Next Article in Journal
PPA1 Deficiency Causes a Deranged Galactose Metabolism Recognizable in Neonatal Screening
Previous Article in Journal
Matrix Selection for the Visualization of Small Molecules and Lipids in Brain Tumors Using Untargeted MALDI-TOF Mass Spectrometry Imaging
 
 
Search for Articles:
Title / Keyword
Author / Affiliation / Email
Journal
Article Type
 
 
Advanced Search
 
Section
Special Issue
Volume
Issue
Number
Page
 
Journals
Metabolites
Volume 13
Issue 11
10.3390/metabo13111140
metabolites-logo
Submit to this Journal Review for this Journal Propose a Special Issue
► Article Menu

Article Menu

share Share announcement Help format_quote Cite question_answer Discuss in SciProfiles thumb_up
...
Endorse textsms
...
Comment
first_page
settings
Order Article Reprints
Font Type:
Arial Georgia Verdana
Font Size:
Aa Aa Aa
Line Spacing:
Column Width:
Background:
Article

Cord Blood Metabolite Profiles and Their Association with Autistic Traits in Childhood

by
Christin S. Kaupper
1,2,
Sophia M. Blaauwendraad
1,2,
Charlotte A. M. Cecil
3,4,5,
Rosa H. Mulder
1,3,
Romy Gaillard
1,2,
Romy Goncalves
1,2,
Ingo Borggraefe
6,
Berthold Koletzko
7 and
Vincent W. V. Jaddoe
1,2,*
1
The Generation R Study Group, Erasmus MC, University Medical Center Rotterdam, 3000 CA Rotterdam, The Netherlands
2
Department of Pediatrics, Sophia’s Children’s Hospital, Erasmus MC, University Medical Center Rotterdam, 3000 CA Rotterdam, The Netherlands
3
Department of Child and Adolescent Psychiatry, Erasmus MC, University Medical Center Rotterdam, 3000 CA Rotterdam, The Netherlands
4
Department of Epidemiology, Erasmus MC, 3000 CA Rotterdam, The Netherlands
5
Molecular Epidemiology, Department of Biomedical Data Sciences, Leiden University Medical Center, 2333 ZC Leiden, The Netherlands
6
Division of Pediatric Neurology, Developmental Medicine and Social Pediatrics, Comprehensive Epilepsy Center for Children and Adolescents, Dr. von Hauner Children’s Hospital, LMU University Hospitals, LMU—Ludwig-Maximilians Universität, 80337 Munich, Germany
7
Division of Metabolic and Nutritional Medicine, Department of Pediatrics, Dr. von Hauner Children’s Hospital, LMU University Hospitals, LMU—Ludwig-Maximilians Universität, 80337 Munich, Germany
*
Author to whom correspondence should be addressed.
Metabolites 2023, 13(11), 1140; https://doi.org/10.3390/metabo13111140
Submission received: 28 September 2023 / Revised: 2 November 2023 / Accepted: 3 November 2023 / Published: 9 November 2023
Browse Figures
Review Reports Versions Notes

Abstract

:
Autism Spectrum Disorder (ASD) is a diverse neurodevelopmental condition. Gene–environmental interactions in early stages of life might alter metabolic pathways, possibly contributing to ASD pathophysiology. Metabolomics may serve as a tool to identify underlying metabolic mechanisms contributing to ASD phenotype and could help to unravel its complex etiology. In a population-based, prospective cohort study among 783 mother–child pairs, cord blood serum concentrations of amino acids, non-esterified fatty acids, phospholipids, and carnitines were obtained using liquid chromatography coupled with tandem mass spectrometry. Autistic traits were measured at the children’s ages of 6 (n = 716) and 13 (n = 648) years using the parent-reported Social Responsiveness Scale. Lower cord blood concentrations of SM.C.39.2 and NEFA16:1/16:0 were associated with higher autistic traits among 6-year-old children, adjusted for sex and age at outcome. After more stringent adjustment for confounders, no significant associations of cord blood metabolites and autistic traits at ages 6 and 13 were detected. Differences in lipid metabolism (SM and NEFA) might be involved in ASD-related pathways and are worth further investigation.

1. Introduction

Autism Spectrum Disorder (ASD) is a diverse neurodevelopmental condition characterized by difficulties in social communication and interaction as well as restrictive and repetitive behavior, interests or activities [1]. Increasing evidence indicates that ASD represents the extreme end of a continuum of autistic traits that are present in the general population [2,3]. The biological mechanisms underlying ASD remain unclear. However, a growing body of evidence suggests that the interplay of genetic and environmental factors may trigger immune and inflammatory responses in early stages of life, which could play a key role in pathophysiological processes leading to ASD [4,5]. These processes might result in metabolic dysregulations that can be observed in cord blood, detectable through metabolomics prior to the manifestation of autistic traits in early childhood. Multiple cross-sectional metabolomics studies among pre-school children with an ASD diagnosis and typically developing children have been conducted and suggested that mitochondrial dysfunction, immune dysregulation, as well as altered amino acid, lipid, and neurotransmitter metabolism could be associated with ASD [6,7,8,9,10,11,12,13,14,15,16,17,18]. Only one previous study assessed the prospective associations of stored dried blood spot metabolites in a small group of 37 newborns that received an ASD diagnosis in childhood and in healthy controls. They did not detect significant differences in metabolite concentrations [19]. The majority of previous studies have been based on small, selected clinical samples using a cross-sectional design, which precluded the possibility of assessing the temporality of associations. In addition, given that social impairments related to ASD exist on a spectrum that extends into the typical range of physiological variations, differences in metabolic composition might also be detectable in fetuses that develop sub-clinical autistic traits in childhood.
We hypothesize that changes in the blood metabolite composition due to early-life metabolic adaptations may be associated with autistic traits in the general population. We tested this hypothesis on 783 children participating in a population-based prospective cohort study from early fetal life onwards. We used a targeted metabolomics approach for amino acids (AA), non-esterified fatty acids (NEFA), phospholipids (PL), sphingomyelins (SM), and carnitines (Carn) from cord blood. These metabolites play an important role in lipid metabolism, signal transduction, oxidative stress, inflammatory responses, and mitochondrial as well as neurotransmitter metabolism and might therefore be involved in ASD pathophysiology [20,21,22,23]. Autistic traits were measured dimensionally at the ages 6 and 13 with the Social Responsiveness Scale.

2. Methods

2.1. Study Design

This study was embedded in the Generation R Study, a prospective population-based cohort from fetal life onwards, designed to identify early environmental and genetic determinants of growth and development. Pregnant women who resided in the Rotterdam area and delivered between April 2002 and January 2006 were eligible for participation. Study approval was obtained by the Medical Ethical Committee of the Erasmus University Medical Center, Rotterdam (MEC 198.782/2001/31). Written informed consent was obtained from all mothers. Umbilical cord blood metabolites were assessed in a subgroup of 921 mother–infant pairs. Of those, 716 and 648 had information on autistic traits at 6 and 13 years, respectively, and 581 had information at both 6 and 13 years (see Flowchart in Figure S1).

