Strain-specific protection of intestinal barrier by heat-treated plant-derived lactic acid bacteria via distinct tight junction modulation

Abstract

Intestinal barrier dysfunction is a key pathological factor in various chronic diseases, including inflammatory bowel disease. This study compared the barrier-protective effects and mechanisms of two heat-treated plant-derived lactic acid bacteria (P-LAB) strains, Lactiplantibacillus plantarum LOC1 isolated from tea leaves and Lacticaseibacillus paracasei RO24 isolated from traditional Japanese pickles, using lipopolysaccharide (LPS)-stimulated intestinal epithelial cell models. Barrier integrity was evaluated in Caco-2 and HT29 cells by measuring transepithelial electrical resistance (TEER) and paracellular permeability to Lucifer Yellow, together with analyses of tight junction-related gene expression, inflammatory cytokines, and global proteomic profiling integrated with Ingenuity Pathway Analysis. LPS exposure significantly impaired barrier function in both cell lines. Pretreatment with either strain significantly suppressed LPS-induced increases in paracellular permeability, whereas TEER recovery was strain- and cell-line-dependent. L. plantarum LOC1 effectively restored TEER in Caco-2 cells, while L. paracasei RO24 exerted significant protective effects primarily in HT29 cells. Mechanistic analyses revealed distinct modes of action between the two strains. L. plantarum LOC1 showed broad anti-inflammatory activity and maintained the expression of JAM-A and occludin, representing a comprehensive barrier-protective strategy. In contrast, L. paracasei RO24 selectively preserved Claudin-1, Claudin-3, and Claudin-7, indicating a structural reinforcement strategy. Proteomic and pathway analyses confirmed reactivation of tight junction signaling by both strains. These findings indicate that heat-treated P-LAB strains can exert strain-specific protective effects on intestinal epithelial barrier function under the present experimental conditions.

INTRODUCTION

The intestinal tract plays a vital role in human health by serving as a physical and immunological barrier to prevent the entry of harmful substances, such as foreign antigens, pathogenic microorganisms, and toxins [1]. This barrier function is supported by intercellular adhesion complexes like tight junctions, adherens junctions, and desmosomes, with tight junctions acting as the crucial frontline structure. Tight junctions are composed of proteins such as TJP-1 (tight junction protein 1), occludin, and the Claudin family of proteins. Their proper expression and localization are essential for maintaining the selective permeability and homeostasis of the intestinal epithelium [2, 3].

Recent studies have shown that the abnormal expression or disruption of these structural proteins can lead to a pathological condition called “leaky gut”, which is implicated in various diseases, including inflammatory bowel disease (IBD), irritable bowel syndrome (IBS), allergies, type 2 diabetes, non-alcoholic fatty liver disease, and even mental disorders such as depression [4,5,6,7].

Given the critical role of the intestinal barrier, the relationship between probiotics and gut barrier function has drawn significant attention. Numerous reports demonstrate that specific strains of lactic acid bacteria and bifidobacteria can enhance barrier function by increasing the transepithelial electrical resistance (TEER) of intestinal epithelial cells and promoting the expression of TJP-1, Claudin, and occludin [8,9,10]. Furthermore, lactic acid bacteria also possess immunomodulatory properties, such as regulating Toll-like receptor signaling and suppressing the production of inflammatory cytokines (IL-6, IL-8, and TNF-α), thereby contributing to maintenance of the intestinal barrier [10].

In recent years, the focus has shifted from dairy-derived to plant-derived lactic acid bacteria (P-LAB) originating from fermented plant foods. Some P-LAB strains have been reported to exhibit acid and bile tolerance that may support gastrointestinal survival, although these properties are highly strain-dependent. In addition, certain P-LAB strains may possess unique functional characteristics associated with adaptation to plant-derived substrates [11]. Traditional fermented foods long consumed in Asia, such as green tea and pickles, are rich sources of diverse P-LAB strains and have attracted attention as functional microbial reservoirs [12]. Research is advancing to understand the effects of these strains on the gut environment and host physiology [13, 14]. Previous studies have reported that the lactic acid bacterial strain Lactiplantibacillus plantarum LOC1, isolated from green tea leaves, exhibits intestinal barrier-enhancing effects [15]. This property is considered extremely important for maintaining a healthy intestinal environment. In contrast, the lactic acid bacterial strain Lacticaseibacillus paracasei RO24, isolated from traditional Japanese pickles, is expected to possess novel functional properties that contribute to human health.

