Sodium Benzoate, a Food Additive and a Metabolite of Cinnamon, Enriches Regulatory T Cells via STAT6-Mediated Upregulation of TGF-β

Multiple sclerosis (MS) is the most common autoimmune demyelinating disease of the CNS. In healthy human beings, autoimmune T cells are normally suppressed by a small population of T cells called regulatory T cells (Tregs) (1). Tregs suppress activation and proliferation of self-reactive T cells and thereby inhibit the immune response of self-reactive T cells against self-antigens (1, 2). However, in MS patients, myelin-reactive T cells overcome the usual restraining mechanism of Tregs, become activated, and target the CNS (3). Tregs are characterized by the presence of transcription factor Foxp3, and therefore CD4⁺CD25⁺ Foxp3⁺ T cells are considered the most common phenotype of Tregs (1, 4). Recent studies suggest that the expression of Foxp3 and the numbers of peripheral CD4⁺CD25⁺ Foxp3⁺ T cells are significantly reduced in relapsing-remitting MS patients compared with control subjects (5). Therefore, upregulation and/or maintenance of Tregs may be beneficial for MS.

Sodium benzoate (NaB), the sodium salt of an aromatic carboxylic acid, is a widely used food preservative and a metabolite of cinnamon. NaB is also of medical importance, as it is a component of Ucephan and Ammonul (both are a combination of sodium benzoate and sodium phenylacetate), Food and Drug Administration (FDA)–approved drugs used in the treatment of hepatic metabolic defects associated with hyperammonemia, such as the urea cycle disorder in children (6–8). The usual FDA-approved dose of NaB for patients with hyperammonemia is 4–10 g/d, which could be increased further during the treatment of acute stage hyperammonemia (6, 9). It has been reported that a minor amount of NaB is also excreted in the urine of humans (10, 11). It is nontoxic and can be administered as a solution in drinking water. It has been reported that a 2% solution of NaB in drinking water is safe for lifelong treatment in mice without any noticeable side effects (12). Recently, we have delineated that NaB is capable of maintaining and/or upregulating Tregs and protecting mice from relapsing-remitting experimental allergic encephalomyelitis (EAE) (13–15). The mechanism by which NaB upregulates Tregs, however, remained unknown.

TGF-β, a multifunctional cytokine, is known to inhibit immune responses through the generation of Tregs by inducing the expression of Foxp3. In this study, we determine that NaB induces the expression of TGF-β in splenocytes via the activation of STAT6 and thereby enriches Tregs. Abrogation of NaB-mediated protection of Tregs and alleviation of EAE by TGF-β neutralization suggest that NaB protects against EAE via the STAT6–TGF-β–Treg pathway.

Bovine myelin basic protein (MBP), l-glutamine, HBSS, and 2-ME were obtained from Life Technologies (Carlsbad, CA). FBS and RPMI 1640 were from Mediatech (Washington, DC). Benzoic acid was purchased from Sigma-Aldrich (St. Louis, MO). Heat-killed Mycobacterium tuberculosis (H37RA) was purchased from Difco Labs. IFA was obtained from Millipore. Rag1 (^−/−) mice (B6.129S7-Rag1tm1Mom/J) were purchased from The Jackson Laboratory (Bar Harbor, ME). Anti–TGF-β Abs were purchased from R&D (Minneapolis, MN). Abs against CD3, CD4, Foxp3, and CD62L were purchased from BD Biosciences (San Jose, CA). Anti-CD73 Abs were obtained from eBioscience (San Diego, CA). For details on Abs, see Supplemental Table I.

Powdered benzoic acid was dissolved in sterile water, and the resulting solution was titrated to pH 7 using sodium hydroxide.

Proteolipid protein (PLP)_139–151–specific 5B6 TCR transgenic (Tg) mice were obtained from Prof. Vijay Kuchroo (Harvard Medical School, Boston, MA). These mice were genotyped by flow cytometry. Briefly, a drop of blood was collected from a tail bleed into 200 μl PBS in a 96-well plate. Samples were spun, RBCs were lysed, and cells were stained with Thy1.1, CD4, and Vβ6. When gated on CD4⁺ cells, the homozygotes were positive for Thy1.1, and ≥90% cells were positive for Vβ6.

