Autism: Where Genetics Meets the Immune Systemdownloads.hindawi.com/journals/specialissues/908965.pdf · 2019. 8. 7. · healing in the skin, in the gastrointestinal, and respiratory

Autism: Where Genetics Meets the Immune System Guest Editors: Antonio M. Persico, Judy Van de Water, and Carlos A. Pardo

Autism Research and Treatment

Autism: Where Genetics Meetsthe Immune System

Autism Research and Treatment

Autism: Where Genetics Meetsthe Immune System

Guest Editors: Antonio M. Persico, Judy Van de Water,and Carlos A. Pardo

Copyright © 2012 Hindawi Publishing Corporation. All rights reserved.

This is a special issue published in “Autism Research and Treatment.” All articles are open access articles distributed under the CreativeCommons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the originalwork is properly cited.

Editorial Board

Jean-Louis Adrien, FranceM. Aman, USAElizabeth Aylward, USAR. F. Berman, USARoberto Canitano, ItalyManuel F. Casanova, USAJamel Chelly, FranceGeraldine Dawson, USAValsamma Eapen, Australia

Mohammad Ghaziuddin, USABarry Gordon, USAH. Jyonouchi, USAConnie Kasari, USAWalter E. Kaufmann, USABryan King, USAN. Kraus, USABennett L. Leventhal, USAM.-Reza Mohammadi, Iran

Klaus-Peter Ossenkopp, CanadaAlan K. Percy, USAMikhail V. Pletnikov, USAHerbert Roeyers, BelgiumJohn A. Sweeney, USAAkio Wakabayashi, UK

Contents

Autism: Where Genetics Meets the Immune System, Antonio M. Persico, Judy Van de Water,and Carlos A. PardoVolume 2012, Article ID 486359, 2 pages

Is Autism a Member of a Family of Diseases Resulting from Genetic/Cultural Mismatches? Implicationsfor Treatment and Prevention, Staci D. Bilbo, John P. Jones, and William ParkerVolume 2012, Article ID 910946, 11 pages

Prenatal and Postnatal Epigenetic Programming: Implications for GI, Immune, and Neuronal Functionin Autism, Mostafa I. Waly, Mady Hornig, Malav Trivedi, Nathaniel Hodgson, Radhika Kini, Akio Ohta,and Richard DethVolume 2012, Article ID 190930, 13 pages

Decreased Levels of EGF in Plasma of Children with Autism Spectrum Disorder, Charity Onore,Judy Van de Water, and Paul AshwoodVolume 2012, Article ID 205362, 4 pages

HLA Immune Function Genes in Autism, Anthony R. Torres, Jonna B. Westover, and Allen J. RosenspireVolume 2012, Article ID 959073, 13 pages

Intracellular and Extracellular Redox Status and Free Radical Generation in Primary Immune Cells fromChildren with Autism, Shannon Rose, Stepan Melnyk, Timothy A. Trusty, Oleksandra Pavliv, Lisa Seidel,Jingyun Li, Todd Nick, and S. Jill JamesVolume 2012, Article ID 986519, 10 pages

High Complement Factor I Activity in the Plasma of Children with Autism Spectrum Disorders,Naghi Momeni, Lars Brudin, Fatemeh Behnia, Berit Nordstrom, Ali Yosefi-Oudarji, Bengt Sivberg,Mohammad T. Joghataei, and Bengt L. PerssonVolume 2012, Article ID 868576, 6 pages

Hindawi Publishing CorporationAutism Research and TreatmentVolume 2012, Article ID 486359, 2 pagesdoi:10.1155/2012/486359

Editorial

Autism: Where Genetics Meets the Immune System

Antonio M. Persico,1, 2 Judy Van de Water,3 and Carlos A. Pardo4

1 Child and Adolescent Neuropsychiatry Unit, University Campus Bio-Medico, Via Alvaro del Portillo 21, 00128 Rome, Italy2 Laboratory of Molecular Psychiatry and Neurogenetics, University Campus Bio-Medico, Via Alvaro del Portillo 21,00128 Rome, Italy

3 Division of Rheumatology/Allergy and Clinical Immunology, UC Davis, 451 Health Science Drive, Suite 6510, GBSF, Davis,CA 95616, USA

4 Department of Neurology, Johns Hopkins University School of Medicine, 600 N. Wolfe Street, Baltimore, MD 21287, USA

Correspondence should be addressed to Antonio M. Persico, [email protected]

Received 24 July 2012; Accepted 24 July 2012

Copyright © 2012 Antonio M. Persico et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Individuals with autism often display immune abnormalitiesin the form of altered cytokine profiles, autoantibodies, andchanges in immune cell function. Are they mere “smoke”or real “fire”? Given what we now know about the cross-talk between the immune and nervous systems, one mustask: are these immune abnormalities simply bystandereffects of genetic/genomic variants directly responsible forabnormal neurodevelopment, or is there a sizable subgroupof patients with autism spectrum disorder (ASD), in which adysfunctional immune system plays a critical mediator rolein the pathogenic chain of events leading to the disorder?Is autism one of many genetic/genomic conditions, Downsyndrome representing the best-known example, accompa-nied by immune abnormalities not primarily responsiblefor changing the trajectory of neurodevelopment? Or couldconceivably an unfortunate combination of genetic/genomicsusceptibility and environmental factors converging onto thesame individual affect neurodevelopment through immune-mediated mechanisms, such as fetomaternal incompatibility,autoimmunity, abnormal neuroinflammation, and so on?

This special issue contains six contributions aimed ataddressing different aspects of this question.

The paper by N. Momeni et al. documents significantlyelevated plasma levels of factor I among ASD childrencompared to controls. Factor I is a plasma enzyme respon-sible for degrading complement factor 3b, which in turnis the major opsonin in the complement system enablingphagocytosis of microbial agents. Higher factor I levels canbe interpreted as either primary (i.e., conferring vulnerability

toward microbial infections) or as a secondary event partof a broader immunomodulatory response aiming to bluntan inflammatory process. F1 plasma levels suggest that thisprocess could play a more relevant role in males and in smallASD children.

The contribution by S. Rose et al. describes reducedglutathione-mediated redox/antioxidant capacity both insideprimary leukocytes and in the plasma of ASD children. Con-sistently with this intracellular and extracellular imbalance ofthe glutathione redox status, intracellular production of freeradicals is also enhanced, especially in 30% of ASD cases.This and several previous studies provide converging evi-dence of excessive oxidative stress likely leading to abnormalmitochondrial function in a similar percentage of autisticindividuals.

Along similar lines, M. I. Waly et al. direct their attentionon the complex array of pathophysiological consequencesproduced by enhanced oxidative stress. In addition toimbalanced glutathione redox status, reduced methioninesynthase activity is particularly interesting, as it resultsin lower availability of S-adenosyl-methionine and con-sequently blunted DNA methylation. Curiously, a similarabnormality underlies autistic features also in two entirelydifferent contexts, namely, children exposed prenatally tovalproic acid and in Rett syndrome girls, carrying MECP2mutations.

The review by S. D. Bilbo et al. presents a stimulatinghypothesis, linking ASD to inflammatory, allergic, andautoimmune diseases. These “hyperimmune” disorders are

2 Autism Research and Treatment

known to share steeply increasing incidence rates over the lastdecades, as well as significant loading in many families withchildren with autism. In the authors’ view, they also share aninappropriate activation of the immune system due to biomedepletion in postindustrial societies, possibly suggestingpreventive strategies based on biome reconstitution.

The paper by C. Onore et al. reports an impressivethreefold decrease in EGF plasma levels among small ASDchildren compared to controls. EGF is involved in woundhealing in the skin, in the gastrointestinal, and respiratorysystems, as well as in the central nervous system where thisneurotrophic factor supports both adult subventricular zoneand midbrain dopaminergic neurons, in addition to stem cellproliferation.

A. R. Torres et al. review immune abnormalities inautism, focusing on HLA gene roles as potential contributorsto autism pathogenesis through several distinct mechanisms.HLA genes or haplotypes have been found associated withseveral autoimmune diseases, but strong associations withautism have also been reported.

These contributions, while not providing the ultimateanswers, do add some new pieces to the autism puzzle,providing further support to the idea that immune abnor-malities may play a pathophysiologically relevant role in asubgroup of families with autistic children. Future studies ofthe interaction of the immune and central nervous system areneeded to clarify in more detail how neuroimmune mecha-nisms influence the enhancement of the neurodevelopmentaldisarray and disturbances of neurobehavioral trajectoriesdetermined by genetic and environmental factors. Theheuristic potential of this line of investigation, both in termsof prevention and clinical therapeutics, ensures continuedefforts until a more definitive answer can be given to ourinitial question, so crucial in today’s autism field.

Antonio M. PersicoJudy Van de Water

Carlos A. Pardo


Review Article

Is Autism a Member of a Family of Diseases Resulting fromGenetic/Cultural Mismatches? Implications forTreatment and Prevention

Staci D. Bilbo,1 John P. Jones,2 and William Parker2

1 Systems and Integrative Neuroscience Group, Department of Psychology and Neuroscience, Duke University, Durham,NC 27710, USA

2 Department of Surgery, Duke University Medical Center, Durham, NC 27710, USA

Correspondence should be addressed to William Parker, [email protected]

Received 12 September 2011; Revised 18 January 2012; Accepted 10 April 2012

Academic Editor: Antonio M. Persico

Copyright © 2012 Staci D. Bilbo et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Several lines of evidence support the view that autism is a typical member of a large family of immune-related, noninfectious,chronic diseases associated with postindustrial society. This family of diseases includes a wide range of inflammatory, allergic,and autoimmune diseases and results from consequences of genetic/culture mismatches which profoundly destabilize the immunesystem. Principle among these consequences is depletion of important components, particularly helminths, from the ecosystem ofthe human body, the human biome. Autism shares a wide range of features in common with this family of diseases, including thecontribution of genetics/epigenetics, the identification of disease-inducing triggers, the apparent role of immunity in pathogenesis,high prevalence, complex etiologies and manifestations, and potentially some aspects of epidemiology. Fortunately, using availableresources and technology, modern medicine has the potential to effectively reconstitute the human biome, thus treating or evenavoiding altogether the consequences of genetic/cultural mismatches which underpin this entire family of disease. Thus, if indeedautism is an epidemic of postindustrial society associated with immune hypersensitivity, we can expect that the disease is readilypreventable.

1. Introduction: Autism as a Member of a LargeFamily of Postindustrial EpidemicsInvolving a Hyperimmune Response

In this paper, we outline a paradigm that points towardautism as one disease among many other well-knowndiseases which all share a common origin and, mostlikely, a common prevention strategy [1]. These emerging,noninfectious diseases are all epidemics of postindustrialculture, which are absent in antiquity in any culture,and absent in present day developing cultures. Membersof this family of disease are invariably associated with ahyperimmune response and can have a very high preva-lence in postindustrial populations, with prevalence oftengreater than 0.1% and sometimes greater than 1.0%. Thesehyperimmune-associated diseases include a wide range ofallergic, autoimmune, and inflammatory diseases such as

lupus, multiple sclerosis, hay fever, appendicitis, chronicfatigue syndrome, inflammatory bowel disease, asthma,celiac disease, type 1 diabetes, Graves’ disease, some typesof eczema, and a wide range of food allergies. Here, basedon a variety of evidence, we argue that autism is yet anothermember of this family of diseases, despite the slownessof the medical community to recognize it as such. Sincethe pathogenesis of at least one form of autism has beendirectly linked to autoantibody production [2], and sinceseven out of every eight cases of severe autism are associatedwith antineuronal antibodies in the serum [3], this ideahas already been demonstrated, at least in part. Further,as described below, autism is associated with a wide rangeof immune abnormalities. Even though some children withautism appear to have “normal” immune systems, it is arguedthat adverse immune events early in development might leadto an autistic phenotype.


A primary consideration in this view of autism as ahyperimmune-associated disease is the connection betweenimmunity and brain development in general, and betweenhyperimmune responses and autism in particular. Althoughthese connections have little bearing on the overall modeldescribing induction of disease by immune hypersensitivityin postindustrial society, the connections provide reason toexpect that the developing brain is sensitive to the samepostindustrial changes in the immune system which areknown to affect virtually all other organs of the human body(e.g., kidneys, pancreas, epidermis, adult nervous tissues,large and small bowel, cardiovascular system, thyroid, lungs,and others). The next four sections will summarize muchof what is currently known regarding those very extensiveconnections.

2. Pervasive Immune SystemAbnormalities in Autism

Immune system abnormalities exist throughout the bodyand brain of autistic children. These include evidenceof brain specific auto-antibodies, altered T, B, and NKcell responses to antigen, altered cytokine production, anincreased incidence of allergies and other autoimmune dis-orders, and functional changes in brain glial cells (microgliaand astrocytes) [4–11]. Microglia are the primary immuno-competent cells of the brain and rapidly respond to anyinfection, injury or other perturbation of homeostasis viaa dynamic process of activation [12]. Notably, increasedmicroglial activation has been observed in several brainregions of autistic patients [13]. Once activated, glia producea wide number of immune signaling molecules, includingcytokines (e.g., interleukin [IL]-1β, tumor necrosis factor[TNF]-α), chemokines (e.g., monocyte chemoattractantprotein [MCP]-1), and other inducible factors (e.g., nitricoxide), which may profoundly influence neural function[14].

3. Immune System-Central NervousSystem Communication

Beyond its traditional role in host defense and tissue repair,the immune system is now considered a diffuse sensoryorgan that works in concert with the endocrine, metabolic,and nervous systems to achieve and maintain homeostasisthroughout the body [15, 16]. In essence, the immunesystem serves as an interface between the human body andthe environment, coordinating the response of the bodyto the environment. Immunocompetent cells are locatedthroughout every organ of the body, including the brain,and regular communication occurs between the centralnervous system and immune tissues during both healthand disease processes. Many excellent reviews have beenwritten on these topics [17–20]. Importantly, bidirectionalcommunication between the brain and immune system hassignificant consequences for plasticity mechanisms withinthe brain, including cognition and emotion, which aremarkedly altered in autism.

4. Glial Cells Direct NormalBrain Development

Microglia are the resident macrophages of the centralnervous system, are associated with the pathogenesis ofmany inflammatory diseases of the brain, and derive fromprimitive yolk sac macrophage precursors, which are ofmesodermal origin and enter the neuroectoderm duringembryogenesis [21]. Early in development, microglia arehighly mobile and primarily amoeboid, consistent withtheir role in the phagocytosis of apoptotic cells [22]. Theexpression of many cytokines within the developing brain,including IL-1β and TGF-β, depends on the presence ofamoeboid microglia [23]. Microglia transform into a highlybranched, ramified morphology by adulthood in most brainregions. This morphological transition occurs in parallelwith neural cell genesis and migration, synaptogenesis, andsynaptic pruning, suggesting functions for microglia ineach of these processes, though these are just beginningto be explored [24–26]. Oligodendrocytes myelinate axonsprimarily during the postnatal period, and disruption oftheir function can be profoundly debilitating as in thecase of periventricular leukomalacia leading to cerebralpalsy [27]. Astrocytes mediate synapse formation withinthe developing brain [28], in part via the secretion ofextracellular matrix proteins called thrombospondins (TSPs)[29, 30]. Alterations in spine density via a putative TSPmechanism are implicated in neurodevelopmental disorderssuch as Down’s and Rett Syndrome [31]. Notably, astrocytematuration marks the end of the perinatal synaptogenicperiod when the brain is most plastic [32].

5. Long-Term Consequences of ImmuneActivation during Early Development

Developmental or “fetal programming” is a growing field ofscience based on evidence that experiences during critical orsensitive periods of perinatal life may modulate or “program”the normal course of development, with the result that adultoutcomes, including behavior, are significantly and oftenpermanently altered [33]. Given the many ways in which theimmune system is important for normal brain development,the capacity for immune-inducing events to influence thelong-term trajectory and function of these processes is likelyprofound, perhaps more so than at any other stage of life[34]. Notably, there is strong evidence that early-life infectioncan permanently alter stress reactivity, disease susceptibility,and notably, vulnerability to cognitive and neuropsychiatricdisorders, including Alzheimer’s disease, Parkinson’s disease,schizophrenia, and autism [35–38]. One of the most dra-matic and direct illustrations of this effect is the induction ofautism-like symptoms in mice by stimulation of the maternalimmune system during pregnancy [39].

Neuroinflammatory-induced alterations in neurogene-sis, migration, axon growth, myelination, synaptogenesis,and dendritic spine density/function are all candidate mech-anisms by which long-term changes in behavior mightresult. Further, the long lifespan of microglia and their

Autism Research and Treatment 3

pattern of central nervous system colonization duringdevelopment suggest that these immune cells may haveparticular significance in terms of neuroinflammatory effectson developmental programming. Diverse early-life eventsmay have enduring consequences for the brain and behaviorof organisms via an influence on these resident immunecells, both by impacting their own intrinsic function as wellas their interactions with ongoing developmental processes[40, 41].

6. Approaching Autism as Another Member ofthe Family of Hyperimmune-AssociatedDiseases: Addressing the Underlying Causes

The study of individual postindustrial associated diseases hasoften failed to enlighten researchers as to the underlyingcause of diseases. While genetic factors associated with theepidemics of postindustrial society abound, these geneticfactors have not changed over many generations, and thusare not responsible for newly emerging epidemics of disease.Almost paradoxically, even when specific triggers are iden-tified, these often cannot be said to cause newly emergingepidemics. For example, peanuts and ragweed pollen aretwo well-known triggers for allergy, but these substanceshave been present in the environment for millennia withouttriggering disease in humans. Similarly, viral infections andpotentially some milk proteins [42] have been identified astriggers for particular autoimmune diseases, but these factorshave been present for much longer than the diseases theytrigger. Thus, as will be described in detail, it is not thegenetics of the disease which point toward a cure, and it isgenerally not identification of the triggers for disease thatis most helpful. Rather, we suggest that the most effectiveand insightful approach toward defeating autism for futuregenerations will be the neutralization of underlying factors inpostindustrial culture which render postindustrial humansuniquely susceptible to disease, despite the presence of thesame genetics and many of the same triggers as found inpreindustrial cultures. These factors have been identifiedas mismatches between normal human genetics and theenvironment of postindustrial culture. With that in mind,an examination of the factors in postindustrial culture whichtrigger the plagues of hyperimmune-associated diseases willbe presented.

7. Mismatches between Normal HumanGenetics and Postindustrial CultureLeading to Hyperimmune Epidemics:Biome Depletion

If indeed autism fits into the family of hyperimmune-associated diseases plaguing postindustrial culture, then thepathogenesis or origins of these diseases merit detailedanalysis in any consideration of autism. As shown in Figure 1,postindustrial culture has several mismatches with normalhuman genetics (known as “evolutionary mismatches”)which have a profoundly destabilizing effect on the human

immune system. This effect is closely intertwined with animpact on metabolism and transport, as will be describedlater. These mismatches of postmodern culture involve someof the most cherished attributes of postmodern culture,including the widespread use of toilets, water treatment facil-ities, indoor working environments, soaps and shampoos,modern medicine, and the loss of factors which demandexercise for survival (e.g., human-eating predators, hardshipin finding food and procuring shelter for most of thepopulation). These mismatches are not a target for therapy,because reversal of any of these mismatches is untenable.Thus, these mismatches can be described as belonging tothe “pretherapeutic zone” in the course of pathogenesis(Figure 1). However, these mismatches lead directly toreadily measurable consequences that, in turn, lead to diseaseif left untended. Fortunately, these direct consequences ofmismatches are readily treated or even avoided altogether.Given that these consequences can be treated or avoided,this part of the pathogenesis of hyperimmune disease canbe described as the “optimal therapeutic zone” (Figure 1).In this paper we describe how mismatches between cultureand genetics/biology have led to specific consequences whichcause immune system hypersensitivity and/or destabiliza-tion. All of these consequences are associated with a widerange of diseases that include allergic, autoimmune, andinflammatory diseases.

The single consequence of genetic/culture mismatchesthat most profoundly impacts hyperimmune-associated dis-ease has been described as a depletion of componentsfrom the natural ecosystem of the human body, or simply“biome depletion” (Figure 1). The use of modern medicine,toilets, water treatment facilities, and cleaning agents in ourculture has led to a substantial alteration or even loss ofliving organisms that have coevolved with humans. Threeprimary components of the biome which are potentiallyimportant have been affected [43]: (a) loss of interactionswith soil bacteria, (b) alterations of the microbiome, thenormal bacteria that are associated with the human body,and (c) loss of helminths, worms that typically inhabitthe gut of mammalian species. Substantial evidence pointstoward all three of these factors as being important instabilization of the human immune system. Although notwidely appreciated, several lines of evidence point verystrongly if not conclusively to the idea that helminths inparticular are necessary for a stable immune system. Thisevidence has been reviewed extensively [1, 43] and can bevery briefly summarized as follows.

(i) Helminths are found almost ubiquitously in prein-dustrial humans [44] and in nondomesticatedmammals, including chimpanzees [45]. Even whenhelminths cannot be found, the immune systemshows signs of stimulation by helminths [46].

(ii) Helminths produce and secrete a number ofmolecules which regulate the host immune system[47].

(iii) Addition of helminths to laboratory animals blocksallergic, inflammatory, and autoimmune diseases[47].


Postindustrial culturetoilets, water treatment,

modern medicine

Biomedepletion

Vitamin Ddeficiency

Uncontrolledstress

Exercise notessential for survival

Indoor work, soap,shampoo, bathing

Pretherapeuticzone

A wide range ofdemands on

transport andPhase I/Phase II

metabolism

Alteration oftransport and

Phase I/Phase IImetabolism

Genetics,epigenetics,“triggers”

Allergy,autoimmunity,

chronicinflammation

Risk ofcognitive

dysfunction

Immunehypersensitivity

and/or destabilization

Metabolicdysfunction,

toxicity issues,oxidative stress

Optimaltherapeutic

zone

Postoptimaltherapeutic

zone

Figure 1: The pathogenesis of epidemics of allergic, inflammatory, and autoimmune disease. Cultural mismatches with human biology (thepretherapeutic zone) have consequences that can be readily avoided (the optimal therapeutic zone). If biome depletion in particular is notavoided, the individual is left susceptible to a wide range of often complex immune-related diseases (the postoptimal therapeutic zone),sometimes associated with imbalances in metabolite transport and/or Phase I/Phase II metabolism. Other consequences such as vitaminD deficiency and uncontrolled stress may, in some but not all cases, add to the problem. Factors contributing to disease which are oftennaturally occurring and generally difficult or impossible to avoid are underlined and fit within the postoptimal therapeutic zone.

