This was a single-centre, randomized, double-blind, placebo-controlled clinical trial. Participants were recruited from the outpatient and ward of China-Japan Friendship Hospital from 13 February 2023, to 13 March 2024. All the subjects who participated in this study voluntarily signed informed consent forms. This study was approved by the Ethics Committee of China-Japan Friendship Hospital (2022-KY-74-1) and is registered in the Chinese Clinical Trial Registry (ChiCTR2200061608). We confirm that our clinical trial follows the CONSORT standards (http://www.consort-state ment.org/) and The Reporting Standard for Clinical Randomized Controlled Trials of Chinese Medicine Compounds (2017 edition) (Cheng et al., 2017).

Patients were recruited from the China-Japan Friendship Hospital. The inclusion criteria were as follows: (1) age 18–75 years; (2) fulfilling the 2016 American College of Rheumatology/European League Against Rheumatism (ACR/EULAR) classification criteria for primary Sjögren’s syndrome (Shiboski et al., 2017); (3) with positive auto-antibodies (antinuclear antibodies (ANA) or anti-Sjögren’s-syndrome-related antigen A (SSA) antibodies or anti-Sjögren’s-syndrome-related antigen B (SSB) antibodies or rheumatoid factor (RF)) and/or high immunoglobulin levels or accelerated erythrocyte sedimentation rates (ESR); (4) compliance with the traditional Chinese medicine syndrome of dampness and dryness; (5) informed and voluntarily signing of the informed consent form. The exclusion criteria were as follows: (1) diagnosis with other connective tissue diseases; (2) using glucocorticoids or biological immunosuppressive agents within the previous 3 months before recruitment; (3) pregnancy, planning for pregnancy or lactation; (4) severe mental illnesses or neurological disorders; (5) severe lesions in heart, brain, lungs, liver, kidneys, or haematopoietic system; (6) visual field defects, fundus lesions, and severe arrhythmias (such as atrial fibrillation, atrioventricular block); and (7) participating in other drug trials. In addition, all participants who completed the informed consent form and were screened for eligibility to enter the RCT were marked as dropouts, regardless of the time or reason for withdrawal, as long as they had not completed the treatment observation period specified in the protocol.

We adopted a superiority testing design method, and used PASS 15.0.5 software for sample size estimation. The HRD group and placebo group were set at a 1:1 ratio. The EULAR Sjögren’s Syndrome Patient-Reported Index (ESSPRI) was used as the primary efficacy endpoint. Previous clinical trial have shown that after intervention with HRD, the mean and standard deviation of ESSPRI scores in pSD patients were 2.53 and 1.20, respectively, and in the placebo group were 0.83 and 1.37, with α = 0.05, β = 0.1, SM = 1, and considering a 20% dropout rate, it was determined that 59 patients needed to be enrolled in each group, resulting in a total sample size of 118 patients.

The participants were randomly assigned to the HRD group or the placebo group at a 1:1 ratio. Participants in the HRD group received HRD and hydroxychloroquine sulfate, whereas those in the placebo group received HRD placebo and hydroxychloroquine sulfate. The participants in both the HRD and placebo groups were each administered one packet of the corresponding substance 30 min after breakfast and one packet each after dinner, whereas 200 mg of hydroxychloroquine sulfate was taken every 30 min after breakfast and 30 min after dinner. The raw materials for HRD were manufactured by Beijing Meikangtang Pharmaceutical Technology Co., Ltd., and their specific compositions are shown in Table 1. The HRD placebo was composed of 10% HRD, a flavouring agent, and an edible pigment. HRD and its placebo were prepared as Chinese herbal decoctions (100 mL/packet) by the Pharmacy Department of China-Japan Friendship Hospital. The hydroxychloroquine sulfate tablets were manufactured by Shanghai Shangyao Zhongxi Pharmaceutical Co., Ltd., and contained 100 mg of active ingredient per tablet (batch No. 119990264).

