Autism Spectrum Disorders — in Search of Mechanistic Biomarkers



Autism spectrum disorders are a group of neuropsychiatric conditions of increasing prevalence. They are initially detected in early development of children. Diagnosis is currently made on the basis of clinical behaviour and cognition. Improvements in accuracy, timeliness and access to diagnosis to help manage the condition is high on the agenda of the autistic communities. A blood test may help for early-stage detection of autism spectrum disorders to focus support where required — particularly when symptoms are most challenging. This article discusses briefly the scientific basis of diagnosis of autism spectrum disorders and recent emergence of candidate blood tests for autism. We conclude that further validation and improvements in understanding of autism spectrum disorders are required to provide the scientific basis and classifier characteristics for accurate and reliable diagnosis by clinical chemistry blood test.

General Information

Keywords: blood test, diagnosis, proteomics, autistic adults, delayed diagnosis

Journal rubric: Research & Diagnosis of ASD

Article type: scientific article


Acknowledgements. NR would like to acknowledge the correspondence and the communication with the autistic community via emails and the websites , that encourage us to write this review.

For citation: Rabbani N., Thornalley P.J. Autism Spectrum Disorders — in Search of Mechanistic Biomarkers. Autizm i narusheniya razvitiya = Autism and Developmental Disorders, 2019. Vol. 17, no. 1, pp. 15–23. DOI: 10.17759/autdd.2019170103.

Full text

            Autism Spectrum Disorders (ASD) are a collection of neuropsychiatric disorders characterized by difficulties in social interactions and interests causing a wide variety of disabilities. These commonly present as speech disturbances, repetitive and/or compulsive behaviors, hyperactivity, anxiety and difficulty of adaptation to new environments, with or without cognitive impairment [1]. The high heterogeneity of the clinical presentation makes diagnosis of ASD difficult and uncertain, particularly at the early stages of development [15]. Discovery and development of robust biomarkers – surrogate indicators of the clinical condition - for diagnosis and progression of severity of ASD would facilitate early diagnosis and prioritization of support to those who likely will otherwise become most severely affected.

Genetic causes of ASD are evident in a minority of cases, with most ASD likely resulting from the combination of environmental factors with a large number of (>1000) contributing causal genetic variants, each associated with a low increase in risk. A study of twins found that genetic variables contribute 35% – 40% toward the risk of developing ASD and that the remaining 60% was associated with prenatal, perinatal, and postnatal environmental factors [5]. For some subjects, a single genetic variant may be sufficient to produce ASD; and in others, complex combinations of many common genetic variants may additively increase the risk of ASD. The largest genome-wide association studies performed to date were statistically underpowered to identify a single genetic variant of genome-wide significance. Further larger studies are required. It is also common in early onset neurodevelopmental disorders that risk factors are either rare with large effect or frequent with small effect [3]. Biological pathways associated with ASD through genetic variants linked to increased susceptibility involve chromatin remodelling, transcription, cytoskeleton dynamics, synaptic function, protein synthesis, damaging modifications and degradation of proteins, and amino acid transporters [2;3]. Transcriptomic, proteomic and metabolomic profiling have been proposed for diagnosis of ASD. Diagnostic performance judged by area under-the-curve of receiver operating characteristic (AUROC) plot of 0.73 – 0.91 [4;11;13]. Typically with ‘omics technologies approaches for diagnosis of a disease of polygenic risk, moving closer to to the phenotype – from genetics, to proteomics and metabolomics, we move closer to causality and diagnosis becomes more secure [9]. Improved diagnostic performance may be achieved with one of a relatively small number combinations of biomarker proteins and/or metabolites linked to the causal mechanisms of ASD.


Why early diagnosis of autism is important


Diagnosis of autism is important for personalized support and care. his is particularly important in the early stages where interventions may help diminish progression of symptoms to undesirable severity; and also because with advancing intellectual development and awareness, subjects with ASD may learn to mask the symptoms - making secure diagnosis challenging [8]. Also, accurate and accessible diagnosis of ASD in parents would provide appropriate clinical support rather than undiagnosed cases be attributed to “bad parenting” and removal of children into the care for social services. For multiple reasons, therefore, the ASD community is deserving of improved service by the clinical diagnostics industry and stakeholders. Some groups within the community of those affected by ASD and their carers have expressed resistance to the development of early-state diagnosis. This may be linked, in part, to fears that if pre-natal tests were possible and developed, social pressure may develop to impose negative fetal section linked to ASD risk predictors. Benefits associated with some ASD phenotypes may thereby be lost in pursuit of avoidance of difficulties of providing care and support for an individual with expected ASD. Such application of ASD diagnosis would be ethically unacceptable. Such fears should not detract from the benefits that diagnosis of ASD may give for appropriate provision of care and support for children and adults affected.


Promise of a blood-based diagnosis for autism

Efforts have been made to develop a blood test for early and more accurate diagnosis. To date only two studies have shown promise and been taken forward to validate the assay.