2.2. Metabolomics Measurements

Umbilical venous cord blood samples were collected by a midwife or an obstetrician (median gestational age at birth: 40.3 weeks; (95% range: 36.9, 42.4)) after birth [24]. Blood samples were transported to the regional laboratory (STAR-MDC), spun, and stored at −80 °C within a maximum of 4 h after collection. For metabolite measurements, the samples were transported on dry ice to the Division of Metabolic and Nutritional Medicine, Dr. von Hauner Children’s Hospital, LMU Munich. There, a targeted metabolomics approach was performed in order to determine the serum concentrations of 193 metabolites (μmol/L), AA, NEFA, PL (including diacyl-lysophosphatidylcholines (PC.aa), acyl-alkyl-lysophosphatidylcholines (PC.ae), acyl-lysophosphatidylcholines (Lyso.PC.a), alkyl-lysophosphatidylcholines (Lyso.PC.e)), SM, and Carn (including free carnitine (Free Carn) and acylcarnitines (Carn.a)). The metabolomics analysis is described in detail elsewhere [24]. In brief, a 1100 high-performance liquid chromatography (HPLC) system (Agilent, Waldbronn, Germany) coupled with a API2000 tandem mass spectrometer (AB Sciex, Darmstadt, Germany) analyzed AA [25]. For the amino acid notation, IUPAC-IUB nomenclature was used [26]. The measurement of PL, NEFA, and Carn was conducted with a 1200 SL HPLC system (Agilent, Waldbronn, Germany) coupled to a 4000QTRAP tandem mass spectrometer from AB Sciex (Darmstadt, Germany) [27,28]. The analytical measurements enable the distinguishment of the total number of double bonds, however, not their position or the distribution of the C-atoms between the fatty acid side chains. For NEFA, PL, and Carn.a, the following notation was used: ‘X:Y’, ‘X’ designates the number of C-atoms of the carbon chains, and ‘Y’ is the total number of double bonds. The ‘a’ is a synonym for acyl chain bound to the backbone via an ester bond (‘acyl-’), whereas ‘e’ stands for an ether bond (‘alkyl’), respectively. For precise measurements, six quality control (QC) samples per batch were consistently measured between study samples. Outliers were excluded, and the coefficients of variation (CV; standard deviation/mean) for each batch (intra-batch) and for all batches (inter-batch) of the QC samples were calculated for each metabolite. We excluded batches with an intra-batch CV higher than 25%, which is in line with previous studies [29,30,31]. Complete metabolite data was excluded for metabolites that had an inter-batch CV higher than 35% or if less than 50% of the batches had passed the QC. The correction of batch effects was carried out by dividing metabolite concentrations by the ratio of intra-batch median and the inter-batch median of the QC samples [31]. If metabolites or participants missed more than 50% of the values, they were excluded [30]. Missing metabolite values of the remaining metabolites and participants were imputed using the Random Forest algorithm (R package missForest) [32]. Further information on the parameters for mass-spectrometry detection and identification of the metabolites can be found in Supplementary Text S1 and Supplementary Table S10.
For the data analysis, individual metabolites were clustered in metabolite groups (AA, NEFA, PC.aa, PC.ae, Lyso.PC.a, Lyso.PC.e, SM, Free Carn, and Carn.a), as well as in metabolite subgroups based on chemical structure and biological relevance (AA: branched-chain amino acids (BCAA), aromatic amino acids (AAA), essential amino acids (EAA), non-essential amino acids (NEAA); NEFA; PC.aa; PC.ae; Lyso.PC.a; Lyso.PC.e; SM: saturated, mono-unsaturated, poly-unsaturated; Carn.a: short-chain, medium-chain, long-chain) [24]. The sum of individual metabolite concentrations per metabolite group was calculated. Moreover, we computed the following ratios: AA ratios, asparagine/aspartic acid (Asn/Asp), and glutamine/glutamic acid (Gln/Glu) as indicators for anaplerosis or replenishing of citric acid cycle metabolites; NEFA.18:1/NEFA.18:0 and NEFA.16:1/NEFA.16:0 ratios as markers of stearoyl-CoA desaturase-1 activity, which is associated with increased fat accumulation and reduced fatty acid oxidation; ΣPC.aa/ΣPC.ae, reflecting oxidative stress, ΣLyso.PC.a/ΣPC.aa as a lipid biomarker of inflammation; (lyso.PC.a.C16:0 + lsyo.PC.a.C18:0)/ΣPC.aa as a proinflammatory biomarker; (lyso.PC.a.C18:1 + lyso.PC.a.C18:2)/ΣPC.aa as an anti-inflammatory biomarker; Carn.a ratios, Carn.a.C16.0/free Carn and Carn.a.C2:0/Carn.a.C16:0 as markers of Carn palmitoyl transferase-1 activity and fatty acid ß-oxidation [33,34,35,36,37,38]. Individual metabolite concentrations, sums, and ratios were square root transformed to obtain normally distributed metabolite concentrations. In order to enable comparison of the effect estimates, individual metabolite concentrations and sums were standardized by calculating standard deviation scores.

2.3. Autistic Traits

At the ages of 6 and 13 years, the primary caregiver of the participating children filled out a short version of the Social Responsiveness Scale (SRS). The SRS is a validated instrument for assessing autistic traits on a continuous scale [39]. It represents the caregiver’s observation of the child’s social behavior during the previous 6 months. The SRS is a useful screening tool to identify children who need further ASD-specific diagnostics [40]. The abridged, 18-item version of the SRS has been shown to be an effective tool for generating results consistent with the full version of the SRS [41]. It is designed for children between the ages of 4 and 18. Each item on the SRS is rated from 0 (never true) to 3 (almost always true), covering social, language, and repetitive behavior, with higher scores indicating greater social impairment [39]. Individual item scores for the SRS were summed and weighted by the number of items completed. We used the SRS as a continuous measure of autistic traits in our analyses.

2.4. Covariates

During pregnancy, data on maternal age at enrollment, educational level (higher education yes/no), smoking (non-smoking, smoked until pregnancy was known, continued smoking during pregnancy), alcohol intake (no intake, until pregnancy was known, continued during pregnancy), and folic acid supplement (yes/no) during pregnancy was gathered using questionnaires. A start of folic acid supplementation during the first 10 weeks of pregnancy or preconceptionally was considered positive. Maternal weight and height at enrollment were measured without shoes, and BMI was calculated. Maternal psychopathology during pregnancy was assessed using the Brief Symptom Inventory (BSI). Maternal vitamin D levels were obtained from serum in pregnancy. Vitamin D levels below 50 nmol/L were considered deficient. Information on sex, gestational age at birth, and birth weight according to European growth charts was obtained from medical records. We calculated gestational age and sex-adjusted birth weight standard deviation scores according to growth charts of Niklasson [42]. Data on the children’s age at outcome was obtained when the primary caregiver filled out the SRS.

2.5. Statistical Analysis

First, we conducted a non-response analysis. We compared mother–child pairs of our subsample with all mother–child pairs with metabolomics data available to identify and take into account possible biases caused by loss to follow-up or exclusion criteria (Supplementary Table S1). We used a Chi-squared test for categorical variables and a Student’s t-test or Mann–Whitney-U test for normally distributed or skewed continuous variables. We explored the correlation between the SRS scores at 6 and 13 years using Spearman’s correlation. Then, we examined the associations of neonatal metabolite ratios, groups, and individual metabolites with the continuous SRS scores at the ages of 6 and 13 using linear regression models. We chose this approach over data reduction or variable selection analysis methods like Principal Component Analysis because our primary goal was to assess the associations of the individual metabolites using regression models on the outcome autistic traits rather than identifying clusters of metabolites that were jointly associated with the outcome. Analyses were only adjusted for child sex and age at outcome in the basic model and additionally adjusted for our preselected confounders in the main model. Confounders were selected based on previous studies and included maternal age, pre-pregnancy BMI, maternal education level, maternal psychopathology, smoking during pregnancy, alcohol consumption during pregnancy, gestational age at birth, and birth weight [13,18,19,43,44]. Several sensitivity analyses were performed. First, we repeated our main analysis, additionally adjusting for vitamin D levels and folic acid supplementation as previous studies suggested associations of folic acid as well as vitamin D deficiency with autistic traits [4]. Second, we explored the influence of individual confounding variables by separately adjusting for each confounder.
As an exploratory analysis, we ran a linear mixed-effects model to investigate whether changes in autistic traits over time are associated with certain metabolites. Models were fit using the lme4 package [45]. Linear mixed-effects models contain numerous positive features, including modeling of random effects and handling of missing time points.
To account for multiple testing, we applied Benjamini–Hochberg correction separately for each model, using an overall false discovery rate (FDR) adjusted p-value of <0.05. To account for missing values of covariates, we performed multiple imputations using the fully conditional specification method, and pooled results from 25 imputed datasets were reported [46]. The proportion of missing values ranged from 6% (vitamin D deficiency) to 17% (folic acid supplement). All statistical tests were 2-sided. Statistical analyses were conducted using R statistical software version 4.2.2 (R Foundation for Statistical Computing, Vienna, Austria).