Animal maintenance and experiments were in accordance with National Institutes of Health guidelines and were approved by the Institutional Animal Care and Use Committee of the Rush University Medical Center, Chicago, IL.

Adoptive transfer of EAE was performed as described previously by us (14, 16–18). Briefly, 4- to 5-wk-old female SJL/J mice were purchased from Harlan Sprague Dawley (Indianapolis, IN). Donor mice were immunized s.c. with 400 μg bovine MBP and 60 μg M. tuberculosis in IFA (14, 16–18). Animals were killed 10–12 d postimmunization (dpi), and the draining lymph nodes were harvested and single-cell suspensions were cultured in RPMI 1640 supplemented with 10% FBS, 50 μg/ml MBP, 50 μM 2-ME, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin. On day 4, cells were harvested and resuspended in HBSS. A total of 2 × 10⁷ viable cells in a volume of 200 μl were injected into the tail vein of naive mice. Pertussis toxin (150 ng per mouse; Sigma-Aldrich) was injected once via the i.p. route on 0 d posttransfer (dpt) of cells. Animals were observed daily for clinical symptoms. Six mice were used in each group. Female mice (4–5 wk old) were randomly selected for any group. Experimental animals were scored by a masked investigator, as follows: 0, no clinical disease; 0.5, piloerection; 1, tail weakness; 1.5, tail paralysis; 2, hind limb weakness; 3, hind limb paralysis; 3.5, forelimb weakness; 4, forelimb paralysis; 5, moribund or death.

PLP_139–151-specific 5B6 TCR Tg mice were obtained from Prof. Vijay Kuchroo (Harvard Medical School, Boston, MA). Female Tg mice (4–5 wk old) were immunized with 10 μg PLP_139–151 in M. tuberculosis in IFA, as described (19).

NaB was mixed with drinking water, and EAE mice were gavaged with 100 μl NaB-containing water at a dose of 100 mg/kg body weight per day, using a gavage needle. Therefore, control EAE mice also received the same amount of water as vehicle via gavage. This dose of NaB is equivalent to or less than the FDA-approved dose (4–10 g/d) of NaB (6, 9).

Spleens isolated from female PLP-TCR Tg mice were placed into a cell strainer and mashed with a syringe plunger. Resulting single-cell suspensions were treated with RBC lysis buffer (Sigma-Aldrich), washed, and cultured in 12-well plates in RPMI 1640 supplemented with 10% FBS, 10 μg/ml PLP_139–151, 50 μM 2-ME, 2 mM l-glutamine, 100 U/ml penicillin, and 100 μg/ml streptomycin in the presence or absence of NaB or sodium formate (NaFO).

Total RNA was isolated from splenocytes using the RNeasy Mini Kit (QIAGEN, Valencia, CA) following the manufacturer’s protocol. To remove any contaminating genomic DNA, total RNA was digested with DNase. Semiquantitative RT-PCR was carried out as described earlier (14, 18, 20), using a RT-PCR kit from Clontech (Mountain View, CA). Briefly, 1 μg total RNA was reverse transcribed using oligo(dT)_12–18 as primer and Moloney murine leukemia virus reverse transcriptase (Clontech) in a 20-μl reaction mixture. The resulting cDNA was appropriately diluted, and diluted cDNA was amplified using titanium Taq DNA polymerase and the following primers. Amplified products were electrophoresed on 1.8% agarose gels and visualized by ethidium bromide staining. The following primers were used: Foxp3, sense, 5′-CAG CTG CCT ACA GTG CCC CTAG-3′, antisense, 5′-CAT TTG CCA GCA GTG GGT AG-3′; CD25, sense, 5′-AGC CAA GTA GGG TGT CTC TCA ACC-3′, antisense, 5′-GCC CAG GAT ACA CAG TGA AGA ACG-3′; CD73, sense, 5′-GCG CAA ACA TTA AGG CTC GGGG-3′, antisense, 5′-GAC CAC AGG CAC CTG CCG TC-3′; CD4, sense, 5′-CCA ACA AGA GCT CAA GGA GAC CAC-3′, antisense, 5′-CGT ACC CTC TTT CCT AGC AAA GGA-3′; CD62L, sense, 5′-AGC CTC TTG CCA GCC AGG GT-3′, antisense, 5′-CCA GCC CCG AGA ATG CGG TG-3′; GAPDH, sense, 5′-GGTGAAGGTCGGTGTGAACG-3′, antisense, 5′-TTGGCTCCACCCTTCAAGTG-3′.

