We recently demonstrated better outcomes in helminth-infected multiple sclerosis (MS) patients, compared with uninfected ones. The present study evaluates the role of TLR2 and retinoic acid (RA) in parasite-driven protection in MS patients. RA serum levels were significantly higher in helminth-infected MS patients than in uninfected MS subjects or healthy controls. Genes involved in RA biosynthesis and metabolism, such as Adh1 and Raldh2, as well as RA receptors and IL-10, were induced in dendritic cells (DCs) via TLR2-dependent ERK signaling. This programmed DCs to induce FOXP3+ T regulatory cells and suppressed production of proinflammatory cytokines (IL-6, IL-12, IL-23, and TNF-α) via induction of suppressor of cytokine signaling 3 (SOCS3), an effect mediated by soluble egg Ag (SEA) obtained from Schistosoma mansoni, and by RA. SEA-activated DCs also inhibited IL-17 and IFN-γ production through autoreactive T cells. These inhibitory effects were abrogated when SOCS3 gene expression was silenced, indicating that SEA-mediated signaling inhibited production of these cytokines by T cells, through a SOCS3-dependent pathway. Overall, helminth-related immunomodulation observed in MS patients was mediated by TLR2- and RA-dependent pathways, through two different mechanisms, as follows: 1) induction of IL-10 and FOXP3+ T regulatory cells, and 2) suppression of proinflammatory cytokine production mediated by SOCS3.
Multiple sclerosis (MS) is an inflammatory demyelinating disease affecting the CNS. Although its etiology remains unknown, several lines of evidence indicate autoimmunity plays a major role in disease susceptibility and development (1). Autoimmune diseases are currently considered to be the result of both genetic susceptibility and environmental factors (2, 3). The genetic component of MS is thought to derive from the presence of common allelic variants in several genes (2). However, discordance of MS between monozygotic twins suggests additional factors, such as environmental modulators, are also involved (4).
Epidemiological data provide strong evidence of a steady rise in MS incidence in developed countries in recent decades, concomitant with decreased incidence of many infectious diseases resulting from antibiotic use, vaccination, or improved hygiene and socioeconomic conditions (5). Supporting these findings, recent investigations demonstrate a dichotomous relationship between regional distribution of MS and that of the parasite Trichuris trichiura, a common human pathogen. MS prevalence appears to fall steeply once a critical threshold of T. trichiura prevalence (∼10%) is exceeded (6). In line with these observations, pre-established Schistosoma mansoni infection in mice or pretreatment with S. mansoni OVA significantly reduces incidence, delays onset, and attenuates clinical course of experimental autoimmune encephalomyelitis (EAE) in mice (7, 8). Likewise, we recently demonstrated that helminth-infected MS patients showed significantly lower number of relapses, minimal changes in disability scores, and significantly lower lesion activity on magnetic resonance imaging (MRI) compared with uninfected individuals with MS (9). Extended follow-up also showed that, after antihelminth drug administration, clinical and radiological activity increased to levels observed in uninfected MS subjects (10).
Retinoic acid (RA) is an active metabolite of vitamin A (also known as retinol), which has multiple isoforms; however, the all-trans isoform predominates in most tissues and is known to be important in a variety of physiological processes, including immunological responses (11). RA mediates these activities by binding to nuclear retinoid receptors, members of the steroid hormone receptor superfamily, and altering transcription activity. Two families of receptors have been identified, RA receptors (RAR) and retinoid X receptors. Each receptor family has at least three subtypes (α, β, and γ), which act as activated ligand transcription factors controlling the expression of a number of target genes (12, 13). In recent years, RA has received particular attention as it has been shown to influence multiple immune cell lineages and to modulate a broad range of immune processes, such as lymphocyte activation and proliferation, Th cell differentiation, tissue-specific lymphocyte homing, and production of specific Ab isotypes (11, 14, 15).
Different combinations of TLRs are expressed on many cells of the immune system. Helminth parasite infections may alter the regulation, function, and levels of expression of these receptors. Our group has provided evidence indicating that surface expression of TLR2 on both B cells and dendritic cells (DCs) is significantly higher in helminth-infected MS patients. Exposure of either cell population to soluble egg Ag (SEA) obtained from S. mansoni resulted in significant upregulation of TLR2 in helminth-infected MS patients, but not in uninfected individuals (16). In different experimental settings, some authors have found similar results, indicating higher expression of TLR2 on PBMCs and enhanced response to TLR2 ligands in parasite-infected individuals, compared with uninfected counterparts (17, 18). However, recent studies using DCs from TLR2−/− and TLR4−/− mice demonstrated that TLR2 and TLR4 were not required for SEA-pulsed DCs to induce anti-inflammatory responses in naive mice, suggesting other receptors, such as C-type lectins, might be implicated in DC response to SEA (19). Discrepancies between these investigations may be due to the presence of different moieties in SEA preparations or, alternatively, to differences in parasite Ag recognition mechanisms between human and mouse TLR2 (20).