3. Results

3.1. Population Characteristics

The mean maternal age at enrollment was 31.8 (±3.9) years, and the median body mass index was 22.4 (kg/m2) (Table 1). Most women were nulliparous (62%), highly educated (66%), did not smoke (79%), and continued drinking alcohol (56%) in pregnancy. A non-response analysis showed that mothers of children included in our analysis were older, had a higher education level, smoked less, drank less alcohol, and used more folic acid supplements during pregnancy as compared to mother–child pairs with metabolomics data available who did not answer the SRS questionnaire (Table S1). Spearman correlation between SRS scores at age 6 and at age 13 suggested a moderate (intercept of 0.58) correlation. Median concentrations of cord blood metabolite groups, individual metabolites, and metabolite ratios are shown in Table S2.

3.2. Neonatal Cord Blood Metabolomics and Autistic Traits at the Ages of 6 and 13

Higher cord blood SM.a.C39.2 concentrations were significantly associated with a 0.6 lower SRS score at 6 years (95% CI (−0.9, −0.31) per SDS increase in SM.a.39.2, FDR-adjusted p-value 0.01) in the basic model (Figure 1A, Table 2), but not in the main model (0.58 lower SRS score (95% CI (−0.84, −0.24) per SDS increase, FDR-adjusted p-value 0.08) (Figure 1B, Table 2). The NEFA.16:1/16:0 ratio was inversely associated with SRS scores in the basic model (0.58 lower SRS score (95% CI (−0.83, −0.26)) per SDS increase in NEFA.16:1/16:0 ratio, FDR-adjusted p-value: 0.02) (Figure 1A, Table 2) but did not remain significant in the main model after multiple testing correction (0.48 lower SRS score (95% CI (−0.77, −0.19)) per SDS increase in NEFA.16:1/16:0 ratio, FDR-adjusted p-value: 0.08) (Table 2). We observed that higher Carn.a.C18.2 concentrations were associated with higher SRS scores at age 6 in the basic (0.48 higher SRS score (95% CI (0.19, 0.77)) per SDS increase in Carn.a.C18.2, FDR-adjusted p-value 0.06) and main model (0.51 higher SRS score (95% CI (0.22, 0.8)) per SDS increase in Carn.a.C18.2, FDR-adjusted p-value 0.08), although it did not survive multiple testing (Table 2). No significant associations between total metabolite ratios, groups, individual concentrations, and autistic traits at 13 years were observed (Figure S1A,B, Tables S3, S4, S6 and S7)). Linear mixed-effects analyses did not yield significant associations between changes in autistic traits over time and metabolites (Table S9).

3.3. Sensitivity Analysis

The additional adjustment for vitamin D deficiency and folic acid supplementation did not influence the associations between cord blood metabolites and SRS scores at age 6 or 13 years (Tables S5 and S8). Adjusting for every confounder separately showed that gestational age had the strongest influence on the association between cord-blood metabolites and autistic traits.

4. Discussion

In this prospective cohort study, we observed some suggestive evidence for associations between SM, NEFA ratio, and autistic traits at age 6 years, but these associations did not remain significant after multiple testing correction and more rigorous covariate adjustment. No significant associations between cord blood metabolites and autistic traits at the ages of 6 and 13 were observed in our main model. Moreover, no effect modification by vitamin D deficiency or folic acid deficiency was observed. Controlling for gestational age at birth showed the biggest effect modification. Finally, no associations between changes in autistic traits over time and metabolites were detected.

