Metabolomics Signatures of serotonin reuptake inhibitor (Escitalopram), serotonin norepinephrine reuptake inhibitor (Duloxetine) and Cognitive Behavior Therapy on Key Neurotransmitter Pathways in Major Depressive Disorder ============================================================================================================================================================================================================================= * Sudeepa Bhattacharyya * Siamak MahmoudianDehkordi * Matthew J Sniatynski * Marina Belenky * Vasant R. Marur * A. John Rush * W. Edward Craighead * Helen S. Mayberg * Boadie W. Dunlop * Bruce S Kristal * Rima Kaddurah-Daouk * Mood Disorder Precision Medicine Consortium ## ABSTRACT Metabolomics provides powerful tools that can inform about heterogeneity in disease and response to treatments. In this study, we employed an electrochemistry-based targeted metabolomics platform to assess the metabolic effects of three randomly-assigned treatments: escitalopram, duloxetine, and Cognitive Behavior Therapy (CBT) in 163 treatment-naïve outpatients with major depressive disorder. Serum samples from baseline and 12 weeks post-treatment were analyzed using targeted liquid chromatography-electrochemistry for metabolites related to tryptophan, tyrosine metabolism and related pathways. Changes in metabolite concentrations related to each treatment arm were identified and compared to define metabolic signatures of exposure. In addition, association between metabolites and depressive symptom severity (assessed with the 17-item Hamilton Rating Scale for Depression [HRSD17]) and anxiety symptom severity (assessed with the 14-item Hamilton Rating Scale for Anxiety [HRSA14]) were evaluated, both at baseline and after 12 weeks of treatment. Significant reductions in serum serotonin level and increases in tryptophan-derived indoles that are gut bacterially derived were observed with escitalopram and duloxetine arms but not in CBT arm. These include indole-3-propionic acid (I3PA), indole-3-lactic acid (I3LA) and Indoxyl sulfate (IS), a uremic toxin. Purine-related metabolites were decreased across all arms. Different metabolites correlated with improved symptoms in the different treatment arms revealing potentially different mechanisms between response to antidepressant medications and to CBT. ## INTRODUCTION Major depressive disorder (MDD) is a serious, disabling, and common condition that is biologically and etiologically heterogeneous[1, 2]. Clinical practice guidelines recommend first step treatment with either an antidepressant medication or an evidence-based psychotherapy, such as cognitive behavior therapy (CBT)[3–5]. Among first-line recommended medications are selective serotonin reuptake inhibitors (SSRIs) and serotonin norepinephrine reuptake inhibitors (SNRIs). Substantial evidence indicates that among outpatients with MDD, medications work as the second step when therapy fails as the first step, and vice versa[6, 7]. Similarly, there is strong evidence that a switch to different medication following a first failed medication trial is likely to be quite effective, demonstrating that mechanistic differences between antidepressants are important to characterize to enable more personalization of treatment [6, 8]. What works well for one patient may not be effective for another, but personalized approaches to MDD treatment are few, resulting in the common trial-and-error approach to treatment selection. Clinicians face a significant challenge in selecting the right treatment for an individual, given the variability in response to antidepressants and psychotherapy. These treatments take weeks to show noticeable benefits[9], which can be discouraging for patients and antidepressants can have side effects ranging from mild to severe[10], impacting a patient’s quality of life and willingness to continue treatment. Furthermore, a significant proportion of patients with MDD do not respond to first-line treatments, necessitating alternative strategies that may include combination therapy, augmentation strategies, or other forms of less traditional treatments[11]. Thus, despite progress in understanding depression and the availability of various treatment options, accurately predicting the most effective treatment for individual patients remains a significant challenge. Additionally, the specific molecular mechanisms that underlie the effectiveness of CBT, how these mechanisms differ from those involved in medication treatments, and why CBT may be more effective for certain patients are not well understood. Collectively, these insights indicate that treatment strategies for MDD, including both CBT and medication, may achieve remission through biologically distinct pathways.