Postpartum depressive disorder (PPD) is a common, disabling, but treatable psychiatric condition (Howard et al., 2014). However, without prompt diagnosis and treatment, maternal suicide and infanticide may be extreme outcomes of PPD (Grigoriadis et al., 2017; Netsi et al., 2018). With a global prevalence of 17.22% (Wang et al., 2021), PPD is a significant maternal and family health burden worldwide. While PPD is most commonly treated with antidepressants and psychological therapies, the efficacy of these approaches varies due to high clinical and functional heterogeneity (Consortium, P.D.A.T.C.a.T.P, 2015; Santos et al., 2017). Indeed, antidepressants have been reported to be only ~42% effective (Brown et al., 2021), and psychotherapy only benefits about a third of PPD patients (Huang et al., 2020). Furthermore, antidepressants have side effects (Carvalho et al., 2016), and any adverse events to the baby during lactation must also be considered (Davanzo et al., 2011). Psychotherapy cannot generally be widely used due to its high cost over long periods of time. An increasing number of PPD patients are seeking safe and effective complementary treatments with few side effects.

Acupuncture is safe and effective in pregnant women (Ormsby et al., 2016; Li et al., 2018; Tong et al., 2019; Li et al., 2020). In a study of 31 meta-analyses and 59 randomized controlled trials, acupuncture was shown to be superior to awaiting treatment, control acupuncture (invasive or non-invasive sham control), and antidepressants in terms of reducing the severity of depression (Hamilton, 1960; Li et al., 2020). Another relatively recent meta-analysis highlighted that acupuncture can significantly reduce Hamilton depression rating (HAMD) scores in PPD patients (Li et al., 2019). However, just like other treatments, the efficacy of acupuncture varies between individuals. PPD therapy urgently requires specific biomarkers to predict therapeutic responses to antidepressant treatments, including acupuncture, so that the correct patients can be prescribed the best treatments at the right time.

Several demographic and clinical therapeutic response predictors to traditional antidepressants in PPD have been reported in robust clinical trials including being white/non-Hispanic (Yonkers et al., 2008), having a major depressive episode within four weeks of delivery (Sharp et al., 2010; Hantsoo et al., 2014), concomitant anxiety symptoms (Cohen et al., 2001), an absence of concomitant psychiatric illness (Yonkers et al., 2008), early response to treatment (Cohen et al., 2001), and improvement within one week after initiation of antidepressants (Appleby et al., 1997). Predictors of non-response included a lifetime history of substance use disorder (Yonkers et al., 2008), concomitant anxiety symptoms (Nonacs et al., 2005), and Hispanic or Black ethnicity (Yonkers et al., 2008). However, the generalizability of these predictors is limited by significant methodological variability including a wide range of studied postpartum periods (2–24 months), comorbid diseases (lifetime alcohol abuse, alcohol dependence, drug abuse, drug dependence, or anxiety disorder) (Nonacs et al., 2005; Suri et al., 2005; Misri et al., 2016), and different severities of depression (minor depression or major depression) (Appleby et al., 1997). In recent years, it has been found that the consistency of quantitative electroencephalographic, the default pattern network with different discrete topological structures in the left and right hemispheres and the variance of the global signal are related to the terminal clinical results of antidepressant treatment of MDD (Hunter et al., 2010; Hou et al., 2016; Zhu et al., 2018), however, the acquisition of the above indicators is undoubtedly complicated, and there are few research results on the efficacy prediction of PPD. Furthermore, predictors focus on clinicodemographic factors and there have been few studies on biomarkers– such as genetic and inflammatory markers (Sharma et al., 2020) – to advance the goal of developing objective and clinically acceptable biomarkers that predict treatment outcomes and guide individualized therapy.

