Translational Psychiatry www.nature.com/tp ARTICLE OPEN Gut microbiota, innate immune pathways, and inflammatory control mechanisms in patients with major depressive disorder Javier R. Caso 1,2,11, Karina S. MacDowell1,2,11, Ana González-Pinto3, Saínza García3, Javier de Diego-Adeliño4, Mar Carceller-Sindreu4, Fernando Sarramea5, Javier Caballero-Villarraso 5, Patricia Gracia-García 6, Concepción De la Cámara7, Luis Agüera8, ✉ María L. Gómez-Lus9, Claudio Alba 10, Juan M. Rodríguez10 and Juan C. Leza 1,2 © The Author(s) 2021 Although alterations in the gut microbiota have been linked to the pathophysiology of major depressive disorder (MDD), including through effects on the immune response, our understanding is deficient about the straight connection patterns among microbiota and MDD in patients. Male and female MDD patients were recruited: 46 patients with a current active MDD (a-MDD) and 22 in remission or with only mild symptoms (r-MDD). Forty-five healthy controls (HC) were also recruited. Psychopathological states were assessed, and fecal and blood samples were collected. Results indicated that the inducible nitric oxide synthase expression was higher in MDD patients compared with HC and the oxidative stress levels were greater in the a-MDD group. Furthermore, the lipopolysaccharide (an indirect marker of bacterial translocation) was higher in a-MDD patients compared with the other groups. 1234567890();,: Fecal samples did not cluster according to the presence or the absence of MDD. There were bacterial genera whose relative abundance was altered in MDD: Bilophila (2-fold) and Alistipes (1.5-fold) were higher, while Anaerostipes (1.5-fold) and Dialister (15- fold) were lower in MDD patients compared with HC. Patients with a-MDD presented higher relative abundance of Alistipes and Anaerostipes (1.5-fold) and a complete depletion of Dialister compared with HC. Patients with r-MDD presented higher abundance of Bilophila (2.5-fold) compared with HC. Thus, the abundance of bacterial genera and some immune pathways, both with potential implications in the pathophysiology of depression, appear to be altered in MDD, with the most noticeable changes occurring in patients with the worse clinical condition, the a-MDD group. Translational Psychiatry (2021)11:645 ; https://doi.org/10.1038/s41398-021-01755-3 INTRODUCTION high-mobility group box 1 (HMGB1)) [7, 9], more studies are Inflammatory processes can be implicated in the development of required to verify the participation of DAMPs in prompting chronic depressive-like symptoms or major depressive disorder (MDD) [1– low-grade inflammation in these pathologies. Consequently, the 3]. It appears that MDD and inflammation are fueling each other: pattern recognition receptors (PRRs) are getting attention. In the inflammation promoting depression and MDD facilitating inflam- central nervous system (CNS) PRRs are primarily expressed by matory reactions [4]. MDD patients usually exhibit alteration of microglia, macrophages and astrocytes and they are either biomarkers of immune dysfunction and data point toward the membrane-bound (Toll-like receptors (TLRs)) or found within the manifestation of all components of an archetypical inflammatory cytoplasm (nucleotide-binding oligomerization domain-like recep- response in MDD [4]. Besides, inflammation is linked with the tors) sensing intracellular signals [10, 11]. clinical severity of mood disorders and remission of MDD is Ligands of PRRs include those related to damage (DAMPs) and connected with the normalization of inflammatory markers [5, 6]. pathogens (pathogen-associated molecular patterns (PAMPs)) and Damage-associated molecular patterns (DAMPs) are endogen- those related to microbes (microbe-associated molecular patterns ous molecules whose production is increased after stressor (MAMPs)) comprising both pathogenic and commensal/symbiotic exposure and tissue damage [7]. In the nonexistence of tissue microbes [7]. The TLRs potential role in the pathogenesis of damage, DAMPs induce the so-called systemic sterile inflamma- psychiatric diseases is being intensely studied [9]. The most tion [8]. Despite some evidence of increased levels of peripheral studied is TLR-4, which responds not only to lipopolysaccharide DAMPs in neuropsychiatric diseases (e.g., heat shock proteins and (LPS), a component of the outer membrane of Gram-negative 1 Department of Pharmacology & Toxicology, Faculty of Medicine, Universidad Complutense de Madrid (UCM), Madrid, Spain. 2Centro de Investigación Biomédica en Red de Salud Mental (CIBERSAM), Instituto de Investigación Hospital 12 de Octubre (Imas12), IUIN-UCM, Madrid, Spain. 3University Hospital of Alava (HUA), University of the Basque Country (UPV/EHU), CIBERSAM, Vitoria-Gasteiz, Spain. 4Department of Psychiatry, Hospital de la Santa Creu i Sant Pau, Institut d’Investigació Biomèdica Sant Pau (IIB-Sant Pau), Universitat Autònoma de Barcelona (UAB), CIBERSAM, Barcelona, Spain. 5Instituto Maimónides de Investigación Biomédica de Córdoba (IMIBIC), Unidad de Gestión Clínica de Salud Mental, Hospital Universitario Reina Sofía, CIBERSAM, Córdoba, Spain. 6Servicio de Psiquiatría, Hospital Universitario Miguel Servet, Departamento de Medicina y Psiquiatría, Universidad de Zaragoza, CIBERSAM, Zaragoza, Spain. 7Servicio de Psiquiatría, Hospital Clínico Universitario, Departamento de Medicina y Psiquiatría, Universidad de Zaragoza, CIBERSAM, Zaragoza, Spain. 8Psychiatry Department, Hospital Universitario 12 de Octubre, UCM, CIBERSAM, Madrid, Spain. 9Department of Medicine—Microbiology Area, Faculty of Medicine, UCM, Madrid, Spain. 10Department of Nutrition and Food Science, Faculty of Veterinary, UCM, Madrid, Spain. 11These authors contributed equally: Javier R. Caso, Karina S. MacDowell. ✉email: jcleza@med.ucm.es Received: 22 June 2021 Revised: 18 November 2021 Accepted: 30 November 2021 J.R. Caso et al. 2 bacteria, but also to DAMPs, resulting in the activation of approved by the Hospitals’ Review Boards and the Complutense inflammatory transcription factors [12, 13]. Ethics Committee (project number PI13/01102). Procedures Preclinical studies indicate a substantial similarity in numerous comply with Spain’s legislation (Ley 14/2007) and the Helsinki inflammatory parameters that are modulated after either LPS or Declaration. All participants signed a written informed consent stressor exposure, which are detectable in the blood and after receiving a complete description of the study. Inclusion/ peripheral tissues [7, 9]. Patterns from gut bacteria can be exclusion criteria are described in Supplementary Material. considered MAMPs or PAMPs, depending on the setting, and they The study included 68 MDD patients and 45 HC. The MDD can induce an immune activation in the CNS [14, 15]. Conse- sample was divided into two groups for additional analyses: one quently, it is worth to investigate the relationship between the composed by patients with a current active depressive episode microbiota–gut–brain (MGB) axis and the immune response according to DSM-IV-TR and a Hamilton Depression Rating Scale ongoing in MDD, especially considering the lack of understanding (HDRS) > 14 (a-MDD) and a separate group of participants about the connection among microbiota, immune response and composed by patients who had responded to treatment and MDD [16, 17]. remained in remission or with only mild symptoms (HDRS 8–14) (r- Microbiota impacts numerous aspects of physiology [18], and MDD). data from experimental animal models and preclinical studies reveal that depression could influence microbiota composition Clinical assessments and suggest that gut microbiota might impact brain function and An ad hoc