[12–14]. We previously have shown that metabolomic profiling is a promising approach that can provide informative insights into the molecular mechanisms of h SSRIs and how they differentially impact the metabolome in patients with MDD[15, 16]. In addition, we have provided insights into rapid metabolomic effects of ketamine[17, 18]. Prior findings suggested that in patients with MDD, the tryptophan pathway and its branches leading to serotonin, melatonin, 5-hydroxyindoleacetate, kynurenine, and indole derivatives appear to be significantly altered compared to healthy controls [15, 19]. The purine pathway, whose regulation is linked to tryptophan metabolism, has also been connected to MDD and other psychiatric illnesses[20]. Branched chain amino acids (BCAAs) that are linked to short chain acylcarnitine biosynthesis and mitochondrial energetics have been implicated in depression in several recent studies[21, 22]. Lower blood levels of BCAAs were associated with depression and increase in their level through dietary intake associated with improved symptoms[23]. However, no randomized control trial studies to our knowledge have directly compared the effects of treatment with psychotherapy versus one or more antidepressant medications on the metabolic pathways of neurotransmitters commonly implicated in MDD, such as serotonin, norepinephrine, or dopamine among MDD patients. In this report, we used a highly sensitive quantitative electrochemistry-based platform to interrogate effect of 12 weeks’ of treatment on key neurotransmitter pathways, including tryptophan, tyrosine and related pathways, among MDD patients randomly assigned to receive escitalopram (an SSRI), duloxetine (an SNRI), and CBT in the Predictors of Remission in Depression to Individual and Combined Treatments (PReDICT) study[24]. This secondary data analysis examined the following aims: 1. Identify baseline metabolites associated with baseline depressive symptoms (assessed with the 17-item Hamilton Rating Scale for Depression [HRSD17][25]) severity and baseline anxiety (assessed with the 14-item Hamilton Rating Scale for Anxiety [HRSA14][26]) severity. 2. Define metabolic signatures for treatment with escitalopram, duloxetine, and CBT; Compare and contrast these signatures. 3. Define metabolic changes associated with changes in depressive symptoms (HRSD17) and changes in anxiety (HRSA14) over 12 weeks of treatment within each treatment arm. ## 2. MATERIALS AND METHODS ### 2.1 Study Design and Participants The design and clinical outcomes of PReDICT have been detailed previously[7, 24, 27]. The PReDICT study was conducted through the Mood and Anxiety Disorders Program of Emory University from 2007-2013 (Clinical trial [NCT00360399](http://medrxiv.org/lookup/external-ref?link_type=CLINTRIALGOV&access_num=NCT00360399&atom=%2Fmedrxiv%2Fearly%2F2024%2F04%2F03%2F2024.04.02.24304677.atom) (08/04/2006) “Predictors of Antidepressant Treatment Response: The Emory CIDAR” [https://clinicaltrials.gov/ct2/show/NCT00360399](https://clinicaltrials.gov/ct2/show/NCT00360399)). The study was approved by the Emory Institutional Review Board. All patients provided written informed consent to participate. The study was conducted in accordance with the 1975 Declaration of Helsinki and its amendments. Briefly, PReDICT aimed to identify predictors and moderators of response to 12 weeks of randomly assigned treatment with duloxetine (30-60 mg/day), escitalopram (10-20 mg/day) or CBT (16 one-hour individual sessions). Eligible participants were adults aged 18-65 with non-psychotic MDD who had never previously been treated for depression. The diagnosis of MDD was made via interview by assessors trained in administering the Structured Clinical Interview for DSM-IV (SCID)[28]. The SCID was also used to diagnose any comorbid psychiatric disorders, including the exclusionary diagnoses of bipolar disorder, psychotic disorder, obsessive compulsive disorder, anorexia nervosa, and current substance abuse or dependence. Additional exclusionary criteria included a neurocognitive disorder, pregnancy, lactation, any uncontrolled general medical condition, or a positive urine drug screen for illicit drugs. Severity of depression at the randomization visit was assessed with the HRSD17. Eligibility required an HRSD17 score ≥18 at the screening visit and ≥15 at the randomization visit, indicative of moderate-to-severe depression. ### 2.2 Metabolomic Profiling and Ratios On the day of randomization (without consideration of diet, time of day or fasting status), 10 cc of whole blood was collected in a red-top silicone-coated tube from an antecubital vein between 8 am and 4 pm. After the blood had clotted for 20 min, the tubes were centrifuged at 1400 g at 4LC for 10 mins. Serum was then pipetted into 1 mL vials and frozen at −80 C until thawed for analyses. #### 2.2.1 High Performance Liquid Chromatograph-Electrochemical detection (HPLC-ECD) Metabolite extraction from human serum and analyses by HPLC-ECD were conducted as described in detail previously,14-20 with minor changes (e.g., specific columns) as noted below. Briefly, the proteins from 125 μL of serum were precipitated by adding 0.5 ml of cold (−20°C) acetonitrile/0.4% glacial acetic acid. After centrifugation for 15 minutes at 4°C, at 15000 x g, the supernatant was transferred into microcentrifuge tubes and dried under vacuum. The dried extract was reconstituted in 100 μL of mobile phase A prior to injection; 50 μL of each sample was injected. Reversed phase LC separation was performed using buffers containing high concentration of salts and pentane sulfonic acid (PSA). Salts were beneficial to maintain conductivity of the EC electrodes, and PSA was used as an ion-pairing reagent to help the retention and separation of hydrophilic metabolites. In this study, we used our standard serum metabolites profiling LC separation method without any modifications to see which 2-oxoacids could be quantified along with other redox metabolites we routinely analyze. Two connected reversed-phase C18 Luna 250 × 4.6 × 5 μm columns (Phenomenex, Torrance, CA) at 30°C were used to improve separation. Mobil