Shaoyao Gancao decoction alleviates functional constipation by inhibiting Escherichia-Shigella expansion, modulating gut microbiota, and suppressing dysbiosis-induced endocannabinoid production: evidence from a self-controlled pilot study

Twenty patients diagnosed with FC were recruited between September and November of 2021. The sample size was consistent with prior pilot microbiome studies, in which cohorts of approximately 20 participants are commonly reported (Jose et al., 2022;Mejía-Granados et al., 2022). The study was approved by the Ethics Committee of the First People’s Hospital of Taicang (KY-2019-201) and conducted in accordance with the Declaration of Helsinki. Written informed consent was obtained from all participants.

The inclusion criteria were as follows: (1) diagnosis of FC according to the Rome III (IV) criteria, (2) age ≥ 18 years; (3) disease duration > 6 months. The exclusion criteria included: (1) organic intestinal diseases, (2) history of intestinal surgery, (3) constipation caused by the long-term use of psychotropic or other medications, (4) use of probiotics, prebiotics, or synbiotics within the previous month, and (5) use of antibiotics within the last month.

Participants received a one-week oral intervention of SGD, administered once daily, one hour before bedtime. Each dose consisted of 5 g of raw licorice and 30 g of raw white peony root, prepared as granules (Jiangsu Jiangyin Tianjiang Pharmaceutical Co., Ltd.) and dissolved in 150 mL of warm water.

Self-reported questionnaires were completed before and after the intervention, assessing weekly spontaneous bowel movement frequency, stool consistency (Bristol Stool Scale), ease of defecation, and perceived overall efficacy. The Defecation Smoothness Score (graded A–C) and Efficacy Score (graded A–D) are defined in Supplementary Table 1. Responder and non-responder classification was based primarily on the objective criterion of stool frequency: participants with >3 bowel movements per week post-intervention were classified as responders, while those with ≤3 bowel movements per week were classified as non-responders. In addition, participants with both an Efficacy Score and Defecation Smoothness Score of C were also categorized as non-responders, regardless of stool frequency.

Fecal samples were collected before treatment and again 3–5 days post-treatment. The samples were immediately stored at −80 °C for subsequent microbiological and SCFA analyses.

SCFA concentrations were measured via gas chromatography as previously described (Mao et al., 2013). Briefly, 1 mL of a fecal slurry (10% w/v) was acidified with 0.2 mL of 25% (w/v) metaphosphoric acid, followed by centrifugation at 14,000 g for 5 min. The resulting supernatant was analyzed using a Shimadzu GC-2010 Plus (Japan) with a DB-FFAP column (0.32 mm × 30 m × 0.5 μm; Agilent Technologies, USA) and a hydrogen flame ionization detector. Crotonic acid was used as the internal standard. Quantified SCFAs included acetic, propionic, isobutyric, butyric, isovaleric, and valeric acids.

Bacterial genomic DNA was extracted from fecal samples using the QIAamp DNA Stool Mini Kit (Qiagen, Germany) following the manufacturer’s instructions. The extracted DNA was stored at −80 °C for sequencing.

DNA from pre- and post-intervention samples was analyzed by Novogene Bioinformatics Technology Co., Ltd. (Beijing, China). The V4–V5 region of the 16S rRNA gene was amplified using primers 515F (5’-GTGCCAGCMGCCGCGGTAA-3’) and 907R (5’-CCGTCAATTCCTTTGAGTTT-3’). Libraries were prepared using the NEBNext Ultra II DNA Library Prep Kit (Illumina, USA), and library quality was assessed via Qubit. Paired-end sequencing (250 bp) was performed on the Illumina NovaSeq platform.

Data processing—including raw read merging, quality filtering, chimera removal, and sequence assembly—was conducted using QIIME2 (version 2022.2). Amplicon sequence variants (ASVs) were generated using DADA2 or the Deblur plugin and taxonomically classified with the Silva database. Alpha diversity (Chao1, Shannon, and Simpson indices) and beta diversity (Bray–Curtis distances) were calculated in QIIME2. PCoA was visualized using ade4 and ggplot2 in R (v4.0.3).

Statistical analysis of beta diversity (Adonis) was performed using vegan, and LEfSe (LDA Effect Size) analysis was carried out in R. Spearman’s rank correlations were used to evaluate associations between microbial taxa, SCFAs, and clinical outcomes. Functional potential was inferred using PICRUSt2 (v2.3.0) against the KEGG database, and microbial phenotypes were predicted using BugBase (Ward et al., 2017;Zhang et al., 2020). The sequencing data are available in the National Microbiology Data Center under the accession number NMDC10018777 (https://nmdc.cn/resource/genomics/sra/detail/NMDC10018777↗).