1.         Multivariate analysis of biomarkers of oxidative stress and DNA methylation ASD prediction


Howsmon et al [6] developed a statistical model utilising biomarkers associated with folic acid-dependent the one carbon metabolic pathway of and trans-sulfuration pathways. Abnormalities in levels of metabolites of these pathways and environment effects have been observed in subjects with ASD. The approach taken in this study to use metabolites already associated with ASD was appropriate.  They developed a multivariate statistical model using 24 analytes measured in blood samples and selected 6 that gave the best classifier discrimination between subjects with and without ASD. Analytes included amino acids, oxidised and reduced glutathione (GSSG and GSH, respectively), 3-nitrotyrosine and 8-hydroxydeoxyguanosine and others from the Integrated Metabolic And Genomic Endeavour (IMAGE) study [10]. They found five variables provided optimum classification:  oxidized glutathione.  GSSG, (GSH + 2 x GSSG)/GSSG, 3-nitrotyrosine, tyrosine, and fraction of total cysteine (cysteine + 2 x cysteine) present as cysteine. In a second study with similar analytes algorithm training optimized to a different set of 5 analytes.  This was followed by validation of the model using data from independent clinical studies, however, data of two of the 5 analytes – stated as the most important variables in the statistical model - were unavailable. The validation set also only included subjects with clinical ASD diagnosis and so the validation only tested for true positive and false negative ASD classifications [7].  This greatly impaired the validation process. There is some doubt, therefore, as to which is the best analyte combination to use. Further validation with the complete set of optimised analytes in an independent clinical subject group including subjects with and without ASD is required. Dissemination of standardised protocols for analyte estimation is also required to facilitate independent corroboration in other laboratories.


2.         Protein glycation and oxidation and related amino acid metabolome


It our own research we took an innovative approach of assessing the diagnostic utility of spontaneous low level glycation, oxidation and nitration modifications of proteins and related amino acid metabolites for diagnosis of ASD [2]. Such biomarkers often shown subtle changes in multiple analytes or a “fingerprint” of metabolic disturbances in clinical conditions [14]. We used a machine learning approach for an objective, data-driven selection of protein damage biomarkers by glycation, oxidation and nitration for optimum diagnosis. We found that children with ASD had increased modifications by selected sugar-derived advanced glycation endproducts (AGEs), Nε-carboxymethyllysine (CML) and Nω-carboxymethyl-arginine (CMA), and increased oxidation damage marker, dityrosine (DT), in plasma protein, with respect to children without ASD. From study of renal handling of amino acids, we found that children with ASD had decreased renal clearance of arginine and CMA with respect to children without ASD. Algorithms to discriminate between children with and without ASD gave strong diagnostic performance with features: plasma protein AGEs - CML, CMA and 3-deoxyglucosone-derived hydroimidazolone (3DG-H), and oxidative damage marker, DT. With these 4 analytes, AUROC was 0.94 and sensitivity and specificity was 92% and 84%, respectively; cf. clinical diagnosis by the well-studied Autism Diagnos­tic Interview-Revised (ADI-R) protocol where sensitivity and specificity is highly variable and lower than achieved in our blood test – ranges of 60 – 90% and 70 – 81%, respectively [15]. We did internal 2-fold cross-validation and repeated this 10 times with independent, random data selection.  Intriguingly, the biomarkers related to glyoxal-mediated modifications of proteins (CML and CMA) and activation of dual oxidase (DUOX) which may implicate lipid peroxidation and abnormal hyper activation of DUOX following a challenge to gastrointestinal immunity in ASD. The effect on renal clearance of arginine and CMA suggested a link to arginine transporters which has been implicated from rare genetic variants previously [12]. All analytes were measured concurrently by liquid chromatography-tandem mass spectrometry (LC-MS/MS) which is readily translatable to implementation in the clinical chemistry laboratory; cf. the test developed by Howsman et al. [6;7] which requires several different analytical runs. This test is applicable to children and adults; further studies to confirm this are planned. The study was a biomarker discovery study and comes with the caveats of requirement for further validation in independent and larger clinical study groups, including subjects with early-stage ASD.


Concluding remarks


Clinical diagnosis of ASD has profound consequences for individuals and their families, friends and carers. Delayed referral of children exhibiting symptoms of ASD to experts in child development and psychiatry for diagnosis also creates profound anxieties for individuals and their support groups. Improved accuracy and access to diagnosis, particularly when symptoms first emerge and with follow-up, is required to provide appropriate personalized care and support. Diagnosis protocols requiring engagement of multiple clinical specialists cannot meet current demand and hence long referral times for families are common. A high performance blood test could meet the clinical demand and engage the clinical diagnostics industry with benefit for all stakeholders. Current and further blood tests in development require rigorous validation and to achieve high accuracy and low false-positive rates before implementation. There are encouraging signs, however, that this may be achievable with a limited number of biomarkers that may also point to the underlying mechanism of ASD.