4.1. Interpretation of Main Findings

The etiology of ASD is complex and poorly understood, but various metabolic pathway disturbances as a result of gene–environment interactions already occurring during the fetal period are believed to contribute to this heterogeneous condition. Cord blood metabolomics is able to capture the metabolic phenotype as a result of the interplay between genetics and environment. Moreover, it presents combined information on both maternal metabolism transferred via the placenta and fetal metabolism [30]. Thus, it may contribute to elucidating the complex etiology of ASD. A review from 2020 that included 10 case–control studies reported repeatedly identified differences in amino acid concentrations (tryptophan, BCAA), pathways of oxidative damage (taurine level elevations), abnormalities in the nitric oxide pathway, and disturbed lipid metabolism (decreased levels of lysolipids, free- and short-/long-chain Carn.a, docosahexaenoic acid, increased levels of sphingosine-1-phosphate), as well as disruptions of the Krebs cycle, oxidative phosphorylation, and respiratory chain among pre-school children with an ASD diagnosis compared to non-ASD-diagnosed controls [9]. Moreover, some studies explored the associations between metabolomics and autistic traits prospectively. An untargeted metabolomics case–control study that analyzed the blood of 30 mothers whose children received an ASD diagnosis and of 30 controls detected, amongst others, differences in histidylglutamate, cinnamoglycine, proline, and adrenoylcarnitine concentrations [43]. Another case–control study analyzed stored mid-pregnancy blood of 52 women whose children later received an ASD diagnosis and of 62 controls [13]. They reported differences in several glycosphingolipid, n-glycan, and pyrimidine metabolism pathways. Together, those studies suggest that alterations in the metabolic composition of the blood might be present in children diagnosed with ASD from early life onwards. However, the metabolites that showed differential abundance varied across the studies. Possible explanations could be the lack of uniformity in the collection and storage of serum samples as well as different methods in obtaining the metabolite concentrations (Nuclear Magnetic Resonance (NMR) vs. liquid chromatography–mass spectrometry/mass spectrometry (LC-MS/MS)). Moreover, some studies, such as ours, used a targeted metabolomics approach with a predefined set of metabolites, whereas other studies used an untargeted approach. Finally, varying ethnicity, age of the study population, and different dietary habits could have influenced the results and, therefore, contributed to its heterogeneity.
We are the first study to examine the prospective associations of cord blood metabolites and autistic traits in childhood dimensionally. No associations between metabolites at birth and autistic traits at ages 6 and 13 were observed in our main model. Autism is a heterogeneous condition with a complex etiology that is only partly elucidated, and not all pathophysiological mechanisms may be detectable through changes in the metabolic composition of the blood. Furthermore, the neurodevelopment of the brain extends after birth, and processes that contributed to an autistic condition in childhood might not have been present at birth. Moreover, the blood metabolome is influenced by contemporary environmental factors such as diet and stress level and may therefore derive stronger associations in cross-sectional settings. A prior study conducted within the Generation R Study revealed that correlations between metabolite measurements at various time points (early pregnancy, cord blood, and at 10 years) were low, ranging between r = −0.10 and r = 0.35 [24]. Observed associations in our study attenuated from 6 to 13 years, and the linear mixed-effects model did not yield associations on changes in autistic traits over time with metabolites. As the SRS is parent-reported, puberty could have caused difficulties in assessing children’s behavior correctly compared to age 6. Further, after the age of 6, other environmental influences could have caused the presence of autistic traits at age 13 and thus weakened its association with metabolites. Our results are in line with a Danish case–control study that investigated the association of dried blood spot metabolites obtained 6 days after the birth of 37 ASD cases and 37 healthy controls [19]. They did not detect significant differences in metabolite concentrations among children with an ASD diagnosis as compared to healthy controls. Corroborating our results, the Danish study also reported that gestational age at birth has a strong confounding effect on the association between metabolomics and ASD. A study that analyzed weekly blood samples from 30 pregnant women was able to predict the gestational age of fetuses from blood metabolite measurements [47]. Therefore, future studies involving cord blood metabolomics should consider gestational age as a confounding variable.
However, our results provide suggestive evidence for an inverse association between the concentration of SM.a.C39.2 at birth and autistic traits at age 6 years. This association was observed after multiple testing correction in our basic model but did not survive correction after adjusting for a wider set of confounders, even though effect sizes were comparable. While this finding should be interpreted with caution, it is noteworthy that SM has previously been implicated in neurodevelopmental outcomes and ASD [17,48,49]. SM is highly abundant in the brain as part of the myelin isolating oligodendrocytes [50]. It comprises cellular membranes and is involved in proliferation, migration, inflammation, and cell survival [51]. A small randomized controlled trial among 24 very-low-birth-weight infants reported improvements in neurodevelopmental tests after feeding them SM-fortified milk [52]. Moreover, an intervention study of 15 ASD-diagnosed children reported correlations between increased urinary levels of SM and improvement of autistic traits after treatment with sulforaphane [53]. However, SM.a.C39.2 is an unusual SM in relation to brain myelination, as most SM-forming myelin has between 18 and 24 C-atoms [54]. Moreover, SM.a.C39.2 contributes only about 0.2% to the total SM in our samples. Therefore, a potential deviation of early-life sphingomyelin formation during brain development could be a target for further exploration in relation to the risk of ASD.
Numerous studies have reported differences in fatty acid concentration, Carn.a, and PL, which are linked to mitochondrial dysfunction, in ASD [43,49,55,56,57]. Carn.a transport acyl groups into the mitochondrial matrix for ß-oxidation. Thus, it is crucial crucial for mitochondrial energy production [20]. Previous studies suggest that the accumulation of Carn.a could be a result of incomplete oxidation of fatty acids due to mitochondrial dysfunction [58]. In our models, although non-significant, a trend between elevated Carn.a.C.18.2 concentrations and higher autistic traits at age 6 was observed in the main model.
This is the first autism-related study to report a decreased NEFA16:1/16:0 ratio associated with higher severity of autistic traits, albeit not significant after multiple testing correction in the main model. The ratio represents a proxy of the stearoyl-CoA desaturase-1 (SCD-1) activity. This enzyme breaks down saturated fatty acids, and decreased levels of SCD-1 have been previously linked to inflammation in, e.g., macrophages and endothelial cells [33,59]. Inflammatory processes have been previously linked to ASD pathophysiology [60]. Moreover, SCD-1 functions to form mono-unsaturated fatty acids, of which relatively large amounts are deposited in brain lipids during early brain growth [54]. Therefore, it is required for normal brain development. Reduced levels of NEFA16:1/16:0 ratio could reflect neurodevelopmental abnormalities occurring during pregnancy, possibly leading to a higher abundance of autistic traits in childhood.
In our population-based study, we did not identify certain metabolites at birth associated with autistic traits in childhood. Although we found some suggestive evidence that decreased SM, NEFA16:1/16:0 concentrations and increased Carn.a concentrations at birth might be predictive of more severe autistic traits in childhood, associations attenuated after adjusting for a wider set of confounders and disappeared at the age of 13. Thus, we conclude that the measured metabolites within our study are not suitable for the prediction of autistic traits in childhood. For future studies, it might be beneficial to explore associations with a wider set of metabolites or even consider an untargeted approach. Moreover, repeated measurements on both metabolomics and autistic traits would enable us to study their association over time.

4.2. Methodological Considerations

Among the strengths of this study is the dimensional measure of autistic traits with the SRS score. Although our dimensional approach does not provide a clinical diagnosis, it allows us to account for children with fewer symptoms and model the full range of symptoms in the general population. It also reduces the risk of outcome misclassification and increases power. Moreover, obtaining measurements of SRS scores at two time points is another strength. It allowed us to study the association of longitudinal trajectories of autistic traits with metabolites. Finally, the availability of sociodemographic and lifestyle information of our participants allowed us to control for a broad range of confounders.
Some limitations in the current study need to be considered. Unlike previous studies that used a case–control design, our study was embedded in a population-based cohort. Although the percentage of participants above the recommended SRS cutoff value (0.9%) corresponds to the ASD prevalence (0.6–2.2%), it is noteworthy that our participant’s autistic traits primarily fall within a subclinical range [4]. Consequently, the identified associations may not be as distinct as in case–control designs. Metabolomics data of the Generation R cohort was available for a subgroup of individuals of Dutch ethnicity, with highly educated, normal-weight mothers who were less likely to drink or smoke during pregnancy. This circumstance might have resulted in a selection bias, causing less severe autistic traits among children. We used a targeted metabolomics approach and analyzed 193 metabolites. By limiting our analyses to a predefined set of metabolites, we might have missed associations of other metabolites with autistic traits. Finally, despite adjusting for many potential confounders, due to the observational nature of our study, we cannot completely rule out the effect of residual confounding.

5. Conclusions

We did not find significant associations between cord blood metabolite profiles and autistic traits at the ages of 6 and 13 in our main model. Our results did provide suggestive evidence for associations between SM and NEFA16:1/16:0 abundance at birth and subclinical autistic traits at age 6. These findings should be regarded as hypothesis-generating and require further investigations.

Supplementary Materials

The following are available online at https://www.mdpi.com/article/10.3390/metabo13111140/s1: Supplementary Table S1: Non-response analysis in singleton live births with and without outcome measurements; Supplementary Table S2: Cord blood metabolite concentrations (N = 783); Supplementary Table S3: Associations of cord blood individual metabolites and metabolites with SRS scores at ages 6 and 13. Basic Model; Supplementary Table S4: Associations of cord blood individual metabolites and metabolite groups with SRS scores at ages 6 and 13. Main Model; Supplementary Table S5: Associations of cord blood individual metabolites and metabolite groups with SRS scores at ages 6 and 13. Adjusted Model; Supplementary Table S6: Associations of cord blood metabolite ratios with SRS scores at age 6 and 13. Basic model; Supplementary Table S7: Associations of cord blood metabolite ratios with SRS scores at age 6 and 13. Main model; Supplementary Table S8: Associations of cord blood metabolite ratios with SRS scores at age 6 and 13. Adjusted model; Supplementary Table S9: p-values of interaction effects of linear mixed-effects models; Supplementary Table S10: Parameters for mass-spectrometry detection and identification; Supplementary Figure S1: Difference in SRS score at age 13 per SDS increase in cord-blood metabolite concentration; Supplementary Text S1: Metabolite Measurements. Reference [61] is cited in Supplementary Materials File.

Author Contributions

Conceptualization, C.S.K., S.M.B., V.W.V.J. and R.G. (Romy Goncalves); methodology, C.S.K., S.M.B., B.K., C.A.M.C., R.H.M. and V.W.V.J.; formal analysis, C.S.K.; writing—original draft preparation, C.S.K. and S.M.B.; writing—review and editing, C.A.M.C., R.H.M., R.G. (Romy Gaillard), R.G. (Romy Goncalves), I.B., B.K. and V.W.V.J.; visualization, C.S.K. and S.M.B.; supervision, S.M.B., V.W.V.J. and B.K.; funding acquisition, B.K., C.A.M.C., R.G. (Romy Gaillard), and V.W.V.J. All authors have read and agreed to the published version of the manuscript.