We present results from several studies specifically conducted to evaluate the role of RA on modulation of various immunological pathways, under helminth-driven protection observed in MS patients.
Materials and Methods
Fifteen patients (10 females and 5 males) with clinical diagnosis of MS according to Poser's criteria and presenting eosinophilia were assessed in a prospective double-cohort study, as previously reported elsewhere (9). Intestinal parasites were present in stool samples from all patients and identified as the cause of eosinophilia. Three patients were infected with Hymenolepis nana, three with T. trichiura, three with Ascaris lumbricoides, three with S. mansoni, two with Strongyloides stercolaris, and one with Enterobius vermicularis. Presence of other endemic parasitoses, including trypanosomiasis, leishmaniasis, amebiasis, giardiasis, and toxoplasmosis, was ruled out in this study population using microscopic stool examination and serological tests. Mean patients’ age was 32.9 ± 7.4 y, Kurtzke expanded disability status scale score was 2.5 ± 0.6, and mean disease duration 8.7 ± 1.9 y. MS diagnosis preceded parasitic intestinal infection diagnosis by 28.3 ± 6.1 mo (range 20–41 months). Fifteen healthy subjects, 15 MS subjects in remission without helminth, and 12 helminth-infected subjects without MS, matched for age, sex, and disease duration, served as controls. Healthy individuals were recruited among family members of helminth-infected MS patients. Prior clinical and neurologic examination as well as standard and hematological laboratory tests ruled out presence of any other underlying conditions. Uninfected MS control subjects included in this study presented clinical course profiles similar to those of other MS patients at our center. Both eosinophil counts and stool examination performed in uninfected MS subjects and healthy controls were negative at study entry, and remained so throughout the duration of the study, indicating these individuals were not asymptomatically infected. Helminths infecting subjects without MS included the following: T. trichiura, three cases; A. lumbricoides, three cases; S. stercolaris, two cases; and S. mansoni, four cases.
No patients had received steroids for at least 3 mo prior to study entry, or other immunomodulatory, immunosuppressive drugs or dietary supplements. Because none of the patients developed clinical disease as a result of the parasite infection, no antiparasite therapy had been prescribed at study entry. After 60–78 mo of follow-up, six patients required antiparasite treatment for helminth symptom exacerbation (fever, abdominal pain, anorexia, weight loss, general malaise, diarrhea, or anemia). None of the patients had clinical evidence of extraintestinal disease (e.g., pulmonary involvement). Two patients infected with A. lumbricoides and two patients infected with T. trichiura received a single dose of albendazole (400 mg), whereas two patients infected with S. stercolaris were treated with two doses of ivermectine (0.2 mg/kg) separated by 2-wk interval.
Cerebrospinal fluid (CSF) samples were available from eight helminth-infected MS patients. None had received antiparasite treatment prior to CSF collection. Ten individuals (7 women and 3 men, mean age 34.7 ± 6.8 y) admitted for extra spinal orthopedic surgery served as controls. In this group, CSF samples were obtained through lumbar tap draw prior to spinal anesthesia injection. A second control group consisting of 10 patients (7 women and 3 men, mean age 37.5 ± 7.9 y) suffering from other inflammatory neurologic diseases (OIND) is as follows: 5 cases of viral encephalitis, 2 cases of aseptic meningitis, and 3 cases of CNS vasculitis were also included.
This study was approved by the Institutional Ethics Committee of the Raúl Carrea Institute for Neurological Research, and participants gave written informed consent.
RA serum level assay
RA serum levels were measured using commercially available ELISA kits (MyBioSource, San Diego, CA), according to manufacturer instructions. Assay sensitivity level was 0.4 ng/ml. Intra- and interassay variation coefficients were 5.1 and 4.8%, respectively.
Myelin basic protein (MBP)83–102, MBP143–168, and myelin oligodendrocyte glycoprotein (MOG)63–87 peptides were synthesized with an automated peptide synthesizer using FAST-MOC chemistry, and expected peptide amino acid composition confirmed by HPLC.
The SEA was prepared aseptically, as previously described elsewhere (21), and used at a concentration of 50 μg/ml.
