Testing the predictive value of peripheral gene expression for nonremission following citalopram treatment for major depression

Systematic review and mega-analysis of the peripheral blood transcriptome in depression implicates dysregulation of lymphoid cells and histones

Background

Depression has been associated with transcriptomic changes in peripheral blood. However, the contribution of specific immune cell subsets or pathways remains unclear, and findings have been variable across previous studies, which have not tended to account for sample cellular composition.

Methods

We performed a systematic review of peripheral blood transcriptome studies in depression. For the five datasets meeting criteria (total N=6,011), we performed harmonized reprocessing and cell-composition-adjusted differential gene and transcript analyses, followed by a bias- and inflation-adjusted weighted Z-score mega-analysis. We investigated the biological pathways and cell subsets implicated by the results. We also performed a sex-stratified gene network mega-analysis using consensus weighted gene co-expression network analysis (WGCNA).

Results

Few genes showed robust differential gene expression (DGE) in depression. Depression was reproducibly associated with decreases in replication-dependent histones, and with a decrease in oxidative phosphorylation pathways in females only. Cell source analyses implicated lymphoid cells (T cells and NK cells) as likely contributors to the depression differential expression signature. WGCNA mega-analysis revealed multiple consensus modules associated with depression, with a PUF60-related module upregulated in both female and male depression in sex-stratified analyses. Two genes predicted to be causally relevant to depression by transcriptome-wide association studies (GPX4 and GYPE) showed significant DGE.

Conclusions

These results are convergent with immunogenetic evidence implicating lymphoid cell dysregulation in depression, while also highlighting histone alterations as a key molecular signature in depression. They also indicate the importance of large-scale datasets for biomarker discovery in the context of heterogeneous disorders like depression.

Artificial intelligence in mental health care: a systematic review of diagnosis, monitoring, and intervention applications

Artificial intelligence (AI) has been recently applied to different mental health illnesses and healthcare domains. This systematic review presents the application of AI in mental health in the domains of diagnosis, monitoring, and intervention. A database search (CCTR, CINAHL, PsycINFO, PubMed, and Scopus) was conducted from inception to February 2024, and a total of 85 relevant studies were included according to preestablished inclusion criteria. The AI methods most frequently used were support vector machine and random forest for diagnosis, machine learning for monitoring, and AI chatbot for intervention. AI tools appeared to be accurate in detecting, classifying, and predicting the risk of mental health conditions as well as predicting treatment response and monitoring the ongoing prognosis of mental health disorders. Future directions should focus on developing more diverse and robust datasets and on enhancing the transparency and interpretability of AI models to improve clinical practice.

Transcriptomics and psychotherapy: An integrative review

Gold-standard psychotherapies like cognitive-behavioral therapy (CBT) show beneficial effects, but patient responses vary, indicating a need to predict and optimize treatment efficacy. Gene expression analysis may offer insights into the interplay between psychosocial processes and biological factors that impact psychopathology and therapeutic response. This integrative review examines 17 studies that assess gene expression in the context of psychotherapy, highlighting innovative frameworks for incorporating gene expression analysis in diagnosis, predicting treatment response, and monitoring treatment progress. Current evidence points to transcriptional control pathways downstream of the hypothalamic-pituitary-adrenal (HPA)-axis and sympathetic nervous system (SNS) signaling pathways, particularly their effects on immune cells (e.g., pro-inflammatory processes and wound healing), as key areas for future research. Higher-level pathway analyses, whether theory-based or empirically driven, appear to offer the most robust framework for future studies. This review also discusses significant limitations of current literature and proposes directions for future research.

Investigation of Association Between Expression of DYX1C1, KIAA0319, and ROBO1 Genes and Specific Learning Disorder in Children and Adolescents

Specific learning disorder (SLD) is prevalent worldwide and is a complex disorder with variable symptoms and significant differences among individuals. Epigenetic markers may alter susceptibility to neurodevelopmental disorders (NDDs). Aberrant expression of protein-coding (mRNA) genes in this pathology shows that the detection of epigenetic molecular biomarkers is of increasing importance in the diagnosis and treatment of individuals with SLD. We compared gene expression level of dyslexia susceptibility 1 candidate gene 1 (DYX1C1), dyslexia-associated protein KIAA0319 (KIAA0319), and roundabout guidance receptor 1 (ROBO1) between children with SLD and healthy children by performing quantitative polymerase chain reaction (qPCR). In addition, we evaluated these gene expressions of severe children with SLD compared to non-severe and male SLD children compared to females. The expression of the DYX1C1, KIAA0319, and ROBO1 genes was statistically significantly upregulated in children with SLD (P < 0.05*). DYX1C1 was also upregulated in severe SLD children (P = 0.03*). In addition, KIAA0319 and ROBO1 genes were differentially expressed in male SLD children compared to females (P < 0.05*). Furthermore, we found that DYX1C1 and ROBO1 genes significantly affect the likelihood of the SLD (respectively, P < 0.001** and P = 0.007*). We expect that the findings provided from this study may contribute to the determination expression level of the relevant genes in the diagnosis, prognosis, and treatment of SLD. In addition, our findings could be a guide for future epigenetics studies on the use of the DYX1C1, KIAA0319, and ROBO1 in therapeutic applications in the SLD.

Multimodal imaging measures in the prediction of clinical response to deep brain stimulation for refractory depression: A machine learning approach

OBJECTIVES

This study compared machine learning models using unimodal imaging measures and combined multi-modal imaging measures for deep brain stimulation (DBS) outcome prediction in treatment resistant depression (TRD).

METHODS

Regional brain glucose metabolism (CMRGlu), cerebral blood flow (CBF), and grey matter volume (GMV) were measured at baseline using 18F-fluorodeoxy glucose (18F-FDG) positron emission tomography (PET), arterial spin labelling (ASL) magnetic resonance imaging (MRI), and T1-weighted MRI, respectively, in 19 patients with TRD receiving subcallosal cingulate (SCC)-DBS. Responders (n = 9) were defined by a 50% reduction in HAMD-17 at 6 months from the baseline. Using an atlas-based approach, values of each measure were determined for pre-selected brain regions. OneR feature selection algorithm and the naïve Bayes model was used for classification. Leave-out-one cross validation was used for classifier evaluation.

RESULTS

The performance accuracy of the CMRGlu classification model (84%) was greater than CBF (74%) or GMV (74%) models. The classification model using the three image modalities together led to a similar accuracy (84%0 compared to the CMRGlu classification model.

CONCLUSIONS

CMRGlu imaging measures may be useful for the development of multivariate prediction models for SCC-DBS studies for TRD. The future of multivariate methods for multimodal imaging may rest on the selection of complementing features and the developing better models.Clinical Trial Registration: ClinicalTrials.gov (#NCT01983904).

An integrated precision medicine approach in major depressive disorder: a study protocol to create a new algorithm for the prediction of treatment response

Major depressive disorder (MDD) is the most common psychiatric disease worldwide with a huge socio-economic impact. Pharmacotherapy represents the most common option among the first-line treatment choice; however, only about one third of patients respond to the first trial and about 30% are classified as treatment-resistant depression (TRD). TRD is associated with specific clinical features and genetic/gene expression signatures. To date, single sets of markers have shown limited power in response prediction. Here we describe the methodology of the PROMPT project that aims at the development of a precision medicine algorithm that would help early detection of non-responder patients, who might be more prone to later develop TRD. To address this, the project will be organized in 2 phases. Phase 1 will involve 300 patients with MDD already recruited, comprising 150 TRD and 150 responders, considered as extremes phenotypes of response. A deep clinical stratification will be performed for all patients; moreover, a genomic, transcriptomic and miRNomic profiling will be conducted. The data generated will be exploited to develop an innovative algorithm integrating clinical, omics and sex-related data, in order to predict treatment response and TRD development. In phase 2, a new naturalistic cohort of 300 MDD patients will be recruited to assess, under real-world conditions, the capability of the algorithm to correctly predict the treatment outcomes. Moreover, in this phase we will investigate shared decision making (SDM) in the context of pharmacogenetic testing and evaluate various needs and perspectives of different stakeholders toward the use of predictive tools for MDD treatment to foster active participation and patients' empowerment. This project represents a proof-of-concept study. The obtained results will provide information about the feasibility and usefulness of the proposed approach, with the perspective of designing future clinical trials in which algorithms could be tested as a predictive tool to drive decision making by clinicians, enabling a better prevention and management of MDD resistance.

The emergence, implementation, and future growth of pharmacogenomics in psychiatry: a narrative review

Psychotropic medication efficacy and tolerability are critical treatment issues faced by individuals with psychiatric disorders and their healthcare providers. For some people, it can take months to years of a trial-and-error process to identify a medication with the ideal efficacy and tolerability profile. Current strategies (e.g. clinical practice guidelines, treatment algorithms) for addressing this issue can be useful at the population level, but often fall short at the individual level. This is, in part, attributed to interindividual variation in genes that are involved in pharmacokinetic (i.e. absorption, distribution, metabolism, elimination) and pharmacodynamic (e.g. receptors, signaling pathways) processes that in large part, determine whether a medication will be efficacious or tolerable. A precision prescribing strategy know as pharmacogenomics (PGx) assesses these genomic variations, and uses it to inform selection and dosing of certain psychotropic medications. In this review, we describe the path that led to the emergence of PGx in psychiatry, the current evidence base and implementation status of PGx in the psychiatric clinic, and finally, the future growth potential of precision psychiatry via the convergence of the PGx-guided strategy with emerging technologies and approaches (i.e. pharmacoepigenomics, pharmacomicrobiomics, pharmacotranscriptomics, pharmacoproteomics, pharmacometabolomics) to personalize treatment of psychiatric disorders.