However, comparative studies investigating the effects of P-LAB strains derived from tea leaves versus those from pickles on intestinal barrier function under identical conditions remain extremely limited. In this study, we focused on the distinct and harsh environments from which these two strains were isolated: tea leaves, which are rich in antimicrobial polyphenols such as catechins [16, 17], and pickles, which are characterized by high salinity and significant osmotic stress. We hypothesized that the differences in adaptation strategies required for these contrasting environments are reflected in their distinct modes of action on intestinal epithelial cells.

The objective of this research was to comprehensively compare the effects of L. plantarum LOC1 and L. paracasei RO24 using LPS-stimulated intestinal epithelial cell models (Caco-2 and HT29 cells). Our analysis encompasses TEER measurements, the expression of tight junction-related genes, the modulation of inflammatory cytokines, and global proteomic profiling integrated with Ingenuity Pathway Analysis (IPA). Through these multifaceted approaches, we aim to elucidate the strain-specific barrier protection mechanisms of each isolate and evaluate their potential for “precision probiotics” tailored to specific pathological conditions and physiological needs.

MATERIALS AND METHODS

P-LAB strain preparation

The two heat-treated P-LAB strains used in this study were L. plantarum LOC1 and L. paracasei RO24. L. plantarum LOC1 (originally isolated from tea leaves) has been previously identified and reported by our group [15]. L. paracasei RO24 (originally isolated from pickles) has been identified as L. paracasei based on morphological and biochemical characteristics, as well as 16S rRNA gene sequencing analysis. These strains were activated by culturing them in MRS broth for 24 hr at 37°C under anaerobic conditions. After cultivation, bacterial cells were harvested by centrifugation at 5,000 × g for 10 min, washed twice with sterile phosphate-buffered saline, and heat-treated at 121°C for 15 min. The heat-treated bacterial cells were then resuspended in serum-free culture medium and adjusted to a final concentration of 1.0 × 10⁸ cells/mL. The inactivated cells were resuspended in serum-free culture medium and adjusted to a final concentration of 1.0 × 10⁸ cells/mL for all treatments. This concentration was selected based on preliminary optimization experiments and previous studies demonstrating biological activity of lactic acid bacteria in intestinal epithelial cell models without apparent cytotoxicity.

Cell culture

Human intestinal epithelial cell lines, Caco-2 and HT29, were obtained from the American Type Culture Collection and maintained at a density of 5.0 × 10⁴ in 75-cm³ flasks in a humidified incubator at 37°C with 5% CO₂. The cells were cultured in Dulbecco’s Modified Eagle Medium supplemented with 10% fetal bovine serum (Daiichi Kagaku, Tokyo, Japan), 1% non-essential amino acids (Gibco BRL, Grand Island, NY, USA), and 1% penicillin/streptomycin (Invitrogen, Thermo Fisher Scientific, Waltham, MA, USA). They were routinely passaged every 3–4 days using 0.25% trypsin-EDTA (FUJIFILM Wako Pure Chemical Co., Osaka, Japan) to maintain a continuous culture. Cells between passages 30 and 40 were used for the experiments.

Transepithelial electrical resistance measurement

For TEER measurements, Caco-2 or HT29 cells were seeded at a density of 1.0 × 10⁵ cells/well onto the apical side of 12-well Transwell^® permeable inserts (0.4 µm pore size, Corning, NY, USA). The medium in both the apical and basolateral chambers was changed every 2–3 days. The formation of a tight epithelial monolayer was confirmed by monitoring the TEER values for 21 days using an epithelial voltohmmeter with a chopstick electrode (Millicell ERS-2, MilliporeSigma, Burlington, MA, USA). The raw resistance values were corrected by subtracting the background resistance of a blank insert and then normalized to the surface area of the membrane (1.12 cm²) to obtain the TEER values in Ω·cm².