protocol was used for demographic and clinical behavior through the MGB axis [19, 20]. Thus, gut microbiota characteristics. Clinical and psychopathological states were alterations could be a contributory factor to the development of assessed by means of the Spanish versions of the HDRS [31], the MDD [21]. These alterations would increase the intestinal barrier Euroquol-5D visual analog scale (EQ-Vas, a measure of self- permeability allowing a bacterial translocation [22], which could perceived health-related quality of life) [32], the perceived stress be related to the inflammatory hypothesis of depression [23]. scale (PSS) [33], list of threatening experiences questionnaire (LTE- Animal studies show that experimental models of depression Q) [34], and the childhood trauma questionnaire-short form (CTQ- affect the intestinal barrier allowing the bacteria translocation and SF) [35]. Rome III criteria for functional gastrointestinal pathology inducing a neuroinflammatory response through TLR-4 activation [36] in the last 3 months were evaluated in all the groups by in the brain [24, 25]. This scenario seems to happen in MDD means of a questionnaire (ten yes/no items) while the psycholo- patients as well, suggesting that bacterial translocation would be gical assessments were performed. related to the inflammatory pathophysiology of the MDD [26, 27]. Data were analyzed using the D’Agostino and Pearson test to Studies examining gut microbiota in MDD and healthy controls assess Gaussian distribution. An unpaired two-tailed t-test was (HC) generally show decreased microbial richness and diversity performed when there were no values in the HC group and a one- [28]. A research has studied active and remitted MDD compared to way ANOVA with a Tukey post-hoc test was employed for controls showing differences in diversity, and in the levels of comparisons between the HC, a-MDD, and r-MDD groups. When specific bacterial taxa, particularly among patients with clinically data did not follow a Gaussian distribution, a nonparametric significant depressive symptoms [29]. However, the association ANOVA with a Kruskal–Wallis and a Dunn’s post-hoc test was between gut microbiota and MDD, and between microbiota and performed. inflammatory markers/processes persists weakly comprehended. An assessment of the fecal microbiota of patients with active MDD Metataxonomic analysis of the fecal samples (a-MDD) and remained in remission or with only mild symptoms Sample collection and DNA extraction. All metataxonomic analysis (r-MDD) compared with HC and the possible actions on the were performed by investigators totally blinded to the groups immune response should be undertaken before drawing definite allocation. Fecal samples were collected in a sterile plastic cup conclusions; these are the main aim and justification of our study, after the participants completed the clinical assessments and were in which we have shown variations of bacteria genera and the kept in an icebox. Samples for bacterial genomic DNA extraction activation of proinflammatory pathways, that can be indirectly were delivered to the laboratory and stored at −80 °C, thus, related to bacterial translocation, with the most noticeable samples were kept on ice for not longer than 1 h. Microbial DNA changes occurring in a-MDD patients. was extracted from fecal aliquots (200 mg) using the QIAamp© Employing samples from MDD patients (including a-MDD and DNA Stool Mini Kit (Qiagen, Hilden, Germany) according to the r-MDD patients) and HC, the aims of this study were: (a) manufacturer’s instructions, with the additional glass-bead beat- identifying whether human fecal microbiota is altered during ing steps on a Mini-beadbeater (FastPrep; Thermo Electron Corp., MDD compared with HC; (b) detecting microbiota signatures Boston, MA, USA). DNA was quantified using a NanoDrop ND-1000 distinctive for MDD and their connections with the possible spectrophotometer (Thermo Electron); integrity and size were inflammatory scenarios present in patients compared with HC; (c) assessed by 1.0% agarose gel electrophoresis on gels containing recognizing TLR-4 and proinflammatory pathways activation in 0.5 mg/mL ethidium bromide. DNA was stored at 20 °C before blood from MDD patients compared with HC; and (d) finding analysis and all samples were processed and sequenced in the whether the possible scenarios detected by the previous points same batch to avoid impacts in the microbial results. are also happening among a-MDD and r-MDD patients as well as compare them with HC. PCR amplification and sequencing. 16S rDNA gene amplicons were amplified following the Illumina protocol for 16S rDNA gene metagenomic sequencing library preparation (part# 15044223 MATERIALS/SUBJECTS AND METHODS Rev. A). The forward primer (TCGTCGGCAGCGTCAGATGTGTATAA- Participants GAGACAGCCTACGGGNGGCWGCAG) and the reverse one (TACGG- Outpatients from multiple hospitals in different cities, with a TAGCAGAGACTTGGTCTGACTACHVGGGTATCTAATCC) were used documented Diagnostic and Statistical Manual of Mental Dis- as previously described [37], generating amplicons targeting the orders, 4th ed., Text Revision (DSM-IV-TR) [30] diagnosis of MDD, V3-V4 hypervariable region of the 16S rDNA gene. An aliquot of through structured clinical interviews performed by psychiatrists, the microbial DNA from each sample (5 ng/μL in 10 mM Tris pH participated in the study. An HC group with a similar age and sex 8.5) was used to initiate the protocol. Libraries were sequenced distribution to the patients was recruited, and clinically evaluated using a 2 × 300 bp paired-end run (MiSeq Reagent kit v3, MS-102- in the same manner as patients, among their relatives and 3001) on a MiSeq Sequencer according to manufacturer’s graduate students and/or hospital staff. Procedures were instructions (Illumina, USA). Translational Psychiatry (2021)11:645 J.R. Caso et al. 3 Quality assessment and taxonomy. Quality assessment was using the D’Agostino and Pearson test to assess Gaussian performed using prinseq-lite program [38]. R1 and R2 from distribution. First, the whole group of MDD patients was Illumina sequencing were joined using fastq-join from ea-tools compared with HC: an unpaired two-tailed t-test was performed. suite [39]. The amplified fragments were clustered in operational Second, a one-way ANOVA with a Tukey post-hoc test was taxonomic units (OTUs) and representative sequences were employed for comparisons between the HC, a-MDD, and r-MDD taxonomically analyzed using RDP_classifier from the Ribosomal groups. When the data did not follow a Gaussian distribution, a Database Project [40]. Raw datasets are available in the E-Prints nonparametric ANOVA with a Kruskal–Wallis test followed by a Complutense repository, ID Code: 64666. Dunn’s post-hoc test was performed. The variance was similar among the groups. A p value < 0.05 was considered statistically Bioinformatic analysis. The bioinformatic analysis was conducted significant in all cases. ANOVAs’ statistical details can be observed combining R (v 3.2.3), QIIME pipelines (v 1.8.0) [41], and Calypso (v in Supplementary Table S2. 