Interferon signaling and hypercytokinemia-related gene expression in the blood of antidepressant non-responders

Only 50% of patients with depression respond to the first antidepressant drug administered. Thus, biomarkers for prediction of antidepressant responses are needed, as predicting which patients will not respond to antidepressants can optimize selection of alternative therapies. We aimed to identify biomarkers that could predict antidepressant responsiveness using a novel data-driven approach based on statistical pattern recognition. We retrospectively divided patients with major depressive disorder into antidepressant responder and non-responder groups. Comprehensive gene expression analysis was performed using peripheral blood without narrowing the genes. We designed a classifier according to our own discrete Bayes decision rule that can handle categorical data. Nineteen genes showed differential expression in the antidepressant non-responder group (n = 15) compared to the antidepressant responder group (n = 15). In the training sample of 30 individuals, eight candidate genes had significantly altered expression according to quantitative real-time polymerase chain reaction. The expression of these genes was examined in an independent test sample of antidepressant responders (n = 22) and non-responders (n = 12). Using the discrete Bayes classifier with the HERC5, IFI6, and IFI44 genes identified in the training set yielded 85% discrimination accuracy for antidepressant responsiveness in the 34 test samples. Pathway analysis of the RNA sequencing data for antidepressant responsiveness identified that hypercytokinemia- and interferon-related genes were increased in non-responders. Disease and biofunction analysis identified changes in genes related to inflammatory and infectious diseases, including coronavirus disease. These results strongly suggest an association between antidepressant responsiveness and inflammation, which may be useful for future treatment strategies for depression.

Transcriptomic Studies of Antidepressant Action in Rodent Models of Depression: A First Meta-Analysis

Antidepressants (ADs) are, for now, the best everyday treatment we have for moderate to severe major depressive episodes (MDEs). ADs are among the most prescribed drugs in the Western Hemisphere; however, the trial-and-error prescription strategy and side-effects leave a lot to be desired. More than 60% of patients suffering from major depression fail to respond to the first AD they are prescribed. For those who respond, full response is only observed after several weeks of treatment. In addition, there are no biomarkers that could help with therapeutic decisions; meanwhile, this is already true in cancer and other fields of medicine. For years, many investigators have been working to decipher the underlying mechanisms of AD response. Here, we provide the first systematic review of animal models. We thoroughly searched all the studies involving rodents, profiling transcriptomic alterations consecutive to AD treatment in naïve animals or in animals subjected to stress-induced models of depression. We have been confronted by an important heterogeneity regarding the drugs and the experimental settings. Thus, we perform a meta-analysis of the AD signature of fluoxetine (FLX) in the hippocampus, the most studied target. Among genes and pathways consistently modulated across species, we identify both old players of AD action and novel transcriptional biomarker candidates that warrant further investigation. We discuss the most prominent transcripts (immediate early genes and activity-dependent synaptic plasticity pathways). We also stress the need for systematic studies of AD action in animal models that span across sex, peripheral and central tissues, and pharmacological classes.

Peripheral proteomic changes after electroconvulsive seizures in a rodent model of non-response to chronic fluoxetine

Major depressive disorder (MDD) is the psychiatric disorder with the highest prevalence in the world. Pharmacological antidepressant treatment (AD), such as selective serotonin reuptake inhibitors [SSRI, i.e., fluoxetine (Flx)] is the first line of treatment for MDD. Despite its efficacy, lack of AD response occurs in numerous patients characterizing Difficult-to-treat Depression. ElectroConvulsive Therapy (ECT) is a highly effective treatment inducing rapid improvement in depressive symptoms and high remission rates of ∼50–63% in patients with pharmaco-resistant depression. Nevertheless, the need to develop reliable treatment response predictors to guide personalized AD strategies and supplement clinical observation is becoming a pressing clinical objective. Here, we propose to establish a proteomic peripheral biomarkers signature of ECT response in an anxio/depressive animal model of non-response to AD. Using an emotionality score based on the analysis complementary behavioral tests of anxiety/depression (Elevated Plus Maze, Novelty Suppressed Feeding, Splash Test), we showed that a 4-week corticosterone treatment (35 μg/ml, Cort model) in C57BL/6JRj male mice induced an anxiety/depressive-like behavior. A 28-day chronic fluoxetine treatment (Flx, 18 mg/kg/day) reduced corticosterone-induced increase in emotional behavior. A 50% decrease in emotionality score threshold before and after Flx, was used to separate Flx-responding mice (Flx-R, n = 18), or Flx non-responder mice (Flx-NR, n = 7). Then, Flx-NR mice received seven sessions of electroconvulsive seizure (ECS, equivalent to ECT in humans) and blood was collected before and after ECS treatment. Chronic ECS normalized the elevated emotionality observed in Flx-NR mice. Then, proteins were extracted from peripheral blood mononuclear cells (PBMCs) and isolated for proteomic analysis using a high-resolution MS Orbitrap. Data are available via ProteomeXchange with identifier PXD037392. The proteomic analysis revealed a signature of 33 peripheral proteins associated with response to ECS (7 down and 26 upregulated). These proteins were previously associated with mental disorders and involved in regulating pathways which participate to the depressive disorder etiology.

Genomic-Analysis-Oriented Drug Repurposing in the Search for Novel Antidepressants

From inadequate prior antidepressants that targeted monoamine neurotransmitter systems emerged the discovery of alternative drugs for depression. For instance, drugs targeted interleukin 6 receptor (IL6R) in inflammatory system. Genomic analysis-based drug repurposing using single nucleotide polymorphism (SNP) inclined a promising method for several diseases. However, none of the diseases was depression. Thus, we aimed to identify drug repurposing candidates for depression treatment by adopting a genomic-analysis-based approach. The 5885 SNPs obtained from the machine learning approach were annotated using HaploReg v4.1. Five sets of functional annotations were applied to determine the depression risk genes. The STRING database was used to expand the target genes and identify drug candidates from the DrugBank database. We validated the findings using the ClinicalTrial.gov and PubMed databases. Seven genes were observed to be strongly associated with depression (functional annotation score = 4). Interestingly, IL6R was auspicious as a target gene according to the validation outcome. We identified 20 drugs that were undergoing preclinical studies or clinical trials for depression. In addition, we identified sarilumab and satralizumab as drugs that exhibit strong potential for use in the treatment of depression. Our findings indicate that a genomic-analysis-based approach can facilitate the discovery of drugs that can be repurposed for treating depression.

Cellular and immunometabolic mechanisms of inflammation in depression: Preliminary findings from single cell RNA sequencing and a tribute to Bruce McEwen

Inflammation is associated with symptoms of anhedonia, a core feature of major depression (MD). We have shown that MD patients with high inflammation as measured by plasma C-reactive protein (CRP) and anhedonia display gene signatures of metabolic reprograming (e.g., shift to glycolysis) necessary to sustain cellular immune activation. To gain preliminary insight into the immune cell subsets and transcriptomic signatures that underlie increased inflammation and its relationship with behavior in MD at the single-cell (sc) level, herein we conducted scRNA-Seq on peripheral blood mononuclear cells from a subset of medically-stable, unmedicated MD outpatients. Three MD patients with high CRP (>3 mg/L) before and two weeks after anti-inflammatory challenge with the tumor necrosis factor antagonist infliximab and three patients with low CRP (≤3 mg/L) were studied. Cell clusters were identified using a Single Cell Wizard pipeline, followed by pathway analysis. CD14⁺ and CD16⁺ monocytes were more abundant in MD patients with high CRP and were reduced by 29% and 55% respectively after infliximab treatment. Within CD14⁺ and CD16⁺ monocytes, genes upregulated in high CRP patients were enriched for inflammatory (phagocytosis, complement, leukocyte migration) and immunometabolic (hypoxia-inducible factor [HIF]-1, aerobic glycolysis) pathways. Shifts in CD4⁺ T cell subsets included ∼30% and ∼10% lower abundance of CD4⁺ central memory (T_CM) and naïve cells and ∼50% increase in effector memory-like (T_EM-like) cells in high versus low CRP patients. T_CM cells of high CRP patients displayed downregulation of the oxidative phosphorylation (OXPHOS) pathway, a main energy source in this cell type. Following infliximab, changes in the number of CD14⁺ monocytes and CD4⁺ T_EM-like cells predicted improvements in anhedonia scores (r = 1.0, p < 0.001). In sum, monocytes and CD4⁺ T cells from MD patients with increased inflammation exhibited immunometabolic reprograming in association with symptoms of anhedonia. These findings are the first step toward determining the cellular and molecular immune pathways associated with inflammatory phenotypes in MD, which may lead to novel immunomodulatory treatments of psychiatric illnesses with increased inflammation.

