Lack of Evidence for Genomic Instability in Autistic Children as Measured by the Cytokinesis-Block Micronucleus Cytome Assay

Lack of Evidence for Genomic Instability in Autistic Children asMeasured by the Cytokinesis-Block Micronucleus Cytome AssayPenelope A.E. Main, Philip Thomas, Manya T. Angley, Robyn Young, Adrian Esterman,Catherine E. King, and Michael F. Fenech

Autism spectrum disorders are a set of neurodevelopmental disorders that are highly hereditable. Increased genomicinstability has been observed in other heritable paediatric neurobiological disorders; therefore, the aim of our study wasto test the hypothesis that DNA damage is increased in children with autism and that B vitamin status may explainvariations in genome integrity between autistic and normal children. We compared 35 children with autism, 27 of theirsiblings without autism and 25 age- and gender-matched community controls for genomic stability using thecytokinesis-block micronucleus cytome (CBMN-cyt) assay, B vitamins and homocysteine, as well as autism-relatedbehaviours. It was found that there were no differences in CBMN-cyt biomarkers between the three groups. Vitamin B2was significantly raised in children with autism and their siblings compared with controls (P = 0.027 and P = 0.016respectively) but there was no difference in other B vitamins or homocysteine. In conclusion, although replication usinga larger cohort is needed, it appears unlikely that genomic instability is a feature of the aetiology of autism. We cannotrule out in utero effects or other types of DNA damage not measured by the CBMN-cyt assay. Autism Res 2014, ••: ••–••.© 2014 International Society for Autism Research, Wiley Periodicals, Inc.

Keywords: autism; genomic instability; DNA damage; B vitamins; behaviour; riboflavin

Introduction

Autism spectrum disorders (ASDs) are a heterogeneousgroup of behaviourally defined disorders characterisedby impaired social interaction, communication deficitsand restricted or repetitive behaviours [AmericanPsychiatric Association, 2000]. Although once consideredrare, current estimates of the median global prevalence ofASDs are about 62/10,000, whereas the median preva-lence of autistic disorder, an ASD that is differentiated byabnormal language development before or around theage of 3 years, is of the order of 17/10,000 [Elsabbaghet al., 2012]. While the exact cause of autism remainsunknown, a strong genetic component is indicated. Copynumber variations [Glessner et al., 2009; Kakinuma &Sato, 2008; Pinto et al., 2010], epigenetic effects[Grafodatskaya, Chung, Szatmari, & Weksberg, 2010;Shanen, 2006] and imprinting [Badcock, 2011] play arole; however, environmental factors such as maternalimmune response [Nordahl et al., 2013], season of con-ception [Zerbo, Issif, Delwiche, Walker, & Hertz-Picciotto,

2011] and age of paternity [Hultman, Sandin, Levine,Lichtenstein, & Reichenberg, 2010] also contribute to theaetiology.

Increased genomic instability events as measured bymicronucleus (MN) frequency have been observed inperipheral tissues of other paediatric neurobiological con-ditions including Down’s syndrome [Maluf & Erdtmann,2001], ataxia telangiectasia [Tomanin et al., 1990] andBloom’s syndrome [Honmaa et al., 2002]. A geneticallyinherited defect in genome maintenance is expected toaffect viability and function of all tissues in the bodyincluding the brain. For example, mutations in theataxia telangiectasia-mutated gene increase chromoso-mal instability, micronuclei, immune dysfunction andneurodegeneration [Savitsky et al., 1995]. Genomic insta-bility has also been shown to be increased under condi-tions of folate and/or vitamin B12 deficiency [Blout et al.,1997; Fenech, 2001; Grafodatskaya et al., 2010], high psy-chosocial stress [Williams & Casanova, 2011b] and withnormal ageing [Fenech, Aitken, & Rinaldi, 1998; Fenech,Dreosti, & Rinaldi, 1997].

From the Sansom Institute of Health Research, University of South Australia, Adelaide, Australia (P.A.E.M., M.T.A., C.E.K.); Department of Animal, Foodand Health Sciences, Commonwealth Scientific and Industrial Research Organisation, Adelaide, Australia (P.A.E.M., P.T., M.F.F.); Finders University ofSouth Australia, Adelaide, Australia (R.Y.); School of Nursing and Midwifery, University of South Australia, Adelaide, Australia (A.E.); Centre for ResearchExcellence in Chronic Disease, James Cook University, Townsville, Australia (A.E.)

Received August 09, 2013; accepted for publication September 26, 2014Address for correspondence and reprints: Penelope A.E. Main, Department of Animal, Food and Health Sciences, Commonwealth Scientific and

Industrial Research Organisation, Gate 13, Kintore Avenue, Adelaide, 5000, Australia. E-mail: [email protected] sponsors: This research was funded through grants from the Channel 7 Children’s Research Foundation (Grant 08 138) (PT), School of

Psychology, Flinders University Research Grants Scheme (10 270) (RY) and the Fleurieu Philanthropy Foundation (Grant 2010) (PM). Additional fundingwas provided by the Commonwealth Scientific and Industrial Organisation Division of Animal, Food and Health Sciences.Published online in Wiley Online Library (wileyonlinelibrary.com)DOI: 10.1002/aur.1428© 2014 International Society for Autism Research, Wiley Periodicals, Inc.

RESEARCH ARTICLE

INSAR 1Autism Research ••: ••–••, 2014

rule out in utero effects or other types of DNA damage not measured by the CBMN-cyt assay. Autism Res 2015, 8:94–104. VC 2014 International Society for Autism Research, Wiley Periodicals, Inc.

Published online 4 November 2014 in Wiley Online Library (wileyonlinelibrary.com)

94 Autism Research 8: 94–104, 2015 INSAR

The main form of folate used by the body,5-methyltetrahydrofolate (5MTHF), is formed in avitamin B2-dependent reaction [Guenther et al., 1999;Leclerc, Sibani, & Rozen, 2005]. 5MTHF donates a methylgroup to homocysteine in a vitamin B12-dependentreaction to form methionine which is converted toS-adenosylmethionine, the main methyl donor in thebody [Zingg & Johnes, 1997]. Folate deficiency results inincreased uracil incorporation into DNA, leading tosingle- and double-strand DNA breaks, chromosomebreakage and ultimately to the formation of MN [Blout &Ames, 1995; Blout et al., 1997; Fenech, 2001].

The cytokinesis-block micronucleus cytome (CBMN-cyt) assay is one of the best validated methods of studyinggenomic instability [Fenech, 2007]. In this assay, DNAdamage events such as (a) MN, a biomarker of chromo-some breakage and/or whole chromosome loss, (b)nucleoplasmic bridges (NPB), a biomarker of DNAmisrepair and/or telomere end fusions, and (c) nuclearbuds (NBUDs), a biomarker for the elimination of ampli-fied DNA and/or DNA repair complexes, are measured inonce-divided cytokinesis-blocked binucleated (BN) cells.Cytotoxicity events are determined by the frequency ofnecrotic and/or apoptotic cells.

