This study was carried out in accordance with the recommendations of Nottingham Research. Ethics Committee 2 with written informed consent from all subjects. All subjects gave written informed consent in accordance with the Declaration of Helsinki. The protocol was approved by the Nottingham Research Ethics Committee 2 (REC Reference 08/H0408/167).

Multiple sclerosis patients attending outpatient clinics at Nottingham University Hospitals were recruited for this study. The study included 35 adult MS patients with clinically definite MS, according to the McDonald Criteria (27), aged 17–57 years (mean: 40.45 ± 10.6). 18 patients had not been treated with any disease-modifying treatments (DMTs), 6 were taking interferon-β, 4 were on Glatiramer acetate, 3 were on Fingolimod, 2 were receiving Dimethyl fumarate, and 2 were receiving corticosteroids. In addition, 25 HC aged 26–58 years (mean: 40.73 ± 10.2) were recruited. A relapse was defined as patient-reported symptoms or objectively observed signs typical of an acute inflammatory demyelinating event in the CNS, lasting at least 24 h (27). HC had no history of autoimmune disease or recent symptomatic infections. There was no significant age difference between MS patients and HC (P = 0.153). All MS patients and HC gave written informed consent prior to blood and urine sampling. The study was approved by the Nottingham Research Ethics Committee and by Nottingham University Hospitals National Health Service Trust Research and Innovation Services. Age and sex-matched samples for the validation cohort were received from University Hospital San Luigi Gonzaga, Orbassano, Turin, Italy which included 20 adult RRMS patients aged between 21 and 55 years (mean 33.35 ± 10.74) and 20 HC aged between 22 and 57 years (mean 33.6 ± 10.97). All the samples in the validation cohort were treatment naïve and there was no age difference between RRMS and HC (P = 0.942). Both patients and controls were Caucasians, mainly from the Piedmont region. Demographic data for the two cohorts are presented in Table 1.

Soluble TLR2 and sTLR4 were measured in serum and urine samples by ELISA (DTLR20, R&D Systems, UK, and SEA753Hu; Cloud-Clone Corp., USA, respectively). For the validation cohort sTLR2 was measured using a different ELISA kit (SEA753Hu; Cloud-Clone Corp., USA). CRP was measured in serum samples by high-sensitivity ELISA (DY1707, R&D Systems).

TLR2 stimulants in urine and serum samples from MS patients and HC were measured using a human embryonic kidney (HEK)-293 cell TLR-transfection assay, as described previously (28). Briefly, cells were transiently transfected with TLR2, CD14, nuclear factor kappa-light-chain-enhancer of activated B cells (NF-κB)-reporter, and thymidine-kinase promoter-driven reporter constructs, then challenged 3 days later with indicated concentrations of defined TLR-ligands, volunteer serums, heat-killed urine samples, or heat-killed bacteria. 18 h later, NF-κB-dependent reporter expression was measured by luminometry and normalized to co-transfected renilla reporter. A standard curve was prepared on each plate by plotting fold NF-κB induction vs concentration using the defined TLR2-ligand Pam₃CSK₄. Because it has been proposed that diverse ligands beyond lipopeptides may stimulate TLR2, the TLR2-stimulating capacity of the samples are presented as biological activities relative to this standard curve and are expressed as ng/ml Pam₃CSK₄-equivalents or bacterial lipopeptide equivalent (BLP) equivalents.

A selection of eight representative UTI-related Gram-positive and Gram-negative bacteria was tested using the same HEK-293-TLR2 transfection system to determine their capacity to stimulate TLR2. All bacteria examined were of hazard group 2 or lower and were studied in a class 2 containment facility. A whole bacterial cell lysate was also prepared from H. pylori (SS1 strain) for assay in the same system. To prepare the lysate, H. pylori cultures from 10 blood agar base 2 plates (Oxoid) cultured for 24 h under microaerobic conditions at 37°C were harvested into ice-cold sterile phosphate-buffered saline. This suspension was then disrupted using a Soniprep 150 sonicator (SANYO, Watford, UK), with 6 × 10s bursts at amplitude of 10 µm, then aliquoted and stored at −20°C. There were approximately 6 × 10⁹ bacteria per ml before sonication, corresponding to 1 × 10⁷ bacteria per ml when diluted, and the protein content of the lysate was 13.8 mg/ml as determined using the BCA protein assay kit (Thermo). The TLR2-stimulating capacities of bacteria and lysates in these experiments are presented as NF-κB fold induction vs cells cultured in medium alone.