There is now mounting evidence supporting a role for the intestinal microbiota in mental health disorders (Rieder et al., 2017; Dubois et al., 2019; Nikolova et al., 2021; McGuinness et al., 2022). This biochemical signaling pathway, also known as the gut-brain axis, is thought to influence cognitive function and mood via neural, metabolic, hormonal, and immune-mediated mechanisms (Foster and McVey Neufeld, 2013). Previous studies (Chung et al., 2019; Zhou et al., 2020; Nikolova et al., 2021) have found differences in the diversity and composition of gut microbial communities between PPD patients and healthy controls. Changes in the intestinal microflora can affect the efficacy of treatment for some diseases (Ma et al., 2019), and intestinal microflora has recently been shown to be a non-invasive diagnostic biomarker for colorectal adenoma and cancer (Liang et al., 2020). In a systematic review, probiotic therapy showed modest benefits in alleviating depressive symptoms in patients with major depressive disorder over four to nine weeks (Alli et al., 2022). Furthermore, Lactobacillus rhamnosus HN001 administered as a probiotic significantly reduced maternal depression and anxiety scores (Slykerman et al., 2017). It is also found that 919 syrup can relieve PPD by regulating the structure and metabolism of intestinal microorganisms and affecting the function of GABA/glutamic acid system in hippocampus (Tian et al., 2021). Additionally, it can be also used to predict responses to cancer immunotherapy in metastatic melanoma patients (Limeta et al., 2020), and dynamic changes in the intestinal microbiota can provide an early prediction of immunotherapy outcomes in patients with hepatocellular carcinoma (Zheng et al., 2019). Recently, some studies have also shown that responses to antipsychotic drugs are related to gut microbiota composition (Schwarz et al., 2018). It is found that the changes of intestinal microbial composition and metabolic function may be related to the response of antidepressants, which provides a potential predictor for the prediction of the curative effect of MDD and can even be used to distinguish MDD from generalized anxiety disorder (Dong et al., 2021; Dong et al., 2022). Therefore, characterizing the nature and impact of the intestinal microbiota on PPD therapy and its value as a biomarker of therapeutic responses would be highly clinically valuable. Acupuncture, as a common complementary alternative therapy, can reduce depressive-like behaviors in chronic unpredictable mild stress (CUMS) rats by regulating intestinal microbes and neurotransmitters (Li et al., 2021). Jiang found that acupuncture can effectively treat all stages of stroke and regulate intestinal flora, thus improving depressive symptoms (Jiang et al., 2023). Therefore, the intestinal microflora may act as clinically relevant biomarkers of therapeutic responses in individuals with mental health diseases, including in those receiving acupuncture.

Here, we first aimed to assess the efficacy of acupuncture in PPD patients. A secondary aim was to identify any differences in the intestinal microbiota in responders and non-responders to acupuncture, with the objective to identify microbiome-based predictors of acupuncture response.

Among 179 patients screened, 88 were enrolled at baseline between March 25 and November 22, 2021. According to patient preference, 60 received acupuncture treatment (acupuncture group) and 28 received no treatment (control group). During the study, nine (10.23%) patients dropped out: five (8.33%) received acupuncture and four (14.29%) had not received acupuncture. Seventy-nine patients completed the eight-week follow-up and assessments ( Figure 1 ).

This trial showed that acupuncture alleviated depressive symptoms in patients with PPD over an 8-week treatment period, and 47.27% patients significantly responded to acupuncture treatment. Additionally, based on gut microbiota profiling, we successfully predicted responses to acupuncture and improvements in clinical symptoms in PPD patients after treatment. Desulfovibrio and Paraprevotella were identified as specific predictive genera.

So far, there have been few reports on the predictors of therapeutic effect of PPD. Pinna speculated the neurosteroid biosynthesis and endogenous cannabinoid system might be able to predict antidepressive treatment, but lack of rigorous experimental studies to confirm this idea (Pinna, 2023). Additionally, there are also some limited explanations, such as the influence of running on cortisol, which can also affect the therapeutic effect of PPD in the later stage (Gobinath et al., 2018). These prediction methods are often incomplete and indirect. Therefore, gut microbiota as a predictor is seem to be more objective.

In the present study, our results further confirm the close relationship between the gut microbiota and mental health disorders. It has long been known that the enterotype distribution varies according to depression status, with Bacteroides enteritidis type 2 more prevalent in depressed patients than healthy controls (Valles-Colomer et al., 2019). It has also been shown that intestinal microbiota disorder is a characteristic of major depressive disorder (MDD) patients (Zheng et al., 2016; Zheng et al., 2020). Duan et al. studied treatment responses to escitalopram in a CUMS mouse depression model, comparing changes in metabolic function before and after treatment, and found that treatment responses were related to microbial composition, providing new insights into the mechanisms underlying variable antidepressant efficacy (Duan et al., 2021). Therefore, the flora structure is closely related to the intrinsic pathobiology of depression, suggesting that intestinal microbial biomarkers may be good predictors of antidepressant treatment responses.