Isolation of DCs and generation of monocyte-derived DCs
PBMCs were isolated from heparinized venous blood through Ficoll-Hypaque (Pharmacia LKB Biotechnology, Piscataway, NJ) density gradient centrifugation, and for some experiments DCs were purified with a blood isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany), in accordance with manufacturer instructions to a purity of 95–98%. Because of the large number of DCs required to perform the experiments described in this investigation (quantities not readily obtainable through venipuncture or leukapheresis), in vitro DC cultures were established from monocytes, generating sufficient cell numbers. Monocyte-derived DCs were therefore generated from PBMCs, as previously described (22). Briefly, monocytes were selected from PBMCs using anti-CD14–coated magnetic beads (Invitrogen, Carlsbad, CA) and cultured in the presence of 100 ng/ml GM-CSF and 50 ng/ml IL-4 (R&D Systems, Minneapolis, MN). On days 2 and 4, half the culture medium was replaced, maintaining the same IL-4 and GM-CSF concentrations. After 5 d in culture, 100 ng/ml LPS was added to stimulate maturation. In some experiments, maturation was achieved using 2 μg/ml CD40L (Amgen, Thousand Oaks, CA). Cells were collected on day 7, and, prior to use in experiments, extensively washed in PBS to remove exogenously added cytokines or residual LPS. CD3+ T lymphocyte contamination was <0.3%. Although monocyte-derived DCs may not be fully representative of their ex vivo isolated counterparts, control experiments performed using both cell populations showed monocyte-derived DCs maintained specific imprinting capacity and characteristics similar to ex vivo isolated DCs, leading us to believe that results observed were not the product of in vitro culture manipulation during DC isolation.
Generation of MBP and MOG peptide-reactive T cell lines
MBP and MOG peptide-specific T cell lines (TCLs) were expanded from peripheral blood, as previously described elsewhere (23). Briefly, 5 × 106 PBMCs were stimulated with optimal concentrations of MBP or MOG peptides (10 μg/ml). After 5–7 d, cells were recultured in fresh medium containing 50 U/ml human rIL-2 (rhIL-2; R&D Systems, Minneapolis, MN) for an additional week. Restimulation cycles were repeated weekly using autologous irradiated PBMCs (3000 rads) as APCs, plus peptide, followed by expansion with rhIL-2. After four cycles of restimulation and expansion, TCLs were evaluated using standard proliferation assays. Cutoff values for positive response were set at a stimulation index >3. All MBP- and MOG-reactive TCLs were >93% CD4+.
Expression of alcohol dehydrogenase 1, retinol dehydrogenase 2, and retinoic receptors on DCs
For quantitative assessment of relative mRNA levels, total RNA was isolated from DCs using TRIzol LS reagent (Invitrogen) following manufacturer instructions. RNA was reverse transcribed using a Moloney murine leukemia virus reverse-transcription kit with random hexamer primers (Invitrogen). Relative levels of alcohol dehydrogenase (Adh1), retinol dehydrogenase (Raldh2), retinoic receptor α, and retinoic receptor γ mRNA were determined by real-time PCR, on an ABI 7000 sequence detection system (Applied Biosystems, Foster City, CA). Values obtained were normalized to the amount of GAPDH. Primer sequences used were as follows: GAPDH, sense 5′-GAAGGTGAAGTCGGAGTC-3′, antisense 5′-GAAGATGGTGATGGGATTTC-3′; Adh1, sense 5′-AACTCCTCTCACTGCTCTCCAC-3′, antisense 5′-GTCACCCTCTTCAGATTGCTTTTCC-3′; Raldh2, sense 5′-TACTTCAGAACGGGAATGACAAAC-3′, antisense 5′-CCAGTGATGTAGGTAAATAAGATAGAGGG-3′; RARα, sense 5′- GGTCGGCGATGGTGAGGGT-3′, antisense 5′-TGGGCAAATACACTACGAACAACAG-3′; RARγ, sense 5′-CGCCGAAGCATCCAGAAGAAC-3′, antisense 5′-GGTGACAGGGTCGTTCGGCG-3′.
Induction of CD4+CD25+FOXP3+ regulatory T cells
To determine the role of SEA and RA in CD4+CD25+FOXP3+ T regulatory (Treg) induction from naive cells, CD4+ T cells were isolated from PBMCs using a CD4+ T cell isolation kit (Miltenyi Biotec). CD25− cells were then isolated by negative selection with CD25 microbeads (Miltenyi Biotec). Purity was determined to be >98% for CD4+CD25− T cells. Next, 5 × 103 DCs were stimulated with SEA (50 μg/ml), and 12 h later, washed and cocultured at 1:10 ratio (DCs to T cells) together with 5 × 104 autologous naive CD4+CD25− T cells. Cultures were stimulated with soluble anti-human CD3 mAb and soluble anti-CD28 Abs (BD Biosciences, San Jose, CA), both at 5 μg/ml, in the presence of rhIL-2 (400 U/ml; R&D Systems) and human rTGF-β (5 ng/ml; R&D Systems), in the presence and absence of retinol (500 nM; Sigma-Aldrich, St. Louis, MO). After 5 d, the number of CD4+CD25+FOXP3+ Treg was evaluated by flow cytometry, using commercially available regulatory T cell staining kits, following manufacturer instructions (eBioscience, San Diego, CA).