All statistical analyses were conducted using GraphPad Prism 9.5.1 (GraphPad Software, Inc., San Diego, CA, USA). Wilcoxon matched-pairs signed-rank tests were used to assess differences before and after the intervention for all patients, including responders and non-responders. Mann–Whitney U tests were used for comparisons between responders and non-responders. A p-value <0.05 was considered statistically significant.

Twenty participants with FC (Rome III (IV) criteria) completed the study. Baseline and post-intervention clinical data are summarized in. 1

Following the SGD intervention, stool frequency and consistency significantly improved (Figures 1A, B). The mean number of bowel movements and Bristol Stool Scale scores increased significantly. Overall, 90% of participants reported some improvement in global efficacy, and 30% reported significant relief (Table 1). Ease of defecation improved in 95% of patients, with 65% showing marked improvement.

According to the predefined criteria (Materials and Methods;), 14 participants were classified as responders, defined as having >3 bowel movements per week post-intervention, yielding an overall efficacy rate of 70%. The remaining six participants were classified as non-responders—five with ≤3 bowel movements/week and one with both global efficacy and defecation smoothness scored as C. Due to the limited number of male participants (n = 2, one in each response group), sex-stratified analyses were not performed. 1

Responders exhibited a significant increase in both stool frequency and consistency, whereas non-responders improved only in stool consistency (Figures 1C, D). Non-responders consistently exhibited lower stool frequency at baseline and after treatment. Among the responders, 100% reported improved defecation ease, and 93% noted enhanced global efficacy (Table 1). In contrast, only 83% of non-responders reported improvements in both measures. These findings indicate that the microbial and metabolic changes observed in responders may be more closely associated with the therapeutic effects of SGD.

Microbial alpha diversity (Figure 2A) did not significantly change following SGD treatment. However, beta diversity (PCoA based on Bray–Curtis distances) showed significant changes in the overall cohort and in responders (Figures 2C, D, Adonis test), suggesting an alteration in community composition. The relative abundance at the genus level is shown in Figure 2B.

LEfSe analysis identified 24 taxa enriched at baseline and 10 taxa predominating after treatment (Figure 3A). Following the SGD intervention, Firmicutes replaced Proteobacteria as the predominant phylum. At the genus level, Escherichia–Shigella was dominant at baseline, along with its higher taxa—Enterobacteriaceae, Enterobacteriales, and Gammaproteobacteria. After treatment, genera such as Veillonella, Roseburia, and Ruminococcus became prevalent. Other genera enriched at baseline included Phascolarctobacterium, Parabacteroides, Marinobacter, Aliidiomarina, and Megasphaera.

Subgroup analysis (Figure 3B) further revealed distinct genus-level patterns between responders and non-responders. Megasphaera was enriched in non-responders alongside Escherichia–Shigella, whereas Marinobacter and Aliidiomarina were characteristic of responders at baseline. Additionally, Oscillibacter, Marvinbryantia, and Faecalicoccus predominated in non-responders, while Eggerthella, Aquamicrobium, Atopobium, Clostridium sensu stricto 1, and Halomonas were signature genera in responders. These findings suggest that differences in baseline microbial composition may underlie differential treatment responses.

After the SGD intervention, the post-treatment microbiota also differed between subgroups. Acidaminococcus, Catenibacterium, Ruminococcus sp. N15.MGS-57, and Bacteroides coprophilus were enriched in responders, whereas Veillonella, Tyzzerella, Roseburia intestinalis, and Clostridium paraputrificum predominated in non-responders. Notably, Ruminococcus sp. N15.MGS-57 was a characteristic post-treatment species in both responders and the overall cohort, with a significant increase observed only among responders (Figure 3C, right). Similarly, Escherichia–Shigella decreased in both groups, but statistical significance was achieved only in responders (Figure 3C, left).

SGD significantly increased total SCFA levels, especially acetate, propionate, and butyrate, while branched-chain SCFAs (isobutyrate, valerate, and isovalerate) showed a decreasing trend (Figure 4). Subgroup analysis showed significant SCFA increases only among responders. The decline in branched-chain SCFAs was more pronounced in non-responders, suggesting that SGD modulates SCFA metabolism more effectively in responsive individuals.

Spearman’s correlation analysis (Figure 5) revealed that baseline-enriched taxa (green) were negatively correlated with SCFA levels and clinical outcomes, whereas post-treatment taxa (red) were positively correlated.