phase A was comprised of 60mM aqueous solution of sodium 1-pentasulfonate (PSA), 0.1% methanol, and 1mg/L citric acid. Mobile phase B consisted of 80:10:10 MeOH/ACN/IPA with 40 mM lithium acetate, 2.0% acetic acid, and 10 mg/L citric acid. The separation was completed over 112 minutes. Detection was done using 16-channel coulometric array detector (i.e., CoulArrayTM, combining four of Hi-E four sensors assembly cells, (ThermoFisher Scientific, Chelmsford, MA) operated with potentials incremented in 60 mV steps (0−900 mV). In our system, Pd electrode is used as a reference electrode. The HPLC-ECD system was controlled by CoulArray software (CoulArray software for Windows, version 3.10). Data analysis was performed using CoulArray Data Station software for Windows. #### 2.2.2 Biocrates P180 Platform Branched-chain amino acids valine and isoleucine were quantified with a targeted metabolomics approach using the AbsoluteIDQ® p180 Kit (BIOCRATES Life Science AG, Innsbruck, Austria), with an ultra-performance liquid chromatography (UPLC)/MS/MS system [Acquity UPLC (Waters), TQ-S triple quadrupole MS/MS (Waters)]. This platform provides measurements of up to 186 endogenous metabolites in quantitative mode (amino acids and biogenic amines) and semi-quantitative mode (acylcarnitines, sphingomyelins, phosphatidylcholines and lysophosphatidylcholines across multiple classes). The AbsoluteIDQ® p180 kit has been fully validated according to European Medicine Agency Guidelines on bioanalytical method validation. The p180 kit includes all quality control (QC) samples, and calibration and internal standards; therefore, quantifications can be directly compared across studies. The quality-control and metabolomic profiling were previously published[16, 29, 30]. In short, metabolites with 1) > 40% of measurements below the lower limit of detection (LOD) and 2) > 30% of SPQC coefficient of variation were excluded from the analysis. To adjust for the batch effects, a correction factor for each metabolite in a specific plate was obtained by dividing the metabolite’s QC global average by QC average within the plate. Missing values were imputed using each metabolite’s LOD/2. #### 2.2.3 Metabolic Ratios To determine which enzymatic processes may be affected by the disease and/or treatment, we expanded metabolic profile with certain ratios of related metabolites within tryptophan, tyrosine, purine and BCAA pathways. These ratios included: 1. Purine Pathway: Uric:Xanthine, Hypoxanthine: Xanthine, Hypoxanthine:Uric 2. Tryptophan Pathway: Tryptophan:Tyrosine, Indole-3-acetic acid:Tryptophan, Indole-3-lactic acid:Tryptophan, Indole-3-propionic acid:Tryptophan, Indoxyl sulfate: Tryptophan, Indoxyl sulfate: Indole-3-propionic acid, 5HT: Tryptophan, Tryptophan: Kynurenine 3. BCAA Pathway: Three ratios were tested: 1) 4-methyl-2-oxopentanoate (b036):Isoleucine, 2) b036:valine and 3) the ratio of b036:3-methyl-2-oxopentanoate (b032). ### 2.3 Depression and Anxiety Symptoms At baseline and after 12 weeks of treatment, severity of depression and anxiety was assessed by blinded raters using the HRSD17 and HRSA14, respectively. ### 2.4 Statistical Analysis Differences in demographic variables and depression scores across the response groups were evaluated using ANOVA and the Pearson Chi-squared test (for categorical variables). Patients with body mass index (BMI) ≥ 30 were labeled as obese vs. BMI<30 as “Non-obese”. All analyses were performed in a metabolite-wise manner in 1) all samples, 2) males, 3) females, 4) obese, 5) non-obese, 6) Hispanic, and 7) non-Hispanic. To assess the significance of a metabolite’s log2 fold-change within each subpopulation, bootstrapping with 1000 repeats were conducted using the “bootstraps” function from the “rsample” package. To assess the association of baseline metabolite levels and the continuous variables HRSD17 and HRSA14, partial correlation analyses were conducted using partial Spearman rank correlation, adjusted for age, sex, and body mass index. Significance of partial correlation were computed using bootstrapping with 1000 repeats. A p-value <0.05 was considered significant. Given the exploratory nature of this initial investigation, no correction for multiple comparisons was made. ## 3. RESULTS ### 3.1 Participant Characteristics (Demographic and Clinical) **Table 1** summarizes the demographic and clinical features of the 163 participants in the PReDICT Study. Of these, 41% of participants were male, and mean (standard deviation) baseline age, HRSD17, and HRSA14 were 39(12) years, 20(3.8) points, and 16 (5.2) points, respectively. Of 163 participants, 131 individuals had both baseline and week 12 electrochemistry-based profile. and 107 individuals had valine and isoleucine measurement at both visits as these two compounds were profiled on the p180 platform. Baseline total HRSD17 scores were highly correlated with HRSA14 scores (Spearman rank correlation rho= 0.61, p = 2.2E-16). View this table: [Table 1:](http://medrxiv.org/content/early/2024/04/03/2024.04.02.24304677/T1) Table 1: Participant Demographic and Clinical Characteristics ### 3.2 Baseline Profiles and Disease Severity We explored metabolite profiles at baseline that correlated with depression and anxiety symptom severity (i.e., HRSD17, and HRSA14) using Spearman partial correlation analysis adjusted for age, sex, and BMI. Estimated partial correlations (rho) and *p*-values were obtained from