Funding

The Generation R Study is financially supported by the Erasmus Medical Center, Rotterdam, the Erasmus University Rotterdam, and the Netherlands Organization for Health Research and Development. VWVJ received a grant from the Netherlands Organization for Health Research and Development (NWO, ZonMw-VIDI 016.136.361) and a European Research Council Consolidator Grant (ERC-2014-CoG-648916). RG received funding from the Dutch Diabetes Foundation (grant number 2017.81.002), the Netherlands Organization for Health Research and Development (NOW, ZonMW, grant number 543003109, NOW, ZonMw 09150172110034)), and from the European Union’s Horizon 2020 research and innovation program under the ERA-NET Cofund action (No. 727565), EndObesity, ZonMW the Netherlands (No. 529051026). Further partial financial support was received from the European Research Council Advanced Grant META-GROWTH ERC-2012-AdG–No. 322605, the German Ministry of Education and Research, Berlin (01EA2203A), German Research Council Ko912/12-1 and INST 409/224-1 FUGG, and Alexander von Humboldt Foundation 3.3-1218469-POL-HFST-E. B.K. is the Else Kröner Senior Professor of Paediatrics at LMU (University of Munich) and is financially supported by the charitable Else Kröner-Fresenius-Foundation, LMU Medical Faculty, and LMU University Hospitals. CAMC is supported by the European Union’s Horizon 2020 Research and Innovation Programme (EarlyCause, grant agreement No. 848158), the HorizonEurope Research and Innovation Programme (FAMILY, grant agreement No. 101057529; HappyMums, grant agreement No. 101057390), and the European Research Council (TEMPO; grant agreement No 101039672). This research was conducted while CAMC was a Hevolution/AFAR New Investigator Awardee in Aging Biology and Geroscience Research.

Institutional Review Board Statement

This study was approved by the Medical Ethics Committee of the Erasmus Medical Center in Rotterdam, the Netherlands (MEC 198.782/200/31, 17 December 2001).

Informed Consent Statement

Informed consent was obtained from all subjects involved in the study.

Data Availability Statement

The data presented in this study are available on request from the corresponding author. The data are not publicly available due to privacy and ethical restrictions.

Acknowledgments

We gratefully acknowledge the contribution of participating mothers, general practitioners, hospitals, midwives, and pharmacies in Rotterdam and of those who contributed to the preparation and analysis of the blood samples for metabolomics analysis.

Conflicts of Interest

The authors declare no conflict of interest.

Abbreviations

AAAmino acids
AAAAromatic amino acids
BCAABranched-chain amino acids
CarnCarnitines
Carn.aAcyl-carnitines
CVCoefficient of variation
EAAEssential amino acids
FDRFalse discovery rate
HPLCHigh-performance liquid chromatography
IUPAC-IUBInternational Union of Pure and Applied Chemistry—International Union of Biochemistry
Lyso.PC.aAcyl-lysophosphatidylcholines
Lyso.PC.eAlkyl-lysophosphatidylcholines
LC/MS-MSLiquid chromatography–mass spectrometry
NEFANon-esterified fatty acids
NEAANon-essential amino acids
NMRNuclear Magnetic Resonance
PCPhosphatidylcholines
PC.aaDiacyl-phosphatidylcholines
PC.aeAcyl-alkyl-phosphatidylcholines
PLPhospholipids
QCQuality control
SDStandard deviation
SDSStandard deviation scores
SGASmall size for gestational age
SMSphingomyelines
SRSSocial Responsiveness Scale

