC-derived exosomes as drug delivery platforms.^31,32 Among these, cytokine priming—particularly co-stimulation with IFN-γ and TNF-α—has emerged as a widely adopted approach. Our team has validated that IFN-γ/TNF-α co-stimulation significantly amplifies the immunosuppressive capacity of hUC-MSCs, marked by upregulated expression of IDO, PD-L1 and IL-10, alongside enhanced inhibition of PBMC proliferation. However, this strategy introduces heterogeneity and adverse effects, including the paradoxical upregulation of proinflammatory IL-6/IL-1β and reduced cell viability, necessitating ongoing mechanistic investigations to refine therapeutic protocols.

First identified in 1996 as a cell adhesion protein induced by neural injury and hence named NINJ1,³³ this molecule has since been found to exhibit high expression in epithelial and myeloid cells. Beyond its roles in neural injury repair,^34,35 tumorigenesis,36, 37, 38 and vascular endothelial function,39, 40, 41 NINJ1 plays a pivotal regulatory role in inflammatory responses.^19,22,42,43 Its molecular mechanisms in inflammation have been partially elucidated, primarily involving modulation of macrophage migration and adhesion.^20,21 Recent breakthroughs reveal NINJ1 as the structural cornerstone of lytic cell death,⁴⁴ serving as the executor of PMR and DAMPs release post-cell death.45, 46, 47, 48 Structurally, NINJ1 is a 16 kDa transmembrane protein featuring 2 transmembrane helices and an extracellular amphipathic helix. Upon membrane tension, NINJ1 dimers undergo conformational changes, assembling into oligomers through amphipathic helix-mediated intra- and intermolecular interactions. These oligomers disrupt membrane integrity via pore formation or lipid dissolution, thereby inducing PMR.¹⁷ Notably, NINJ1 has been validated as the executioner of PMR across diverse cell death modalities, including pyroptosis,¹⁷ apoptosis,¹⁷ ferroptosis,⁴⁹ necrosis,⁵⁰ and secondary necrosis.⁵¹

In this study, we observed a significant upregulation of NINJ1 expression in hUC-MSCs following IFN-γ and TNF-α co-stimulation priming. Notably, NINJ1 knockdown effectively reversed the priming-induced elevation of pro-inflammatory genes such as IL-6 and IL-1β. Mechanistically, we identified NINJ1 as the central executor mediating PMR and DAMPs release in co-stimulated hUC-MSCs. Genetic interference of NINJ1 substantially attenuated PMR progression and DAMPs efflux by inhibiting post-mortem lytic processes.

Following the initial elucidation of NINJ1-mediated inflammatory regulation mechanisms, our research focus shifted toward therapeutic validation. We designed and conducted in vitro and in vivo experiments to investigate the role of NINJ1 in the anti-inflammatory functionality of hUC-MSCs. The murine monocyte/macrophage leukemia cell line RAW264.7 served as a critical in vitro model, as it reliably mimics inflammatory responses when activated by stimuli such as LPS. This activation triggers the release or upregulation of key inflammatory mediators including TNF-α, IL-6 and IL-1β, making RAW264.7 cells an ideal tool for studying inflammatory processes and screening anti-inflammatory agents.⁵² Through co-culture experiments with RAW264.7 cells, we demonstrated that NINJ1 interference significantly enhances the inflammation-suppressing capacity of hUC-MSCs under IFN-γ and TNF-α co-stimulation.

Sepsis, a systemic inflammatory response syndrome predominantly triggered by trauma or infection,⁵³ is characterized by an early surge of hallmark inflammatory cytokines such as TNF-α, IL-6, and IL-1β.⁵⁴ These cytokines enter the bloodstream, inducing the aggregation and activation of macrophages, monocytes, neutrophils, and platelets, thereby perpetuating a cascade of inflammatory mediators that drive systemic inflammation. Uncontrolled inflammation ultimately leads to shock, multi-organ dysfunction, and failure, contributing to its alarmingly high morbidity and mortality rates and posing a grave threat to human health.⁵⁵ MSCs, owing to their unique biological advantages—particularly their potent anti-inflammatory and immunomodulatory properties—hold significant promise in sepsis treatment. However, current clinical trials reveal suboptimal therapeutic outcomes with MSC-based interventions, underscoring the urgent need for novel strategies to enhance efficacy.56, 57, 58, 59 In our study, a murine sepsis model was established via cecal CLP, followed by intravenous administration of hUC-MSCs. The results demonstrated that NINJ1 interference significantly augments the anti-inflammatory therapeutic effects of hUC-MSCs in sepsis management.

