Blood serum samples were obtained from SCZ patients (n = 26) and non-psychiatric controls (n = 13). Patients with schizophrenia were recruited at the A.O.U. “Federico II” hospital of Naples over 6 months and diagnosed according to the Diagnostic and Statistical Manual of mental disorders, Fifth Edition (DSM-5) (80). Inclusion criteria for patients were: age 18–60 years; no evidence of worsening psychotic symptoms in the previous 6 months; absence of other major systemic, psychiatric (e.g., addictive disorders, frequent substance use in the 6 months prior to enrollment, etc.), or neurological disorders. Healthy controls were sex-matched individuals with no history of neurological, psychiatric, or systemic conditions or family psychiatric history. SCZ patients were divided into two groups according to treatment resistance: non-treatment-resistant schizophrenia (nTRS; n = 13) and treatment-resistant schizophrenia (TRS; n = 13). The treatment resistance condition was defined as a failure of at least two different antipsychotic regimens, each administered for > 6 weeks and at an optimal dose, according to the modified Treatment Response and Resistance in Psychosis Working Group Consensus criteria (81). All TRS patients were under treatment with clozapine while nTRS patients were treated with different conventional antipsychotics, such as olanzapine, risperidone, haloperidol, amisulpride, promazine, paliperidone and aripiprazole. Clinical data were collected within 1 month from the blood sample and included the severity of psychotic symptoms measured by the Positive and Negative Syndrome Scale (PANSS) (82) and cognitive performances assessed by the Brief Assessment of Cognition in Schizophrenia (BACS) (83). Demographic characteristics are reported in Table 1 . Phlebotomy was conducted by a psychiatric nurse; collection was performed in fasting status, in the morning before breakfast. Serum was separated by centrifugation and stored at −80°C until analysis. Written informed consent was obtained from all subjects, according to the Declaration of Helsinki. The study was approved by the Ethics Committee of the University “Federico II” of Naples (protocol number: 195/19).

Blood serum samples were obtained from two different Italian hospitals: A.O.U. “Federico II”, Naples, Italy (ASD, n = 33; Control, n = 6) and Istituto Giannina Gaslini, Genoa, Italy (ASD, n = 20; Control, n = 24). Participants from A.O.U. “Federico II” were consecutive samples of children and adolescents, along 6 months, referred to the Department of Pediatrics — Unit of Child and Adolescent Neuropsychiatry, for an evaluation in a clinical hypothesis or revaluation of ASD. All the subjects received a full assessment, including a complete history (pregnancy, childbirth, psychomotor development), structured clinical interviews and validated observations [Autism Diagnostic Observation Schedule-2 (84), Griffiths Mental Development Scale (85) or Leiter International Performance Test-Revised (86), Vineland Adaptive Behavior Scales—II edition (87)]. Diagnosis of ASD was formulated according to DSM-5 (80).

About 60 subjects were evaluated; study participants included 33/60 ASD subjects, whose parents signed an informed consent form to participate in the study. Subjects were aged between 18 and 189 months, both males (n = 28) and females (n = 5). Inclusion criteria were a clinical diagnosis of ASD, less than 18 years of age; exclusion criteria included: epilepsy diagnosis or other neurological disorders; psychiatric comorbidity (e.g. obsessive-compulsive disorder, psychosis, etc.), other chronic diseases (e.g. chronic intestinal diseases, malabsorption, etc.).

Six healthy typically developed subjects were recruited as a control group; inclusion criteria were the absence of psychiatric diagnosis, less than 18 years of age. For the control group, the same exclusion criteria were used.

The enrolled subjects followed routine clinical procedures for outpatients, from which data were collected. Each patient was also investigated by blood samples, as per routine procedures during clinical evaluation. Blood samples were collected in the hospital for both ASD and control children. Phlebotomy was conducted by a pediatric nurse, collection was made in fasting status, in the morning before breakfast. Serum was separated by centrifugation and stored at −80°C until analysis. The study was conducted according to the principles of the Declaration of Helsinki; ethical approval was obtained by the Ethics Committee of the University Federico II of Naples (220/18). Written informed consent was collected from parents or legal guardians of enrolled children for both clinical information collection and data acquisition and treatment.

Participants from Istituto Giannina Gaslini were consecutive samples of children and adolescents, referred to the Child Neuropsychiatry Unit Day Hospital for a third-level neuroradiological, biochemical, metabolic and genetic evaluation in a clinical diagnosis of ASD. All the subjects received a full assessment, including a complete history (pregnancy, childbirth, psychomotor development), structured clinical interviews and validated observations [Autism Diagnostic Observation Schedule-2 (84), Autism Diagnostic Interview (Lord et al., 1994), Griffiths Mental Development Scale (85), Wechsler Intelligence Scale for Children – fourth edition (Wechsler D., 2003) or the Wechsler Preschool and Primary Scale of Intelligence - III edition - WPPSI-III (Wechsler D., 2002), Vineland Adaptive Behavior Scales—II edition (87)]. Diagnosis of ASD was formulated according to DSM-5 (80).