Measurement of cytokine production
MBP and MOG peptide-specific TCLs were stimulated with the cognate Ag, and with DCs previously incubated with SEA, or SEA plus retinol during 12 h. Forty-eight hours later, supernatants were collected, and IL-17 and IFN-γ production was assessed by ELISA using commercially available kits (R&D Systems) and following manufacturer instructions. To measure LPS- or SEA-stimulated DC cytokine production, cells were cultured in the presence and in the absence of retinol. Supernatants were collected 48 h poststimulation, and IL-6, IL-10, IL-12 p70, IL-23, IL-27, and TNF-α were measured using commercially available ELISA kits, purchased from R&D Systems following manufacturer instructions.
Small interfering RNA technique
TLR, SOCS1, and SOCS3 expression were silenced using the small interfering RNA (siRNA) technique. DCs were transfected with 25 nM siRNA for TLR2, TLR5, TLR9, SOCS1, and SOCS3, or with scrambled nonsilencing control oligonucleotides using TransIT-TKO siRNA transfection reagent (Mirus Bio, Madison, WI), according to manufacturer instructions. The sequence of siRNA strands used to silence the TLR2 gene was sense 5′-GCACUUUAUAUUCACUUACtt-3′, antisense 5′-GUAAGUGAAUAUAAAGUGCtc-3′; for TLR5 was sense 5′-GGAGCAAUUUCCAACUUAUtt-3′, antisense 5′-AUAAGUUGGAAAUUGCUCCtt-3′; for TLR9 was sense 5′-GACGGCAACUGUUAUUACAtt-3′, antisense 5′-UGUAAUAACAGUUGCCGUCtt-3′; for SOCS3 was sense, 5′-CCAAGAACCUGCGCAUCCAdTdT-3′, antisense, 5′-UGGAUGCGCAGGUUCUUGGdTdT-3′; and for SOCS1 was sense 5′-GCAUCCGCGUGCACUUUCAUU-3′, antisense 5′-AAUGAAAGUGCACGCGGAUGC-3′.
Evaluation of MAPK signaling
p38, ERK1/2, and JNK1/2 activities were evaluated using commercially available ELISA kits (Invitrogen). Briefly, DCs (1.5–2 × 106) were collected on day 7 and cultured for pre-established periods of time with different stimuli. ELISA tests were performed following the manufacturer instructions. For inhibition studies, DCs were incubated with the specific ERK1/2 inhibitor, U0126 (2 μM; Calbiochem), for 1 h before stimulation.
Differences observed between groups were evaluated using the Mann-Whitney U test. The p values < 0.05 were considered statistically significant.
RA serum levels are increased in parasite-infected MS patients
As illustrated in Fig. 1, RA serum levels were significantly higher in parasite-infected MS patients (11.2 ± 0.7 ng/ml), compared with levels in healthy controls (3.7 ± 0.3 ng/ml; p < 0.0001), in uninfected MS patients (3.3 ± 0.3 ng/ml; p < 0.0001), or in treated parasite-infected MS patients (3.7 ± 0.4 ng/ml; p < 0.0001).
The limited number of patients infected with each different type of parasite precludes appropriate statistical analysis regarding impact of specific helminth infections on RA serum levels and immunological responses.
SEA induces expression of Adh1, Raldh2, and RARs on DCs
Previous studies have reported expression in DCs of genes involved in the biosynthesis and metabolism of RA, such as Adh1, Raldh2, Adh5, and Raldh1 (24, 25). Locally produced RA can act on immune cells in autocrine or paracrine fashion by binding to nuclear receptors of RAR or retinoid X receptor families. RAR proteins are ubiquitously expressed on immune cells (26). We therefore studied expression and regulation of Adh1, Raldh2 mRNA by DCs, and examined RARα and RARγ expression on these cells. RT-PCR analysis demonstrated that DCs constitutively expressed low levels of both enzymes, and that levels of expression were significantly upregulated (9- to 12-fold) following stimulation with SEA (p < 0.0001; Fig. 2A, 2B). Likewise, both ex vivo isolated as well as monocyte-derived DCs expressed low levels of RARα and RARγ mRNA. After SEA activation, an 8- to 10-fold increase in DC expression of both RARs was observed (p < 0.0001; Fig. 2C, 2D).
Because cell exposure to SEA resulted in significant TLR2 upregulation, we next examined the possible role of TLR2 in the induction of Adh1, Raldh2, or RARs in DCs. Silencing of TLR2 using the siRNA technique led to significant inhibition of Adh1, Raldh2, and RAR expression induced by SEA in DCs (p < 0.0001). Furthermore, neither the TLR9 ligand CpG 1668 nor the TLR5 ligand flagellin induced the enzymes. In contrast, Pam3Cys, another TLR2-specific ligand, did induce substantial expression of Adh1 and Raldh2, as well as RARα and RARγ (p = 0.001; Fig. 2A–D).