8.84) [42]. Estimates of intrasample diversity were made at a rarefaction depth of 27,000 reads per sample. Alpha diversity was assessed with the Shannon diversity index, which considers the RESULTS number and evenness of microbial species. Differences between Demographic and clinical characteristics of the sample groups were assessed using either Kruskal–Wallis tests for three Patients were white Caucasian and most of the participants were groups comparison (control, a-MDD, and r-MDD) or Wilcoxon rank married and had secondary or primary education. The mean age sum tests for pairwise and two groups (control and MDD) was 43.98 years and 77.77% females, and 46 patients were a-MDD comparisons. Distance matrices containing a dissimilarity value (mean age 42.1 years, 78.26% females) and 22 were r-MDD (mean for each pairwise sample comparison were performed to evaluate age 45.85 years, 77.27% females). These groups were compared the beta diversity and to plot patterns of bacterial community with 45 HC (mean age 44.72 years, 75.5% females). There were no diversity. For the quantitative (relative abundance) and qualitative differences among the groups in age nor in gender. The Rome III (presence/absence) analyses, the Bray–Curtis dissimilarity and criteria indicated that none of the participants in the study binary Jaccard distance indices were used, respectively. Principal presented functional gastrointestinal disorders. All participants coordinates analysis (PCoA) was used to visually display patterns presented a body mass index under 30. of beta diversity through the distance matrices. The PERMANOVA The average number of months since the onset of the disease analysis with 999 permutations was performed to reveal was 125.73 in a-MDD and 160.91 in r-MDD patients. statistically significant differences (p < 0.05). Differences in genera The a-MDD group showed higher HDRS, PSS, and LTE-Q scores were compared by using either Kruskal–Wallis tests for variables than the r-MDD group. The r-MDD group showed higher levels in with three groups comparison (control, a-MDD, and r-MDD) or the EQ-Vas scale and, consequently, a higher health-related Wilcoxon rank sum tests for pairwise and control vs. MDD quality of life than patients with a-MDD (66.58 vs. 38.78, comparisons. respectively) although their values were still far from the ones in To correct for multiple comparisons, Bonferroni-adjusted the HC group (89.82). Both MDD groups had higher scores in the significance levels were set for each analysis. The bar chart with childhood trauma scales when compared with HC (Table 1). the cladogram was performed with the Hclust hierarchical cluster analysis with complete linkage method from the Calypso online Metataxonomic analysis of fecal samples software. Samples were submitted to metataxonomic profiling (n = 113). Globally, 451 representative OTUs were retrieved from 6,436,207 Biochemical determinations in plasma and peripheral blood high-quality-filtered sequences. The phyla Firmicutes, Bacteroi- mononuclear cells (PBMCs) detes, Proteobacteria, Actinobacteria, and Verrucomicrobia Specimen collection and preparation. Venous blood samples accounted for 99.36% of the sequencing data (Supplementary (10 mL) were collected in the morning, between 8 a.m. and 10 Fig. S1). a.m., after overnight fasting. Blood tubes were centrifuged (641 × Bacterial diversity (Shannon index) was not different among the g for 10 min at 4 °C). Plasma samples were collected and stored at different groups (Supplementary Fig. S2) of patients (HC, a-MDD, −80 °C. The rest of the sample was 1:2 diluted in culture medium and r-MDD) (p value > 0.05) and no statistical differences were (RPMI 1640, LifeTech) and a gradient with Ficoll-Paque (GE found in relation to the 10 most abundant genera (Supplementary Healthcare) was used to isolate mononuclear cells by centrifuga- Table S3). tion (800 × g for 40 min at room temperature [RT]). The PBMC layer There was no clear separation between the subjects according was aspired, suspended in RPMI and centrifuged (1116 × g for to the relative abundance of the 20 most abundant genera (TSS; 10 min at RT). The supernatant was removed, and the mono- total-sum normalization and the Hclust with complete linkage nuclear cell-enriched pellet was stored at −80 °C. method) (Fig. 1A). At the OTUs level, the PCoA plots of the Bray–Curtis dissimilarity matrix revealed that the samples did not Determinations in plasma. Biochemical parameters in plasma cluster according to their diagnosis (Fig. 1B), while the analysis of were measured using commercially available kits and following similarity (PERMANOVA) showed that there was no statistical the manufacturers’ instructions (details in Supplementary Material). difference between the three groups of subjects (p = 0.089). The binary Jaccard distance matrix (presence/absence) indicated that Determinations in PBMCs. PBMC samples were fractionated in samples did not cluster according to the diagnosis (p = 0.131) cytosolic and nuclear extracts using a method which provides a (Fig. 1C). high purity nuclear fraction [43]. Protein expression analyses were The abundance of sequences from the genera Bilophila and performed in cytosolic extracts, except for the two transcription Alistipes was higher in MDD patients (both MDD groups factors studied (i.e., NF-κBp65 and PPARγ), whose expression levels combined) compared with HC (Fig. 1D, E). Oppositely, the were analyzed in nuclear extracts (details in Supplementary abundance of sequences from the genera Anaerostipes and Material). Antibodies’ identifiers provided by the Resource Dialister was lower in MDD patients (Fig. 1F, G). Identification Portal can be seen in Supplementary Table S1. When the subgroups were analyzed, a-MDD patients showed higher abundance of sequences belonging to the genus Alistipes Statistical analysis of the biochemical determinations (Fig. 1I) while lower levels of genera Anaerostipes and Dialister Data are expressed as mean ± standard error of the mean. The (Fig. 1J, K). Patients with r-MDD presented higher presence of the ROUT method was used to identify outliers. Data were analyzed genus Bilophila (Fig. 1H). Translational Psychiatry (2021)11:645 J.R. Caso et al. 4 B). However, the plasma levels of the HMGB1 protein were lower Table 1. Sociodemographic and clinical characteristics of the subjects. in the MDD group (Fig. 2C). Healthy Acute Remission MDD controls (HC), MDD (a- (r-MDD), n = 22 Comparison between a-MDD, r-MDD, and HC. The a-MDD group n = 45 MDD), n = 46 presented not altered expression of TLR-4 (Fig. 2D), higher LPS Sex (women), n (%) 34 (75.5%) 36 17 (77.27%) plasma levels (Fig. 2E), and no changes in the HMGB1 plasma (78.26%) levels (Fig. 2F) compared with HC. The r-MDD group displayed Age (years) 44.72 42.10 45.85 higher TLR-4 expression (Fig. 2D), no changes in LPS plasma levels Employment status (Fig. 2E), and lower HMGB1 plasma levels compared with HC Full-time job 34 20 9 (Fig. 2F). Part-time job 0 0 0 Effects of MDD on intra- and intercellular inflammatory Student 3 1 1 parameters Unemployed 5 8 9 MDD patients vs. HC. The ratio between the activated (phos- Retired 1 1 2 phorylated) ERK and the ERK total form was decreased in PBMCs On sick leave 2 16 1 from MDD patients compared with HC (Fig. 3A). N/A 0 0 0 Patients with MDD did not show differences in the protein Educational level levels of the nuclear factor kappaB (NFκB) nor in the plasma levels Primary education 16 14 14 of C-reactive protein (CRP) and interleukin (IL)-6 compared with Secondary education 10 18 3 HC (Fig. 3B–D). University studies 18 13 4 Comparison between a-MDD, r-MDD, and HC. Patients with N/A 1 2 1 a-MDD did not show changes when compared with HC in any Relationship status of the intra- and intercellular inflammatory parameters analyzed Single 12 10 6 (Fig. 3E–H). Patients with r-MDD presented lower levels of the Married (or solid 24 27 13 pERK/ERK ratio in PBMCs (Fig. 3E) and higher plasma levels of IL-6 relationship) (Fig. 3H) compared with HC. The r-MDD group presented lower Divorced 8 5 2 plasma CRP levels compared with the a-MDD group (Fig. 3G). N/A 1 4 1 Onset of the disease n.a. 125.73 160.91### Effects of the MDD on the cyclooxygenase-2 pathway, on the (months) oxidative/nitrosative response, and on counterbalancing Previous treatments mechanisms On antidepressant, n (%) 0 38 (82.6) 20 (90.9) MDD patients vs. HC. MDD patients presented lower levels of On others, n (%) 0 17 (36.9) 2 (9.09) COX-2 expression in PBMCs (Fig. 4A) compared with HC, without HDRS 0.08 21.17*** 10.54**,### changes in the proinflammatory PGE2 and the anti-inflammatory EQ-Vas 89.82 38.78*** 66.58***,### 15d-PGJ2 plasma levels (Fig. 4B, C). The anti-inflammatory nuclear PSS 19.47 33.82*** 27.35***,### factor PPARγ expression was not affected between the different LTE-Q 0.81 1.26*** 1.02* groups (Fig. 4D). Moreover, MDD patients displayed higher Childhood trauma isoform of the nitric oxide synthase (iNOS) protein levels in CTQ-SF total scores 28.74 40.19*** 46.22***,# PBMCs (Fig. 4I) but no changes when analyzing TBARS, SOD and GPx enzymatic activities (Fig. 4J–L). CTQ-EA (emotional abuse) 5.27 8.9*** 8.81***,## CTQ-PA (physical abuse) 5.22 6.58*** 6.26*** Comparison between a-MDD, r-MDD, and HC. The COX-2 pathway CTQ-SA (sexual abuse) 5.04 6.77*** 6.4***,## was not altered in the a-MDD and r-MDD groups not showing any CTQ-EN (emotional 6.79 10.38*** 10.63***,## difference between them nor with the HC group (Fig. 4E–H). neglect) Both a-MDD and r-MDD groups presented higher iNOS CTQ-PN (Physical neglect) 5.35 7.28*** 7.53***,## expression in PBMCs (Fig. 4M). The a-MDD group shown higher Functional gastrointestinal Negative Negative Negative TBARS levels compared with HC (Fig. 4N). There were no pathology (Rome III criteria) differences in the activity of SOD and GPx among groups Smoking, cig/day (n users) 14 (11) 9.8 (19) 6.36 (5) (Fig. 4O, P). Alcohol, SDU/week (n users) 9.7 (22) 2.4 (11) 1.75 (4) Other (n users) 1 2 0 An unpaired two-tailed t-test was performed when comparing two groups. DISCUSSION For more than two groups comparisons, a one-way ANOVA with a Tukey post- This study presents evidence about differences in some bacterial hoc test was employed, and in those cases in which the data did not follow a genera and changes in the innate immune response that could be Gaussian distribution, a nonparametric ANOVA with a Kruskal–Wallis test associated to MDD diagnosis. Patients in an active depressive followed by a Dunn’s post-hoc test was performed. Bold values indicates episode, with higher depressive scores and lower health-related statistical significant p values. *p < 0.05, **p < 0.01, ***p < 0.001 vs. HC; #p < 0.05, ##p < 0.01, ###p < 0.001 vs. a-MDD. quality of life, showed some different profile in gut microbiota that HDRS Hamilton Depression Rating Scale, EQ-Vas Euroquol-5D visual analog scale, could be linked to the changes in the inflammatory markers also PSS perceived stress scale, LTE-Q list of threatening experiences questionnaire, detected in this group compared to patients with a depressive CTQ-SF childhood trauma questionnaire-short form, SDU standard drink unit. episode in remission or recovered from a depressive episode. Disentangling these patterns of variations and their relationships with different phases of the illness could be particularly valuable to understand further neurobiological mechanisms involved in MDD. Effects of the MDD on the innate immune pathway Fecal samples did not cluster according to the presence or the MDD patients vs. HC. The TLR-4 protein expression in PBMCs and absence of MDD. In addition, bacterial diversity was not different LPS plasma levels did not show differences among groups (Fig. 2A, among the groups. A recent systematic review has revealed the Translational Psychiatry (2021)11:645 J.R. Caso et al. 5 Fig. 1 Metataxonomic analysis and abundance of bacteria genera in fecal samples. A Comparison of the relative abundance of the 20 most abundant genera (TSS; total-sum normalization) in the three groups of subjects. The dendrogram was based on genera similarity between the samples. Hclust with complete linkage method from the Calypso online software was used to compute the hierarchical clustering. PCoA plots of bacterial profiles based on Bray–Curtis similarity analysis (relative abundance) (B) and on the Jaccard’s coefficient for binary data (presence/ absence) (C) from the three groups of subjects (blue triangle, control group; red circles, a-MDD group; orange squares, r-MDD group). The value given on each axis label represents the percentage of the total variance explained by that axis. Relative abundance of the genera Bilophila (D), Alistipes (E), Anaerostipes (F), and Dialister (G) in MDD patients compared with HC. Patients with r-MDD presented an increased presence of sequences belonging to the genus Bilophila (H). Patients with a-MDD showed increased abundance of sequences belonging to the genus Alistipes (I), while those corresponding to the genera Anaerostipes and Dialister were decreased (J, K). Differences in sample group genera were compared by using either the Wilcoxon rank test for variables with two groups or the Kruskal–Wallis test for variables with more than two groups. To correct for multiple comparisons, Bonferroni-adjusted significance levels were set for each analysis. *p < 0.05, **p < 0.01 vs. healthy controls (HC). existence of a wide disparity of results among human case–control higher abundance of Bilophila and Alistipes could be responsible studies on the relationships between MDD and fecal microbiota for the effects on the immune response detected in this group. [44]. Therefore, there is not a minimal consensus regarding neither There are other mechanisms through which these two bacterial microbial diversity, relative abundance, nor directionality of genera may be involved in MDD, one of them being a possible differences in taxa associated with MDD. Such heterogeneity of relationship of Bilophila with mood-related disorders. In particular, results may be due to differences in the studied populations as mice subjected to nerve injury presented an anhedonia-like well as the different methodologies used in the different studies. phenotype and increased stool levels of this genus [46]. Another Our study comprised a white Caucasian population, so further conceivable mechanism would be based in the fact that Alistipes research is warranted, and a definitive conclusion about the species are indole-positive and, therefore, may affect tryptophan effects of MDD in bacterial diversity cannot be reached at present. availability [47]. As tryptophan is the precursor of serotonin (5-HT), Genera Bilophila (phylum Proteobacteria) and Alistipes (phylum higher levels of Alistipes may disrupt the serotonergic system Bacteroidetes) were higher and Anaerostipes and Dialister (phylum balance. Firmicutes) were lower in the feces of patients with MDD when Bilophila presented higher levels in r-MDD and Alistipes in compared with HC. Furthermore, sequences corresponding to a-MDD patients, compared with HC, indicating that some Alistipes were higher, while those of Anaerostipes and Dialister elements of the microbiota are changing through the MDD were lower in a-MDD patients. Finally, r-MDD patients presented course with a possible shift toward a percentage increase in the higher abundance of Bilophila compared with HC. To our Gram-negative bacterial ratio, although we cannot infer this knowledge, this is the first report, showing a higher Bilophila presumed shift based in our results and further research is abundance in fecal samples from patients recovered from MDD. A needed. In this case, a Gram-negative genus related to the higher amount of Alistipes had been previously found in relation to impairment of the serotonergic system (i.e., Alistipes) would be MDD patients [29, 45]. higher in the active phase of