Enteropathogenic infections modulate intestinal serotonin transporter (SERT) function by activating Toll-like receptor 2 (TLR-2) in Crohn’s disease

Serotonin (5-hydroxytryptamine [5-HT]) is an intestinal neuromodulator that regulates several essential enteric physiological functions such as absorption or secretion of fluids, and peristaltic reflexes. Availability of the intestinal 5-HT is dependent on serotonin transporter (SERT), which uptakes 5-HT and facilitates its degradation. Interestingly, Toll-like receptor 2 (TLR-2) is co-localized with 5-HT, which suggests a possible impact of neuroendocrine cells in the inflammatory response through TLR-2 activation. Serum 5-HT levels were measured in 80 Crohn’s disease (CD) patients and 40 healthy control subjects. Additionally, fully differentiated Caco-2 monolayers were infected with Mycobacteria paratuberculosis (MAP), L. monocytogenes, or M. smegmatis in the presence of exogenous 5-HT at different concentrations. Cells were subsequently harvested and used for measuring SERT activity, RNA isolation followed by RT-PCR, protein quantification, and tissue damage markers (DHE, LDH, GSH and MDA). TLR-2 intracellular signaling pathways were assessed by pre-incubating Caco-2 monolayers with selective blockers of ERK, cAMP/PKA, p38 MAPK, and 5-HT3 receptors. MAP-infected CD patients (N = 40) had higher serum 5-HT levels (462.95 ± 8.55 ng/mL, N = 40) than those without MAP infection (385.33 ± 10.3 ng/mL, N = 40). TLR-2 activation by enteropathogenic bacteria inhibited SERT activity in the presence of exogenous 5-HT by up to 50%. These effects were increasing gradually over 72 h, and MAP infection had the greatest effect on SERT inhibition when cells were exposed to 5-HT in a concentration dependent manner. Additionally, inhibition of SERT activity was accompanied with higher levels of pro-inflammatory cytokines (TNF-α, IL-6, IL-8) and oxidative stress markers (DHE, LDH and MDA), whereas SERT expression and protein level were downregulated. Consequently, inhibition of TLR-2 and p38 MAPK signaling or blocking 5-HT3 receptors restored SERT activity and reduced the production of pro-inflammatory cytokines, as reflected by the downregulation of oxidative stress and tissue damage markers. The involvement of TLR-2 in the intestinal pathology might be concluded not only from its innate immune role, but also from its effect on modulating the intestinal serotonergic response. Ultimately, regulating the intestinal serotonergic system can be therapeutically exploited to mitigate other enteropathogenic infections, which will help in understanding the gut-microbiome-brain connection.


Higher serum 5-HT level in MAP-infected CD patients compared to CD patients without MAP
infection or healthy controls. We measured 5-HT serum level in 40 healthy control subjects, 40 MAP positive CD patients and 40 MAP negative CD patients. Our data indicate a significant higher serum level of 5-HT in CD patients (N = 80) in comparison to healthy controls (N = 40) [424.14 ± 9.42 ng/mL vs. 211.69 ± 8.11 ng/mL, respectively]. Moreover, CD patients who were MAP infected had a significantly higher serum level of 5-HT (462.95 ± 8.55 ng/mL), when compared to MAP negative CD patients (385.33 ± 10.3 ng/mL) (Fig. 1).

