Patients included in the study were recruited at the Addiction treatment facilities of the Institute of Neuropsychiatry and Addiction of Parc de Salut Mar, in Barcelona. A total of 110 patients were informed about the study. Of them, 23 patients refused to participate, 36 patients did not meet inclusion criteria, and 26 patients did not complete the study. Healthy controls were healthy volunteers, were recruited from a database of healthy subjects willing to participate in medical research projects. For this study, some subjects were postgraduate university students. Thus, a total of 35 subjects participated in the study; 16 cases diagnosed of CUD and primary MDD, 9 cases diagnosed of CUD and CUD-induced MDD, and 10 healthy controls (Table 1). All diagnoses were done according DSM-IV-TR criteria (23). Inclusion criteria included both genders, older than 18 years, Caucasian origin, and body mass indices between 19 and 29 kg/m². In the MDD groups (primary or CUD-induced), the most recent episode should be in remission, and Hamilton Depression Rating Scale (HDRS) (24) scores should be less than 6. In the CUD groups, subjects should maintain at least 4 weeks of abstinence, confirmed by negative urine controls. Exclusion criteria included: cognitive or language limitations that precluded evaluations; pregnant or breastfeeding women and any medical problem that could interfere in the study procedures; in the comorbid CUD/MDD groups: any psychiatric disorder in Axis I (DSM-IV-TR) other than MDD, and/or any substance use disorders other than cocaine or nicotine use disorder; in the healthy control group: any psychiatric disorder in Axis I (DSM-IV-TR), family history of depressive disorder and any substance use disorders (except nicotine use disorder).

The clinical protocol was approved by the local Research Ethical Committee (CEIC-Parc de Salut Mar, Barcelona, Spain) and the study was conducted in accordance with the Declaration of Helsinki and Spanish laws concerning clinical research. Healthy controls were financially compensated. All subjects gave the written, informed consent prior to participation in the study. The neurochemical protocol was also approved by the Ethics Committee of Clinical Investigation (CEIC-CAIB) and developed following the guidelines of the University of the Balearic Islands (UIB).

A close-ended questionnaire was used to record patients’ sociodemographic characteristics, family history, medical assessment including serological status (human immunodeficiency virus and hepatitis C virus) history of substance use and previous psychiatric treatment. Depression severity was evaluated with the Spanish version of the HDRS (25).

Substance and non-substance use disorders were diagnosed according to DSM-IV-TR criteria (23), using the Spanish version of the PRISM-IV (26).

The PRISM is a semi-structured interview designed to evaluate current and life-time DSM-IV-TR disorders and has showed good to excellent validity and test–retest reliability for primary (MDD) and substance-induced major depression with kappa ranging from 0.66 to 0.75 (26, 27).

Patients and healthy controls participated in the ATD test (22), which was performed in a randomized, double-blind, within-subject, crossover and placebo-controlled design. Prior to the study a validation and standardization of the ATD test was performed in healthy controls. The two experimental sessions were conducted with at least a 1-week washout period between them (ATD and non-ATD sessions). In female subjects, the menstrual cycle was controlled and the first session was conducted during the follicular phase to control possible differences in the results due to the involvement of the serotonergic system in hormonal changes.

For the 24 h prior to each session, subjects consumed a low Trp diet (22) and they were fasting on the day of the laboratory sessions. Before starting the sessions, urine was screened for cocaine and other substances to exclude consumption prior to testing. During one session, in randomized form, the subjects were given an amino acid mixture lacking Trp (ATD session; i.e., l-alanine 4.1 g, l-arginine 3.7 g, l-cysteine 2.0 g, glycine 2.4 g, l-histidine 2.4 g, l-isoleucine 6.0 g, l-leucine 10.1 g, l-lysine 6.7 g, l-methionine 2.3 g, l-proline 9.2 g, l-phenylalanine 4.3 g, l-serine 5.2 g, l-threonine 4.9 g, l-tyrosine 5.2 g, and l-valine 6.7 g) and in the other session a similar amino acid mixture but containing Trp (non-ATD; 3.0 g) (19). Blood samples were obtained prior, during and after the test. With this technique, free and total plasma Trp levels are markedly reduced 3–5 h after the depleting Trp drink. A Trp-enriched diet (high Trp/large neutral amino acids, LNAA, ratio), given at 6 h led to the recovery of plasma Trp to basal levels at 24 h (22).

During both sessions, venous blood was obtained by inserting a venous catheter in the non-dominant hand from healthy volunteers (n = 10) and CUD/MDD patients (n = 25, Table 1) before starting the test (basal), at maximum plasma Trp depletion (5 h after induction) and at recovery of plasma Trp levels (24 h after induction) (see Table 1). At each time, aliquots of 5 ml blood from each subject were collected in plastic tubes containing citrate-dextrose solution (ACD, 0.06 mol/l sodium citrate, 0.38 mol/l citric acid, and 0.136 mol/l dextrose) as anticoagulant (8:1 vol/vol). Platelets were isolated by centrifugation (350 × g for 15 min at room temperature) on 5 ml of one-step Human Platelets™ Cell Separation Medium (5 mmol/l Tricine/sodium hydroxide buffer, pH 7.0, containing 12% NycodenzR, and 0.56% sodium chloride; Nycomed Pharma AS, Oslo, Norway). Then, the translucent platelet layer was recovered, mixed with 2 vol of 0.9% NaCl, and recentrifuged (600 g for 5 min) to pellet the platelets that were stored at −80°C until the biochemical assays.