MDD, while a Gram-negative genus Bilophila and Alistipes are Gram-negative bacteria, and conse- associated with prolonged pathologies (i.e., Bilophila), including quently, the LPS from their membrane can stimulate the innate inflammatory bowel disease (IBD), would be higher in the r-MDD immune system via TLR-4 activation [12] after an intestinal patients. dysfunction allowing bacterial translocation. Importantly, TLR-4 Anaerostipes and Dialister showed lower levels in MDD activation induces depressive-like behaviors in animal models, and compared with HC. Anaerostipes is a Gram-positive anaerobic it has been proposed as a crucial factor in the inflammatory theory genus from the phylum Firmicutes. Anaerostipes species can of depression [9, 23]. Here, plasma LPS levels are higher in the metabolize carbohydrates producing butyrate, a short-chain fatty group with the worse clinical condition (a-MDD), and thus, the acid that plays a key role for gut homeostasis [48, 49] with Translational Psychiatry (2021)11:645 J.R. Caso et al. 6 Fig. 2 Effects of the MDD on the innate immune pathway. TLR-4 protein expression in PBMCs (A) and LPS plasma levels (B) were not affected by the MDD. Plasma levels of HMGB1 decreased in the MDD group (C). TLR-4 expression was not altered in the a-MDD group and was increased in the r-MDD group compared with HC (D). LPS plasma levels augmented in the a-MDD and were not affected in the r-MDD group compared with HC (E). HMGB1 plasma levels decreased in the r-MDD group compared with HC (F). Data are expressed as mean ± standard error of the mean (SEM). In the western blots the densitometric data of the respective bands of interest are normalized by β-actin (lower band). Samples were from parallel experiments and that gels/blots were processed in parallel. An unpaired two-tailed t-test was performed when comparing two groups. For more than two groups comparisons, a one-way ANOVA with a Tukey post-hoc test was employed, and in those cases in which the data did not follow a Gaussian distribution, a nonparametric ANOVA with a Kruskal–Wallis test followed by a Dunn’s post-hoc test was performed. *p < 0.05, **p < 0.01 vs. healthy controls (HC). immunosuppressive and anti-inflammatory functions [44]. Thus, Anaerostipes and Dialister declined in a-MDD patients. A the lower Anaerostipes abundance could explain the effects on the previous report found that Dialister is relatively more abundant inflammatory and oxidative/nitrosative stress responses detected in HC than in a-MDD patients [29]. This reduction in a-MDD could in MDD. Moreover, Anaerostipes improves depressive-like beha- be supporting the protective nature of these genera against viors in stressed mice and through an improvement of the depression, through mechanisms mentioned above (e.g., immu- serotonergic system balance and increasing trophic factors nosuppression and prevention of bacterial translocation) and expression [50]. The lower levels of Anaerostipes observed in our clearly deserves further consideration. Dialister was not completely patients with MDD could be related to a critical mechanism depleted in r-MDD suggesting a protective nature of this genus. involved in the pathophysiology of MDD at the microbiota level. We aimed to recognize the potential role of TLR-4 in our However, we are still far from being able to establish causality, and scenario as one of the predominant sources of the TLR-4 further research is warranted. activation is the altered gut microbiota [21] and the subsequent Dialister is a Gram-negative genus from the phylum Firmicutes. shifts in microbiota through increased intestinal barrier perme- The relative significance as well as broad microbiological data on ability that has been suggested in patients with MDD [27, 54]. Dialister species in human clinical samples remains rather scarce. It MDD patients present lower plasma levels of HMGB1, an has been positively correlated with spondyloarthritis activity [51]. important mediator in innate immunity, characterized as an Some Dialister species are capable of generating both acetate and alarmin or danger signal involved in the LPS-induced depressive- propionate, and in Crohn’s disease, a condition in which there is like behavior [55]. The reduction in the HMGB1 levels could seem an increased permeability of the intestinal mucosal barriers counterintuitive and contrary to the inflammatory hypothesis; allowing bacterial translocation, a decrease of Dialister has been however, when splitting the groups, this reduction is mostly reported [52]. Curiously, a recent study has shown that Dialister is accounted by the low levels detected in the r-MDD group. Hence, depleted in MDD, even after correcting for the confounding reduced HMGB1 levels among our MDD patients could be related effects of antidepressants, and that is positively associated with to the recovery process. Further research is necessary, but the quality-of-life indicators [53]. In fact, Dialister was considered as a HMGB1 decrease could be explaining the higher TLR-4 expression potential psychobiotic genus and, consequently, as a main target detected in r-MDD; it is plausible that as the plasma levels of the for follow-up research. Our data show a Dialister depletion in MDD, ligand (HMGB1) are dropping, the expression levels of the which would agree with the view of a genus negatively associated receptor are climbing, as a compensatory mechanism. However, with depression and positively associated with a healthy condition the LPS raise detected in patients with a-MDD could also cause the for the host. TLR-4 expression increment: r-MDD patients have been a-MDD Translational Psychiatry (2021)11:645 J.R. Caso et al. 7 Fig. 3 Effects of MDD on intra- and intercellular inflammatory parameters. The ratio between the activated ERK (pERK) and the ERK total form decreased in PBMCs from patients with MDD compared with HC (A). Patients with MDD did not show differences in the protein levels of NF-κB nor in the plasma levels of CRP and IL-6 compared with HC (B–D). a-MDD patients did not show differences when compared with HC (E–H). pERK/ERK ratio in PBMCs decreased (E) and plasma IL-6 increased (H) in r-MDD patients compared with HC. Plasma CRP levels in the r-MDD group decreased compared with the a-MDD group (G). Data are expressed as mean ± standard error of the mean (SEM). In the western blots the densitometric data of the respective bands of interest are normalized by β-actin or by GAPDH (lower band). Samples were from parallel experiments and that gels/blots were processed in parallel. An unpaired two-tailed t-test was performed when comparing two groups. For more than two groups comparisons, a one-way ANOVA with a Tukey post-hoc test was employed, and in those cases in which the data did not follow a Gaussian distribution, a nonparametric ANOVA with a Kruskal–Wallis test followed by a Dunn’s post-hoc test was performed. *p < 0.05, **p < 0.01, ***p < 0.001 vs. healthy controls (HC); #p < 0.05 vs. a-MDD. formerly, and thus, the higher LPS levels during their a-MDD depression than in mild/moderate depression [59]. High levels of period could be activating the TLR-4, causing the TLR-4 increment CRP are associated with antidepressant treatment resistance and detected in their r-MDD phase. lower remission rates [60]. Our data showed lower CRP levels in LPS presented higher levels only in a-MDD compared with HC. r-MDD. This would agree with this view of CRP as an indicator of LPS from Gram-negative bacteria is the main ligand of the TLR-4 current severity and a potential predictor of antidepressant response. and a parameter widely related to depressive-like behavior [9]. A Meta-analyses indicate