Most studies employ a multifaceted approach to characterizing the gut microbiota, usually including measures of both α- and β-diversity. α-diversity is commonly used as a surrogate of community stability and function, which are thought to be beneficial to the host (Shade, 2017). Jiang et al. found that the intestinal microflora α-diversity was higher in antidepressant drug non-responders than responders in MDD patients compared with healthy controls (Jiang et al., 2015). In addition, the α-diversity of the gut microbiota was not significantly different in MDD patients with different treatment responses (Dong et al., 2022). In our study, we found that there were no significant differences in diversity between PPD patients who did and did not response to acupuncture, nor were there differences in the abundance and uniformity of the gut microbiota between the two groups. This mirrors the inconsistent results of previous studies, and the specific reasons underlying these differences need further study.

β-diversity reflects relationships between samples by analyzing the species composition and abundance (Anderson et al., 2011). Our β-diversity analysis showed that responders were significantly separated from non-responders, and the responder group had a more similar species composition. Kelly et al. and Zheng et al. both reported significant differences in β-diversity between individuals with depression and healthy controls (Kelly et al., 2016; Zheng et al., 2016). In the CUMS-induced depression study in mice, the β-diversity was also different between non-responders and responders (Duan et al., 2021), as was the β-diversity in patients who did and did not benefit from anti-programmed cell death protein 1 (PD-1) immunotherapy (Mao et al., 2021). The latter study found that the intestinal microflora affected the spectrum of immunotherapy-related adverse events, with high species diversity and relative abundance perhaps protective against immunotherapy-related adverse events (Mao et al., 2021).

We further analyzed and identified specific genera associated with acupuncture treatment responses. At the genus level, Desulfovibrio and Paraperevotella were enriched in responders, consistent with previous studies reporting a high abundance of Paraprevotella and Desulfovibrio at the genus level in patients with depression (Naseribafrouei et al., 2014; Chen et al., 2018a; Chen et al., 2018b). Desulfovibrio are present in the oral and intestinal tract of approximately 50% of people, where they release hydrogen sulfide as a product of sulfate reduction (Devereux et al., 1990). Hydrogen sulfide is involved in the natural prevention of many digestive tract diseases (Pires et al., 2006), and there is a well-established link between desulfurization bacteria and individual intestinal diseases (Verstreken et al., 2012). For example, Scanlan et al. found significantly more desulfurization bacteria in the feces of colon cancer patients than healthy people (Scanlan et al., 2009), and similarly Rowan et al. found a significantly higher relative abundance of desulfurization bacteria in the intestinal tracts of patients with ulcerative colitis than those of healthy controls (Rowan et al., 2010). Additionally, we discovered that Paraprevotella was enriched in the gut microbiota of patients responding well to acupuncture. Paraprevotella belongs to the Prevotellaceae family, and another family member Prevotella is associated with a healthy plant-based diet and probiotic use (Ley, 2016). Prevotella can also act as an opportunistic pathogen associated with periodontal and dental inflammation, intestinal inflammation, rheumatoid arthritis, and bacterial vaginitis (Arweiler and Netuschil, 2016; Randis and Ratner, 2019; Bertelsen et al., 2021; Jia et al., 2021).

Desulfovibrio and Paraperevotella have different potential pathogenic mechanisms. For example, Desulfovibrio organisms co-cultured with human oral epidermoid carcinoma (KB) cells increased interleukin (IL)-6 production, implicating them in immune responses (Bisson-Boutelliez et al., 2010). Colonization of the intestine with Prevotella leads to metabolic changes in the microbiota that reduce IL-18 production (Iljazovic et al., 2021), thus aggravating intestinal inflammation and possibly leading to systemic autoimmunity. Furthermore, Prevotella can damage intestinal mucosal barrier function by producing sulfatase, which induces and degrades mucus, thus helping itself and other harmful bacteria to access intestinal epithelial cells to generate local inflammation (Wright et al., 2000). In addition, these two genera as Gram-negative bacteria might help to explain the role of microbiota in the development/maintenance of depression. Gram-negative bacteria contain lipopolysaccharides in the outer cell membrane leaflet (Al Bander et al., 2020), and lipopolysaccharides interacts with macrophages and stimulates immune responses through pro-inflammatory cytokine release. Supporting this, increased levels of proinflammatory cytokines including IL-1β and IL-6 and decreased levels of anti-inflammatory cytokines including IL-4 and IL-10 have been detected in people living with depression (Berk et al., 2013; Wong et al., 2016).