Some evidence suggests that children with autistic dis-order may be more susceptible to environmental toxins[Grandjean & Landrigan, 2006; Lawler, Croen, Grether,& Van de Water, 2004; Slotkin, Vevin, & Seidler, 2006]and/or exhibit defects in folate or glutathione metabo-lism [Adams et al., 2011; James et al., 2004, 2006; Pascaet al., 2009; Ramaekers, Blau, Sequeria, Nassogne, &Quadros, 2007] which are associated with increasedDNA damage [Fenech, 2001; Grafodatskaya et al., 2010;Liu et al., 2006; Lucero et al., 2000]. Furthermore, asgenomic instability is increased in other early onsetneurobiological disorders as previously mentioned[Honmaa et al., 2002; Maluf & Erdtmann, 2001;Tomanin et al., 1990], it was hypothesised that childrenwith autistic disorder may have increased genome insta-bility compared with children without autism and thatnutritional factors involved in DNA metabolism maydiffer compared with controls.

Methods

Study Design and Participants

Participants for this case–control study were recruitedthrough presentations at Autism South Australia parentsupport groups (PM), a letter to parents of childrenenrolled in the Flinders University Early InterventionResearch Program for Children with Autism, advertise-ments in the Autiser (a quarterly hard copy and electronicpublication produced for Autism SA clients) and the Mes-senger (a free local community newspaper). An article

was also placed in the Messenger and the University ofSouth Australia Magazine and an item about the studyappeared on the local Channel 7 television news pro-gramme. Controls were recruited through the Catholicschools system and flyers placed around local universi-ties, hospitals and medical research institutes. Parentswere provided with an information sheet and wererequired to sign an informed consent form for their chil-d(ren) to be enrolled into the study. Controls receivedremuneration for their participation.

Inclusion criteria

Cases. Cases were defined as children aged < 16 yearswho had received a diagnosis of autistic disorder usingDiagnostic and statistical manual of mental disorders: Textrevision (DSM-IVTR) [American Psychiatric Association,2000] by a multidisciplinary team of qualified healthprofessionals (e.g. paediatrician, psychologist, speechtherapist, occupational therapist).

Siblings. Siblings were defined as children with (a) asibling who had received a diagnosis of autistic disorderusing DSM-IVTR by a multidisciplinary team and (b) didnot display any social or communication deficits asscreened using the Social Communication Questionnaire(SCQ).

Controls. Controls were defined as children aged < 16years with (a) no family history of autism or pervasivedevelopmental disorder according to the DSM-IVTR, i.e.there were no siblings or first-degree relatives (includingcousins) with a diagnosis of ASD and (b) who did notdisplay any social or communication deficits as screenedusing the SCQ.

Exclusion criteria

All participants. Children who received supplementa-tion with high-dose vitamins (such as 400 mcg or morefolic acid or 200 mcg or more vitamin B12 or 64.5 mcg/kginjected methyl cobalamin daily) or drugs (such assodium valproate) that affect the folate/methionine/trans-sulphuration pathways were excluded from thestudy.

Cases. Children diagnosed with autistic disorder but notby a multidisciplinary team or who were diagnosed withautistic disorder but who presented with intellectualdisability caused by a comorbid diagnosis, e.g. Down’ssyndrome were excluded from the study.

Siblings. Children who had deficits in communication,socialisation and/or stereotypic/repetitive behaviours;intellectual disability; one or more comorbidities; geneticor a chromosomal abnormality were excluded from thestudy.

INSAR2 Main et al./Genomic instability in autismINSAR Main et al./Genomic instability in autism 95

Controls. Children who had received a diagnosis ofautism or who were not developing normally wereexcluded from the study.

All study procedures were approved by the Universityof South Australia’s Human Research Ethics Committee,The Flinders University of South Australia’s ClinicalResearch Ethics Committee, the Commonwealth Scien-tific and Industrial Organisation’s Animal, Food andHuman Health Research Ethics Committee and AutismSouth Australia’s Professional Practice Committee.

Demographic, Physical and Psychology Measures

Demographic information including date of birth,supplementation, medication, and socio-economic statuswas collected at the time of blood collection. Blood speci-mens were obtained during a visit to the child(ren)’shome after an overnight fast. Samples (∼ 20 mL) werewithdrawn from a cubital vein into blood tubes (contain-ing heparin, ethylenediaminetetraacetic acid (EDTA) oranticoagulant gel depending on the analysis the bloodwould be subject to) and immediately stored on ice beforebeing transported either to an Australian National Asso-ciation of Testing Authorities accredited pathology labo-ratory (either Gribbles or the Institute of Medical andVeterinary Science) or to the laboratories at the Animal,Food and Health Science Division at the CommonwealthScientific and Industrial Research Organisation, Adelaide.Serum (for measuring homocysteine and vitamins) wasseparated by centrifugation at 4200 rpm for 10 min at8°C within 1 hr of collection.

B vitamin and homocysteine status were measuredby accredited laboratories using previously describedmethods. Serum and erythrocyte folate and vitamin B12status were determined using chemiluminescent micro-particle immunoassay kits [Abbott Diagnostics Division,2010a, 2010b]. Vitamin B2 status was determined usinga validated high performance liquid chromatography(HPLC) kit [ChromSystems Diagnostics, 2011] andhomocysteine concentration was determined using adirect chemiluminescent competitive immunoassay kit[Siemans Healthcare Diagnostics, 2008].

Parents completed the Telethon Institute for ChildResearch’s Food Frequency Questionnaire for each childenrolled in the study which has been validated in 14 yearolds [Ambrosini, de Klerk, O’Sullivan, Beilin, & Oddy,2009]. The diet, macronutrient and micronutrient intakewas analysed using the Foodworks2009 program [XyrisSoftware, 2009].

In addition, a comprehensive psychometric assessmentof each participant was undertaken by a psychologistunder the direction of RY within 1 week of a blood draw.Parents completed the SCQ [Rutter, Bailey, & Lord, 2003],a screening instrument for ASDs and the Vineland Adap-tive Behaviour Scale (VABS) [Carter et al., 1998] which

measures adaptive functioning. Children were adminis-tered the Peabody Picture Vocabulary Test (PPVT)-ThirdEdition [Dunn & Dunn, 1981] and Raven Coloured Pro-gressive Matrices (RCPM) [Raven, Raven, & Court, 2004]which measure non-verbal IQ. As the RCPM is designedfor children aged 5 years or more, it was not administeredto younger children. In addition, case diagnosis was con-firmed using the Childhood Autism Rating Scale (CARS)which can distinguish children with autism from thegeneral population including other developmental disor-ders, as well as indicate the severity of presentation[Schopler, Reichler, & Renner, 1980].

Determination of the CBMN-cyt Assay Biomarkers

The CBMN-cyt assay was performed using duplicateblood cultures. Five hundred microlitres of heparinisedblood was added to 4.5 mL pre-warmeds (37°C) RPMI-1640 cell culture medium (Sigma, St Louis, MO, USA,Australia) supplemented with 10% fetal bovine serum(FBS) (Trace Biosciences, Brisbane, Australia), 1% sodiumpyruvate (Sigma) and 1% glutamine (Sigma) in 30 mLsterile polycarbonate sample containers tubes(Technoplas, St Marys, SA, Australia) and incubated in ahumidified atmosphere with 5% CO2 at 37°C (MCO-17AIC, Sanyo, Japan) (with the caps loose) until equili-brated.