On day 21, the differentiated cell monolayers were assigned to four experimental groups: 1) vehicle control, 2) LPS group (exposed to 1 ng/mL LPS from Escherichia coli O111:B4 [Sigma–Aldrich, St. Louis, MO, USA] for 24 hr), 3) LPS + L. plantarum LOC1 group, and 4) LPS + L. paracasei RO24 group. In the heat-treated P-LAB groups, cells were pre-treated with the respective heat-treated P-LAB strain for 6 hr. Subsequently, LPS (1 ng/mL) was added without removing bacterial cells, and the cells were co-incubated for an additional 24 hr. Thus, the total bacterial exposure time was 30 hr. The TEER values were measured before and after the treatments, and the results were expressed as a percentage of the initial TEER value.

Lucifer Yellow permeability assay

For the Lucifer Yellow (LY) permeability analysis, Caco-2 and HT29 cells were subjected to the same treatment schedule as used for TEER measurement (6 hr of bacterial pretreatment followed by 24 hr of co-incubation with LPS). After completion of treatment, permeability was assessed using LY. To evaluate the recovery of paracellular barrier integrity, an LY permeability assay was performed following the bacterial treatment. After the 24 hr of incubation with the heat-treated P-LAB strains, the medium in the apical and basolateral chambers was removed. A 100 µM LY solution (Sigma-Aldrich, St. Louis, MO, USA) prepared in phenol red-free medium was added to the apical chamber (0.5 mL), while the basolateral chamber was filled with 1.5 mL of fresh medium without LY. The cells were then incubated at 37°C for 1 hr.

Following incubation, 100 µL of the medium was sampled from the basolateral chamber, and the fluorescence intensity was measured using a microplate reader (filter settings: excitation 485 nm, emission 535 nm). A standard curve was generated using LY solutions ranging from 0 to 100 µM to calculate the concentration of LY in each basolateral sample. The permeability was expressed as the percentage of the total amount of LY added to the apical side that had permeated to the basolateral side, according to the following formula:

Permeability (%) = (Total amount in basolateral chamber / Total amount added to apical chamber) × 100.

The results were used to quantitatively assess the extent of barrier disruption by LPS and the subsequent restorative effects of the heat-treated P-LAB strains.

RNA isolation

For the gene expression analysis, Caco-2 and HT29 cells were seeded onto 6-well plates at a density of 2.0 × 10⁵ cells/well and cultured until they reached 80% confluence. Cells were divided into the same four experimental groups described for the TEER assay. Briefly, cells were pretreated with each heat-treated P-LAB strain for 6 hr, followed by co-treatment with LPS (1 ng/mL) for 24 hr. Total RNA was extracted immediately after treatment using TRIzol™ reagent (Thermo Fisher Scientific, Waltham, MA, USA) according to the manufacturer’s protocol. RNA concentration and purity were assessed with a NanoDrop 2000 spectrophotometer (Thermo Fisher Scientific) using the A260/A280 and A260/A230 ratios, respectively. Complementary DNA (cDNA) was synthesized from 1 µg of total RNA using a High-Capacity cDNA Reverse Transcription Kit (Applied Biosystems, Thermo Fisher Scientific, Waltham, MA, USA).

Real-time PCR

Quantitative real-time PCR was performed using a SYBR^® Green PCR Master Mix (Applied Biosystems) on an Applied Biosystems StepOnePlus Real-Time PCR System. Each 20 µL reaction mixture contained 10 µL of 2× SYBR^® Green Master Mix, 0.4 µL each of forward and reverse primer (final concentration, 0.2 µM), 1 µL of the cDNA template (corresponding to approximately 50 ng of reverse-transcribed RNA), and nuclease-free water to adjust the final volume. The thermal cycling conditions were as follows: initial denaturation at 95°C for 10 min, followed by 40 cycles of denaturation at 95°C for 15 sec and annealing/extension at 60°C for 1 min. A melting curve analysis was conducted at the end of amplification to verify the specificity of the products. Specific primers were used for tight junction-related genes (JAM-A, Claudin-1, Claudin-3, Claudin-7, and occludin) and pro-inflammatory cytokines (IL-1β, IL-6, IL-8, and TNF-α). Gene expression levels were normalized to the housekeeping gene β-actin, and relative quantification was calculated using the comparative Ct method (2^−ΔΔCt). The sequences of all primers used in this study are listed in Table 1 [18].

Table 1. Primers used in this study.