The relative expression of each gene with respect to GAPDH was measured after scanning the bands with a FluorChem 8800 Imaging System (Alpha Innotech, San Leandro, CA).

Real-time PCR analysis was performed using the ABI PRISM 7700 Sequence Detection System (Applied Biosystems, Foster City, CA), as described earlier (14, 18, 20). Briefly, reactions were performed in 96-well optical reaction plates on cDNA equivalent to 50 ng DNase-digested RNA in a volume of 25 μl, containing 12.5 μl TaqMan Universal Master Mix and optimized concentrations of FAM-labeled probe, forward and reverse primers following the manufacturer’s protocol. All primers and FAM-labeled probes for mouse genes and GAPDH were obtained from Applied Biosystems. The mRNA expressions of respective genes were normalized to the level of GAPDH mRNA. Data were processed by ABI Sequence Detection System 1.6 software and analyzed by ANOVA.

Chromatin immunoprecipitation (ChIP) assays were performed using a kit (Millipore), as described earlier (21, 22). Briefly, after 3 h of stimulation, cells were fixed by adding formaldehyde, and cross-linked adducts were resuspended and sonicated. ChIP was performed on the cell lysate by overnight incubation at 4°C with 2 μg either anti-STAT6 or anti-STAT1 Abs, followed by incubation with protein G agarose (Santa Cruz Biotechnology) for 2 h. IgG was also run as a control. The beads were washed and incubated with elution buffer. To reverse the cross-linking and purify the DNA, precipitates were incubated in a 65°C incubator overnight and digested with proteinase K. DNA samples were then purified and precipitated, and precipitates were washed with 75% ethanol, air dried, and resuspended in Tris-EDTA buffer. The following primers were used for amplification of chromatin fragments of the mouse TGF-β gene: sense, 5′-CCA CCG CCT AGG TCC CCA C-3′; antisense, 5′-GGG AGT TGT TGA AGG GTC-3′.

Splenocytes (1 × 10⁶ cells/well) were grown in RPMI 1640complete medium without antibiotics. STAT6 small interfering RNA (siRNA) (ThermoFisher Scientific) was diluted in 100 μl Opti-MEM I reduced serum medium at a concentration of 0.25 μg per well. Similarly, Lipofectamine 2000 (ThermoFisher Scientific) reagent was diluted in 100 μl Opti-MEM medium and incubated for 5 min at room temperature in a separate Eppendorf tube. Then siRNA and Lipofectamine 2000 were combined, and after 20 min of reaction, the siRNA–Lipofectamine 2000 complex was added to each well. Cells were treated 48 h after transfection and harvested for subsequent analysis.

For immunofluorescence staining (21, 22), rabbit polyclonal anti–TGF-β (1:100) and goat polyclonal anti-Iba1 (1:100) (Supplemental Table I) were used. The samples were mounted and observed under an Olympus IX81 fluorescence microscope. Counting analysis was performed using Olympus Microsuite V software with the help of a touch counting module. After acquiring images under a ×20 objective lens, images were further analyzed as follows. Before counting cells, the entire image area was calibrated with the help of a rectangular box available in the touch counting panel. Once the area of the image was measured, the touch counting program was applied to count the number of fluorescent signals using the simple mouse click method.

Production of TGF-β in culture supernatant was monitored by ELISA, using an assay kit from eBioscience (San Diego, CA).

Two-color flow cytometry was performed as described previously (15, 23). Briefly, 1 × 10⁶ splenocytes suspended in flow staining buffer were incubated at 4°C with appropriately diluted FITC-labeled Ab to CD4 (Supplemental Table I) for 30 min, washed, and resuspended in fixation and permeabilization solution. Following incubation in the dark for 30 min, cells were washed, blocked with test Fc block (anti-mouse CD16/32) in permeabilization buffer, and subsequently incubated with appropriately diluted PE-labeled Abs to Foxp3 at 4°C in the dark. After incubation, the cell suspension was centrifuged, washed three times, and resuspended in flow staining buffer. The cells then were analyzed through FACS (BD Biosciences, San Jose, CA). Cells were gated based on morphological characteristics. Apoptotic and necrotic cells were not accepted for FACS analysis.

