Fecal microbiota as a predictor of acupuncture responses in patients with postpartum depressive disorder

Background There are several clinical and molecular predictors of responses to antidepressant therapy. However, these markers are either too subjective or complex for clinical use. The gut microbiota could provide an easily accessible set of biomarkers to predict therapeutic efficacy, but its value in predicting therapy responses to acupuncture in patients with depression is unknown. Here we analyzed the predictive value of the gut microbiota in patients with postpartum depressive disorder (PPD) treated with acupuncture. Methods Seventy-nine PPD patients were enrolled: 55 were treated with acupuncture and 24 did not received any treatment. The 17-item Hamilton depression rating scale (HAMD-17) was used to assess patients at baseline and after eight weeks. Patients receiving acupuncture treatment were divided into an acupuncture-responsive group or non-responsive group according to HAMD-17 scores changes. Baseline fecal samples were obtained from the patients receiving acupuncture and were analyzed by high-throughput 16S ribosomal RNA sequencing to characterize the gut microbiome. Results 47.27% patients responded to acupuncture treatment and 12.5% patients with no treatment recovered after 8-week follow-up. There was no significant difference in α-diversity between responders and non-responders. The β-diversity of non-responders was significantly higher than responders. Paraprevotella and Desulfovibrio spp. were significantly enriched in acupuncture responders, and these organisms had an area under the curve of 0.76 and 0.66 for predicting responder patients, respectively. Conclusions Paraprevotella and Desulfovibrioare may be useful predictive biomarkers to predict PPD patients likely to respond to acupuncture. Larger studies and validation in independent cohorts are now needed to validate our findings.


Introduction
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.

Study design
This was a prospective cohort study approved by the Ethics Committee of Shenzhen Hospital of Traditional Chinese Medicine [K2020-027-01].The study was registered with the Chinese Clinical Trial Registry (http://www.chictr.org.cn/index.aspx;ChiCTR210 0041687).Patients with PPD were recruited from the Shenzhen Traditional Chinese Medicine Hospital and Shenzhen Maternity & Child Healthcare Hospital (Shenzhen, China).All procedures used in this study conformed to the ethical standards of national and institutional human experimental committees and the Declaration of Helsinki.All subjects supplied written informed consent (Graphic Abstract).

Participant recruitment
Patients were initially screened for PPD using the Edinburgh Postnatal Depression Scale (EPDS) and then further evaluated using the 17-item Hamilton depression rating scale (HAMD-17) by physicians.All patients were assigned into acupuncture treatment group or no treatment group according to their own preference.

Diagnostic criteria
PPD was diagnosed by the evaluating physician according to the Fifth Edition of the Diagnosis and Statistics of Mental Illness (DSM-V) (Battle, 2013;First, 2013).Patients needed to meet five or more of the following symptoms, including at least the first or second symptoms, and the symptoms should have lasted for at least two weeks: (1) low mood and depressive emotion; (2) lack of interest in or loss of enjoyment in activities; (3) significant weight gain or loss; (4) poor sleep, insomnia, or lethargy; (5) psychomotor excitement or retardation; (6) a feeling of fatigue or weakness; (7) a sense that life is worthless, self-accusation, or self-guilt; (8) decline in cognition or difficulty concentrating; and (9) recurrent thoughts of death.

Inclusion and exclusion criteria
The inclusion criteria were: (1) patients between 20 and 49 years of age; (2) a diagnosis of PPD made by a psychiatrist; (3) illness appearing within a year of delivery; (4) HAMD-17 scores between 7 and 24; and (5) providing informed consent, voluntarily participating in the study, and able to complete the assessment instrument.
Exclusion criteria were: (1) severe psychiatric disorders such as bipolar affective disorder and schizophrenia; (2) mental disorder due to brain diseases or for other reasons, and unable to understand the contents of the questionnaire and cannot be effectively evaluated; (3) pregnancy; (4) patients with a HAMD suicide score >2 points; (5) anyone attempting suicide in the past year; and (6) anyone taking antibiotics or probiotics in the past month.