Gene	Forward primer (5′ → 3′)	Reverse primer (5′ → 3′)
Housekeeping gene
β-actin	CTG GAA CGG TGA AGG TGA CA	AAG GGT CTT CTC TAA CAT GCA
Inflammatory cytokine genes
IL-1β	AGG CTG CCG GGA CTC ACA GCA	TGA GGC CCA AGG CCA CAG GT
IL-6	GCA CTG GCA GAA AAC AAC CT	TCA AAC TCC AAA GAC CAG TGA
IL-8	CTC TTG GCA GCC TTC CTG ATT	ACT CTC AAT CAC TCT CAG TTC T
TNFα	CCC AGG GAC CTC TCT CTA ATC	ATG GGC TAC AGG CTT GTC ACT
Tight junction & adhesion genes
JAM A	GTG CCT TCA GCA ACT TCT CC	GAG CCG ATA TCC GTT TGG TC
Claudin 1	TTT CCT GCT ACA ACA ATC CTC TCC	GTT GTT TTT CGG GGA CAG GAA AC
Claudin 3	TGT GCT GTT CCT TCT GCG CCC	CTT AGA CGA AGT CCA TGG CGG
Claudin 7	CTG CAA AAT GTA CGA CTC GGT G	GCA AGA CCT GCC ACG ATG AAA A
Occludin	CCC ATC TGA CTA TGT GGA AAG A	AAA CCC GTT GTC ATT CAC TTT G

Protein sample preparation and proteomic analysis

For proteomic analysis, HT29 cells were seeded in 100 mm dishes at a density of 1.0 × 10⁶ cells/dish and cultured to 80% confluence. HT29 cells were subjected to the same treatment schedule used for gene expression analysis (6 hr of bacterial pretreatment followed by 24 hr of LPS co-treatment). Cells were harvested immediately after treatment for proteomic analysis. Protein samples were then extracted using 0.5% sodium dodecanoate in 100 mM Tris-HCl (pH 8.5) with sonication (Bioruptor UCD-200, Sonic Bio Co., Ltd., Kanagawa, Japan) for 15 min. Protein concentrations were determined using a BCA assay kit (Thermo Fisher Scientific) and normalized to 150 ng/µL. For digestion, 3 µg of protein was reduced with 10 mM dithiothreitol (50°C, 30 min), alkylated with 30 mM iodoacetamide (room temperature, 30 min in dark), and quenched with 60 mM cysteine (10 min). The digested samples were then diluted in 50 mM Tris-HCl (pH 8.0) and digested overnight at 37°C using 500 ng Trypsin/Lys-C mix (Promega, Madison, WI, USA). Following acidification with 5% trifluoroacetic acid and desalting using C18-StageTips, the peptides were dried and redissolved in 3% acetonitrile/0.1% trifluoroacetic acid.

Peptides were analyzed via nano LC-MS/MS using an UltiMate 3000 RSLCnano system coupled to an Orbitrap Fusion Lumos mass spectrometer (Thermo Fisher Scientific). Separation was performed on a 75 μm × 30 cm nano-capillary column at 50°C with a 90-min gradient at a rate of 200 nL/min. Data-independent acquisition (DIA) was performed using overlapping windows (495–785 m/z for MS1, >200 m/z for MS2) at a resolution of 30,000 with optimized collision energies.

Data processing was conducted using Scaffold DIA (Proteome Software) with a human spectral library generated by Prosit. The search parameters included trypsin digestion with ≤1 missed cleavage, 10 ppm mass tolerance, and cysteine carbamidomethylation. Protein identification required a <1% false discovery rate for both peptides and proteins. Quantification utilized the EncyclopeDIA algorithm selecting the four highest-quality fragment ions per peptide, with statistical analysis performed using Perseus v1.6.15.0.

Functional enrichment and pathway analysis

To elucidate the biological significance of the identified protein changes and their associated pathways, the proteomic datasets were further analyzed using IPA (QIAGEN). The protein accession numbers, fold changes, and p-values calculated via Perseus were uploaded to the IPA software. Canonical Pathway Analysis was performed to identify significantly enriched biological processes, with a specific focus on “Tight Junction Signaling”.

Statistical analysis

All experiments were performed using at least three independent biological replicates, each with technical triplicates unless otherwise stated. Data are presented as the mean ± standard error of the mean (SEM). Statistical analysis was performed using Tukey’s HSD multiple comparison test. All statistical analyses were conducted using the GraphPad Prism software (version 9.0, GraphPad Software, Boston, MA, USA).