The HRD was constructed as follows: (1) Rehmannia glutinosa Libosch (15 g), Lilium lancifiolium Thunb (10 g), Poria cocos (Schw.) Wolf (15 g), Perilla frutescens (L.) Britt (10 g), Hordeum vulgare L. (10 g) and Morus alba L. (5 g) herbs were mixed and soaked in cold water for 15 min (2) After boiling for 20 min, 200 mL of liquid was poured out from the filter residue. (3) Cold water was added to immerse the medicinal residue again, and the herbs were boiled for 20 min. Two hundred millilitres of liquid was then poured from the filtered residue. (4) The two liquids are mixed and divided evenly into two portions, morning and evening, to be taken warm after meals.

The preparation process for the HRD placebo was as follows: (1) A 0.02‰ bittering agent (sucrose octaacetate), 1.26‰ colouring agent (caramel pigment), 5.2‰ polyglucose, and 0.1‰ citric acid were added to a container, water was added, and the mixture was stirred evenly. (2) A total of 10% raw HRD material solution was added and mixed well to create a simulated substance that was similar in colour, odour, and appearance to HRD. (3) The substance was mixed thoroughly and boiled before distribution and packaging.

Upon enrolment, the participants underwent a comprehensive collection of demographic data and physical examinations. Subsequently, symptom and disease activity assessments and exocrine gland tests were performed, along with laboratory tests, including complete blood cell counts, hepatic and renal function tests, tests for assessing markers of immune inflammation, and electrocardiography.

All participants performed self-assessments of their symptoms every 4 weeks and kept a diary of their self-ratings. At the 12th week, they returned to the hospital for a comprehensive evaluation. Adverse reactions were continuously monitored and documented throughout the medication period. At the 24th week, follow-up was still continued to assess their symptom scores.

The participants were randomized by experts in research methodology from the Clinical Research Data and Project Management Platform at China-Japan Friendship Hospital using SAS 9.4 statistical software through a simple random method. Each participant was assigned a random number, which was kept in a sealed envelope by a designated individual not involved in the clinical trial of the research team. The participants, physicians and statisticians were unable to access any information that could disclose the group allocation of the participants through any means. When a participant was enrolled, he or she was contacted to provide them with the random number.

The primary outcome measure was the assessment of the salivary flow rate.

The secondary outcome measures included clinical scale indicators, obtained through visual analogue scales (VAS) (including the oral dryness VAS score, dry eye VAS score, fatigue VAS score, and limb pain VAS score), at all time points (the 4th, 8th, 12th, and 24th week). The overall symptom assessment was evaluated using the ESSPRI at all time points. Additionally, disease activity was assessed using the EULAR Sjögren’s Syndrome Disease Activity Index (ESSDAI) and the clinical ESSDAI (ClinESSDAI). Exocrine gland function was assessed with Schirmer’s test. In addition, immune-inflammatory indicators such as ESR, C-reactive protein (CRP) level, immunoglobulin G (IgG) level, IgA level and IgM level were measured. Quality of life was assessed using the short form-36 health survey (SF-36) and Sjögren’s tool for assessing response (STAR). Unless indicated otherwise, the above indicators were evaluated only at week 12.

Peripheral blood samples were collected from included patients who gave additional informed consent before and after intervention. Meanwhile, samples were also collected from healthy volunteers with matched gender and age. All healthy volunteers were recruited from the Health Examination Center of China-Japan Friendship Hospital, and their samples were only collected at the time of enrollment. Leukocytes from the peripheral blood were isolated using a density gradient centrifugation, followed by RNA extraction from the isolated leukocytes with TRIzol reagent.

The RNA concentration and purity of samples were measured with a NanoDrop One ultramicro spectrophotometer (Thermo Fisher Scientific, Waltham, MA, United States). The experiments and data analysis for m6A and m5C RNA modifications were conducted by Aksomics through their liquid chromatography–mass spectrometry (LC-MS) mRNA modification detection service. Five postintervention samples each from the placebo-treated and HRD-treated patients, as well as five from healthy volunteers, were selected from the samples for RNA modification regulator analysis. The experimental procedures and data analysis were conducted by Aksomics using a Human Epitranscriptomics PCR array.