(iv) Addition of helminths to humans cures inflamma-tory bowel disease [48] and stops the progression ofmultiple sclerosis [49].

(v) The effects of biome depletion on the immunesystem are not necessarily instantaneous becauselong-term effects of immune interactions with thebiome can span decades or, through epigenetics, evengenerations [1]. Nevertheless, the epidemiology ofhyperimmune disease reveals an inverse correlationbetween helminths and disease.

(vi) Humans with helminths and rodents colonized withhelminths have profoundly less reactive immune sys-tems than do either without helminths [44, 50, 51].

(vii) A long coevolutionary history of helminths and ver-tebrates, in conjunction with the potent immunoreg-ulatory effects of the former, is consistent with theidea that the human immune system is literally,physically dependent on helminths [1].

We propose that biome depletion is changing theimmune system at a population-wide level, which inturn will undoubtedly impact the brain, particularly dur-ing sensitive periods of development. Importantly, thismodel does not compete with any prior hypotheses orknown risk factors for neuroinflammatory conditions, butrather complements them. For instance, biome deple-tion may set the stage for exaggerated levels of neu-roinflammation, which interacts with other known riskfactors or triggers (psychological or metabolic stress,

infection, genetics or epigenetics) in the induction ofautism.

8. Mismatches between Normal HumanGenetics and Post-Industrial CultureLeading to Hyperimmune Epidemics:Vitamin D Deficiency

Based on the available information outlined above, itseems very likely that biome depletion is by far the mostsignificant and most profoundly influential consequenceof genetic/cultural mismatches in postindustrial society.However, other consequences which apparently exacerbatethe effects of biome depletion are evident. For example,vitamin D deficiency has reached epidemic proportions inpostindustrial society, is known to impact inflammationduring early development [52], and, like biome depletion,has been linked to a spectrum of allergic, autoimmune, andinflammatory diseases [53–55]. Further, epidemiology andother circumstantial evidence link vitamin D deficiency toautism [56, 57]. Interestingly, the widely publicized asso-ciation between rainfall and autism [58] can be accountedfor, perhaps more than in part, by decreased exposure tosunshine in areas with high rainfall, and the subsequentimpact of lower vitamin D levels.

Vitamin D deficiency is another consequence ofpostindustrial culture’s mismatches with human biology(Figure 1). Vitamin D production requires exposure of oilson the body’s surface to sunlight. The resulting photochem-ical reaction and production of vitamin D is greatly reduced


by components of postindustrial culture that reduce expo-sure to sunlight (e.g., indoor work environments, sunscreen).

The association of vitamin D deficiency with a spectrumof hyperimmune-associated diseases places this consequenceof biology/culture mismatch alongside biome depletion as anunderlying agent destabilizing immune system function inpostindustrial populations. However, vitamin D deficiency isan ancient problem, as indicated by the ancient occurrenceof rickets, whereas epidemics of hyperimmune-associateddiseases are more recent in nature. Thus, it seems likelythat vitamin D deficiency is a contributor to hyperimmune-associated disease, but not sufficient by itself to lead todisease. Further, the prevalence of vitamin D deficiency inpostindustrial society is somewhat less than the prevalenceof biome depletion: although the prevalence of vitamin Ddeficiency is substantial, with one study finding 24% ofindividuals deficient (≤20 ng/mL) with an additional 34%having levels that were insufficient (≤29 ng/mL) [59], theprevalence of biome depletion in postindustrial populationsis essentially 100%. Thus, it is expected that the impact ofvitamin D deficiency on immunity in postmodern culturemay be a matter of exacerbating the problems associated withbiome depletion more so than a significant problem whenconsidered in isolation.

Some mechanisms by which vitamin D might stabilizethe immune system have been elucidated. Vitamin D acts asa signaling molecule to enhance immune cell function in thepresence of infection and is important for maintaining thenormal interface between the microbiome and the immunesystem [60]. However, the mechanisms by which vitamin Dinteracts within the body are complex, with the moleculebeing intertwined either directly or indirectly in virtuallyall aspects of human biology. The active form of vitaminD binds and activates the Vitamin D receptor, altering theexpression of a variety of genes involved in such diverseprocesses as cell growth and proliferation, bone remodeling,and calcium homeostasis [61]. It has been estimated thatvitamin D binds human DNA in more than 2,500 placesand changes the expression of more than 200 genes [62].Thus, while it is known that vitamin D is required for normalimmune function, it is likely that vitamin D deficiencycould cause a general destabilization of the human biomeindependent of the immune system, particularly during earlydevelopment. Fortunately, regardless of the extent to whichvitamin D deficiency destabilizes the human biome, dietarysupplements are readily available, and thus this consequenceof culture/biology mismatch is easily avoided.

9. Mismatches between Normal HumanGenetics and Postindustrial CultureLeading to Hyperimmune Epidemics:Other Factors

The above discussion points out two factors, primarilybiome depletion and secondarily vitamin D deficiency,which generally destabilize the human immune system.These factors can be identified by their association witha wide range of hyperimmune-associated diseases. Other

consequences of postindustrial culture can also be foundwhich are associated with a wide range of disease, but limitsin space prevent a complete discussion. For example, chronicuncontrollable psychosocial stress is yet another factor that isgenerally associated with a range of hyperimmune-associateddiseases that include allergy, autoimmunity, and inflamma-tion [63–66]. Such stress derives from a complex range ofbiology/culture mismatches, including, for example, the lossof physical exercise [67–69] as a requirement for survival inpostindustrial culture (Figure 1). As another example of con-sequences of biology/culture mismatches which destabilizesimmune function, deprivation from mother’s milk duringinfancy (not included in Figure 1) is associated with a widerange of allergic and autoimmune conditions [70–74]. Thisgenetic/culture mismatch, like other mismatches, involves anaspect of postindustrial society which should not be reversed(the survival of infants despite the unavailability of mother’smilk), and thus fits into the “pretherapeutic zone” ofFigure 1. The extent to which factors such as uncontrollablestress and lack of mother’s milk can contribute to diseasein the presence of an otherwise normal biome remains tobe elucidated, although, as outlined above, biome depletionappears to be the single most important factor contributingto the increased incidence of hyperimmune-associated dis-eases.

10. Metabolism and Transport of Toxins andHyperimmune-Associated Disease

The connection between metabolism and immunity isimportant in a consideration of hyperimmune-associateddiseases. A variety of enzymes have been identified which areuseful in the biotransformation of endogenous (producedby the body or the associated biome) as well as exogenous(pharmaceutical) substrates. This system of enzymes pro-vides the body with a means of processing a wide rangeof substances, some of which are toxins, and is profoundlyaffected by the immune system [75]. The metabolic portionof this system is readily divided into two components:Phase I biotransformation includes the oxidative reactionsof the microsomal cytochrome P450 monooxygenase systemthat is expressed in the liver and gastrointestinal tract,and to some extent in a variety of extrahepatic tissues.Phase II biotransformation includes conjugation reactions ofsubstrates with endogenous cofactors such as glucuronate,sulphate, acetate, or glycine.

In addition to Phase I and Phase II biotransformation,the transport of their substrates and byproducts is also animportant component of this system. A variety of membranespanning polypeptides have been identified in the liver,gastrointestinal tract, kidney as well as other tissues in thebody, each responsible for the active trafficking of substratesacross cell membranes. These transport proteins are in factregulated by the same pathways that regulate the Phase I andII enzymes, and they can be thought of as being regulated intandem [76–80].

Immune activation results in a downregulation of PhaseI components [81, 82], transporters [83, 84] and potentially


Phase II components in what might be viewed as a tradeoffbetween immunity and metabolism. In time of duress, whenthe immune system is geared to inflammation and defense,the energy devoted to metabolism/transport may need tobe temporarily diverted to host defense. This is apparentlyan effective strategy, but as a consequence, chronic immuneactivation could result in metabolic difficulties as outlinedin Figure 1. Unfortunately, vitamin D deficiency is expectedto make this problem worse, since vitamin D is requiredto upregulate both Phase I and Phase II components [85]of metabolism. Further, accumulation of toxins as a resultof reduced Phase I, Phase II, or transport activity mayalso stimulate the immune system, adding further to thepropensity for hyperimmune activity.

11. Implications of Autism as a Member ofthe Family of Hyperimmune-Associated,Postindustrial Diseases

The implications of this model are several-fold and deservecareful consideration. The consideration of some possibleassertions that are not supported by this model is probablyequally as important.

(i) This model indicates that prophylactic normalizationof the biome will result in a profound decreasein the incidence of autism. Results with otherhyperimmune-associated diseases have been promis-ing, although a true normalization of the biome asa means of prevention has never been attempted forany disease.

(ii) This model is consistent with the idea that a varietyof factors affecting immunity and/or metabolism/transport may strongly influence the incidence ofautism.

(iii) There are several potential therapeutic approachesthat this model does not suggest which will necessar-ily be successful. For example, this model does notpredict that helminths will necessarily work as a curefor autism. Whether helminths can be used to treatautism will depend on the extent to which autism isreversible and the extent to which ongoing problemswith biome stability can be reversed.

(iv) Although porcine whipworms are currently beingused to treat the effects of biome depletion in somestudies, the model described in this paper doesnot predict that the porcine whipworm will preventautism, even prophylactically. (Although the use ofporcine whipworms is currently underway for a fewclinical studies, the use of porcine whipworms as ameans of prophylaxis is probably not economicallyfeasible because the costs of giving the necessarybiweekly doses would be prohibitive. This problemis not expected when using helminths adapted tohumans, since one dose lasts for years.) It might, butthis potentially depends on the extent to which thisspecies recapitulates the presence of a normal biome.

Thus, if porcine whipworms do not work, it does notsuggest that the hypothesis is wrong.

(v) This model does not suggest that helminths arethe only organisms involved in biome depletion.Normalization of the microbiome may also be acritical factor. In the face of modern practices inobstetrics and of widespread antibiotic use, this is anissue that must be considered.

(vi) Pharmaceuticals may never prove adequate. Thecomplexity of the immune system and the effects ofbiome depletion suggest that efforts to readjust thesystem using drugs may be naive and overly opti-mistic. Biome reconstitution may be the easiest andmost straight forward way to bring the system intohomeostasis. Similarly, metabolic factors may noteasily be compensated for by pharmaceuticals. Nor-malization and regulation rather than compensationand suppression are probably the better approaches.

(vii) Genetics may be relatively unimportant from aclinical perspective because it may not provide ameans to prevent or treat disease. Triggers may alsobe unimportant from a clinical perspective because(a) they may simply be unavoidable, and (b) evenif triggers for a particular disease can be identifiedand avoided, such efforts may simply decrease theprevalence of one disease at the expense of anincreased prevalence of another.

12. The Required Approach

Clinical trials are urgently needed to conclusively test avariety of questions. The approach needs to be extensiveand systematic rather than timid and piecemeal. The testsrequired will be on a large scale, although this scale palesin comparison to the current amount of effort and energybeing poured into research on allergic, inflammatory, andautoimmune issues. This approach to medicine will notlikely be supported by any committee of experts containingmembers with vested interests in traditional immunologicalapproaches (either in humans or in rodents) or in thepharmaceutical industry. Rather, this approach is much morelikely to be appreciated by those individuals with a broadunderstanding of biology and whose primary interest is inimproving health care and, ultimately, the health of thepopulation as a whole. With that in mind, the followingquestions need to be answered.

(i) Can autism be cured or effectively treated withbiome reconstitution, or only prevented by biomereconstitution?

(ii) How important for children is the biome of themother prior to and during pregnancy, and whilebreastfeeding? (Does prevention require reconstitu-tion of the mother’s biome?)

(iii) Which helminths or combination of helminths isboth safe and effective? Although some concernsdealing with the use of helminths for therapy have


been expressed, these objections are easily dealt with[1]. Several candidate species are well adapted tohumans, asymptomatic in low numbers, and thus areobvious candidates for initial and immediate test-ing. Further, new technologies might be developedwhich could improve on species that are naturallyoccurring. For example, irradiation of organisms toachieve sterility and eliminate the possibility of trans-mission, or cultivation of human-specific helminthsin ultraclean and immunodeficient rodents or evenin vitro might improve the utility of helminths forwidespread medical use.

(iv) What medical conditions such as a suppressedimmune system, anemia, or coagulopathy might becontraindications for biome reconstitution?

(v) Does optimal biome reconstitution vary with humangenotype or other factors such as age, gender, andbody size/composition?

(vi) In addition to the restoration of helminths to thebiome, what steps should be taken for restoration andmaintenance of the microbiome?

(vii) How will a reconstituted biome affect other areasof medicine, including aging, immunosuppression,infectious disease, cancer biology, and vaccine tech-nology?

13. Is Autism a Postindustrial,Hyperimmune-Associated Epidemic?

The idea that autism is a result of evolutionary mismatchesis predicated on the idea that autism is epidemic in postin-dustrial society. Thus, if the incidence of autism is the samein preindustrial societies as it is in postindustrial societies,then biome depletion is not a factor in the pathogenesisof autism, and biome reconstitution will not affect theincidence of autism. The epidemiology of autism has beenhotly debated. Some changes in diagnostic criteria andawareness of autism have certainly affected changes in thereported incidence of autism over time. However, debatesregarding the changing diagnosis of autism over the past fewdecades have little bearing on the idea that biome depletionprofoundly influences autism. Rather, the pertinent questionis whether the prevalence of autism has changed over thepast 150 years as various cultures have transitioned frompre to postindustrial over the course of several generations.Although it may be very difficult if not impossible to estimatethe prevalence of autism prior to the industrial revolutionin countries that are currently postindustrial, it might stillbe possible to examine the prevalence of autism in somepreindustrial societies. One very recent study by Schieve etal. is particularly informative in this regard: using a phonesurvey technique to obtain data on 4,690 Hispanic children,a “striking heterogeneity” in the prevalence of autismspectrum disorder between US-born Hispanic children with2 US-born parents (autism spectrum disorder prevalence2.39%) and otherwise similar children with 2 foreign-born parents (autism spectrum disorder prevalence 0.31%;

P = 0.05) that autism is a consequence of evolutionarymismatches in postindustrial society and is consistent withthe idea that the parental immune status is important inthe development of autism. Further assessment of autismin preindustrial cultures may, however, prove unnecessary;the easiest way to evaluate the idea that autism is relatedto biome depletion is probably to conduct the experiment.Whether prophylactic biome reconstitution does or doesnot affect the incidence of autism will answer all questionsconclusively, just as the effects of biome reconstitution[49, 87] and vitamin D supplementation [88] on multiplesclerosis, if confirmed by future studies, will effectively endlong-standing debates about the etiology of that disease.

The idea that autism is a result of evolutionary mis-matches is predicated on the idea that autism often occursin genetically “normal” individuals. Thus, if autism canbe attributed to genetic factors which dictate the presenceof disease even in individuals living in a preindustrialsociety, then biome depletion is probably not a factor inthe pathogenesis of autism, and biome reconstitution willprobably not affect the incidence of autism. Certainly thereare cases where autism is associated with genetic mutationswhich are deleterious, such as fragile X syndrome, but atpresent most individuals with autism appear geneticallynormal [89]. However, it remains unknown what percentageof autism can be attributed directly to genetics, and the ideathat biome depletion and other evolutionary mismatchesaffect mutation rates is speculative and unexplored. Again,the most effective approach to resolving the debate regardingthe role of genetics may be to prophylactically reconstitutethe biome of a sample population and await the results.

The epidemiology and the genetics of autism areinextricably woven together. It is anticipated that one oftwo pictures will emerge. Either a postindustrial epidemicof autism with normal genetics will be identified, orautism will be eventually defined as a “normal” part ofthe human condition, potentially associated with certaingenes or combinations of genes that predispose to diseaseregardless of the environment. In the former view, autismis due to an evolutionary mismatch, whereas in the latterview, autism is due to evolution itself. It is of criticalimportance not to accept the latter view prematurely sincesuch acceptance may disable efforts at compensating forevolutionary mismatches or discourage work on identifyingpostindustrial triggers for disease, with potentially tragicconsequences. Fortunately, several lines of evidence argueagainst the idea that autism is a “normal” part of humanbiology. First, a general consideration of the nature of humanevolution argues that evolution is not responsible for autism.The social interactions disrupted by autism are fundamentalto survival of not only humans, but also a wide range ofmammalian and nonmammalian species, pointing to theancient evolutionary origins of those interactions. Given theintense selection pressures on early human populations, itseems unlikely that an error rate of 1% in social interactionswould have developed and then persisted for many tensof thousands of years. However, the idea that autism wastolerated during human evolution because it is essentially atradeoff, an inherent “side-effect” of human brain function


that is required for survival, cannot be ruled out. Asecond line of evidence pointing in favor of autism as aresult of evolutionary mismatch rather than evolution isthe study, described above, showing that parental birth ina postindustrial culture seems to predispose offspring toautism [86]. Third, the observations (a) that the immunesystem is profoundly destabilized by postindustrial cultureand (b) that brain development is closely tied with immunefunction, when considered together, argue strongly for theidea that it is an evolutionary mismatch, not evolution,which is responsible for autism. Indeed, to the extentthat human brain development is strongly tied to theimmune system, it would be difficult to explain the ideathat postindustrial culture and the resulting consequencesof evolutionary mismatches such as biome depletion donot profoundly affect brain development. Finally, the strongassociation of autism with autoantibody production [3]and other immune abnormalities points toward the etiologyof autism as straight forward: another epidemic in thelong list of hyperimmune associated epidemics caused bybiome depletion in postindustrial society. The fact that thisparticular disease affects primarily the brain rather thanother organs or tissues undoubtedly leads to the tremendouscomplexity in the pathology associated with autism, but thiscomplexity should not blind the medical community to thesimple roots that apparently underlie epidemics of all allergicand autoimmune diseases, or to the apparent solutions thatcan currently be implemented to prevent those diseases [1].

14. Summary/Conclusions

Autism has many of the characteristic features ofhyperimmune-associated diseases which result frombiome depletion and other consequences of biology/culturalmismatches in postindustrial culture, potentially sharinga common cause and potentially a common solution withthose diseases. The identification of triggers for autism(e.g., association with viral infection during pregnancy[90]) parallels that of other diseases in this family. However,the trigger that causes disease is often unrelated to thecultural mismatches with human biology which destabilizethe human biome and make the trigger dangerous. Theassociation of autism with genetic and probably epigeneticfactors also parallels that of other hyperimmune-associateddiseases, but offers little in the hope of prevention, treatment,or a cure.

The ecosystem of the human body, the human biome,is a tightly interwoven collection of factors which includesimmunity, brain function, sunlight, hormones, metabolism,transport of metabolites, the microbiome, and helminths. Itstands firmly to reason that if one or more of these com-ponents are profoundly altered or even deleted, the entiresystem may become destabilized. This view of ecosystemshas long been appreciated by ecologists [91]. Although somehyperimmune diseases might be eliminated if the respectivetriggers for those diseases are identified and avoided, suchas approach cannot be successful in eliminating the overallproblem; roughly 50% of children in postindustrial societyhave some form of chronic health condition [92], many

millions are affected by a wide range of allergic, inflam-matory, and autoimmune disorders, and inflammation as aresult of biome destabilization may play a key role in suchcommon maladies as heart disease and dementia. With thatin mind, the need for normalization of the biome cannot beoveremphasized, regardless of particular diseases that mightbe prevented by identifying and avoiding the environmentaltrigger. Thus, while it is hoped that a trigger that causesautism might be confidently identified in the near future, andnew cases of autism might be greatly reduced as a result, it ishypothesized that normalization of the human biome will benecessary to avoid the full milieu of chronic immune-relateddiseases that plague postindustrial society.

Acknowledgments

The authors gratefully acknowledge the contribution of theResearch Coordination Network in Ecological Immunologyfor facilitating discussion, and the Coalition for SafeMindsfor providing the publication fee. The authors also thankSharon L. Brenner, Zoie E. Holzknecht, and Susanne Meza-Keuthen for helpful discussion.

References

[1] S. D. Bilbo, G. A. Wray, S. E. Perkins, and W. Parker,“Reconstitution of the human biome as the most reasonablesolution for epidemics of allergic and autoimmune diseases,”Medical Hypotheses, vol. 77, no. 4, pp. 494–504, 2011.

[2] V. T. Ramaekers, N. Blau, J. M. Sequeira, M. C. Nassogne, andE. V. Quadros, “Folate receptor autoimmunity and cerebralfolate deficiency in low-functioning autism with neurologicaldeficits,” Neuropediatrics, vol. 38, no. 6, pp. 276–281, 2007.

[3] G. A. Mostafa and L. Y. Al-Ayadhi, “The relationship betweenthe increased frequency of serum antineuronal antibodiesand the severity of autism in children,” European Journal ofPediatric Neurology. In press.

[4] P. Ashwood and A. J. Wakefield, “Immune activation ofperipheral blood and mucosal CD3+ lymphocyte cytokineprofiles in children with autism and gastrointestinal symp-toms,” Journal of Neuroimmunology, vol. 173, no. 1-2, pp. 126–134, 2006.

[5] P. Ashwood, S. Wills, and J. Van De Water, “The immuneresponse in autism: a new frontier for autism research,” Journalof Leukocyte Biology, vol. 80, no. 1, pp. 1–15, 2006.

[6] R. L. Blaylock, “A possible central mechanism in autismspectrum disorders, part 2: immunoexcitotoxicity,” AlternativeTherapies in Health and Medicine, vol. 15, no. 1, pp. 60–67,2009.

[7] R. L. Blaylock and A. Strunecka, “Immune-glutamatergicdysfunction as a central mechanism of the autism spectrumdisorders,” Current Medicinal Chemistry, vol. 16, no. 2, pp.157–170, 2009.

[8] N. C. Derecki, E. Privman, and J. Kipnis, “Rett syndromeand other autism spectrum disordersbrain diseases of immunemalfunction,” Molecular Psychiatry, vol. 15, no. 4, pp. 355–363,2010.