RESULTS

Analysis of TEER and LY permeability

To investigate the protective effects of two heat-treated P-LAB strains against lipopolysaccharide (LPS)-induced epithelial barrier dysfunction, we measured TEER, an indicator of barrier integrity, and LY permeability, an indicator of paracellular transport, using Caco-2 and HT29 cells.

LPS stimulation (1 ng/mL, 24 hr) significantly reduced TEER by approximately 35% in both Caco-2 and HT29 cells (Fig. 1A, 1B). Consistent with this, LY permeability significantly increased in both cell lines, confirming LPS-induced barrier dysfunction (Fig. 1C, 1D). Pretreatment with each heat-treated P-LAB strain (10⁸ cells/mL) exhibited protective effects against this LPS-induced disruption; however, the efficacy varied depending on the evaluation metric and cell line. In Caco-2 cells, only L. plantarum LOC1 significantly inhibited the LPS-induced TEER reduction, restoring values to near-control levels. In contrast, L. paracasei RO24 showed no significant improvement in TEER in Caco-2 cells. Regarding LY permeability, both strains significantly suppressed the LPS-induced increase in both Caco-2 and HT29 cells. Tukey’s HSD multiple comparison test revealed no statistically significant difference between the protective effects of L. plantarum LOC1 and L. paracasei RO24 on LY permeability in either cell line.

Fig. 1.

Changes of the transepithelial electrical resistance (TEER) and Lucifer Yellow (LY) permeability in Caco-2 and HT29 cells following lipopolysaccharide (LPS) stimulation alone or pretreatment with two heat-treated plant-derived lactic acid bacteria (P-LAB) strains, L. plantarum LOC1 and L. paracasei RO24. TEER was measured as an indicator of barrier integrity (A, B), and LY permeability was assessed as an indicator of paracellular transport (C, D). Statistical analysis was performed using Tukey’s HSD multiple comparison test. Data are presented as the mean ± standard error of the mean. *p<0.05 vs. control; #p<0.05 vs. LPS group. (A, C) Caco-2 cells; (B, D) HT29 cells.

Expression analysis of inflammation-related factors and tight junction-related genes

To elucidate the molecular basis of the barrier-protective effects, we measured the mRNA expression levels of pro-inflammatory cytokines (IL-1β, IL-6, IL-8, and TNF-α) and tight junction-related proteins (JAM-A, Claudin-1, Claudin-3, Claudin-7, occludin) using real-time PCR.

LPS stimulation markedly increased the mRNA expression of all tested pro-inflammatory cytokines in both cell lines (Fig. 2A, 2B). While L. plantarum LOC1 extensively suppressed the induction of these cytokines in both lines, L. paracasei RO24 also showed generally similar anti-inflammatory trends, although its inhibitory effects were weaker and did not consistently reach statistical significance. Specifically, L. paracasei RO24 failed to significantly suppress IL-8 expression in Caco-2 cells, and no significant inhibitory effects were observed for IL-6 and TNF-α in HT29 cells. These findings suggest that the anti-inflammatory effects of L. paracasei RO24 were more modest and less statistically robust than those observed for L. plantarum LOC1.

Fig. 2.

Changes of the mRNA expression of key pro-inflammatory cytokines (IL-1β, IL-6, IL-8, TNF-α) and tight junction-related genes (JAM-A, Claudin-1, Claudin-3, Claudin-7, occludin) in Caco-2 and HT29 cells following lipopolysaccharide (LPS) stimulation alone or pretreatment with two heat-treated plant-derived lactic acid bacteria (P-LAB) strains, L. plantarum LOC1 and L. paracasei RO24. Statistical analysis was performed using Tukey’s HSD multiple comparison test. Data are presented as the mean ± standard error of the mean. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001. (A) Caco-2 cells; (B) HT29 cells.

Furthermore, clear strain-specific differences were observed in the gene expression profiles of tight junction-related genes (Fig. 2A, 2B). In Caco-2 cells, L. plantarum LOC1 significantly upregulated the expression of JAM-A and occludin, whereas L. paracasei RO24 exhibited no significant effect on these specific genes, which aligns with the TEER results (Fig. 2A). However, it is noteworthy that even in Caco-2 cells, L. paracasei RO24 showed a tendency to induce higher expression levels of the claudin family (Claudin-1, Claudin-3, Claudin-7) compared with L. plantarum LOC1. The potent impact of both strains on the claudin family was similarly observed in HT29 cells (Fig. 2B). In contrast, no significant differences were found between the two strains regarding the expression of JAM-A or occludin in this cell line.

