DBA/1J female mice, 8–12 weeks old, were supplied by Biogem SCARL (Ariano Irpino, Italy). C57BL/6 female mice and additional female DBA/1J mice were purchased from Charles River (Wilmington, MA, United States) and housed in the animal facility of the University of Perugia. CIA was elicited by intradermal immunization with 50 μL/mouse collagen type II (Sigma, C9301-5MG, Lot#016M4158V; St. Louis, MO, United States) emulsified in complete Freund’s adjuvant (Sigma, F5881-10ML, Lot#SLBQ1106V) on day 1 and in incomplete Freund’s adjuvant (Sigma, F5506-10ML, Lot#SLBL9742V) on day 31. Arthritis was induced in 15 mice (3 groups of 5 mice each), while 3 mice composed the untreated negative control group. An additional 12 mice (4 mice/group) were immunized and utilized for gene expression profiling. Complete details of CIA mouse generation are provided in the Supplementary Material. Details of the CIA experimental procedure, with the arthritis score for each mouse, are given in Supplementary Material. Animal experiments were performed after approval by the ethical committee of the University of Perugia.

Leaves of A. tonkinensis were harvested in Hanoi, Vietnam in October 2015, and the voucher species (No. AT-2015) was deposited in the Institute of Chemistry, Vietnam Academy of Science and Technology. Briefly, freshly harvested leaves were air-dried for 2 days and oven-dried at 40–50°C in a forced-air oven. To obtain the decoction, 5 g dried A. tonkinensis leaves were boiled 3 times in a total volume of 100 mL (5 g/100 mL final concentration) distilled water. The decoction was then given to mice ad libitum instead of water (At-treated mice). For in vitro studies, the decoction was prepared by boiling dried leaves in RPMI-1640 medium and filtering the total extract with a 0.22-μm syringe filter (5 g/100 mL final concentration). The TAT-2 compound was isolated as previously described (Pozzesi et al., 2011).

One knee joint per mouse was removed, fixed in 10% (v/v) formalin for 24 h, decalcified in 5% (v/v) trichloroacetic acid for 7 days, dehydrated, embedded in paraffin, sectioned into 3- to 4-μm-thick sections, and stained with hematoxylin and eosin.

The Qiagen Inflammatory Response and Autoimmunity RT² Profiler PCR Array was utilized according to the manufacturer’s protocol (Valencia, CA, United States). Briefly, the conversion of experimental mRNA to first-strand cDNA was accomplished using the RT² First Strand Kit. Next, cDNA was added to the appropriate RT² SYBR Green Mastermix, which was aliquoted into wells of the RT² Profiler PCR Array. Relative gene expression was determined using the ΔΔCt method.

RNA was isolated using the Qiagen RNeasy Plus Micro kit, and conversion of total RNA to cDNA was performed with the QuantiTect Reverse Transcription kit (Qiagen). RT-PCR was performed with the ABI-7300 Real-Time Cycler (Applied Biosystems, Foster City, CA, United States), and amplification was achieved using the TaqMan Assay (Mm00439618 for IL-17a, Mm04209823 for IL-22, Mm01261022 for RORC, Mm00438864 for FasL, and Mm00434256 for IL-2 (Thermo Fisher Scientific, Waltham, MA, United States). The ΔΔCt method was used to determine the expression levels of IL-17a, IL-22, IL-2, FasL, and RORC).

CD4⁺ T cells were isolated (>95% purity, see Supplementary Material) from single-cell suspensions of mouse spleens using the Naive CD4⁺ T Cell Isolation kit (Miltenyi Biotec, Bergisch Gladbach, Germany). After isolation, cells were seeded on 24-well plates at a concentration of 1 × 10⁶ cells/well and maintained in RPMI 1640 medium with 10% fetal bovine serum, 100 U/mL penicillin, and 100 μg/mL streptomycin at 37°C and with 5% CO₂ for 6 days. CD4⁺ T cells were activated with anti-mouse CD3 and anti-mouse CD28 (eBioscience, Thermo Fisher Scientific). For Th17 polarization, we used anti-mouse IL-12, IL-23, p40 (eBioscence), anti-mouse IL-4, (eBioscence), recombinant murine IL-6 (Peprotech, Rocky Hill, NJ, United States), and recombinant human TGF-β1 (Peprotech).