References

  1. First, M.B. Diagnostic and statistical manual of mental disorders, 5th edition, and clinical utility. J. Nerv. Ment. Dis. 2013, 201, 727–729. [Google Scholar] [CrossRef] [PubMed]
  2. Constantino, J.N.; Todd, R.D. Autistic traits in the general population: A twin study. Arch. Gen. Psychiatry 2003, 60, 524–530. [Google Scholar]
  3. Lundström, S.; Chang, Z.; Råstam, M.; Gillberg, C.; Larsson, H.; Anckarsäter, H.; Lichtenstein, P. Autism spectrum disorders and autistic like traits: Similar etiology in the extreme end and the normal variation. Arch. Gen. Psychiatry 2012, 69, 46–52. [Google Scholar] [CrossRef] [PubMed]
  4. Lord, C.; Elsabbagh, M.; Baird, G.; Veenstra-Vanderweele, J. Autism spectrum disorder. Lancet 2018, 392, 508–520. [Google Scholar] [PubMed]
  5. Bjørklund, G.; Meguid, N.A.; El-Bana, M.A.; Tinkov, A.A.; Saad, K.; Dadar, M.; Hemimi, M.; Skalny, A.V.; Hosnedlová, B.; Kizek, R.; et al. Oxidative Stress in Autism Spectrum Disorder. Mol. Neurobiol. 2020, 57, 2314–2332. [Google Scholar] [PubMed]
  6. Mesleh, A.G.; Abdulla, S.A.; El-Agnaf, O. Paving the Way toward Personalized Medicine: Current Advances and Challenges in Multi-OMICS Approach in Autism Spectrum Disorder for Biomarkers Discovery and Patient Stratification. J. Pers. Med. 2021, 11, 41. [Google Scholar] [PubMed]
  7. Shen, L.; Liu, X.; Zhang, H.; Lin, J.; Feng, C.; Iqbal, J. Biomarkers in autism spectrum disorders: Current progress. Clin. Chim. Acta 2020, 502, 41–54. [Google Scholar]
  8. Glinton, K.E.; Elsea, S.H. Untargeted Metabolomics for Autism Spectrum Disorders: Current Status and Future Directions. Front. Psychiatry 2019, 10, 647. [Google Scholar] [PubMed]
  9. Likhitweerawong, N.; Thonusin, C.; Boonchooduang, N.; Louthrenoo, O.; Nookaew, I.; Chattipakorn, N.; Chattipakorn, S.C. Profiles of urine and blood metabolomics in autism spectrum disorders. Metab. Brain Dis. 2021, 36, 1641–1671. [Google Scholar]
  10. Qureshi, F.; Adams, J.B.; Audhya, T.; Hahn, J. Multivariate Analysis of Metabolomic and Nutritional Profiles among Children with Autism Spectrum Disorder. J. Pers. Med. 2022, 12, 923. [Google Scholar]
  11. Needham, B.D.; Adame, M.D.; Serena, G.; Rose, D.R.; Preston, G.M.; Conrad, M.C.; Campbell, A.S.; Donabedian, D.H.; Fasano, A.; Ashwood, P.; et al. Plasma and Fecal Metabolite Profiles in Autism Spectrum Disorder. Biol. Psychiatry 2021, 89, 451–462. [Google Scholar] [PubMed]
  12. Smith, A.M.; King, J.J.; West, P.R.; Ludwig, M.A.; Donley, E.L.R.; Burrier, R.E.; Amaral, D.G. Amino Acid Dysregulation Metabotypes: Potential Biomarkers for Diagnosis and Individualized Treatment for Subtypes of Autism Spectrum Disorder. Biol. Psychiatry 2019, 85, 345–354. [Google Scholar] [PubMed]
  13. Ritz, B.; Yan, Q.; Uppal, K.; Liew, Z.; Cui, X.; Ling, C.; Inoue, K.; von Ehrenstein, O.; Walker, D.I.; Jones, D.P. Untargeted Metabolomics Screen of Mid-pregnancy Maternal Serum and Autism in Offspring. Autism Res. 2020, 13, 1258–1269. [Google Scholar] [CrossRef]
  14. Panjwani, A.A.; Ji, Y.; Fahey, J.W.; Palmer, A.; Wang, G.; Hong, X.; Zuckerman, B.; Wang, X. Maternal Dyslipidemia, Plasma Branched-Chain Amino Acids, and the Risk of Child Autism Spectrum Disorder: Evidence of Sex Difference. J. Autism Dev. Disord. 2020, 50, 540–550. [Google Scholar]
  15. Smith, A.M.; Natowicz, M.R.; Braas, D.; Ludwig, M.A.; Ney, D.M.; Donley, E.L.R.; Burrier, R.E.; Amaral, D.G. A Metabolomics Approach to Screening for Autism Risk in the Children’s Autism Metabolome Project. Autism Res. 2020, 13, 1270–1285. [Google Scholar] [PubMed]
  16. Orozco, J.S.; Hertz-Picciotto, I.; Abbeduto, L.; Slupsky, C.M. Metabolomics analysis of children with autism, idiopathic-developmental delays, and Down syndrome. Transl. Psychiatry 2019, 9, 243. [Google Scholar]
  17. Kelly, R.S.; Boulin, A.; Laranjo, N.; Lee-Sarwar, K.; Chu, S.H.; Yadama, A.P.; Carey, V.; Litonjua, A.A.; Lasky-Su, J.; Weiss, S.T. Metabolomics and Communication Skills Development in Children; Evidence from the Ages and Stages Questionnaire. Metabolites 2019, 9, 42. [Google Scholar] [CrossRef]
  18. Ristori, M.V.; Mortera, S.L.; Marzano, V.; Guerrera, S.; Vernocchi, P.; Ianiro, G.; Gardini, S.; Torre, G.; Valeri, G.; Vicari, S.; et al. Proteomics and Metabolomics Approaches towards a Functional Insight onto AUTISM Spectrum Disorders: Phenotype Stratification and Biomarker Discovery. Int. J. Mol. Sci. 2020, 21, 6274. [Google Scholar] [CrossRef]
  19. Courraud, J.; Ernst, M.; Svane Laursen, S.; Hougaard, D.M.; Cohen, A.S. Studying Autism Using Untargeted Metabolomics in Newborn Screening Samples. J. Mol. Neurosci. 2021, 71, 1378–1393. [Google Scholar] [PubMed]
  20. Dambrova, M.; Makrecka-Kuka, M.; Kuka, J.; Vilskersts, R.; Nordberg, D.; Attwood, M.M.; Smesny, S.; Sen, Z.D.; Guo, A.C.; Oler, E.; et al. Acylcarnitines: Nomenclature, Biomarkers, Therapeutic Potential, Drug Targets, and Clinical Trials. Pharmacol. Rev. 2022, 74, 506–551. [Google Scholar]
  21. El-Ansary, A.; Chirumbolo, S.; Bhat, R.S.; Dadar, M.; Ibrahim, E.M.; Bjørklund, G. The Role of Lipidomics in Autism Spectrum Disorder. Mol. Diagn. Ther. 2020, 24, 31–48. [Google Scholar] [CrossRef] [PubMed]
  22. Brister, D.; Rose, S.; Delhey, L.; Tippett, M.; Jin, Y.; Gu, H.; Frye, R.E. Metabolomic Signatures of Autism Spectrum Disorder. J. Pers. Med. 2022, 12, 1727. [Google Scholar] [CrossRef]
  23. Taniguchi, M.; Okazaki, T. The role of sphingomyelin and sphingomyelin synthases in cell death, proliferation and migration-from cell and animal models to human disorders. Biochim. Biophys. Acta 2014, 1841, 692–703. [Google Scholar] [CrossRef] [PubMed]
  24. Voerman, E.; Jaddoe, V.W.V.; Uhl, O.; Shokry, E.; Horak, J.; Felix, J.F.; Koletzko, B.; Gaillard, R. A population-based resource for intergenerational metabolomics analyses in pregnant women and their children: The Generation R Study. Metabolomics 2020, 16, 43. [Google Scholar] [CrossRef] [PubMed]
  25. Harder, U.; Koletzko, B.; Peissner, W. Quantification of 22 plasma amino acids combining derivatization and ion-pair LC-MS/MS. J. Chromatogr. B Anal. Technol. Biomed. Life Sci. 2011, 879, 495–504. [Google Scholar] [CrossRef]
  26. IUPAC-IUB Joint Commission on Biochemical Nomenclature (JCBN). Nomenclature and symbolism for amino acids and peptides. Recommendations. Biochem. J. 1984, 219, 345–373. [Google Scholar]
  27. Hellmuth, C.; Weber, M.; Koletzko, B.; Peissner, W. Nonesterified fatty acid determination for functional lipidomics: Comprehensive ultrahigh performance liquid chromatography-tandem mass spectrometry quantitation, qualification, and parameter prediction. Anal. Chem. 2012, 84, 1483–1490. [Google Scholar] [CrossRef]
  28. Uhl, O.; Fleddermann, M.; Hellmuth, C.; Demmelmair, H.; Koletzko, B. Phospholipid Species in Newborn and 4 Month Old Infants after Consumption of Different Formulas or Breast Milk. PLoS ONE 2016, 11, e0162040. [Google Scholar] [CrossRef]
  29. Rauschert, S.; Mori, T.A.; Beilin, L.J.; Jacoby, P.; Uhl, O.; Koletzko, B.; Oddy, W.H.; Hellmuth, C. Early Life Factors, Obesity Risk, and the Metabolome of Young Adults. Obesity 2017, 25, 1549–1555. [Google Scholar] [CrossRef]
  30. Hellmuth, C.; Uhl, O.; Standl, M.; Demmelmair, H.; Heinrich, J.; Koletzko, B.; Thiering, E. Cord Blood Metabolome Is Highly Associated with Birth Weight, but Less Predictive for Later Weight Development. Obes. Facts 2017, 10, 85–100. [Google Scholar] [CrossRef]
  31. Shokry, E.; Marchioro, L.; Uhl, O.; Bermúdez, M.G.; García-Santos, J.A.; Segura, M.T.; Campoy, C.; Koletzko, B. Impact of maternal BMI and gestational diabetes mellitus on maternal and cord blood metabolome: Results from the PREOBE cohort study. Acta Diabetol. 2019, 56, 421–430. [Google Scholar] [CrossRef] [PubMed]
  32. Wei, R.; Wang, J.; Su, M.; Jia, E.; Chen, S.; Chen, T.; Ni, Y. Missing Value Imputation Approach for Mass Spectrometry-based Metabolomics Data. Sci. Rep. 2018, 8, 663. [Google Scholar] [CrossRef] [PubMed]
  33. Sampath, H.; Ntambi, J. Role of stearoyl-CoA desaturase in human metabolic disease. Future Lipidol. 2008, 3, 163–173. [Google Scholar] [CrossRef]
  34. Squeri, A. Coronary Artery Disease—New Insights and Novel Approaches; IntechOpen: London, UK, 2012. [Google Scholar] [CrossRef]
  35. Kirchberg, F.F.; Brandt, S.; Moß, A.; Peissner, W.; Koenig, W.; Rothenbacher, D.; Brenner, H.; Koletzko, B.; Hellmuth, C.; Wabitsch, M. Metabolomics reveals an entanglement of fasting leptin concentrations with fatty acid oxidation and gluconeogenesis in healthy children. PLoS ONE 2017, 12, e0183185. [Google Scholar] [CrossRef] [PubMed]
  36. Molnos, S.; Wahl, S.; Haid, M.; Eekhoff, E.M.W.; Pool, R.; Floegel, A.; Deelen, J.; Much, D.; Prehn, C.; Breier, M.; et al. Metabolite ratios as potential biomarkers for type 2 diabetes: A DIRECT study. Diabetologia 2018, 61, 117–129. [Google Scholar] [CrossRef]
  37. Pickens, C.A.; Vazquez, A.I.; Jones, A.D.; Fenton, J.I. Obesity, adipokines, and C-peptide are associated with distinct plasma phospholipid profiles in adult males, an untargeted lipidomic approach. Sci. Rep. 2017, 7, 6335. [Google Scholar] [CrossRef]
  38. Zhang, W.; Sun, G.; Aitken, D.; Likhodii, S.; Liu, M.; Martin, G.; Furey, A.; Randell, E.; Rahman, P.; Jones, G.; et al. Lysophosphatidylcholines to phosphatidylcholines ratio predicts advanced knee osteoarthritis. Rheumatology 2016, 55, 1566–1574. [Google Scholar] [CrossRef]
  39. Constantino, J.N.; Davis, S.A.; Todd, R.D.; Schindler, M.K.; Gross, M.M.; Brophy, S.L.; Metzger, L.M.; Shoushtari, C.S.; Splinter, R.; Reich, W. Validation of a brief quantitative measure of autistic traits: Comparison of the social responsiveness scale with the autism diagnostic interview-revised. J. Autism Dev. Disord. 2003, 33, 427–433. [Google Scholar] [CrossRef]
  40. Van den Dries, M.A.; Guxens, M.; Pronk, A.; Spaan, S.; El Marroun, H.; Jusko, T.A.; Longnecker, M.P.; Ferguson, K.K.; Tiemeier, H. Organophosphate pesticide metabolite concentrations in urine during pregnancy and offspring attention-deficit hyperactivity disorder and autistic traits. Environ. Int. 2019, 131, 105002. [Google Scholar] [CrossRef]