Statistical significance was accessed via an unpaired two-tailed t test using Sigma plot software. The results are interpreted based on the normally distributed data points (two-tailed), consideration of the null hypothesis (type 1 error), and a p value < 0.05 was considered significant.

Earlier we demonstrated that NaB stimulates Tregs in MBP-primed splenocytes and suppresses adoptive transfer of EAE in mice (14). Similarly, NaB was also capable of enriching CD4⁺Foxp3⁺ Tregs in PLP-primed splenocytes isolated from PLP-TCR Tg mice (Fig. 1A). These results were specific, as NaFO—having a similar structure but without the benzene ring—was unable to protect or upregulate CD4⁺Foxp3⁺ Tregs in Tg mice (Fig. 1A). To directly prove that the enrichment of Tregs by NaB is actually due to induction of Tregs, we used MOG-primed Rag1 (^−/−) splenocytes. By two-color FACS analysis, we found CD4⁺Foxp3⁺ Tregs in wild-type (WT), but not Rag1 (^−/−), splenocytes (Supplemental Fig. 1). MOG priming decreased CD4⁺Foxp3⁺ Tregs in WT splenocytes, and treatment by NaB, but not NaFO, resulted in the upregulation of CD4⁺Foxp3⁺ Tregs in WT splenocytes (Supplemental Fig. 1). However, we did not see any upregulation of CD4⁺Foxp3⁺ Tregs in Rag1 (^−/−) splenocytes (Supplemental Fig. 1), indicating the true induction of Tregs by NaB. Because Foxp3⁺ Tregs usually express CD62L and CD73, we also analyzed the surface expression of these molecules by FACS. Similar to the upregulation of Foxp3, NaB, but not NaFO, enriched CD4⁺CD62L⁺ (Fig. 1B) and CD4⁺CD73⁺ (Fig. 1C) Tregs in PLP-primed splenocytes. The mean fluorescence intensities (MFIs) of gated populations of cells confirm our FACS results for CD62L (Fig. 1D) and CD73 (Fig. 1E). We also analyzed the mRNA expression of different Treg markers. The mRNA expression of Foxp3, CD73, CD62L, and CD25 decreased upon PLP priming, and NaB treatment rescued the expression of these molecules in a time-dependent manner, as evidenced by our semiquantitative RT-PCR (Supplemental Fig. 2A) and real-time PCR (Supplemental Fig. 2B). The upregulation of Foxp3, CD73, CD62L, and CD25 mRNAs was evident in as little as 6 h of NaB treatment, which further increased with subsequent hours of treatment (Supplemental Fig. 2A, 2B). A dose-dependent study showed that NaB was capable of restoring and/or upregulating the expression of these Treg-specific molecules at a dose of 200 μM (Supplemental Fig. 2C, 2D). In contrast, neither PLP priming nor NaB treatment had an effect on the mRNA expression of CD4 (Supplemental Fig. 2A, 2D), suggesting that these results are not due to any alteration in CD4⁺ T cells. In cell culture experiments, NaB shows efficacy at higher concentrations because NaB interacts with glycine to form hippuric acid (10, 11) and glycine is present in cell culture media.

Next, we investigated mechanisms by which NaB enriched Tregs. Because TGF-β is known to play an important role in the generation of Tregs from CD4⁺CD25⁻ precursors (24), we examined the effect of NaB on the expression of TGF-β. The level of TGF-β was monitored in supernatants by ELISA. As is evident from Fig. 2A and 2B, restimulation of splenocytes isolated from PLP-TCR Tg mice with PLP reduced the production of TGF-β. However, NaB treatment dose-dependently increased the production of TGF-β in PLP-primed splenocytes (Fig. 2A). Time-dependent studies indicated that NaB was capable of increasing the production of TGF-β significantly within 6 h of treatment (Fig. 2B). Western blot analysis of splenocytes showed that NaB induced both dimeric and monomeric forms of TGF-β in PLP-primed splenocytes (Fig. 2C, 2D).