Analysis of mechanisms via proteomics and IPA

To globally validate the phenotypic and genetic findings at the protein level, DIA proteomic analysis was performed using HT29 cells. A heat map summarizing relative protein abundance changes across the four experimental groups is shown in Supplementary Fig. 1. The heat map includes representative tight junction-related proteins (JAM-A, Claudin family proteins, OCLN, TJP family proteins, and CGN), together with signaling molecules associated with inflammation (NF-κB, TNF, EMT, and MYLK), cytoskeletal regulation (RAC1, CDC42, TIAM1, and CTNNB1), and epithelial polarity/barrier regulation (PALS1, PTEN, and SMURF1).

LPS treatment reduced the abundance of multiple tight junction-related proteins, including JAM-A, occludin, TJP family proteins, and Claudin family proteins, while simultaneously increasing inflammatory and barrier-disruptive signaling-related factors such as NF-κB, TNF, EMT-related proteins, and MYLK. Treatment with L. plantarum LOC1 or L. paracasei RO24 partially restored their protein abundance profiles in a strain-dependent manner.

Representative tight junction-related proteins corresponding to the targets analyzed in the gene expression study are presented in Fig. 3, enabling direct comparison between mRNA- and protein-level responses. Detailed proteomic analysis revealed strain-specific regulatory patterns. L. plantarum LOC1 preferentially maintained the abundances of JAM-A and occludin, whereas L. paracasei RO24 showed relatively stronger preservation of Claudin family proteins (Fig. 4). In addition, broader proteomic trends observed in Supplementary Fig. 1 suggested that L. plantarum LOC1 more extensively suppressed inflammation-associated signaling proteins, while L. paracasei RO24 preferentially preserved structural barrier-related proteins.

Fig. 3.

Relative protein abundance of representative tight junction-related proteins corresponding to the targets analyzed in the gene expression study (JAM-A, Claudin-1, Claudin-3, Claudin-7, and occludin). Protein levels were quantified by DIA-MS-based proteomic analysis in HT29 cells following LPS stimulation alone or pretreatment with two heat-treated plant-derived lactic acid bacteria (P-LAB) strains, L. plantarum LOC1 and L. paracasei RO24. Statistical analysis was performed using Tukey’s HSD multiple comparison test. Data are presented as the mean ± standard error of the mean. *p<0.05, **p<0.01, ***p<0.001, ****p<0.0001.

Fig. 4.

IPA-based prediction of tight junction signaling pathway activity in HT29 cells. Ingenuity Pathway Analysis (IPA) was used to evaluate predicted activation states of the “Tight Junction Signaling” pathway based on proteomic data. (A) LPS-treated cells compared with untreated control, indicating predicted pathway suppression. (B) LPS + L. plantarum LOC1 group compared with LPS alone, indicating predicted restoration of pathway activity. (C) LPS + L. paracasei RO24 group compared with LPS alone, indicating predicted restoration of pathway activity. The interpretation of colors, relationship lines, and predicted activation/inhibition states is summarized in the prediction legend included within the figure. Predictions are based on coordinated protein abundance changes identified in the proteomic dataset.

Furthermore, correlation analysis between the relative mRNA expression levels (log2 fold change) and protein abundances (log2 fold change) of major tight junction-related proteins demonstrated significant positive correlation (R²=0.4669, p=0.005; Supplementary Fig. 2). This result supports the general consistency between transcriptomic and proteomic trends observed in the present study.

To further interpret these coordinated protein abundance changes at the pathway level, IPA was performed. IPA indicated that the “Tight Junction Signaling” pathway showed a negative z-score under LPS treatment, suggesting predicted suppression relative to the untreated control. This pathway shifted toward positive z-scores following treatment with either L. plantarum LOC1 or L. paracasei RO24, suggesting partial restoration of barrier-related signaling (Fig. 4A–4C).