Platelet samples were washed in 500 µl of 0.9% NaCl, isolated by centrifugation (3,000 × g for 15 min at 4°C) and prepared (with minor modifications) for Western blot analysis as described (28, 29). Briefly, the pellet was sonicated in 1 ml of 50 mmol/l Tris buffer, pH 7.5, containing 1 mmol/l ethylenediaminetetracetic acid, 2 mmol/l magnesium chloride (MgCl₂), as well as 10 µl/ml of protease and phosphatase inhibitor cocktails (P8340 and P2850; Sigma-Aldrich). The suspension was centrifuged (16,000 × g for 30 min at 4°C), the resulting pellet (platelet membranes) thoroughly resuspended in protease/phosphatase inhibitor-free Tris buffer. Then, 60 µl of 160 mmol/l Tris, pH 6.8, with 8% sodium dodecyl sulfate (SDS) were added and the mixture heated for 5 min, followed by 5 min sample chilling on ice. Bicinchoninic acid assay was used to determine the sample protein concentration, which was adjusted to 3–4 µg/µl with protease/phosphatase inhibitor-free Tris buffer. Finally, the platelet samples were mixed (9:1, vol/vol) with electrophoresis loading buffer (300 mmol/l Tris, pH 6.8, 20% SDS, 70% glycerol, and 0.03% bromophenol blue) and boiled for 3 min.

The contents of platelet 5-HT_2A receptor forms (monomer and dimer) and IRAS/nischarin (I₁-IR candidate) were assessed by standard SDS-polyacrylamide gel electrophoresis (PAGE)/Western blot procedures as described [e.g., Ref. (28)]. In brief, platelet proteins (10 µg) were separated by SDS–PAGE on 10% polyacrylamide minigels (6 cm × 8 cm; Bio-Rad, CA, USA) and transferred on nitrocellulose membranes (Protran BA 85; Sigma-Aldrich, MO, USA). The membranes were incubated overnight with specific primary antibodies (Ab; dilution range 1/666 to 1/10,000) in blocking solution (PBS containing 5% nonfat-dried milk, 0.5% bovine serum albumin and 0.2% Tween20), followed by 1 h incubation with the appropriate horseradish peroxidase-conjugated secondary antibody (1/5,000 diluted anti-rabbit or anti-mouse Ab, #7074 and #7076; Cell Signaling, MA, USA). The primary Abs used were as follows: anti-5-HT_2A receptor Ab, #RA24288, lot: 401281 (Neuromics, MN, USA); anti-NISCH Ab, #56849, lot: GR111152 (Abcam, UK); anti-β-actin Ab, #A1978, lot: 065M4837V (Sigma-Aldrich).

The 5-HT_2A receptor antibody used [validated in brain tissue of 5-HT_2A receptor KO mice (30)], immunodetected the platelet monomeric receptor as a ~60 kDa band and a ~120 kDa peptide which most probably represents (detected at twice the molecular mass) the homodimeric form of the 5-HT_2A receptor (31, 32) resistant to the denaturalizing assay conditions (see Figure 1). In addition to the possibility of 5-HT_2A receptor heterodimerization (31, 32), more recent studies have demonstrated that this serotonin receptor subtype can form functional homodimers resistant to the reducing agent dithiothreitol (33, 34). The IRAS/nischarin antibody has also been validated in previous studies (15, 16). Enhanced chemiluminescence detection system (Amersham, UK) was used to visualize the bound antibody on autoradiographic film (Amersham) and the resulting immunoreactive bands were quantified (integrated optical density, IOD) by densitometric scanning (GS-800 densitometer; Bio-Rad).

Comorbid CUD/MDD subjects were matched to healthy volunteers (control, C; 2–3 comorbid CUD/MDD subjects per C) (see Table 1) to assess basal differences of target proteins between these groups. In all subject groups, the platelet samples from the ATD test (5 h/Trp depletion, 24 h/Trp recovery) were compared with the corresponding subjects’ basal value in the same gel and the content (IOD) of target proteins was normalized to that of β-actin. Whenever possible, platelet samples (10 µg) for the studied subjects were assessed twice.