Urine samples were tested using Combur-Test Strip dipsticks (Cobas, Roche, Switzerland). Positive results for leukocytes, nitrite, protein, and erythrocytes were considered potential signs of infection.

Flow cytometric analyses were carried out using frozen peripheral blood mononuclear cells (PBMC) from RRMS patients collected during relapse and remission. After thawing, the cells were cultured for 5 h (short-term stimulation) with phorbol 12-myristate 13-acetate (0.1 µg/ml), ionomycin (1 µg/ml), and Brefeldin-A (10 µg/ml, added only in the last hour). After stimulation and washing, cells were stained with Live/Dead^® Fixable Blue Dead Cell Stain Kit (ThermoFisher Scientific) for 30 min at room temperature. Cells were then stained extracellularly with anti-CD4 (BD Pharmingen) and anti-TLR2 (eBioscience, Clone TL2.1). Cells were acquired using an LSR II flow cytometer (BD Biosciences), collecting a minimum of 50,000 events in each sample, and analyzed using Kaluza software (Beckman Coulter, version 1.5). Gating strategy ungated PBMCs were plotted on SSC/FSC plot and then lymphocytes were gated by excluding dead cells/debris. Then, from the lymphocytes, live cells were gated. Then, the percentage of live lymphocytes that were CD4+/TLR2+ were selected.

Peripheral blood mononuclear cells from healthy volunteers were treated with 4 µg/ml of Calpain (Natural human Calpain 1 protein, AbCam, Product no. ab91019) for 1 h at 37°C and then stained with anti-CD4 and anti-TLR2 antibodies and analyzed by flow cytometry as above. Gating strategy ungated PBMCs were plotted on SSC/FSC plot and then lymphocytes were gated by excluding dead cells/debris. Then, from the lymphocytes, CD4+ cells were gated and then percentage of TLR2+ cells from CD4+ population were identified.

The data are presented as mean (±SD or SEM). All statistical analyses were performed using Prism 7.01 (GraphPad Software, San Diego, CA, USA). Data which were not normally distributed were log transformed before comparison of groups using the Mann–Whitney U test and Wilcoxon matched-pairs signed rank test (for paired samples). p Values <0.05 were considered significant.

As it has been reported that serum sTLR2 is a biomarker of recent infection, and of disease activity in other conditions, we measured sTLR2 levels in serum samples from MS patients in relapse and remission and compared them with those from HC (29, 30). We observed significantly higher levels of sTLR2 in the serum samples of MS patients both during relapse (0.87 ng/ml) and remission (0.85 ng/ml) when compared with HC (0.58 ng/ml) (Figure 1A). However, there was no significant difference in the sTLR2 values between paired MS relapse and remission samples (n = 14, Figure 1B). Since, both TLR2 and TLR4 expression has been reported to be elevated in the PBMCs from MS patients and soluble forms for both of these receptors (sTLR2 and sTLR4) were detected in various infectious and non-infectious inflammatory conditions, we also measured sTLR4 levels in the serum samples (31, 32). We found that sTLR4 levels were not significantly different between groups (1.52, 2.14, and 1.69 ng/ml) for HC, MS relapse, and MS remission, respectively (Figures 1C,D). Serum samples were also tested for the presence of TLR2 stimulants, but these were below the limit of detection by the HEK293-TLR2 transfection assay (data not shown).

Similar to sTLR2, we found significantly higher levels of serum CRP during MS relapse compared to HC, although, there was no significant difference between paired relapse and remission samples (Figures 1E,F). This raised the possibility that sTLR2 may be increased in relation to the acute phase reaction (APR). However, there was no significant correlation between serum levels of CRP and sTLR2 or sTLR4, and an in vitro model of the APR (interleukin-1β treated hepatocarcinoma cell line cells), revealed that sTLR2 is not released by stimulated hepatocytes (data not shown). Together, these findings suggest that the observed increase in serum sTLR2 in MS patients is likely not directly caused by the APR.