Mitotic cell division was stimulated by the addition ofphytohemagglutinin (Abbott Murex, Dartford, UK) (finalconcentration of media 202.4 μg/mL). Following incuba-tion for 44 hr, cytochalasin B (Sigma) was added to thecell culture (final concentration of 6.0 μg/mL) to inhibitcytokinesis and the culture incubated for a further 24 hr,after which lymphocytes were harvested from the wholeblood culture.

Lymphocytes were isolated by overlaying the bloodculture onto Ficoll-Paque and centrifuging at 400 × g for30 min at 18–20°C. The lymphocyte layer located at theinterface of the Ficoll-Paque was removed and placed in afresh tube using a sterile plugged Pasteur pipette. Theleukocyte suspension was diluted with three times itsvolume of Hanks Balanced Buffer Solution (HBSS)(Thermo-Fisher, Scientific, Scoresby, Vic, Australia) andcentrifuged at 180 × g for 10 min. The supernatant wasdiscarded and the cell pellet resuspended in HBSS. Thesuspension of cells was centrifuged at 100 × g for 10 minat room temperature. The supernatant was discarded andthe cells resuspended in 300 μL culture medium to which22.4 μL dimethyl sulphoxide (Sigma) was added to facili-tate cell disaggregation.

For each duplicate culture, cells were harvested bypipetting 120 μL cell suspension into a cyto-centrifugecup (Shandon Products, Astmoor Runcom, Cheshire,UK) attached to a cyto-centrifuge filter card (ThermoElectron, Scoresby, Vic, Australia, USA) and microscope

3Main et al./Genomic instability in autismINSAR96 Main et al./Genomic instability in autism INSAR

slide and spun for 5 min at 600 rpm using a ShandonScientific cyto-centrifuge (Shandon Products) so thattwo spots were obtained for each culture. Slides were airdried for 30 min before being fixed and stained using acommercial kit (Diff Quick, Lab-Aids, Narrabeen, NSW,Australia), then air dried overnight before sealing withcover slips using DePex (VWR, BDH, Prolabo, Poole,Dorset, UK).

Samples were assigned a code prior to scoring by PT.CBMN-cyt biomarkers were measured visually by anexperienced scorer who did not have access to thecode, using a bright field microscope (Leica 20EB;Wetzlar, Germany) at 1000× magnification under oilimmersion. Established criteria were used for scoringDNA damage and cytotoxicity biomarkers as measured inthe CBMN-cyt assay [Fenech, 2007]. Briefly, 250 cells werescored from each spot (total 500 cells) to determinemono-nucleated, BN, multinucleated, apoptotic andnecrotic cell ratios. One thousand BN cells were scoredfrom each spot (total 2000 BN cells) for the followingDNA damage biomarkers: BN cells with MN (MNed cells),BN cells with NPB and BN cells with NBUDs.

Statistical Analysis

The distribution of continuous variables was checkedusing the Shapiro–Wilk test. The data are presented asmeans and standard deviations if normally distributedand as medians with interquartile ranges if skewed. Thestatistical significance of differences in biomarkers or psy-chology profile for age- and gender-matched case–controlpairs were determined using conditional logistic regres-sion. Robust Poisson regression was used instead whereconvergence was not achieved. Logistic regression withcluster analysis to take into account family relationshipwas used to determine the significance of differencesbetween case–sibling and sibling–control groups. Eachgroup included a variable number of children from theone family. The statistical analysis was performed usingthe clogit or logit, or vce cluster commands in the Statastatistical package version 11 (StataCorp) [StataCorp,2009]. A P value of < 0.05 was considered to be signifi-cant prior to a Holm–Bonferroni correction being appliedto correct for multiple comparisons. A chi-squared testwas applied to determine whether the proportion of par-ticipants with a serum folate reading of 46 μg/L or morewas significantly different between cohorts.

Multiple linear regression if normally distributed, orquantile regression if skewed, was applied to evaluate therelationship between the B vitamins/homocysteine andCBMN-cyt biomarkers and B vitamins/homocysteine/CBMN-cyt and autistic behavioural trait/psychologicalmeasurements for each cohort. The statistical analysis wasperformed using the regress command in the Stata statis-tical package version 11 [StataCorp, 2009] if the indepen-

dent variable was normally distributed and qreg if it wasskewed. A Holm–Bonferroni correction was applied whereP < 0.05 to correct for multiple comparisons. The coeffi-cient of determination (R2) is presented in the text toprovide an indication of the strength of the measuredassociation(s). The power of the study to detect a differ-ence of one micronuclei was determined using GraphPadStatMate version 2.00 for Windows [GraphPad, 2011].

Results

Recruitment

Sixty-one children with autistic disorder enrolled in ourstudy of which 15 were ineligible and 11 withdrew.Reasons for ineligibility were: not diagnosed with autisticdisorder by a multidisciplinary team (1), taking anti-epileptic drugs (2), supplementation with high dosevitamin B12 (5) or folate (7). Reasons for withdrawalincluded: change of address (1), child became too old forthe study (1), parent became too busy (2), child devel-oped needle phobia (3) and spouse/child withdrewconsent (4). Parents of a child(ren) with autistic disorderwho was eligible for participation were approached toenrol the child’s siblings in the study. All siblings enrolledin the study were eligible for inclusion and there wereno withdrawals. Twenty-eight children with no familyhistory of pervasive developmental disorders wereenrolled in the study as community controls of whichthree withdrew because the family moved (2) or the childbecame too old for the study (1). The number of controlswas sufficient to detect a difference of one MN at 85%power (two tailed) [GraphPad, 2011].

Participant Characteristics

Thirty-five children diagnosed with autistic disorder, 27unaffected siblings and 25 age- and gender-matched com-munity controls participated in the study. The mean ageand standard deviation of cases was 7.57 ± 2.92 years,siblings were 9.31 ± 4.81 years and controls 8.56 ± 2.84years. Thirty-three cases (94.3%), 15 siblings (55.6%) and23 controls (92%) were male. While there was no signifi-cant difference between cases and controls in age orgender (since they were matched), there was a significantdifference in gender but not age between cases and sib-lings (P = 0.002, R2 = 0.1626) and siblings and controls(P = 0.048, R2 = 0.1325). All participants were of Cauca-sian descent and there was no significant difference insocio-economic status (parental income) between thethree groups (data not shown).

There was no significant difference in supplementa-tion with multivitamin or mineral only preparationsbetween groups. Multivitamins were taken by 12 cases(34.3%), 9 siblings (33.3%) and 9 controls (36.0%) and

INSAR4 Main et al./Genomic instability in autismINSAR Main et al./Genomic instability in autism 97

mineral only supplements by 8 cases (22.9%), 2 siblings(7.4%) and 3 controls (12%). Polyunsaturated fatty acidswere taken by 16 cases (46%), 4 siblings (15%) and 5controls (20%), and probiotics were taken by 4 cases(11%), 2 siblings (7%) and no controls. Other supple-ments included amino acids, homeopathic preparationsand digestive enzymes which were only taken by someof the cases.