bootstrapping and are presented in **Table 2** for the overall samples and in **Supplementary Fig 1** for stratified analyses. We identified 14 metabolites and ratios at baseline that were significantly correlated with depression severity and/or anxiety severity. As expected, due to a high correlation between these two outcomes (rho=0.61), the direction of correlations was similar. View this table: [Table 2:](http://medrxiv.org/content/early/2024/04/03/2024.04.02.24304677/T2) Table 2: Significant Spearman Partial Correlation of Baseline Metabolites/Ratios and Disease Severity. Spearman Partial correlations were adjusted for age, sex, and baseline BMI. Partial correlations, p-values and 95% confidence intervals were calculated using bootsrtaping with 1000 repetitions. *: p<0.05; **: p<0.01; \***|: p<0.001 In *the tryptophan pathway*, tryptophan showed a positive correlation with both depression and anxiety symptom severity, while indoxyl sulfate and kynurenine were only significantly positively correlated with anxiety symptom. In *the phenylalanine/tyrosine pathway*, tyrosine showed positive correlations with both depression and anxiety symptoms similar to tryptophan. Higher levels of homovanillic acid were associated with less severe anxiety symptoms. In *the cysteine/methionine pathway*, cystine, an oxidized dimeric form of the amino acid cysteine, was significantly positively associated with greater severity of both depression and anxiety symptoms; methionine was positively associated with depression symptoms and higher levels of S-adenosyl-homocysteine, which is a precursor of all homocysteine, was associated with greater anxiety symptoms. In *the glycolysis/energy metabolism pathway*, higher levels of pyruvate showed a significant positive correlation with higher anxiety symptoms. 3-methyl-2-oxopentanoate and 4-methyl-2-oxopentanoate, two metabolites of the BCAAs valine and isoleucine, were also positively correlated with more severe depression and anxiety scores. The ratio of I3PA:TRP showed a negative correlation with depression severity and IS:TRP showed a positive significant correlation with total anxiety score, suggesting decreased production of I3PA and increased production of indoxyl sulfate from tryptophan in patients with more severe symptoms. ### 3.3 Metabolite Signatures of 12 Weeks of treatment with antidepressant medication or CBT All three treatments triggered significant changes in metabolites implicated in neurotransmitter related pathways. However, when compared to CBT, the changes observed in the medication arms were more similar **(Figure 1).** In addition, we noted similar signatures across different strata (men, women, Hispanic, non-Hispanic, obese, non-obese) of the study (**Supplementary Fig 2**). ![Figure 1:](http://medrxiv.org/https://www.medrxiv.org/content/medrxiv/early/2024/04/03/2024.04.02.24304677/F1.medium.gif) [Figure 1:](http://medrxiv.org/content/early/2024/04/03/2024.04.02.24304677/F1) Figure 1: Signature of Exposure. Forest plot shows 12 weeks log2 fold change and 95% bootstrap confidence intervals within each treatment group. P-values were computed using bootstrapping with 1000 repetitions. A negative value represents decrease in metabolite levels and a positive value shows an increase. *In the tryptophan pathway*, the largest impact was observed in serotonin levels which decreased significantly in both the escitalopram and duloxetine arms, while no change was observed in the CBT arm. The ratio of serotonin to tryptophan was also significantly lower in the medication arms but remained unchanged in CBT. The indole containing metabolites of tryptophan, primarily generated by the gut bacteria such as indoxyl sulfate, indole 3-lactic acid and indole 3-propionic acid were all increased by medications but not CBT. The ratios of each of the indoles to tryptophan were also higher in the medication arms but showed no change in CBT. *In the purine pathway,* hypoxanthine, xanthine and uric acid decreased or trended to decrease in all three arms. Specifically, uric acid and xanthine in the CBT arm, hypoxanthine in the duloxetine arm, and xanthine in the escitalopram arm showed significant decreases with treatment. *In the phenylalanine/tyrosine pathway*, homovanillic acid showed small increases in the medication arms, which was statistically significant with duloxetine. Vinylmandelic acid, a metabolic by-product of norepinephrine and epinephrine, showed significant decrease in the CBT arm but not in the medication arms. Among other molecules, salicylic acid (SA) and its active metabolite 2,5-dihydroxybenzoic acid, also known as gentisic acid, showed significant decreases in the CBT arm. SA was also decreased with exposure in the duloxetine arm but not in the escitalopram arm. ### 3.3 Metabolic system changes in relation to depressive and anxious symptom change for each treatment We compared the metabolite changes that correlated with improved symptoms of depression and anxiety distress upon treatment with medications and CBT treatments. **Table 3** lists the metabolites that showed changes in their serum levels associated with changes in symptom severity. View this table: [Table 3:](http://medrxiv.org/content/early/2024/04/03/2024.04.02.24304677/T3) Table 3: Significant Associations Between %Change in Disease Severity Score and Log2 Fold Change Metabolites/Ratios. Linear regression models were adjusted for age, sex, and baseline BMI. *p*-values and 95% confidence intervals were calculated using bootstrapping with 1000 repetitions. Positive beta values indicate that decreases in metabolites or ratios changes were associated with improvement in symptoms; negative beta values indicate that increases in metabolites or ratios were associated with improvement in symptoms. Shaded boxes reflect significant associations.