Next, we examined whether NaB was capable of protecting and/or upregulating TGF-β in vivo in the spleen of EAE mice. Immunization of PLP-TCR Tg mice with 10 μg PLP strongly induced clinical symptoms of EAE (19). Therefore, Tg mice were immunized with PLP, and from the onset of the acute phase (4 dpi), mice were treated with either NaB or NaFO via gavage until 12 dpi (acute phase), followed by monitoring of the level of TGF-β by immunofluorescence. Induction of EAE decreased the level of TGF-β in spleen, which was increased by NaB treatment (Fig. 2E). These results were specific, as NaFO treatment remained unable to increase the level of splenic TGF-β (Fig. 2E). Double-labeling analysis showed that TGF-β was mostly expressed by Iba1-positive macrophages in the spleen (Fig. 2E).

Because NaB increased the expression of TGF-β in PLP-primed splenocytes, we examined whether NaB was also capable of upregulating TGF-β in normal splenocytes. Dose-dependent studies showed that NaB could increase the mRNA expression of TGF-β at different doses, exhibiting maximum efficacy at 200 μM (Fig. 3A, 3B). These results were further corroborated by Western blot (Fig. 3C, 3D). Similarly, time-dependent analysis indicated that NaB-mediated increase in TGF-β mRNA was maximum at 6 h (Fig. 3E, 3F). Immunofluorescence analysis of splenocytes also showed that NaB increased the level of TGF-β in Iba1-positive cells (Fig. 3G, 3H). Together, these results suggest that NaB was capable of increasing the expression of TGF-β in both normal and Ag-primed splenocytes.

Next, to investigate mechanisms by which NaB increased the expression of TGF-β, we analyzed the Tgfb promoter by MatInspector. Mouse Tgfb promoter harbors consensus sequences for binding of many transcription factors, such as NF-κB, C/EBPβ, STAT6, and so on. However, NaB does not induce the activation of NF-κB and C/EBPβ. In fact, NaB suppresses the activation of NF-κB and C/EBPβ in microglia (25). Therefore, we examined the role of STAT6 (Fig. 4A) in NaB-mediated upregulation of TGF-β. At first, we investigated whether NaB was capable of inducing the activation of STAT6. In the case of IL-4 signaling, STAT6 undergoes phosphorylation followed by its entry into the nucleus and recruitment to the target gene promoter (26). We found that NaB induced the phosphorylation of STAT6 without increasing the level of total STAT6 in normal splenocytes (Fig. 4B, 4C). STAT6 phosphorylation was clearly visible at 30 and 60 min of NaB stimulation (Fig. 4B, 4C). Similarly, NaB, but not NaFO, also induced the phosphorylation of STAT6 without increasing the level of total STAT6 in PLP-specific splenocytes (Supplemental Fig. 3A, 3B). Consistent with the fact that activated STAT6 enters into the nucleus, NaB induced nuclear translocation of STAT6 in normal splenocytes at 60 min of treatment or later (Fig. 4D, 4E). These results were specific, as we did not see any nuclear translocation of STAT1 upon NaB treatment (Fig. 4D).

Because NaB treatment led to the activation of STAT6, we examined whether NaB induced the recruitment of STAT6 to the Tgfb gene promoter. After immunoprecipitation of chromatin fragments by Abs against STAT6, we were able to amplify 83 bp fragments flanking the STAT6 binding site in NaB-treated, but not in control untreated or in NaFO-treated, normal splenocytes (Fig. 4F, 4G). Similarly, NaB, but not NaFO, also induced the recruitment of STAT6 to the Tgfb promoter in PLP-specific splenocytes (Supplemental Fig. 3C, 3D). In contrast, no amplification product was observed in any of the immunoprecipitates obtained with control IgG in either normal (Fig. 4F, 4G) or PLP-specific splenocytes (Supplemental Fig. 3C, 3D), suggesting the specificity of these interactions.

Because NaB treatment led to the activation of STAT6 and recruitment of STAT6 to the Tgfb gene promoter, next we examined whether STAT6 was, in fact, required for NaB-mediated upregulation of TGF-β. Therefore, we employed siRNA-mediated knockdown of STAT6. STAT6, but not control, siRNA suppressed the expression of STAT6 protein in normal splenocytes (Fig. 5A). Accordingly, STAT6, but not control, siRNA inhibited NaB-mediated upregulation of TGF-β mRNA (Fig. 5B, 5C) and protein (Fig. 5D) expression. These results are also supported by TGF-β ELISA from supernatant (Fig. 5E). These results suggest that NaB requires STAT6 for the upregulation of TGF-β in normal splenocytes.