[9] A. M. Enstrom, C. E. Onore, J. A. Van de Water, and P.Ashwood, “Differential monocyte responses to TLR ligandsin children with autism spectrum disorders,” Brain, Behavior,and Immunity, vol. 24, no. 1, pp. 64–71, 2010.


[10] I. Maezawa and L. W. Jin, “Rett syndrome microglia damagedendrites and synapses by the elevated release of glutamate,”Journal of Neuroscience, vol. 30, no. 15, pp. 5346–5356, 2010.

[11] C. A. Pardo, D. L. Vargas, and A. W. Zimmerman, “Immunity,neuroglia and neuroinflammation in autism,” InternationalReview of Psychiatry, vol. 17, no. 6, pp. 485–495, 2005.

[12] W. J. Streit, “Microglia as neuroprotective, immunocompetentcells of the CNS,” GLIA, vol. 40, no. 2, pp. 133–139, 2002.

[13] D. L. Vargas, C. Nascimbene, C. Krishnan, A. W. Zim-merman, and C. A. Pardo, “Neuroglial activation andneuroinflammation in the brain of patients with autism,”Annals of Neurology, vol. 57, no. 1, pp. 67–81, 2005.

[14] W. J. Streit, “Microglia and neuroprotection: implications forAlzheimer’s disease,” Brain Research Reviews, vol. 48, no. 2, pp.234–239, 2005.

[15] A. J. Husband, “The immune system and integrated home-ostasis,” Immunology and Cell Biology, vol. 73, no. 4, pp. 377–382, 1995.

[16] L. Vitkovic, J. Bockaert, and C. Jacque, “’Inflammatory’cytokines’ neuromodulators in normal brain?” Journal ofNeurochemistry, vol. 74, no. 2, pp. 457–471, 2000.

[17] R. Dantzer and K. W. Kelley, “Twenty years of researchon cytokine-induced sickness behavior,” Brain, Behavior, andImmunity, vol. 21, no. 2, pp. 153–160, 2007.

[18] S. F. Maier and L. R. Watkins, “Cytokines for psychologists:implications of bidirectional immune-to-brain communi-cation for understanding behavior, mood, and cognition,”Psychological Review, vol. 105, no. 1, pp. 83–107, 1998.

[19] F. Mignini, V. Streccioni, and F. Amenta, “Autonomic inner-vation of immune organs and neuroimmune modulation,”Autonomic and Autacoid Pharmacology, vol. 23, no. 1, pp. 1–25, 2003.

[20] D. Wrona, “Neural-immune interactions: an integrative viewof the bidirectional relationship between the brain andimmune systems,” Journal of Neuroimmunology, vol. 172, no.1-2, pp. 38–58, 2006.

[21] F. Ginhoux, M. Greter, M. Leboeuf et al., “Fate mappinganalysis reveals that adult microglia derive from primitivemacrophages,” Science, vol. 330, no. 6005, pp. 841–845, 2010.

[22] P. Rezaie and D. Male, “Mesoglia and microglia—a historicalreview of the concept of mononuclear phagocytes withinthe central nervous system,” Journal of the History of theNeurosciences, vol. 11, no. 4, pp. 325–374, 2002.

[23] D. Giulian, D. G. Young, J. Woodward, D. C. Brown, and L. B.Lachman, “Interleukin-1 is an astroglial growth factor in thedeveloping brain,” Journal of Neuroscience, vol. 8, no. 2, pp.709–714, 1988.

[24] S. Rakic and N. Zecevic, “Programmed cell death in thedeveloping human telencephalon,” European Journal of Neu-roscience, vol. 12, no. 8, pp. 2721–2734, 2000.

[25] W. J. Streit, “Microglia and macrophages in the developingCNS,” NeuroToxicology, vol. 22, no. 5, pp. 619–624, 2001.

[26] G. A. Garden and T. Moller, “Microglia biology in health anddisease,” Journal of Neuroimmune Pharmacology, vol. 1, no. 2,pp. 127–137, 2006.

[27] M. J. Bell and J. M. Hallenbeck, “Effects of intrauterineinflammation on developing rat brain,” Journal of NeuroscienceResearch, vol. 70, no. 4, pp. 570–579, 2002.

[28] E. M. Ullian, S. K. Sapperstein, K. S. Christopherson, and B. A.Barres, “Control of synapse number by glia,” Science, vol. 291,no. 5504, pp. 657–661, 2001.

[29] K. S. Christopherson, E. M. Ullian, C. C. A. Stokeset al., “Thrombospondins are astrocyte-secreted proteins that

promote CNS synaptogenesis,” Cell, vol. 120, no. 3, pp. 421–433, 2005.

[30] C. Eroglu, N. J. Allen, M. W. Susman et al., “Gabapentinreceptor alpha2delta-1 is a neuronal thrombospondin receptorresponsible for excitatory CNS synaptogenesis,” Cell, vol. 139,no. 2, pp. 380–392, 2009.

[31] O. Garcia, M. Torres, P. Helguera, P. Coskun, and J. Busciglio,“A role for thrombospondin-1 deficits in astrocyte-mediatedspine and synaptic pathology in down’s syndrome,” PLoSONE, vol. 5, no. 12, Article ID e14200, 2010.

[32] C. M. Muller and J. Best, “Ocular dominance plasticity in adultcat visual cortex after transplantation of cultured astrocytes,”Nature, vol. 342, no. 6248, pp. 427–430, 1989.

[33] L. Bennet and A. Gunn, “The fetal origins of adult mentalillness,” in Early Life Origins of Health and Disease (Advancesin Experimental Medicine and Biology), M. Wintour-Coghlanand J. Owens, Eds., pp. 204–211, Springer, New York, NY,USA, 2006.

[34] D. Rice and S. Barone Jr, “Critical periods of vulnerability forthe developing nervous system: evidence from humans andanimal models,” Environmental Health Perspectives, vol. 108,supplement 3, pp. 511–533, 2000.

[35] M. Hornig, H. Weissenbock, N. Horscroft, and W. I. Lipkin,“An infection-based model of neurodevelopmental damage,”Proceedings of the National Academy of Sciences of the UnitedStates of America, vol. 96, no. 21, pp. 12102–12107, 1999.

[36] K. B. Nelson and R. E. Willoughby, “Infection, inflammationand the risk of cerebral palsy,” Current Opinion in Neurology,vol. 13, no. 2, pp. 133–139, 2000.

[37] P. Rantakallio, P. Jones, J. Moring, and L. Von Wendt,“Association between central nervous system infections duringchildhood and adult onset schizophrenia and other psychoses:a 28-year follow-up,” International Journal of Epidemiology,vol. 26, no. 4, pp. 837–843, 1997.

[38] L. Shi, S. H. Fatemi, R. W. Sidwell, and P. H. Patterson,“Maternal influenza infection causes marked behavioral andpharmacological changes in the offspring,” Journal of Neuro-science, vol. 23, no. 1, pp. 297–302, 2003.

[39] N. V. Malkova, C. Z. Yu, E. Y. Hsiao, M. J. Moore, and P.H. Patterson, “Maternal immune activation yields offspringdisplaying mouse versions of the three core symptoms ofautism,” Brain, Behavior, and Immunity, vol. 26, no. 4, pp.607–616, 2012.

[40] S. D. Bilbo and J. M. Schwarz, “Early-life programming oflater-life brain and behavior: a critical role for the immunesystem,” Frontiers in Behavioral Neuroscience, vol. 3, p. 14,2009.

[41] S. D. Bilbo, S. H. Smith, and J. M. Schwarz, “A lifespanapproach to neuroinflammatory and cognitive disorders: acritical role for glia,” Journal of Neuroimmune Pharmacology,pp. 1–18, 2011.

[42] R. B. Elliott, H. E. Wasmuth, N. J. Bibby, A. K. H. Macgibbon,and J. P. Hill, “The effect of beta-casein polymorphism onthe stimulation of diabetes in the non-obese diabetic mouseand effect of beta-casomorphin-7 on immune cell activity,”International Dairy Journal, vol. 8, p. 580, 1998.

[43] G. A. W. Rook, “Review series on helminths, immunemodulation and the hygiene hypothesis: the broader impli-cations of the hygiene hypothesis,” Immunology, vol. 126, no.1, pp.3–11, 2009.

[44] G. Borkow, Q. Leng, Z. Weisman et al., “Chronic immune acti-vation associated with intestinal helminth infections results inimpaired signal transduction and anergy,” Journal of ClinicalInvestigation, vol. 106, no. 8, pp. 1053–1060, 2000.


[45] M. P. Muehlenbein, “Parasitological analyses of the male chim-panzees (Pan troglodytes schweinfurthii) at Ngogo, KibaleNational Park, Uganda,” American Journal of Primatology, vol.65, no. 2, pp. 167–179, 2005.

[46] M. Scaglia, M. Tinelli, R. Revoltella et al., “Relationshipbetween serum IgE levels and intestinal parasite load inAfrican populations,” International Archives of Allergy andApplied Immunology, vol. 59, no. 4, pp. 465–468, 1979.

[47] J. P. Hewitson, J. R. Grainger, and R. M. Maizels, “Helminthimmunoregulation: the role of parasite secreted proteinsin modulating host immunity,” Molecular and BiochemicalParasitology, vol. 167, no. 1, pp. 1–11, 2009.

[48] R. W. Summers, D. E. Elliott, K. Qadir, J. F. Urban Jr, R.Thompson, and J. V. Weinstock, “Trichuris suis seems to besafe and possibly effective in the treatment of inflammatorybowel disease,” American Journal of Gastroenterology, vol. 98,no. 9, pp. 2034–2041, 2003.

[49] J. Correale and M. Farez, “Association between parasiteinfection and immune responses in multiple sclerosis,” Annalsof Neurology, vol. 61, no. 2, pp. 97–108, 2007.

[50] A. P. Devalapalli, A. Lesher, K. Shieh et al., “Increased levelsof IgE and autoreactive, polyreactive IgG in wild rodents:implications for the hygiene hypothesis,” Scandinavian Journalof Immunology, vol. 64, no. 2, pp. 125–136, 2006.

[51] A. Lesher, B. Li, P. Whitt et al., “Increased IL-4 productionand attenuated proliferative and pro-inflammatory responsesof splenocytes from wild-caught rats (Rattus norvegicus),”Immunology and Cell Biology, vol. 84, no. 4, pp. 374–382, 2006.

[52] N. Q. Liu, A. T. Kaplan, V. Lagishetty et al., “Vitamin D and theregulation of placental inflammation,” Journal of Immunology,vol. 186, no. 10, pp. 5968–5974, 2011.

[53] R. M. Lucas, A. L. Ponsonby, K. Dear et al., “Sun exposure andvitamin D are independent risk factors for CNS demyelina-tion,” Neurology, vol. 76, no. 6, pp. 540–548, 2011.

[54] S. Sharief, S. Jariwala, J. Kumar, P. Muntner, and M. L.Melamed, “Vitamin D levels and food and environmentalallergies in the United States: results from the National Healthand Nutrition Examination Survey 2005-2006,” Journal ofAllergy and Clinical Immunology, vol. 127, no. 5, pp. 1195–1202, 2011.

[55] M. F. Holick, “Vitamin D deficiency,” The New England Journalof Medicine, vol. 357, pp. 266–281, 2007.

[56] M. J. Dealberto, “Prevalence of autism according to maternalimmigrant status and ethnic origin,” Acta Psychiatrica Scandi-navica, vol. 123, no. 5, pp. 339–348, 2011.

[57] J. J. Cannell, “Autism and vitamin D,” Medical Hypotheses, vol.70, no. 4, pp. 750–759, 2008.

[58] M. Waldman, S. Nicholson, N. Adilov, and J. Williams,“Autism prevalence and precipitation rates in California,Oregon, and Washington counties,” Archives of Pediatrics andAdolescent Medicine, vol. 162, no. 11, pp. 1026–1034, 2008.

[59] V. Gilsanz, A. Kremer, A. O. Mo, T. A. L. Wren, and R. Kremer,“Vitamin D status and its relation to muscle mass and musclefat in young women,” Journal of Clinical Endocrinology andMetabolism, vol. 95, no. 4, pp. 1595–1601, 2010.

[60] A. L. Kau, P. P. Ahern, N. W. Griffin, A. L. Goodman,and J. I. Gordon, “Human nutrition, the gut microbiomeand the immune system,” Nature, vol. 474, no. 7351, pp.327–336, 2011.

[61] M. H. de Borst, R. A. de Boer, R. P. Stolk, J. P. J. Slaets, B. H. R.Wolffenbuttel, and G. Navis, “Vitamin D deficiency: universalrisk factor for multifactorial diseases?” Current Drug Targets,vol. 12, no. 1, pp. 97–106, 2011.

[62] S. V. Ramagopalan, A. Heger, A. J. Berlanga et al., “A ChIP-seq defined genome-wide map of vitamin D receptor binding:associations with disease and evolution,” Genome Research,vol. 20, no. 10, pp. 1352–1360, 2010.

[63] J. P. Gouin, L. V. Hantsoo, and J. K. Kiecolt-Glaser, “Stress,negative emotions, and inflammation,” in Handbook of SocialNeurosciences, J. T. C. J. Decety, Ed., pp. 814–829, John Wileyand Sons, New York, NY, USA, 2011.

[64] J. K. Kiecolt-Glaser, K. L. Heffner, R. Glaser et al., “Howstress and anxiety can alter immediate and late phase skin testresponses in allergic rhinitis,” Psychoneuroendocrinology, vol.34, no. 5, pp. 670–680, 2009.

[65] R. Glaser and J. K. Kiecolt-Glaser, “Stress damages immunessytem and health,” Discovery Medicine, vol. 5, pp. 165–169,2005.

[66] P. Møller, H. Wallin, and L. E. Knudsen, “Oxidative stressassociated with exercise, psychological stress and life-stylefactors,” Chemico-Biological Interactions, vol. 102, no. 1, pp.17–36, 1996.

[67] E. Puterman, J. Lin, E. Blackburn, A. O’Donovan, N. Adler,and E. Epel, “The power of exercise: buffering the effect ofchronic stress on telomere length,” PLoS ONE, vol. 5, no. 5,Article ID e10837, 2010.

[68] P. Salmon, “Effects of physical exercise on anxiety, depression,and sensitivity to stress: a unifying theory,” Clinical PsychologyReview, vol. 21, no. 1, pp. 33–61, 2001.

[69] U. Rimmele, R. Seiler, B. Marti, P. H. Wirtz, U. Ehlert,and M. Heinrichs, “The level of physical activity affectsadrenal and cardiovascular reactivity to psychosocial stress,”Psychoneuroendocrinology, vol. 34, no. 2, pp. 190–198, 2009.

[70] T. G. Merrett, M. L. Burr, B. K. Butland, J. Merrett, F.G. Miskelly, and E. Vaughan-Williams, “Infant feeding andallergy: 12-month prospective study of 500 babies born intoallergic families,” Annals of Allergy, vol. 61, no. 6, pp. 13–20,1988.

[71] R. K. Chandra, S. Puri, and A. Hamed, “Influence of maternaldiet during lactation and use of formula feeds on developmentof atopic eczema in high risk infants,” British Medical Journal,vol. 298, no. 6693, pp. 228–230, 1989.

[72] S. Koletzko, P. Sherman, M. Corey, A. Griffiths, and C. Smith,“Role of infant feeding practices in development of Crohn’sdisease in childhood,” British Medical Journal, vol. 298, no.6688, pp. 1617–1618, 1989.

[73] G. Dick, “President’s address. The etiology of multiple sclero-sis,” Proceedings of the Royal Society of Medicine, vol. 69, no. 8,pp. 611–615, 1976.

[74] D. Fava, R. D. G. Leslie, and P. Pozzilli, “Relationship betweendairy product consumption and incidence of IDDM inchildhood in Italy,” Diabetes Care, vol. 17, no. 12, pp. 1488–1490, 1994.

[75] L. B. Ray, “Linking inflammation and metabolism,” Science’sSTKE, vol. 386, Article ID tw165, 2007.

[76] M. Assem, E. G. Schuetz, M. Leggas et al., “Interactionsbetween hepatic Mrp4 and Sult2a as revealed by the constitu-tive androstane receptor and Mrp4 knockout mice,” Journal ofBiological Chemistry, vol. 279, no. 21, pp. 22250–22257, 2004.

[77] O. Burk, K. A. Arnold, A. K. Nussler et al., “Antimalar-ial artemisinin drugs induce cytochrome P450 and MDR1expression by activation of xenosensors pregnane X receptorand constitutive androstane receptor,” Molecular Pharmacol-ogy, vol. 67, no. 6, pp. 1954–1965, 2005.

[78] A. Geick, M. Eichelbaum, and O. Burk, “Nuclear receptorresponse elements mediate induction of Intestinal MDR1 by


Rifampin,” Journal of Biological Chemistry, vol. 276, no. 18, pp.14581–14587, 2001.

[79] H. R. Kast, B. Goodwin, P. T. Tarr et al., “Regulation ofmultidrug resistance-associated protein 2 (ABCC2) by thenuclear receptors pregnane X receptor, farnesoid X-activatedreceptor, and constitutive androstane receptor,” Journal ofBiological Chemistry, vol. 277, no. 4, pp. 2908–2915, 2002.

[80] E. G. Schuetz, A. H. Schinkel, M. V. Relling, and J. D. Schuetz,“P-glycoprotein: a major determinant of rifampicin-inducibleexpression of cytochrome P4503A in mice and humans,”Proceedings of the National Academy of Sciences of the UnitedStates of America, vol. 93, no. 9, pp. 4001–4005, 1996.

[81] K. W. Renton, “Alteration of drug biotransformation andelimination during infection and inflammation,” Pharmacol-ogy and Therapeutics, vol. 92, no. 2-3, pp. 147–163, 2001.

[82] A. P. Beigneux, A. H. Moser, J. K. Shigenaga, C. Grunfeld,and K. R. Feingold, “Reduction in cytochrome P-450 enzymeexpression is associated with repression of CAR (constitutiveandrostane receptor) and PXR (pregnane X receptor) inmouse liver during the acute phase response,” Biochemical andBiophysical Research Communications, vol. 293, no. 1, pp. 145–149, 2002.

[83] G. Hartmann, A. K. Y. Cheung, and M. Piquette-Miller,“Inflammatory cytokines, but not bile acids, regulate expres-sion of murine hepatic anion transporters in endotoxemia,”Journal of Pharmacology and Experimental Therapeutics, vol.303, no. 1, pp. 273–281, 2002.

[84] E. Siewert, C. G. Dietrich, F. Lammert et al., “Interleukin-6regulates hepatic transporters during acute-phase response,”Biochemical and Biophysical Research Communications, vol.322, no. 1, pp. 232–238, 2004.

[85] C. S. Song, I. Echchgadda, Y. K. Seo et al., “An essentialrole of the CAAT/enhancer binding protein-α in the vita-min D-induced expression of the human steroid/bile acid-sulfotransferase (SULT2A1),” Molecular Endocrinology, vol. 20,no. 4, pp. 795–808, 2006.

[86] L. A. Schieve, S. Boulet, S. J. Blumberg et al., “Associationbetween parental nativity and autism spectrum disorderamong US-born non-Hispanic white and Hispanic children,2007 National Survey of Children’s Health,” Disability andHealth Journal, vol. 5, no. 1, pp. 18–25, 2012.

[87] J. Correale, M. Farez, and G. Razzitte, “Helminth infectionsassociated with multiple sclerosis induce regulatory B cells,”Annals of Neurology, vol. 64, no. 2, pp. 187–199, 2008.

[88] J. M. Burton, S. Kimball, R. Vieth et al., “A phase I/IIdose-escalation trial of vitamin D3 and calcium in multiplesclerosis,” Neurology, vol. 74, no. 23, pp. 1852–1859, 2010.

[89] Y. Shen, K. A. Dies, I. A. Holm et al., “Clinical genetic testingfor patients with autism spectrum disorders,” Pediatrics, vol.125, no. 4, pp. e727–e735, 2010.

[90] H. O. Atladottir, P. Thorsen, L. Østergaard et al., “Maternalinfection requiring hospitalization during pregnancy andautism spectrum disorders,” Journal of Autism and Develop-mental Disorders, vol. 40, no. 12, pp. 1423–1430, 2010.

[91] L. S. Mills, M. F. Soule, and D. F. Doak, “The keystone-speciesconcept in ecology and conservation,” Bioscience, vol. 43, no.4, pp. 219–224, 1993.

[92] C. D. Bethell, M. D. Kogan, B. B. Strickland, E. L. Schor, J.Robertson, and P. W. Newacheck, “A national and state profileof leading health problems and health care quality for USchildren: key insurance disparities and across-state variations,”Academic Pediatrics, vol. 11, no. 3, pp. S22–S33, 2011.


Research Article

Prenatal and Postnatal Epigenetic Programming: Implications forGI, Immune, and Neuronal Function in Autism

Mostafa I. Waly,1 Mady Hornig,2 Malav Trivedi,3 Nathaniel Hodgson,3 Radhika Kini,3

Akio Ohta,3 and Richard Deth3

1 Department of Food Science and Nutrition, Sultan Qaboos University, Alkoudh 123, Muscat, Oman2 Department of Epidemiology, Columbia University, New York, NY 10032, USA3 Department of Pharmaceutical Sciences, Northeastern University, Boston, MA 02115, USA

Correspondence should be addressed to Richard Deth, [email protected]

Received 23 January 2012; Accepted 3 May 2012


Copyright © 2012 Mostafa I. Waly et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

Although autism is first and foremost a disorder of the central nervous system, comorbid dysfunction of the gastrointestinal(GI) and immune systems is common, suggesting that all three systems may be affected by common molecular mechanisms.Substantial systemic deficits in the antioxidant glutathione and its precursor, cysteine, have been documented in autism inassociation with oxidative stress and impaired methylation. DNA and histone methylation provide epigenetic regulation of geneexpression during prenatal and postnatal development. Prenatal epigenetic programming (PrEP) can be affected by the maternalmetabolic and nutritional environment, whereas postnatal epigenetic programming (PEP) importantly depends upon nutritionalsupport provided through the GI tract. Cysteine absorption from the GI tract is a crucial determinant of antioxidant capacity,and systemic deficits of glutathione and cysteine in autism are likely to reflect impaired cysteine absorption. Excitatory aminoacid transporter 3 (EAAT3) provides cysteine uptake for GI epithelial, neuronal, and immune cells, and its activity is decreasedduring oxidative stress. Based upon these observations, we propose that neurodevelopmental, GI, and immune aspects of autismeach reflect manifestations of inadequate antioxidant capacity, secondary to impaired cysteine uptake by the GI tract. Genetic andenvironmental factors that adversely affect antioxidant capacity can disrupt PrEP and/or PEP, increasing vulnerability to autism.