*: p<0.05; **: p<0.01; \***|: p<0.001 Among the metabolites of *tryptophan pathway*, indoxyl sulfate and kynurenine increased with improvements in depression and anxiety symptoms in the duloxetine arm. The ratios of IS:TRP and KYN:TRP were significantly negatively correlated with depression symptom severity in the duloxetine arm. *In the purine pathway*, Hypoxanthine and xanthine were decreased in the CBT arm and were significantly correlated with improvements in depression and anxiety symptom severity. Xanthine also showed significant positive correlation with depression symptom severity in the duloxetine arm but not in the escitalopram arm. The ratio of Uric acid:Xanthine was increased with improvements in depression symptom severity in both CBT and duloxetine arms. *The branched-chain amino acid (BCAA),* Increase in valine was significantly associated with improvements in depressive symptom severity both in CBT and duloxetine arms and also correlated with decrease in anxiety symptom severity in the duloxetine arm and showed a similar increasing trends (non-significant at p<0.05) in the CBT arm. In contrast, decrease in two metabolites of the BCAA biosynthesis pathway, 3-methyl-2-oxopentanoate and 4-methyl-2-oxopentanoate were associated with improvements in depressive symptoms in the duloxetine arm. The ratios of 4-methyl-2-oxopentanoate/valine and 4-methyl-2-oxopentanoate/Isoleucine were positively correlated with depressive symptom severity in the duloxetine arm. ## 4. DISCUSSION In this study, we applied a targeted liquid-chromatography coupled to electrochemical coulometric array (LCECA) metabolomics platform to compare metabolite profiles in serum samples from MDD patients in the three arms of the PreDICT study. The targeted metabolites were amino acids and redox metabolites primarily related to neurotransmitter pathways. Among the highlights of our findings in the MDD patients was the replication of our metabolomic analyses from other depression samples, that antidepressant medications significantly lowered serum serotonin levels; further, and the ratio of serotonin to its parent metabolite, tryptophan, after 12 weeks of treatment. Our results, for the first time showed that this decrease in serotonin levels was not observed in the CBT arm. We and others[15, 16, 31–33] have shown that plasma 5-HT levels decreased after SSRI treatment, and that higher baseline 5HT levels were associated with better treatment response. The exact mechanism associated with this robust decrease in circulating serotonin with the administration of SSRIs are not known even though several mechanisms including 5-HT1A autoreceptor hypofunctioning[32, 34] and potential SSRI-induced blockade of platelet uptake of 5-HT from cells in the gut[32] have been suggested. The other major differentiating metabolomic effect of the treatments was the significant shift towards the indole acids branch of the tryptophan metabolic pathway in the medication arms but not in CBT. Indole 3-lactic acid, indole 3-propionic acid and indoxyl sulfate, and their ratios to tryptophan, all increased significantly with 12 weeks of drug exposure, but were not observed in the CBT arm. This shift in the tryptophan metabolic pathway towards increased production of the gut-microbiota-synthesized indoles with exposure to antidepressants has been reported previously by our group in multiple studies[15, 35]. These observations highlight the differences in the mode of action of the drugs and the psychotherapy even though both are efficacious in the treatment and remission of MDD. Indoxyl sulfate is a known uremic toxin that has been shown, by our group, to be associated with increased severity of depression and anxiety symptoms in depression patients[35]. Given these associations, the finding that treatment with either escitalopram or duloxetine, significantly increased the serum levels of indoxyl sulfate was surprising. This increase in uremic toxin can be interpreted as potential side-effect of the drugs via modulation of gut microbiome. More investigations are needed to confirm this increase in indoxyl sulfate as a possible consequence of these antidepressant medications. Across all three treatments there were statistically significant, or trend-level decreases of purine pathway metabolites, including hypoxanthine, xanthine, and uric acid and they were associated with improved symptoms in CBT and duloxetine arms. These results may indicate that the purine pathway metabolites are associated with or are reflective of treatment response in these patients. This corroborates our previous findings on a different MDD cohort exposed to either citalopram/escitalopram[15] or sertraline (other SSRIs) or placebo[36], wherein we had observed significant decreases or decreasing trends in xanthine and hypoxanthine levels in both drugs or placebo groups, with exposure. It is possible that decreases in these purine metabolites indicate reduction in oxidative stress by direct or indirect inhibition of the xanthine oxidase enzyme that generates vascular oxidative stress through