  1. American Psychiatric Association: Diagnostic And Statistical Manual Of Mental Disorders [DSM-5]. Washington, DC: American Psychiatric Publishing, 2013.
  2. Anwar A, Abruzzo P.M., Pasha S., Rajpoot K., Bolotta A., Ghezzo A., Marini M., Posar A., Visconti P., Thornalley P.J., Rabbani N. Advanced glycation endproducts, dityrosine and arginine transporter dysfunction in autism — a source of biomarkers for clinical diagnosis. Molecular Autism, 2018, vol. 9:3.
  3. Bourgeron T. Current knowledge on the genetics of autism and propositions for future research. Comptes Rendus Biologies [Biology Reports], 2016, vol. 339, pp. 300—307.
  4. Diémé B., Mavel S., Blasco H., Tripi G., Bonnet-Brilhault F., Malvy J., Bocca C., Andres C.R., Nadal-Desbarats L., Emond P. Metabolomics Study of Urine in Autism Spectrum Disorders Using a Multiplatform Analytical Methodology. Journal of Proteome Research, 2015, vol. 14, pp. 5273—5282.
  5. Hallmayer J., Cleveland S., Torres A. et al. Genetic heritability and shared environmental factors among twin pairs with autism. Archives of General Psychiatry, 2011, vol. 68, pp. 1095—1102.
  6. Howsmon D.P., Kruger U., Melnyk S., James S.J., Hahn J. Classification and adaptive behavior prediction of children with autism spectrum disorder based upon multivariate data analysis of markers of oxidative stress and DNA methylation. PLOS Computational Biology, 2017, vol. 13. doi:10.1371/journal. pcbi.1005385
  7. Howsmon D.P., Vargason T., Rubin R.A., Delhey L., Tippett M., Rose S., Bennuri S.C., Slattery J.C., Melnyk S., James S.J., Frye R.E., Hahn J. Multivariate techniques enable a biochemical classification of children with autism spectrum disorder versus typically-developing peers: A comparison and validation study. Bioengineering & translational medicine, 2018, vol. 3, pp. 156—165.
  8. Hull L., Petrides K.V., Allison C., Smith P., Baron-Cohen S., Lai M.-C., Mandy W. “Putting on My Best Normal”: Social Camouflaging in Adults with Autism Spectrum Conditions. Journal of Autism and Developmental Disorders, 2017, vol. 47, pp. 2519—534.
  9. Jenkinson C.P., Göring H.H.H., Arya R., Blangero J., Duggirala R., DeFronzo R.A. Transcriptomics in type 2 diabetes: Bridging the gap between genotype and phenotype. Genomics Data, 2016, vol. 8, pp. 25—6.
  10. Melnyk S., Fuchs G.J., Schulz E., Lopez M., Kahler S.G., Fussell J.J., Bellando J., Pavliv O., Rose S., Seidel L., Gaylor D.W., James S.J. Metabolic Imbalance Associated with Methylation Dysregulation and Oxidative Damage in Children with Autism. Journal of Autism and Developmental Disorders, 2012, vol. 42, pp. 367—377.
  11. Momeni N., Bergquist J., Brudin L., Behnia F., Sivberg B., Joghataei M.T., Persson B.L. A novel bloodbased biomarker for detection of autism spectrum disorders. Translational Psychiatry, 2012, vol. 2. doi:10.1038/tp.2012.19
  12. Nava C., Rupp J., Boissel J.-P., Mignot C., Rastetter A., Amiet C., Jacquette A., Dupuits C., Bouteiller D., Keren B., Ruberg M., Faudet A., Doummar D., Philippe A., Périsse D., Laurent C., Lebrun N., Guillemot V., Chelly J., Cohen D., Héron D., Brice A., Closs E.I., Depienne C. Hypomorphic variants of cationic aminoacid transporter 3 in males with autism spectrum disorders. Amino Acids, 2015, vol. 47, pp. 2647—2658.
  13. Scherer S.W., Dawson G. Risk factors for autism: translating genomic discoveries into diagnostics. Human Genetics, 2011, vol. 130, pp. 123—148.
  14. Thornalley P.J. Quantitative screening of protein glycation, oxidation, and nitration adducts by LCMS/MS: protein damage in diabetes, uremia, cirrhosis, and Alzheimer’s disease. In Dalle-Donne I., Scaloni A., Butterfield D.A. (eds.) Redox Proteomics. Hoboken: Wiley, 2006. Pp. 681—28.
  15. Zwaigenbaum L., Penner M. Autism spectrum disorder: advances in diagnosis and evaluation. The BMJ, 2018, vol. 361, k1674.

Information About the Authors

Naila Rabbani, PhD, Reader of Experimental Systems Biology, Warwick Medical School and Zeeman Institute for Systems Biology & Infectious Disease Epidemiology Research; University of Warwick, Deputy Director, Research Technology Platform — Proteomics, University of Warwick, Coventry, U.K. Clinical Sciences Research Laboratories; University Hospital, Coventry,, Coventry, Great Britain, e-mail:

Paul J. Thornalley, PhD, Direttore del Centro per la Ricerca sul Diabete, Diabetes Research Center, Qatar Biomedical Research Institute (QBRI), Hamad Bin Khalifa University; Qatar Foundation, Doha, Qatar, e-mail:



Total: 3412
Previous month: 80
Current month: 85


Total: 914
Previous month: 3
Current month: 5