Interventions
Patients in the acupuncture group were treated with acupuncture therapy by an acupuncturist with a doctor's license and at least three years of clinical experience.Before patient enrollment, all acupuncturists participated in standardized operating procedure training, including locating the acupoints and needle manipulation.
Patients in the no treatment group didn't receive any therapy.

Clinical outcomes
The clinical outcome was the response rate.The HAMD is a commonly used scale for clinical evaluation of depressive state (Hamilton, 1960).Depressive symptoms of PPD patients were assessed by HAMD-17 scale (17 items, scored from 0 to 52, higher scores representing more severe the depressive symptoms).
Patients were defined as responders if the HAMD-17 score reduced by ≥50% or the HAMD-17 score was <7 after treatment.Patients were defined as non-responders if the reduction in HAMD-17 score was <50% (Keller, 2003).

Fecal samples collection
Fecal samples of PPD participants in acupuncture treatment group were collected once at baseline and placed in sterile plastic cups, then frozen at −80°C immediately after defecation.The details of fecal sample collection are described elsewhere (Zhou et al., 2019).

DNA extraction and 16S ribosomal RNA gene sequencing
DNA was extracted using the MOBIO PowerSoil ® DNA Separation Kit according to the manufacturer's instructions, and stored at −80°in Tris-EDTA buffer solution before microbial MiSeq sequencing.The V4 region of 16S rRNA gene was amplified by PCR with primers 515F (5'-GTGYCAGCMGCCGCGGTAA-3') and 806R (5'-GGACTACNVGGGTWTCTAAT-3'), along with barcode sequences, as previously described (Zhou et al., 2020).PCR mixtures contained 1 ml of each forward and reverse primer (10mM), 1 ml of template DNA, 4 ml of dNTPs (2.5mM), 5 ml of 10× EasyPfu Buffer, 1 ml of Easy Pfu DNA Polymerase (2.5 U/ml), and 1 ml of double-distilled water in a 50-ml reaction volume.Thermal cycling consisted of an initial denaturation step at 95°for 5min, followed by 30 cycles of denaturation at 94°for 30 s, annealing at 60°for 30 s, and extension at 72°for 40 s, with a final extension step at 72°for 4min.Amplicons were run for each sample on an agarose gel.Expected band size for 515f-806r was ∼300-350 bp.Amplicons were quantified with Quant-iT PicoGreen dsDNA Assay Kit (P11496; Thermo Fisher Scientific, Waltham, MA, USA) according to manufacturer's instructions.The amplicon library for high-throughput sequencing on the Illumina MiSeq V3 reagent PE150 (300 cycles) platform was combined to an equal amount and subsequently quantified using KAPA Library Quantification Kit (KK4824; Illumina, Inc., San Diego, CA, USA) according to manufacturer's protocols.

High-throughput sequencing of 16S ribosomal RNA gene and microbial analysis
High-throughput sequencing analysis was performed using Quantitative Insights into Microbial Ecology (QIIME) 2.0 according to the manufacturer's instructions.Raw Illumina read data were deposited into tags, reads belonging to each sample were separated with barcodes, and low-quality reads were removed.The processed tags were clustered into amplicon sequence variants (ASVs) using the commonly used 97% similarity threshold.ASVs were assigned to taxa by matching to the SILVA database.A phylogenetic tree of representative sequences was constructed.a-   Secondly, the Bray-Curtis dissimilarity and unweighted unifrac calculated by principal coordinate analyses were used for bdiversity indices.To further identify specific bacteria as biomarkers at the genus level, linear discriminant analysis effect size (LEfSe) was applied through the Huttenhower Lab Galaxy Server (Segata et al., 2011) after taxa summaries were reformatted.LEfSe settings were as previously described (Zhou et al., 2020), and systemic forms with a linear discriminant analysis (LDA) cutoff of 2.0 and a P < 0.05 in the built-in rank sum test were considered statistically significant.Finally, biomarker data (specific bacteria) calculated by LEfSe were further analyzed by receiver operator characteristic (ROC) curve analysis, and area under the curve (AUC) was used to assess the ROC effect.The cut-off value associated with optimal sensitivity and specificity was used to distinguish acupuncture responders and non-responders.