  41. Lyall, K.; Hosseini, M.; Ladd-Acosta, C.; Ning, X.; Catellier, D.; Constantino, J.N.; Croen, L.A.; Kaat, A.J.; Botteron, K.; Bush, N.R.; et al. Distributional Properties and Criterion Validity of a Shortened Version of the Social Responsiveness Scale: Results from the ECHO Program and Implications for Social Communication Research. J. Autism Dev. Disord. 2021, 51, 2241–2253. [Google Scholar] [CrossRef]
  42. Niklasson, A.; Albertsson-Wikland, K. Continuous growth reference from 24th week of gestation to 24 months by gender. BMC Pediatr. 2008, 8, 8. [Google Scholar] [CrossRef] [PubMed]
  43. Hollowood-Jones, K.; Adams, J.B.; Coleman, D.M.; Ramamoorthy, S.; Melnyk, S.; James, S.J.; Woodruff, B.K.; Pollard, E.L.; Snozek, C.L.; Kruger, U.; et al. Altered metabolism of mothers of young children with Autism Spectrum Disorder: A case control study. BMC Pediatr. 2020, 20, 557. [Google Scholar] [CrossRef] [PubMed]
  44. Yap, I.K.; Angley, M.; Veselkov, K.A.; Holmes, E.; Lindon, J.C.; Nicholson, J.K. Urinary metabolic phenotyping differentiates children with autism from their unaffected siblings and age-matched controls. J. Proteome Res. 2010, 9, 2996–3004. [Google Scholar] [CrossRef] [PubMed]
  45. Bates, D.M.M.; Bolker, B.; Walker, S. Linear Mixed-Effects Models Using ‘Eigen’ and S4; CRAN—R Project: Vienna, Austria, 2023. [Google Scholar]
  46. Van Buuren, S. Multiple imputation of discrete and continuous data by fully conditional specification. Stat. Methods Med. Res. 2007, 16, 219–242. [Google Scholar] [CrossRef]
  47. Liang, L.; Rasmussen, M.H.; Piening, B.; Shen, X.; Chen, S.; Röst, H.; Snyder, J.K.; Tibshirani, R.; Skotte, L.; Lee, N.C.; et al. Metabolic Dynamics and Prediction of Gestational Age and Time to Delivery in Pregnant Women. Cell 2020, 181, 1680–1692.e1615. [Google Scholar] [CrossRef]
  48. Wang, L.; Zheng, R.; Xu, Y.; Zhou, Z.; Guan, P.; Wu, Y.; Zhou, J.; Cheng, Z.; Zhang, L. Altered Metabolic Characteristics in Plasma of Young Boys with Autism Spectrum Disorder. J. Autism Dev. Disord. 2022, 52, 4897–4907. [Google Scholar] [CrossRef]
  49. Wang, H.; Liang, S.; Wang, M.; Gao, J.; Sun, C.; Wang, J.; Xia, W.; Wu, S.; Sumner, S.J.; Zhang, F.; et al. Potential serum biomarkers from a metabolomics study of autism. J. Psychiatry Neurosci. 2016, 41, 27–37. [Google Scholar] [CrossRef]
  50. Gault, C.R.; Obeid, L.M.; Hannun, Y.A. An overview of sphingolipid metabolism: From synthesis to breakdown. Adv. Exp. Med. Biol. 2010, 688, 1–23. [Google Scholar]
  51. Schneider, N.; Hauser, J.; Oliveira, M.; Cazaubon, E.; Mottaz, S.C.; O’Neill, B.V.; Steiner, P.; Deoni, S.C.L. Sphingomyelin in Brain and Cognitive Development: Preliminary Data. eNeuro 2019, 6. [Google Scholar] [CrossRef]
  52. Tanaka, K.; Hosozawa, M.; Kudo, N.; Yoshikawa, N.; Hisata, K.; Shoji, H.; Shinohara, K.; Shimizu, T. The pilot study: Sphingomyelin-fortified milk has a positive association with the neurobehavioural development of very low birth weight infants during infancy, randomized control trial. Brain Dev. 2013, 35, 45–52. [Google Scholar] [CrossRef]
  53. Bent, S.; Lawton, B.; Warren, T.; Widjaja, F.; Dang, K.; Fahey, J.W.; Cornblatt, B.; Kinchen, J.M.; Delucchi, K.; Hendren, R.L. Identification of urinary metabolites that correlate with clinical improvements in children with autism treated with sulforaphane from broccoli. Mol. Autism 2018, 9, 35. [Google Scholar] [CrossRef] [PubMed]
  54. Staellberg-Stenhagen, S.; Svennerholm, L. Fatty Acid Composition of Human Brain Sphingomyelins: Normal Variation with Age and Changes during Myelin Disorders. J. Lipid Res. 1965, 6, 146–155. [Google Scholar] [CrossRef]
  55. Lv, Q.Q.; You, C.; Zou, X.B.; Deng, H.Z. Acyl-carnitine, C5DC, and C26 as potential biomarkers for diagnosis of autism spectrum disorder in children. Psychiatry Res. 2018, 267, 277–280. [Google Scholar] [CrossRef] [PubMed]
  56. Hassan, M.H.; Desoky, T.; Sakhr, H.M.; Gabra, R.H.; Bakri, A.H. Possible Metabolic Alterations among Autistic Male Children: Clinical and Biochemical Approaches. J. Mol. Neurosci. 2019, 67, 204–216. [Google Scholar] [CrossRef] [PubMed]
  57. Barone, R.; Alaimo, S.; Messina, M.; Pulvirenti, A.; Bastin, J.; Group, M.I.-A.; Ferro, A.; Frye, R.E.; Rizzo, R. A Subset of Patients with Autism Spectrum Disorders Show a Distinctive Metabolic Profile by Dried Blood Spot Analyses. Front. Psychiatry 2018, 9, 636. [Google Scholar] [CrossRef]
  58. McGill, M.R.; Li, F.; Sharpe, M.R.; Williams, C.D.; Curry, S.C.; Ma, X.; Jaeschke, H. Circulating acylcarnitines as biomarkers of mitochondrial dysfunction after acetaminophen overdose in mice and humans. Arch. Toxicol. 2014, 88, 391–401. [Google Scholar] [PubMed]
  59. Liu, X.; Strable, M.S.; Ntambi, J.M. Stearoyl CoA desaturase 1: Role in cellular inflammation and stress. Adv. Nutr. 2011, 2, 15–22. [Google Scholar] [CrossRef]
  60. Rossignol, D.A.; Frye, R.E. Evidence linking oxidative stress, mitochondrial dysfunction, and inflammation in the brain of individuals with autism. Front. Physiol. 2014, 5, 150. [Google Scholar] [CrossRef]
  61. Uhl, O.; Glaser, C.; Demmelmair, H.; Koletzko, B. Reversed phase LC/MS/MS method for targeted quantification of glycerophospholipid molecular species in plasma. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 2011, 879, 3556–3564. [Google Scholar] [CrossRef]
Metabolites 13 01140 g001aMetabolites 13 01140 g001b
Figure 1. Associations of individual cord-blood metabolites with SRS scores at the age of 6 years. (A) basic model, (B) main model. Values represent the estimated change in the SRS score associated per SDS increase of a single cord blood metabolite (µmol/L) from linear regression models. (A) is adjusted for sex and age at outcome; (B) is adjusted for sex, age at outcome, maternal BMI, maternal psychopathologies, education level, smoking during pregnancy, alcohol intake during pregnancy, gestational age at birth, and birthweight. Labeled values represent significant associations (FDR-adjusted p-values < 0.05). Corresponding numerical values are shown in Table S4. AA amino acids, NEFA non-esterified fatty acids, PC.aa diacyl-phosphatidylcholines, PC.ae acyl-alkyl-phosphatidylcholines, lyso.PC.a. acyl-lysophosphatidylcholines, lyso.PC.e alkyl-lysophosphatidylcholines, Carn.a acylcarnitines, SM sphingomyelines. Each color represents a different metabolite group, e.g., red represents “AA”.
Figure 1. Associations of individual cord-blood metabolites with SRS scores at the age of 6 years. (A) basic model, (B) main model. Values represent the estimated change in the SRS score associated per SDS increase of a single cord blood metabolite (µmol/L) from linear regression models. (A) is adjusted for sex and age at outcome; (B) is adjusted for sex, age at outcome, maternal BMI, maternal psychopathologies, education level, smoking during pregnancy, alcohol intake during pregnancy, gestational age at birth, and birthweight. Labeled values represent significant associations (FDR-adjusted p-values < 0.05). Corresponding numerical values are shown in Table S4. AA amino acids, NEFA non-esterified fatty acids, PC.aa diacyl-phosphatidylcholines, PC.ae acyl-alkyl-phosphatidylcholines, lyso.PC.a. acyl-lysophosphatidylcholines, lyso.PC.e alkyl-lysophosphatidylcholines, Carn.a acylcarnitines, SM sphingomyelines. Each color represents a different metabolite group, e.g., red represents “AA”.
Metabolites 13 01140 g001aMetabolites 13 01140 g001b
Table 1. General characteristics of the study population.
Table 1. General characteristics of the study population.
CharacteristicsTotal Sample
n = 783
Maternal characteristics
Age at enrolment mean (±SD), years31.8 (3.9)
Education level, high, n (%)510 (65.1%)
Pre-pregnancy body mass index, median (95% range), kg/m222.4 (18.5, 34.0)
Smoking, n (%)
Never smoked during pregnancy555 (79.1%)
Smoked until pregnancy was known63 (9.0%)
Continued smoking during pregnancy84 (11.9%)
Alcohol use, n (%)
No alcohol consumption during pregnancy207 (29.6%)
Alcohol consumption until pregnancy was known103 (14.7%)
Alcohol consumption continued during pregnancy389 (55.7%)
Psychopathology, median (95% range)0.1 (0, 1)
Folic acid supplements use, yes, n (%) *600 (92.9%)
Vitamin D deficiency, yes, n (%) **221 (30.2%)
Fetal characteristics
Fetal sex, female, n (%)372 (47.5%)
Gestational age at birth in weeks, median (95% range)40.3 (36.9, 42.4)
Birthweight in grams, mean (±SD)3541 (494.0)
Birthweight < 2500 g, n (%)16 (2.0%)
Birthweight 2500 to 4500 g, n (%)744 (95.0%)
Birthweight > 4500 g, n (%)23 (3.0%)
Child characteristics
6 years visit
Age at visit in years, median (95% range)5.9 (5.7, 6.8)
SRS autistic traits score, median (95% range)3.0 (0, 12)
13 years visit
Age at visit in years, median (95% range)13.5 (13, 14.4)
SRS autistic traits score, median (95% range)4.0 (0, 15)
SRS: Social Responsiveness Scale. Values presented as mean (±standard deviation (SD), median (interquartile range (95% range)), or number of participants (valid%). Psychopathology: GSI score; Number of missing per covariate: smoking, n = 81 (10%); alcohol use, n = 84 (11%); psychopathology, n = 51 (7%); education, n = 7 (1%); folic acid supplement n = 137 (17%), vitamin D, n = 50 (6%). * Start of folic acid supplementation preconception or in first 10 weeks of pregnancy = “yes”. ** Vitamin D deficiency: <50 nmol/L.
Table 2. Association of cord-blood metabolites and SRS scores at age 6 and 13.
Table 2. Association of cord-blood metabolites and SRS scores at age 6 and 13.
SRS Age 6SRS Age 13
Basic ModelMain ModelBasic ModelMain Model
Beta p-Value *Beta p-Value *Betap-Value *Beta p-Value *
Sphingomyelines
SM.a.C.39.2		−0.600.01−0.540.08−0.050.880.020.99
Carnitines
Carn.a.C.18.2		0.480.060.510.080.280.510.310.69
NEFA Ratio
NEFA 16:1/16:0		−0.550.02−0.480.08−0.380.40−0.320.69
* FDR-corrected p-value. Basic model: adjusted for gender and age at outcome; main model: adjusted for gender, age at outcome, maternal BMI, maternal psychopathologies, education level, smoking during pregnancy, alcohol intake during pregnancy, gestational age at birth, and birthweight.
Disclaimer/Publisher’s Note: The statements, opinions and data contained in all publications are solely those of the individual author(s) and contributor(s) and not of MDPI and/or the editor(s). MDPI and/or the editor(s) disclaim responsibility for any injury to people or property resulting from any ideas, methods, instructions or products referred to in the content.