Transcriptional biomarkers of response to pharmacological treatments in severe mental disorders: A systematic review

Variation in the expression level and activity of genes involved in drug disposition and action in tissues of pharmacological importance have been increasingly investigated in patients treated with psychotropic drugs. Findings are promising, but reliable predictive biomarkers of response have yet to be identified. Here we conducted a PRISMA-compliant systematic search of PubMed, Scopus and PsycInfo up to 12 September 2020 for studies investigating RNA expression levels in cells or biofluids from patients with major depressive disorder, schizophrenia or bipolar disorder characterized for response to psychotropic drugs (antidepressants, antipsychotics or mood stabilizers) or adverse effects. Among 5497 retrieved studies, 123 (63 on antidepressants, 33 on antipsychotics and 27 on mood stabilizers) met inclusion criteria. Studies were either focused on mRNAs (n = 96), microRNAs (n = 19) or long non-coding RNAs (n = 1), with only a minority investigating both mRNAs and microRNAs levels (n = 7). The most replicated results include genes playing a role in inflammation (antidepressants), neurotransmission (antidepressants and antipsychotics) or mitochondrial function (mood stabilizers). Compared to those investigating response to antidepressants, studies focused on antipsychotics or mood stabilizers more often showed lower sample size and lacked replication. Strengths and limitations of available studies are presented and discussed in light of the specific designs, methodology and clinical characterization of included patients for transcriptomic compared to DNA-based studies. Finally, future directions of transcriptomics of psychopharmacological interventions in psychiatric disorders are discussed.

Whole blood transcriptional signatures associated with rapid antidepressant response to ketamine in patients with treatment resistant depression

Ketamine has rapid and sustained antidepressant effects in patients with treatment-resistant depression (TRD). However, the underlying mechanisms of action are not well understood. There is increasing evidence that TRD is associated with a pro-inflammatory state and that ketamine may inhibit inflammatory processes. We thus investigated whole blood transcriptional profiles related to TRD and gene expression changes associated with treatment response to ketamine. Whole blood was collected at baseline (21 healthy controls [HC], 26 patients with TRD) and then again in patients with TRD 24 hours following a single intravenous infusion of ketamine (0.5 mg/kg). We performed RNA-sequencing and analyzed (a) baseline transcriptional profiles between patients with TRD and HC, (b) responders vs. non-responders before ketamine treatment, and (c) gene expression signatures associated with clinical improvement. At baseline, patients with TRD compared to HC showed a gene expression signature indicative of interferon signaling pathway activation. Prior to ketamine administration, the metabotropic glutamate receptor gene GRM2 and the ionotropic glutamate receptor gene GRIN2D were upregulated in responders compared to non-responders. Response to ketamine was associated with a distinct transcriptional signature, however, we did not observe gene expression changes indicative of an anti-inflammatory effect. Future studies are needed to determine the role of the peripheral immune system in the antidepressant effect of ketamine.

Large-Scale Transcriptomics Studies Provide Insight Into Sex Differences in Depression

Major depressive disorder (MDD) is a leading cause of disability, affecting more than 300 million people worldwide. We first review the well-known sex difference in incidence of MDD, with women being twice as likely to be diagnosed as men, and briefly summarize how the impact of MDD varies between men and women, with sex differences in symptoms, severity, and antidepressant drug response. We then attempt to deconstruct the biological bases for MDD and discuss implications for sex differences research. Next, we review findings from human postmortem studies, both from selected candidate gene studies and from well-powered, unbiased transcriptomics studies, which suggest distinct, and possibly opposite, molecular changes in the brains of depressed men and women. We then discuss inherent challenges of research on the human postmortem brain and suggest paths forward that rely on thoughtful cohort design. Although studies indicate that circulating gonadal hormones might underlie the observed sex differences in MDD, we discuss how additional sex-specific factors, such as genetic sex and developmental exposure to gonadal hormones, may also contribute to altered vulnerability, and we highlight various nuances that we believe should be considered when determining mechanisms underlying observed sex differences. Altogether, this review highlights not only how various sex-specific factors might influence susceptibility or resilience to depression, but also how those sex-specific factors might result in divergent pathology in men and women.

Machine learning in the prediction of depression treatment outcomes: a systematic review and meta-analysis

BACKGROUND

Multiple treatments are effective for major depressive disorder (MDD), but the outcomes of each treatment vary broadly among individuals. Accurate prediction of outcomes is needed to help select a treatment that is likely to work for a given person. We aim to examine the performance of machine learning methods in delivering replicable predictions of treatment outcomes.

METHODS

Of 7732 non-duplicate records identified through literature search, we retained 59 eligible reports and extracted data on sample, treatment, predictors, machine learning method, and treatment outcome prediction. A minimum sample size of 100 and an adequate validation method were used to identify adequate-quality studies. The effects of study features on prediction accuracy were tested with mixed-effects models. Fifty-four of the studies provided accuracy estimates or other estimates that allowed calculation of balanced accuracy of predicting outcomes of treatment.

RESULTS

Eight adequate-quality studies reported a mean accuracy of 0.63 [95% confidence interval (CI) 0.56-0.71], which was significantly lower than a mean accuracy of 0.75 (95% CI 0.72-0.78) in the other 46 studies. Among the adequate-quality studies, accuracies were higher when predicting treatment resistance (0.69) and lower when predicting remission (0.60) or response (0.56). The choice of machine learning method, feature selection, and the ratio of features to individuals were not associated with reported accuracy.

CONCLUSIONS

The negative relationship between study quality and prediction accuracy, combined with a lack of independent replication, invites caution when evaluating the potential of machine learning applications for personalizing the treatment of depression.

Transcriptomic signatures of psychomotor slowing in peripheral blood of depressed patients: evidence for immunometabolic reprogramming

Inflammation impacts basal ganglia motor circuitry in association with psychomotor retardation, a key symptom of major depression (MD). We previously reported associations between circulating protein inflammatory biomarkers and psychomotor slowing as measured by neuropsychological tests probing psychomotor speed in patients with MD. To discover novel transcriptional signatures in peripheral blood immune cells related to psychomotor slowing, microarray data were analyzed in a primary cohort of 88 medically-stable, unmedicated, ambulatory MD patients. Results were confirmed and extended in a second cohort of 57 patients with treatment resistant depression (TRD) before and after anti-inflammatory challenge with the tumor necrosis factor antagonist infliximab versus placebo. Composite scores reflecting pure motor and cognitive-motor processing speed were linearly associated with 403 and 266 gene transcripts in each cohort, respectively (|R| > 0.30, p < 0.01), that were enriched for cytokine signaling and glycolysis-related pathways (p < 0.05). Unsupervised clustering in the primary cohort revealed two psychomotor slowing-associated gene co-expression modules that were enriched for interferon, interleukin-6, aerobic glycolysis, and oxidative phosphorylation pathways (p < 0.05, q < 0.1). Transcripts were predominantly derived from monocytes, plasmacytoid dendritic cells, and natural killer cells (p's < 0.05). In infliximab-treated TRD patients with high plasma C-reactive protein concentrations (>5 mg/L), two differential co-expression modules enriched for oxidative stress and mitochondrial degradation were associated with improvements in psychomotor reaction time (p < 0.05). These results indicate that inflammatory signaling and associated metabolic reprogramming in peripheral blood immune cells are associated with systemic inflammation in depression and may affect relevant brain circuits to promote psychomotor slowing.

Effective Antidepressant Chronotherapeutics (Sleep Deprivation and Light Therapy) Normalize the IL-1β:IL-1ra Ratio in Bipolar Depression

Background

Mood disorders associate with peripheral markers of low-grade inflammation, among which circulating levels of interleukin-1β (IL-1β) consistently predict diagnosis and poor outcomes. Antidepressant chronotherapeutics (total sleep deprivation and light therapy, TSD+LT) prompts response in drug-resistant bipolar depression, but its effect on peripheral inflammation were never assessed. Here we explored the effects of TSD+LT on IL-1β signaling.

Methods

We studied the ratio between IL-1β and its receptor antagonist (IL-1β:IL1ra) in 33 healthy participants, and in 26 inpatients with a major depressive episode in course of Bipolar Disorder, before and after treatment with three cycles of repeated TSD+LT, interspersed with sleep recovery nights, administered during 1 week. Treatment effects of mood and on IL-1β:IL1ra were analyzed in the context of the Generalized Linear Model (GLM).

Results

At baseline, patients had higher IL-1β, IL1ra, and IL-1β:IL1ra than controls. Treatment significantly decreased IL-1β:IL1ra, by decreasing IL-1β and increasing IL1ra, the effect being proportional to baseline levels and normalizing values. Patients with higher baseline levels showed the highest decrease in IL-1β:IL-1ra, which associated with the immediate antidepressant response at the first cycle; while patients with lower baseline values showed negligible changes in the IL-1β:IL-1ra, unrelated to treatment response.

Conclusion

We observed a parallel change of inflammatory biomarkers and severity of depression after chronotherapeutics, suggesting that a reduction in inflammation associated with depression could contribute to the mechanism of action of TSD+LT, and warranting interest for controlled studies addressing the role of inflammation in the recovery from bipolar depression.

Changes in RNA expression levels during antidepressant treatment: a systematic review

More than a third of patients treated with antidepressants experience treatment resistance. Furthermore, molecular pathways involved in antidepressant effect have yet to be fully understood. Therefore, we performed a systematic review of clinical studies that examined changes in RNA expression levels produced by antidepressant treatment. Literature search was performed through April 2021 for peer-reviewed studies measuring changes in mRNA or non-coding RNA levels before and after antidepressant treatment in human participants following PRISMA guidelines. Thirty-one studies were included in qualitative synthesis. We identified a large amount of heterogeneity between the studies for genes/RNAs measured, antidepressants used, and treatment duration. Of the six RNAs examined by more than one study, expression of the brain-derived neurotrophic factor (BDNF) gene and genes in the inflammation pathway, particularly IL-1β, were consistently reported to be altered by antidepressant treatment. Limitations of this review include heterogeneity of the studies, possibility of positive publication bias, and risk of false-negative findings secondary to small sample sizes. In conclusion, our systematic review provides an updated synthesis of RNA expression changes produced by antidepressant treatment in human participants, where genes in the BDNF and inflammatory pathways were identified as potential targets of antidepressant effect. Importantly, these findings also highlight the need for replication of the included studies in multiple strong, placebo-controlled studies for the identification of evidence-based markers that can be targeted to improve treatment outcomes.