Analysis of clinical data
The demographic and clinical outcomes were analyzed using the SPSS 22.0 software (IBM Statistics, Armonk, NY, USA).Normally distributed data were analyzed using Student's t-test, while nonparametric data were analyzed using the Mann-Whitney U-test with data expressed as medians with interquartile ranges (IQR).Categorical data were compared using the chi-squared test.A Pvalue < 0.05 was considered statistically significant.

Results
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).

Clinical characteristics of acupuncture therapy and control patients
The baseline demographic and clinical characteristics are shown in Table 2.There were no significant differences in age, body mass index (BMI), number of days postpartum, number of parturitions, duration of disease, delivery mode, family history, and EPDS or HAMD-17 scores between those receiving acupuncture and those not receiving acupuncture at baseline.47.27% responded to acupuncture treatment in the acupuncture group, and 12.5% patients not receiving treatment recovered after 8week follow up.This difference was significant (P = 0.003) (Table 3).

Comparison of HAMD reduction rate between groups
Compared with the HAMD reduction rate between two groups, the results showed reduction rate in acupuncture group was superior than that in control group.This difference was significant (P <0.001) (Table 4).

HAMD changes between patients receiving acupuncture and controls
Compared with baseline, HAMD scores in the control group did not significantly decrease (P = 0.113).However, the HAMD score decreased significantly in patients receiving acupuncture treatment (P <0.001) (Table 5).
3.2 Characteristics of the gut microbiota between responders and non-responders before treatment

Baseline clinical characteristics between responders and non-responders
There were no significant differences in age, BMI, number of days postpartum, number of parturitions, length of disease, delivery mode, family history, and EPDS or HAMD-17 scores between responders and non-responders (Table 6).

Sequencing characteristics
A total of 55 samples from all recruited subjects were sequenced on an Illumina MiSeq sequencer.For downstream analysis, 2259092 qualified reads from 2373462 raw reads were filtered.

Gut microbial diversity changes in acupuncture responders and non-responders
We next used different diversity indices (evenness, Faith PD, observed species, Shannon diversity) to assess gut microbial a-diversity.There were no significant differences in diversity between acupuncture responders and non-responders (P = 0.7856, P = 0.4276, P = 0.6679, and P = 0.7208, respectively).However, the gut microbial diversity, as estimated by evenness, Faith PD, observed species, and Shannon diversity, tended to be higher in responders than non-responders (Figure 2).
To better understand differences in overall community composition between the samples, we calculated Bray-Curtis distances and unweighted UniFrac distances, which were both higher in non-responders than responders (P = 0.0065 and P = 5.5e -05 ) and between groups (P = 0.019 and P = 0.0005) (Figures 3A, C), To further demonstrate differences in species diversity between samples, we applied the Principal Coordination Analysis and non-metric multidimensional scaling (Figures 3B, D).The gut microbial composition was similar between groups, with a    tendency to being more centralized in the responder group than in the non-responder group, although this was not statistically significant (P = 0.959 and P = 0.911).

Specific genera associated with acupuncture treatment responses
LEfSe analysis (p <0.05, LDA > 2) was used to identify specific bacteria associated with acupuncture treatment responses.g_Desulfovibrio, g_Paraprevotella, and Paraprevotella_xylaniphila were enriched in the responder group (Figure 5).
Having identified these three genera (biomarkers), we performed ROC curve analysis to evaluate their predictive accuracy.The area under the curve (AUC) was 0.76 and 0.66 for g_Paraprevotella and g_Desulfovibrio, respectively.The AUC of combining genera g_Paraprevotella and g_Desulfovibrio was 0.65 (Figure 6).