To gain insight into potential signaling mechanisms mediating Adh1, Raldh2, and RAR induction in DCs by SEA, we investigated the role of MAPK signaling pathways by assessing p38, ERK1/2, and JNK1/2 phosphorylation. As shown in Fig. 2E, SEA induced ERK phosphorylation in DCs to levels 8–10 times above baseline. ERK phosphorylation activity was abrogated when cells were preincubated with U0126, a specific inhibitor of ERK1/2 activation. In contrast, SEA had no measurable effect on p38 or JNK1/2 phosphorylation (Fig. 2E). As positive control, LPS stimulation of DCs led to phosphorylation of all three MAPK species. In line with these observations, induction of Adh1 and Raldh2, as well as RAR mRNA expression, was largely abrogated by inhibitors against ERK (Fig. 2A–D). In summary, SEA induced Adh1, Raldh2, RARα, and RARγ on DCs via TLR2-mediated activation of ERK. SEA effects did not differ between infected MS patients and helminth-infected subjects without MS (Supplemental Fig. 1).
SEA and retinol induce FOXP3+ Treg cells
Because SEA induced enzymes involved in RA synthesis, we set out to determine whether these DCs metabolized retinol to RA and whether they induced CD4+CD25+FOXP3+ Treg cells. DCs were stimulated with SEA, and 12 h later washed and cocultured with autologous naive CD4+CD25− T cells, as described in 2Materials and Methods. As shown in Fig. 3A, SEA-treated DCs significantly increased CD4+CD25+FOXP3+ Treg cell percentages (p = 0.0001). In contrast, CpG 1668–stimulated DCs had no effect. Treg cell numbers induced by SEA-treated DCs were further increased in the presence of retinol. To determine whether this effect was mediated through RA synthesis, DCs were stimulated with SEA both in the presence and in the absence of the Raldh inhibitor disulfiram. After 12 h, DCs were washed and cocultured with CD4+CD25− T cells in the presence of SEA plus retinol. Inhibition of de novo RA synthesis in DCs suppressed SEA-induced Treg cells. Furthermore, SEA-stimulated DCs treated with the ERK1/2 inhibitor U0126 restored CD4+CD25+FOXP3+ Treg cell percentages to levels observed prior to SEA treatment. Overall, these results indicate that SEA induced DCs to express RA synthesis enzymes, stimulating the development of FOXP3+ Treg cells. These results are also consistent with the effects of TLR2-mediated ERK signaling, inducing Raldh2 in DCs.
Th cell differentiation depends on a unique combination of stimulation and subsequent activation of diverse transcription factors. Cooperative activation of NFAT and Smad3 leads to CD4+CD25+FOXP3+ Treg cell induction (27), and cooperation between STAT3 and Smad3 switches on Th17 cell induction (28). We therefore set out to investigate whether parasite Ag induced Smad3 phosphorylation in T cells, through RA-mediated effects. After coculture of CD4+CD25− with SEA-treated DCs, as well as with untreated DCs, Smad3 phosphorylation was measured using ELISA kits in nuclear extracts of CD4+ T cells. Smad3 phosphorylation was significantly increased after coculture of CD4+CD25− T cells with SEA-treated DCs (Fig. 3B). In contrast, coculture of CD4+CD25− T cells with untreated DCs did not affect phospho-Smad3 levels. This effect was reproduced by culturing naive T cells with RA, or with retinol blocked either by the RAR antagonist LE135, or by Raldh inhibitor disulfiram, respectively (p < 0.0001; Fig. 3B). Neither SEA, RA, nor retinol affected Smad2 phosphorylation under these conditions (data not shown). We subsequently examined whether SEA and RA influenced STAT3 phosphorylation. As shown in Fig. 3C, STAT3 phosphorylation was significantly decreased in naive T cells after coculture with SEA-treated DCs (p < 0.0001). Conversely, untreated DCs (control) did not affect phospho-STAT3 levels. Similar results were observed when naive T cells were cultured in the presence of RA or of retinol. Addition of LE135 or of disulfiram abrogated this effect, resulting in significant STAT3 phosphorylation. Overall, these findings demonstrate that SEA, through RA-dependent pathways, may play an important role in modulating the balance between CD4+CD25+FOXP3+ Treg cells and Th17 cell differentiation.
Notably, as illustrated in Fig. 3D, significantly higher levels of CD4+CD25+FOXP3+ Treg cells were detected in CSF from helminth-infected MS patients (15.2 ± 1.02%), compared with levels observed in healthy controls (5.7 ± 0.43%; p < 0.0001), uninfected MS patients (5.4 ± 0.39; p < 0.0001), or patients with OIND (5.5 ± 0.5%; p < 0.0001). Values were even higher than peripheral blood levels (p = 0.009; Fig. 3E).