A large-scale genome-wide gene expression analysis in peripheral blood identifies very few differentially expressed genes related to antidepressant treatment and response in patients with major depressive disorder

A better understanding of the biological factors underlying antidepressant treatment in patients with major depressive disorder (MDD) is needed. We perform gene expression analyses and explore sources of variability in peripheral blood related to antidepressant treatment and treatment response in patients suffering from recurrent MDD at baseline and after 8 weeks of treatment. The study includes 281 patients, which were randomized to 8 weeks of treatment with vortioxetine (N = 184) or placebo (N = 97). To our knowledge, this is the largest dataset including both gene expression in blood and placebo-controlled treatment response measured by a clinical scale in a randomized clinical trial. We identified three novel genes whose RNA expression levels at baseline and week 8 are significantly (FDR < 0.05) associated with treatment response after 8 weeks of treatment. Among these genes were SOCS3 (FDR = 0.0039) and PROK2 (FDR = 0.0028), which have previously both been linked to depression. Downregulation of these genes was associated with poorer treatment response. We did not identify any genes that were differentially expressed between placebo and vortioxetine groups at week 8 or between baseline and week 8 of treatment. Nor did we replicate any genes identified in previous peripheral blood gene expression studies examining treatment response. Analysis of genome-wide expression variability showed that type of treatment and treatment response explains very little of the variance, a median of <0.0001% and 0.05% in gene expression across all genes, respectively. Given the relatively large size of the study, the limited findings suggest that peripheral blood gene expression might not be the best approach to explore the biological factors underlying antidepressant treatment.

Machine learning to predict clinical remission in depressed patients after acute phase selective serotonin reuptake inhibitor treatment

OBJECTIVE

Selective serotonin reuptake inhibitors (SSRIs) are suggested as the first-line treatment for patients with major depressive disorder (MDD), but the remission rate is unsatisfactory. We aimed to establish machine learning models and explore variables available at baseline to predict the 8-week outcome among patients taking SSRIs.

METHODS

Data from 400 patients were used to build machine learnings. The last observation carried forward approach was used to determine the remitter/non-remitter status of the patients at week 8. Using least absolute shrinkage and selection operator (LASSO) to select features, we built 4 different machine learning algorithms including gradient boosting decision tree, support vector machine (SVM), random forests, and logistic regression with five-fold cross-validation. Then, we adopted Shapley additive explanations (SHAP) values to interpret the model output.

RESULTS

The remission rate is 67.8%. We obtained 78 features from the baseline characteristics, including 25 sociodemographic characteristics, 31 clinical features, 15 psychological traits and 7 neurocognitive functions, and 13 of these features were selected to establish SVM. The accuracy of the SVM prediction is 74.49%, reaching an average area under the curve of 0.734±0.043. The sensitivity is 0.899±0.038 with a positive predictive value of 0.776±0.028. The specificity is 0.422±0.091 with a negative predictive value of 0.674±0.086. According to the SHAP values, neurocognitive functions and anxiety and hypochondriasis symptoms were important predictors.

CONCLUSION

Our study supports the utilization of machine learning approaches with inexpensive and highly accessible variables to accurately predict the 8-week treatment outcome of SSRIs in patients with MDD.

Machine learning and bioinformatic analysis of brain and blood mRNA profiles in major depressive disorder: A case-control study

This study analyzed gene expression messenger RNA data, from cases with major depressive disorder (MDD) and controls, using supervised machine learning (ML). We built on the methodology of prior studies to obtain more generalizable/reproducible results. First, we obtained a classifier trained on gene expression data from the dorsolateral prefrontal cortex of post-mortem MDD cases (n = 126) and controls (n = 103). An average area-under-the-receiver-operating-characteristics-curve (AUC) from 10-fold cross-validation of 0.72 was noted, compared to an average AUC of 0.55 for a baseline classifier (p = .0048). The classifier achieved an AUC of 0.76 on a previously unused testing-set. We also performed external validation using DLPFC gene expression values from an independent cohort of matched MDD cases (n = 29) and controls (n = 29), obtained from Affymetrix microarray (vs. Illumina microarray for the original cohort) (AUC: 0.62). We highlighted gene sets differentially expressed in MDD that were enriched for genes identified by the ML algorithm. Next, we assessed the ML classification performance in blood-based microarray gene expression data from MDD cases (n = 1,581) and controls (n = 369). We observed a mean AUC of 0.64 on 10-fold cross-validation, which was significantly above baseline (p = .0020). Similar performance was observed on the testing-set (AUC: 0.61). Finally, we analyzed the classification performance in covariates subgroups. We identified an interesting interaction between smoking and recall performance in MDD case prediction (58% accurate predictions in cases who are smokers vs. 43% accurate predictions in cases who are non-smokers). Overall, our results suggest that ML in combination with gene expression data and covariates could further our understanding of the pathophysiology in MDD.

Higher baseline interleukin-1β and TNF-α hamper antidepressant response in major depressive disorder

Raised pro-inflammatory immune/inflammatory setpoints, leading to an increased production of peripheral cytokines, have been associated with Major Depressive Disorder (MDD) and with failure to respond to first-line antidepressant drugs. However, the usefulness of these biomarkers in clinical psychopharmacology has been questioned because single findings did not translate into the clinical practice, where patients are prescribed treatments upon clinical need. We studied a panel of 27 inflammatory biomarkers in a sample of 108 inpatients with MDD, treated with antidepressant monotherapy for 4 weeks upon clinical need in a specialized hospital setting, and assessed the predictive effect of baseline peripheral measures of inflammation on antidepressing efficacy (response rates and time-lagged pattern of decrease of depression severity) using a machine-learning approach with elastic net penalized regression, and multivariate analyses in the context of the general linear model. When considering both categorical and continuous measures of response, baseline levels of IL-1β predicted non-response to antidepressants, with the predicted probability to respond being highly dispersed at low levels of IL-1β, and stratifying toward non-response when IL-1β is high. Significant negative effects were also detected for TNF-α, while IL-12 weakly predicted response. These findings support the usefulness of inflammatory biomarkers in the clinical psychopharmacology of depression, and add to ongoing research efforts aiming at defining reliable cutoff values to identify depressed patients in clinical settings with high inflammation, and low probability to respond.

Major Depressive Disorder Is Associated With Differential Expression of Innate Immune and Neutrophil-Related Gene Networks in Peripheral Blood: A Quantitative Review of Whole-Genome Transcriptional Data From Case-Control Studies

BACKGROUND

Whole-genome transcription has been measured in peripheral blood samples as a candidate biomarker of inflammation associated with major depressive disorder.

METHODS

We searched for all case-control studies on major depressive disorder that reported microarray or RNA sequencing measurements on whole blood or peripheral blood mononuclear cells. Primary datasets were reanalyzed, when openly accessible, to estimate case-control differences and to evaluate the functional roles of differentially expressed gene lists by technically harmonized methods.

RESULTS

We found 10 eligible studies (N = 1754 depressed cases and N = 1145 healthy controls). Fifty-two genes were called significant by 2 of the primary studies (published overlap list). After harmonization of analysis across 8 accessible datasets (n = 1706 cases, n = 1098 controls), 272 genes were coincidentally listed in the top 3% most differentially expressed genes in 2 or more studies of whole blood or peripheral blood mononuclear cells with concordant direction of effect (harmonized overlap list). By meta-analysis of standardized mean difference across 4 studies of whole-blood samples (n = 1567 cases, n = 954 controls), 343 genes were found with false discovery rate <5% (standardized mean difference meta-analysis list). These 3 lists intersected significantly. Genes abnormally expressed in major depressive disorder were enriched for innate immune-related functions, coded for nonrandom protein-protein interaction networks, and coexpressed in the normative transcriptome module specialized for innate immune and neutrophil functions.

CONCLUSIONS

Quantitative review of existing case-control data provided robust evidence for abnormal expression of gene networks important for the regulation and implementation of innate immune response. Further development of white blood cell transcriptional biomarkers for inflamed depression seems warranted.

Predicting second-generation antidepressant effectiveness in treating sadness using demographic and clinical information: A machine learning approach

INTRODUCTION

Current guidelines for choosing antidepressant medications involve a trial-and-error process. Most patients try multiple antidepressants before finding an effective antidepressant. This study uses demographic and clinical information to create models predicting effectiveness of different antidepressants in treating sadness in a nationally representative sample of US adults.

METHODS

A secondary analysis of the Collaborative Psychiatric Epidemiology Survey (CPES) was performed. Participants with or without a mental health diagnosis who reported sadness as a symptom, and were taking fluoxetine (n=156), sertraline (n=224), citalopram (n=91), paroxetine (n=156), venlafaxine (n=69), bupropion (n=92), or trazadone (n=26) within the past year were included. Two sets of principal component analyses (PCAs) and logistic regressions were performed: one determined associations between symptom clusters and antidepressant effectiveness for sadness, and the other created models to predict effectiveness. Both PCAs controlled for psychiatric and medical diagnoses, substance use, psychiatric medications, alternative treatments, and demographics.

RESULTS

Anxiety was associated with ineffectiveness of fluoxetine in treating sadness. Low mood scores were associated with ineffectiveness of paroxetine and venlafaxine, and fatigue was associated with ineffectiveness of sertraline. The models for predicting drug effectiveness had a mean accuracy of 83% and internal validity of 72%.