To validate the sTLR2 elevation in the serum of RRMS patients, we performed ELISA tests in the sera of an independent cohort of patient and HC from a different population (Piedmont, North-Western Italy). There was a similar upward trend of mean sTLR2 levels in the serum samples of RRMS patients (0.907 ng/ml) compared to HC (0.836 ng/ml) and the values were almost in the same concentration range as in the discovery data set although it did not reach statistical significance (P = 0.6061) (Figures 2A,B). We, therefore, pooled the data from our original discovery cohort and the validation cohort to see the overall trend. In the pooled data, sTLR2 levels were 0.888 and 0.696 ng/ml in the RRMS (n = 41) and HC (n = 44), respectively, and the analysis retained the original significance (P = 0.0006, Mann–Whitney test) (Figure 2C). We also measured the levels of sTLR4 in the serum samples from the validation cohort and the trend was the same as observed in our discovery cohort, with values that were very low or below the detection limit (data not shown).

As we reported recently that Tregs from RRMS patients are more susceptible to TLR2-stimulation than those of HC (19), we explored whether CD4+ T cells may have potential to contribute to the circulating pool of sTLR2. The percentage of CD4+ T cells surface expression of TLR2 was not significantly different between relapse and remission (Figures 3A–C). However, previous studies have shown that extracellular calpain, a cysteine protease released by T lymphocytes, has the ability to cleave TLR2, leading to increased levels of sTLR2 in the extracellular environment both in mice and humans (26). Treatment with exogenous calpain led to a non-significant reduction in TLR2 surface expression on CD4+ T cells (data not shown). However, we were unable to confirm this trend by ELISA, as sTLR2 levels were below the limit of detection in the supernatants of T cells cultured in the presence of calpain (data not shown).

Soluble TLR2 may be released in response to excessive TLR2-stimulation (25) and so we sought evidence of potential sources of TLR2-stimulants that may be of relevance to MS. As UTI is a common complication of MS (33), we used urine dipstick tests to seek evidence of UTI in MS patients. Although levels of protein and erythrocytes were significantly increased in urine of MS patients compared to HC (Table 2), there were no significant differences in overall positivity for markers of infection between the groups, using this method. To gain further insight into whether MS patients may be exposed to increased TLR2-stimulants as a result of subclinical UTI not detectable by dipstick, we used a HEK293-TLR2 transfection assay calibrated with the synthetic ligand Pam₃CSK₄ to quantify the biological activities of TLR2-stimulants present in urine of HC and MS patients, as described previously (28). These measurements revealed that TLR2-stimulant activities were far greater in urine samples of MS patients compared to HC (1.72, 126.2, and 55.5 ng/ml), relative to Pam₃CSK₄-equivalent activity, for HC, MS relapse, and MS remission, respectively [P = 0.0004, P = 0.0020 (Figures 4A,B)].

Soluble TLR2 and sTLR4 levels were also measured in urine samples of HC and MS patients. Urinary sTLR2 levels were higher than measured in serum, but unlike serum sTLR2 levels, there were no statistically significant differences between HC and MS groups (Figures 4C,D). sTLR4 levels were below the limit of detection in most urine samples and the mean values were also not significantly different between the groups (Figures 4E,F).

These findings suggested potential involvement of UTI-related bacteria in the stimulation of TLR2 in MS patients. To explore which types of bacteria may contribute to such activity, a panel of representative UTI-relevant bacteria was tested for their ability to trigger TLR2-signaling using the HEK293-TLR2 transfection system. All the Gram-negative organisms tested (Escherichia coli, Klebsiella pneumoniae, Pseudomonas aeruginosa, and Proteus vulgaris) were potent stimulators of TLR2-signaling (Figure 5A). However, while two of the Gram-positive organisms stimulated strong TLR2-signaling (Staphylococcus aureus and Staphylococcus epidermidis), two induced little or no TLR2-signaling (Enterococcus faecalis and Streptococcus pyogenes). As infection with the gastric pathogen H. pylori has also been associated with MS (34), the TLR2-stimulating capacity of H. pylori lysate was assessed, confirming that this organism also contains abundant TLR2-stimulants (Figure 5B). Thus, diverse bacteria responsible for infections which may be relevant to MS, of both Gram-positive and Gram-negative origin, are potent triggers of TLR2 signaling.