MAP infection has the highest effect on inhibiting SERT activity in the presence of exogenous 5-HT.
To test the direct effect of bacterial infection on SERT activity, fully differentiated Caco-2 monolayers . Values were pre-tested for normal distribution using the Kolmogorov-Smirnov normality test. Significance among experiments was assessed by two-way analysis of variance (ANOVA) followed by Bonferroni correction test. *P value < 0.05 compared to healthy controls. # P value < 0.05 compared to healthy controls or CD MAP negative patients. www.nature.com/scientificreports/ were infected with MAP, L. monocytogenes and M. smegmatis in the presence of different levels of 5-HT (0, 100, 250, 500 ng/mL). SERT activity was measured at three time points every 24 h and values were normalized to control group without infection or 5-HT exposure. We noticed that MAP infection had the greatest significant effect on reducing SERT activity when cells were exposed to 250 ng/mL (78.3 ± 3.2%, 69.5 ± 1.7% and 65.2 ± 1.4% at 24, 48 and 72 h, respectively) and 500 ng/mL (71.4 ± 1.5%, 64.3 ± 2.1% and 54.7 ± 1.5% at 24, 48 and 72 h, respectively) of 5-HT ( Fig. 2A). Caco-2 monolayers infected with L. monocytogenes showed significant reduction of SERT activity when cells were exposed to 500 ng/mL for 48 and 72 h (75.3 ± 1.6% and 69.8 ± 1.8% at 48 and 72 h, respectively) (Fig. 2B). However, M. smegmatis infection did not have a significant effect on SERT activity (Fig. 2C). Our results indicate that exposure to 5-HT in all tested bacterial infections upregulates TLR-2 and pro-inflammatory cytokines expression and protein level in a concentration dependent manner (Fig. 3). Consequently, stimulating inflammation resulted in downregulation of SERT and IL-10 expression. These results were significant only in the presence of 5-HT at 500 ng/mL in all groups including cells exposed to 5-HT without bacterial infection, whereas MAP infection induced inflammation and downregulated SERT significantly at 100, 250 and 500 ng/mL of 5-HT treatment (Table 1). Specifically, in Caco-2 monolayers exposed to 500 ng/mL of 5-HT, expression and protein level of TNF-α, IL-6 and IL-8 were about 3 folds higher in MAP-infected monolayers compared to control cells without infection or 5-HT treatment. In contrast, expression and protein level of SERT and IL-10 were progressively decreasing significantly following MAP infection in monolayers exposed to 5-HT at 100, 250 and 500 ng/ mL. Since MAP infection caused the highest level of inflammation and inhibition of SERT expression in Caco-2 monolayers exposed to an excess level of 5-HT, we further performed fluorescence staining assay for SERT detection in the presence of 5-HT at 500 ng/mL following 24 h of MAP infection. Similarly, 5-HT exposure in MAP infected monolayers showed the highest effect on reducing SERT fluorescence in comparison to MAP infection alone or to 5-HT without MAP infection, which indicates that 5-HT further reduces SERT localization caused by MAP infection (Fig. 4).  5C). Addition of 5-HT to the bacterial infection progressively augmented LDH and MDA levels while decreasing GSH activity in a concentration dependent manner, and the result was significant in all tested groups at 500 ng/mL. Since MAP infection caused the highest level of oxidative stress in Caco-2 monolayers exposed to an excess level of 5-HT, we further per-  , TLR-2 (E) and SERT (F) in Caco-2 monolayers exposed to exogenous 5-HT levels (0, 100, 250 and 500 ng/ mL) for 24 h. Values were pre-tested for normal distribution using the Kolmogorov-Smirnov normality test.

MAP infection induces the highest level of oxidative stress and tissue damage in
Significance among experiments was assessed by two-way analysis of variance (ANOVA) followed by Bonferroni correction test. All RT-PCR experiments were performed in triplicates. Data are presented as Mean ± SD. *P value < 0.05. **P value < 0.001. www.nature.com/scientificreports/ or to 5-HT without MAP infection, which indicates that 5-HT increases oxidative stress level caused by MAP infection (Fig. 6).