LIMITATIONS

CPES data were collected from 2001-2003, so newer drugs were not included. Effectiveness was for sadness, so results are not directly comparable to studies using overall depressive symptom reductions as outcomes.

CONCLUSION

Since fewer than 50% of patients currently respond to their first antidepressant, this model could provide modest improvement to choosing starting antidepressants in treating sadness.

Response to fluoxetine in children and adolescents: a weighted gene co-expression network analysis of peripheral blood

The inconclusive and non-replicated results of pharmacogenetic studies of antidepressant response could be related to the lack of acknowledgement of its mechanism of action. In this scenario, gene expression studies provide and interesting framework to reveal new candidate genes for pharmacogenetic studies or peripheral biomarkers of fluoxetine response. We propose a system biology approach to analyse changes in gene expression induced by eight weeks of treatment with fluoxetine in peripheral blood. 21 naïve child and adolescents participated in the present study. Our analysis include the identification of gene co-expression modules, using Weighted Gene Co-expression Network Analysis (WGCNA), followed by protein-protein interaction (PPi) network construction coupled with functional annotation. Our results revealed two modules of co-expression genes related to fluoxetine treatment. The constructed networks from these modules were enriched for biological processes related to cellular and metabolic processes, cell communication, immune system processes, cell death, response to stimulus and neurogenesis. Some of these processes, such as immune system, replicated previous findings in the literature, whereas, neurogenesis, a mechanism proposed to be involved in fluoxetine response, had been identified for first time using peripheral tissues. In conclusion, our study identifies several biological processes in relation to fluoxetine treatment in peripheral blood, offer new candidate genes for pharmacogenetic studies and valuable markers for peripheral moderator biomarkers discovery.

Predicting treatment effects in unipolar depression: A meta-review

There is increasing interest in clinical prediction models in psychiatry, which focus on developing multivariate algorithms to guide personalized diagnostic or management decisions. The main target of these models is the prediction of treatment response to different antidepressant therapies. This is because the ability to predict response based on patients' personal data may allow clinicians to make improved treatment decisions, and to provide more efficacious or more tolerable medications to the right patient. We searched the literature for systematic reviews about treatment prediction in the context of existing treatment modalities for adult unipolar depression, until July 2019. Treatment effect is defined broadly to include efficacy, safety, tolerability and acceptability outcomes. We first focused on the identification of individual predictor variables that might predict treatment response, and second, we considered multivariate clinical prediction models. Our meta-review included a total of 10 systematic reviews; seven (from 2014 to 2018) focusing on individual predictor variables and three focusing on clinical prediction models. These identified a number of sociodemographic, phenomenological, clinical, neuroimaging, remote monitoring, genetic and serum marker variables as possible predictor variables for treatment response, alongside statistical and machine-learning approaches to clinical prediction model development. Effect sizes for individual predictor variables were generally small and clinical prediction models had generally not been validated in external populations. There is a need for rigorous model validation in large external data-sets to prove the clinical utility of models. We also discuss potential future avenues in the field of personalized psychiatry, particularly the combination of multiple sources of data and the emerging field of artificial intelligence and digital mental health to identify new individual predictor variables.

Challenges and Future Prospects of Precision Medicine in Psychiatry

Precision medicine is increasingly recognized as a promising approach to improve disease treatment, taking into consideration the individual clinical and biological characteristics shared by specific subgroups of patients. In specific fields such as oncology and hematology, precision medicine has already started to be implemented in the clinical setting and molecular testing is routinely used to select treatments with higher efficacy and reduced adverse effects. The application of precision medicine in psychiatry is still in its early phases. However, there are already examples of predictive models based on clinical data or combinations of clinical, neuroimaging and biological data. While the power of single clinical predictors would remain inadequate if analyzed only with traditional statistical approaches, these predictors are now increasingly used to impute machine learning models that can have adequate accuracy even in the presence of relatively small sample size. These models have started to be applied to disentangle relevant clinical questions that could lead to a more effective management of psychiatric disorders, such as prediction of response to the mood stabilizer lithium, resistance to antidepressants in major depressive disorder or stratification of the risk and outcome prediction in schizophrenia. In this narrative review, we summarized the most important findings in precision medicine in psychiatry based on studies that constructed machine learning models using clinical, neuroimaging and/or biological data. Limitations and barriers to the implementation of precision psychiatry in the clinical setting, as well as possible solutions and future perspectives, will be presented.

Time Course of Changes in Peripheral Blood Gene Expression During Medication Treatment for Major Depressive Disorder

Changes in gene expression (GE) during antidepressant treatment may increase understanding of the action of antidepressant medications and serve as biomarkers of efficacy. GE changes in peripheral blood are desirable because they can be assessed easily on multiple occasions during treatment. We report here on GE changes in 68 individuals who were treated for 8 weeks with either escitalopram alone, or escitalopram followed by bupropion. GE changes were assessed after 1, 2, and 8 weeks of treatment, with significant changes observed in 156, 121, and 585 peripheral blood gene transcripts, respectively. Thirty-one transcript changes were shared between the 1- and 8-week time points (seven upregulated, 24 downregulated). Differences were detected between the escitalopram- and bupropion-treated subjects, although there was no significant association between GE changes and clinical outcome. A subset of 18 genes overlapped with those previously identified as differentially expressed in subjects with MDD compared with healthy control subjects. There was statistically significant overlap between genes differentially expressed in the current and previous studies, with 10 genes overlapping in at least two previous studies. There was no enrichment for genes overexpressed in nervous system cell types, but there was a trend toward enrichment for genes in the WNT/β-catenin pathway in the anterior thalamus; three genes in this pathway showed differential expression in the present and in three previous studies. Our dataset and other similar studies will provide an important source of information about potential biomarkers of recovery and for potential dysregulation of GE in MDD.

Epigenome-wide association study of depression symptomatology in elderly monozygotic twins

Depression is a severe and debilitating mental disorder diagnosed by evaluation of affective, cognitive and physical depression symptoms. Severity of these symptoms strongly impacts individual's quality of life and is influenced by a combination of genetic and environmental factors. One of the molecular mechanisms allowing for an interplay between these factors is DNA methylation, an epigenetic modification playing a pivotal role in regulation of brain functioning across lifespan. The aim of this study was to investigate if there are DNA methylation signatures associated with depression symptomatology in order to identify molecular mechanisms contributing to pathophysiology of depression. We performed an epigenome-wide association study (EWAS) of continuous depression symptomatology score measured in a cohort of 724 monozygotic Danish twins (346 males, 378 females). Through EWAS analyses adjusted for sex, age, flow-cytometry based blood cell composition, and twin relatedness structure in the data we identified depression symptomatology score to be associated with blood DNA methylation levels in promoter regions of neuropsin (KLK8, p-value = 4.7 × 10^-7) and DAZ associated protein 2 (DAZAP2, p-value = 3.13 × 10^-8) genes. Other top associated probes were located in gene bodies of MAD1L1 (p-value = 5.16 × 10^-6), SLC29A2 (p-value = 6.15 × 10^-6) and AKT1 (p-value = 4.47 × 10^-6), all genes associated before with development of depression. Additionally, the following three measures (a) DNAmAge (calculated with Horvath and Hannum epigenetic clock estimators) adjusted for chronological age, (b) difference between DNAmAge and chronological age, and (c) DNAmAge acceleration were not associated with depression symptomatology score in our cohort. In conclusion, our data suggests that depression symptomatology score is associated with DNA methylation levels of genes implicated in response to stress, depressive-like behaviors, and recurrent depression in patients, but not with global DNA methylation changes across the genome.

Machine learning in mental health: a scoping review of methods and applications

BACKGROUND

This paper aims to synthesise the literature on machine learning (ML) and big data applications for mental health, highlighting current research and applications in practice.

METHODS

We employed a scoping review methodology to rapidly map the field of ML in mental health. Eight health and information technology research databases were searched for papers covering this domain. Articles were assessed by two reviewers, and data were extracted on the article's mental health application, ML technique, data type, and study results. Articles were then synthesised via narrative review.

RESULTS

Three hundred papers focusing on the application of ML to mental health were identified. Four main application domains emerged in the literature, including: (i) detection and diagnosis; (ii) prognosis, treatment and support; (iii) public health, and; (iv) research and clinical administration. The most common mental health conditions addressed included depression, schizophrenia, and Alzheimer's disease. ML techniques used included support vector machines, decision trees, neural networks, latent Dirichlet allocation, and clustering.

CONCLUSIONS

Overall, the application of ML to mental health has demonstrated a range of benefits across the areas of diagnosis, treatment and support, research, and clinical administration. With the majority of studies identified focusing on the detection and diagnosis of mental health conditions, it is evident that there is significant room for the application of ML to other areas of psychology and mental health. The challenges of using ML techniques are discussed, as well as opportunities to improve and advance the field.