Next, we investigated whether NaB also required STAT6 for increasing the expression of TGF-β in neuroantigen-primed splenocytes. As with control splenocytes, siRNA knockdown of STAT6 also inhibited NaB-mediated increase in TGF-β mRNA in PLP-primed splenocytes (Fig. 5F, 5G). To examine whether the effect of NaB was Ag specific, we also used MBP-primed splenocytes. As with PLP-primed splenocytes, siRNA knockdown of STAT6 (Fig. 5H) also suppressed NaB-mediated upregulation of TGF-β mRNA (Fig. 5I, 5J) and protein (Fig. 5K, 5L) in MBP-primed T cells. Because STAT6 is a signature transcription factor for Th2 cells and is responsible for the transcription of Th2 cell–specific molecules, such as GATA3 and IL-10, we also determined whether STAT6 knockdown suppressed the expression of GATA3 and IL-10 in PLP-primed splenocytes. As reported earlier in the case of MBP-primed splenocytes (14), NaB treatment also increased the expression of IL-10 and GATA3 mRNAs in PLP-primed splenocytes (Fig. 5F, 5G). However, siRNA knockdown of STAT6 attenuated NaB-mediated increase in GATA3 and IL-10 mRNAs in PLP-primed splenocytes (Fig. 5F, 5G). These results suggest that STAT6 is responsible for the generation of both Th2 cells and Tregs upon NaB treatment.

Because TGF-β is involved in the differentiation of non-Tregs into Tregs and NaB treatment increased the production of TGF-β, we examined the role of TGF-β in NaB-mediated upregulation of Tregs. NaB upregulated the mRNA expression of Foxp3, CD73, CD62L, and CD25 in PLP-primed splenocytes (Fig. 6A, 6B). Similarly, NaB treatment also increased the Foxp3⁺CD4⁺ population of T cells in splenocytes, as is evident from the FACS dot plot (Fig. 6C). This finding is also corroborated by the MFI of Foxp3 (Fig. 6D). To neutralize the functions of TGF-β, we used functional blocking Abs against TGF-β. Interestingly, TGF-β neutralizing Abs nullified NaB-mediated upregulation of Foxp3, CD73, CD62L, and CD25 mRNAs (Fig. 6A, 6B) and enrichment of the Foxp3⁺CD4⁺ population of T cells (Fig. 6C, 6D). In contrast, control IgG had no such abrogating effect (Fig. 6), suggesting the specificity of the effect.

Next, we investigated whether NaB treatment protected Tregs in EAE mice via TGF-β. EAE was induced in female SJL/J mice by adoptive transfer of MBP-primed T cells, as described by us in many studies (14, 16–19). During autoimmune insults, Tregs become both numerically and functionally defective. Therefore, as expected, we observed significant reduction in the Foxp3⁺CD4⁺ population of T cells in EAE splenocytes as compared with normal splenocytes (Fig. 7A, 7B). Treatment of EAE mice with NaB, but not NaFO, led to the protection and/or upregulation of the Foxp3⁺CD4⁺ population in splenocytes (Fig. 7A, 7B). However, cotreatment of EAE mice with TGF-β neutralizing Abs abrogated NaB-mediated enrichment and/or protection of the Foxp3⁺CD4⁺ population of T cells (Fig. 7A, 7B). In contrast, control IgG had no such abrogating effect, suggesting the specificity of the effect. Similarly, our dual immunohistochemical studies with CD3 and Foxp3 further substantiate the role of TGF-β in NaB-mediated upregulation of Foxp3 in vivo in the spleen of EAE mice at the protein level (Fig. 7C). These results clearly show that NaB treatment protects Tregs in EAE mice via TGF-β.