Inhibition of 5-HT 3 receptors or p38 MAPK pathway reduce inflammation and oxidative stress by restoring SERT activity in MAP-infected Caco-2 monolayers.
In order to further analyze the effects of TLR-2 activation on modulating SERT activity and inflammation, we used specific inhibitors of intracellular pathways involved in TLR-2 signaling, including TLR-2 inhibitor [MMG 11 (5 ug (1 uM)]. Additionally, we tested the anti-inflammatory effects of 5-HT 3 receptor antagonist [ondansetron (40 ng/mL)] to show that excess 5-HT is required for sustained inflammatory process caused by MAP infection. In comparison to MAP-infected Caco-2 monolayers exposed to 500 ng/mL 5-HT for 24 h, inhibition of TLR-2 lead to the highest significant decrease in TNF-α, IL-6 and IL-8 (by 2.26, 2.18 and 2.17 folds, respectively), while it significantly restored expression of IL-10 and SERT (2.58 and 2.06 folds, respectively) ( Fig. 7A). Similarly, inhibition of p38 MAPK and blocking 5-HT 3 receptors by ondansetron had a significant effect on downregulating the expression of TNF-α, IL-6 and IL-8 by nearly 2 folds, while these treatments significantly restored expression of IL-10 and SERT (Fig. 7A). In contrast, inhibition of ERK and cAMP/PKA pathways did not have a significant effect on modulating cytokines or SERT expression. The results of modulating gene expression were correspondent with the quantified protein level ( Table 2). As shown in Table 3, significant reduction in LDH activity and MDA level, in addition to higher GSH and SERT activities were noticed following inhibition of TLR-2 and p38 MAPK or by blocking 5-HT 3 receptors using physiological level (40 ng/mL) of ondansetron.
Moreover, since the phosphorylated form (p-p38 MAPK) mediates p38 MAPK activity, we measured the ratio of p-p38 MAPK/p38 MAPK pathway following MAP infection in the presence of different levels of 5-HT (0, 100, 250, 500 ng/mL). The results obtained have suggested that MAP infection without 5-HT treatment increases the ratio of p-p38 MAPK/p38 MAPK by about two folds, while 5-HT progressively increases phosphorylation of p38 MAPK following MAP infection in a concentration dependent manner, reaching about four folds higher when cells were exposed to 500 ng/mL (Fig. 7B).

Discussion
The intestinal mucosal barrier selectively absorbs essential nutrients while preventing the entry of harmful toxins and pathobionts 21 . Moreover, the intestinal epithelium participates in an active immune response by recognizing specific microorganism-associated molecular patterns (MAMPs) through TLRs, which carefully distinguishes commensal from pathogenic bacteria 22 . In this context, TLR-2 plays an integral role in intestinal homeostasis, by recognizing Gram-positive bacteria and mycobacterial lipopeptide 23 . Heterodimeric expression of TLR1 or TLR6 is required for TLR-2 activation, which generates a signal to tolerate or eradicate bacteria, depending on its pathogenicity 22 . Therefore, dysregulation in TLR-2 activity has been implicated in several GI conditions such as IBD and IBS [24][25][26][27][28] .
Recent studies have demonstrated that TLR-2 signaling pathway regulates the intestinal serotonergic system 29 . The presence of 5-HT in the GI tract plays a major role in gut motility, neuronal reflexes and fluid secretion. Table 1. Effects of enteropathogenic bacterial infections on protein level of TNF-α, IL-6, IL-10, IL-8, TLR-2 and SERT in Caco-2 monolayers exposed to exogenous 5-HT levels (0, 100, 250 and 500 ng/mL) for 24 h. Values were pre-tested for normal distribution using the Kolmogorov-Smirnov normality test. Significance among experiments was assessed by two-way analysis of variance (ANOVA) followed by Bonferroni correction test. All ELISA experiments were performed in triplicates. *P value < 0.05.