In addition, upstream regulator analysis suggested reduced activation of inflammatory signaling networks associated with NF-κB and cytokine-related responses in the L. plantarum LOC1-treated group compared with the LPS group. These IPA predictions were derived from coordinated changes in the DIA proteomic dataset rather than direct measurement of pathway activity.

DISCUSSION

In addition to its fundamental roles in digestion and nutrient absorption, the intestinal tract functions as a critical biological defense system by acting as both a physical and immunological barrier [1]. This intestinal barrier function is primarily maintained by tight junctions formed between adjacent intestinal epithelial cells. Disruption of this barrier leads to increased intestinal permeability, a condition commonly referred to as “leaky gut”. Leaky gut has been implicated in a wide range of diseases, including IBD and IBS, positioning the maintenance and restoration of intestinal barrier integrity as an important therapeutic target [19,20,21].

In recent years, increasing attention has been directed toward heat-treated bacterial cells, generally defined as non-viable bacterial cells or their cellular components that may confer physiological benefits [22, 23]. Compared with live probiotics, heat-treated bacterial cells offer advantages in terms of safety and storage stability, while retaining the ability to induce host responses through bacterial structural components. Under certain conditions, heat-treated bacterial cells have been reported to exhibit functional effects comparable to, or even exceeding, those of live bacteria [24, 25]. In this study, we investigated the strain-specific effects of heat-treated P-LAB cells on LPS-induced intestinal epithelial barrier dysfunction, with a particular focus on differences in tight junction modulation between L. plantarum LOC1 and L. paracasei RO24.

LPS is known to impair intestinal barrier function through a dual mechanism involving activation of inflammatory signaling pathways and disruption of tight junction architecture in intestinal epithelial cells [26, 27]. Consistent with previous reports, LPS stimulation in the present study significantly upregulated the expression of major pro-inflammatory cytokine genes, including IL-1b, IL-6, IL-8, and TNF-a. These findings are in agreement with established evidence that intestinal epithelial cells recognize LPS as a pathogen-associated molecular pattern, leading to activation of inflammatory cascades centered on NF-kB signaling [28, 29]. Such inflammatory responses have been shown to induce downregulation and mislocalization of tight junction-related proteins, thereby contributing to increased epithelial permeability [30].

Indeed, LPS stimulation in our model resulted in reduced expression of key tight junction components, including JAM-A, occludin, and TJP-1. These molecules are essential for maintaining intercellular adhesion and regulating paracellular permeability [31,32,33,34], and their downregulation is considered a molecular hallmark of leaky gut pathophysiology [35, 36]. Taken together, these observations indicate that the experimental model employed in this study appropriately reflects both inflammatory activation and tight junction disruption, validating its use as an in vitro intestinal barrier dysfunction model.

TEER analysis revealed that both L. plantarum LOC1 and L. paracasei RO24 exerted protective effects against LPS-induced barrier impairment; however, these effects were clearly cell line-dependent. L. paracasei RO24 significantly suppressed LPS-induced TEER reduction in HT29 cells but failed to show a comparable effect in Caco-2 cells. This discrepancy can be explained by differences in tight junction-related protein dependency between the two cell lines. Caco-2 cells exhibit a highly differentiated epithelial phenotype and maintain robust barrier function through a dense tight junction network primarily composed of JAM-A and occludin [7, 37]. In our gene expression analyses, L. plantarum LOC1 broadly preserved the expression of these tight junction components, whereas the effects of L. paracasei RO24 were limited. This likely accounts for the insufficient barrier recovery observed with L. paracasei RO24 in Caco-2 cells.

In contrast, barrier function in HT29 cells is known to rely heavily on Claudin family proteins. Epithelial electrical resistance and ion selectivity are largely determined by the specific combination of Claudin subtypes expressed [34]. In HT29 cells in particular, Claudin-1 has been identified as a key regulator of cellular phenotype and b-catenin signaling [38], while Claudin-3 and Claudin-7 play crucial roles in cell adhesion and barrier formation [39, 40]. In the present study, L. paracasei RO24 maintained the expression of these Claudin proteins in HT29 cells at levels comparable to, or exceeding, those observed with L. plantarum LOC1. This Claudin-focused mode of action likely underlies the selective barrier-protective effect of L. paracasei RO24 in HT29 cells.