After the discovery of IL-23, Th17 cells are considered to play a more active role than Th1 cells in the disease process of EAE and MS (27, 28). Of interest, Tregs release IL-35 to control the proliferation of autoimmune Th17 cells, and thus an inverse relationship exists between Tregs and Th17 cells (29). Because NaB treatment enriched Tregs, we examined whether NaB was also capable of regulating Th17 cells in EAE mice. Whereas induction of EAE increased the level of CD4⁺IL-17⁺ T cells in splenocytes (Fig. 8A), NaB markedly suppressed EAE-induced upregulation of the CD4⁺IL-17⁺ T cell population (Fig. 8A). In contrast, NaFO treatment had no such suppressive effects on CD4⁺IL-17⁺ T cells (Fig. 8A). Because NaB upregulated Tregs via TGF-β and Tregs are known to suppress Th17 cells, we also examined whether NaB requires TGF-β to suppress the Th17 response in EAE mice. Of note, cotreatment of EAE mice with TGF-β neutralizing Abs, but not control IgG, negated NaB-mediated inhibition of CD4⁺IL-17⁺ T cells (Fig. 8A). MFI analysis of IL-17 (Fig. 8B) within the CD4⁺ population also supported this finding.

Next, to investigate the functional significance of NaB-mediated increase in TGF-β further, we assessed whether NaB protected mice from the clinical symptoms of EAE via TGF-β. We induced EAE in female PLP-TCR Tg mice by PLP immunization (Fig. 9A) and female SJL/J mice by adoptive transfer of MBP-primed T cells (Fig. 9B). During NaB treatment, the function of TGF-β was blocked in vivo in EAE mice by anti–TGF-β neutralizing Ab. At first, we examined the appropriate dose of anti–TGF-β neutralizing Ab to be used in EAE mice. Although anti–TGF-β Ab markedly aggravated the clinical symptoms of EAE at a dose of 50 μg per mouse, we did not observe any such increase in clinical score at a dose of 20 μg per mouse (Supplemental Fig. 4). Therefore, to investigate the role of TGF-β in NaB-mediated protection of EAE, we used TGF-β at a dose of 20 μg per mouse. As is evident from Fig. 9, NaB treatment alleviated clinical symptoms of EAE in both PLP-TCR Tg mice (Fig. 9A) and adoptively transferred female SJL/J mice (Fig. 9B). However, functional blocking of the anti–TGF-β Ab almost completely abrogated the NaB-mediated protective effect on EAE mice in both cases (Fig. 9A, 9B). This result was specific, as control IgG had no such effect (Fig. 9A, 9B). Together, these results suggest that NaB protects EAE via TGF-β.

MS is an autoimmune disorder of the CNS in which myelin components are particularly targeted by the immune system, resulting in demyelination of axons and associated debilitating symptoms that vary over time. Despite intense investigations, no effective therapy is available for this disease. It has been shown that there is a significant decrease in the number of CD4⁺FOXP3⁺ T cells as well as in the expression level of Foxp3 in relapsing-remitting MS and other lymphoproliferative autoimmune disorders (5, 30, 31). Hence, the upregulation of FOXP3⁺ Tregs might be useful for suppressing the activation of autoimmune T cells and controlling autoimmune disorders. Earlier we have delineated that NaB, a metabolite of cinnamon and an FDA-approved drug against urea cycle disorders in children, upregulates Tregs and inhibits the disease process of EAE in mice.

The mechanisms by which NaB protects or induces Tregs under autoimmune conditions, however, are poorly understood. Because TGF-β is an important inducer of Tregs, we examined the role of TGF-β in NaB-mediated protection and/or upregulation of Tregs. In this study, we show that NaB treatment enriches Tregs via upregulation of TGF-β. Our conclusion is based on the following findings: First, NaB increased the expression of TGF-β mRNA and protein in normal splenocytes. Second, PLP priming reduced the expression of TGF-β in splenocytes. However, NaB treatment increased the expression of TGF-β in PLP-primed splenocytes. Third, induction of EAE reduced the level of TGF-β in spleen. However, NaB treatment was capable of restoring or increasing the level of TGF-β in the spleen of EAE mice. Fourth, PLP priming reduced Tregs in splenocytes and NaB upregulated Tregs in PLP-primed splenocytes. However, neutralization of TGF-β abrogated NaB-mediated upregulation of Tregs in PLP-primed splenocytes. Fifth, induction of EAE reduced the population of CD4⁺Foxp3⁺ T cells, and NaB treatment inhibited the loss of CD4⁺Foxp3⁺ T cells in EAE mice. However, treatment of EAE mice with neutralizing Abs against TGF-β abolished NaB-mediated protection of Tregs. Sixth, although TGF-β promotes Tregs, in the presence of IL-6, TGF-β is known to promote the differentiation of Th17 cells (32). However, earlier we have shown that NaB suppresses the induction of IL-6 (25). Consistently, in this study, we show that treatment of EAE mice with NaB led to the inhibition of CD4⁺IL-17⁺ T cells in EAE mice via TGF-β. Finally, anti–TGF-β Ab also neutralized the protective effect of NaB against EAE, indicating that NaB protects mice from EAE via TGF-β.