Share and Cite

MDPI and ACS Style

Kaupper, C.S.; Blaauwendraad, S.M.; Cecil, C.A.M.; Mulder, R.H.; Gaillard, R.; Goncalves, R.; Borggraefe, I.; Koletzko, B.; Jaddoe, V.W.V. Cord Blood Metabolite Profiles and Their Association with Autistic Traits in Childhood. Metabolites 2023, 13, 1140. https://doi.org/10.3390/metabo13111140

AMA Style

Kaupper CS, Blaauwendraad SM, Cecil CAM, Mulder RH, Gaillard R, Goncalves R, Borggraefe I, Koletzko B, Jaddoe VWV. Cord Blood Metabolite Profiles and Their Association with Autistic Traits in Childhood. Metabolites. 2023; 13(11):1140. https://doi.org/10.3390/metabo13111140

Chicago/Turabian Style

Kaupper, Christin S., Sophia M. Blaauwendraad, Charlotte A. M. Cecil, Rosa H. Mulder, Romy Gaillard, Romy Goncalves, Ingo Borggraefe, Berthold Koletzko, and Vincent W. V. Jaddoe. 2023. "Cord Blood Metabolite Profiles and Their Association with Autistic Traits in Childhood" Metabolites 13, no. 11: 1140. https://doi.org/10.3390/metabo13111140

Note that from the first issue of 2016, this journal uses article numbers instead of page numbers. See further details here.

Article Metrics

Back to TopTop