CD4+CD25+FOXP3+ Treg cell function induced by SEA-treated DCs was further investigated by testing ability of these cells to suppress proliferative responses and IFN-γ secretion in CD4+CD25− cells. To this end, CD4+CD25− T cells were stimulated with anti-CD3 and anti-CD28 mAb, whereas increasing numbers of autologous induced CD4+CD25+FOXP3+ Treg cells were added. The CD4+CD25+FOXP3+ Treg cells suppressed proliferation of CD4+CD25− T cells in response to PHA, and to plate-bound anti-CD3 stimulation (63 ± 15% at ratio 1:10), as well as the production of IFN-γ (53 ± 25% at ratio 1:10) by indicator CD4+CD25− T cells. Collectively, these results suggest that SEA-treated DCs induce CD4+CD25+FOXP3+ Treg cells, which retain immunosuppressive activity and may thus be considered bona fide Treg cells. Because of the low number of CD4+CD25+FOXP3+ Treg cells in CSF, it was not possible to investigate suppressive capacity of these cells.
DCs stimulated with SEA produce lower amounts of IL-6, IL-12, IL-23, and TNF-α and higher amounts of IL-10
We previously demonstrated that SEA significantly suppressed LPS-induced IL-6, IL-12, and TNF-α production, and enhanced that of IL-10 and TGF-β by DCs (16). To determine whether RA contributes to these effects, we stimulated DCs with SEA in the presence as well as in the absence of retinol, and assessed production of IL-6, IL-10, IL-12, IL-23, IL-27, and TNF-α. SEA-treated DCs produced lower amounts of IL-6, IL-12, IL-23, and TNF-α (p = 0.01). Addition of retinol further reduced levels of these cytokines (p = 0.01; Fig. 4A–D). Inhibition of RARs using the antagonist LE135 was associated with significant increase in production of all of these cytokines (p < 0.0001), and specific neutralization of IL-10Rs produced significant enhancement of proinflammatory cytokines in response to SEA (p < 0.0001; Fig. 4A–D). In contrast, SEA-exposed DCs significantly enhanced IL-10 production, and addition of retinol further increased IL-10 levels (p < 0.01; Fig. 4E). Inhibition of RARs was also associated with a significant reduction of IL-10 production (p < 0.0001), suggesting that enhanced secretion of IL-10 by DCs exposed to SEA is mediated by RA. Notably, stimulation of DCs with SEA or SEA plus retinol did not result in significant differences in IL-27 secretion compared with unstimulated DCs (data not shown).
Similar results were observed in infected MS patients and helminth-infected subjects without MS (Supplemental Fig. 2).
SEA and retinol induce SOCS3 expression
Next, we investigated the mechanisms by which RA suppressed proinflammatory cytokines. There is accumulating evidence that suppressor of SOCS, particularly SOCS1 and SOCS3, may suppress inflammatory reactions under pathological conditions in which proinflammatory cytokines play an important role (29). Moreover, previous investigations have demonstrated that reduction in DC proinflammatory cytokine production mediated by RA was associated with increase in SOCS3 expression (25). Similarly, in our system, SOCS3 expression levels were significantly upregulated following SEA stimulation (p < 0.0001; Fig. 5A). Addition of retinol to the culture system further increased SOCS3 expression 10- to 15-fold. In contrast, addition of Raldh inhibitor disulfiram, or of the RAR antagonist LE135, significantly reduced SOCS3 expression (p < 0.001; Fig. 5A). Overall, these results suggest that SEA-induced reduction of proinflammatory cytokines is, at least in part, mediated by RA. No changes were observed in SOCS1 expression under identical culture conditions (data not shown). To assess the role of different TLRs on SOCS3 expression during SEA stimulation, TLR2, TLR5, and TLR9 genes were silenced before stimulation using the siRNA technique. Enhancement of SOCS3 expression mediated by SEA was not present when TLR2 expression was silenced (p < 0.001; Fig. 5B). In contrast, no effects were observed when TLR5 and TLR9 were silenced. These results indicate that SOCS3 expression induced by SEA in DCs is TLR2 dependent.
Similar findings were observed in both infected MS patients and helminth-infected subjects without MS (Supplemental Fig. 3).
SEA inhibits proinflammatory cytokine production through a SOCS3-dependent pathway
To better understand the role of SOCS3 in SEA suppression of proinflammatory cytokines by DCs, we used the siRNA technique to inhibit its expression. Suppression of IL-23, IL-6, IL-12, and TNF-α disappeared when SOCS3 expression was silenced (p < 0.001). In contrast, silencing of SOCS1 had no effect (Fig. 6A–D). To further assess the role of SOCS3 on T cell activation during helminth infection, DCs from infected MS patients were stimulated for 12 h with SEA, then cocultured with MBP or MOG peptide-specific TCLs, and stimulated with specific Ags. IL-17 and IFN-γ secretion was measured after 48 h using ELISA. SEA-activated DCs significantly inhibited IL-17 and IFN-γ production (p < 0.001). As observed for DCs, inhibitory effects on production of both cytokines were abrogated when SOCS3 (but not SOCS1) gene expression was silenced using the siRNA technique (Fig. 6E, 6F), indicating that SEA-mediated signaling inhibited IL-17 and IFN-γ production by T cells, also through a SOCS3-dependent pathway (Fig. 7).