1. Introduction

Neurological and behavioral symptoms implicate abnormalbrain development as a core pathophysiological feature ofautism, but increasing evidence also indicates immune [1–8]and gastrointestinal (GI) abnormalities in a significant subset[1, 8–12]. This triad of dysfunctional systems provides notonly a more complete description of autism and autism spec-trum disorders (ASDs), but also an important opportunityto consider the mechanisms which could result in the sharedinvolvement of these three particular systems, especiallyin the context of development. Recent elucidation of thecentral role of epigenetic regulation of gene expression indevelopment presents a molecular framework within whichprenatal and postnatal maturation of neuronal, immune, andGI systems can be viewed.

As all cells possess the same DNA, their differentiationinto various cell types reflects stable suppression of somegenes and activation of others, accomplished in large partby epigenetic regulation. Such regulation involves reversiblemodifications of both DNA nucleotides and histone proteins[13]. These modifications or epigenetic marks favor ordisfavor formation of nucleosomes, DNA/histone complexesthat maintain genes in a compacted, inactive state (hete-rochromatin). Methylation of DNA is the most fundamentalepigenetic mark, enabling binding of proteins containingmethylDNA binding domains, such as methyl CpG bindingprotein 2 (MeCP2). Proteins with methylDNA bindingdomains recruit other proteins, including those capable ofmodifying histones at their tail regions [14]. Multiple sitesand forms of histone modification (e.g., methylation, acety-


lation, and phosphorylation) make this a highly complexmode of regulation, affected by diverse signaling pathwaysthat adjust gene expression to local cellular metabolicconditions [15]. Whereas certain epigenetic marks are tosome degree reversible, others are not readily reversed and,once in place, these may be sustained for an entire lifespanand/or be transmitted across generations through germlinemodifications [16]. Accordingly, epigenetic marking or pro-gramming early in development has the potential to exertlong-lasting effects [17].

There is considerable evidence that epigenetic pro-gramming is highly sensitive to changes in the cellularenvironment, as broadly defined [18]. Indeed, it appears thatepigenetic regulation is a widely utilized adaptive mechanismto allow cells to maintain a favorable metabolic statusunder different conditions, including differential exposure tophysiological substances (e.g., hormones, neurotransmitters,or growth-regulation factors), xenobiotics (e.g., pollutants,toxic chemicals), or even infectious agents (e.g., bacteria orviruses, fungi or parasites) [19]. Epigenetic regulation mayfacilitate adaptation to changes in the cellular environmentthrough stable alterations of cellular phenotype, potentiallyresulting in progressive differentiation and maturation dur-ing fetal and possibly postnatal development [20].

During prenatal development, nutritional support tothe fetus is provided via the transplacental circulationas a function of the available maternal nutriture, withmetabolic consequences for both the mother and developingchild. During postnatal development, nutritional support isprovided by oral food intake into the GI tract: breast-milk-or cow-milk-based formula at first, followed by introductionof other foods, usually in a controlled, gradual manner, butthe provision of adequate nutritional resources to supportfurther development of the newborn is not assured. Thus,the prenatal/postnatal transition is a critical juncture inmetabolic adaptation. This transition is particularly linkedto aerobic metabolism, as the delivery of oxygen via newbornlungs and its ultimate rate of utilization must be balanced bythe available antioxidant capacity of body tissues. Epigeneticregulation offers an opportunity for adaptation during thismetabolic transition, allowing the level of ongoing aerobicmetabolism in each organ system, tissue, and cell type tobe maintained in homeostatic equilibrium with antioxidantmetabolism.

We use the term postnatal epigenetic programming(PEP) to describe the ongoing adaptive changes in geneexpression occurring in response to the transition from fetalto postnatal metabolism, as distinct from prenatal epigeneticprogramming (PrEP), at term that recognizes the dynamicoccurrence of similar changes which occur in utero. Whileit is clear that autism can result from a variety of geneticand environmental factors, by focusing on factors affectingantioxidant and methylation status in the developing GItract, immune system, and brain, we hope to illuminate thepathological mechanisms underlying ASDs and other relateddisorders.

2. Materials and Methods

2.1. Purification of Regulatory T Cells. To purify CD4+

CD25+ regulatory T cells, spleen cells and lymph nodecells from C57BL/6 mice were combined as a cell source.The cells were labeled with FITC-conjugated anti-CD24and anti-CD8 mAbs and with anti-FITC microbeads. Afterthe removal of CD8+ T cells and CD24-expressing B cellsusing an AutoMACS separator (Miltenyi Biotec, Auburn,CA), regulatory T cells were purified by positive selection ofCD25+ cells using PE-conjugated anti-CD25 mAb and anti-PE microbeads. All antibodies except for anti-CD25 mAbwere from BD Biosciences (San Jose, CA). Other materialswere from Miltenyi Biotec. Purity of CD4+ CD25+ cells washigher than 95 %.

2.2. RNA Isolation. RNA was isolated using the RNAqueous-4PCR kit (Ambion, Naugatuck, CT), then treated withDNase to eliminate contaminating genomic DNA, andquantified using an ND-1000 NanoDrop spectrophotometer.

2.3. cDNA Synthesis. cDNA synthesis was performed usingthe First strand cDNA synthesis (Roche, Nutley, NJ). 1 μgRNA was used as template, along with 1 mM dNTP mix,60 uM random hexamer primers and dH20 to make a finalvolume of 13 μL. Samples were denatured at 65◦C for 5 min-utes and then placed on ice. Transcriptor RT (20 units/μL),Protector RNase inhibitor (40 U/μL), 5x Transcriptor ReverseTranscriptase Reaction Buffer, and dH20 in a final volumeof 7 μL were used to bring the final volume to 20 μL. Thiswas followed by incubation at 25◦C for 10 min followed by30 min incubation at 55◦C. The reverse transcriptase enzymewas inhibited by incubating at 85◦C for 5 min and thencooling on ice.

2.4. Primers. Primers were designed using the InvitrogenOligoPerfect Designer to have approximately 50–60% GCcontent, an annealing temperature of 60◦C, and a lengthof 20 bases. Glyceraldehyde phosphate dehydrogenase(GAPDH) expression was used for normalization. Primersfor EAAT3 (designated EAAC1 in mice) and GAPDH wereEAAT3-Forward: 5′-TTACAGCCACCGCTGCCAGC-3′;EAAT3-Reverse: 5′-GGCAGCCCCACAGCACTCAG-3′;GAPDH-Forward: 5′-TCTCCACACCTATGGTGCAA-3′;GAPDH-Reverse: 5′-CAAGAAACAGGGGAGCTGAG-3′.

2.5. qRT-PCR Analysis. qRT-PCR was performed on dupli-cate samples using the LightCycler 480 from Roche. 5 μL ofcDNA was used for qRTPCR, along with 10 μM sense andantisense primers, 10 μL SYBR Green I Master from Roche,and dH20 in a final volume of 20 μL. The following thermalparameters were used: incubation for 5 min at 95◦C, followedby 45 cycles of 95◦C for 10 sec, 60◦C for 20 sec and 72◦C for30 sec, and lastly a single cycle of 95◦C for 5 sec, 1 min at 65◦Cand 97◦C for the melting curve. No template controls wererun on each plate and dissociation curves were generated todetermine any nonspecific products. Data was normalized to


GAPDH, and the ΔCt second derivative method was used foranalysis of EAAT3 expression.

2.6. MS Assay. Brain cortex samples were obtained fromthimerosal-treated or untreated C57BL/6 and SJL/J mice, asper the previously published protocol [21]. MS activity inmouse cortex was determined by measuring incorporationof radiolabel from [methyl-14C]methyltetrahydrofolate intomethionine, as previously described [22]. Enzyme assayswere performed under anaerobic conditions by bubblingnitrogen gas through stoppered vials for 1 hour at 37◦C andterminated by heating at 98◦C for 2 min after which sampleswere cooled on ice. Radiolabeled methionine was separatedfrom unreacted radiolabeled methylfolate by passing thereaction mixture through an anion exchange column. Thecolumn was washed with 2 mL of water and the aqueoussamples were collected and counted. Reported values arenormalized to protein content and corrected for the countsobserved in control assays in which sample enzyme wasomitted.

2.7. GSH Measurement. A 100 μL aliquot of cortex super-natant was incubated with 2 μL of monochlorobimane(25 mmol/L) and 2 μL of glutathione-S-transferase reagentwas added, as provided by a commercial kit (Biovision,Mountain View, CA). After a 30 min incubation at 37◦C,fluorescence was read at 380 nm excitation and 460 nmemission. GSH content was determined by comparison withvalues from a standard curve using freshly prepared GSH.

2.8. Statistical Methods. Statistical analysis was carried outusing Graph Pad Prism version 5.01. Data was expressedas mean ± standard error of the mean (SEM). Student’st-test was conducted to determine the differences betweenindividual groups.

3. Results and Discussion

3.1. Antioxidant Capacity and Aerobic Metabolism. Aerobicmetabolism is unavoidably accompanied by the risk ofoxidative damage, most notably caused by reactive oxygenspecies (ROS) generated as by-products of mitochondrialoxidative phosphorylation and ATP production. Increasedmetabolic activity requires higher rates of ATP productionand increased ROS formation. Maintenance of redox equilib-rium requires inactivation of ROS before these by-productscan oxidize cellular components, as well as repair of oxidizedbiomolecules (proteins, lipids, DNA, and RNA). Catalase,superoxide dismutase and the selenoprotein glutathioneperoxidase inactivate ROS, and a wide array of redox-activeproteins and enzymes participate in the repair of oxidizedmolecules and the maintenance of thiols in their reducedstate, with glucose-derived NADPH serving as the source ofreducing electrons [23, 24].

As elegantly reviewed by Schafer and Buettner [25], theequilibrium between reduced and oxidized forms of glu-tathione (GSH and GSSG, resp.) is the primary determinant

of intracellular redox status. The relatively high concentra-tion of GSH within the cell imposes a pervasive influence onessentially all aspects of metabolism. Although all cell typesand tissues rely upon the same core antioxidant mechanisms,variations on this central theme are recognized. For example,the intracellular level of GSH in liver hepatocytes is close to10 mM [25], whereas in neurons it is 0.2 mM, some 50-foldlower [26].

As the rate of ROS formation must be quantitativelybalanced by the available antioxidant capacity, a multitudeof mechanisms have evolved to restrict aerobic metabolismwhen antioxidant status is low and vice versa. One prominentexample involves glutathionylation of cysteine residues incomplex I of the mitochondrial electron transport chain,which limits the flow of electrons, thereby restricting down-stream ROS formation [27]. In glutathionylation, GSSG,the oxidized form of glutathione, reacts with a reducedcysteine, resulting in GSH release and modification of theprotein in proportion to the prevailing redox state [28]. Thisregulatory mechanism not only controls ROS formation butalso provides an auxiliary pathway for glutathione reductaseactivity and production of GSH. Additional intracellularcontrol over the balance between ROS and antioxidantreservoirs is achieved by the capacity for reversal of the glu-tathionylation of complex I by the action of glutaredoxin 2,in conjunction with the selenoprotein thioredoxin reductase,using NADPH-derived electrons [27].

More than ten studies have reported significantly lowerplasma levels of GSH in autism, accompanied by lowerlevels of cysteine [29–39]. Among these studies, the averagedecrease in GSH was 37%, representing a substantial reduc-tion in antioxidant capacity. Several studies also reported asignificant decrease in the GSH to GSSG ratio [8–20, 22–36],indicative of oxidative stress, consistent with other reportsof increased biomarkers of oxidative stress [29, 32, 40].One consequence of persistent oxidative stress, an increasein glutathionylation of complex I of the mitochondrialelectron transport chain, provides a potential explanationfor the mitochondrial dysfunction frequently reported inautism [7]. Genetic factors affecting mitochondrial functionmay interact with oxidative stress. For example, SLC25A12,the gene encoding the Ca2+-dependent aspartate-glutamatecarrier isoform 1 (AGC1), has been identified as a putativeautism susceptibility gene [41]. AGC1 facilitates the supplyof NADH for oxidative phosphorylation, and elevated Ca2+

levels will, therefore, increase ROS formation via increasedAGC1 activity. Higher Ca2+-dependent AGC1 transport rateswere demonstrated in autistic subjects [41], and it has beenproposed that dyregulated Ca2+ levels may be a broadlyinfluential factor in causing autism [42].

3.2. Intestinal Absorption of Cysteine. Starting from birth,the systemic availability of cysteine to support GSH synthesisis dependent upon absorption of the food-derived sulfur-containing amino acids cysteine and methionine from theGI tract (Figure 1). Adequate GI uptake of selenium isalso critical for maintaining GSH in its reduced state.Intestinal epithelial cells express different transporters on


GI tract Blood Brain

Bloodbrain barrier

GIepithelial cells

LiverCysteine Cystine

Cysteine CysteineCysteine

Cystine Cystine

Astrocytes

Cystine GSH

Neurons

Cysteine

CysteineGSHDietaryprotein

Methionine

GSH GSHEAAT3

EAAT3

Figure 1: Absorption and systemic distribution of dietary cysteine and methionine. GI epithelial cells take up cysteine and methionine,with EAAT3 being particularly important for cysteine uptake in the distal ileum. A portion of methionine is converted to cysteine via themethionine cycle and transsulfuration of homocysteine. Absorbed cysteine is taken up by the liver, which releases oxidized cystine. Cystine,but not cysteine, is able to traverse the blood brain barrier and is taken up by astroctyes, which convert it to GSH. Cysteine from astrocyte-derived GSH is available for EAAT3-mediated neuronal uptake, supporting neuronal GSH synthesis.

their luminal brush border surface that are critical foruptake of sulfur and selenium-containing amino acids. Theseamino acids are then subject to intracellular metabolismbefore being ultimately released on the contralateral surface,either in their free form or incorporated into proteins orpeptides, including GSH. Availability of transported cysteineor selenium can, therefore, affect the GSH levels and redoxstatus of the epithelial cells themselves, as well as that ofthe adjacent cells located within the intestinal wall, includingimmune cells.

Among amino acid transporters in GI epithelial cells,EAAT3 (excitatory amino acid transporter 3, EAAC1) isselective for cysteine and was initially cloned from GIepithelial cells [43]. Subsequent studies revealed that EAAT3is most prominently expressed in the small intestine, espe-cially in the terminal ileum [43, 44], a prominent site ofinflammation in subjects with autism [45]. The highest levelswere found in crypt cells and lower villus regions, the locusof multipotent stem cells that sustain the epithelial lining ofthe gut [44]. EAAT3 is a member of a family of Na+- and H+-dependent amino acid transporters named for their ability totransport glutamate and aspartate, but EAAT3 is unique in itspreference for cysteine. EAAT3 exists in equilibrium betweenthe endoplasmic reticulum and the plasma membrane,and its transport activity is regulated by the PI3 kinasesignaling pathway, similar to insulin regulation of glucosetransporters [46]. In resting cells, about 75% of EAAT3resides inside the cell, where it is inactive, but activationof either PI3 kinase or protein kinase C increases the

proportion of active transporters in the plasma membrane.Oxidative stress inhibits EAAT3 transport activity [47], andseveral cysteine residues have been identified as being criticalfor this mode of regulation, raising the possibility thatthey might be targets for glutathionylation. Oxidative stressaffecting gut epithelium would, therefore, decrease EAAT3-dependent absorption of cysteine, with the local and systemicconsequences of lower GSH levels.

Cells express a multitude of adaptive mechanisms thatallow them to maintain homeostatic redox equilibrium,of which the Keap1/Nrf2 system is a prominent example.Nrf2 (nuclear factor erythroid 2-like 2) is a transcriptionfactor protein with a high turnover rate, that is, tightlybound to Keap-1 (Kelch-like ECH-associated protein 1);however, in response to a more oxidative redox state, Nrf2is released from Keap-1 and moves to the nucleus where itpromotes expression of a large number of genes via bindingto antioxidant response element (ARE) sites [48]. It hasbeen recently demonstrated that Keap1/Nrf2 mediates up-regulation of EAAT3 in response to oxidative stress [49].

Similar to autism, GI disorders such as celiac diseaseand inflammatory bowel disease have an autoimmunecomponent [50], and diets excluding gluten and/or casein(i.e., GF/CF diet) are frequently beneficial in both conditions.Prompted by the recognition that digestion of both glutenand casein yields peptides with opiate activity [51, 52], werecently examined the activity of these peptides on EAAT3-mediated cysteine uptake and redox status in human GIepithelial and neuronal cells and found that they significantly


Methioninesynthase

HCY

MET

SAH

SAM

Methylationreactions

e.g.DNA

methylationATP

Adenosine

MethylTHF

THF

Cystathionine

Cysteine

GSH

D4HCY

D4SAM

D4SAH

D4MET

ATP

MethylTHF

THF

Phospholipidmethylation

Adenosine

Dopamine

Cysteine

PI3-kinase(+)

Partially blocked in neuronal cells

EAAT3

Astrocyte

Cysteinylglycine GSH

GSSG

Growthfactors

Oxidative stressor

gluten/caseinCystine

GSH

Cysteine

Neuron

(−)

(−)

γ-glutamylcysteine

>200

PP + PiPP + Pi

Figure 2: Redox and methylation pathways in neurons. The amino acid cysteine is rate-limiting for glutathione (GSH) synthesis andit is provided either by uptake from astrocyte-derived cysteine or by transsulfuration of homocysteine (HCY). The methionine cycleof methylation (lower right) depends upon both dietary methionine (MET) and remethylation of HCY by methionine synthase. Sinceformation of HCY from S-adenosylhomocysteine (SAH) is reversible and SAH inhibits methylation, decreased methionine synthase activity(e.g., caused by oxidative stress) both augments GSH synthesis and inhibits methylation reactions. Thus, redox status and methylationactivity are closely linked. The D4 dopamine receptor carries out a cycle of dopamine-stimulated phospholipid methylation which iscompletely dependent upon methionine synthase and is therefore also sensitive to oxidative stress. Growth factors promote EAAT3 activityby increasing its location on the cell surface. EAAT3 activity is inhibited by oxidative stress and by food-derived opiate peptides.

inhibited uptake [53]. Their action was associated witha decrease in GSH levels and the SAM/SAH ratio, alongwith alteration in the expression of redox-related genes.Important differences, however, were observed between pep-tides derived from bovine versus human casein, which maycontribute to the well-recognized benefits of breastfeeding.

3.3. Postnatal Epigenetic Regulation (PEP). Methylation ofDNA and histones is fundamental to epigenetic regulationof gene expression, providing factors affecting methylationcapacity with an opportunity to affect development. S-adenosylmethionine (SAM) is the donor of methyl groupsto more than 200 methyltransferase reactions, includingDNA and histone methyltransferases [54], whereas its de-methylated form, S-adenosylhomocysteine (SAH), inhibitsmethylation by virtue of its competition with SAM. TheSAM to SAH ratio determines methylation capacity forall methylation reactions, providing an exceptionally broadinfluence over cellular metabolism.

The SAM/SAH ratio is responsive to the activity ofmethionine synthase (MS), the folate, and vitamin B12-dependent enzyme that converts homocysteine (HCY) tomethionine (MET) as part of the methionine cycle ofmethylation (Figure 2). Its vitamin B12 (cobalamin) cofactorrenders MS activity highly sensitive to oxidative stress;changes in cellular redox status (shifts toward reducing or

oxidizing conditions) are able to exert epigenetic effectsvia their influence on MS and the SAM/SAH ratio. Recog-nizing this metabolic relationship, we previously proposeda “redox/methylation hypothesis” of autism, linking thisneurodevelopmental disorder to oxidative stress induced byxenobiotic exposures [55], and now we apply this perspectiveto epigenetic programming of GI and immune function.

During oxidative stress associated with GI inflammation,impaired EAAT3 activity will decrease availability of cys-teine for GSH synthesis, leading to increased ROS levels,diminished MS activity, a decrease in the SAM/SAH ratioand a global decrease in methylation capacity, altering thepattern of DNA and histone methylation in epithelial cells.The distribution of EAAT3 suggests that this inflammation-induced epigenetic response will be most prominent in cryptstem cells of the terminal ileum, where it can influence boththe proliferation and the functionality of the GI epithelium.Notably, the terminal ileum is also a critical location for folateand vitamin B12 absorption [56, 57].

Postnatal infant nutrition is most commonly provided byeither human breast milk or bovine milk. Human breast milkis comparatively rich in sulfur-containing amino acids, andcolostrum contains 2.5-fold higher levels versus mature milk[58]. A comparison of plasma cysteine levels in 12-week-old infants found significantly higher levels for breast-fedversus formula-fed infants [59]. Thus, breast feeding appears


to provide a superior supply of the essential raw materials forsynthesis of the antioxidant GSH. As this source of cysteineis made available, EAAT3 and the factors that modulate itsactivity serve as critical regulators of the extent to whichcysteine may be absorbed.

Breast milk contains a number of different growth fac-tors, including insulin-like growth factor-1 (IGF-1), epider-mal growth factor (EGF), vascular endothelial growth factor(VEGF), hepatocyte growth factor (HGF), and transform-ing growth factor-β1 and β2 (TGF-β1/β2), among others[60, 61]. These factors activate PI3 kinase and mitogen-activated protein (MAP) kinase signaling pathways to achievetheir growth-promoting effects, which are associated withincreased metabolic activity and its attendant increase inROS production. We recently demonstrated the abilityof IGF-1 and other PI3 kinase-activating growth factorsto stimulate EAAT3-mediated cysteine uptake in culturedhuman neuronal cells, accompanied by an increase in GSHlevels, an increased SAM/SAH ratio, and increased DNAmethylation [62]. HGF, acting via the MAP kinase pathway,increases expression of the two enzymes which convertcysteine to GSH [63]. Thus, growth factors promote GSHsynthesis via both increased cysteine uptake and transcrip-tional effects, corresponding to short-term and longer-termeffects, respectively. These observations imply that breast-milk-derived growth factors can stimulate cysteine uptakeand increase GSH in the GI epithelium, coincident withthe initiation of postnatal digestion. The extent to whichantioxidant resources (i.e., cysteine and GSH) are madeavailable to the rest of the body depends upon the efficiencyof this process.