5-HT concentration (ng/mL)
TNF-α ± SD (pg/mL) IL-6 ± SD (pg/mL) IL-8 ± SD (pg/mL) IL-10 ± SD (pg/mL) TLR-2 ± SD (ng/mL) SERT ± SD (pg/mL) www.nature.com/scientificreports/ A compelling evidence has suggested that 5-HT is involved in CD pathogenesis, which is characterized by the prevalence of enterochromaffin cells, higher 5-HT mucosal content, increased activity of tryptophan hydroxylase I (TPH-1), and more importantly, lower expression of SERT, a critical regulator of the intestinal 5-HT availability 10 . In the present study, we have confirmed that serum 5-HT level is significantly higher among CD patients in comparison to healthy subjects, and MAP-infected CD patients had a higher level of 5-HT than those without MAP infection. We demonstrated that TLR-2 activation in fully differentiated Caco-2 monolayers by enteropathogenic bacteria (MAP, L. monocytogenes and M. smegmatis) inhibits SERT activity in the presence of exogenous 5-HT. These effects increased gradually over 72 h, and MAP infection had the greatest effect on SERT inhibition when www.nature.com/scientificreports/ cells were exposed to 5-HT in a concentration dependent manner. Additionally, inhibition of SERT activity was accompanied with higher level of pro-inflammatory cytokines (TNF-α, IL-6, IL-8) and oxidative stress markers (LDH and MDA), whereas SERT expression and protein level were downregulated. These results indicate that TLR-2 activation may repress the intestinal SERT function and availability, leading to excessive accumulation of extracellular 5-HT in the GI tract. Consequently, inhibition of TLR-2 restored SERT activity and reduced the production of pro-inflammatory cytokines, which was reflected by the decrease in oxidative stress and tissue damage markers. Recently, it has been shown that 5-HT accumulation activates colonic nicotinamide adenine dinucleotide phosphate (NADPH) oxidase, which generates reactive oxygen species (ROS) 30 . Furthermore, we previously reported that MAPinfected CD patients had a higher level of glutathione peroxidase (GPx) activity, indicating higher oxidative stress level 31 . Therefore, accumulation of extracellular 5-HT in the presence of MAP infection further induced the pro-inflammatory process and caused higher level of tissue damage. Ultimately, this could affect the intestinal tight junctions that are essential in regulating the physical intestinal barrier, leading to paracellular movement of 5-HT across the intestinal epithelium to the blood vessels, where it can trigger complications in multiple organ-systems (Fig. 8).
In order to investigate the underlying mechanism by which TLR-2 activation inhibits SERT activity, we used specific inhibitors of intracellular pathways involved in TLR-2 signaling. Recent studies have shown that cell signaling associated ERK pathway modulates TLR-2 signaling in macrophages and neuronal cells 32,33 . In addition, the cAMP/PKA pathway has been involved in the TLR-2 immune response 34 . However, we noticed that inhibition of the ERK and cAMP/PKA pathways did not have a significant effect on modulating pro-inflammatory cytokine release or restoring SERT activity. In contrast, inhibition of the p38 MAPK pathway had a significant effect on downregulating inflammation and restoring SERT activity. The results obtained have suggested that 5-HT progressively increases the phosphorylation of MAPK following MAP infection in a concentration dependent manner, which was demonstrated by increasing the ratio of p-p38 MAPK/p38 MAPK. The role of p38 MAPK signaling in modulating SERT function is controversial, since p38 MAPK activation does not have an effect on SERT activity in the brain, but it enhances SERT activity in neuronal cells 35,36 . Our results agree with two previous studies, showing that p38 MAPK activation inhibits SERT activity in the intestinal epithelial tissue 29,37 .
The presence of 5-HT 3 receptors in the intestinal mucosa mediates a variety of GI processes, including motility and fluids secretion 5 . Therefore, 5-HT 3 antagonists effectively counteract chemotherapy-induced emesis 6 . Since 5-HT 3 receptors are expressed in macrophages and T-cells, new reports have shed light on additional potential anti-inflammatory applications for this class of drugs, which confirms the role of these receptors in inflammatory and immune responses 8,9 . In this study, we blocked 5-HT 3 receptors by using ondansetron, which significantly suppressed MAP-induced inflammation and mitigated the effects of had a significant effect of excess 5-HT exposure. Furthermore, this treatment significantly restored SERT expression and function. Collectively, these results indicate that using 5-HT 3 receptor antagonists may have a potential role in suppressing the intestinal inflammatory response to pathogenic bacteria.
It is worth mentioning that higher SERT expression was found in the human small intestine compared with the colon 38 . Interestingly, SERT expression demonstrates remarkable variation across different regions of the human small intestine, with highest expression in the ileum, followed by the duodenum and the jejunum 38 . Low SERT expression in the colon may point out an alternate mechanism for 5-HT uptake by different monoamine transporters, including norepinephrine transporter (NET), organic cation transporter (OCT), and dopamine transporter (DAT) 38 . Whether the accumulation of 5-HT in the colonic mucosa induces expression of SERT or other monoamine transporters in the colonic epithelial cells, needs further investigation. www.nature.com/scientificreports/ Since SERT is dependent on the membrane potential created by the sodium-potassium adenosine triphosphatase (Na + /K + -ATPase) to function properly, increasing the activity of this enzyme in the intestines could serve as a potential pharmacological target when SERT activity is attenuated. Despite the myriad advantages provided by Caco-2 cell culture, it does not always recapitulate the normal physiology and lineage development of the native human intestinal epithelium. Therefore, exploring SERT activity in 3-dimensional organotypic intestinal epithelium matrix will be useful to further understand 5-HT's involvement in the intestinal pathophysiology. www.nature.com/scientificreports/ In summary, our results demonstrate that TLR-2 activation by enteropathogenic infections downregulate SERT expression and function leading to reduction in 5-HT reuptake. As a result, excessive accumulation of intestinal 5-HT induces inflammation and causes tissue damage. These effects are ostensibly mediated by the p38 MAPK signaling pathway and the activation of 5-HT 3 receptors, since inhibiting p38 MAPK pathway or blocking 5-HT 3 receptors restored SERT activity and reduced inflammation. This indicates that TLR-2 signaling modulates the innate and the serotonergic responses, thereby directing the course of intestinal pathologies. Ultimately, regulating the intestinal serotonergic system could provide therapeutic approaches in enteropathogenic infections, which will help in understanding the gut-microbiome-brain connection. www.nature.com/scientificreports/