Because TGF-β is a multifunctional molecule with important functions in development, cancer, wound healing, immunity, and so on, characterization of intracellular pathways evoked to transduce the signal from the cell surface to the nucleus for induction of the TGF-β gene, which is poorly understood, is an active area of investigation. Analysis of the mouse Tgfb gene promoter shows that it has consensus sequences for binding of several transcription factors, such as NF-κB, C/EBPβ, and STAT6. However, NaB is unable to induce the activation of NF-κB and C/EBPβ (25). Therefore, we examined the role of STAT6 in NaB-mediated upregulation of TGF-β. STAT6, belonging to the signal transducer and activator of transcription family of proteins, is critical for a number of responses in T cells, including the development of Th2 cells and IL-4–stimulated proliferative responses (33, 34). Induction of nuclear localization of STAT6, but not STAT1α, in splenocytes by NaB suggests that NaB is capable of activating STAT6, but not STAT1α. Abrogation of NaB-mediated upregulation of TGF-β in normal as well as PLP-primed splenocytes by siRNA knockdown of STAT6 indicates the involvement of STAT6 in NaB-mediated increase in TGF-β. Furthermore, NaB also induced the recruitment of STAT6 to a STAT6-responsive element on the Tgfb promoter.

At present, the mechanisms by which NaB leads to the activation of STAT6 in splenocytes are not known. In the case of IL-4 signaling, it binds to its cognate receptor, the type I IL-4R consisting of the IL-4R α-chain (IL-4Rα) and the common γ-chain (26). Binding of IL-4 to IL-4Rα leads to dimerization of the two receptor subunits followed by phosphorylation of tyrosine residues within IL-4Rα by Janus kinases. STAT6 monomers are known to bind via their SH2 domains to the phosphotyrosine residues within the intracellular portion of IL-4Rα and become phosphorylated by Janus kinases (26, 35). Subsequently, phosphorylated STAT6 monomers dimerize and then translocate into the nucleus, where they regulate the expression of IL-4 target genes (36). Upon translocation, STAT6 preferentially binds to STAT6 motifs in which the palindromic half-sites TTC/GAA are divided by four arbitrary bases (36). Therefore, it is possible that NaB treatment engages IL-4Rα to activate STAT6 in splenocytes. However, further studies are necessary to verify such possibilities. Earlier we have shown that NaB is also capable of switching the differentiation of MBP-primed T cells from Th1 to Th2 mode (14), which could be due to NaB-mediated activation of STAT6.

Although, in general, Th2 cells and Tregs differ by function, and in surface as well as nuclear markers, these two cell types share some overlapping functions as well. For example, both Th2 cells and Tregs secrete IL-10 (3, 37, 38). Therefore, sometimes Tregs are also expressed as IL-10–producing Tregs (3, 38). The underlying mechanisms are poorly understood. Our results suggest a possible mechanism for such overlap. It is known that STAT6 binds to the promoter of GATA3 (39), the driver of Th2 phenotype, and the IL-10 gene. The expression of GATA3 and Th2 cytokines, including IL-4 and IL-10, is diminished in STAT6 null mice (36, 40). Because STAT6 is a crucial transcription factor for the transcription of both Th2 cell–specific molecules and Treg-inducing TGF-β, STAT6 could be one of the reasons for overlapping function between Th2 cells and Tregs, serving as a common link between these two cell types.

In summary, we have demonstrated that the cinnamon metabolite NaB enriches Tregs via STAT6-mediated production of TGF-β and thereby protects mice from EAE. NaB is typically dosed for adults with urea cycle disorders at 2–5 g orally twice daily (9, 41), which is equivalent to the dose (100 mg/kg body weight per day) used in this study. Therefore, our results highlight a novel immunomodulatory role of NaB and suggest that NaB and/or its parent compound may be explored for therapeutic intervention in MS and other demyelinating disorders.

The authors have no financial conflicts of interest.