A Spectra System HPLC system (Thermo Separation, San Jose, CA, United States), fitted with a quaternary pump module (P4000), an online degasser, and a diode array detector (DAD) SpectroSystem UV 6000lp (Thermo Separation.) was used. Analytes were separated using a reversed-phase Agilent Zorbax ODS column (5-μm particle size, 3.0 × 150 mm i.d.; Agilent Technologies, Milan, Italy), coupled to a 20 × 4.6 mm C18 guard column (Blasi et al., 2016). Gradient elution with a flow rate of 1 mL/min was used. The mobile phase consisted of the following: (A) water containing 0.1% formic acid and (B) acetonitrile containing 0.1% formic acid. The initial mobile phase composition was 95% A. The percentage of B was linearly increased to 20% at 30 min and to 55% at 50 min. Finally, the percentage of B was reduced to 5% and the column re-equilibrated to the initial conditions for 7 min. The injection volume was 20 μL. Detection was performed online using DAD in the range from 200 to 700 nm. Chromatograms were acquired and data were handled using Xcalibur software version 1.2 (Finnigan Corporation 1998–2000, San Jose, CA, United States). A standard solution containing catechin, kaempferol, kaempferol-3-O-glucoside, quercetin, quercetin-3-β-D-glucoside (all obtained from Extrasynthese, Genay, France), and TAT-2 (isolated by us) was used to identify and quantify the analytes. Calibration curves were generated using three injections at different concentrations ranging from 1.5 μg/mL to 120.0 μg/mL. TAT-6 was identified by comparing the UV-Vis spectrum with data from the literature (Dang et al., 2009) and quantified using the calibration curve for TAT-2.

Statistical significance was determined using the Mann-Whitney U-test or Student’s t-test as specified in the figure legends. Differences were considered statistically significant according to the following criteria: ^∗p < 0.05; ^∗∗p < 0.01; ^∗∗∗p < 0.001.

To determine the chemical composition of the At water extract (decoction) both qualitatively and quantitatively, three independent HPLC experiments were performed. Seven compounds were identified, and the quantities of each compound were consistent among the three experiments (Figure 1A). The most abundant compounds were TAT-2 and TAT-6, with no significant difference in quantity between the two compounds (mean ± standard deviation, 1225.0 ± 16.5 μg/mL and 142.8 ± 17.0 μg/mL, respectively, p = 0.29) (Figure 1B). Thus, quality monitoring showed that At water extract composition was quite constant among experiments.

Mice were classified into the following groups: (1) not immunized or treated with At (Control) (n = 3); (2) immunized to induce CIA but not treated with At (CIA) (n = 5); (3) immunized and treated with At from the first day of immunization [CIA + At (from day 1)](n = 5); and (4) immunized and treated with At after development of CIA [CIA + At (from day 43)](n = 5). Figure 2A shows that At consumption was generally consistent, with a peak at week 4 (the time of the second immunization): the mean decoction (50 g/L) consumption was 4.4 mL/day/mouse. Consumption of At decoction by CIA mice was higher than that of water by CIA mice, as measured in our sister experiment (see Supplementary Material). The animals’ body weight increased 8–10% by the end of the study, with no significant differences among the four groups (Figure 2B).

Arthritis induction was scored following the criteria outlined in Supplementary Table S1. Starting at day 32, signs of arthritis appeared, with scores ranging from 1 to 2. Arthritis development became more evident on day 43, with scores ranging from 1 to 4. We further monitored CIA progression through weekly calculation of the arthritis score and observed a markedly increased arthritis score in the CIA-induced mice compared to controls not attributable to normal ankle growth (Figure 3A). Concomitantly with CIA development, we noted clinical signs of suffering that did not compromise the animals’ overall welfare, likely due to limb pain. As the experimental period continued, CIA mice became less active and more easily manipulated. Figure 3B presents the mean arthritis scores of the experimental mice from CIA onset (day 32). At day 46, scores significantly differed between CIA vs. CIA + At (from day 1) and vs. CIA + At (from day 43) mice (p < 0.001 and p < 0.05, respectively), with approximately 50% decreased scores in At-treated mice compared to untreated mice. Although arthritis development became evident in CIA mice at day 32, administration of At from either day 1 or day 43 significantly reduced CIA progression as evaluated by joint edema.

Figure 3C illustrates joint histology in control and experimental mice. Cartilage was intact and smooth in healthy mice, and no inflammatory infiltrate was present. Mice with CIA exhibited altered cartilage with many visible niches and dramatic inflammatory infiltrate. When At decoction was administered before CIA development, cartilage was intact and smooth as in the control mice, and only a mild inflammatory infiltration was observed. In mice treated with At after CIA development, cartilage was altered similarly to that of CIA untreated mice, but the inflammatory infiltrate was reduced. Thus, At decoction prevents CIA if given before appearance of clinical signs and limits CIA progression after arthritis development, although At does not reverse CIA-related cartilage damage.