Phascolarctobacterium and Parabacteroides were significantly negatively correlated with acetate, propionate, total SCFAs, stool frequency, and Bristol scores. Faecalicoccus pleomorphus, prominent in non-responders, was negatively correlated with propionate, total SCFAs, and Bristol scores.

In contrast, Ruminococcus sp. N15.MGS-57, B. coprophilus, and Acidaminococcus (dominant in responders post-treatment) showed strong positive correlations with SCFAs and clinical outcomes. These results suggest that SGD promotes a microbiome shift favoring beneficial taxa and metabolic profiles in responders.

BugBase analysis predicted reductions in aerobic, biofilm-forming, potentially pathogenic, and oxidative stress–tolerant bacteria after SGD intervention (Figure 6), particularly in responders. These reductions suggest partial restoration of a healthy anaerobic microbial environment.

At baseline, responders had elevated aerobes (Figure 6A) and decreased facultative anaerobes (Figure 6B), while non-responders showed increases in both aerobes and facultative anaerobes with reduced obligate anaerobes, indicating more severe anaerobic dysbiosis in the latter.

PICRUSt2 functional prediction revealed downregulation of 13 KEGG level 3 pathways and upregulation of three pathways post-treatment (Figure 7). The downregulated pathways included bacterial invasion of epithelial cells, biosynthesis of unsaturated fatty acids, arachidonic and linoleic acid metabolism, and retrograde endocannabinoid signaling. The upregulated pathways included spliceosome function, teichoic acid biosynthesis, and the one-carbon pool by folate. These findings suggest that SGD may mitigate the epithelial barrier disruption, lipid metabolic dysregulation, and aberrant nervous system signaling associated with FC.

At baseline, patients exhibited pronounced dysbiosis, with Escherichia–Shigella predominating, particularly among non-responders. This aligns with previous reports associating Escherichia–Shigella with FC (Yao et al., 2023) and constipation in autistic individuals (Strati et al., 2017). Higher-level taxa—including Proteobacteria, Gammaproteobacteria, Enterobacteriales, and Enterobacteriaceae—were also enriched, consistent with known signatures of dysbiosis linked to disrupted anaerobic conditions (Shin et al., 2015). Notably, Escherichia–Shigella remained elevated post-intervention in non-responders, who showed minimal symptomatic improvement, reinforcing its role in persistent FC (Ohkusa et al., 2019; Pan et al., 2022).

Intestinal inflammation is a key driver of dysbiosis (Lynch and Pedersen, 2016). Pro-inflammatory cytokines such as interleukin (IL)-6 and IL-1β promote metabolic shifts in epithelial cells, from mitochondrial β-oxidation to anaerobic glycolysis; this increases luminal oxygen and nitrate availability (Byndloss and Bäumler, 2018; Byndloss et al., 2017), which favors the growth of Escherichia coli (Winter et al., 2013; Wei et al., 2022). SGD has demonstrated anti-inflammatory effects in other models, including reduction of IL-1β and IL-6 in polycystic ovary syndrome (Chang et al., 2021). White peony root and licorice—two key components of SGD —exert immunomodulatory effects by inhibiting IL-6, nitric oxide (NO) production, and iNOS expression (He and Dai, 2011; Yang et al., 2015). Since elevated levels of IL-6 and IL-1β are also found in individuals with constipation (Wei et al., 2022; Cıralı et al., 2018), SGD’s anti-inflammatory properties likely contribute to the restoration of a balanced microbiota and the improvement of intestinal function.

Functional prediction analysis revealed increased retrograde endocannabinoid signaling at baseline. Endocannabinoids serve as lipid mediators that modulate excitatory neurotransmission in the gut by activating CB1 receptors (CB1Rs), thereby suppressing acetylcholine release and intestinal motility (Bisogno et al., 2003; Bashashati et al., 2015). Certain gut bacteria can produce endocannabinoid-like molecules (Degli Esposti et al., 2015); Escherichia–Shigella is known to generate glycerophospholipid-derived metabolites (Li et al., 2020). Moreover, Shigella spp. has been implicated in intractable FC via production of docosapentaenoic acid (DPA), a precursor to the endocannabinoid analog DPEA (Chen et al., 2022; Sihag and Di Marzo, 2022). The enrichment of pathways related to unsaturated fatty acid and arachidonic acid metabolism further supports the idea that there is elevated microbial endocannabinoid activity in FC. Suppression of Escherichia–Shigella by SGD likely reduces these metabolites, alleviating endocannabinoid-mediated disruptions in gut motility.