Materials and methods
Clinical samples. We previously collected serum from peripheral blood samples (4.0 mL K 2 -EDTA tube) of 100 CD patients (CDAI ≥ 220 and ≤ 450) and 40 healthy control subjects acquired from Dr. Naser's laboratory at the University of Central of Florida. MAP infection status was determined by IS900 PCR as described earlier 39 , and then we randomly selected 40 MAP positive and 40 MAP negative CD patients for this study. Clinical serum 5-HT levels were determined by serotonin ELISA assay (Eagle Biosciences, Amherst, NH). The average age of CD patients was 45.6 ± 12.4, and the average age for healthy controls was 25 ± 5.1, respectively. The study was approved by the University of Central Florida Institutional Review Board #IRB00001138. Each subject completed and signed a written consent form before samples were collected.
Cell culture. The in vitro part of this study was carried out in the human enterocyte-like cell line Caco-2 (ATCC HTB-37). This cell line has been described as an excellent intestinal model to study enteropathogenic infections and numerous GI disorders. Caco-2 cells (Passage #5) were cultured in ATCC-formulated Eagle's minimum essential medium supplemented with 20% fetal bovine serum and maintained at 37 °C in a humidified 5% CO 2 incubator. Cells were grown in 12-well plates at a density of 4 × 10 4 cells per well until confluency and differentiation have been reached.  Table 2. Effects of several inhibitors involved in TLR-2 signaling pathway on protein level of TNF-α, IL-6, IL-8, IL-10 and SERT in MAP-infected Caco-2 monolayers in the presence of excess level (500 ng/mL) of 5-HT following 24 h of incubation. Values were pre-tested for normal distribution using the Kolmogorov-Smirnov normality test. Significance among experiments was assessed by two-way analysis of variance (ANOVA) followed by Bonferroni correction test. All ELISA experiments were performed in triplicates. *P value < 0.05.