Medication taken by cases included a laxative (1), ananti-yeast agent (1), an angiotensin-converting enzymeinhibitor (1) risperidone (1) lithium oroate (1), oxytocinand bethanechol (1) and asthma medication (3). Foursiblings took regular medication comprising insulin (1)and asthma medication (3) and one community controltook asthma medication.

Psychological Profile

Autism diagnosis and severity was confirmed for casesusing the CARS, in which a score of 30 or more indicatesa diagnosis of autism. The median and interquartile rangeof CARS score was 32.5 (30.5–35) with the range being29–39.5 (Table 1).

The SCQ was administered to all participants in orderto be assured that none of the siblings or communitycontrols had an ASD. A score of 15 or higher indicatessignificant social and communication problems warrant-ing a diagnosis. The median and interquartile range forcases was 17 (12–21), for siblings 1.5 (1–4) and controls 3(1–5) (Table 1). Cases obtained significantly higher scoreson the SCQ than controls (P < 0.0001) and siblings(P < 0.0001, R2 = 0.7780); however, there was no signifi-cant difference between siblings and controls.

Each participant was also assessed for adaptive func-tioning using the VABS for which a higher score meansthat the child has a greater ability to function in everydaylife. The median and interquartile range for cases was67.5 (61–76), siblings 100.5 (95–110) and controls 109.0(99–116) (Table 1). Scores from cases were significantly

lower than for siblings (P = < 0.0001, R2 = 0.6768) andcontrols (P = 0.006, R2 = 0.6899); however, there was nosignificant difference between siblings and controls.

Two measures of IQ were administered as autistic dis-order presents with intellectual disability in about 50% ofcases [Morgan et al., 2002]. The PPVT measures receptivelanguage and is designed for children aged 2.5 years orolder. One case and two siblings were too young toadminister the test and a further four cases were unable tocomplete the test. The scores are presented as age- andgrade-based percentiles (Table 1). The median andinterquartile range for cases were 7 (0–37), siblings 61(45–86) and controls 68 (55–91). Cases scored signifi-cantly below siblings (P = < 0.016, R2 = 0.5907) and con-trols (P < 0.0001, R2 = 0.2815); however, there was nosignificant difference between siblings and controls.

The RCPM measures non-verbal IQ and is designed forchildren aged 5 years or over. Two cases, five siblings andone control were too young to do the test and a furtherten cases were unable to complete the test. The unad-justed median and interquartile range for cases were 23(16–23), siblings 32 (26–35) and controls 30 (26–33)(Table 1). Cases scored significantly lower than siblings(P = < 0.001, R2 = 0.1732) and controls (P = 0.012,R2 = 0.3845); however, there was no difference betweensiblings and controls.

B Vitamins and Homocysteine

Vitamin B2 was significantly higher in cases and siblingsthan controls (P = 0.027, R2 = 0.3345 and P = 0.016,R2 = 0.1572 respectively) although the association was nolonger significant after Holm–Bonferroni adjustment(Table 2). There was no significant difference betweencases and siblings. The means and standard deviations ofvitamin B2 were 344.8 ± 55.7 nmol/L FAD for cases,339.9 ± 35.8 nmol/L flavin adenine dinucleotide (FAD)for siblings and 303.7 ± 38.2 nmol/L FAD for controls.Vitamin B2 was significantly higher in male

Table 1. Psychological Profile (Median and Interquartile Ranges, P values for Group Comparisons)

Case Sibling Control P values

n = 35 n = 27 n = 25Case cfcontrol

Case cfsibling

Sibling cfcontrol

Childhood Autism Rating Scale 32.5 (30.5–35) – –Social Communication Questionnaire 17 (12–21) 1.5 (1–4) 3.0 (1–5) < 0.0001 < 0.0001 0.867Vineland Adaptive Behaviour Scale 67.5 (61–76) 100.5 (95–110) 109.0 (99–116) 0.006 < 0.0001 0.565Peabody Picture Vocabulary Testa 7 (0–37) 61 (45–86) 68 (55–91) 0.016 < 0.0001 0.546Raven Coloured Progressive Matricesb 23 (16–23) 32 (26–35) 30 (26–33) 0.012 0.001 0.578

cf, compared to.aPercentile score.bTotal score.Bold, statistically significant after Holm–Bonferroni adjustment.

5Main et al./Genomic instability in autismINSAR98 Main et al./Genomic instability in autism INSAR

(354.15 ± 23.58) than female siblings (324.33 ± 41.09)(P = 0.034). Supplementation status did not affectvitamin B2 levels.

Erythrocyte folate, vitamin B12 and homocysteinelevels were not significantly different between groups(Table 2). The median and interquartile ranges for eryth-rocyte folate, serum vitamin B12 and serum homocyste-ine were: 952.5 (827–1156) nmol/L, 429 (353–588)pmol/L and 6.9 (5.7–8.75) μmol/L respectively for cases;875 (784–1065) nmol/L, 405 (285–500) pmol/L and 6.7(5.6–7.7) for siblings; and 929 (767–1117) nmol/L, 422(356–491) pmol/L and 5.7 (5.0–7.0) μmol/L for controls.Gender and supplementation status did not affect eryth-rocyte folate, vitamin B12 or homocysteine levels.

Unfortunately, the upper limit for reporting serumfolate by both outsourced laboratories was 46 nmol/L;therefore, we were unable to calculate the median orinterquartile range for serum folate. Thirty-five partici-pants (18/35 cases, 14/27 siblings and 3/25 controls)had ≥ 46 nmol/L for serum folate. Chi-squared analysisshowed that the proportions of samples with serum folateat 46 nmol/L or higher was significantly differentbetween cases and controls (P = 0.0016), and betweensiblings and controls (P = 0.0028) but not between casesand siblings (Table 3).

CBMN-cyt Assay

There was no significant difference between groups forbiomarkers of cytotoxicity (Table 4). The median and

interquartile ranges for apoptosis were 1.3% (1.0–1.8) forcases; 1.6% (0.7–2.5) for siblings; and 1.1% (0.6–2.1)for controls. For necrosis, they were: 9.3% (6.6–10.6) forcases; 7.8% (6.1–10.0) for siblings; and 8.0% (6.3–14.5)for controls.

Likewise, there were no significant differences betweencohorts for any DNA damage biomarkers (Table 4). Themedian and interquartile ranges for the number of BNcells with MN were: 1.5 (1.0–2.5) for cases; 1.0 (0.5–2.0)for siblings; and 1.0 (0.5–2.0) for controls; for BN cellswith NPB, they were: 1.0 (0.5–1.5) for cases; 0.5 (0.0–1.0)for siblings; and 1.0 (0.5–1.6) for controls; and thenumber of BN cells with NBUDs were: 2.5 (1.5–3.0) forcases; 1.0 (1.0–4.0) for siblings; and 2.5 (1.5–3.5) forcontrols.