Discussion
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   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 aand bdiversity.a-diversity is commonly used as a surrogate of

A B FIGURE 4
Microbiome composition differences at the phylum (A) and genus (B) levels between the two groups.PPD_R, PPD patients who are responsive to acupuncture treatment; PPD_NR, PPD patients who are not responsive to acupuncture treatment.
community stability and function, which are thought to be beneficial to the host (Shade, 2017).Jiang et al. found that the intestinal microflora a-diversity was higher in antidepressant drug non-responders than responders in MDD patients compared with healthy controls (Jiang et al., 2015).In addition, the a-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.b-diversity reflects relationships between samples by analyzing the species composition and abundance (Anderson et al., 2011).Our b-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 b-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 bdiversity was also different between non-responders and responders (Duan et al., 2021), as was the b-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 immunotherapyrelated 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  ROC curves using fecal microbiota to distinguish responders from non-responders.AUC, area under the curve; ROC, receiver operating characteristics.
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 wellestablished 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 cocultured 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-1b 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, a7nAChR 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 a7nAChR (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.

Limitations
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.

Conclusion
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 middle of the posterior hairline, on the middle of the top of the head Yintang (EX-HN3) On the forehead, between the brows Zhongwan (RN12) On the anterior median line, 4 cun superior to the umbilicus Qihai (RN6) On the anterior median line, 1.5 cun caudal to the umbilicus Guanyuan (RN4) On the anterior median line, 3 cun caudal to the umbilicus Neiguan (PC6) 2 cun proximal to the processus styloideus radii, between the tendons of the palmaris longus and the flexor carpi radialis Shenmen (HT7) In the wrist, the ulnar end of the transverse striation of the carpal palmar, and the radial depression of the flexor tendon of the ulnar carpal Hegu (LI4) On the highest point at m. interosseus dorsalis Sanyinjiao (SP6) 3 cun proximal to the medial malleolus Taichong (LR3) Between metatarsal I and II, just distal to the caput FIGURE 1Trial flow diagram.

2
FIGURE 2Microbial a-diversity analyses.Evenness index (A), Faith PD (B), observed species (C), Shannon diversity (D).PPD_R, PPD patients who are responsive to acupuncture treatment; PPD_NR, PPD patients who are not responsive to acupuncture treatment.
FIGURE 3 Microbial b-diversity analyses.The Wilcoxon rank-sum test analysis (A) and principal coordinates analysis plots (B) of the fecal microbiomes based on the Bray-Curtis distance; Wilcoxon rank-sum test analysis (C) and principal coordinates analysis plots (D) based on the Unweighted-UniFrac distance metric.PPD_R, PPD patients who are responsive to acupuncture treatment; PPD_NR, PPD patients who are not responsive to acupuncture treatment.
FIGURE 5 Differences in bacterial taxa between acupuncture responders and non-responders.Cladogram showing the most differentially abundant taxa identified by LEfSe.Red indicates clades enriched in the responder group (A).Comparisons of gut microbiota between acupuncture responders and non-responders (B).Only genera meeting a linear discriminant analysis score threshold >2 are shown.PPD_R, PPD patients who are responsive to acupuncture treatment; PPD_NR, PPD patients who are not responsive to acupuncture treatment.

TABLE 1
Localization of acupoints selected in this trial.
PD indices were calculated by Wilcoxon rank sum test.For b-diversity indices, firstly, Wilcoxon rank sum test was used to analyze the inter-group and intra-group differences.The former indicated differences in microbial composition between samples within the same group; the later indicate the differences in microbial composition of pairwise samples from different groups.

TABLE 2
Clinical characteristics of patients in the acupuncture therapy and the control groups.
*A, Natural childbirth; B, Cesarean section; C, Women miscarried but pregnant for more than seven months.

TABLE 4
Comparison of HAMD reduction rate between acupuncture group and control group.

TABLE 5
Comparison of HAMD score before and after treatment in two groups.

TABLE 6
Clinical characteristics of acupuncture responders and non-responders.
*A, Natural childbirth; B, Cesarean section; C, The Women miscarried who are but were pregnant for more than seven months.