Study participants included 20 ASD subjects aged between 3 years and 6 months and 11 years and 4 months, 19 males and 1 female. Inclusion criteria were a clinical diagnosis of ASD, less than 18 years of age; while exclusion criteria were the presence of other psychiatric diagnosis, epilepsy or other chronic diseases. Twenty-four developing normal children were recruited as a control group; inclusion criteria were the absence of psychiatric diagnosis, less than 18 years of age. For the control group, the same exclusion criteria were used.

Blood collection was made in fasting status, in the morning. Serum was separated by centrifugation and stored at −80°C until analysis. The study was conducted according to the principles of the Declaration of Helsinki; ethical approval was obtained by the Ethics Committee of the Liguria Region (N. CET - Liguria: 437/2023 - DB id 13411). Demographic characteristics of ASD and control individuals from both hospitals’ cohorts, such as age and sex distribution, are reported in Table 2 . The hospitals involved in the study were chosen for the presence of SCZ outpatient clinics including a referral center for treatment-resistant psychosis (adult patients at University Federico II - Psychiatry Section) and for being regional centers of child psychiatry referral (Child Neuropsychiatry at Istituto Giannina Gaslini of Genoa and A.O.U. Federico II Neuropsychiatry Section of Naples).

Serum samples were mixed in a 1:10 dilution with HPLC-grade methanol (900 µL) and centrifuged at 13,000 x g for 10 min. Supernatants were dried at 45°C and suspended in 0.2 M trichloroacetic acid (TCA). Samples were then neutralized with 0.2 M NaOH and subjected to pre-column derivatization with o-phthaldialdehyde/N-acetyl-L-cysteine in 50% methanol. To resolve diastereoisomer derivatives, two types of columns were used: a ZORBAX Eclipse Plus C8 5-μm reversed-phase column (Agilent, 4.6x250 mm) and a Symmetry C8 5 μm reversed-phase column (Waters, 4.6x250mm); the separation was performed under isocratic conditions (0.1 M sodium acetate buffer, pH 6.2, 1% tetrahydrofuran, 1 mL/min flow rate). A washing step in 0.1 M sodium acetate buffer, 3% tetrahydrofuran and 47% acetonitrile, was performed after every single run. Identification and quantification of D-Asp, L-aspartate (L-Asp), L-glutamate (L-Glu), L-asparagine (L-Asn), D-Ser, L-serine (L-Ser), L-glutamine (L-Gln) and glycine (Gly) were based on retention times and peak areas and then compared with those associated with external standards ( Figure 1A ). Peak’s identity was confirmed by the selective degradation of the D-enantiomers by RgDAAO M213R variant ( Figure 1A ) (88). Ten μg of the enzyme was added to the samples, incubated at 30°C for at least 3 h, and subsequently derivatized. Amino acids concentration in the serum were expressed as µM. D-amino acid/total amino acid ratio was expressed as a percentage (%). Quantification of enantiomers was based on peak areas using calibration curves for each enantiomer.

Data for clinical characteristics are reported as medians with the respective interquartile range (first-third quartile). The hypothesis of normality was assessed with the Shapiro-Wilk test. For variables not normally distributed, statistical analyses were performed using their log-transformed values. To identify potential confounders, we compared the demographic characteristics between patients and controls using two-sample (two-tailed) t-tests for age and Chi-square tests with Yates’ correction for sex. ANCOVA models (controlling for statistically different covariates between groups) were adopted to assess significant differences in amino acid concentrations between cases and controls. The p-values from ANCOVA models were corrected for multiple testing, following Bonferroni’s method. ANCOVAs were followed by Tukey post-hoc comparisons. To estimate the effect magnitude of significant outcomes, we computed the partial eta square (η²p), which provides a quantifiable measure of the proportion of variance in the dependent variable that is associated with an independent variable, while controlling for other variables. For Chi-square and t-test statistics, the Cohen’s w and d were computed as effect sizes, respectively. All statistical analyses were conducted with RStudio R version 4.1.2.