Low-Grade Inflammation as a Predictor of Antidepressant and Anti-Inflammatory Therapy Response in MDD Patients: A Systematic Review of the Literature in Combination With an Analysis of Experimental Data Collected in the EU-MOODINFLAME Consortium

Low-grade inflammation plays a role not only in the pathogenesis of major depressive disorder (MDD) but probably also in the poor responsiveness to regular antidepressants. There are also indications that anti-inflammatory agents improve the outcomes of antidepressants. Aim: To study whether the presence of low-grade inflammation predicts the outcome of antidepressants, anti-inflammatory agents, or combinations thereof. Methods: We carried out a systematic review of the literature on the prediction capability of the serum levels of inflammatory compounds and/or the inflammatory state of circulating leukocytes for the outcome of antidepressant/anti-inflammatory treatment in MDD. We compared outcomes of the review with original data (collected in two limited trials carried out in the EU project MOODINFLAME) on the prediction capability of the inflammatory state of monocytes (as measured by inflammatory gene expression) for the outcome of venlafaxine, imipramine, or sertraline treatment, the latter with and without celecoxib added. Results: Collectively, the literature and original data showed that: 1) raised serum levels of pro-inflammatory compounds (in particular of CRP/IL-6) characterize an inflammatory form of MDD with poor responsiveness to predominately serotonergic agents, but a better responsiveness to antidepressant regimens with a) (add-on) noradrenergic, dopaminergic, or glutamatergic action or b) (add-on) anti-inflammatory agents such as infliximab, minocycline, or eicosapentaenoic acid, showing-next to anti-inflammatory-dopaminergic or lipid corrective action; 2) these successful anti-inflammatory (add-on) agents, when used in patients with low serum levels of CRP/IL-6, decreased response rates in comparison to placebo. Add-on aspirin, in contrast, improved responsiveness in such "non-inflammatory" patients; 3) patients with increased inflammatory gene expression in circulating leukocytes had a poor responsiveness to serotonergic/noradrenergic agents. Conclusions: The presence of inflammation in patients with MDD heralds a poor outcome of first-line antidepressant therapies. Immediate step-ups to dopaminergic or glutamatergic regimens or to (add-on) anti-inflammatory agents are most likely indicated. However, at present, insufficient data exist to design protocols with reliable inflammation parameter cutoff points to guide such therapies, the more since detrimental outcomes are possible of anti-inflammatory agents in "non-inflamed" patients.

Applications of machine learning algorithms to predict therapeutic outcomes in depression: A meta-analysis and systematic review

BACKGROUND

No previous study has comprehensively reviewed the application of machine learning algorithms in mood disorders populations. Herein, we qualitatively and quantitatively evaluate previous studies of machine learning-devised models that predict therapeutic outcomes in mood disorders populations.

METHODS

We searched Ovid MEDLINE/PubMed from inception to February 8, 2018 for relevant studies that included adults with bipolar or unipolar depression; assessed therapeutic outcomes with a pharmacological, neuromodulatory, or manual-based psychotherapeutic intervention for depression; applied a machine learning algorithm; and reported predictors of therapeutic response. A random-effects meta-analysis of proportions and meta-regression analyses were conducted.

RESULTS

We identified 639 records: 75 full-text publications were assessed for eligibility; 26 studies (n=17,499) and 20 studies (n=6325) were included in qualitative and quantitative review, respectively. Classification algorithms were able to predict therapeutic outcomes with an overall accuracy of 0.82 (95% confidence interval [CI] of [0.77, 0.87]). Pooled estimates of classification accuracy were significantly greater (p < 0.01) in models informed by multiple data types (e.g., composite of phenomenological patient features and neuroimaging or peripheral gene expression data; pooled proportion [95% CI] = 0.93[0.86, 0.97]) when compared to models with lower-dimension data types (pooledproportion=0.68[0.62,0.74]to0.85[0.81,0.88]).

LIMITATIONS

Most studies were retrospective; differences in machine learning algorithms and their implementation (e.g., cross-validation, hyperparameter tuning); cannot infer importance of individual variables fed into learning algorithm.

CONCLUSIONS

Machine learning algorithms provide a powerful conceptual and analytic framework capable of integrating multiple data types and sources. An integrative approach may more effectively model neurobiological components as functional modules of pathophysiology embedded within the complex, social dynamics that influence the phenomenology of mental disorders.

C-reactive protein and response to lurasidone in patients with bipolar depression

Prior studies suggest that the inflammatory biomarker c-reactive protein (CRP) holds promise for predicting antidepressant response in patients with major depressive disorder. The objective of this study was to evaluate whether CRP might similarly predict antidepressant responses to lurasidone in patients with bipolar I depression. Serum CRP concentration was measured prior to, and following, 6 weeks of treatment in 485 outpatients with bipolar I depression. Patients were randomized to receive monotherapy with lurasidone 20-60 mg/day (N = 161), lurasidone 80-120 mg/day (N = 162) or placebo (N = 162). CRP was assessed using the wide-range CRP assay (wr-CRP). The primary efficacy endpoint was change from baseline to week 6 in Montgomery-Åsberg Depression Rating Scale (MADRS) score. Mixed models and statistical interaction tests were applied to investigate the moderating effects of pre-treatment wr-CRP on clinical endpoints. CRP was evaluated as a log-transformed continuous variable and by clinically-relevant cut-points. Increasing pre-treatment wr-CRP level predicted a larger overall antidepressant response to lurasidone, as well as an increased response for a number of individual depressive symptoms. These moderating effects of pre-treatment wr-CRP remained significant after adjustment for potential confounds (e.g. baseline BMI and weight change). Treatment with lurasidone did not affect serum concentrations of CRP compared to placebo during the study. Elevated CRP level prior to treatment was associated with an enhanced clinical response to lurasidone in patients with bipolar I depression. If confirmed in future studies, CRP may represent a clinically useful diagnostic and predictive biomarker supporting a precision medicine approach to the treatment of bipolar depression.

Differentially expressed genes related to major depressive disorder and antidepressant response: genome-wide gene expression analysis

Treatment response to antidepressants is limited and varies among patients with major depressive disorder (MDD). To discover genes and mechanisms related to the pathophysiology of MDD and antidepressant treatment response, we performed gene expression analyses using peripheral blood specimens from 38 MDD patients and 14 healthy individuals at baseline and at 6 weeks after the initiation of either selective serotonin reuptake inhibitor (SSRI) or mirtazapine treatment. The results were compared with results from public microarray data. Seven differentially expressed genes (DEGs) between MDD patients and controls were identified in our study and in the public microarray data: CD58, CXCL8, EGF, TARP, TNFSF4, ZNF583, and ZNF587. CXCL8 was among the top 10 downregulated genes in both studies. Eight genes related to SSRI responsiveness, including BTNL8, showed alterations in gene expression in MDD. The expression of the FCRL6 gene differed between SSRI responders and nonresponders and changed after SSRI treatment compared to baseline. In evaluating the response to mirtazapine, 21 DEGs were identified when comparing MDD patients and controls and responders and nonresponders. These findings suggest that the pathophysiology of MDD and treatment response to antidepressants are associated with a number of processes, including DNA damage and apoptosis, that can be induced by immune activation and inflammation.

Progress in Elucidating Biomarkers of Antidepressant Pharmacological Treatment Response: A Systematic Review and Meta-analysis of the Last 15 Years

BACKGROUND

Antidepressant drugs are widely prescribed, but response rates after 3 months are only around one-third, explaining the importance of the search of objectively measurable markers predicting positive treatment response. These markers are being developed in different fields, with different techniques, sample sizes, costs, and efficiency. It is therefore difficult to know which ones are the most promising.

OBJECTIVE

Our purpose was to compute comparable (i.e., standardized) effect sizes, at study level but also at marker level, in order to conclude on the efficacy of each technique used and all analyzed markers.

METHODS

We conducted a systematic search on the PubMed database to gather all articles published since 2000 using objectively measurable markers to predict antidepressant response from five domains, namely cognition, electrophysiology, imaging, genetics, and transcriptomics/proteomics/epigenetics. A manual screening of the abstracts and the reference lists of these articles completed the search process.

RESULTS

Executive functioning, theta activity in the rostral Anterior Cingular Cortex (rACC), and polysomnographic sleep measures could be considered as belonging to the best objectively measured markers, with a combined d around 1 and at least four positive studies. For inter-category comparisons, the approaches that showed the highest effect sizes are, in descending order, imaging (combined d between 0.703 and 1.353), electrophysiology (0.294-1.138), cognition (0.929-1.022), proteins/nucleotides (0.520-1.18), and genetics (0.021-0.515).

CONCLUSION

Markers of antidepressant treatment outcome are numerous, but with a discrepant level of accuracy. Many biomarkers and cognitions have sufficient predictive value (d ≥ 1) to be potentially useful for clinicians to predict outcome and personalize antidepressant treatment.

Comparison of familial and non-familial suicidal behaviors among people with major depressive disorder: Testing the discriminative predicting role of high-yield clinical variables

BACKGROUND

Suicidal behavior in first-degree relatives of people diagnosed with major depressive disorder (MDD) increases the risk of suicidal behavior. Such an effect may be the result of genetic risk factors or environmental ones, including imitation, or both. Surprisingly few studies have examined this question and thus, there still is little known about the effect of first-degree family history of suicidal behavior on the type of suicidal behavior and profile of risk factors related to the diathesis for suicidal behavior. Even less is known about intra-familial risk transmission.

METHODS

Patients with MDD (n = 252) experiencing a current major depressive episode and who had a previous suicide attempt were studied. Those with and without a family history of first-degree relatives who had made a suicide attempt or died by suicide were compared across clinical and suicide-related characteristics.

RESULTS

Suicide attempters with (FDR⁺, n = 59) and without a first-degree relative with suicide attempt or suicide (FDR^-, n = 193) were similar in terms of type or frequency of suicide attempts, level of lifetime aggression and impulsivity, age of onset of depression and age at first suicide attempt.

LIMITATIONS

Cross-Sectional study. Lack of additional external validators.

