Unraveling the metabolomic architecture of autism in a large Danish population-based cohort

Filip Ottosson; Francesco Russo; Anna Abrahamsson; Nadia MacSween; Julie Courraud; Kristin Skogstrand; Olle Melander; Ulrika Ericson; Marju Orho-Melander; Arieh S Cohen; Jakob Grove; Preben Bo Mortensen; David M Hougaard; Madeleine Ernst

doi:10.1186/s12916-024-03516-7

Unraveling the metabolomic architecture of autism in a large Danish population-based cohort

BMC Med. 2024 Jul 19;22(1):302. doi: 10.1186/s12916-024-03516-7.

Authors

Filip Ottosson^#^{1

2}, Francesco Russo^#^{3

4}, Anna Abrahamsson³, Nadia MacSween³, Julie Courraud^{3

4

5

6}, Kristin Skogstrand^{3

4}, Olle Melander⁷, Ulrika Ericson⁷, Marju Orho-Melander⁷, Arieh S Cohen^{3

4

8}, Jakob Grove^{4

9

10

11}, Preben Bo Mortensen^{4

12

13}, David M Hougaard^{3

4}, Madeleine Ernst^{14

15}

Affiliations

¹ Section for Clinical Mass Spectrometry, Danish Center for Neonatal Screening, Department of Congenital Disorders, Statens Serum Institut, Copenhagen, Denmark. faot@ssi.dk.
² The Lundbeck Foundation Initiative for Integrative Psychiatric Research, Copenhagen, Denmark. faot@ssi.dk.
³ Section for Clinical Mass Spectrometry, Danish Center for Neonatal Screening, Department of Congenital Disorders, Statens Serum Institut, Copenhagen, Denmark.
⁴ The Lundbeck Foundation Initiative for Integrative Psychiatric Research, Copenhagen, Denmark.
⁵ Laboratory of Analytical Chemistry, Department of Chemistry, National and Kapodistrian University of Athens, 15771, Panepistimiopolis, ZografouAthens, Greece.
⁶ Department of Clinical Therapeutics, School of Medicine, National and Kapodistrian University of Athens, Alexandra Hospital, 11528, Athens, Greece.
⁷ Department of Clinical Sciences, Lund University, Malmö, Sweden.
⁸ Testcenter Denmark, Statens Serum Institut, Copenhagen, Denmark.
⁹ Department of Biomedicine - Human Genetics, Aarhus University, Aarhus, Denmark.
¹⁰ Bioinformatics Research Center, Aarhus University, Aarhus, Denmark.
¹¹ Center for Genomics and Personalized Medicine, Aarhus, Denmark.
¹² NCRR - National Centre for Register-Based Research, Aarhus University, Aarhus, Denmark.
¹³ CIRRAU - Centre for Integrated Registerbased Research at Aarhus University, Aarhus, Denmark.
¹⁴ Section for Clinical Mass Spectrometry, Danish Center for Neonatal Screening, Department of Congenital Disorders, Statens Serum Institut, Copenhagen, Denmark. maet@ssi.dk.
¹⁵ The Lundbeck Foundation Initiative for Integrative Psychiatric Research, Copenhagen, Denmark. maet@ssi.dk.

^# Contributed equally.

Abstract

Background: The prevalence of autism in Denmark has been increasing, reaching 1.65% among 10-year-old children, and similar trends are seen elsewhere. Although there are several factors associated with autism, including genetic, environmental, and prenatal factors, the molecular etiology of autism is largely unknown. Here, we use untargeted metabolomics to characterize the neonatal metabolome from dried blood spots collected shortly after birth.

Methods: We analyze the metabolomic profiles of a subset of a large Danish population-based cohort (iPSYCH2015) consisting of over 1400 newborns, who later are diagnosed with autism and matching controls and in two Swedish population-based cohorts comprising over 7000 adult participants. Mass spectrometry analysis was performed by a timsTOF Pro operated in QTOF mode, using data-dependent acquisition. By applying an untargeted metabolomics approach, we could reproducibly measure over 800 metabolite features.

Results: We detected underlying molecular perturbations across several metabolite classes that precede autism. In particular, the cyclic dipeptide cyclo-leucine-proline (FDR-adjusted p = 0.003) and the carnitine-related 5-aminovaleric acid betaine (5-AVAB) (FDR-adjusted p = 0.03), were associated with an increased probability for autism, independently of known prenatal and genetic risk factors. Analysis of genetic and dietary data in adults revealed that 5-AVAB was associated with increased habitual dietary intake of dairy (FDR-adjusted p < 0.05) and with variants near SLC22A4 and SLC22A5 (p < 5.0e - 8), coding for a transmembrane carnitine transporter protein involved in controlling intracellular carnitine levels.

Conclusions: Cyclo-leucine-proline and 5-AVAB are associated with future diagnosis of autism in Danish neonates, both representing novel early biomarkers for autism. 5-AVAB is potentially modifiable and may influence carnitine homeostasis.

Keywords: 5-Aminovaleric acid betaine; Autism; Cyclo-leucine-proline; Dried blood spot; Metabolome; Metabolomics; Neonatal.

MeSH terms

Adult
Autistic Disorder* / blood
Autistic Disorder* / epidemiology
Autistic Disorder* / genetics
Betaine / blood
Cohort Studies
Denmark / epidemiology
Female
Humans
Infant, Newborn
Male
Metabolome
Metabolomics* / methods

Substances

Betaine