Human breast milk is also rich in selenium, and breast-fed infants have higher selenium levels than formula-fedinfants [64]. A significant proportion (up to 30%) of breastmilk selenium is in the form of glutathione peroxidase (GPx),which is hydrolyzed to selenocysteine during digestion.The importance of GPx activity in breast milk may be tocombat oxidation, especially of vulnerable omega-3 fattyacids. However, GPx appears to also represent an importantreservoir for delivery of selenium to the developing infant. Asa structural analog of cysteine, breast-milk-derived seleno-cysteine can also be transported by EAAT3 and contributesto the antioxidant capacity and redox status [65].

As MS activity and the SAM/SAH ratio are responsiveto redox status, it is reasonable to propose that the post-natal transition to independent nutrition is associated withepigenetic responses and that growth factor stimulation ofcysteine uptake is a key event in this transition, which wecall postnatal epigenetic programming or PEP. Although PEPis an ongoing body-wide process affecting the developmentof all tissues, but we specifically propose that the GItract plays a uniquely important role as the source ofnutritional support for metabolism, and that availability ofantioxidant resources is the critical determinant of the levelof metabolic activity that may be achieved without cellulardamage. Epigenetic regulation of gene expression provides amolecular mechanism by which the functional state of the

GI tract can be matched to the ongoing level of metabolicactivity. In other words, the supply of antioxidant must bematched to the demand for antioxidant, and the rate ofpostnatal growth and development must be restricted so asto not exceed the supply provided by the GI tract.

As a developmental disorder in which the availabilityof GSH is reported to be reduced by almost 40%, autismcan be viewed as a syndrome resulting from an imbalancebetween antioxidant supply and demand—essentially, a stateof oxidative stress that interrupts the normal epigeneticallybased program of development. In the absence of a large-scale prospective study, it is not possible to determinewhether the autism-associated GSH deficit is present frombirth or begin during postnatal development. The formercase would likely reflect a significant maternal deficit inantioxidant resources, and the finding that plasma levelsof GSH and the SAM/SAH ratio are significantly decreasedin mothers of autistic children [36] is consistent withsuch a possibility. Under the latter scenario, the imbalancemight develop in a progressive manner during postnataldevelopment if escalating metabolic activity and ROS pro-duction outstripped the ability of the GI tract to providesufficient antioxidant, and inhibition of EAAT3-mediatedcysteine uptake in response to developing oxidative stresswould also exacerbate this self-reinforcing cycle. Postnatalexposure to agents which interfere with antioxidant capacity,including mercury, lead, pesticides, and other xenobiotics,could shift the redox balance sufficiently to initiate develop-mental regression secondary to impaired methylation and itsepigenetic consequences.

Microbes within the GI tract compete for nutrientresources in a normally symbiotic manner, but dysbiosiscan intervene when the normal array of organisms isdisplaced or skewed. GI dysfunction, which is common inautism [8–12], may be associated with a failure of the GIepithelium to absorb antioxidant nutrients, such as cysteineor selenocysteine. Increased availability of these criticalantioxidant nutrients for intestinal microbes may encouragethe growth of previously restricted microorganisms, whilediminishing the growth of others. Once established, thedysbiotic microbial population will decrease availability ofantioxidant nutrients for absorption by further competingwith the host, perpetuating the antioxidant deficit.

The importance of PEP varies by cell and tissue types,depending upon the extent of the developmental matu-ration of specific cells and organs at birth. Delaying thedevelopment of selected neural networks (e.g., cortex andhippocampus) until after birth allows their programmingto be guided by experience, in contrast to other networks(e.g., brainstem respiratory centers) which must functionimmediately upon birth. Indeed, unique features of humanbrain allow PEP to be dynamically exploited throughout thelifespan, providing the capacity for memory [66]. In a similarmanner, maturation of the immune system is restraineduntil birth so as to optimize its response to environmentalexposures, while at the same time minimizing its poten-tial for autoimmunity. While this perspective encompasses


the triad of neuronal, immune, and GI symptom domains inautism, the prominence of each component may vary amongindividuals, reflecting, in part, genetic differences in thevulnerability to oxidative stress and impaired methylation.

3.4. Redox Signaling in Human Brain. As previously noted,cysteine availability is limited for GSH synthesis. However,the level of cysteine in cerebrospinal fluid (CSF) is onlyone tenth the level in blood [67], meaning that the rawmaterial for making antioxidants is much scarcer in thebrain. Indeed, the GSH level in neurons is the lowestreported for any cell type [25], despite the fact that thebrain utilizes oxygen at a ten times higher rate than othertissues. As illustrated in Figure 2, neurons obtain theircysteine from neighboring astrocytes, which release GSHthat is subsequently broken down to cysteine, which is thenavailable for uptake by neurons, depending upon the levelof EAAT3 activity, which is controlled by growth factors.Growth factors utilize redox status to regulate neuronalfunction via the intermediate involvement of MS, a formof “redox signaling.” By allowing more cysteine into cellsand increasing GSH synthesis, growth factors increase MSactivity and increase DNA methylation. These increases inMS activity and DNA methylation lead to epigenetic changesin gene expression as well as many other metabolic effectsthrough >200 methylation reactions, all affected by theSAM/SAH ratio. The increase in antioxidant level also allowscells to safely increase their oxidative metabolism (that is,their mitochondrial activity); the resultant increase in ATPsupports a higher level of neuronal activity. In this way, redoxsignaling not only regulates gene expression but also controlsthe level of neuronal function.

A second brain-specific feature that augments redoxsignaling is a brain-specific limitation in the level of trans-sulfuration activity, which restricts conversion of HCY tocysteine [68]. This unique feature contributes to the lowGSH level in neurons, making them more responsive togrowth factor-induced EAAT3 activation. Together with thelimited availability of extracellular cysteine, this makes epi-genetic regulation exceptionally dynamic. Along with othermethylation-dependent consequences of redox signaling,the long-term nature of epigenetic effects endows neuronswith the ability to create “metabolic memories” that arecoordinated with experience via the input of sensory systems.This view is consistent with recent studies identifying thecentral role of epigenetic changes in learning and memoryformation [66, 69]. Compromised function of these redox-dependent systems could contribute to the types of cognitivefunction deficits and neurodevelopmental delays that areprimary to autism.

The restricted availability of GSH in the brain is partlycompensated for by the presence of alternate mechanismsfor sustaining antioxidant capacity, including an increaseddependence upon selenoproteins. The brain has developedthe capacity to retain selenium under conditions of scarcity,even when other tissues become depleted. Nonetheless,

reliance on and retention of selenoproteins are an insufficientmeans of maintaining antioxidant defenses under certainconditions, such as exposure to mercury. Selenoproteinsare exquisitely sensitive to inhibition by mercury, and anymercury that penetrates the brain will interfere with redoxsignaling and disrupt normal epigenetic regulation [70, 71].Very tight binding of mercury by selenoproteins, especiallyto selenoprotein P, contributes to the long-term retentionof mercury in the brain and its accumulation across thelifespan, and it has been proposed that selenoprotein P, whichcontains ten selenocysteine residues, has evolved to serve aspecific role to protect against mercury toxicity [72].

In addition to conversion of HCY to MET, MS carriesout a second reaction, providing methyl groups to the D4dopamine receptor; the D4 receptor, when stimulated bydopamine, subsequently transfers these methyl groups tomembrane phospholipids—a unique activity carried outby the D4 dopamine receptor [73]. Genetic variants ofthe D4 receptor (such as the 7-repeat variant) have beenlinked to novelty-seeking behavior and are important riskfactors for attention-deficit/hyperactivity disorder (ADHD)[74]. Although the role of dopamine-stimulated phospho-lipid methylation remains incompletely understood, we firstproposed that it facilitates attention by modulating thefrequency of neural networks, promoting their synchro-nization [75]; others subsequently confirmed dopamine-stimulated phospholipid methylation to be critically involvedin neuronal synchronization during attention [76]. As D4receptor-mediated phospholipid methylation is absolutelydependent upon MS activity, it is thus also dependent onEAAT3, raising the possibility that ADHD might involve adecrease in MS activity secondary to oxidative stress.

3.5. Redox and the Immune Response. In the past severalyears, a previously unappreciated role for redox regulationof the immune response has been elucidated, involvingextrusion of GSH from activated antigen presenting cells(APCs, e.g., dendritic cells, macrophages, or B cells), fol-lowed by release of its cysteine, which is then taken up by Tcells [77] This relationship is analogous to the relationshipbetween astrocytes and neurons illustrated in Figure 2.Uptake of this cysteine by naıve CD4+ CD25− effector Tcells (Teff) leads to their activation and proliferation, inassociation with increased GSH synthesis and the epigeneticconsequences thereof. Activated Teff cells release cytokinesthat attract other cells and promote inflammation, as well asstimulate B cell activation and antibody production (HelperT cells). However, the extent of Teff cell activation is limitedby CD4+ CD25+ Foxp3+ regulatory T cells (Treg) that com-pete with Teff cells for available extracellular cysteine, therebysuppressing immune and autoimmune responses [77].

Recognizing this newly described mode of T cell regu-lation, we hypothesized that EAAT3 might be involved incysteine uptake. To evaluate this possibility, we used qRT-PCR to determine whether EAAT3 was expressed in murineT cells and compared the level of expression in T cell subsets,


0

0.5

1

1.5

2

EA

AT

3 ex

pres

sion

nor

mal

ized

to G

AP

DH

UNSEP CD4+ T cells CD8+ T cells B cells

∗

(a)

EA

AT

3 ex

pres

sion

nor

mal

ized

to G

AP

DH

0

2

4

Treg CD4+

∗

(b)

Figure 3: Expression of EAAT3 in murine lymphocyte subsets.Spleen and lymph node-derived lymphocytes were separated bymagnetic cell sorting, followed by qRT-PCR analysis for EAAT3(EAAC1) expression. (a) EAAT3 expression was significantly higher(P < 0.05) in CD4+ T cells as compared to unseparated lymphocytes(UNSEP), CD8+ T cells, or B cells. (b) EAAT3 expression wassignificantly higher (P < 0.05) in FoxP3+ Treg cells.

separated on the basis of surface epitopes by magnetic cellsorting. As illustrated in Figure 3(a), EAAT3 expression wasdetected in unseparated cells, as well as in CD4+ and CD8+ Tcells and B cell populations, with the highest expression levelin CD4+ cells. Further analysis showed that the expressionof EAAT3 by Treg cells is almost two-fold higher thanthat in unseparated CD4+ cells (Figure 3(b)). Thus, EAAT3

0

10

20

30

40

C57BL6/JThimerosal C57BL6/J

SJL/JThimerosal SJL/J

GSH

(μ

g/m

g pr

otei

n)

∗∗ ∗∗

Figure 4: GSH levels in frontal cortex of SJL/J and C57BL6/J micewith or without thimerosal treatment. GSH levels were significantlylower (P < 0.05) in cortex of SJL/J autoimmune-prone micecompared to C57BL6/J strain mice. GSH levels were not affectedby thimerosal treatment.

expression by immune cells provides a mechanism for themto take up cysteine in a competitive manner, and higherEAAT3 expression in Treg cells may underlie the ability ofthis cell subset to limit the immune response and restrictautoimmunity.

The SJL/J strain of mice has been widely employed as amodel for autoimmunity, but these mice also exhibit deficitsin GSH, similar to those reported in autistic children. InSJL/J mice, the deficit is due to mutations in the enzymeswhich synthesize GSH; factors underlying GSH deficitsin autism are as yet unknown. In a previous study, weshowed that the thiol-reactive compound thimerosal causedgrowth impairment, as well as neurodevelopmental andneurochemical effects in SJL/J mice, but not in C57BL6/Jmice, including upregulated brain levels of EAAT3 [21]. Tofurther characterize the role of redox and methylation statusin autoimmunity, we measured both GSH levels and MSactivity in the liver and brain of control and thimerosal-treated SJL/J mice, as compared to C57BL6/J mice. GSHlevels were significantly lower in both liver and brain of SJL/Jas compared with the levels found in C57BL6/J mice, butthey did not differ as a function of thimerosal treatmentin either strain (Figure 4). Methionine synthase activitywas also significantly lower in SJL/J versus C57BL6/J mice,but not different in thimerosal-treated animals (Figure 5).These results confirm the occurrence of low GSH in theautoimmune-prone SJL/J strain, along with the novel findingthat MS activity is reduced, although neonatal exposure tothimerosal did not alter these features. Expression of EAAT3was significantly lower in CD4+ T-cells from SJL/J miceversus C57BL6/J mice, suggesting that autoimmunity is asso-ciated with impaired capacity for cysteine uptake (Figure 6).Together these findings provide support for the roles ofoxidative stress and impaired methylation in autoimmunity.

Abnormal immune function, including, but not limitedto, increased autoimmunity, has been extensively docu-mented in autism. Of particular relevance are autoantibodies


MethylB12 HydroxoB12

0

5

10

15

20

MS

acti

vity

(pm

ol/m

in/m

g pr

otei

n)

Thimerosal C57BL6/JC57BL6/J

Thimerosal SJL/JSJL/J

∗

∗

∗∗ ∗∗

∗∗

Figure 5: Methionine synthase activity in cortex of SJl/J andC57BL6/J mice with or without thimerosal treatment. Methioninesynthase activity was measured in the presence of either methylco-balamin (MethylB12) or hydroxocobalamin (HydroxoB12). Activitywas significantly lower in SJL/J autoimmune-prone mice comparedto the C57BL6/J strain, irrespective of thimerosal treatment.Activity was higher in the presence of methylcobalamin. ∗P < 0.02compared to C57BL6/J group, ∗∗P < 0.001 compared to C57BL6/Jgroup.

targeting brain and neuronal proteins [2, 3, 5] as well asautoantibodies targeting the folate receptor [78–80]. Thelatter were initially linked to the relatively rare cerebralfolate deficiency syndrome [78] but recently were found tobe present in a high proportion (75%) of autistic subjects[80], and a casein-free diet greatly reduced the circulatinglevel of anti-folate receptor antibodies [79]. The numberof circulating Treg cells was reported to be significantlylower in autistic subjects, with 73% showing abnormally lowlevels [4]. The Treg cell decrease was more prominent inthe more severe cases, as well as among individuals withallergic manifestations such as asthma and atopic dermatitis.Although autistic subjects with low Treg cells did notmanifest clinical or laboratory-based signs of autoimmunity(arthritis or arthralgia, elevated erythrocyte sedimentationrate, or leucopenia), the frequency of autoimmunity washigher in their families. The same researchers found ele-vated markers of oxidative stress (increased plasma F2-isoprostane and/or decreased GPx activity) in 89% of autisticsubjects and evidence of antineuronal antibodies in 55%,leading them to suggest a link between oxidative stressand autoimmunity [5]. Notably, expression of Foxp3, thetranscription factor which defines Treg cells, is epigeneticallyregulated, with increased expression when DNA methylationis decreased [81], and Treg cell levels are elevated inmale, but not female, SJL mice [82]. As illustrated inFigure 7, these observations suggest that autoimmunity inautism may be related to inflammation in the distal ileum,where impaired activity of EAAT3 in Treg cells can leadto formation of autoantibodies directed against the folatereceptor.

0

0.5

1

1.5

2

C57 SJL

EA

AT

3 ex

pres

sion

C57BL6/JSJL/J

∗

Figure 6: Expression of EAAT3 in CD4+ T-cells from C57BL6/Jand SJL/J mice. Spleen and lymph node-derived lymphocytes wereseparated by FACS analysis cell sorting, followed by qRT-PCRanalysis for EAAT3 (EAAC1) expression. EAAT3 expression wassignificantly higher in CD4+ T-cells from C57BL6/J mice versusautoimmune-prone SJL/J mice (P < 0.05).

4. Conclusions

The prominence of neurological, GI, and immune symptomsin autism suggests a shared mechanism of dysregulation thatmay relate to the significant deficits in systemic reservoirsof antioxidant GSH reported in this neurodevelopmentaldisorder. Impaired GI absorption of cysteine, the essentialGSH precursor, is proposed as a crucial factor in thepathogenesis of this triad of symptoms, wherein reducedcysteine availability leads to a condition of local and sys-temic oxidative stress, and subsequent disruption of normalepigenetic regulation of gene expression. In some cases, theextent of impaired cysteine absorption may be severe enoughto result in overt GI inflammation, while in other casesthe restriction may only alter immune and/or neurologicaldevelopment and function. Awareness of the redox-basedlinkage between GI, brain, and immune systems serves toilluminate a number of diseases whose origins may reflectthe critically important roles of PrEP and PEP. Indeed,adaptive epigenetic responses to changes in redox status arelikely to play a critical role in diseases arising across thelifespan, especially those that can be traced to environmentalexposures that interfere with antioxidant homeostasis.

Acknowledgment

This report was supported in part by a research grant fromSafeMinds to R Deth.


Terminal part of ileocolic artery

Cecal branches

Ileal branches

Appendicularartery

Vermiform process

Terminalalalll par pa a tileoiil colic c arter

Cecaal brbbb an

Ileal brb

Apparte

f

Ileum

Cecum

Mucosal immune cells

GIlumenBody

CYS

Se-CYSCYS Se-CYS

GSH

Se-proteins

GSSG

Redox status

Epigenetic regulation

EAAT3

GI epithelial cell

GSHCYS

Growth factors (+)

Anti-FR-Ab

APC

T-cell

Tregcell

GSH

CYS Redoxstatus

FR

Terminalileum

Opiate peptides (−)NaıveCD4+

Figure 7: Cysteine and selenocysteine uptake in the terminal ileum regulates autoimmunity. EAAT3 transports diet-derived cysteine(CYS) and selenocysteine (Sel-CYS) into GI epithelial cells in the terminal ileum. Within the cell, cysteine is directed to GSH synthesisand selenocysteine is incorporated into selenoproteins, which maintain GSH in its reduced state. EAAT3 activity is regulated by multiplefactors, including intracellular redox status, Nrf2-dependent transcription, growth-factor-dependent translocation to the cell surface, andthe inhibitory effects of casein and gluten-derived opiate peptides. Systemic availability of cysteine and GSH depends upon these terminalileum events, as does the local mucosal environment. Immune cells in the terminal ileum can be a source of auto-antibodies, includinganti-folate receptor (FR) antibodies, which are commonly present in subjects with autism [80]. Under oxidative stress conditions, EAAT3-mediated uptake of cysteine by Treg cells is impaired, limiting their suppression of naıve CD4+ T cells, and increasing autoimmune responsesto antigens present locally, such as the folate receptor.

References

[1] P. Ashwood, A. Anthony, F. Torrente, and A. J. Wake-field, “Spontaneous mucosal lymphocyte cytokine profilesin children with autism and gastrointestinal symptoms:mucosal immune activation and reduced counter regulatoryinterleukin-10,” Journal of Clinical Immunology, vol. 24, no. 6,pp. 664–673, 2004.

[2] C. A. Pardo, D. L. Vargas, and A. W. Zimmerman, “Immunity,neuroglia and neuroinflammation in autism,” InternationalReview of Psychiatry, vol. 17, no. 6, pp. 485–495, 2005.

[3] P. Goines and J. van de Water, “The immune system’s role inthe biology of autism,” Current Opinion in Neurology, vol. 23,no. 2, pp. 111–117, 2010.

[4] G. A. Mostafa, A. Al Shehab, and N. R. Fouad, “Frequencyof CD4+CD25 high regulatory T cells in the peripheral bloodof Egyptian children with autism,” Journal of Child Neurology,vol. 25, no. 3, pp. 328–335, 2010.

[5] G. A. Mostafa, E. S. El-Hadidi, D. H. Hewedi, and M. M.Abdou, “Oxidative stress in Egyptian children with autism:relation to autoimmunity,” Journal of Neuroimmunology, vol.219, no. 1-2, pp. 114–118, 2010.

[6] C. Onore, M. Careaga, and P. Ashwood, “The role ofimmune dysfunction in the pathophysiology of autism,” Brain,Behavior, and Immunity, vol. 26, no. 3, pp. 383–392, 2012.

[7] D. A. Rossignol and R. E. Frye, “A review of researchtrends in physiological abnormalities in autism spectrumdisorders: immune dysregulation, inflammation, oxidative

stress, mitochondrial dysfunction and environmental toxicantexposures,” Molecular Psychiatry, vol. 17, no. 4, pp. 389–401,2012.

[8] H. Jyonouchi, L. Geng, D. L. Streck, and G. A. Toruner,“Children with autism spectrum disorders (ASD) who exhibitchronic gastrointestinal (GI) symptoms and marked fluctua-tion of behavioral symptoms exhibit distinct innate immuneabnormalities and transcriptional profiles of peripheral blood(PB) monocytes,” Journal of Neuroimmunology, vol. 238, pp.73–80, 2011.

[9] B. L. Williams, M. Hornig, T. Buie et al., “Impaired car-bohydrate digestion and transport and mucosal dysbiosis inthe intestines of children with autism and gastrointestinaldisturbances,” PLoS ONE, vol. 6, no. 9, Article ID e24585,2011.

[10] T. Buie, G. J. Fuchs III, G. T. Furuta et al., “Recommendationsfor evaluation and treatment of common gastrointestinalproblems in children with ASDs,” Pediatrics, vol. 125, no. 1,pp. S19–S29, 2010.

[11] T. Buie, D. B. Campbell, G. J. Fuchs et al., “Evaluation,diagnosis, and treatment of gastrointestinal disorders inindividuals with ASDs: a consensus report,” Pediatrics, vol.125, no. 1, pp. S1–S18, 2010.

[12] R. N. Nikolov, K. E. Bearss, J. Lettinga et al., “Gastrointestinalsymptoms in a sample of children with pervasive developmen-tal disorders,” Journal of Autism and Developmental Disorders,vol. 39, no. 3, pp. 405–413, 2009.


[13] A. P. Feinberg, “Epigenetics at the epicenter of modernmedicine,” Journal of the American Medical Association, vol.299, no. 11, pp. 1345–1350, 2008.

[14] J. M. Lasalle and D. H. Yasui, “Evolving role of MeCP2 in Rettsyndrome and autism,” Epigenomics, vol. 1, pp. 119–130, 2009.