The t-test was conducted to assess the impact of HRD on the m6A and m5C RNA modifications in pSD patients. Differential expression gene (DEG) analysis was carried out according to a screening criterion of |logFC| >1, adjusted P < 0.05, to identify key regulators in the experiment. The target genes of these key regulators were subsequently searched in the RM2Target database (http://rm2target.canceromics.org), and validated target genes were selected for GO enrichment analysis. Finally, Spearman analysis was employed to evaluate the associations between key regulators and patient clinical characteristics.

GraphPad Prism eight was used for statistical analysis. Count data are described as [n (%)]. Normally distributed measurement data are described as ( x ¯ ± s ), whereas data not conforming to a normal distribution are described as [M (P25, P75)]. Intergroup comparisons of count data were performed with the chi-square test, Yates’ corrected chi-square test, or Fisher’s exact test, as appropriate. If the data could be described with a parametric model, the two-sample independent t-test was employed for between-group comparisons; otherwise, the Mann‒Whitney U test was used. For within-group sample data that could be described with a parametric model, the paired t-test (two-tailed) was used for analysis; if not, the Wilcoxon matched-pairs signed-rank test was employed. Repeated measures indicators were evaluated with a mixed-effect model for repeated measures (MMRM) using the “nlme” R software package. A two-sided P value <0.05 was considered to indicate statistical significance.

From 13 February 2023, to 13 March 2024, a total of 121 potential participants at the China-Japan Friendship Hospital were assessed for eligibility, among whom 118 participants entered the RCT and were assigned to the HRD group or placebo group at a 1:1 ratio (Figure 1). In the HRD group, one patient withdrew from the trial before taking the medicine because a thyroid tumor was discovered simultaneously. In placebo group, one patient withdrew from the trial before taking the medicine due to personal reason. Therefore, 116 patients were included in this study. However, three patients in HRD group and two patients in placebo group withdrew from the trial after taking the medicine due to personal reasons. There were 111 patients included in per protocol analysis set. No adverse reaction symptoms occurred during the medication period of the above five patients who have withdrawn from the trial. All the 116 patients completed the assessment of symptom scores at the 24th week. There were no statistically significant differences in the baseline levels of each outcome indicator between the two groups (Table 2).

Table 3 lists all outcome results for the HRD group and placebo group. After 12 weeks of HRD treatment, the salivary flow rate was greater than at baseline (P = 0.002), whereas no differences from baseline were observed in the placebo group (P = 0.678). The salivary flow rate from week 0 to week 12 in the HRD group (median improvement: 0.009) significantly differed from that in the placebo group (median improvement: 0.000), P = 0.049 (Figure 2).

In terms of secondary outcome measures, the clinical scale indicators in the HRD group, including the oral dryness VAS score (P < 0.001), dry eye VAS score (P < 0.001), dryness VAS score (P < 0.001), fatigue VAS score (P < 0.001), limb pain VAS score (P = 0.003) and ESSPRI score (P < 0.001), all decreased from baseline to week 12. After treatment, the HRD group showed significant improvements in the oral dryness (median improvement: 1.0 for the HRD group and 0.0 for the placebo group, P < 0.001), dry eye (median improvement: 1.0 for the HRD group and 0.0 for the placebo group, P = 0.004), dryness (median improvement: 1.0 for the HRD group and 0.0 for the placebo group, P = 0.001), and limb pain VAS scores (median improvement: 0.5 for the HRD group and 0.0 for the placebo group, P = 0.008), and the ESSRRI (median improvement: 1.0 for the HRD group and 0.0 for the placebo group, P = 0.001) relative to the placebo group, whereas there was no difference in the fatigue VAS score between the two groups (median improvement: 1.0 for the HRD group and 1.0 for the placebo group, P = 0.238) (Figures 3a,b). MMRM analysis revealed that the fatigue VAS scores of the HRD group and the placebo group gradually decreased over the treatment period but rebounded during the follow-up period with similar trends. Moreover, there was no group and time point interaction effect. However, during the treatment period, other clinical scale indicators in the HRD group gradually decreased over time. At the 12th week, the decrease in the HRD group was more significant than that in the placebo group, and significant group and time point interaction effects were observed over the entire trial period (Table 4; Figure 3c).