Association Between B Vitamins/Homocysteine and theCBMN-cyt Biomarkers

Apoptosis was positively associated with erythrocytefolate in cases (P < 0.001, R2 = 0.3427) (Table 5). Necrosiswas positively associated with vitamin B12 in controls(P < 0.001), R2 = 0.1078) and vitamin B2 in siblings(P = 0.012, R2 = 0.2114).

BN cells with MN were positively associated withvitamin B12 (P = 0.030, R2 = 0.1197) in cases. BN cellswith NBUDs were negatively associated with vitamin B2in cases (P < 0.001, 0.1570). There were no significantassociations between homocysteine and the CBMN-cytbiomarkers for any cohort.

Table 2. Comparison of B Vitamins and Homocysteine (Median and IQ Range, P values for Group Comparison)

Case Sibling Control

n = 35 n = 27 n = 25Case cfcontrol

Case cfsibling

Sibling cfcontrol

Serum folate (μg/L) 46 (37.2–46.0) 46 (39.6–46.0) 38.1 (35.9–41.0) 0.090 0.437 0.169Erythrocyte folate (nmol/L packed erythrocytes) 952.5 (827–1156) 875 (784–1065) 929 (767–1117) 0.819 0.685 0.768Serum vitamin B12 (pmol/L) 429 (353–588) 405 (285–500) 422 (356–491) 0.427 0.225 0.907Serum homocysteine (μmol/L) 6.9 (5.7–8.8) 0.133 0.291 0.341Whole blood vitamin B2 (nmol/L FAD)a 344.8 ± 55.7 339.8 ± 35.8 303.7 ± 38.2 0.027 0.659 0.016

cf, compared to.aMean ± standard deviation.FAD, flavin adenine dinucleotide.Bold, statistically significant before Holm–Bonferroni adjustment.

Table 3. Chi-Squared Test for the Maximum Serum Folate Reported by the Laboratory (nmol/L)

Case Sibling ControlCase cfcontrol

Case cfsibling

Sibling cfcontrol

Serum folate < 46 14 11 12 0.7492 0.9849 0.7766Serum folate ≥ 46 18 14 13

cf, compared to.

INSAR6 Main et al./Genomic instability in autismINSAR Main et al./Genomic instability in autism 99

Supplementation/Psychometric Measures and VitaminB/CBMN-cyt Status

There was a significant positive association betweensupplementation with multivitamins and the numberof BN cells with MN (P = 0.008, R2 = 0.0794) in cases(Table 6). A significant negative association was observedin controls between supplementation with mineralonly preparations and serum folate levels (P < 0.0001,R2 = 0.5101). There were no significant associationsbetween vitamin B/CBMN-cyt status for any psychomet-ric measure.

Discussion

The main finding of this study is that there was nosignificant difference in genome stability biomarkers

between children with autistic disorder relative to con-trols or siblings without autism (Table 3). This is consis-tent with in vitro modelling using lymphoblastoid celllines which showed no difference in genomic instabilityin autistic disorder cases and their non-autistic siblingswith or without a hydrogen peroxide or nitrosative stresschallenge [Main, Thomas, & Fenech, 2013], suggestingthat it is unlikely that DNA metabolism and repair aredeficient in autism. Furthermore, the DNA damage levelsinclusive of MN observed in our study are low comparedwith those for other paediatric neurological disorders[Honmaa et al., 2002; Maluf & Erdtmann, 2001; Tomaninet al., 1990] or young South Australian adults [Fenechet al., 1998], again indicating that it is unlikely thatintrinsic susceptibility to genome damage is a cause ofautism. We cannot exclude the possibility that (a)genome damage events occurring in utero may havecontributed to the neurological pathologies in autistic

Table 4. CBMN-cyt Biomarkers (Median and Interquartile Range, P values for Group Comparisons)

Case

%CVa

Sibling

%CVa

Control

%CVaCase cfcontrol

Case cfsibling

Sibling cfcontroln = 35 n = 27 n = 25

Apoptosis (%) 1.3 (1.0–1.8) 43.0 1.6 (0.7–2.5) 54.4 1.1 (0.6–2.1) 83.2 0.747 0.125 0.611Necrosis (%) 9.3 (6.6–10.6) 19.2 7.8 (6.1–10.0) 26.4 8.0 (6.3–14.5) 26.7 0.203 0.158 0.099BN with MN/1000 BN cells 1.5 (1.0–2.5) 47.1 1.0 (0.5–2.0) 70.7 1.0 (0.5–2.0) 47.1 0.839 0.369 0.889BN with NPB/1000 BN cells 1.0 (0.5–1.5) 58.9 0.5 (0.0–1.0) 14.1 1.0 (0.5–1.6) 70.7 0.275 0.825 0.727BN with NBUD/1000 BN cells 2.5 (1.5–3.0) 47.1 1.0 (1.0–4.0) 47.1 2.5 (1.5–3.5) 47.1 0.066 0.675 0.708

cf, compared to.aMedian coefficient of variation for duplicate measures.BN, binucleated cells; MN, micronuclei; NPB, nucleoplasmic bridges; NBUD, nuclear buds.Bold, statistically significant after Holm–Bonferroni adjustment.

Table 5. Association Between B Vitamins and Cytokinesis Block Micronucleus Cytome Assay Biomarkers

Cohort B vitamin OR Std error P value 95% confidence interval Pseudo R2

Apoptosis Case Erythrocyte folate 0.0023 0.0004 < 0.001 0.0015 0.0031 0.3453Necrosis Control Vitamin B12 0.0401 0.0097 < 0.001 0.0201 0.06–1 0.1078

Sibling Vitamin B2 0.0318 0.0117 0.012 0.0077 0.0559 0.2114Case Serum folate 0.1709 0.0611 0.009 0.0462 0.2955 0.1806

BN with MN/1000 BN cells Case Vitamin B12 0.0022 0.0010 0.030 0.0002 0.0042 0.1197BN with NBUD/1000 BN cells Sibling Serum folate −0.3030 0.0506 < 0.001 −0.4077 −0.1983 0.0201

Case Vitamin B2 −0.0143 0.0021 < 0.001 −0.0186 −0.0100 0.1570

BN, binucleated cells; MN, micronuclei; NBUD, nuclear buds.Bold, statistically significant after Holm–Bonferroni adjustment.

Table 6. Association Between Supplementation and Vitamin B/CBMN Cytome Biomarkers

Supplement Cohort Coefficient Std error P value 95% confidence interval Adjusted R2

Multivitamin CaseBN with MN/1000 BN 1.5 0.5327 0.008 0.4162 2.5838 0.0794

Mineral only ControlSerum folate −13.9 2.8990 < 0.0001 −19.8970 −7.9030 0.5101

BN, binucleated cells; MN, micronuclei.Bold, significant after Holm–Bonferroni adjustment.

7Main et al./Genomic instability in autismINSAR100 Main et al./Genomic instability in autism INSAR

disorder, and/or (b) that subtle damage to specific regionsof the genome such as the subtelomere or the mitochon-drial deletions may have occurred; however, these eventswere not investigated in the current study.