that IL-6 levels are elevated in the blood possible cause for this increment could be the dysfunction of the of MDD patients [2, 58], suggesting that IL-6 levels might serve as intestinal barrier and the ensuing bacterial translocation. Interest- a predictive biomarker. As IL-6 acts on so many diverse tissues ingly, there are higher levels of some Gram-negative genera (i.e., throughout the organism, dysregulation of this cytokine can Alistipes) in the stool of the a-MDD patients. precipitate a multitude of events relevant to depression [61]. In MDD patients the ratio pERK/ERK in PBMCs presented lower Importantly, IL-6 has many functions within the immune system values compared with HC. The r-MDD group presented lower levels depending on the type of organ it is acting upon and its signaling of the pERK/ERK ratio in PBMCs and plasma CRP and higher plasma is complex and resulting in both inflammatory and anti- levels of IL-6. ERK is a downstream element activated after TLR-4 inflammatory cascades [62]. Our results show higher IL-6 levels stimulation, implicated in numerous signaling cascades wherein in the r-MDD group compared with HC. A possible explanation various extracellular stimuli induce inflammation [13]. Several studies could be the actions of the antidepressants, as in vitro studies associate the ERK cascade with the etiology and treatment of employing whole blood cultures from MDD patients show that depression [56]. Postmortem brain studies indicate that individuals some antidepressants induce an increment in the peripheral levels who committed suicide have ERK abundance and activity. Con- of several cytokines, including IL-6 [63]. Another possibility could versely, in experimental studies the ERK cascade is activated by be that the IL-6 detected would be acting as anti-inflammatory antidepressants, and apparently, its antidepressant effects involve through its classical signaling. neurotrophic and growth factors [57]. Here, the lower ERK levels in The inducible iNOS expression is higher in MDD patients MDD patients could be the result of its expression levels in the compared with HC. This increment is maintained in both a-MDD r-MDD group, which is the group with a better clinical condition. It and r-MDD groups, but the TBARS, a marker of damage produced could be possible that antidepressant treatments have stimulated by oxidative stress, is only higher in the a-MDD group, and the two this pathway and when achieving the response phase, this pathway antioxidant enzymes analyzed were not affected by the disease. had been exhausted. In addition, it could be possible that the Oxidative/nitrosative stress is increased in MDD [64] and it can be a lessening of this MAPK is resulting from the achievement of a better final consequence of the inflammation detected in this disease [65]. clinical condition. Whatever the case may be, further research is Preclinical research shows that the oxidative/nitrosative stress caused needed to fully understand the role of ERK in the MDD. by translocated bacteria, as well as the parameters involved in this CRP has been measured in numerous prior studies of MDD [58]. stress regulation, are implicated in the actions of a depression model There is heterogeneity of effect size between studies that may be on the CNS and the depressive-like behavior [25]. More research is attributable to clinical variability, with higher CRP in severe warranted, but oxidative/nitrosative stress appears to be a crucial Translational Psychiatry (2021)11:645 J.R. Caso et al. 8 Fig. 4 Effects of the MDD on the cyclooxygenase-2 pathway, on the oxidative/nitrosative response, and on counterbalancing mechanisms. Patients with MDD presented lower levels of COX-2 expression in PBMCs (A) compared with HC, without changes in PGE2 (B) and in 15d-PGJ2 (C) plasma levels. The MDD did not affect the PPARγ expression (D). The a-MDD and r-MDD groups did not show any difference between them nor with the HC group (E–H). The MDD increased iNOS protein levels in PBMCs (I) but no affected the activity of TBARS, SOD and GPx (J–L). a-MDD and r-MDD groups presented increased iNOS expression in PBMCs (M). The a-MDD group shown increased TBARS levels compared with HC (N). There were no differences in the activity of SOD and GPx among groups (O, P). Data are expressed as mean ± standard error of the mean (SEM). In the western blots the densitometric data of the respective bands of interest are normalized by β-actin or by GAPDH (lower band). Samples were from parallel experiments and that gels/blots were processed in parallel. An unpaired two- tailed t-test was performed when comparing two groups. For more than two groups comparisons, a one-way ANOVA with a Tukey post-hoc test was employed, and in those cases in which the data did not follow a Gaussian distribution, a nonparametric ANOVA with a Kruskal–Wallis test followed by a Dunn’s post-hoc test was performed. *p < 0.05, ***p < 0.001 vs. healthy controls (HC). mechanism through which the inflammation caused by the potential hampered by the lack of a dedicated reference database of gut bacterial translocation would be affecting the clinical evolution of microbial neuroactive metabolic potential [53]. MDD patients. Similarly, more factors involving the antioxidant In summary, there are bacterial genera with potential machinery should be studied. implications in the pathophysiology of depression, whose This study has some limitations. One is the number of samples abundance is altered in the feces of MDD patients. Besides, caused by the difficulties to obtain a high number of this kind of a-MDD and r-MDD patients present differences among them that samples from well-characterized patients. This limitation could are worthy of further consideration, being the a-MDD patients explain the absence of significant results when the bacteria the group showing the most noticeable changes. In addition, diversity was analyzed. Anyhow, the number of samples allowed bacterial translocation (inferred by the increment in the plasma to get enough statistical power to discover differences in bacteria levels of LPS) could be affecting the immune response, and the genera and in biochemical parameters. Another limitation is the oxidative/nitrosative stress seems to be a crucial mechanism absence of a clear way to relate the theoretical bacterial through which the inflammation would be affecting the clinical translocation with the immune system, although the plasma LPS evolution of MDD patients. It is not clear whether stimulation of levels and the TLR-4 activation in PBMCs could be considered the immune parameters associated with MDD is a precipitating indirect indicators. Microbiota composition is influenced by other event of the disorder or a process within MDD. However, data factors such as lifestyle and medication use, and this is also a seem to point to an inflammatory process in those patients with limitation. Finally, there is a common limitation to the microbiota a more severe condition, recurrent symptoms, and/or treatment and CNS studies, namely that the practical elucidation of resistance. Thus, identifying compositional microbiota markers metagenomes in a MGB axis context persists challenging and is characteristic for MDD patients and their links with the potential Translational Psychiatry (2021)11:645 J.R. Caso et al. 9 inflammatory states present in those patients could represent a 26. Maes M, Kubera M, Leunis JC. The gut-brain barrier in major depression: intestinal promising therapeutic approach, especially considering that mucosal dysfunction with an increased translocation of LPS from gram negative dietary manipulations can have an impact on the gut microbiota, enterobacteria (leaky gut) plays a role in the inflammatory pathophysiology of potentially facilitating therapeutic interventions in some MDD depression. Neuro Endocrinol Lett. 2008;29:117–24. 