reactive oxygen species production by catalyzing the hypoxanthine to xanthine to urate synthesis pathway. Branched chain amino acids like valine, leucine and isoleucine have been suggested to play crucial roles in the development of depression through their activation of the mammalian target of rapamycin (mTor) pathway[21]. Our results in this study show that that at baseline higher levels of two intermediates of the BCAA biosynthesis pathway, 3-methyl-2-oxopentanoate and 4-methyl-2-oxopentanoate were correlated with more severe depressive and anxiety symptoms and a decrease in these two metabolites over the course of 12 weeks of CBT or duloxetine treatment was associated with improved symptoms. In contrast, an increase in valine with exposure to CBT or duloxetine were correlated with improvements in depressive symptoms. The decrease in these intermediates potentially may be due to increased synthesis of the downstream BCAA in response to duloxetine exposure. None of these changes were observed in the escitalopram arm suggesting that even though the drugs were more similar in their associations with metabolite alterations compared to CBT, they differed in their impact on the BCAA biosynthesis pathway. There are several limitations to our study. The first limitation is that more insights could be gained by a dose response study. A single dose of the medications used in the experiment may not be sufficient to fully understand the effects or potential benefits of the treatment being studied. The second limitation is that we did not have catecholamines profiled. Catecholamines are a group of hormones that include metabolites like adrenaline and noradrenaline. Profiling catecholamines could provide important information about their involvement in the biochemical processes being studied; this is especially relevant because our comparisons included a SSRI and a SNRI, the latter largely associated with the catecholamine metabolic pathway. Without these data, it is challenging to fully understand the interactions and effects of the treatment. The final limitation is the inability to see the link between BCAA (branched-chain amino acids) biochemistry and other metabolites. BCAAs compromise a group of essential amino acids that play important roles in various biochemical processes in the body. However, their interactions with other metabolites, such as other amino acids or metabolic intermediates, may provide further insights into their functions and potential effects. Without studying these links, our understanding of the overall biochemical processes was limited. ## CONFLICT OF INTEREST Dr. Dunlop has received research support from, Boehringer Ingelheim, Compass, NIMH, and Usona Institute, and has served as a consultant to Biohaven, Cerebral Therapeutics, Myriad Neuroscience, and Otsuka. Dr. Craighead receives research support from the NIH; is a board member of Hugarheill ehf, an Icelandic company dedicated to the prevention of depression; receives book royalties from John Wiley; and is supported by the Mary and John Brock Foundation, the Pitts Foundation, and the Fuqua family foundations. He is a consultant to the George West Mental Health Foundation and a member of the Scientific Advisory Boards of AIM for Mental Health, Galen Mental Health, and the ADAA. A. John Rush has received consulting fees from Compass Inc., Curbstone Consultant LLC, Emmes Corp., Evecxia Therapeutics, Inc., Holmusk Technologies, Inc., ICON, PLC, Johnson and Johnson (Janssen), Liva-Nova, MindStreet, Inc., Neurocrine Biosciences Inc., Otsuka-US; speaking fees from Liva-Nova, Johnson and Johnson (Janssen); and royalties from Wolters Kluwer Health, Guilford Press and the University of Texas Southwestern Medical Center, Dallas, TX (for the Inventory of Depressive Symptoms and its derivatives). He is also named co-inventor on two patents: U.S. Patent No. 7,795,033: Methods to Predict the Outcome of Treatment with Antidepressant Medication, Inventors: McMahon FJ, Laje G, Manji H, Rush AJ, Paddock S, Wilson AS; and U.S. Patent No. 7,906,283: Methods to Identify Patients at Risk of Developing Adverse Events During Treatment with Antidepressant Medication, Inventors: McMahon FJ, Laje G, Manji H, Rush AJ, Paddock S. Rima Kaddurah-Daouk is an inventor on key patents in the field of Metabolomics and hold equity in Metabolon, a biotech company in North Carolina. In addition, she holds patents licensed to Chymia LLC and PsyProtix with royalties and ownership. All other authors reported no biomedical financial interests or potential conflicts of interest. ## AUTHOR CONTRIBUTIONS S.B. and S.M. did analyses of data and helped write the manuscript; B.K. and his team generated biochemical data and wrote its methods and helped with interpretation of findings; B.W.D., W.E.C. and A.J.R. helped with interpretation of findings and clinical relevance; H.S. helped with background literature searches. R.K.D. is PI for project and helped with concept development, study design, data interpretation and connecting biochemical and clinical data, and with the writing of the manuscript. ## FUNDING This work was funded by grant support to Dr. Rima Kaddurah-Daouk (PI) through NIH grants R01MH108348, R01AG046171 and U01AG061359. Metabolomics data is provided by the Mood Disorders Precision Medicine Consortium (MDPMC). The MDPMC is led by Dr. Kaddurah-Daouk at Duke University in partnership with a large number of academic institutions. As