The correlation between B vitamins or homocysteinewith CBMN-cyt assay biomarkers did not show consistenttrends across groups and are therefore difficult to inter-pret. It was interesting, however, to observe a positivecorrelation between BN cells with MN and vitamin B12 orsupplementation with multivitamins in cases but not insiblings or controls. This result is the opposite to expec-tations based on cross sectional and B vitamin supple-mentation studies in adults [Thomas, Wu, Dhillon, &Fenech, 2011] suggesting that nutritional requirementsfor genome stability in children with autism may differfrom their non-autistic siblings or controls.

The marginally increased level of vitamin B2 observedin cases and their non-autistic siblings are intriguing.Increased vitamin B2 was reported in children withautistic disorder compared with published literature in aprevious study that used a microbiological assay(P = 0.0029) [Siva Sankar, 1979]; however, a more recentstudy found no significant difference in children withASDs relative to controls using a similar method [Adamset al., 2011]. In our study, vitamin B2 was negativelyassociated with NBUDs, a biomarker of gene amplifica-tion and/or the removal of DNA repair complexes, butthis was only evident in cases (P < 0.0001, R2 = 0.1570).This finding is consistent with an in vitro study conductedin our laboratory which showed: (a) high vitamin B2 andfolic acid concentration decreased the number of NBUDsrelative to conditions of low folate, irrespective of ribo-flavin status and (b) that NBUDs were significantlydecreased in TT homozygotes relative to CC homozygotesfor the MTHFR C677T mutation [Kimura, Umegaki,Higuchi, Thomas, & Fenech, 2004]. It is possible thatcases and siblings had a higher frequency of the TThomozygote genotype of MTHFR C677T. Although wedid not determine the genotype for MTHFR C677T forparticipants in our study, other researchers have reportedan increased frequency of the TT homozygote in childrenwith ASD [Boris, Goldblatt, Goalanko, & James, 2004; Liuet al., 2011; Mohammed et al., 2009].

Dietary intake, differences in intestinal permeabilityand differences in microflora of the proximal intestine arefactors that may also explain the increase in vitamin B2in cases and their siblings without autism. While therewas no significant difference in dietary intake of vitaminB2 (including supplementation) between cohorts, fibreintake was lower in children with autistic disorder andtheir siblings (P = 0.072, R2 = 0.1391 and P < 0.0001,R2 = 0.2057 respectively) (data not shown) suggestingthe possibility of a longer intestinal transit time anddecreased fermentative bacteria. A number of studieshave reported changes in intestinal flora [Adams,

Johansen, Powell, Quig, & Rubin, 2011; Finegold et al.,2002; Wang et al., 2011; Williams et al., 2011a] and/orincreased intestinal permeability in children withautism [D’Eufemia et al., 1996; Horvath, Papadimitriou,Rabsztyn, Drachenberg, & Tildon, 2000] suggesting thatthe higher level of vitamin B2 seen in children withautism and their siblings may reflect altered gutmicrobiota, longer transit time and/or bowel hyper-permeability. Increased levels of riboflavin are not patho-logical as riboflavin has a low level of absorption and isreadily excreted in urine [Rivlin, 2006]. Furthermore,there was no association between riboflavin and anybehavioural measure in this study.

As shown by other research groups [Adams et al., 2007,2011; Melnyk et al., 2012; Pasca et al., 2009], there wereno significant differences in folate (serum or erythrocyte),vitamin B12 or homocysteine between groups (Table 2).Standard laboratory tests are unable to discriminatebetween the different chemical forms of folate or vitaminB12 in body tissues, so it is possible that, although nosignificant difference in the overall level of these micro-nutrients was reported, some children with autism mayhave lower levels of the active forms of these micronutri-ents (5MTHF or the reduced active form of cobalaminrespectively) than children without autism.

The strengths of our study are that we used age- andgender-matched cases and community controls, linkedthe data to objective behavioural measures and used acomprehensive and well-validated DNA damage andcytotoxicity assay, i.e. the CBMN-cyt assay. The mainlimitation is the small size of the study which lacked thepower to assess the impact of genotype that may contrib-ute to differences in genomic stability and B vitaminsin metabolism [Boris et al., 2004; James et al., 2006]. Afurther limitation is that the coefficients of variation forduplicate measures of DNA damage markers and apopto-sis were sometimes higher than the recommended 40%for BN cells with MN [Fenech, 2007] because the CBMN-cyt measures were low (including BN cells with MN) andsometimes zero.

Further studies are needed to test the following hypoth-eses: (a) lymphocytes from children with autistic disorderexhibit significant differences in the expression of genesthat regulate the cell cycle compared with children withno family history of autism; and (b) increased wholeblood vitamin B2 in children with autistic disorder andtheir siblings is due to differences in gastrointestinal bac-teria, intestinal permeability and/or increased transittime.

Conclusion

Based on the results of this study, it is unlikely that DNAmetabolism and repair are substantially deficient in

INSAR8 Main et al./Genomic instability in autismINSAR Main et al./Genomic instability in autism 101

autism. The increased levels of riboflavin in the bloodfrom families with autistic disorder is an interesting newobservation that requires further investigation to deter-mine whether it can be replicated and to identify poten-tial mechanisms.

Statement of Author Contributions

PAEM was responsible for execution of the study, dataanalysis and interpretation, and all aspects of the manu-script. MFF was responsible for conceptualisation, plan-ning and assistance in interpreting the results and criticalreview of the paper. PT was responsible for grant writing,and assisted with planning, interpretation of the studyfindings and critical review of the paper, CEK assistedwith critical review of the paper, MTA played a key role inplanning and initiating the study, and AE providedadvice on the statistical aspects of the project. RY pro-vided advice on the psychological aspects of the project,as well as overseeing administration of the psychologytests. All authors read and approved the final manuscript.

Acknowledgments

The authors would like to express their gratitude to thefamilies in Adelaide and on the Fleurieu Peninsula whoseparticipation made this study possible. We also acknowl-edge the invaluable contribution of our colleagues,Abbey-Leigh Dalton and Maria Berekally, who conductedthe phlebotomy, Carolyn Salisbury who conducted theblood culture and scored the slides for the CBMN-cytbiomarkers, Melissa Rodi and Candy Fong who con-ducted the psychometric tests and scored them, and Mar-garet Miller and Pennie Taylor who assisted in theanalysis of the food frequency questionnaires as well asDavid Topping for advice along the way. There are noconflicts of interest.

References

Abbott Diagnostics Division. (2010a). Architect system B12 Ref7K61. Lisnamuck, Longford, Ireland: Abbott Ireland Diagnos-tics Division.

Abbott Diagnostics Division. (2010b). Architect System FolateRef 1P74. Longford, Ireland: Abbott Ireland DiagnosticsDivision.

Adams, J.B., Audhya, T., McDonough-Means, S., Rubin, R.A.,Quig, D., et al. (2011). Nutritional and metabolic status ofchildren with autism vs. neurotypical children, and the asso-ciation with autism severity. Nutrition and Metabolism, 8,1–32.

Adams, J.B., Johansen, L.J., Powell, L.D., Quig, D., & Rubin, R.A.(2011). Gastrointestinal flora and gastrointestinal status inchildren with autism—comparisons to typical children and

correlation with autism severity. BMC Gastroenterology, 11,22.