We recruited SCZ patients (n = 26) and non-psychiatric control subjects (n = 13) in A.O.U. “Federico II” Hospital ( Figure 1 ; Table 1 ) to measure by HPLC the serum levels of L-Glu and Gly, which in the brain represent, respectively, the main excitatory amino acid and a major NMDAR co-agonist (together with D-Ser), as well as L-Gln and L-Asn, which participate to L-Glu and L-Asp biosynthesis, respectively. Specifically, SCZ patients were subdivided into nTRS and TRS groups (n = 13/condition) based on the assessment of symptoms persistence after at least two different conventional antipsychotic regimens (see Materials and Methods). No statistically significant differences were found in sex [χ² (2, N = 39) = 5.0556, p = 0.0798, Cohen’s w = 0.36], while significant age-dependent variations were observed [median (min; max) of years: Ctrl = 28 (21; 40), nTRS = 47 (22; 60), TRS = 34 (25; 57), F(2, 36) = 10.688, p = 0.0002, η²p = 0.37, one-way ANOVA; Table 1 ] among groups. Based on this, to assess the changes in amino acid levels, we used an ANCOVA model considering the effect of age as a confounding factor.

We found significant differences in L-Glu serum content among groups [F(2, 35) = 5.552, p = 0.0082, η²p = 0.24], evidencing an increase of this amino acid in both nTRS and TRS patients, compared to controls ( Figures 1A, B ; Supplementary Table 1 ). However, such L-Glu serum variation did not survive after correction with the Bonferroni multiple comparisons method ( Supplementary Table 1 ). Similarly, no alterations among groups were found for L-Gln levels, L-Gln/L-Glu ratio, L-Asn and Gly levels ( Figures 1A, C–F ; Supplementary Table 1 ). Overall, our analysis revealed no significant changes in L-Glu, L-Gln, L-Asn and Gly serum levels, as well as L-Gln/L-Glu serum ratio, in both nTRS and TRS patients, compared to their non-psychiatric controls.

After we measured the levels of D-Ser, D-Asp and their respective precursors, L-Ser and L-Asp, the latter being also one of the major NMDAR agonists in the CNS. Statistical analysis revealed significant alteration in D-Asp levels in nTRS and TRS patients, compared to control individuals [F(2, 35) = 8.397, p = 0.001, η²p = 0.324], which survived also after correction with Bonferroni multiple comparisons ( Figures 2A ; Supplementary Table 1 ). The following Tukey post-hoc test highlighted that both nTRS and TRS patients displayed significantly reduced D-Asp levels, compared to controls (Ctrl vs nTRS, p = 0.0104; Ctrl vs TRS, p = 0.0177), while no significant alteration was observed between nTRS and TRS groups (TRS vs nTRS, p = 0.976) ( Figure 2A ; Supplementary Table 1 ). Conversely, we found comparable levels of L-Asp [F(2, 35) = 2.794, p = 0.0749] and D-Asp/total Asp ratio [F(2, 35) = 0.476, p = 0.6255] among nTRS, TRS and control subjects ( Figures 2B, C ; Supplementary Table 1 ).

Statistical analysis also revealed significant differences in D-Ser serum levels among nTRS, TRS patients and control subjects [F(2, 35) = 6.322, p = 0.0045, η²p = 0.265; Figure 2D ; Supplementary Table 1 ), which was confirmed after correction with Bonferroni multiple comparisons ( Supplementary Table 1 ). The following Tukey post-hoc analysis evidenced a significant decrease of D-Ser in TRS but not in nTRS patients, compared to controls (Ctrl vs nTRS, p = 0.1078; Ctrl vs TRS, p = 0.0103; Figure 2D ; Supplementary Table 1 ). However, no D-Ser changes were found between TRS and nTRS patients (TRS vs nTRS, p = 0.569; Figure 2D ; Supplementary Table 1 ). Also in this case, the deregulation was confined to the D-enantiomer levels, as L-Ser levels did not significantly change among groups [F(2, 35) = 2.900, p = 0.0683; Figure 2E ; Supplementary Table 1 ]. Despite the decrease being specific for the D-Ser, we found an unaltered D-Ser/total Ser ratio [F(2, 35) = 1.757, p = 0.1875; Figure 2F ; Supplementary Table 1 ].

Altogether, our analyses showed selective reductions in D-Asp serum levels in both nTRS and TRS patients, and in D-Ser serum levels only in TRS group, compared to non-psychiatric control subjects. Conversely, no alterations were found in their respective L-enantiomers, L-Asp and L-Ser, among nTRS, TRS and control individuals.