SEA effects did not differ between infected MS patients and helminth-infected subjects without MS (Supplemental Fig. 4).
The long life span of helminths is evidence enough of just how accomplished these organisms are at immune evasion. Several studies in humans and in animal models have shown that chronic helminth infections trigger prominent anti-inflammatory network development, leading to attenuation of Ag-specific immune responses to both the parasite and unrelated pathogens. These findings suggest parasites can not only suppress host immune responses directed against them, but can also exert bystander suppression against third-party Ags. Following this premise, we recently demonstrated that helminth-infected MS patients experienced milder disease course compared with noninfected MS subjects (9). Parasite-driven protection was associated with induction of Treg cells secreting suppressive cytokines IL-10 and TGF-β, as well as of CD4+CD25+FOXP3+ T cells displaying significant suppressive function. In addition to the development of Treg cells, helminth infections also induced regulatory B cells in MS patients, capable of dampening the immune response through production of IL-10 (30). These findings provide evidence to support autoimmune downregulation secondary to parasite infections in MS patients, through regulatory T and B cell action, with effects extending beyond simple response to an infectious agent. Evidence of regulatory mechanisms present during helminth infections is now emerging, offering potential explanations as to why infected hosts exhibit altered immune responses to bystander Ags. Helminths may thus increase regulatory cell numbers or activity, either by generating new cells or by activating or expanding existing populations.
The last few decades have witnessed a major effort to identify factors associated with MS onset and progression (3). Among potentially relevant environmental factors investigated to date, it has been suggested that lipophilic vitamin metabolites, such as those of vitamins A and D, may have immunomodulatory properties, and therefore influence the course and severity of MS (11, 14). The immunological and neurotrophic effects of vitamin A metabolites (31) make a causal relationship between vitamin A levels and MS disease activity biologically plausible. A reduction of retinol levels in MS patients compared with patients with noninflammatory neurologic diseases and with healthy controls has already been reported (32, 33). Furthermore, a recent study described inverse association between increasing levels of s-retinol and MRI lesion activity in MS patients, finding s-retinol levels predicted MRI outcomes during subsequent months of follow-up (34). This is consistent with observations in different animal models. Oral administration of retinoid prevented development of EAE and improved clinical course, even when given after the onset of disease (35). Likewise, RA prolonged survival in mice with lupus nephritis (36) and reduced murine collagen-induced arthritis development (37).
Although recent studies have highlighted an important role for RA in the induction of Treg cells in the gut, its role in systemic immune responses is poorly understood. The present data demonstrate that helminth-mediated immunomodulation observed in MS patients is, at least in part, exerted by TLR2- and RA-dependent pathways, through two different mechanisms, as follows: 1) induction of IL-10 and of FOXP3+ Treg cells, and 2) suppression of proinflammatory cytokine production, mediated mainly by SOCS3 (Fig. 7).
Recent insights into the role of RA in the promotion and regulation of multiple immunological pathways draw attention to the influence of RA on immunity. For example, mucosal CD103+ DCs express RA synthesizing enzymes, and are able to induce molecules such as CCR9 and α4β7 on conventional T cells involved in directing gut trophism (38). In addition, RA produced by DCs from the GALT, in synergy with TGF-β, induces strong extrathymic FOXP3+ Treg cell differentiation, and inhibits Th17 differentiation (39, 40). These effects can be reversed by blocking RARs, indicating the capacity of the gut to induce tolerogenic outcomes through a RA-dependent pathway. Likewise, recent studies have demonstrated that other tissues that constitute environmental interfaces (e.g., skin, lungs) resemble the gut in that they also contain DCs that constitutively produce RA to induce FOXP3+ Treg cells (41). RA has also been implicated in the generation of IgA-secreting B cells, adding further evidence to support a multifactorial role for RA in mucosal immunity (42).
Vitamin A metabolites, particularly RA, also modulate more specific functional aspects of immune responses, such as the Th1-Th2 balance. Thus, vitamin A deficiency correlates with decreased Th2 cell responses (43) and, conversely, vitamin A supplementation blocks the production of Th1 cell cytokines (44). These effects of vitamin A on Th1 and Th2 cell differentiation are mediated by RA. In fact, RA promotes Th2 cell differentiation by inducing IL-4 gene expression (45), as well as Th2 cell–promoting transcription factors, such as GATA3 and STAT6 (44, 46). Moreover, RA blocks the expression of the Th1 cell master regulator T-bet (44, 46). In contrast, recent experiments using Th1-driving intracellular parasite T. gondii have shown that RA is required for the generation of Th1 and Th17 responses in the gut (47). To reconcile these conflicting observations, it has been proposed that RA may act to aid T cell activation during early stages of the immune response, but downregulate effector responses at later stages. Although a mechanism for this process has not yet been elucidated, it could be due to changes in expression of either Raldh or RAR during inflammatory responses, which could in turn lead to activation of distinct signaling pathways (47).