[15] Y. Natsume-Kitatani, M. Shiga, and H. Mamitsuka, “Genome-wide integration on transcription factors, histone acetylationand gene expression reveals genes Co-regulated by histonemodification patterns,” PLoS ONE, vol. 6, no. 7, Article IDe22281, 2011.

[16] M. K. Skinner and C. Guerrero-Bosagna, “Environmentalsignals and transgenerational epigenetics,” Epigenomics, vol. 1,pp. 111–117, 2009.

[17] F. Perera and J. Herbstman, “Prenatal environmental expo-sures, epigenetics, and disease,” Reproductive Toxicology, vol.31, no. 3, pp. 363–373, 2011.

[18] R. Feil and M. F. Fraga, “Epigenetics and the environment:emerging patterns and implications,” Nature Reviews Genetics,vol. 13, pp. 97–109, 2012.

[19] Z. Hochberg, R. Feil, M. Constancia et al., “Child health,developmental plasticity, and epigenetic programming,”Endocrine Reviews, vol. 32, no. 2, pp. 159–224, 2011.

[20] J. M. LaSalle, “A genomic point-of-view on environmentalfactors influencing the human brain methylome,” Epigenetics,vol. 6, no. 7, pp. 862–869, 2011.

[21] M. Hornig, D. Chian, and W. I. Lipkin, “Neurotoxic effects ofpostnatal thimerosal are mouse strain dependent,” MolecularPsychiatry, vol. 9, no. 9, pp. 833–845, 2004.

[22] M. I. Waly, K. K. Kharbanda, and R. C. Deth, “Ethanol lowersglutathione in rat liver and brain and inhibits methioninesynthase in a cobalamin-dependent manner,” Alcoholism, vol.35, no. 2, pp. 277–283, 2011.

[23] D. P. Jones and Y. M. Go, “Redox compartmentalization andcellular stress,” Diabetes, Obesity and Metabolism, vol. 12, no.2, pp. 116–125, 2010.

[24] E. S. J. Arner and A. Holmgren, “Physiological functions ofthioredoxin and thioredoxin reductase,” European Journal ofBiochemistry, vol. 267, no. 20, pp. 6102–6109, 2000.

[25] F. Q. Schafer and G. R. Buettner, “Redox environment of thecell as viewed through the redox state of the glutathione disul-fide/glutathione couple,” Free Radical Biology and Medicine,vol. 30, no. 11, pp. 1191–1212, 2001.

[26] X. Sun, A. Y. Shih, H. C. Johannssen, H. Erb, P. Li, and T.H. Murphy, “Two-photon imaging of glutathione levels inintact brain indicates enhanced redox buffering in developingneurons and cells at the cerebrospinal fluid and blood-braininterface,” Journal of Biological Chemistry, vol. 281, no. 25, pp.17420–17431, 2006.

[27] S. M. Beer, E. R. Taylor, S. E. Brown et al., “Glutaredoxin2 catalyzes the reversible oxidation and glutathionylationof mitochondrial membrane thiol proteins: implications formitochondrial redox regulation and antioxidant defense,”Journal of Biological Chemistry, vol. 279, no. 46, pp. 47939–47951, 2004.

[28] I. Dalle-Donne, R. Rossi, G. Colombo, D. Giustarini, andA. Milzani, “Protein S-glutathionylation: a regulatory devicefrom bacteria to humans,” Trends in Biochemical Sciences, vol.34, no. 2, pp. 85–96, 2009.

[29] J. B. Adams, T. Audhya, S. McDonough-Means et al.,“Nutritional and metabolic status of children with autismvs. neurotypical children, and the association with autismseverity,” Nutrition and Metabolism, vol. 8, article 34, 2011.

[30] J. B. Adams, M. Baral, E. Geis et al., “The severity of autismis associated with toxic metal body burden and red blood cell

glutathione levels,” Journal of Toxicology, vol. 2009, Article ID532640, 7 pages, 2009.

[31] E. Pastural, S. Ritchie, Y. Lu et al., “Novel plasma phospholipidbiomarkers of autism: mitochondrial dysfunction as a putativecausative mechanism,” Prostaglandins Leukotrienes and Essen-tial Fatty Acids, vol. 81, no. 4, pp. 253–264, 2009.

[32] Y. Al-Gadani, A. El-Ansary, O. Attas, and L. Al-Ayadhi,“Metabolic biomarkers related to oxidative stress and antiox-idant status in Saudi autistic children,” Clinical Biochemistry,vol. 42, no. 10-11, pp. 1032–1040, 2009.

[33] S. P. Pasca, E. Dronca, T. Kaucsar et al., “One carbonmetabolism disturbances and the C677T MTHFR genepolymorphism in children with autism spectrum disorders,”Journal of Cellular and Molecular Medicine, vol. 13, no. 10, pp.4229–4238, 2009.

[34] D. A. Geier, J. K. Kern, C. R. Garver et al., “Biomarkers ofenvironmental toxicity and susceptibility in autism,” Journal ofthe Neurological Sciences, vol. 280, no. 1-2, pp. 101–108, 2009.

[35] D. A. Geier, J. K. Kern, C. R. Garver, J. B. Adams, T. Audhya,and M. R. Geier, “A prospective study of transsulfurationbiomarkers in autistic disorders,” Neurochemical Research, vol.34, no. 2, pp. 386–393, 2009.

[36] S. Jill James, S. Melnyk, S. Jernigan, A. Hubanks, S. Rose, andD. W. Gaylor, “Abnormal transmethylation/transsulfurationmetabolism and DNA hypomethylation among parents ofchildren with autism,” Journal of Autism and DevelopmentalDisorders, vol. 38, no. 10, pp. 1966–1975, 2008.

[37] S. J. James, S. Melnyk, S. Jernigan et al., “Metabolic endophe-notype and related genotypes are associated with oxidativestress in children with autism,” American Journal of MedicalGenetics B, vol. 141, no. 8, pp. 947–956, 2006.

[38] S. J. James, P. Cutler, S. Melnyk et al., “Metabolic biomarkers ofincreased oxidative stress and impaired methylation capacityin children with autism,” The American Journal of ClinicalNutrition, vol. 80, no. 6, pp. 1611–1617, 2004.

[39] S. Melnyk, G. J. Fuchs, E. Schulz et al., “Metabolic imbal-ance associated with methylation dysregulation and oxidativedamage in children with autism,” Journal of Autism andDevelopmental Disorders, vol. 42, no. 3, pp. 367–377, 2011.

[40] A. Chauhan and V. Chauhan, “Oxidative stress in autism,”Pathophysiology, vol. 13, no. 3, pp. 171–181, 2006.

[41] L. Palmieri, V. Papaleo, V. Porcelli et al., “Altered calciumhomeostasis in autism-spectrum disorders: evidence frombiochemical and genetic studies of the mitochondrial aspar-tate/glutamate carrier AGC1,” Molecular Psychiatry, vol. 15,no. 1, pp. 38–52, 2010.

[42] V. Napolioni, A. M. Persico, V. Porcelli, and L. Palmieri, “Themitochondrial aspartate/glutamate carrier AGC1 and calciumhomeostasis: physiological links and abnormalities in autism,”Molecular Neurobiology, vol. 44, no. 1, pp. 83–92, 2011.

[43] Y. Kanai and M. A. Hediger, “Primary structure and functionalcharacterization of a high-affinity glutamate transporter,”Nature, vol. 360, no. 6403, pp. 467–471, 1992.

[44] T. Iwanaga, M. Goto, and M. Watanabe, “Cellular distributionof glutamate transporters in the gastrointestinal tract ofmice. An immunohistochemical and in situ hybridizationapproach,” Biomedical Research, vol. 26, no. 6, pp. 271–278,2006.

[45] A. J. Wakefield, P. Ashwood, K. Limb, and A. Anthony, “Thesignificance of ileo-colonic lymphoid nodular hyperplasia inchildren with autistic spectrum disorder,” European Journal ofGastroenterology and Hepatology, vol. 17, no. 8, pp. 827–836,2005.


[46] K. Aoyama, M. Watabe, and T. Nakaki, “Regulation ofneuronal glutathione synthesis,” Journal of PharmacologicalSciences, vol. 108, no. 3, pp. 227–238, 2008.

[47] D. Trotti, N. C. Danbolt, and A. Volterra, “Glutamatetransporters are oxidant-vulnerable: a molecular link betweenoxidative and excitotoxic neurodegeneration?” Trends in Phar-macological Sciences, vol. 19, no. 8, pp. 328–334, 1998.

[48] S. Singh, S. Vrishni, B. K. Singh, I. Rahman, and P. Kakkar,“Nrf2-ARE stress response mechanism: a control point inoxidative stress-mediated dysfunctions and chronic inflam-matory diseases,” Free Radical Research, vol. 44, no. 11, pp.1267–1288, 2010.

[49] C. Escartin, S. J. Won, C. Malgorn et al., “Nuclear factorerythroid 2-related factor 2 facilitates neuronal glutathionesynthesis by upregulating neuronal excitatory amino acidtransporter 3 expression,” Journal of Neuroscience, vol. 31, no.20, pp. 7392–7401, 2011.

[50] S. Caja, M. Maki, K. Kaukinen, and K. Lindfors, “Antibodies inceliac disease: implications beyond diagnostics,” Cellular andMolecular Immunology, vol. 8, no. 2, pp. 103–109, 2011.

[51] F. R. Huebner, K. W. Lieberman, R. P. Rubino, and J. S.Wall, “Demonstration of high opioid-like activity in isolatedpeptides from wheat gluten hydrolysates,” Peptides, vol. 5, no.6, pp. 1139–1147, 1984.

[52] H. Teschemacher, “Opioid receptor ligands derived from foodproteins,” Current Pharmaceutical Design, vol. 9, no. 16, pp.1331–1344, 2003.

[53] M. S. Trivedi, N. W. Hodgson, and R. C. Deth, “Casein andgluten-derived peptides modulate cysteine uptake and cellularthiol levels via activation of opiate receptors,” Society forNeuroscience Meeting. In press.

[54] T. C. Petrossian and S. G. Clarke, “Uncovering the humanmethyltransferasome,” Molecular and Cellular Proteomics. Inpress.

[55] R. Deth, C. Muratore, J. Benzecry, V. A. Power-Charnitsky,and M. Waly, “How environmental and genetic factorscombine to cause autism: a redox/methylation hypothesis,”NeuroToxicology, vol. 29, no. 1, pp. 190–201, 2008.

[56] T. Ganesan, M. H. Khadra, J. Wallis, and D. E. Neal, “VitaminB12 malabsorption following bladder reconstruction or diver-sion with bowel segments,” ANZ Journal of Surgery, vol. 72, no.7, pp. 479–482, 2002.

[57] S. Vasilopoulos, K. Saiean, J. Emmons et al., “Terminal ileumresection is associated with higher plasma homocysteine levelsin Crohn’s disease,” Journal of Clinical Gastroenterology, vol.33, no. 2, pp. 132–136, 2001.

[58] M. E. McNally, S. A. Atkinson, and D. E. C. Cole, “Contribu-tion of sulfate and sulfoesters to total sulfur intake in infantsfed human milk,” Journal of Nutrition, vol. 121, no. 8, pp.1250–1254, 1991.

[59] J. C. Miner, E. Bisse, C. P. Aebischer, A. Bell, H. Wieland, andJ. Lutschg, “Assessment of vitamin B-12, folate, and vitaminB-6 status and relation to sulfur amino acid metabolism inneonates,” The American Journal of Clinical Nutrition, vol. 72,no. 3, pp. 751–757, 2000.

[60] T. Ozgurtas, I. Aydin, O. Turan et al., “Vascular endothelialgrowth factor, basic fibroblast growth factor, insulin-likegrowth factor-I and platelet-derived growth factor levels inhuman milk of mothers with term and preterm neonates,”Cytokine, vol. 50, no. 2, pp. 192–194, 2010.

[61] M. S. Murphy, “Growth factors and the gastrointestinal tract,”Nutrition, vol. 14, no. 10, pp. 771–774, 1998.

[62] N. Hodgson, M. Waly, and R. C. Deth, “EAAT3-mediatedcysteine uptake in human neuroblastoma cells is PI3 kinase-dependent,” in Proceedings of the Society for NeuroscienceMeeting, November 2008.

[63] H. Yang, N. Magilnick, M. Xia, and S. C. Lu, “Effects ofhepatocyte growth factor on glutathione synthesis, growth,and apoptosis is cell density-dependent,” Experimental CellResearch, vol. 314, no. 2, pp. 398–412, 2008.

[64] J. G. Dorea, “Selenium and breast-feeding,” British Journal ofNutrition, vol. 88, no. 5, pp. 443–461, 2002.

[65] S. D Watts, D. Torres-Salazar, C. B. Divito, and S. G. Amara,“EAAT3 cysteine transport: potential mechanism of cysteinebinding and translocation,” in Proceedings of the Society forNeuroscience Meeting, November 2009.

[66] J. J. Day and J. D. Sweatt, “Epigenetic mechanisms incognition,” Neuron, vol. 70, no. 5, pp. 813–829, 2011.

[67] A. Castagna, C. Le Grazie, A. Accordini et al., “Cerebrospinalfluid S-adenosylmethionine (SAMe) and glutathione concen-trations in HIV infection: effect of parenteral treatment withSAMe,” Neurology, vol. 45, no. 9, pp. 1678–1683, 1995.

[68] H. H. Tallan, S. Moore, and W. H. Stein, “L-cystathionine inhuman brain,” The Journal of Biological Chemistry, vol. 230,no. 2, pp. 707–716, 1958.

[69] J. Graff, D. Kim, M. M. Dobbin, and T. Li-Huei, “Epigeneticregulation of gene expression in physiological and patholog-ical brain processes,” Physiological Reviews, vol. 91, no. 2, pp.603–649, 2011.

[70] C. M. L. Carvalho, E. H. Chew, S. I. Hashemy, J. Lu, andA. Holmgren, “Inhibition of the human thioredoxin system:a molecular mechanism of mercury toxicity,” Journal ofBiological Chemistry, vol. 283, no. 18, pp. 11913–11923, 2008.

[71] N. V. C. Ralston and L. J. Raymond, “Dietary selenium’sprotective effects against methylmercury toxicity,” Toxicology,vol. 278, no. 1, pp. 112–123, 2010.

[72] K. T. Suzuki, C. Sasakura, and S. Yoneda, “Binding sites for the(Hg-Se) complex on selenoprotein P,” Biochimica et BiophysicaActa, vol. 1429, no. 1, pp. 102–112, 1998.

[73] A. Sharma, M. L. Kramer, P. F. Wick et al., “D4 dopaminereceptor-mediated phospholipid methylation and its implica-tions for mental illnesses such as schizophrenia,” MolecularPsychiatry, vol. 4, no. 3, pp. 235–246, 1999.

[74] J. M. Swanson, M. Kinsbourne, J. Nigg et al., “Etiologicsubtypes of attention-deficit/hyperactivity disorder: brainimaging, molecular genetic and environmental factors and thedopamine hypothesis,” Neuropsychology Review, vol. 17, no. 1,pp. 39–59, 2007.

[75] A. Y. Kuznetsova and R. C. Deth, “A model for modulation ofneuronal synchronization by D4 dopamine receptor-mediatedphospholipid methylation,” Journal of Computational Neuro-science, vol. 24, no. 3, pp. 314–329, 2008.

[76] T. Demiralp, C. S. Herrmann, M. E. Erdal et al., “DRD4and DAT1 polymorphisms modulate human gamma bandresponses,” Cerebral Cortex, vol. 17, no. 5, pp. 1007–1019,2007.

[77] Z. Yan, S. K. Garg, and R. Banerjee, “Regulatory T cellsinterfere with glutathione metabolism in dendritic cells andT cells,” Journal of Biological Chemistry, vol. 285, no. 53, pp.41525–41532, 2010.


[79] V. T. Ramaekers, J. M. Sequeira, N. Blau, and E. V. Quadros, “Amilk-free diet downregulates folate receptor autoimmunity in


cerebral folate deficiency syndrome,” Developmental Medicineand Child Neurology, vol. 50, no. 5, pp. 346–352, 2008.

[80] R. E. Frye, J. M. Sequeira, E. V. Quadros, S. J. James, and D. A.Rossignol, “Cerebral folate receptor autoantibodies in autismspectrum disorder,” Molecular Psychiatry. In press.

[81] J. K. Polansky, L. Schreiber, C. Thelemann et al., “Methylationmatters: Binding of Ets-1 to the demethylated Foxp3 gene con-tributes to the stabilization of Foxp3 expression in regulatory Tcells,” Journal of Molecular Medicine, vol. 88, no. 10, pp. 1029–1040, 2010.

[82] S. Hussain, S. J. Kirwin, and S. A. Stohlman, “Increased Tregulatory cells lead to development of Th2 immune responsein male SJL mice,” Autoimmunity, vol. 44, no. 3, pp. 219–228,2011.


Research Article

Decreased Levels of EGF in Plasma of Children with AutismSpectrum Disorder

Charity Onore,1, 2 Judy Van de Water,2, 3 and Paul Ashwood1, 2

1 Department of Medical Microbiology and Immunology, University of California, Davis, USA2 The M.I.N.D. Institute, University of California, Davis, Sacramento, CA 95817, USA3 Division of Rheumatology, Allergy and Clinical Immunology, University of California, Davis, USA

Correspondence should be addressed to Paul Ashwood, [email protected]

Received 14 September 2011; Revised 3 November 2011; Accepted 16 December 2011


Copyright © 2012 Charity Onore et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Autism Spectrum Disorder (ASD) is a neurodevelopmental disorder estimated to affect 1 in 110 children in the U.S., yet thepathology of this disorder is not fully understood. Abnormal levels of several growth factors have been demonstrated in adultswith ASD, including epidermal growth factor (EGF) and hepatocyte growth factor (HGF). Both of these growth factors serveimportant roles in neurodevelopment and immune function. In this study, concentrations of EGF and HGF were assessed in theplasma of 49 children with ASD aged 2–4 years old, and 31 typically developing controls of a similar age as part of the AutismPhenome Project (APP). Levels of EGF were significantly reduced in the ASD group compared to typically developing controls(P = 0.003). There were no significant differences in HGF levels in young children with ASD and typically developing controls.EGF plays an important role in regulating neural growth, proliferation, differentiation and migration, and reduced levels of thismolecule may negatively impact neurodevelopment in young children with ASD.

1. Introduction

Autism Spectrum Disorder (ASD) is a developmental disor-der characterized by impairments in social interaction andcommunication, and the presence of restricted behaviors orinterests [1]. According to the most current CDC estimate,ASD affects 1 in 110 children in the US [2], yet the patho-physiology of the disorder is largely unknown. Recently, sev-eral growth factors have been found to be dysregulated in asubstantial proportion of adults with ASD [3].

In the central nervous system, growth factors can regulatethe processes of neuronal growth, differentiation, and pro-liferation, as well as regulating neuronal survival, neuronalmigration, and the formation or elimination of synapses [3].In addition to their central function in regulating neuro-development, current literature has also illustrated the dualnature of many growth factors as immune modulators andhighlighted their involvement in crosstalk between the im-mune system and the central nervous system (CNS) [4–7].Many studies suggest the presence of aberrant immuneactivity in ASD, in the CNS [8, 9] and in the periphery

[10–12], which may be influenced by atypical growth factoractivity. Growth factor dysregulation may contribute to ASDpathology by directly affecting CNS development, and/or byaugmenting immune function.

Epidermal growth factor (EGF) and hepatocyte growthfactor (HGF) are both involved in the growth and prolifera-tion of several cell types, including neurons and glia of theCNS. EGF is present at high levels in the central nervoussystem (CNS) and plays a critical role in controlling pro-liferation and differentiation of nervous tissue during neu-rogenesis [13, 14]. In addition to the function of EGF as aCNS growth factor, it is a central factor in promoting woundhealing. EGF is expressed at sites of injury and inhibits theactivity of nitric oxide synthase, preventing inflammation[4, 15]. EGF deficiency results in several neurological,gastrointestinal, dermal, and pulmonary abnormalities inanimal models [16]. An increased frequency of EGF singlenucleotide polymorphisms has been reported in ASD, as wellas lower plasma EGF levels in adults with autism [17, 18].

HGF also plays a dual role as a neurological growth factorand an immune modulator. HGF can modulate immune


responses by signaling through the MET receptor on antigenpresenting cells, which results in a tolerogenic phenotypewith reduced proinflammatory cytokine production andcellular activity [5]. HGF is also essential for normal neu-rodevelopment, and disruption of HGF signaling results incomplex alterations in GABAergic neuron development inthe forebrain of animal models [19]. Lower levels of HGFin sera of autistic adults have been described [20], as wellas decreased expression of the HGF receptor in postmortembrain samples [21].

To determine if there exists a differential profile forperipheral blood growth factor levels in ASD, we analyzedplasma EGF and HGF in well-characterized young childrenaged 2–4 years old with a diagnosis of ASD and unrelatedtypically developing children who were frequency-matchedfor age. In addition, levels of plasma EGF and HGF wereinvestigated for any associations with clinical behavioral anddevelopmental outcomes.

2. Methods and Materials

2.1. Subjects and Behavioral Assessments. Eighty study par-ticipants aged between 2–4 years of age were recruited aspart of the APP [22]. Participants consisted of 49 childrenwith ASD (median age of 2.88 years, interquartile range2.66–3.41 years, 42 males) and 31 typically developing (TD)children (median age of 2.96 years, interquartile range2.85–3.27, 20 males). Diagnostic instruments included theAutism Diagnostic Observation Schedule-Generic (ADOS-G) [23] and the Autism Diagnostic Interview-Revised (ADI-R) [24]. All diagnostic assessments were conducted ordirectly observed by trained, licensed clinical psychologistswho specialize in autism and had been trained accordingto research standards for these tools. Inclusion criteria forASD were taken from the diagnostic definition of ASDin young children formulated and agreed upon by theCollaborative Programs of Excellence in Autism. Inclusioncriteria for TD controls included developmental scoreswithin two standard deviations of the mean on all subscalesof the MSEL. Exclusion criteria for TD controls includeda diagnosis of mental retardation, pervasive developmentaldisorder or specific language impairment, or any knowndevelopmental, neurological, or behavioral problems. TDchildren were screened and excluded for autism with theSocial Communication Questionnaire (scores > 11) (SCQ—Lifetime Edition) [25].