Then we measured the levels of the same neuroactive amino acids and their precursors in the serum of pediatric ASD patients and control subjects recruited in two different Italian Hospitals (Istituto Giannina Gaslini: ASD, n = 20, Ctrl, n = 24; A.O.U. “Federico II”: ASD, n = 33; Ctrl, n = 6). First, we analyzed the cohort of ASD patients and control subjects from Istituto Giannina Gaslini. Before proceeding with statistical comparisons, we assessed potential imbalance in clinical variables, such as sex and age, between ASD patients and control individuals. No statistically significant differences were found in sex (χ² (1, N = 44) = 0.009, p = 0.926), while significant variations between groups were observed in age [median (min; max) of years: Ctrl = 9.8 (3.1; 24.8) vs ASD = 6.1 (3.5; 11.3), t(42) = -3.414, p = 0.001, Cohen’s d = -1.03; Student’s t test] ( Table 2 ). Based on this, to assess amino acid variations between ASD and control subjects, we used ANCOVA model, including age as a confounder. Statistical analysis revealed no significant alterations in L-Glu [F(1, 41) = 0.769, p = 0.3855] and L-Gln [F(1, 41) = 2.127, p = 0.1524] levels, L-Gln/L-Glu ratio [F(1, 41) = 0.285, p = 0.5961], as well as L-Asn [F(1, 41) = 1.785; p = 0.1889] and Gly [F(1, 41) = 2.143; p = 0.1508] levels between the two diagnosis groups ( Figures 3A–E , Supplementary Table 2 ).

Then, we analyzed the cohort of ASD patients and control subjects recruited from A.O.U. “Federico II” hospital. We observed significant differences in both sex (χ² (1, N = 39) = 9.060; p = 0.003, Cohen’s w = 0.56) and age [median (min; max) of years: Ctrl = 13 (5; 17) vs ASD = 5.4 (0.4; 15.7); t(36) = -3.075, p = 0.004, Cohen’s d = -1.37; Student’s t test] ( Table 2 ). For this reason, we evaluated the differences between groups by ANCOVA models, including both age and sex as confounders. HPLC analysis revealed unaltered levels of each of the analyzed amino acids [L-Glu: F(1, 34) = 3.307, p = 0.0778; L-Gln: F(1, 34) = 1.504, p = 0.2285; L-Gln/L-Glu ratio: F(1, 34) = 0.170, p = 0.6825; L-Asn F(1, 34) = 1.188, p = 0.2834; Gly: F(1, 34) = 0.044, p = 0.8346] ( Figures 3F–J , Supplementary Table 2 ) in ASD patients, compared to control subjects. Collectively, our results show no significant alterations in L-Glu, L-Gln, L-Asn and Gly serum levels, as well as L-Gln/L-Glu serum ratio, in both cohorts of ASD patients analyzed, compared to their respective control individuals.

Finally, we measured the serum content of D-Ser, L-Ser, D-Asp and L-Asp. Statistical analysis in ASD patients and control subjects from Istituto Giannina Gaslini revealed no significant differences in D-Asp (F(1, 41) = 0.046, p = 0.8317), L-Asp (F(1, 41) = 1.443, p = 0.2366), and D-Asp/total Asp ratio (F(1, 41) = 0.561, p = 0.4583) between ASD patients and control subjects ( Figures 4A–C , Supplementary Table 2 ). Likewise, we found comparable levels between diagnoses also for D-Ser (F(1, 41) = 0.235, p = 0.6301), L-Ser (F(1, 41) = 0.659, p = 0.4218) and D-Ser/total Ser ratio (F(1, 41) = 0.058, p = 0.8113) ( Figures 4D–F , Supplementary Table 2 ). Finally, we analyzed ASD patients and control subjects from A.O.U. “Federico II”. We found unaltered D-Asp levels between diagnoses (F(1, 41) = 0.036, p = 0.8507; Figure 4G , Supplementary Table 2 ). Conversely, we showed a slight but significant L-Asp levels increase in ASD patients, compared to controls (F(1, 34) = 4.449, p = 0.0424; η²p = 0.12; Figure 4H , Supplementary Table 2 ), which did not survive after correction with Bonferroni multiple comparisons method ( Supplementary Table 2 ). In line with unaltered D-Asp and L-Asp levels, D-Asp/total Asp ratio did not significantly change between diagnosis groups (F(1, 34) = 0.0001 p = 0.9954; Figure 4I , Supplementary Table 2 ). Again, ANCOVA analysis revealed comparable D-Ser and L-Ser levels [F(1, 34) = 0.027, p = 0.8704], L-Ser [F(1, 34) = 3.132, p = 0.0858], and D-Ser/total Ser ratio [F(1, 34) = 1.347; p = 0.2539] ( Figures 4J–L , Supplementary Table 2 ) between ASD patients and control individuals. Our overall HPLC analysis revealed no significant changes in D-Asp, L-Asp, D-Ser and L-Ser serum levels, as well as D-Asp/total Asp and D-Ser/total Ser serum ratios, in both cohorts of ASD patients analyzed, compared to their respective control individuals.