CONCLUSIONS

Contrary to our hypothesis and the concept of "genetic anticipation", a first-degree family history of suicide attempt or suicide in currently depressed attempters with MDD was not associated with a range of clinical and suicide-related characteristics. Longitudinal studies incorporating external validators and potential biological markers may advance this area of research.

Transcriptomic and epigenomic biomarkers of antidepressant response

BACKGROUND

Antidepressant treatment is associated with a high rate of poor response, and thus, biomarker development is warranted.

METHODS

We aimed to synthesize studies investigating gene expression, small RNAs, and epigenomic biomarkers of antidepressant response. We conducted a narrative review of the literature.

RESULTS

Firstly, we detailed the challenges involved, in terms of biological tissues, relevant study time frames, and mandatory statistical tools. Secondly we synthesized results obtained in gene expression studies, focusing mainly on genome-wide studies, particularly small non-coding RNA, including micro-RNA and other small RNA species. In addition, we reviewed the potential biomarkers of antidepressant response arising from studies investigating DNA methylation variation and histone modifications.

LIMITATIONS

We did not conduct a meta-analysis due to the heterogeneity of the study.

CONCLUSION

Although promising, the field of gene expression and epigenomic biomarkers of antidepressant response is still in its infancy, and needs further development to define useful biomarkers in clinical practice.

Convergence and divergence of neurocognitive patterns in schizophrenia and depression

BACKGROUND

Neurocognitive impairments are frequently observed in schizophrenia and major depressive disorder (MDD). However, it remains unclear whether reported neurocognitive abnormalities could objectively identify an individual as having schizophrenia or MDD.

METHODS

The current study included 220 first-episode patients with schizophrenia, 110 patients with MDD and 240 demographically matched healthy controls (HC). All participants performed the short version of the Wechsler Adult Intelligence Scale-Revised in China; the immediate and delayed logical memory of the Wechsler Memory Scale-Revised in China; and seven tests from the computerized Cambridge Neurocognitive Test Automated Battery to evaluate neurocognitive performance. The three-class AdaBoost tree-based ensemble algorithm was employed to identify neurocognitive endophenotypes that may distinguish between subjects in the categories of schizophrenia, depression and HC. Hierarchical cluster analysis was applied to further explore the neurocognitive patterns in each group.

RESULTS

The AdaBoost algorithm identified individual's diagnostic class with an average accuracy of 77.73% (80.81% for schizophrenia, 53.49% for depression and 86.21% for HC). The average area under ROC curve was 0.92 (0.96 in schizophrenia, 0.86 in depression and 0.92 in HC). Hierarchical cluster analysis revealed for MDD and schizophrenia, convergent altered neurocognition patterns related to shifting, sustained attention, planning, working memory and visual memory. Divergent neurocognition patterns for MDD and schizophrenia related to motor speed, general intelligence, perceptual sensitivity and reversal learning were identified.

CONCLUSIONS

Neurocognitive abnormalities could predict whether the individual has schizophrenia, depression or neither with relatively high accuracy. Additionally, the neurocognitive features showed promise as endophenotypes for discriminating between schizophrenia and depression.

Pharmacogenomics in the treatment of mood disorders: Strategies and Opportunities for personalized psychiatry

Personalized medicine (personalized psychiatry in a specific setting) is a new model towards individualized care, in which knowledge from genomics and other omic pillars (microbiome, epigenomes, proteome, and metabolome) will be combined with clinical data to guide efforts to new drug development and targeted prescription of the existing treatment options. In this review, we summarize pharmacogenomic studies in mood disorders that may lay the foundation towards personalized psychiatry. In addition, we have discussed the possible strategies to integrate data from omic pillars as a future path to personalized psychiatry. So far, the progress of uncovering single nucleotide polymorphisms (SNPs) underpinning treatment efficacy in mood disorders (e.g., SNPs associated with selective serotonin re-uptake inhibitors or lithium treatment response in patients with bipolar disorder and major depressive disorder) are encouraging, but not adequate. Genetic studies have pointed to a number of SNPs located at candidate genes that possibly influence response to; (a) antidepressants COMT, HTR2A, HTR1A, CNR1, SLC6A4, NPY, MAOA, IL1B, GRIK4, BDNF, GNB3, FKBP5, CYP2D6, CYP2C19, and ABCB1 and (b) mood stabilizers (lithium) 5-HTT, TPH, DRD1, FYN, INPP1, CREB1, BDNF, GSK3β, ARNTL, TIM, DPB, NR3C1, BCR, XBP1, and CACNG2. We suggest three alternative and complementary strategies to implement knowledge gained from pharmacogenomic studies. The first strategy can be to implement diagnostic, therapeutic, or prognostic genetic testing based on candidate genes or gene products. The second alternative is an integrative analysis (systems genomics approach) to combine omics data obtained from the different pillars of omics investigation, including genomics, epigenomes, proteomics, metabolomics and microbiomes. The main goal of system genomics is an identification and understanding of biological pathways, networks, and modules underlying drug-response. The third strategy aims to the development of multivariable diagnostic or prognostic algorithms (tools) combining individual's genomic information (polygenic score) with other predictors (e.g., omics pillars, neuroimaging, and clinical characteristics) to finally predict therapeutic outcomes. An integration of molecular science with that of traditional clinical practice is the way forward to drug discoveries and novel therapeutic approaches and to characterize psychiatric disorders leading to a better predictive, preventive, and personalized medicine (PPPM) in psychiatry. With future advances in the omics technology and methodological developments for data integration, the goal of PPPM in psychiatry is promising.

Differential Peripheral Proteomic Biosignature of Fluoxetine Response in a Mouse Model of Anxiety/Depression

The incorporation of peripheral biomarkers in the treatment of major depressive disorders (MDD) could improve the efficiency of treatments and increase remission rate. Peripheral blood mononuclear cells (PBMCs) represent an attractive biological substrate allowing the identification of a drug response signature. Using a proteomic approach with high-resolution mass spectrometry, the present study aimed to identify a biosignature of antidepressant response (fluoxetine, a Selective Serotonin Reuptake Inhibitor) in PBMCs in a mouse model of anxiety/depression. Following determination of an emotionality score, using complementary behavioral analysis of anxiety/depression across three different tests (Elevated Plus Maze, Novelty Suppressed Feeding, Splash Test), we showed that a 4-week corticosterone treatment (35 μg/ml, CORT model) in C57BL/6NTac male mice induced an anxiety/depressive-like behavior. Then, chronic fluoxetine treatment (18 mg/kg/day for 28 days in the drinking water) reduced corticosterone-induced increase in emotional behavior. However, among 46 fluoxetine-treated mice, only 30 of them presented a 50% decrease in emotionality score, defining fluoxetine responders (CORT/Flx-R). To determine a peripheral biological signature of fluoxetine response, proteomic analysis was performed from PBMCs isolated from the “most” affected corticosterone/vehicle (CORT/V), corticosterone/fluoxetine responders and non-responders (CORT/Flx-NR) animals. In comparison to CORT/V, a total of 263 proteins were differently expressed after fluoxetine exposure. Expression profile of these proteins showed a strong similarity between CORT/Flx-R and CORT/Flx-NR (R = 0.827, p < 1e^-7). Direct comparison of CORT/Flx-R and CORT/Flx-NR groups revealed 100 differently expressed proteins, representing a combination of markers associated either with the maintenance of animals in a refractory state, or associated with behavioral improvement. Finally, 19 proteins showed a differential direction of expression between CORT/Flx-R and CORT/Flx-NR that drove them away from the CORT-treated profile. Among them, eight upregulated proteins (RPN2, HSPA9, NPTN, AP2B1, UQCRC2, RACK-1, TOLLIP) and one downregulated protein, TLN2, were previously associated with MDD or antidepressant drug response in the literature. Future preclinical studies will be required to validate whether proteomic changes observed in PBMCs from CORT/Flx-R mice mirror biological changes in brain tissues.

Translational Identification of Transcriptional Signatures of Major Depression and Antidepressant Response

Major depressive disorder (MDD) is a highly prevalent mental illness whose therapy management remains uncertain, with more than 20% of patients who do not achieve response to antidepressants. Therefore, identification of reliable biomarkers to predict response to treatment will greatly improve MDD patient medical care. Due to the inaccessibility and lack of brain tissues from living MDD patients to study depression, researches using animal models have been useful in improving sensitivity and specificity of identifying biomarkers. In the current study, we used the unpredictable chronic mild stress (UCMS) model and correlated stress-induced depressive-like behavior (n = 8 unstressed vs. 8 stressed mice) as well as the fluoxetine-induced recovery (n = 8 stressed and fluoxetine-treated mice vs. 8 unstressed and fluoxetine-treated mice) with transcriptional signatures obtained by genome-wide microarray profiling from whole blood, dentate gyrus (DG), and the anterior cingulate cortex (ACC). Hierarchical clustering and rank-rank hypergeometric overlap (RRHO) procedures allowed us to identify gene transcripts with variations that correlate with behavioral profiles. As a translational validation, some of those transcripts were assayed by RT-qPCR with blood samples from 10 severe major depressive episode (MDE) patients and 10 healthy controls over the course of 30 weeks and four visits. Repeated-measures ANOVAs revealed candidate trait biomarkers (ARHGEF1, CMAS, IGHMBP2, PABPN1 and TBC1D10C), whereas univariate linear regression analyses uncovered candidates state biomarkers (CENPO, FUS and NUBP1), as well as prediction biomarkers predictive of antidepressant response (CENPO, NUBP1). These data suggest that such a translational approach may offer new leads for clinically valid panels of biomarkers for MDD.