such, the investigators within the MDPMC, not listed specifically in this publication’s author’s list, provided data along with its pre-processing and prepared it for analyses, but did not participate in analyses or writing of this manuscript. A complete listing of MDPMC investigators can be found at: [https://sites.duke.edu/mdpmc/files/2022/06/MDPMC-Members2022.pdf](https://sites.duke.edu/mdpmc/files/2022/06/MDPMC-Members2022.pdf). The PReDICT study was supported by NIH grants P50-MH077083 (PI Mayberg), R01-MH080880 (PI Craighead), UL1-RR025008 (PI Stevens), M01-RR0039 (PI Stevens) and the Fuqua family foundations. A preprint of this manuscript is available on medRxiv. ## Supporting information Sup Fig- 1 [[supplements/304677_file04.pdf]](pending:yes) Sup Fig-2 [[supplements/304677_file05.pdf]](pending:yes) ## Data Availability All data produced in the present study are available upon reasonable request to the authors. ## CONTRIBUTION TO THE FIELD STATEMENT This study contributes to the field of major depressive disorder by identifying and mapping biochemical signature associated with 3 different therapies. **Supplementary Figure 1:** Forest Plots Showing Baseline Association (and 95% confidence interval) of Clinical Outcomes and Metabolite Levels within each Strata. **Supplementary Figure 2:** Forest Plots Showing 12 Weeks change and 95% confidence Interval of metabolites in response to treatment within each Strata ## ACKNOWLEDGEMENTS We acknowledge the assistance of Ms. Lisa Howerton (Duke). ## ABBREVIATIONS HRSA14 : 14-item Hamilton Anxiety Rating Scale HRSD17 : 17-item Hamilton Depression Rating Scale MDD : Major Depressive Disorder PReDICT : Predictors of Remission in Depression to Individual and Combined Treatments study * Received April 2, 2024. * Revision received April 2, 2024. * Accepted April 3, 2024. * © 2024, Posted by Cold Spring Harbor Laboratory This pre-print is available under a Creative Commons License (Attribution-NonCommercial-NoDerivs 4.0 International), CC BY-NC-ND 4.0, as described at [http://creativecommons.org/licenses/by-nc-nd/4.0/](http://creativecommons.org/licenses/by-nc-nd/4.0/) ## References 1. 1.He, J., et al., Regional metabolic heterogeneity in anterior cingulate cortex in major depressive disorder: A multi-voxel (1)H magnetic resonance spectroscopy study. J Affect Disord, 2022. 318: p. 263–271. 2. 2.Suseelan, S. and G. Pinna, Heterogeneity in major depressive disorder: The need for biomarker-based personalized treatments. Adv Clin Chem, 2023. 112: p. 1–67. 3. 3.Malhi, G.S., et al., The 2020 Royal Australian and New Zealand College of Psychiatrists clinical practice guidelines for mood disorders: Major depression summary. Bipolar Disord, 2020. 22(8): p. 788–804. 4. 4.Malhi, G.S., et al., NICE guideline for depression ranks short-term psychodynamic psychotherapy (STTP). Bipolar Disord, 2022. 24(5): p. 467–471. 5. 5.Malhi, G.S., et al., The management of depression: the evidence speaks for itself. Br J Psychiatry, 2023. 222(3): p. 97–99. 6. 6.Thase, M.E., et al., Cognitive therapy versus medication in augmentation and switch strategies as second-step treatments: a STAR*D report. Am J Psychiatry, 2007. 164(5): p. 739–52. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.ajp.164.5.739&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=17475733&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F04%2F03%2F2024.04.02.24304677.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=000246370300014&link_type=ISI) 7. 7.Dunlop, B.W., et al., Benefits of Sequentially Adding Cognitive-Behavioral Therapy or Antidepressant Medication for Adults With Nonremitting Depression. Am J Psychiatry, 2019. 176(4): p. 275–286. 8. 8.Ionescu, D.F., J.F. Rosenbaum, and J.E. Alpert, Pharmacological approaches to the challenge of treatment-resistant depression. Dialogues Clin Neurosci, 2015. 17(2): p. 111–26. 9. 9.Kulikov, A.V., et al., Interplay between the key proteins of serotonin system in SSRI antidepressants efficacy. Expert Opin Ther Targets, 2018. 22(4): p. 319–330. 10. 10.Salwan, A., et al., Side effects of antidepressants, in Side Effects of Drugs Annual. 2022, Elsevier. p. 25–36. 11. 11.Scott, F., et al., Systematic review and meta-analysis of augmentation and combination treatments for early-stage treatment-resistant depression. J Psychopharmacol, 2023. 37(3): p. 268–278. 12. 12.Dunlop, B.W., et al., Functional Connectivity of the Subcallosal Cingulate Cortex And Differential Outcomes to Treatment With Cognitive-Behavioral Therapy or Antidepressant Medication for Major Depressive Disorder. Am J Psychiatry, 2017. 174(6): p. 533–545. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.ajp.2016.16050518&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=28335622&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F04%2F03%2F2024.04.02.24304677.atom) 13. 13.Lam, R.W., Review: antidepressants and psychotherapy may be equally effective for promoting remission in major depressive disorder. Evid Based Ment Health, 2003. 6(2): p. 45. 14. 14.Etkin, A., et al., A cognitive-emotional biomarker for predicting remission with antidepressant medications: a report from the iSPOT-D trial. Neuropsychopharmacology, 2015. 40(6): p. 1332–42. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/npp.2014.333&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=25547711&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F04%2F03%2F2024.04.02.24304677.atom) 15. 