27. Maes M, Kubera M, Leunis JC, Berk M. Increased IgA and IgM responses against patients [66]. gut commensals in chronic depression: further evidence for increased bacterial translocation or leaky gut. J Affect Disord. 2012;141:55–62. 28. Socala K, Doboszewska U, Szopa A, Serefko A, Wlodarczyk M, Zielinska A, et al. REFERENCES The role of microbiota-gut-brain axis in neuropsychiatric and neurological dis- 1. Liu Y, Ho RC, Mak A. Interleukin (IL)−6, tumour necrosis factor alpha (TNF-alpha) orders. Pharmacol Res. 2021;172:105840. and soluble interleukin-2 receptors (sIL-2R) are elevated in patients with major 29. Jiang H, Ling Z, Zhang Y, Mao H, Ma Z, Yin Y, et al. Altered fecal microbiota depressive disorder: a meta-analysis and meta-regression. J Affect Disord. composition in patients with major depressive disorder. Brain Behav Immun. 2012;139:230–9. 2015;48:186–94. 2. Dowlati Y, Herrmann N, Swardfager W, Liu H, Sham L, Reim EK, et al. A meta- 30. American Psychiatric Association. Diagnostic and statistical manual of mental analysis of cytokines in major depression. Biol Psychiatry. 2010;67:446–57. disorders. 4th ed. Washington DC: American Psychiatric Association; 2000. 3. Kohler-Forsberg O, C NL, Hjorthoj C, Nordentoft M, Mors O, Benros ME. Efficacy of 31. Ramos-Brieva JA, Cordero-Villafafila A. A new validation of the Hamilton Rating anti-inflammatory treatment on major depressive disorder or depressive symp- Scale for Depression. J Psychiatr Res. 1988;22:21–8. toms: meta-analysis of clinical trials. Acta Psychiatr Scand. 2019;139:404–19. 32. Badia X, Roset M, Montserrat S, Herdman M, Segura A. [The Spanish version o 4. Bauer ME, Teixeira AL. Inflammation in psychiatric disorders: what comes first? fEuroQol: a description and its applications. European Quality of Life scale]. Med Ann N Y Acad Sci. 2019;1437:57–67. Clin. 1999;112:79–85. 5. Valkanova V, Ebmeier KP, Allan CL. CRP, IL-6 and depression: a systematic review 33. Remor E. Psychometric properties of a European Spanish version of the Perceived and meta-analysis of longitudinal studies. J Affect Disord. 2013;150:736–44. Stress Scale (PSS). Span J Psychol. 2006;9:86–93. 6. Cattaneo A, Gennarelli M, Uher R, Breen G, Farmer A, Aitchison KJ, et al. Candidate 34. Motrico E, Moreno-Kustner B, de Dios Luna J, Torres-Gonzalez F, King M, Nazareth I, genes expression profile associated with antidepressants response in the GEN- et al. Psychometric properties of the list of threatening experiences_LTE and its DEP study: differentiating between baseline ‘predictors’ and longitudinal ‘targets’. association with psychosocial factors and mental disorders according to different Neuropsychopharmacology. 2013;38:377–85. scoring methods. J Affect Disord. 2013;150:931–40. 7. Fleshner M, Frank M, Maier SF. Danger signals and inflammasomes: stress-evoked 35. Hernandez A, Gallardo-Pujol D, Pereda N, Arntz A, Bernstein DP, Gaviria AM, et al. sterile inflammation in mood disorders. Neuropsychopharmacology. 2017;42:36–45. Initial validation of the Spanish childhood trauma questionnaire-short form: 8. Ratajczak MZ, Pedziwiatr D, Cymer M, Kucia M, Kucharska-Mazur J, Samochowiec factor structure, reliability and association with parenting. J Interpers Violence. J. Sterile inflammation of brain, due to activation of innate immunity, as a culprit 2013;28:1498–518. in psychiatric disorders. Front Psychiatry. 2018;9:60. 36. Drossman DA, Dumitrascu DL. Rome III: new standard for functional gastro- 9. Garcia-Bueno B, Caso JR, Madrigal JL, Leza JC. Innate immune receptor Toll-like intestinal disorders. J Gastrointestin Liver Dis. 2006;15:237–41. receptor 4 signalling in neuropsychiatric diseases. Neurosci Biobehav Rev. 37. Klindworth A, Pruesse E, Schweer T, Peplies J, Quast C, Horn M, et al. Evaluation of 2016;64:134–47. general 16S ribosomal RNA gene PCR primers for classical and next-generation 10. Walsh JG, Muruve DA, Power C. Inflammasomes in the CNS. Nat Rev Neurosci. sequencing-based diversity studies. Nucleic Acids Res. 2013;41:e1. 2014;15:84–97. 38. Schmieder R, Edwards R. Quality control and preprocessing of metagenomic 11. Garate I, Garcia-Bueno B, Madrigal JL, Caso JR, Alou L, Gomez-Lus ML, et al. Stress- datasets. Bioinformatics. 2011;27:863–4. induced neuroinflammation: role of the Toll-like receptor-4 pathway. Biol Psy- 39. Aronesty E. ea-utils: command-line tools for processing biological sequencing chiatry. 2013;73:32–43. data. ea-utils; 2011. https://expressionanalysis.github.io/ea-utils/. 12. Takeda K, Kaisho T, Akira S. Toll-like receptors. Neurosci Biobehav Rev. 40. Cole JR, Wang Q, Cardenas E, Fish J, Chai B, Farris RJ, et al. The Ribosomal 2016;64:134–47. Database Project: improved alignments and new tools for rRNA analysis. Nucleic 13. Guha M, Mackman N. LPS induction of gene expression in human monocytes. Acids Res. 2009;37:D141–5. Cell Signal. 2001;13:85–94. 41. Caporaso JG, Kuczynski J, Stombaugh J, Bittinger K, Bushman FD, Costello EK, 14. He C, Narayanan PK, Fort MM. Assessment of the performance of three multiplex et al. QIIME allows analysis of high-throughput community sequencing data. Nat array panels for the detection of circulating cytokines and chemokines in naive, Methods. 2010;7:335–6. LPS, and SEB-treated cynomolgus macaques. Brain Behav Immun. 2013;28:54–62. 42. Zakrzewski M, Proietti C, Ellis JJ, Hasan S, Brion MJ, Berger B, et al. Calypso: a user- 15. Maslanik T, Mahaffey L, Tannura K, Beninson L, Greenwood BN, Fleshner M. The friendly web-server for mining and visualizing microbiome-environment inter- inflammasome and danger associated molecular patterns (DAMPs) are impli- actions. Bioinformatics. 2017;33:782–3. cated in cytokine and chemokine responses following stressor exposure. Brain 43. Garcia-Bueno B, Bioque M, Mac-Dowell KS, Barcones MF, Martinez- Behav Immun. 2013;28:54–62. Cengotitabengoa M, Pina-Camacho L, et al. Pro-/anti-inflammatory dysregula- 16. Foster JA, McVey Neufeld KA. Gut-brain axis: how the microbiome influences tion in patients with first episode of psychosis: toward an integrative inflam- anxiety and depression. Trends Neurosci. 2013;36:305–12. matory hypothesis of schizophrenia. Schizophr Bull. 2014;40:376–487. 17. Dinan TG, Cryan JF. The impact of gut microbiota on brain and behaviour: 44. Cheung SG, Goldenthal AR, Uhlemann AC, Mann JJ, Miller JM, Sublette ME. implications for psychiatry. Curr Opin Clin Nutr Metab Care. 2015;18:552–8. Systematic review of gut microbiota and major depression. Front Psychiatry. 18. Clemente JC, Ursell LK, Parfrey LW, Knight R. The impact of the gut microbiota on 2019;10:34. human health: an integrative view. Cell. 2012;148:1258–70. 45. Naseribafrouei A, Hestad K, Avershina E, Sekelja M, Linlokken A, Wilson R, et al. 19. Cryan JF, Dinan TG. Mind-altering microorganisms: the impact of the gut Correlation between the human fecal microbiota and depression. Neurogas- microbiota on brain and behaviour. Nat Rev Neurosci. 2012;13:701–12. troenterol Motil. 2014;26:1155–62. 20. Naveed M, Zhou QG, Xu C, Taleb A, Meng F, Ahmed B, et al. Gut-brain axis: a 46. Yang C, Fang X, Zhan G, Huang N, Li S, Bi J, et al. Key role of gut microbiota in matter of concern in neuropsychiatric disorders…! Prog Neuropsychopharmacol anhedonia-like phenotype in rodents with neuropathic pain. Transl Psychiatry. Biol Psychiatry. 2021;104:110051. 2019;9:57. 