Adams, M., Lucock, M., Stuart, J., Fardell, S., Baker, K., & Ng, X.(2007). Preliminary evidence for involvement of the folategene polymorphism 19bp deletion-DHFR in occurrence ofautism. Neuroscience Letters, 422, 24–29.

Ambrosini, G.L., de Klerk, N.H., O’Sullivan, T.A., Beilin, L.J., &Oddy, W.H. (2009). The reliability of a food frequency ques-tionnaire for use among adolescents. European Journal ofClinical Nutrition, 63, 1251–1259.

American Psychiatric Association (2000). Diagnostic and statis-tical manual of mental disorders: Text revision (DSM-IVTR)(4th ed.). Washington, DC: American Psychiatric Association.

Badcock, C. (2011). The imprinted brain: How genes set thebalance between autism and psychosis. Epigenomics, 3, 345–359.

Blout, B.C., & Ames, B.N. (1995). DNA damage in folate defi-ciency. Baillieres Clinical Haematology, 8, 461–478.

Blout, B.C., Mack, M.M., Wehr, C.M., MacGregor, J.T., Hiatt,R.A., et al. (1997). Folate deficiency causes uracil misin-corporation into human DNA and chromosome breakage:Implications for cancer and neuronal damage. Proceedings ofthe National Academy of Sciences of the United States ofAmerica, 94, 3290–3295.

Boris, M., Goldblatt, A., Goalanko, J., & James, S.J. (2004). Asso-ciation of MTHFR gene variants with autism. Journal ofAmerican Physicians and Surgeons, 9, 106–108.

Carter, A.S., Volkmar, F.R., Sparrow, S.S., Wang, J.-J., Lord, C.,et al. (1998). The Vineland adaptive behavior scales: Supple-mentary norms for individuals with autism. Journal ofAutism and Developmental Disorders, 28, 287–302.

ChromSystems Diagnostics. (2011). HPLC for vitamin B2 inwhole blood Order number 37000. Munich, Germany:ChromSystems Diagnostics.

D’Eufemia, P., Celli, M., Finocchiaro, R., Pacifico, L., Viozzi, L.,et al. (1996). Abnormal intestinal permeability in childrenwith autism. Acta Paediatrica, 85, 1076–1079.

Dunn, L.M., & Dunn, D.M. (1981). Peabody Picture VocabularyTest—Revised.

Elsabbagh, M., Divan, G., Koh, G.-J., Kim, Y.S., Kauchali, S., et al.(2012). Global prevalence of autism and other pervasivedevelopmental disorders. Autism Research, 5, 160–179.

Fenech, M. (2001). The role of folic acid and vitamin B12 ingenomic stability of human cells. Mutation Research, 475,57–67.

Fenech, M. (2007). Cytokinesis-block micronucleus cytomeassay. Nature Protocols, 2, 1084–1104.

Fenech, M., Aitken, C., & Rinaldi, J. (1998). Folate, vitamin B12,homocysteine status and DNA damage in young Australianadults. Carcinogenesis, 19, 1163–1171.

Fenech, M.F., Dreosti, I.E., & Rinaldi, J.R. (1997). Folate, vitaminB12, homocysteine status and chromosome damage rate inlymphocytes of older men. Carcinogenesis, 18, 1329–1336.

Finegold, S.M., Molitoris, D., Song, Y., Liu, C., Vaisanen, M.L.,et al. (2002). Gastrointestinal microflora studies in late-onsetautism. Clinical Infectious Diseases, 35, S6–S16.

Glessner, J.T., Wang, K., Cai, G., Korvatska, O., Kim, C.E., et al.(2009). Autism genome-wide copy number variation revealsubiquitin and neuronal genes. Nature, 459, 569–573.

9Main et al./Genomic instability in autismINSAR102 Main et al./Genomic instability in autism INSAR

Grafodatskaya, D., Chung, B., Szatmari, P., & Weksberg, R.(2010). Autism spectrum disorders and epigenetics. Journal ofthe American Academy of Child and Adolescent Psychiatry,49, 794–809.

Grandjean, P., & Landrigan, P.J. (2006). Developmental neuro-toxicity of industrial chemicals. The Lancet, 368, 2167–2177.

GraphPad. (2011). StatMate version 2.0 for Windows. SanDiego, CA: GraphPad. Retrieved from http://www.graphpad.com/scientific-software/prism/

Guenther, B.H., Sheppard, C.A., Tran, P., Rozen, R., Matthews,R.G., & Ludwig, M.L. (1999). The structure and propertiesof methylenetetrahydrofolate reductase from Escherichiacoli suggests how folate ameliorates human hyperhomo-cysteinemia. Nature Structural and Molecular Biology, 6,359–365.

Honmaa, M., Tadokoro, S., Sakamoto, H., Tanabe, H., Sugimoto,M., et al. (2002). Chromosomal instability in B-lymphoblasotoid cell lines from Werner and Bloom syn-drome patients. Mutation Research, 520, 15–24.

Horvath, K., Papadimitriou, J.C., Rabsztyn, A., Drachenberg, C.,& Tildon, J.T. (2000). Gastrointestinal abnormalities in chil-dren with autistic disorder. Journal of Pediatrics, 135 559–563.

Hultman, C.M., Sandin, S., Levine, S.Z., Lichtenstein, P., &Reichenberg, A. (2010). Advancing age and risk of autism:New evidence from a population-based study and a meta-analysis of epidemiological studies. Molecular Psychiatry, 16,1203–1212.

James, S.J., Cutler, P., Melnyk, S., Jernigan, S., Janak, L., et al.(2004). Metabolic biomarkers of increased oxidative stressand impaired methylation capacity in children with autism.American Journal of Clinical Nutrition, 80, 1611–1617.

James, S.J., Melnyk, S., Jernigan, S., Cleves, M.A., Halsted, C.H.,et al. (2006). Metabolic endophenotype and related geno-types are associated with oxidative stress in children withautism. American Journal of Medical Genetics B Neuropsy-chiatric Genetics, 141, 947–956.

Kakinuma, H., & Sato, H. (2008). Copy number variations asso-ciated with autism spectrum disorder. Pharmacogenomics, 9,1143–1154.

Kimura, M., Umegaki, K., Higuchi, M., Thomas, P., & Fenech, M.(2004). Methylenetetrahydrofolate reductase C677T poly-morphism, folic acid and riboflavin are important determi-nants of genome stability in cultured human lymphocytes.Journal of Nutrition, 134, 48–56.

Lawler, C.P., Croen, L.A., Grether, J.K., & Van de Water, J. (2004).Identifying environmental contributions to autism: Provoca-tive clues and false leads. Mental Retardation and Develop-mental Disabilities Research Reviews, 10, 292–302.

Leclerc, D., Sibani, S., & Rozen, R. (2005). Molecular biologyof methylenetetrahydrofolate reductase (MTHFR) and over-view of mutation/polymorphisms. In P.M. Ueland & R. Rozen(Eds.), MTHFR polymorphisms and disease (pp. 1–20).Georgetown, TX: Landes Bioscience.