Notably, while the knockdown of NINJ1 significantly ameliorated short-term inflammatory markers in septic models, this therapeutic effect failed to translate into long-term survival benefits, a discrepancy likely attributable to multifactorial mechanisms. Firstly, the transient efficacy of siRNA-mediated gene silencing resulted in gradually diminished NINJ1 suppression over time. Secondly, the complex inflammatory milieu in sepsis, characterized by high levels of inflammatory factors (e.g., IL-1β, HMGB1)⁶⁰ perpetuated NINJ1 transcriptional activation, elevating its messenger RNA abundance and partially counteracting siRNA-mediated degradation. Furthermore, accelerated apoptosis of exogenous MSCs under intense inflammatory stress61, 62, 63 compromised their sustained immunomodulatory capacity. Most critically, NINJ1 knockdown may potentially triggered compensatory activation of alternative signaling pathways, thereby amplifying cellular response complexity. These findings not only elucidate the dissociation between transient anti-inflammatory effects and enduring survival outcomes but also delineate actionable strategies for therapeutic refinement—engineering NINJ1-knockout stem cells or developing targeted delivery systems may surmount current limitations and achieve durable efficacy.

Research evidence demonstrates an inverse correlation between MSCs passage number and cellular senescence, with advanced-passage cells exhibiting diminished therapeutic potential due to restricted proliferation, attenuated differentiation capacity, and dysregulated gene expression profiles.64, 65, 66, 67 Notably, NINJ1 has been identified as a downstream target of p53, modulating p53 expression and its dependent cellular survival and senescence pathways.³⁶ Building upon these findings, our experimental design rigorously standardized passage number variables. Although systematic exploration of NINJ1 functional heterogeneity across different passages of hUC-MSCs was beyond this study's scope, the exclusive use of passage 5–6 cells effectively mitigated functional interference from advanced-passage cells exhibiting marked senescence-associated impairments. Crucially, this passage selection aligns with clinically translational standards (P3–P6) widely adopted in therapeutic development,68, 69, 70 ensuring both experimental reproducibility and technical compatibility for potential clinical translation.

In summary, the co-stimulation strategy employing IFN-γ and TNF-α represents a validated approach to amplify the anti-inflammatory capacity of hUC-MSCs. Within this priming framework, our findings demonstrate that NINJ1 impairs the anti-inflammatory function of hUC-MSCs through dual pro-inflammatory mechanisms—inducing the expression of pro-inflammatory genes such as IL-1β and IL-6, and mediating PMR to facilitate the release of DAMPs, as illustrated in Fig. 6. Targeted genetic knockdown of NINJ1 significantly enhances their anti-inflammatory therapeutic efficacy by suppressing these pro-inflammatory mediators, thereby amplifying the immunomodulatory capacity of hUC-MSCs. However, the mechanistic intricacies underlying this phenomenon, particularly how NINJ1 knockdown modulates cytokine crosstalk and downstream immunomodulatory pathways, remain incompletely elucidated. Future studies should further investigate the underlying molecular mechanisms by which NINJ1 influences the immunomodulatory functions of hUC-MSCs. This could involve generating NINJ1-knockout engineered hUC-MSCs lines to systematically elucidate the functional roles of NINJ1, thereby providing mechanistic clarity and potential targets for optimizing MSC-based therapies.

Fig. 6.

Schematic diagram of the mechanism by which IFN-γ and TNF-α synergistically enhance the expression of NINJ1 and impair the anti-inflammatory function of MSCs.

IFN-γ: interferon-gamma; TNF-α: tumor necrosis factor-alpha; IL: interleukin; NINJ1: nerve injury-induced protein 1; IFNGR: interferon gamma receptor; TNFR: tumor necrosis factor receptor; hUC-MSC: human umbilical cord mesenchymal stem cell; DAMPs: damage-associated molecular patterns; TLRs: toll-like receptors.

CRediT authorship contribution statement

Wang Hu: Writing – original draft, Validation, Software, Methodology, Formal analysis. Guomei Yang: Writing – original draft, Validation, Software, Methodology. Luoquan Ao: Validation, Software, Investigation. Peixin Shen: Validation, Data curation. Mengwei Yao: Validation, Methodology. Yuchuan Yuan: Funding acquisition. Jiaoyue Long: Writing – review & editing, Methodology. Zhan Li: Writing – review & editing, Supervision, Methodology, Conceptualization. Xiang Xu: Writing – review & editing, Supervision, Project administration, Funding acquisition, Conceptualization.

Ethical statement

The animal experiments were approved by Laboratory Animal Welfare and Ethics Committee of the Army Medical University (Approval No.: AMUWEC20237429)

Funding

This research was supported by the National Natural Science Foundation of China (No.82302436), and Science Foundation of State Key laboratory of Trauma and Chemical Poisoning (No. 2024K003).

Declaration of competing interest

The authors declare that they have no competing interests.