Genetic Contributions of Inflammation to Depression

This paper describes the effects of immune genes genetic variants and mRNA expression on depression's risk, severity, and response to antidepressant treatment, through a systematic review on all papers published between 2000 and 2016. Our results, based largely on case-control studies, suggest that common genetic variants and gene-expression pathways are involved in both immune activation and depression. The most replicated and relevant genetic variants include polymorphisms in the genes for interleukin (IL)-1β, IL-6, IL-10, monocyte chemoattractant protein-1, tumor necrosis factor-alpha, C-reactive protein, and phospholipase A2. Moreover, increased blood cytokines mRNA expression (especially of IL-1β) identifies patients that are less likely to respond to conventional antidepressants. However, even for the most replicated findings there are inconsistent results, not only between studies, but also between the immune effects of the genetic variants and the resulting effects on depression. We find evidence that these discrepant findings may be explained, at least in part, by the heterogeneity of the depression immunophenotype, by environmental influences and gene × environment interactions, and by the complex interfacing of genetic variants with gene expression. Indeed, some of the most robust findings have been obtained in patients developing depression in the context of treatment with interferon-alpha, a widely used model to mimic depression in the context of inflammation. Further 'omics' approaches, through GWAS and transcriptomics, will finally shed light on the interaction between immune genes, their expression, and the influence of the environment, in the pathogenesis of depression.

Predisposition to treatment response in major depressive episode: A peripheral blood gene coexpression network analysis

Antidepressant efficacy is insufficient, unpredictable and poorly understood in major depressive episode (MDE). Gene expression studies allow for the identification of significantly dysregulated genes but can limit the exploration of biological pathways. In the present study, we proposed a gene coexpression analysis to investigate biological pathways associated with treatment response predisposition and their regulation by microRNAs (miRNAs) in peripheral blood samples of MDE and healthy control subjects. We used a discovery cohort that included 34 MDE patients that were given 12-week treatment with citalopram and 33 healthy controls. Two replication cohorts with similar design were also analyzed. Expression-based gene network was built to define clusters of highly correlated sets of genes, called modules. Association between each module's first principal component of the expression data and clinical improvement was tested in the three cohorts. We conducted gene ontology analysis and miRNA prediction based on the module gene list. Nine of the 59 modules from the gene coexpression network were associated with clinical improvement. The association was partially replicated in other cohorts. Gene ontology analysis demonstrated that 4 modules were associated with cytokine production, acute inflammatory response or IL-8 functions. Finally, we found 414 miRNAs that may regulate one or several modules associated with clinical improvement. By contrast, only 12 miRNAs were predicted to specifically regulate modules unrelated to clinical improvement. Our gene coexpression analysis underlines the importance of inflammation-related pathways and the involvement of a large miRNA program as biological processes predisposing associated with antidepressant response.

MicroRNA-326 acts as a molecular switch in the regulation of midbrain urocortin 1 expression

BACKGROUND

Altered levels of urocortin 1 (Ucn1) in the centrally projecting Edinger-Westphal nucleus (EWcp) of depressed suicide attempters or completers mediate the brain's response to stress, while the mechanism regulating Ucn1 expression is unknown. We tested the hypothesis that microRNAs (miRNAs), which are vital fine-tuners of gene expression during the brain's response to stress, have the capacity to modulate Ucn1 expression.

METHODS

Computational analysis revealed that the Ucn1 3' untranslated region contained a conserved binding site for miR-326. We examined miR-326 and Ucn1 levels in the EWcp of depressed suicide completers. In addition, we evaluated miR-326 and Ucn1 levels in the serum and the EWcp of a chronic variable mild stress (CVMS) rat model of behavioural despair and after recovery from CVMS, respectively. Gain and loss of miR-326 function experiments examined the regulation of Ucn1 by this miRNA in cultured midbrain neurons.

RESULTS

We found reduced miR-326 levels concomitant with elevated Ucn1 levels in the EWcp of depressed suicide completers as well as in the EWcp of CVMS rats. In CVMS rats fully recovered from stress, both serum and EWcp miR-326 levels rebounded to nonstressed levels. While downregulation of miR-326 levels in primary midbrain neurons enhanced Ucn1 expression levels, miR-326 overexpression selectively reduced the levels of this neuropeptide.

LIMITATIONS

This study lacked experiments showing that in vivo alteration of miR-326 levels alleviate depression-like behaviours. We show only correlative data for miR-325 and cocaine- and amphetamine-regulated transcript levels in the EWcp.

CONCLUSION

We identified miR-326 dysregulation in depressed suicide completers and characterized this miRNA as an upstream regulator of the Ucn1 neuropeptide expression in midbrain neurons.

Transcriptomics in Interferon-α-Treated Patients Identifies Inflammation-, Neuroplasticity- and Oxidative Stress-Related Signatures as Predictors and Correlates of Depression

Owing to the unique opportunity to assess individuals before and after they develop depression within a short timeframe, interferon-α (IFN-α) treatment for chronic hepatitis C virus (HCV) infection is an ideal model to identify molecular mechanisms relevant to major depression, especially in the context of enhanced inflammation. Fifty-eight patients were assessed prospectively, at baseline and monthly over 24 weeks of IFN-α treatment. New-onset cases of depression were determined using the Mini International Neuropsychiatric Interview (MINI). Whole-blood transcriptomic analyses were conducted to investigate the following: (1) baseline gene expression differences associated with future development of IFN-α-induced depression, before IFN-α, and (2) longitudinal gene expression changes from baseline to weeks 4 or 24 of IFN-α treatment, separately in those who did and did not develop depression. Transcriptomics data were analyzed using Partek Genomics Suite (1.4-fold, FDR adjusted p⩽0.05) and Ingenuity Pathway Analysis Software. Twenty patients (34%) developed IFN-α-induced depression. At baseline, 73 genes were differentially expressed in patients who later developed depression compared with those who did not. After 4 weeks of IFN-α treatment, 592 genes were modulated in the whole sample, representing primarily IFN-α-responsive genes. Substantially more genes were modulated only in patients who developed depression (n=506, compared with n=70 in patients who did not), with enrichment in inflammation-, neuroplasticity- and oxidative stress-related pathways. A similar picture was observed at week 24. Our data indicate that patients who develop IFN-α-induced depression have an increased biological sensitivity to IFN-α, as shown by larger gene expression changes, and specific signatures both as predictors and as correlates.

Systematic review of blood transcriptome profiling in neuropsychiatric disorders: guidelines for biomarker discovery

INTRODUCTION

The utility of blood for genome-wide gene expression profiling and biomarker discovery has received much attention in patients diagnosed with major neuropsychiatric disorders. While numerous studies have been conducted, statistical rigor and clarity in terms of blood-based biomarker discovery, validation, and testing are needed.

METHODS

We conducted a systematic review of the literature to investigate methodological approaches and to assess the value of blood transcriptome profiling in research on mental disorders. We were particularly interested in statistical considerations related to machine learning, gene network analyses, and convergence across different disorders.

RESULTS

A total of 108 peripheral blood transcriptome studies across 15 disorders were surveyed: 25 studies used a variety of machine learning techniques to assess putative clinical viability of the candidate biomarkers; 11 leveraged a higher-order systems-level perspective to identify gene module-based biomarkers; and nine performed analyses across two or more neuropsychiatric phenotypes. Notably, ~50% of the surveyed studies included fewer than 50 samples (cases and controls), while ~75% included less than 100.

CONCLUSIONS

Detailed consideration of statistical analysis in the early stages of experimental planning is critical to ensure blood-based biomarker discovery and validation. Statistical guidelines are presented to enhance implementation and reproducibility of machine learning and gene network analyses across independent studies. Future studies capitalizing on larger sample sizes and emerging next-generation technologies set the stage for moving the field forwards. Copyright © 2016 John Wiley & Sons, Ltd.

Whole-genome expression analysis reveals genes associated with treatment response to escitalopram in major depression

The reasons for variability in treatment response in major depressive disorder (MDD) are not fully understood, but there is accumulating evidence suggesting that therapeutic outcomes of antidepressants can be influenced by genetic factors. In the present study we applied the microarray Illumina platform for whole genome expression profiling in depressive patients treated with escitalopram medication in order to identify genes underlying response to antidepressant treatment. The initial study sample consisted of 135 outpatients with major depressive disorder (mean age 31.1±11.6 years, 68% females) treated with escitalopram 10-20mg/day for 12 weeks, from which 87 patients (55 females) were included in gene expression analyzing. The gene expression profiles were measured on peripheral blood cells at baseline, at week 4 and at the end of treatment (week 12) using BeadChips Illumina. The fold change was used to demonstrate rate of changes in average gene expressions between studied groups. Statistical analyses were performed using the false discovery rate (FDR). The most interesting gene, which showed the predictive effect on treatment outcome by delineating low dose responders and treatment-resistant patients at the beginning of medication, was NLGN2, belonging to a family of neuronal cell surface proteins and involving in synapse formation. In addition, the several gene clusters, related to immune response, signal transduction and neurotrophin pathway, have distinguished responders from non-responders at the week 4 of treatment. After 4 weeks of escitalopram treatment (10mg/day), the YWHAZ gene has showed the highest transcriptional change in responders as compared with non-responders. Finally, at the end of the treatment we noticed that at least three genes (NR2C2, ZNF641, FKBP1A) have been strongly associated with resistance to escitalopram. Thus the results of this study support that exploration of peripheral gene expression is a useful tool in the further identification of novel genetic biomarkers for antidepressant treatment response.