Expression profiles of genes involved in autoimmunity were generated from mRNA extracted from joints of CIA and [CIA + At (from day 1)] mice (Figure 4A). We identified 29 significantly modulated genes, including one upregulated gene encoding CCAAT/enhancer binding protein β, whose expression increased approximately 3-fold with At treatment. The 28 downregulated genes included those encoding 9 chemokines (CCL1, CCL20, CCL22, CCL3, CCL4, CXCL1, CXCL5, and CXCL9), 6 chemokine receptors (CCR1, CCR4, CCR7, CXCR1, CXCR2, and CXCR4), 10 cytokines and cytokine receptors (IL-10, IL-17a, IL-18, IL-1a, IL-1b, IL-1RN, IL-22, IL-23r, IL-6), 3 proteins related to apoptosis (FasL, Ripk2, and TNFSF14), and the kininogen KNG1. Expression of genes encoding IL-17 and IL-22 was not detected in At-treated joints, and CCL20 and IL-6 expression decreased more than 10 fold (Figure 4A). These four downregulated genes attracted our attention because they are involved in the differentiation and/or function of Th17 cells. Specifically, the proinflammatory cytokines IL-17 and IL-22 are normally produced by Th17 cells (Pelaia et al., 2012), IL-6 promotes differentiation of Th0 cells to Th17 cells (Pelaia et al., 2012), and CCL20 enhances Th17 cell migration (Kaneko et al., 2018). These gene expression profiles suggest At differentially attenuates Th17 cell-associated immune responses in the joints of CIA mice.

Our gene expression results prompted us to further evaluate expression of Th17 cytokines and T cell activation markers in the lymph nodes of control and experimental mice via RT-PCR. We specifically measured expression of genes encoding IL-2, an activated T cell cytokine (Pozzesi et al., 2007), IL-17, IL-22, and FasL, a marker of T cell activation and apoptosis. Figure 4B shows that expression of IL-17 was significantly inhibited in all At-treated mice, whereas IL-2, IL-22, and FasL were significantly inhibited in the day 1 At-treated mice only. These expression patterns reflect the partial protective effect of At administered in mice after appearance of CIA signs observed in our histological analysis.

To determine the effects of At ingestion in healthy DBA1J mice (same types of animals used for the CIA in vivo experiments), At decoction was administered to these mice at the same concentration used for the aforementioned CIA in vivo experiments. After 10 days, the mice were sacrificed, and their spleen, lymph node, and joint cells were harvested. RNA was extracted for further analysis. No macroscopic differences were observed in the mice or their organs between water-treated and At-treated mice. There were no significant differences between the number of spleen and lymph node cells from water- or At-treated mice. As shown in Figure 5A, RT-PCR analyses showed no significant differences in mRNA expression of IL-6, IL-17, or IL-22 in the spleen, lymph node, or joint cells between animal groups. Similarly, no significant differences were detected between water- and At-treated mice in the differentiation of isolated CD4⁺ T cells to Th17 cells, although more variability was observed in the At decoction group (Figure 5B). Thus, 10 days of At decoction was not associated with any significant differences compared to 10 days of water, suggesting that At decoction has no acute toxic or other effects in healthy DBA1J mice.

To determine whether inhibition of Th17 cells is a specific property of At water extract, we also analyzed the role of the decoction in the in vitro development of Th17 cells. The results of these experiments were not meant to replace the in vivo results but to provide further support for the findings of the in vivo CIA model suggesting that At decoction attenuates Th17 responses. Splenic CD4⁺ T lymphocytes were isolated from mice and cultured in the following groups: (1) in anti-CD3-adhered plates with anti-CD28 antibodies to stimulate T cell activation; (2) the same as group 1 with the addition of IL-6 and TGF-β to promote differentiation to Th17 cells and the addition of anti-IL-4 and anti-IL-12 monoclonal antibodies (mAbs) to block differentiation to Th1 and Th2 cells, respectively; and (3) same as group 2 with the addition of At. Expression of IL-17, IL-22, and RORC, a Th17 cell-specific transcription factor, in the presence or absence of At was then measured by RT-PCR. As shown in Figure 6A, At significantly decreased the total number of Th17 cells compared to those in untreated cells. Th17 polarization was successful in control cells, but not in At-treated cells, as indicated by the At-mediated inhibition of IL-17 and RORC (Figure 6B). Thus, At inhibited the in vitro differentiation of Th17 cells from splenic CD4⁺ T cells.

Since TAT-2 (Pozzesi et al., 2011) is one of the two major components of the At decoction (Figure 1), the effect of TAT-2 on differentiation of Th17 cells from splenic CD4⁺ T lymphocytes was assayed. As shown in Figure 6C, TAT-2 at concentrations of 0.01 and 0.001 mg/mL decreased expression of IL-17 and RORC, and as the primary component of the decoction, likely contributed to the reduced Th17 differentiation in At-treated mice.