With respect to disease activity, glandular assessment, and serological parameters, after 12 weeks of HRD treatment, the ESSDAI (P = 0.022) and ClinESSDAI (P = 0.011) were lower than at values, whereas no such differences were observed in the placebo group (P = 0.094 and P = 0.112, respectively). There was no difference in the ESSDAI or the ClinESSDAI between the two groups at week 0, but there was a significant difference in the ClinESSDAI between the HRD group and the placebo group at week 12 (median values: 0.0 and 2.0, respectively; P = 0.038) (Figure 3e).

In terms of glandular function, in addition to the primary outcome measure of the salivary flow rate, there were also significant changes in the HRD group in the results of Schirmer’s test. Compared with those at week 0, the results for the HRD group were significantly greater at week 12, including those of the left eye (P = 0.017) and right eye (P = 0.001) and their mean values (P = 0.001), whereas those of the placebo group were not different from those at baseline. The change in the Schirmer’s test result from week 0 to week 12 in the HRD group (median improvement: 0.5) significantly differed from that of the placebo group (median improvement: 0.0), P = 0.005, both for the left eye (median improvement: 0.0 and 0.0, respectively, P = 0.015) and the right eye (median improvement: 1.0 and 0.0, respectively, P = 0.013) (Figure 3d).

None of the serological parameters were significantly different between the two groups before treatment (Supplementary Figure 1).

The scores for each dimension in the assessment of patients’ quality of life are shown in Table 5. The HRD group showed significant differences between week 0 and week 12 in four dimensions: emotional function (P = 0.048), physical pain (P = 0.026), overall health (P = 0.007), and health change (P < 0.001), whereas the placebo group only showed improvement in the health change dimension after 12 weeks of treatment compared with baseline (P = 0.001) (Supplementary Figure 2).

Among the 116 patients included in this study, three patients in HRD group and two patients in placebo group refused to undergo the examination and assessment at the 12th week due to personal reasons. Therefore, the STAR scale was used to evaluate a total of 111 participants in the two groups, with six individuals in the HRD group and three in the placebo group meeting the standard of clinical efficacy. No significant difference was observed between the groups.

In the study, a total of 11 participants experienced adverse events, with three occurring in the HRD group and eight in the placebo group. In the HRD group, two participants reported diarrhea and one reported nausea and vomiting. In the placebo group, three participants reported diarrhea, two reported nausea, and three reported rash. All adverse events resolved spontaneously and were considered unrelated to the intervention. There was no significant difference in the incidence of adverse event between the groups.

To further investigate the epigenetic mechanisms underlying the effects of HRD treatment in pSD, a total of 93 peripheral blood samples from patients and healthy volunteers who gave additional informed consent were collected in this study. After excluding patients who only provided blood samples once, 70 peripheral blood samples from patients before and after treatment (including 18 patients in HRD group and 17 patients in placebo group) and 18 peripheral blood samples from 18 healthy volunteers were collected. All of the 88 samples were used to examine the levels of m6A and m5C in the PBMCs. Compared with healthy volunteers, individuals in both the HRD group and the placebo group presented with lower m6A levels and higher m5C levels before treatment (P < 0.05). Compared with the corresponding levels at baseline, patients with pSD had increased m6A levels (P = 0.003) and decreased m5C levels (P = 0.003) after HRD treatment, whereas no significant differences were observed in the placebo group posttreatment (Table 6; Figures 4a,b).

We compared the differentially expressed regulators among healthy volunteers, the HRD group and the placebo group and found that HNRNPA2B1 and TET1 were significantly upregulated in the placebo group; moreover, both regulators were significantly downregulated after HRD treatment. Further comparison of the differentially expressed regulators between the HRD group and healthy subjects revealed that TET1 was still upregulated, but there was no significant difference in the expression of HNRNPA2B1 (Figure 4c). GO enrichment analysis of the target genes revealed that they were primarily enriched in “apoptosis”, which also included gland development and other features related to pSD (Figure 4d). Finally, Spearman correlation analysis revealed that the expression of HNRNPA2B1 was significantly correlated with the serum RF level in patients with HRD (r = 0.975, P < 0.01) (Figure 4e).