Recently, both in vitro and in vivo experiments have demonstrated that RA suppresses Th17 cell differentiation, reciprocally promoting generation of Treg cells expressing FOXP3 through RARα binding, associated to downregulation of ROR-γt, a key transcription factor for Th17 cell development (40, 48). In agreement with our results, possible mechanisms to explain these observations include the following: suppression of IL-6– and IL-23–driven signaling, as well as Smad3-dependent enhanced TGF-β signaling (49). Previous investigations have demonstrated that RA was insufficient to induce FOXP3 in the absence of TGF-β, indicating that responsiveness to TGF-β is a prerequisite for RA to induce Treg cell differentiation (49). Likewise, blockade of TGF-β–induced Smad2/3 phosphorylation diminished TGF-β–mediated Treg cell generation and abrogated the additive effect of RA (50). In contrast, RA increased Smad3 expression in activated CD4+ T cells (49, 51). Overall, these findings indicate that RA and TGF-β synergistically promote enhanced Smad3 activity, increasing mutual signaling to further enhance FOXP3 expression. Alternatively, RA concentration may be a factor in determining RA signaling outcome. For example, reports have shown that high RA concentrations can block Th17 cell differentiation (40), whereas low RA concentrations seem to be essential for Th17 cell differentiation (52). These associations suggest RA synthesis pathways may be manipulated to shift the balance between Treg and effector cells, and in turn influence immunopathology. Similarly, in vivo RA suppresses the inflammatory response and ameliorates course of autoimmune diseases in different animal models, including trinitrobenzene sulphonic acid–induced colitis (53), rheumatoid arthritis (37), type I diabetes (54), and EAE (35, 49). In the latter, reduced Th17 differentiation and Th1 responses (49), combined with development of Th2 phenotype T cells (35), were considered responsible for disease resolution.
TLR regulation may be altered by helminth parasites, modifying function as well as levels of expression. In previously published work, we provided evidence that surface expression of TLR2 on both B cells and DCs was significantly higher in helminth-infected MS patients compared with uninfected MS subjects or healthy controls. Moreover, cells exposed to SEA resulted in significant TLR2 upregulation, reduced DC production of proinflammatory cytokines, and enhanced TGF-β as well as IL-10 production (16). These patterns of DC response are similar to those observed after stimulation using Giardia extracts (55). The results we now present demonstrate that genes involved in RA biosynthesis and metabolism, such as Adh1 and Raldh2, as well as RARs, can be induced in DCs via TLR2-dependent ERK signaling, which programs DCs to induce FOXP3+ Treg cells and suppresses the production of proinflammatory cytokines via induction of SOCS3. Notably, DCs from TLR2−/− mice exhibited significantly lower Raldh activity compared with their wild-type counterparts (56). These cells were significantly impaired in their ability to induce gut-homing T cells as well as FOXP3+ Treg cells, indicating that TLR2-mediated signals are key for the acquisition of functional properties by DCs, including their capacity to produce RA, which is critical to imprint immune functions. Consistent with these findings, it has been recently demonstrated that zymosan, a yeast cell wall derivative recognized by TLR2, induced Treg cells and suppressed IL-23– and Th17/Th1-mediated autoimmune responses in vivo, through RA-mediated pathways (25).
Parasites inhabit immune competent hosts for long periods, and can therefore develop regulatory networks, generating strong anti-inflammatory responses, which may on one hand restrict host tissue damage, but on the other also enhance parasite survival. Our data provide evidence that helminths and their Ags modulate TLR2 expression and consequently the innate immune response, preventing host tissue damage. TLR2 expression and function can, however, have bystander effects such that responses to nonhelminth pathogens are also dampened, inducing a protective effect against autoimmune diseases such as MS. The present results revealed RA as a crucial mediator in the maintenance of this anti-inflammatory milieu.
This work was supported by an internal grant from the Dr. Raúl Carrea Institute for Neurological Research, Foundation for the Fight against Infant Neurological Illnesses (to J.C.).
The online version of this article contains supplemental material.
Abbreviations used in this article:
experimental autoimmune encephalomyelitis
myelin basic protein
myelin oligodendrocyte glycoprotein
magnetic resonance imaging
other inflammatory neurologic disease
soluble egg Ag
small interfering RNA
suppressor of cytokine signaling
T cell line
Jorge Correale is a board member of Merck–Serono Argentina, Biogen–Idec Latin America, Genzyme Latin America, and Merck–Serono Latin America. Dr. Correale has received reimbursement for developing educational presentations for Merck–Serono Argentina, Merck–Serono Latin America, Biogen–Idec Argentina, and Teva–Tuteur Argentina, Teva–Ivax, as well as professional travel/accommodations stipends. Mauricio F. Farez has received professional travel and accommodations stipends from Merck–Serono Argentina.