2.2. Measurement of EGF. Peripheral blood was collected inacid-citrate-dextrose Vacutainers (BD Biosciences, San Jose,CA). The plasma fraction was immediately harvested bycentrifugation and stored as aliquots at −80◦C until the dateof assay. Plasma levels of EGF were determined by HumanEGF enzyme-linked immunosorbant assay (ELISA) kit (R&DSystems, Minneapolis, MN). Plasma levels of HGF weredetermined by Human HGF enzyme-linked immunosorbantassay (ELISA) kit (R&D Systems). The assays were performedaccording to the protocols provided by the manufacturer,and all samples were assayed in duplicate. Optical density

was measured on a Wallac Victor3 multilabel-plate reader(PerkinElmer, Boston, MA) at 450 nm.

2.3. Statistical Analysis. Statistical analysis to compare levelsof growth factors between ASD and TD groups was con-ducted with unpaired Student’s t-test. All analyses were con-ducted with GraphPad Prism statistical software (GraphPadSoftware Inc., San Diego, CA).

3. Results

Plasma levels of EGF were approximately 3-fold lower inthe ASD group (23.1 ± 6.2 pg/mL) compared with the TDgroup (76.9±20.0 pg/mL) (P = 0.003). Plasma levels of HGFwere not statistically different between children with ASD(423.5± 20.9 pg/mL) and TD controls (436.6± 19.2 pg/mL)(Figure 1).

4. Discussion

In this study, we investigated the levels of EGF and HGF,two growth factors involved in neurodevelopment. Previousstudies have found decreased levels of EGF in high function-ing adults with autism [18], but no studies have looked atchildren with ASD who are close to the onset of the disorder.We found that levels of plasma EGF were significantlyreduced in young children with ASD as compared withsimilarly age-matched typically developing control children.This finding is consistent with that seen in adults with highfunctioning autism and may suggest that a deficiency in EGFis persistent throughout the time course of ASD.

Current research has demonstrated that EGF is involvedin growth, differentiation, and maintenance of several tissuesincluding the CNS and the gastrointestinal tract (GI) [14].Notably, abnormalities of both the CNS and the GI have beenreported in ASD [26]. In the CNS, EGF serves as a potentneurotrophic factor, and in vivo studies have demonstratedthe effect of EGF in promoting proliferation, differentiation,survival, and migration of multipotent neural progenitorcells and the differentiation of these cells into astrocytes andneurons [27]. In this paper, we were limited to investigatingEGF in plasma and not in cerebral spinal fluid. However,EGF rapidly transports across the blood-brain barrier [28],suggesting peripheral EGF levels could be representative oflevels in the CNS that may impact neurodevelopment.

In the GI tract, EGF is necessary for normal developmentof the intestinal mucosa, and mice deficient in EGF receptorsuffer from symptoms similar to those of necrotizing ente-rocolitis, with gradual destruction of villi, become severelymalnourished, and typically die before postnatal day 8[16]. EGF promotes wound healing in animal models ofulcerative colitis [29, 30]. Although GI symptoms affect alarge proportion of children with ASD, the exact nature andextent of GI inflammation in ASD are still controversial [31–34]. Interestingly, preliminary investigations in this studyshow that EGF levels are associated with increased bloatingin children with ASD, as well as with raw, standard, andpercentile rank scores on the Peabody picture vocabulary


ASD0

25

50

75

100

ASDControl

Control

P = 0.0032

EG

F (p

g/m

L)

(a)

ASD0

100

200

300

400

500

ASDControl

Control

HG

F (p

g/m

L)

(b)

Figure 1: Plasma levels of EGF and HGF. Levels of EGF and HGF in peripheral blood plasma from ASD participants and similarly agedtypically developing controls as measured by ELISA (data is shown as mean and SEM). EGF levels are significantly lower in ASD subjects incomparison to typically developing controls (P = 0.003). HGF levels did not differ significantly between the two groups. Significance wasdetermined by two-tailed unpaired Student’s t-test.

test-III and with the composite and nonverbal developmentquotient of the Mullen’s test (P < 0.05, data not shown).However, the meaning of this data is not clear and we willvalidate these findings in a larger replication cohort.

It has been previously reported that HGF was decreasedin the plasma of adults with autism [20]; however, we did notfind a significant difference between children with ASD andtypically developing children for plasma HGF levels. Thismay be due to age differences between the participants of ourstudy and those in the previous reports. To our knowledge,the relationship between age and serum HGF levels has notbeen thoroughly established, but there is evidence that levelsmay decrease with age [35].

Collectively, our data suggest that reduced levels of EGFare present in ASD. The roles of EGF and ASD require furtherresearch to better elucidate the relationship between thispotent growth factor and ASD pathophysiology.

References

[1] APA, Diagnostic and Statistical Manual of Mental Disorders,American Psychiatric Publishing, Inc., Arlington, Va, USA,2000.

[2] C. Rice, “Prevalence of autism spectrum disorders—autismand developmental disabilities monitoring network, UnitedStates, 2006,” Morbidity and Mortality Weekly Report, vol. 58,no. 10, pp. 1–20, 2009.

[3] T. Nickl-Jockschat and T. M. Michel, “The role of neurotroph-ic factors in autism,” Molecular Psychiatry, vol. 16, pp. 478–490, 2011.

[4] D. E. Heck, D. L. Laskin, C. R. Gardner, and J. D. Laskin, “Epi-dermal growth factor suppresses nitric oxide and hydrogenperoxide production by keratinocytes. Potential role for nitricoxide in the regulation of wound healing,” Journal of BiologicalChemistry, vol. 267, no. 30, pp. 21277–21280, 1992.

[5] K. Okunishi, M. Dohi, K. Nakagome et al., “A novel role ofhepatocyte growth factor as an immune regulator throughsuppressing dendritic cell function,” Journal of Immunology,vol. 175, no. 7, pp. 4745–4753, 2005.

[6] J. A. Vega, O. Garcıa-Suarez, J. Hannestad, M. Perez-Perez, andA. Germana, “Neurotrophins and the immune system,” Jour-nal of Anatomy, vol. 203, no. 1, pp. 1–19, 2003.

[7] R. Tabakman, S. Lecht, S. Sephanova, H. Arien-Zakay, and P.Lazarovici, “Interactions between the cells of the immune andnervous system: neurotrophins as neuroprotection mediatorsin CNS injury,” Progress in Brain Research, vol. 146, pp. 387–401, 2004.

[8] D. L. Vargas, C. Nascimbene, C. Krishnan, A. W. Zimmerman,and C. A. Pardo, “Neuroglial activation and neuroinflamma-tion in the brain of patients with autism,” Annals of Neurology,vol. 57, no. 1, pp. 67–81, 2005.

[9] X. Li, A. Chauhan, A. M. Sheikh et al., “Elevated immune re-sponse in the brain of autistic patients,” Journal of Neuroim-munology, vol. 207, no. 1-2, pp. 111–116, 2009.

[10] P. Ashwood, P. Krakowiak, I. Hertz-Picciotto, R. Hansen, I.N. Pessah, and J. Van de Water, “Altered T cell responses in


children with autism,” Brain, Behavior, and Immunity, vol. 25,no. 5, pp. 840–849, 2011.

[11] A. M. Enstrom, C. E. Onore, J. A. Van de Water, and P.Ashwood, “Differential monocyte responses to TLR ligandsin children with autism spectrum disorders,” Brain, Behavior,and Immunity, vol. 24, no. 1, pp. 64–71, 2010.

[12] P. Ashwood, P. Krakowiak, I. Hertz-Picciotto, R. Hansen, I.Pessah, and J. Van de Water, “Elevated plasma cytokines inautism spectrum disorders provide evidence of immune dys-function and are associated with impaired behavioral out-come,” Brain, Behavior, and Immunity, vol. 25, no. 1, pp. 40–45, 2011.

[13] C. J. Xian and X. F. Zhou, “Roles of transforming growth fac-tor-α and related molecules in the nervous system,” MolecularNeurobiology, vol. 20, no. 2-3, pp. 157–183, 1999.

[14] C. J. Xian and X. F. Zhou, “EGF family of growth factors:essential roles and functional redundancy in the nerve system,”Frontiers in Bioscience, vol. 9, pp. 85–92, 2004.

[15] S. Pastore and F. Mascia, “Novel acquisitions on the immuno-protective roles of the EGF receptor in the skin,” Expert Reviewof Dermatology, vol. 3, no. 5, pp. 525–527, 2008.

[16] P. J. Miettinen, J. E. Berger, J. Meneses et al., “Epithelial im-maturity and multiorgan failure in mice lacking epidermalgrowth factor receptor,” Nature, vol. 376, no. 6538, pp. 337–341, 1995.

[17] T. Toyoda, K. Nakamura, K. Yamada et al., “SNP analyses ofgrowth factor genes EGF, TGFβ-1, and HGF reveal haplotypicassociation of EGF with autism,” Biochemical and BiophysicalResearch Communications, vol. 360, no. 4, pp. 715–720, 2007.

[18] K. Suzuki, K. Hashimoto, Y. Iwata et al., “Decreased serum lev-els of epidermal growth factor in adult subjects with high-functioning autism,” Biological Psychiatry, vol. 62, no. 3, pp.267–269, 2007.

[19] P. Levitt, K. L. Eagleson, and E. M. Powell, “Regulation of ne-ocortical interneuron development and the implications forneurodevelopmental disorders,” Trends in Neurosciences, vol.27, no. 7, pp. 400–406, 2004.

[20] G. Sugihara, K. Hashimoto, Y. Iwata et al., “Decreased serumlevels of hepatocyte growth factor in male adults with high-functioning autism,” Progress in Neuro-Psychopharmacologyand Biological Psychiatry, vol. 31, no. 2, pp. 412–415, 2007.

[21] D. B. Campbell, R. D’Oronzio, K. Garbett et al., “Disruption ofcerebral cortex MET signaling in autism spectrum disorder,”Annals of Neurology, vol. 62, no. 3, pp. 243–250, 2007.

[22] E. Singer, ““Phenome” project set to pin down subgroups ofautism,” Nature Medicine, vol. 11, no. 6, p. 583, 2005.

[23] C. Lord, M. Rutter, S. Goode et al., “Autism diagnostic obser-vation schedule: a standardized observation of communicativeand social behavior,” Journal of Autism and DevelopmentalDisorders, vol. 19, no. 2, pp. 185–212, 1989.

[24] C. Lord, M. Rutter, and A. L. Couteur, “Autism diagnosticinterview-revised: a revised version of a diagnostic interviewfor caregivers of individuals with possible pervasive devel-opmental disorders,” Journal of Autism and DevelopmentalDisorders, vol. 24, no. 5, pp. 659–685, 1994.

[25] S. K. Berument, M. Rutter, C. Lord, A. Pickles, and A. Bailey,“Autism screening questionnaire: diagnostic validity,” BritishJournal of Psychiatry, vol. 175, pp. 444–451, 1999.

[26] M. Careaga, J. Van de Water, and P. Ashwood, “Immunedysfunction in autism: a pathway to treatment,” Neurothera-peutics, vol. 7, no. 3, pp. 283–292, 2010.

[27] C. G. Craig, V. Tropepe, C. M. Morshead, B. A. Reynolds, S.Weiss, and D. Van Der Kooy, “In vivo growth factor expansion

of endogenous subependymal neural precursor cell popula-tions in the adult mouse brain,” Journal of Neuroscience, vol.16, no. 8, pp. 2649–2658, 1996.

[28] W. Pan and A. J. Kastin, “Entry of EGF into brain is rapid andsaturable,” Peptides, vol. 20, no. 9, pp. 1091–1098, 1999.

[29] R. J. Farrell, “Epidermal growth factor for ulcerative colitis,”New England Journal of Medicine, vol. 349, no. 4, pp. 395–397,2003.

[30] A. Sinha, J. Nightingale, K. P. West, J. Berlanga-Acosta, andR. J. Playford, “Epidermal growth factor enemas with oralmesalamine for mild-to-moderate left-sided ulcerative colitisor proctitis,” New England Journal of Medicine, vol. 349, no. 4,pp. 350–357, 2003.

[31] K. Horvath and J. A. Perman, “Autism and gastrointestinalsymptoms,” Current Gastroenterology Reports, vol. 4, no. 3, pp.251–258, 2002.

[32] P. Ashwood, A. Anthony, A. A. Pellicer, F. Torrente, J. A.Walker-Smith, and A. J. Wakefield, “Intestinal lymphocytepopulations in children with regressive autism: evidence forextensive mucosal immunopathology,” Journal of ClinicalImmunology, vol. 23, no. 6, pp. 504–517, 2003.

[33] P. Ashwood, A. Anthony, F. Torrente, and A. J. Wakefield,“Spontaneous mucosal lymphocyte cytokine profiles in chil-dren with autism and gastrointestinal symptoms: mucosal im-mune activation and reduced counter regulatory interleukin-10,” Journal of Clinical Immunology, vol. 24, no. 6, pp. 664–673,2004.

[34] P. Ashwood and A. J. Wakefield, “Immune activation of pe-ripheral blood and mucosal CD3+ lymphocyte cytokine pro-files in children with autism and gastrointestinal symptoms,”Journal of Neuroimmunology, vol. 173, no. 1-2, pp. 126–134,2006.

[35] R. Ramirez, D. Hsu, A. Patel et al., “Over-expression of hepat-ocyte growth factor/scatter factor (HGF/SF) and the HGF/SFreceptor (cMET) are associated with a high risk of metastasisand recurrence for children and young adults with papillarythyroid carcinoma,” Clinical Endocrinology, vol. 53, no. 5, pp.635–644, 2000.


Review Article

HLA Immune Function Genes in Autism

Anthony R. Torres,1 Jonna B. Westover,1 and Allen J. Rosenspire2

1 Center for Persons with Disabilities, Utah State University, 6804 Old Main Hill, Logan, UT 84322, USA2 Department of Immunology and Microbiology, School of Medicine, Wayne State University, 221 Lande Building, Detroit,MI 48201, USA

Correspondence should be addressed to Anthony R. Torres, [email protected]

Received 10 October 2011; Accepted 11 November 2011

Academic Editor: Judy Van de Water

Copyright © 2012 Anthony R. Torres et al. This is an open access article distributed under the Creative Commons AttributionLicense, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properlycited.

The human leukocyte antigen (HLA) genes on chromosome 6 are instrumental in many innate and adaptive immune responses.The HLA genes/haplotypes can also be involved in immune dysfunction and autoimmune diseases. It is now becoming apparentthat many of the non-antigen-presenting HLA genes make significant contributions to autoimmune diseases. Interestingly, it hasbeen reported that autism subjects often have associations with HLA genes/haplotypes, suggesting an underlying dysregulationof the immune system mediated by HLA genes. Genetic studies have only succeeded in identifying autism-causing genes ina small number of subjects suggesting that the genome has not been adequately interrogated. Close examination of the HLAregion in autism has been relatively ignored, largely due to extraordinary genetic complexity. It is our proposition that geneticpolymorphisms in the HLA region, especially in the non-antigen-presenting regions, may be important in the etiology of autismin certain subjects.

1. Autism

Leo Kanner first described autism in 1943 [1] after finding 11children with common symptoms of obsessiveness, stereo-typy, and echolalia at Johns Hopkins University. Autismremained an esoteric disorder for several decades untilphysicians and parents connected these symptoms with anincreasing number of patients. It is important to note thatthe diagnostic criteria have been modified over the yearsto include a broader category of symptoms, thus increasingthe number of children diagnosed with the disorder, nowreferred to as Autism Spectrum Disorder (ASD) [2]. Current-ly, the Centers for Disease Control and Prevention (CDC)states that the incidence of ASD is 1 out of 110 children inthe United States [3]. The severity of ASD varies greatly withthe most severe forms, much like Kanner autism, displayinglanguage regression, seizures, and lower IQ. Altevogt et al.[4] have suggested that autism, or more properly ASD, is nota single disorder, but a collection of similar disorders eachwith different characteristics and perhaps etiologies.

Even after several decades of research, there is muchdebate around the world on the etiology of ASD. It is clearthat ASD results from abnormal brain development in either

the prenatal period or infancy stage of life. Exposure tomercury, maternal viral infections, autoimmune disorders,and the inheritance of certain gene combinations have beenimplicated in the etiology. Unfortunately, none of these areashave given clear answers as to the etiology. Fortunately,psychologists have made significant strides in treating chil-dren and it appears the earlier behavioral treatment starts,the better the outcome. Nevertheless, medical researcherscontinue to search for the cause(s) of ASD. This paperdiscusses a possible role for the immune system, and inparticular immune function genes in the human leukocyteantigen region (HLA), as a research area that should be moreclosely investigated.

2. Infections

One of the first areas of interest in the 1960s and 1970s wasthe search for an infectious agent that might be involved inthe etiology of ASD [5]. During this time, there were manycase reports in the literature that suggested an associationbetween congenital rubella infection and resultant autisticbehaviors. However, after decades of research no definite role


for infectious agents in autism etiology has been confirmed.On the other hand, these endeavors have led to observationsthat perhaps the immune system was involved in autism, andevidence continues to mount that immune abnormalities areindeed associated with ASD.

3. Familial Studies

Studies indicating familial clustering and the increases ofASD in twins have been interpreted by many as an indicationof genetic predisposition. Twin studies show the concordancerates of monozygotic twins at 36–96%, whereas dizygotictwins are 0–24% concordant, resulting in an estimated herit-ability of autism at >90% [6–8]. Additionally, family studieshave shown autism to have familial aggregation with 3–8% ofsubsequently born siblings either being autistic or showingsome form of pervasive developmental disorder (PDD) [9,10]. This is a 3- to 8-fold increase in risk for siblings over thegeneral population. A more recent study gave an estimate of18.7% sibling recurrence risk for ASD, a 20-fold increase overthe general population [11]. It is important to note, however,that family data should be looked at with great caution, asindividuals living in the same household will have similarexposures to microorganisms and environmental chemicals.Taking this into account, a recent paper based on data fromtwin pairs estimates the genetic heritability for ASD to be14–67% [12]. However, this idea is somewhat controversialas many in the research community continue to feel that theresults from family studies are indicative of a strong geneticetiology [13, 14].

4. Genetics

Many of the early genetic studies involved the examination ofmicrosatellites throughout the human genome in an attemptto find genomic regions that would associate with autism.Overall, this approach was not very fruitful and researchersquickly started to examine single-nucleotide polymorphisms(SNPs) as techniques advanced [15]. With this approachmany researchers proposed that multiple candidate geneswere associated with ASD. Unfortunately, studies proposingcandidate genes were often contradictory and proved to beunreliable [15]. One of the most interesting genetic findingsin ASD is the association of autism with the MET receptortyrosine kinase gene located on chromosome 7q31 as METsignaling participates in gastrointestinal repair, immunefunction, neocortical, and cerebellar growth [16]. It is impor-tant to mention that the autism MET associations havebeen replicated by other research groups [17]. This autismassociation showed a relative risk of 2.27 which is muchlower than the relative risk for HLA gene/allele associationsdiscussed below.

It is very reasonable to believe that deletions or dupli-cations of genetic regions, which can cause lower or higherlevels of gene expression, could produce pathological phe-notypes. Consequently, newer approaches examining copynumber variation (CNV) and microarray analyses of 500,000SNPs or more have been in vogue for several years in the

study of ASD. These newer approaches have identified CNVmutation differences in genes involved in neuronal celladhesion and ubiquitin degradation as being associated withASD [18]; however, these results have yet to be replicated byother researchers.

The neurexin-1 gene has been associated with a variety ofdevelopmental disorders including ASD [19]. The neurexin-1 gene (2p16.3) is one of the largest genes in the humangenome with 24 exons in 1.1 Mb. With two independentpromotors there can be over 1,000 neurexin isoforms gener-ated from the 24 exons in different cells or tissue. Anothergene of interest is the contactin-associated protein-like 2(CNTNAP2) gene that was shown to be associated inOld-Order Amish subjects with intractable epilepsy andASD [20]. Three other groups have now confirmed theinvolvement of CNTNAP2 in ASD [21–23].

Both the neurexin-1 and the CNTNAP2 genes are in-volved in synaptic function. Although these approaches haveshown a strong association of certain genes with ASD,only a small percentage of subjects with ASD have thesemutations. For example, the neurexin-1gene is found in onlyabout 0.5% of autism cases and 0.2% of controls. The 90%missing inheritance may be largely due to marked geneticheterogeneity, suggesting that different ASD phenotypesshould be examined separately [24–26]. Recent genetic re-search has also associated numerous immune function geneswith autism [27–30]. A large study that examined SNP datafrom several genomewide scans on 3,130 subjects with schiz-ophrenia found that the 5 most significant SNP markers arefound across the HLA region [31]. It appears that some ofmissing inheritance, at least in schizophrenia, was uncoveredin the HLA region and we suggest a similar finding will beconfirmed/uncovered in autism.

5. Immune Abnormalities in Autism

It is no surprise to see immune gene associations in ASD, asnumerous researchers have reported immune abnormalitiesin autism for over 20 years. It has become increasingly obvi-ous that inflammatory processes are associated with autism.Blood levels of the inflammatory cytokines IL-6, INF-γ, andTNF-αwere shown to be elevated in autistic individuals com-pared to controls [32, 33]. Later, seminal work by Vargas etal. [34] utilized direct morphological analysis and immuno-histologic techniques to show an active neuroinflammatoryprocess in the cerebral cortex, white matter, and in particularthe cerebellum of ASD patients that was dependent onactivation of microglia and astroglia. Cytokine profilingdemonstrated that neuroinflammation was accompanied byupregulation of the macrophage chemoattractant protein(MCP-1) and TGF-beta in brain tissue, and that MCP-1 wasalso upregulated in cerebral spinal fluid [34]. More recentwork has directly demonstrated that aside from blood, IL-6, TNF-α, and INF-γ are elevated in ASD brains, along withthe other inflammatory cytokines GM-CSF and IL-8 [35].Most recently, it has been reported that the importantinflammatory-associated transcription factor, nuclear factorkappa-light-chain-enhancer of activated B cells (NF-κB) is


Table 1: A list of proteins against which autoantibodies have beenfound in the serum of subjects with autism.