Liu, A.-L., Lu, W.-Q., Wang, Z.-Z., Chen, W.-H., Lu, W.-H.,et al. (2006). Elevated levels of urinary 8-hydroxy-2′-deoxyguanosine, lymphocytic micronuclei, and serum gluta-thione S-transferase in workers exposed to coke ovenemissions. Environmental Health Perspectives, 114, 673–677.

Liu, K., Solehdin, F., Cohen, I.L., Maripaz, G., Gonzalez, M.G.,et al. (2011). Population- and family-based studies associatethe MTHFR gene with idiopathic autism in simplex families.Journal of Autism and Developmental Disorders, 41, 938–944.

Lucero, L., Pastor, S., Suarez, S., Durban, R., Gomez, C., et al.(2000). Cytogenetic biomonitoring of Spanish greenhouseworkers exposed to pesticides: Micronuclei analysis in periph-eral blood lymphocytes and buccal epithelial cells. MutationResearch, 464, 255–262.

Main, P.A.E., Thomas, P., & Fenech, M. (2013). Necrosis isincreased in lymphoblastoid cell lines from children withautism compared to their non-autistic siblings under condi-tions of oxidative and nitrosative stress. Mutagenesis, 28 475–484.

Maluf, S.W., & Erdtmann, B. (2001). Genomic instability inDown Syndrome and Fanconi anemia assessed by micronu-clei analysis and single cell gel electrophoresis. Cancer Genet-ics and Cytogenetics, 124, 71–75.

Melnyk, S., Fuchs, G.J., Schulz, E., Lopez, M., Kahler, S.G., et al.(2012). Metabolic imbalance associated with methylationdysregulation and oxidative damage in children with autism.Journal of Autism and Developmental Disorders, 42, 367–377.

Mohammed, N.S., Jain, J.M., Chintakindi, K.P., Singh, R.P., Naik,U., & Akella, R.R. (2009). Aberrations in folate metabolicpathway and altered susceptibility to autism. PsychiatricGenetics, 19, 171–176.

Morgan, C.N., Roy, M., Nasr, A., Chance, P., Hand, M., et al.(2002). A community survey establishing the prevalence rateof autistic disorder in adults with learning disability. ThePsychiatrist, 26, 127–130. doi:10.1192/pb.26.4.127

Nordahl, C.W., Braunschweig, D., Iosif, A.-M., Lee, A., Rogers, S.,et al. (2013). Maternal autoantibodies are associated withabnormal brain enlargement in a subgroup of children withautism spectrum disorder. Brain, Behavior, and Immunity,30, 61–65.

Pasca, S.P., Dronca, E., Kaucsar, T., Craciun, E.C., Endreffy, E.,et al. (2009). One carbon metabolism disturbances and theC677T MTHFR gene polymorhism in children with autismspectrum disorders. Journal of Cellular and Molecular Medi-cine, 13, 4229–4238.

Pinto, D., Pagnamenta, A.T., Klei, L., Anney, R., Merico, D., et al.(2010). Functional impact of global rare copy number varia-tion in autism spectrum disorders. Nature, 466, 368–372.

Ramaekers, V.T., Blau, N., Sequeria, J.M., Nassogne, M.C., &Quadros, E.V. (2007). Folate receptor autoimmunity and cere-bral folate deficiency in low-functioning autism with neuro-logical defects. Neuropediatrics, 38, 276–281.

Raven, J., Raven, J.C., & Court, J.H. (2004). Manual for Raven’sProgressive Matrices and Vocabulary Scales.

Rivlin, R. (2006). Riboflavin. In B. Bowman & R.M. Russell (Eds.),Present knowledge in nutrition (Vol. 1, pp. 250–259). Wash-ington DC: ISLI Press.

Rutter, M., Bailey, A., & Lord, C. (2003). Social CommunicationQuestionnaire.

Savitsky, K., Bar-Shira, A., Gilad, S., Rotman, G., Ziv, Y., et al.(1995). A single ataxia telangiectasia gene with a productsimilar to PI-3 kinase. Science, 268, 1749–1753.

INSAR10 Main et al./Genomic instability in autismINSAR Main et al./Genomic instability in autism 103

Schopler, E., Reichler, R.J., & Renner, B.R. (1980). Toward objec-tive classification of childhood autism: Childhood AutismRating Scale (CARS). Journal of Autism and DevelopmentalDisorders, 10, 91–103.

Shanen, N.C. (2006). Epigenetics of autism spectrum disorders.Human Molecular Genetics, 15, R138–R150.

Siemans Healthcare Diagnostics. (2008). ADVIA Centaur Homo-cysteine Ref 124578. Tarrytown, NY: Siemans HealthcareDiagnostics.

Siva Sankar, D.V. (1979). Plasma levels of folates, riboflavin,vitamin B6 and ascorbate in severely disturbed children.Journal of Autism and Developmental Disorders, 9,73–82.

Slotkin, T.A., Vevin, E.D., & Seidler, F.J. (2006). Comparativedevelopmental neurotoxicity of organophosphate insecti-cides: Effects on brain development are separable fromsystemic toxicity. Environmental Health Perspectives, 114,746–751.

StataCorp. (2009). Stata Statistical Software: Release 11. CollegeStation, TX: StataCorp LP.

Thomas, P., Wu, J., Dhillon, V., & Fenech, M. (2011). Effect ofdietary intervention on human micronucleus frequency inlymphocytes and buccal cells. Mutagenesis, 26, 69–76.

Tomanin, R., Sarto, F., Mazzotti, D., Giacomelli, L., Raimondi, F.,& Trevisan, C. (1990). Louis-Bar syndrome: Spontaneous and

induced chromosomal aberrations in lymphocytes andmicronuclei in lymphocytes, oral mucosa and hair root cells.Human Genetics, 85, 31–38.

Wang, L., Christophersen, C.T., Sorich, M.J., Gerber, J.P., Angley,M.T., & Conlon, M.A. (2011). Low relative abundancesof the mucolytic bacterium Akkermansia muciniphila andBifidobacterium spp. in feces of children with autism.Applied and Environmental Microbiology, 77, 6718–6721.

Williams, B.L., Hornig, M., Buie, T., Bauman, M.L., Paik, M.C.,et al. (2011a). Impaired carbohydrate digestion and transportand mucosal dysbiosis in the intestines of children withautism and gastrointestinal disturbances. PLoS ONE, 6, e24585.

Williams, E.L., & Casanova, M.F. (2011b). Above genetics:Lesions from cerebral development in autism. TranslationalNeuroscience, 2, 106–120.

Xyris Software. (2009). Foodworks2009. High Gate Hill, Queen-sland, Australia: Xyris Software.

Zerbo, O., Issif, A.-M., Delwiche, L., Walker, C., &Hertz-Picciotto, I. (2011). Month of conception and risk ofautism. Epidemiology (Cambridge, Mass.), 22, 469–475.

Zingg, J.M., & Johnes, P.A. (1997). Genetic and epigeneticaspects of DNA methylation on genome expression, evolu-tion, mutation and carcinogenesis. Carcinogenesis, 18, 869–882.

11Main et al./Genomic instability in autismINSAR104 Main et al./Genomic instability in autism INSAR