21. Zheng P, Zeng B, Zhou C, Liu M, Fang Z, Xu X, et al. Gut microbiome remodeling 47. Song Y, Kononen E, Rautio M, Liu C, Bryk A, Eerola E, et al. Alistipes onderdonkii sp. induces depressive-like behaviors through a pathway mediated by the host’s nov. and Alistipes shahii sp. nov., of human origin. Int J Syst Evol Microbiol. metabolism. Mol Psychiatry. 2016;21:786–96. 2006;56:1985–90. 22. Jiang W, Wu N, Wang X, Chi Y, Zhang Y, Qiu X, et al. Dysbiosis gut microbiota 48. Eeckhaut V, Van Immerseel F, Pasmans F, De Brandt E, Haesebrouck F, Ducatelle associated with inflammation and impaired mucosal immune function in intes- R, et al. Anaerostipes butyraticus sp. nov., an anaerobic, butyrate-producing bac- tine of humans with non-alcoholic fatty liver disease. Sci Rep. 2015;5:8096. terium from Clostridium cluster XIVa isolated from broiler chicken caecal content, 23. Miller AH, Raison CL. The role of inflammation in depression: from evolutionary and emended description of the genus Anaerostipes. Int J Syst Evol Microbiol. imperative to modern treatment target. Nat Rev Immunol. 2016;16:22–34. 2010;60:1108–12. 24. Garate I, Garcia-Bueno B, Madrigal JL, Bravo L, Berrocoso E, Caso JR, et al. Origin 49. Schwiertz A, Hold GL, Duncan SH, Gruhl B, Collins MD, Lawson PA, et al. Anae- and consequences of brain Toll-like receptor 4 pathway stimulation in an rostipes caccae gen. nov., sp. nov., a new saccharolytic, acetate-utilising, butyrate- experimental model of depression. J Neuroinflammation. 2011;8:151. producing bacterium from human faeces. Syst Appl Microbiol. 2002;25:46–51. 25. Martin-Hernandez D, Caso JR, Bris AG, Maus SR, Madrigal JL, Garcia-Bueno B, et al. 50. Sun J, Wang F, Hong G, Pang M, Xu H, Li H, et al. Antidepressant-like effects of Bacterial translocation affects intracellular neuroinflammatory pathways in a sodium butyrate and its possible mechanisms of action in mice exposed to depression-like model in rats. Neuropharmacology. 2016;103:122–33. chronic unpredictable mild stress. Neurosci Lett. 2016;618:159–66. Translational Psychiatry (2021)11:645 J.R. Caso et al. 10 51. Picchianti-Diamanti A, Rosado MM, D’Amelio R. Infectious agents and inflam- AUTHOR CONTRIBUTIONS mation: the role of microbiota in autoimmune arthritis. Front Microbiol. JRC conceived and performed experiments involving the characterization of the 2017;8:2696. inflammatory pathways, interpreted the data, and wrote the manuscript. KSM 52. Joossens M, Huys G, Cnockaert M, De Preter V, Verbeke K, Rutgeerts P, et al. performed experiments involved in the characterization of the inflammatory pathways Dysbiosis of the faecal microbiota in patients with Crohn’s disease and their and contributed to the writing of the manuscript. AG-P, SG, JdD-A, MC-S, FS, JC-V, PG-G, unaffected relatives. Gut. 2011;60:631–7. CDlC, and LA recruited the patients and controls and performed the clinical and 53. Valles-Colomer M, Falony G, Darzi Y, Tigchelaar EF, Wang J, Tito RY, et al. The psychological assessments. MLG-L interpreted the microbiological results. CA and JMR neuroactive potential of the human gut microbiota in quality of life and performed the bioinformatics analyses and interpreted the microbiological results. JCL depression. Nat Microbiol. 2019;4:623–32. designed the study, interpreted the data, and contributed to the writing of the 54. Berk M, Williams LJ, Jacka FN, O’Neil A, Pasco JA, Moylan S, et al. So depression is manuscript. All authors read and approved the final version of the manuscript. an inflammatory disease, but where does the inflammation come from? BMC Med. 2013;11:200. 55. Wu TY, Liu L, Zhang W, Zhang Y, Liu YZ, Shen XL, et al. High-mobility group box-1 FUNDING was released actively and involved in LPS induced depressive-like behavior. J This work was supported by grants from the Spanish Ministries of Health-ISCIII (FIS Psychiatr Res. 2015;64:99–106. PI13/01102) and MICINN-FEDER Funds (SAF2016-75500-R and PID2019-109033RB-I00) 56. Tanis KQ, Duman RS. Intracellular signaling pathways pave roads to recovery for to JCL and by CIBERSAM. JRC is a Ramon y Cajal Researcher (Spanish Ministry of mood disorders. Ann Med. 2007;39:531–44. Science and Innovation and FEDER). JdD-A was supported by the Catalan 57. Mercier G, Lennon AM, Renouf B, Dessouroux A, Ramauge M, Courtin F, et al. MAP Intensification Programme (PERIS, SLT008/18/00092; Generalitat de Catalunya). kinase activation by fluoxetine and its relation to gene expression in cultured rat astrocytes. J Mol Neurosci. 2004;24:207–16. 58. Haapakoski R, Mathieu J, Ebmeier KP, Alenius H, Kivimaki M. Cumulative meta- analysis of interleukins 6 and 1beta, tumour necrosis factor alpha and C-reactive COMPETING INTERESTS protein in patients with major depressive disorder. Brain Behav Immun. The authors declare no competing interests. 2015;49:206–15. 59. Chamberlain SR, Cavanagh J, de Boer P, Mondelli V, Jones DNC, Drevets WC, et al. Treatment-resistant depression and peripheral C-reactive protein. Br J Psychiatry. ADDITIONAL INFORMATION 2019;214:11–19. Supplementary information The online version contains supplementary material 60. Zhang J, Yue Y, Thapa A, Fang J, Zhao S, Shi W, et al. Baseline serum C-reactive available at https://doi.org/10.1038/s41398-021-01755-3. protein levels may predict antidepressant treatment responses in patients with major depressive disorder. J Affect Disord. 2019;250:432–8. Correspondence and requests for materials should be addressed to Juan C. Leza. 61. Hodes GE, Menard C, Russo SJ. Integrating Interleukin-6 into depression diagnosis and treatment. Neurobiol Stress. 2016;4:15–22. Reprints and permission information is available at http://www.nature.com/ 62. Wolf J, Rose-John S, Garbers C. Interleukin-6 and its receptors: a highly regulated reprints and dynamic system. Cytokine. 2014;70:11–20. 63. Munzer A, Sack U, Mergl R, Schonherr J, Petersein C, Bartsch S, et al. Impact of Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims antidepressants on cytokine production of depressed patients in vitro. Toxins. in published maps and institutional affiliations. 2013;5:2227–40. 64. Black CN, Bot M, Scheffer PG, Cuijpers P, Penninx BW. Is depression associated with increased oxidative stress? A systematic review and meta-analysis. Psy- choneuroendocrinology. 2015;51:164–75. 65. Moylan S, Berk M, Dean OM, Samuni Y, Williams LJ, O’Neil A, et al. Oxidative & Open Access This article is licensed under a Creative Commons nitrosative stress in depression: why so much stress? Neurosci Biobehav Rev. Attribution 4.0 International License, which permits use, sharing, 2014;45:46–62. adaptation, distribution and reproduction in any medium or format, as long as you give 66. David LA, Maurice CF, Carmody RN, Gootenberg DB, Button JE, Wolfe BE, et al. appropriate credit to the original author(s) and the source, provide a link to the Creative Diet rapidly and reproducibly alters the human gut microbiome. Nature. Commons license, and indicate if changes were made. The images or other third party 2014;505:559–63. material in this article are included in the article’s Creative Commons license, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons license and your intended use is not permitted by statutory ACKNOWLEDGEMENTS regulation or exceeds the permitted use, you will need to obtain permission directly The authors would like to thank Dr. Pedro Holgado Madera, Dr. Carmen Ortega from the copyright holder. To view a copy of this license, visit http://creativecommons. Bernardo, and Dr. Carlo Alemany for their collaboration with the clinical aspects of org/licenses/by/4.0/. the study and the Fundación para el Fomento de la Investigación Biomédica y Sanitaria de la Comunidad Valenciana (FISABIO) for their services in the microbiota © The Author(s) 2021 studies. Translational Psychiatry (2021)11:645