2-Arachidonoylglycerol (2-AG), the most abundant endocannabinoid (Bashashati et al., 2015;Reichardt et al., 2014), can participate in an autocrine loop. In the mouse liver, 2-AG activates CB1R, inducing estrogen-related receptor gamma (ERRγ), which upregulates diacylglycerol lipase (DAGL), the enzyme responsible for 2-AG synthesis—thereby sustaining high 2-AG levels (Jung et al., 2020). Inhibition of DAGLα restores gut motility in FC models (Bashashati et al., 2015), suggesting that a similar mechanism may operate in the gut.

Given the presence of CB1Rs and DAGLα throughout the enteric nervous system (Bashashati et al., 2015) and ERRγ expression in the ileum and colon (Modica et al., 2010), we hypothesize that microbial-derived endocannabinoids trigger a 2-AG autocrine loop that sustains elevated endocannabinoid levels and inhibits gut motility. This mechanism could explain post-infectious and laxative-refractory constipation phenotypes.

SGD significantly elevated fecal SCFAs—acetate, propionate, and butyrate—all of which are diminished in patients with FC and increased following fecal microbiota transplantation (Mars et al., 2020;Pan et al., 2022;Su et al., 2022;Tian et al., 2020).

Prior to intervention, the microbiota was dominated by asaccharolytic bacteria, such as Phascolarctobacterium, Parabacteroides, Megasphaera, and F. pleomorphus, which are poorly associated with SCFA production (Ikeyama et al., 2020;Li et al., 2023;Reichardt et al., 2014;De Maesschalck et al., 2014).

Post-intervention, responders exhibited enrichment of SCFA-producing taxa, including Bacteroides coprophilus, Ruminococcus sp. N15.MGS-57, Acidaminococcus, and Catenibacterium—each associated with fermentation and SCFA generation (Grondin et al., 2017;Su et al., 2024;Jumas-Bilak et al., 2007;Kageyama and Benno, 2000).

SCFAs activate PPARγ, support epithelial mitochondrial β-oxidation, and maintain hypoxic conditions essential for gut homeostasis (Byndloss and Bäumler, 2018;Byndloss et al., 2017). Thus, SGD facilitated a microbial shift toward fermentative, SCFA-producing bacteria, contributing to colonic homeostasis.

Ruminococcus sp. N15.MGS-57 may interfere with the 2-AG loop by modulating estrogen levels. Through β-glucuronidase activity, Ruminococcus increases circulating estrogen levels by deconjugating inactive estrogen forms (Zhao et al., 2022;Hu et al., 2023;Baker et al., 2017). Estrogen signaling can intersect with ERRγ pathways (Saito and Cui, 2018), potentially disrupting the ERRγ-DAGL-2AG feedback loop. SGD itself has been shown to elevate estradiol in animal models (Chang et al., 2021), further supporting this mechanism.

Additionally, B. coprophilus—a bile acid–resistant bacterium involved in primary bile acid biotransformation (Staley et al., 2017)—may regulate bile acid metabolism, thereby influencing lipid homeostasis and indirectly suppressing the 2-AG loop. Given that secondary bile acids accelerate colonic transit and inhibit E. coli colonization (Li et al., 2021;Li et al., 2018), these bacteria may synergistically restore motility via hormonal and metabolic regulation.

BugBase analysis revealed distinct dysbiosis patterns between responders and non-responders. Responders exhibited an imbalance between aerobes and facultative anaerobes, while non-responders displayed a broader dysbiosis with a marked loss of obligate anaerobes, suggesting a pan-enteric motility disorder potentially involving both the small and large intestines (Bassotti et al., 1996). Baseline LEfSe analysis identified Escherichia–Shigella and Clostridia- and Erysipelotrichia-related taxa as dominant in non-responders. In contrast, responders harbored aerobic and nitrate-respiring taxa such as Pseudomonadales, Rhizobiales, Desulfovibrionales, and Marinobacter, of which the latter two taxa use nitrate as an electron acceptor (Chua et al., 2018). These organisms may compete with Escherichia–Shigella for nitrate and oxygen, limiting their overgrowth. These findings indicate that baseline microbial composition may predict therapeutic responsiveness to SGD.

A key limitation of this study is its relatively small sample size, which may reduce statistical power and limit the generalizability of the findings. Nevertheless, this work was intentionally designed as a pilot, self-controlled study to generate preliminary clinical and mechanistic evidence that could inform the design of larger randomized controlled trials. Despite the small cohort size, the consistent improvements observed in both clinical outcomes and multi-omics profiles strengthen the internal validity of our conclusions. Future investigations involving larger, multicenter cohorts—ideally incorporating randomized placebo-controlled or crossover designs—will be essential to validate these findings and to refine responder stratification based on baseline microbial signatures.