The program GraphPad PRISM™, version 6.0 (GraphPad Software, CA, USA) was used for data analysis. All data sets are expressed as mean values ± SEM. Prior to statistical analyses, the parametric data were tested for normality (Kolmogorov–Smirnov normality test) and inspected for possible outliers (Grubb’s test; critical value: Z > 1.96; GraphPad Software at www.graphpad.com/quickcalcs/grubbs1.cfm). Detected outliers for a particular target protein were discarded from further analysis and the final number of cases (healthy controls, CUD/MDD) analyzed for target proteins is indicated in the corresponding figure legend. The experimental design of this study incorporated the pairing of comorbid CUD/MDD subjects with healthy controls, matched for different variables (mainly gender and age, see Table 1), and consequently a statistical test that takes pairing into account was performed. Analysis of covariance indicated that sex distribution (reduced number of females) and age did not influence the results (data not shown). A two-tailed one-sample t-test (with N − 1 degrees of freedom), which compares the experimental mean (percentage of control) with a hypothetical mean of 100, was used for the statistical evaluations (35). The effects of ATD on target proteins in different subject groups were evaluated by Student’s two-tailed t-test. The level of significance was set at p < 0.05 for all statistical evaluations.

The main sociodemographic and clinical characteristics of the sample are described in Table 1. As expected, after the induction of the ATD test, the plasma level of the amino acid Trp was markedly reduced at 5 h (>85%) in all subject groups (p < 0.01 for each group, Table 1).

The basal content of monomeric 5-HT_2A receptor form was markedly reduced in platelets of comorbid CUD/MDD subjects (all patients: 43%), when compared with matched controls (100%), and this receptor down-regulation was independent of the presence or absence of AD medication (decreases in AD−: 47% and in AD+: 40%) (Figure 1A, upper panel). In contrast, the basal content of dimeric 5-HT_2A receptor form in platelets of comorbid CUD/MDD subjects, regardless of AD treatments, was not significantly different to that in control subjects (Figure 1A, lower panel).

When compared with the corresponding platelet 5-HT_2A receptor basal value (100%), ATD had no significant effect on monomeric receptor contents in healthy controls (5 and 24 h; Figure 1B, upper panel). In contrast, ATD further decreased 5-HT_2A receptor density in unmedicated comorbid CUD/MDD subjects (decreases in AD−: 26–46% at 5–24 h) (Figure 1B, upper panel), whereas AD medication was associated with an upregulation of 5-HT_2A monomeric receptor form (increases in AD+: 28–17% at 5–24 h; Figure 1B, upper panel).

In healthy controls, ATD increased the content of platelet 5-HT_2A receptor dimer (44–46% at 5–24 h; Figure 1B, lower panel). In contrast, the dimeric form of platelet 5-HT_2A receptor was found reduced after ATD in unmedicated comorbid CUD/MDD subjects (decreases in AD−: 23–13% at 5–24 h) (Figure 1B, lower panel). The ATD had no significant effects on the contents of 5-HT_2A receptor dimer in comorbid CUD/MDD subjects following AD treatment at 5–24 h (Figure 1B, lower panel).

Similar analyses were performed between patients with primary MDD or CUD-induced MDD Comparison of these two subgroups of subjects revealed no significant differences for the density of platelet monomeric and dimeric 5-HT_2A receptor forms at basal level and during/after ATD (data not shown).

Imidazoline receptor antisera selected/nischarin, a putative I₁-IR, was immunodetected in human platelets as a single 167 kDa protein (Figure 2). Even though no basal differences were found for platelet IRAS/nischarin content in comorbid CUD/MDD subjects when compared with controls (100%, Figure 2A), its expression was regulated in a dissimilar manner in the different subject subgroups throughout the ATD test (Figure 2B). Compared to the corresponding basal value (100%), ATD had no effect on platelet IRAS/nischarin in the case of healthy controls but significantly reduced its expression at 24 h in CUD-induced MDD (40%) subjects (Figure 2B)., Conversely, CUD with primary MDD was associated with increased (11–24%) platelet IRAS/nischarin contents in the ATD test (Figure 2B). This discrepant regulation originated significant differences for the IRAS/nischarin contents at 24 h between the two groups of subjects with CUD, those with primary and those with CUD-induced MDD (Figure 2B). It is to note that these alterations appeared to be independent of AD treatment (data not shown).

Study limitations are noted. The relatively small sample size within subject groups was the main limiting factor. In this study, arising from strict recruitment criteria for patients with CUD without any other substance use disorder than nicotine (especially alcohol use disorder, which frequently coexists with CUD) and patients without active cocaine use in the last 4 weeks. The latter was very difficult to recruit mainly in the CUD-induced MDD group, which has previously been associated with higher relapse rates than Primary MDD (hazard ratios 6.5 vs 2.7 respectively) (44) emphasizing the need to improve diagnosis and appropriate treatment in these patients. Finally, sex differences in biomarkers (45) could be assessed due to the small number of female participants.

Strengths of the study are also noted. The diagnostic procedures that were used (PRISM) has better reliability and validity for drug dependence and MDD diagnoses (primary and substance induced) in comparison to other available procedures (25, 26), but its systematic use in clinical practice is difficult because it needs well trained professionals and lasts about 1–3 h to be administrated. PRISM is a good instrument for research, but not for usual clinical practice. Thus, there is a need of biomarkers that can facilitate the diagnosis in an easy and valid way. Also, these biomarkers would improve the knowledge of neurobiological mechanisms of both types of depression associated to CUD and facilitate the appropriate treatment. In addition, comparable results (Trp depletion: 86–87%) have been reported in studies using a similar protocol in pure MDD (19, 33).