15.Bhattacharyya, S., et al., Metabolomic signature of exposure and response to citalopram/escitalopram in depressed outpatients. Transl Psychiatry, 2019. 9(1): p. 173. 16. 16.MahmoudianDehkordi, S., et al., Alterations in acylcarnitines, amines, and lipids inform about the mechanism of action of citalopram/escitalopram in major depression. Transl Psychiatry, 2021. 11(1): p. 153. 17. 17.Singh, B., et al., Metabolomic signatures of intravenous racemic ketamine associated remission in treatment-resistant depression: A pilot hypothesis generating study. Psychiatry Res, 2022. 314: p. 114655. 18. 18.Rotroff, D.M., et al., Metabolomic signatures of drug response phenotypes for ketamine and esketamine in subjects with refractory major depressive disorder: new mechanistic insights for rapid acting antidepressants. Transl Psychiatry, 2016. 6(9): p. e894. 19. 19.Badawy, A.A., S. Dawood, and S. Bano, Kynurenine pathway of tryptophan metabolism in pathophysiology and therapy of major depressive disorder. World J Psychiatry, 2023. 13(4): p. 141–148. 20. 20.Ali-Sisto, T., et al., Purine metabolism is dysregulated in patients with major depressive disorder. Psychoneuroendocrinology, 2016. 70: p. 25–32. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1016/j.psyneuen.2016.04.017&link_type=DOI) 21. 21.Baranyi, A., et al., Branched-Chain Amino Acids as New Biomarkers of Major Depression - A Novel Neurobiology of Mood Disorder. PLoS One, 2016. 11(8): p. e0160542. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1371/journal.pone.0160542&link_type=DOI) 22. 22.Whipp, A.M., et al., Branched-chain amino acids linked to depression in young adults. Front Neurosci, 2022. 16: p. 935858. 23. 23.Koochakpoor, G., et al., Dietary intake of branched-chain amino acids in relation to depression, anxiety and psychological distress. Nutr J, 2021. 20(1): p. 11. 24. 24.Dunlop, B.W., et al., Effects of Patient Preferences on Outcomes in the Predictors of Remission in Depression to Individual and Combined Treatments (PReDICT) Study. Am J Psychiatry, 2017. 174(6): p. 546–556. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1176/appi.ajp.2016.16050517&link_type=DOI) 25. 25.Hamilton, M., A rating scale for depression. J Neurol Neurosurg Psychiatry, 1960. 23(1): p. 56–62. [FREE Full Text](http://medrxiv.org/lookup/ijlink/YTozOntzOjQ6InBhdGgiO3M6MTQ6Ii9sb29rdXAvaWpsaW5rIjtzOjU6InF1ZXJ5IjthOjQ6e3M6ODoibGlua1R5cGUiO3M6MzoiUERGIjtzOjExOiJqb3VybmFsQ29kZSI7czo0OiJqbm5wIjtzOjU6InJlc2lkIjtzOjc6IjIzLzEvNTYiO3M6NDoiYXRvbSI7czo1MDoiL21lZHJ4aXYvZWFybHkvMjAyNC8wNC8wMy8yMDI0LjA0LjAyLjI0MzA0Njc3LmF0b20iO31zOjg6ImZyYWdtZW50IjtzOjA6IiI7fQ==) 26. 26.Hamilton, M., The assessment of anxiety states by rating. Br J Med Psychol, 1959. 32(1): p. 50–5. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1111/j.2044-8341.1959.tb00467.x&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=13638508&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F04%2F03%2F2024.04.02.24304677.atom) [Web of Science](http://medrxiv.org/lookup/external-ref?access_num=A1959CCE5400005&link_type=ISI) 27. 27.Dunlop, B.W., et al., Predictors of remission in depression to individual and combined treatments (PReDICT): study protocol for a randomized controlled trial. Trials, 2012. 13: p. 106. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1186/1745-6215-13-106&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=22776534&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F04%2F03%2F2024.04.02.24304677.atom) 28. 28.First, M.B., et al., SCID-I: User’s guide for the Structured Clinical Interview for DSM-IV Axis I disorders. 1997: American Psychiatric Press. 29. 29.St John-Williams, L., et al., Targeted metabolomics and medication classification data from participants in the ADNI1 cohort. Sci Data, 2017. 4: p. 170140. 30. 30.Bhattacharyya, S., et al., Pilot Study of Metabolomic Clusters as State Markers of Major Depression and Outcomes to CBT Treatment. Front Neurosci, 2019. 13: p. 926. 31. 31.Gupta, M., et al., TSPAN5, ERICH3 and selective serotonin reuptake inhibitors in major depressive disorder: pharmacometabolomics-informed pharmacogenomics. Mol Psychiatry, 2016. 21(12): p. 1717–1725. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1038/mp.2016.6&link_type=DOI) 32. 32.Holck, A., et al., Plasma serotonin levels are associated with antidepressant response to SSRIs. J Affect Disord, 2019. 250: p. 65–70. 33. 33.Sun, Y., et al., The relationship between plasma serotonin and kynurenine pathway metabolite levels and the treatment response to escitalopram and desvenlafaxine. Brain Behav Immun, 2020. 87: p. 404–412. 34. 34.Lemonde, S., et al., Association of the C(-1019)G 5-HT1A functional promoter polymorphism with antidepressant response. Int J Neuropsychopharmacol, 2004. 7(4): p. 501–6. [CrossRef](http://medrxiv.org/lookup/external-ref?access_num=10.1017/S1461145704004699&link_type=DOI) [PubMed](http://medrxiv.org/lookup/external-ref?access_num=15447813&link_type=MED&atom=%2Fmedrxiv%2Fearly%2F2024%2F04%2F03%2F2024.04.02.24304677.atom) 35. 35.Brydges, C.R., et al., Indoxyl sulfate, a gut microbiome-derived uremic toxin, is associated with psychic anxiety and its functional magnetic resonance imaging-based neurologic signature. Sci Rep, 2021. 11(1): p. 21011. 36. 36.Kaddurah-Daouk, R., et al., Pretreatment metabotype as a predictor of response to sertraline or placebo in depressed outpatients: a proof of concept. Transl Psychiatry, 2011. 1(7): p. e26-.