Treatment
TNF-α ± SD (pg/mL) IL-6 ± SD (pg/mL) IL-8 ± SD (pg/mL) IL-10 ± SD (pg/mL) SERT ± SD (pg/mL)  Table 3. Effects of several inhibitors involved in TLR-2 signaling pathway on the activities of SERT, LDH, GSH and MDA level in MAP-infected Caco-2 monolayers in the presence of excess level (500 ng/mL) of 5-HT following 24 h of incubation. Values were pre-tested for normal distribution using the Kolmogorov-Smirnov normality test. Significance among experiments was assessed by two-way analysis of variance (ANOVA) followed by Bonferroni correction test. All SERT, LDH, GSH, and MDA activity experiments were performed in triplicates. *P value < 0.05.

Treatment SERT activity ± SD (% of control) LDH activity ± SD (% of control) GSH activity ± SD (% of control) MDA level ± SD (uM)
Control (MAP + 500 ng/mL 5-HT) 71. www.nature.com/scientificreports/ according to manufacturer's protocol. The assay utilizes a fluorescent substrate that mimics the biogenic 5-HT and is taken into the cell through SERT, resulting in increased fluorescence intensity. SERT activity values were presented as percentage activity of control group without bacterial infection or 5-HT treatment.

RNA extraction, reverse transcription and real-time PCR.
Following each specific infection/ treatment, total RNA was extracted and purified from Caco-2 monolayers using TRIzol isolation protocol (Abcam, Cambridge, MA). Thermal cycler (MyGene Series Peltier) was used for cDNA synthesis as described previously 40   www.nature.com/scientificreports/ CA) and TLR2 [RayBiotech, Peachtree Corners, GA)] following manufacturer's instructions. All ELISA experiments were performed in triplicates, and the absorbance was read at 450 nm wavelength.

Tissue damage and oxidative stress measurement. Markers of oxidative stress in fully differentiated
Caco-2 monolayers including lactate dehydrogenase (LDH), glutathione (GSH), and malondialdehyde (MDA) were tested following each specific infection/treatment by using specific assay kits (ThermoFisher, Waltham, MA) per manufacturer's protocol as we described earlier 41 .
Treatment with TLR-2, ERK, p38MAPK and cAMP/PKA inhibitors. To confirm the role of TLR-2 activation in downregulating SERT activity following bacterial infection, we blocked TLR-2 receptors using selective human TLR-2 antagonist (Novus Biologicals, Littleton, CO) to rule out the role of other TLRs in regulating SERT function. Caco-2 monolayers were pre-incubated with 5 ug/mL of TLR2 inhibitor for 30 min prior to MAP infection and 5-HT treatment. We also assessed TLR-2 intracellular signaling pathways in depth by preincubating Caco-2 monolayers with selective blockers of ERK ( Statistical analysis. Statistical analysis was performed using GraphPad Prism 7.02 software. All data collected in this study were pre-tested for normal distribution using the Kolmogorov-Smirnov normality test. Significance among experiments was assessed by two-way analysis of variance (ANOVA) followed by Bonferroni correction test. For experiments performed at 3 time points over 72 h, significance was determined by repeated measures ANOVA. P value < 0.05 and a 95% confidence interval (CI) were used for the assessment of differences in all experiments. P values < 0.001 were also mentioned when achieved. Data were presented as Mean ± SD.
Patient consent for publication. Obtained.
Ethics approval. University of Central Florida Institutional Review Board (IRB00001138).

Data availability
All experiments were performed in accordance with relevant guidelines and regulations. Raw data is available upon request.