As hypothesised, this study showed that L. plantarum MB452 altered the expression levels of tight junction-related genes in healthy intestinal epithelial cells. Of the tight junction bridging proteins, occludin mRNA abundance was higher in the presence of L. plantarum MB452. The over-expression of the occludin protein has been linked to increased TEER , and based on the findings of this study, increased occludin gene expression may contribute to the ability of L. plantarum MB452 to enhance tight junction integrity. In support of this, genes encoding for the occludin-associated plaque proteins, ZO-1 and ZO-2 and cingulin, also had increased expression levels in the presence of L. plantarum MB452. The zonula occludens bind to the cytoplasmic end of occludin and form the scaffolding to link occludin to the actin cytoskeleton . The increased expression levels of the occludin, ZO-1, ZO-2 and cingulin genes appeared to correlate with increased intensity of their proteins as indicated by the fluorescent microscopy images.
In contrast, other genes that had increased transcript levels in the presence of L. plantarum MB452 are known to be involved in tight junction disassembly. The gene encoding ITCH, an ubiquitin-ligase molecule, had increased expression levels in the presence of L. plantarum MB452; however, the ITCH protein is known to contribute to the degradation of occludin . The increased expression of the ITCH gene may lead to an increase in the turnover of occludin protein and, therefore, may have contributed to the increased occludin mRNA noted in this data. The gene encoding the SNAI1 protein also had increased expression in the presence of L. plantarum MB452; however, the SNAI1 protein is known to bind to occludin and claudin genes promoters suppressing their expression . Although these two genes, ITCH and SNAI1, have been linked to tight junction disassembly, 17 out of the 19 tight junction-related genes with increased expression levels in response to L. plantarum MB452 exposure contribute to tight junction stability; therefore, the cumulative effect would most likely be enhanced intestinal barrier function.
The 'tightness' of tight junctions is commonly thought to be, at least partly, due to claudins, which are a set of bridging proteins; however, none of the claudin genes were differentially expressed in response to L. plantarum MB452. Decreases in the abundance of claudin-2, -3 and -4 proteins (measured using western blotting) have been associated with a decrease in TEER . Another study showed that a decrease in TEER was associated with altered cellular localisation of claudin-1 and -5, but not altered abundance , so it is possible that L. plantarum MB452 may have altered the distribution of claudin proteins without changing gene expression and/or protein abundance.
The results of this study showed that L. plantarum MB452 enhanced the expression of 19 genes involved in the tight junction signalling pathway in healthy cells. A previous study showed that L. plantarum CGMCC 1258 is able to protect against the disruption of four tight junction proteins caused by Enteroinvasive E. coli ATCC 43893 (serotype O124:NM) . However, another study looking at the effect of L. plantarum ATCC202195 on the expression of genes in Caco-2 cells challenged with Enteroinvasive E. coli ATCC43893 (serotype O124:NM) did not report any changes in tight junction gene expression . This suggests that the L. plantarum protection against tight junction disruption was not due to it altering host gene expression, and was likely due to it inhibiting the action of the pathogen in that study.
The ability to enhance the expression of tight junction-related genes is not common to all L. plantarum strains. In addition to the study that showed that L. plantarum ATCC202195 nullifies changes in Caco-2 cell gene expression induced by Enteroinvasive E. coli ATCC43893 (serotype O124:NM) , other published studies investigating whole genome expression in response to L. plantarum have shown that, L. plantarum WCFSI induces increased expression of genes involved in lipid metabolism and cellular growth and development in healthy human duodenum  and L. plantarum (strain not given) alters the NF-κB pathway to limit inflammatory responses in healthy human duodenum . However, in these published studies only a few tight junction-related genes had altered expression levels when exposed to L. plantarum, for example increased expression of the ZO-2 gene, so they are unlikely to contribute to changes in tight junction integrity, compared to the changes in 19 tight junction genes induced by L. plantarum MB452 reported in this study. This is not surprising since strains of L. plantarum can have differing effects on intestinal barrier function in vitro, from neutral (cause no increase in TEER) to beneficial (cause substantial increase in TEER; unpublished results), and thus, it is likely that different strains may also have different effects on epithelial cell gene expression.
The observed increase in intestinal barrier function induced by L. plantarum MB452 may also be, at least partly, due to changes in intestinal epithelial cell gene expression that have an indirect effect on tight junction stability. Eight genes encoding for tubulins had lower expression levels in response to L. plantarum MB452. A high turnover in tubulin synthesis has been linked to the disassembly of tight junctions ; thus, the reduced expression levels of these genes may account for the positive effect of L. plantarum MB452 on intestinal barrier function. Similarly, seven genes encoding for proteasome subunits had lower mRNA abundance in the presence of L. plantarum MB452. Proteasomes, which are large protein complexes responsible for breaking down surplus or damaged proteins, have previously been linked to tight junction degradation, and proteasome inhibitors can prevent degradation of occludin  and ZO-2 . The reduction in proteasome gene expression induced by L. plantarum MB452 may be an additional mechanism by which tight junction integrity is enhanced.
Several of the tight junction-related genes with altered expression induced by L. plantarum MB452 may also be involved in reducing cell proliferation. For example, ZO-1, which had increased gene expression in the presence of L. plantarum MB452, is a 'dual location protein' involved in the regulation of cell proliferation. The ZO-1 protein binds to the CSDA protein (also known as ZONAB) and sequesters it to tight junctions, and removal of the CSDA protein from nucleus in this way results in a reduction in the CDK4 protein . Therefore, an increase in ZO-1 gene expression may lead to a decrease in CDK4 gene expression as seen here (Figure 3), which highlights the link between the formation of tight junctions and a reduction in cell proliferation . Additionally, L. plantarum MB452 reduced the expression of the CPSF2 gene, which encodes a protein which is part of the CSTF-CPSF-SYMPK complex, that regulates cell-cycle related gene expression and promotes cell proliferation . Together with the decreased expression of tubulin genes, these effects of L. plantarum MB452 on the ZO-1, CDK4 and CPSF2 genes may lead to decreased cell proliferation and contribute to the reported anti-proliferative effect of the VSL#3 product .
L. plantarum MB452 did not alter the expression levels of other genes and pathways that have been affected by some probiotic bacteria, such as the NF-κB pathway , PPARγ [40, 41], innate immune response pathway , or human β defensin-2 . This indicates that, unlike some other probiotic bacteria, L. plantarum MB452 does not seem to exert its beneficial effect by regulating host immune responses in healthy intestinal cells.
In this study using L. plantarum MB452 alone, only certain effects previously associated with VSL#3 were observed. VSL#3 is a mixture of L. plantarum, L. casei, L. acidophilus, L. delbrueckii subspecies bulgaricus, B. longum, B. breve, B. infantis and Streptococcus thermophilus, and is likely that each bacterial species has a range of effects. A previous study indicated that of the bacterial strains present in VSL#3, the culture supernatant of B. infantis was associated with the greatest increase in TEER across Caco-2 cells compared to untreated controls . Of the VSL#3 lactobacilli, L. plantarum MB452 produced the supernatant with the greatest effect of TEER, which is in agreement with previous work that showed the beneficial effects of L. plantarum MB452 supernatant . Other studies indicated that the anti-inflammatory effects of VSL#3 are, at least partially, due to VSL#3 bifidobacteria decreasing the abundance of the pro-inflammatory cytokine IL-8  and L. casei in VSL#3 reducing the abundance the pro-inflammatory cytokine interferon gamma-induced protein 10 . The genes encoding for these cytokines were not altered in response to L. plantarum MB452 in the present study.