An increasing number of studies show that the occurrence and development of depression are closely related to inflammation and immunity (Simmons and Broderick, 2005; Maes et al., 2012; Kelly et al., 2015), Inflammatory cytokines and kynurenine pathway have been found as potential therapeutic targets for PPD, because the increase of plasma IL-6 and IL-8 and the decrease of serotonin, IL-2 and quinolinic acid are related to the severity of depressive symptoms, which increases the risk of PPD (Achtyes et al., 2020). These results indicate that the increased level of some inflammatory biomarkers in PPD patients means that the disease is related to the impaired adaptability of the immune system (Bränn et al., 2020).

Therefore, the high expression of these genera in PPD patients may correspond to increased levels of inflammatory biomarkers, and several studies have shown a strong association between persistent inflammatory responses and antidepressant therapy resistance (Carvalho et al., 2013). Electroacupuncture can downregulate inflammatory factors such as IL-6 in the hippocampus of depressed rats, suggesting that electroacupuncture may relieve depression through immune regulation (Guo et al., 2014; Yue et al., 2018). Indeed, α7nAChR is activated by acetylcholine released from cholinergic nerve endings and is a key target for inhibiting pro-inflammatory cytokines release by macrophages (Stakenborg et al., 2017). Acupuncture can reduce inflammatory cytokine production through the vagus nerve by activating α7nAChR (Yang et al., 2021). Acupuncture can also regulate the interaction between the gut microbiota and the brain-gut axis, inhibit proinflammatory cytokine production, alter the number and proportion of the gut microbiota, restore its stability, improve intestinal barrier function, and further adjust body function (Jang et al., 2020; Wang et al., 2020). In acupuncture treatment of PPD, the unique regulation mechanism of immune intestinal flora also played an important role. Finally, we found that acupuncture might inhibit inflammation and improve depression via two pathways: (1) inhibiting the release of inflammatory cytokines by activating the vagus nerve; and (2) regulating the brain-gut axis through the intestinal microflora, some predecessors put forward the same argument previously. (Yang et al., 2022).Therefore, accumulation of Desulfovibrio and Paraperevotella in the intestinal tracts of responsive patients could mediate the immune response induced by acupuncture to better regulate and alleviate depressive symptoms. The conclusion of our research results accords with the above conclusion, which can be understood as that acupuncture has played a better and more sensitive role in the flora of the responders.

The study has several limitations. The sample size of present study was relatively small, and further studies in larger sample sizes are needed to confirm the findings with more advanced analyses methods, such as machine learning methods.

In conclusion, Paraprevotella and Desulfovibrio predicted early responses to antidepressants in patients with PPD receiving acupuncture. These results may help clinicians optimize their management of individual PPD patients in the future. Baseline enrichment and metabolism of Paraprevotella and Desulfovibrio intestinal microbiota in PPD patients were related to treatment outcomes. These findings pave the way for a new approach to personalize and maximize the efficacy of acupuncture treatment in PPD patients and provide potential new and accurate biomarkers for managing PPD patients.

The original contributions presented in the study are publicly available. This data can be found here: https://www.ncbi.nlm.nih.gov/, with accession number PRJNA976190.

The studies involving humans were approved by Ethics Committee of Shenzhen Traditional Chinese Medicine Hospital. The studies were conducted in accordance with the local legislation and institutional requirements. The participants provided their written informed consent to participate in this study.

Y-MZ and J-JY contributed equally. M-P and Z-XY are the corresponding authors. Y-MZ conceived and planned the experiments. Y-MZ and J-JY wrote the manuscript. Y-QX, X-MM, YYXZ, Y-HG, CC and X-XH executed the experiments. M-P and Z-XY contributed to revise the final manuscript. All authors contributed to the article and approved the submitted version.