Proteomic analysis combined with IPA-based integrative pathway analysis further clarified the distinct action strategies of the two strains. L. plantarum LOC1 was predicted by IPA to attenuate several upstream inflammatory signaling pathways activated by LPS, including networks associated with NF-kB and pro-inflammatory cytokine signaling. These findings should be interpreted as pathway-level predictions derived from coordinated proteomic changes rather than direct measurement of TNF-a or IL-6 protein abundance [28, 41]. Such comprehensive anti-inflammatory activity may function as a “preventive” barrier protection strategy by averting widespread disruption of tight junction components. The present study demonstrated clear strain-specific differences in barrier-protective and molecular responses between L. plantarum LOC1 and L. paracasei RO24. Although these strains were originally isolated from distinct plant-associated environments, the current dataset does not allow conclusions regarding whether the observed functional differences are attributable to isolation source, species background, or other strain-specific genetic factors. Therefore, any relationship between ecological origin and functional phenotype should be considered hypothetical at this stage. Future studies including multiple strains from diverse sources within comparable taxonomic backgrounds will be required to evaluate this possibility.

In contrast, L. paracasei RO24 showed a tendency to preferentially support the maintenance and reinforcement of specific tight junction-related proteins, especially members of the Claudin family, rather than broadly suppressing inflammatory signaling. These strain-specific and structure-dependent modes of action underscore the importance of selecting appropriate bacterial strains based on the molecular pathology underlying intestinal barrier dysfunction.

Furthermore, by integrating real-time PCR-based mRNA analysis with DIA-MS-based proteomic profiling, we demonstrated a correlation (R2=0.4669) between transcript and protein levels for key tight junction-related factors. This finding supports the notion that the barrier-protective effects of heat-treated P-LAB strains are not attributable to passive physical shielding but instead involve active biological responses regulated at both transcriptional and translational levels within host cells [9, 42]. The signaling alterations identified by IPA provide molecular-level support for this coordinated host response and extend the analysis beyond descriptive observations [43].

Collectively, the strain-specific and tight junction component-dependent barrier-protective effects demonstrated in this study can be understood as functional characteristics of heat-treated bacterial cells that do not rely on bacterial colonization or proliferation. The observation that distinct strains selectively target different tight junction disruption patterns supports the validity of a precision heat-treated bacterial cell framework, in which heat-treated bacterial cells are applied in a pathology-dependent manner [44].

This study is based on in vitro analyses and does not fully recapitulate the complex in vivo environment, including the intestinal microbiota and mucosal immune system. Validation of these findings in appropriate animal models, including FITC-dextran permeability assays and tissue-level tight junction analyses, will be an important next step toward confirming the physiological relevance of the strain-specific protective effects observed in this study. Moreover, the present study employed an LPS-induced barrier dysfunction model as a standardized inflammatory stimulus. Although this model enables reproducible evaluations of barrier impairment, other inflammatory stimuli such as TNF-a or IL-6 may induce partially distinct signaling pathways and epithelial responses. Future studies using cytokine-mediated barrier dysfunction models would therefore be valuable to broaden the biological relevance of the present findings. Additionally, only a single bacterial concentration was evaluated in the present study. Although this design was suitable for direct strain comparison under standardized conditions, future dose-response studies would provide further mechanistic insight. Furthermore, only heat-treated bacterial cells were examined in the present study, and direct comparison with viable cells of the same strains was not performed. Therefore, the current findings should be interpreted as evidence of biological activity of heat-treated bacterial cells under the present experimental conditions rather than proof of effects unique to heat-treated bacterial cells. Future studies directly comparing viable and heat-treated forms will be important to clarify the contribution of bacterial viability. Nevertheless, our results provide important guidance for future in vivo studies and the development of functional materials by emphasizing strain selection strategies tailored to specific tight junction impairment patterns. Future studies involving immunofluorescence-based localization analysis of tight junction proteins will be valuable to further confirm whether the transcriptomic and proteomic preservation observed in this study is accompanied by structural maintenance at the cellular level.

DATA AVAILABILITY STATEMENT

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

AUTHORS’ CONTRIBUTIONS

YT drafted and edited the manuscript. JY and IS conceived the study and collected information. All authors contributed to the article and approved the submitted version.

CONFLICT OF INTEREST

The authors declare that the research was conducted in the absence of any commercial or financial relationships that could be construed as a potential conflict of interest.