Protein Reference

Transglutaminase 2 [39]

45 and 62 kDa proteins in cerebellum [40, 41]

Voltage dependent anion channel (VDAC) [42]

Hexokinase-1 [42]

Mitochondrial protein [43]

Antimitochondrial DNA auto-antibodies [43]

Nuclear proteins [44]

52 kDa protein in hypothalamus and thalamus [45]

43–48 kDa protein in the hypothalamus [45]

Folate receptor [46]

Brain-derived neuro trophic factor [47]

HSP90 [48]

Myelin basic protein (MBP) [38, 49]

Myelin-associated glycoprotein (MAG) [50]

Myelin oligodendrocyte glycoprotein (MOG) [50]

Neuron-axon filament protein [51]

Glial fibrillary acidic protein [51]

upregulated in both blood [36] and brain tissue [37] ofautistic individuals. Other immune abnormalities such asautoantibodies started to be reported in 1993 [38] (Table 1).

Autoantibodies to myelin basic protein have been notedby at least a couple of researchers [38, 52]. An increasein autoantibody reactivity has been reported against otherbrain proteins in ASD including nerve growth factor [53],brain endothelium [54], cerebellar proteins [40], and sero-tonin 5-HT receptors [55] and transglutaminase-2, a proteinimportant in synaptic stabilization [39]. Croonenberghs etal. [56] noted a significant increase in gamma globulinespecially of the IgG2 and IgG4 subclasses in children withautism over a control population. Autoantibodies to severaluncharacterized brain-specific proteins have been reportedin the plasma of autistic individuals. In particular, westernblot analysis has shown the presence of IgG autoantibodiestargeting a protein of approximately 52 kDA located in thehypothalamus and thalamus of adult brain [45]. Otherautoantibodies targeting three brain proteins of 43–48 kDAlocated in the hypothalamus have also been observed in theserum of autistic individuals [45]. Autoantibodies targetingcerebellar proteins of 45 and 62 kDA have been associatedwith ASD [40]. These autoantibodies may be specific tocerebellar Golgi cells, which are GABAergic interneurons[41]. Autoantibodies reactive to human brain proteins in the36–39 and 61 kDA range have also been found in the sera ofmothers of autistic children. [57, 58].

It is important to note that while autoantibodies associ-ated with ASD may be biomarkers, they may not necessarilybe pathologic in and of themselves. Central tolerance refersto the process whereby immature lymphocytes are negativelyselected based on the ability of their antigen receptors (theB cell receptor or BCR for B cells and the T-cell receptoror TCR for T cells) to recognize self-antigens [59–61]. Until

fairly recently, it was believed that most immature lym-phocytes recognizing self-antigens (autoimmune repertoire)were normally neutralized by virtue of central tolerancebefore they could mature, and that peripheral tolerancewould insure the removal of any self-reactive lymphocyteescaping central tolerance. Peripheral tolerance refers to theprocess whereby mature self-reactive lymphocytes whichhave escaped central tolerance are eliminated, largely byCD95-mediated apoptosis [62]. Thus responses to foreignantigens were viewed as normal, while anti-self responseswere considered necessarily pathologic. However, with therealization that many self-reactive lymphocytes survive cen-tral and peripheral tolerance, this view has had to bemodified [49]. Limited immune responses to self-antigens(autoimmunity) are now understood to be normal and notnecessarily pathologic [63].

Upon stimulation of the system with a pathogen, cognatelymphocyte clones representative of the antiforeign reper-toire normally expand and mature, providing a protectiveimmune response. On the other hand, it is the expansion andmaturation of those clones representative of the autoimmunerepertoire that leads to autoimmune disease. In other words,the developing immune system can be characterized as bal-ancing production between antiforeign (protective) and anti-self (autoimmune) repertoires. While a beneficial function ofnaturally occurring, low level autoimmune antibodies, alsoreferred to as natural antibodies (NAs), remains a matter ofdebate, it appears as if the repertoire of NA is reflective ofthe susceptibility to develop specific autoimmune diseases[64, 65]. Because central tolerance of T cells depends to alarge extent upon the strength of the TCR interaction withan autoantigen or an HLA class I or II complex [59], theNA repertoire will to a large extent depend upon the HLAhaplotype, with some haplotypes favoring autoantibodiestargeting one antigen over another. For instance, in the caseof ASD we have found an association with low level antibodyresponses to tissue transglutaminase, and that this responseappears linked to the HLA-DR3/DQ2 and DR7/DQ2 hap-lotypes [39]. It is likely that the other autoantibodies notedabove as being associated with ASD may be linked to differenthaplotypes.

Aside from autoantibodies and altered cytokine levelsthere appear to be other immune abnormalities associatedwith ASD. It was noted that there were decreased numbersof T-lymphocytes and an altered ratio of suppressor T-lymphocytes to helper T-lymphocytes [66, 67] and altered T-lymphocytes responses in children with autism [68]. Warrenet al. [69] reported that subjects with autism had reducedNK cell killing in the standard K562 target cell cytotoxicityassay. This observation of decreased NK cell killing has beenrepeated by at least a couple of other research teams [28, 70].One research group [71] observed that monocyte counts andneopterin levels were increased in autistic children comparedto gender and age-matched healthy controls suggesting thatthe immune system was overactivated in the ASD group.Another elegant set of experiments involved the stimulationof cultured monocytes with several toll-like-receptor (TLR)ligands. The monocytes from subjects with ASD had signif-icant increases or decreases in proinflammatory cytokines


Table 2: Genes and alleles in the HLA region.

HLA genes Number HLA alleles Number

HLA class I genes 6 HLA class I alleles 2215

HLA class II genes 12 HLA class II alleles 986

HLA class I-like genes 2 HLA class I-like alleles 94

Non-HLA genes 112

Total genes 132

HLA class I genes A/B/C/E/F/G

HLA class I-like genes MICA/MICB

HLA class II genes DRA/DRB/DQ/DP/DM/DO

depending on the particular TLR ligand added to the cell cul-ture media [72].

One large-scale study found that the frequency ofautoimmune disorders in the families with autistic chil-dren was found to be higher than in control subjects,especially mothers of autistic children [73]. Another groupdemonstrated that autoimmune diseases were increasedsignificantly in families with ASD compared with those ofhealthy control subjects [74], suggesting a link between thedisorders. Croen et al. [75] showed that maternal psoriasisdiagnosed around the time of pregnancy is significantlyassociated with a subsequent diagnosis of autism in the child.Additionally, they showed a 2-fold increase in risk for a childhaving ASD if the mother was diagnosed and with asthma orallergies during pregnancy. An association between a familyhistory of type 1 diabetes mellitus (T1DM) and infantileautism as well as a significant association between maternalhistories of either rheumatoid arthritis (RA) or celiac diseaseand ASDs was noted by Atladottir et al. [76].

6. Autism HLA Genetics

HLA is the name for the major histocompatibility complex(MHC) in humans and HLA and MHC are often usedinterchangeably in the literature. The HLA region on chro-mosome 6p21 (about 4× 106 bp) is of major interest in basicresearch as well as medicine as genes/proteins in this regionare involved in many biological processes such as histocom-patibility, inflammation, ligands for immune cell receptors,and the complement cascade. The HLA region has 20 typicalHLA genes and 112 nontypical HLA genes (Table 2) that areinherited together as frozen blocks of DNA called ancestralor extended haplotypes. Complete DNA sequences havebeen published for 8 of the more common ancestral hap-lotypes in an effort to expedite basic and disease research[77]. It should be mentioned that smaller haplotypes canalso be constructed for genes that are linked. HLA genesalso play a role in reproduction, pregnancy maintenance,mate selection, and even kin recognition [78, 79] and havebeen associated with over 100 diseases/disorders includingautism. The proteins encoded by HLA genes are ligands,receptors, cytokines, signaling factors, heat shock proteins,transcription regulators, and so forth. Current research isincreasingly demonstrating a role for HLA proteins in neuralcell interactions, synaptic function, cerebral hemispheric

specialization, central nervous system (CNS) development[80–84], and even neurological disorders [85]. The genesof the HLA region are shown in Figure 1. Shiina et al.[86, 87] have published two excellent reviews on the HLAsuper-locus. Not only is there extraordinary complexity inthe HLA genes, there are extensive haplotype-related tran-scriptional differences [88]. It has been shown that geneticmechanisms outside of the non-antigen-binding HLA genesin the ancestral haplotype 8.1 (also referred to as COX) areassociated with susceptibility to many autoimmune diseases[89]. It is our proposition that HLA genes/proteins shouldbe more carefully examined due to increasing evidence ofautoimmune type associations in autism.

7. HLA Associations in ASD

It was suggested over 30 years ago by Stubbs and Magenis[90] that the HLA region might be important in autism.Warren et al. [91] first reported that the HLA ancestralhaplotype 44.1 (B44-SC30-DR4) was associated with autismwith a relative risk of 7.9. That result was confirmed ina separate case/control population [92]. Interestingly, theindividual components of AH 44.1 (A2-B44-SC30-DR4)include a deleted C4B gene and DRβ1∗0401, both of whichhave been shown independently to be significantly associatedwith ASD [93, 94]. Examination with different geneticmarkers than those used by Warren suggested that certainHLA haplotypes are associated with autism in Sardinian andItalian families [95, 96].

Warren et al. [97] reported that the shared epitope-binding pocket (DRβ1∗0401, ∗0404, and ∗0101) in the thirdhypervariable region of DRβ1 has a strong association withautism. A relative risk of 19.8 for autism was reported forsubjects with one of the two extended HLA haplotypes. Bothof these haplotypes have many allelic similarities especiallythe DRβ1∗0401.

AH 44.1 (HLA-A2, Cw5, B44, Bf∗S, C2∗C, C4A3,C4BQ0, DRβ1∗0401, DQB1∗0301).

AH 62.1 (HLA-A2, Cw3, B15, Bf∗S, C2∗C, C4A3, C4B3,DRβ1∗0401, DQB1∗0302).

The shared epitope has been associated with severalautoimmune diseases such as rheumatoid arthritis, psoriaticarthritis, and systemic lupus erythematosus [98]. Torres etal. [93] confirmed the association of the HLA-DR4 allele andalso found that the DR13 and DR14 alleles occurred less


Chr. 6

Telo

mer

eTe

lom

ere

Shor

t ar

mC

entr

omer

eLo

ng

arm

Cla

ss I

reg

ion

Cla

ss I

reg

ion C

lass

II

regi

on

Cla

ss I

II r

egio

n

Ext

ende

d cl

ass

II r

egio

n

αbl

ock

βbl

ock

κbl

ock

Telomeric side

Centromeric side

Figure 1: Gene map of the human leukocyte antigen (HLA) region. The major histocompatibility complex (MHC) gene map correspondsto the genomic coordinates of 29677984 (GABBR I) to 33485635 (KIFC1) in the human genome build 36.3 of the National Center forBiotechnology Information (NCBI) map viewer. The regions separated by arrows show the HLA subregions such as extended class I, classicalclass I, class III, classical class II, and extended class II regions from telomere (left and top side) to centromere (right and bottom side). White,gray, striped and black boxes show expressed genes, gene candidates, noncoding genes and pseudogenes, respectively. The location of thealpha, beta, and kappa blocks containing the cluster of duplicated HLA class I genes in the class I region are indicated. (Reprinted withpermission from the Journal of Human Genetics [87].)

often in subjects with autism, suggesting a possible protectivemechanism. Interestingly, the DR13 allele was inherited lessfrequently than expected from the mothers. Associationswith autism and the DR4 allele have since been confirmed in

three additional research groups. Lee et al. [99] demonstratedthat boys with autism and their mothers had a significantlyhigher frequency of DR4 than normal control subjects (oddsratios 4.20 and 5.54, resp.), suggesting that a maternal-fetal


immune interaction could be involved in autism. Johnsonet al. [100] reported significant transmission disequilibriumfor HLA-DR4 (odds ratio 4.67) from maternal grandparentsto mothers of children with autism which also suggests amaternal-fetal interaction for HLA-DR4. It has been recentlyshown in Han Chinese that the HLA-DRβ1 allele frequenciesincluding DR4 are different in subjects with autism versuscontrol subjects [101].

It was reported 20 years ago that subjects with autismhad a significant increase in the C4B null allele (C4Bgene deletion) compared to control subjects [91]. After thisobservation, it was also noted that subjects with autism hada significant deficiency in the plasma C4B protein [102].These initial findings of an increase in the deletion of theC4B gene was supported by examining a new population ofsubjects with autism [94]. Mostafa and Shehab [103] haverecently reported a significant increase in the deletion of theC4B gene in the Egyptian population. They also reported anincreased risk when there was a family history of autoim-mune diseases in the autism population. Descriptions ofseveral non-antigen-binding HLA genes that are associatedwith autoimmune diseases are discussed below and listedin Table 3.

8. Non-Antigen-Binding HLA Genes in Class I

Although MICA/MICB genes are structurally similar to clas-sical antigen-binding HLA class I genes, they encode proteinsthat interact with different T-cell receptors in response tostress as infection or neoplastic transformation. The twogenes are located on the centromeric end of the class I regionnear HLA-B at the border of the class III region (Figure 1)[87]. MICA/MICB proteins are ligands for NKG2D recep-tors on NK cell and gamma/delta T-cell receptors. Adap-tive immunity involves three lymphocyte populations (Bcells, alpha/beta T cells, and gamma/delta T cells). Gam-ma/delta T cells represent a small population of T cells thatpossess a T-cell receptor that is distinct from the typicalalpha/beta T cell. They are concentrated in the intestinal mu-cosa and appear to have a prominent role in recognizingsmall bacterial phosphoantigens and undefined antigenspresented by MICA/MICB proteins. Gamma/delta T cellshave potent cytotoxic activity and have been considered alink between innate and adaptive immunity. Polymorphismsin MICA/MICB genes have been associated with T1DM, AD,and SLE independent of DRβ1 alleles. There are several genesassociated with psoriasis (PSORSI locus). It is unclear howthe risk is spread among these genes as they are closely linked.Two genes RNF39 and TRIM39 in the class I region have beenassociated with Behcet’s disease.

9. Non-Antigen-Binding HLA Genes in Class III

Tumor necrosis factor-alpha (TNF-α) is a proinflammatory,multifunctional cytokine that plays important roles in cellphysiology. It is synthesized in numerous cells includingmacrophages, NK cells, T cells, mast cells, osteoblasts, gran-ulocytes, smooth muscle cells, fibroblasts, and keratinocytes.

In the CNS, TNF-α is made in microglia cells, astrocytes, andneurons [141]. In the normal state, TNF-α drives acute andchronic inflammatory responses that leads to the removal ofinjurious stimuli and the restoration of homeostasis. TNF-α is necessary for neural cell differentiation and neuronmaturation and there is evidence that it is critical in thenormal brain for proper synaptic function.

There is extensive research that implicates TNF-α as a keymediator in disease progression, inflammation, blood-brain-barrier deterioration, and even cell death [142]. Elevated lev-els are present in numerous neurological disorders includingmultiple sclerosis, Alzheimer’s disease, Parkinson’s disease,ischemia, traumatic brain injury, and as mentioned above,ASD [141]. It is unclear if TNF-α contributes to the diseasestate or the higher concentrations limit neuronal injury.There are three adjacent genes: lymphotoxin alpha and beta(LTA and LTB) and leukocyte-specific transcript-1 (LST1)in the TNF-α block. LTA and LTB are proinflammatorycytokines like TNF-α, and LST1 plays a role in inflammatoryand infectious diseases. It has been suggested that vaccine-induced immunity changes with certain haplotypes in thesegenes [143]. Two other genes important in immune functionare adjacent to the TNF-α block. Nuclear factor of kappa lightpolypeptide gene enhancer in B-cells inhibitor-like 1(NF-κBIL1) on the telomeric side of TNA-α is an inhibitor of NF-κB, an important pleiotropic immune system transcriptionfactor that is upregulated in ASD [144, 145]. In addition, NF-κBIL1 has been associated with several autoimmune diseases(Table 3). The natural cytotoxicity trigger receptor three(NCR3) gene on the centromeric side encodes proteins thatare ligands for activating receptors on NK cells that recognizetumor cells [146]. Typical HLA-B and -C proteins are alsoligands for KIR receptors on NK cells.

There are numerous heat shock proteins (HSPs), alsoreferred to as chaperones, that assist the folding of newlysynthesized proteins as well as those that have been unfoldedor denatured [147]. HSPs are also important in cell protectivefunctions and in inhibiting the apoptosis cascade. They arenamed by their molecular weights (HSP100, HSP90, HSP70,HSP60, and smaller HSPs). Although the proteins appearin normal cellular functions, they are induced to higherlevels in trauma, epilepsy, neurodegenerative diseases, andother injuries. An underlying feature among AD, Parkinson’sdisease, spinocerebellar ataxia, and other neurodegenerativediseases is the accumulation of misfolded proteins andHSPs are being studied because of their role in foldingand refolding proteins [147]. There is intense interest inHSPs as they appear to protect neurons [148] and one mustremember that postmitotic neurons are unable to dilutemisfolded or aggregated proteins through division.

There are three HSP70 genes (HSPA1L, HSPA1A, andHSPA1B) in the class III HLA region that are adjacent butseparate genes. HSP70 proteins have been demonstrated tostimulate IL-6 and TNF-α production, activate microglialcells, and stimulate phagocytosis [148]. HSP70 proteins arealso important in autoimmunity by enhancing antigen pre-sentation in both HLA class I and HLA class II systems [147,149]. Peptides that are associated with HSP70 at the time of


Table 3: Non-classical HLA genes associated with autoimmune diseases.

Gene HGNC gene number Autoimmune disease References

Extended class I OR2H2 8253 SLE [104]

Class I

RNF39 18064 Behcet’s disease [105]

TRIM39 10065 Behcet’s disease [105]

PSORS1 locus 9573 Systemic sclerosis, Psoriasis [106–111]

MICA 7090 T1DM, AD, SLE [112–114]

MICB 7091 SLE [104]

Class III

BAT1–BAT5 13917–21 Alzheimer’s, AIDS [115, 116]

NFKBIL1 7800 Sjogren’s syndrome, SLE, RA [117, 118]

TNF Block 11892Alzheimer’s, Psoriasis, Autoimmunehepatitis, Sarcoidosis

[115, 119–121]

AIF1 352 T1DM [122]

HSP70 genes 5232–4 MS [123]

Complement genes 1248, 1324, 1323 SLE, myasthenia gravis, T1DM [124–126]

SKIV2L 10898 SLE [127]

ATF6B (CREBL1) 2349 SLE [104]

NOTCH4 7884 Systemic sclerosis [128]

C6orf10 13922 SLE [104]

BTNL2 1142 Ulcerative colitis, Sarcoidosis [129, 130]

Class II

TAP2 44 Psoriasis [131]

PSMB8 9545 Hypersensitivity pneumonitis [132]

Psoriasis [133]

TAP1 43 Vitiligo [134]

PSMB9 9546 Psoriasis, Vitiligo [133, 134]

HLA-DM 4934, 4935Psoriasis, Antiphospholipidauto-antibodies, RA, SLE, T1DM

[131, 135–138]

HLA-DO 4936, 4937 common variable immunodeficiency [139]

Class II ExtendedHSD17B8 3554 SLE [104]

DAXX 2681 MS [140]

Abbreviations: (HGNC) HUGO Gene Nomenclature Committee (http://www.genenames.org/); (SLE) systemic lupus erythematosus; (AIDS) acquiredimmune deficiency syndrome; (T1DM) type 1 diabetes mellitus; (AD) Addison’s disease; (RA) rheumatoid arthritis; (MS) multiple sclerosis, PSORS1 psoriasislocus genes (CDSN, PSORS1C1, PSORS1C2, CCHCR1, POU5F1, PSORS1C3), TNF Block genes (LTA, TNF, LTB, LST1), HSP70 genes (HSPA1L, HSPA1A,HSPA1B), Complement genes (C2, C4B, C4A).

T-cell presentation have been shown to be more antigenicand therefore involved in autoimmunity [149, 150].

Another gene that is independently associated withautoimmune diseases outside of non-antigen-binding HLAalleles is allograft inflammatory factor-1 (AIF1). The SNPrs2269475 C > T in AIF1 has been associated with RA.The T allele was significantly higher in the RA patients andthere was no significant linkage disequilibria between theAIF1 SNP and DRβ1 alleles. Anticyclic citrullinated peptideantibodies commonly used to monitor RA were significantlyincreased in carriers with the T allele [151] but not the Callele.

10. Non-Antigen-Binding HLA Genes in Class II

Although HLA-DM and -DQ proteins have structures likeclassical antigen-binding HLA proteins, they work in thecytoplasm and not at the cell surface antigen-presenting

HLA proteins. DM stabilizes and edits the peptide repertoirepresented by DQ proteins by catalyzing CLIP release. Theassociations of DQ2 with T1DM and celiac disease have beenknown for several decades; however, the biochemistry behindthese associations has not been elucidated. It is now knownthat DQ2 is a poor substrate for DM and it has been proposedthat antigen presentation in the thymus and periphery can beaffected by impaired DQ-DM interactions so as to promoteautoimmune disease [152]. HLA-DO is another nonclassicalclass II HLA protein that is involved in the loading of peptidesto HLA-DR proteins by modulating the function of HLA-DM [153].

There are two genes in the class II region that encode pro-teins involved in the transport of antigen from the cytoplasmto the endoplasmic reticulum for binding to class I HLAproteins: transporter associated with antigen-processing 1and 2 (TAP1 and TAP2). Several research groups havereported associations of TAP2 genes with SLE independent ofclassical HLA alleles [154]. This interaction is very interesting


as it means that class II genes can affect class I antigenbinding.

There are two other genes in the extended class IIregion that have been associated with autoimmune dis-eases (Table 3). Hydroxysteroid 17-beta dehydrogenase 8(HSD17β8) is important in regulating the concentrations ofactive estrogens and androgens and high levels of estrogenare well known to be associated with SLE and other autoim-mune diseases. The second gene, death-associated protein 6(DAXX), is a very important protein that interacts with avariety of proteins in the nucleus and cytoplasm. Perhapsmost importantly, it interacts with the death receptor Fas.Engstrom et al. [155] published a paper describing decreasedexpression of Fas on CD4 + T lymphocytes but higher serumlevels of soluble Fas in ASD.

11. Summary

There is mounting evidence that the immune system playsa role in the pathogenesis of ASD in certain individuals.This evidence comes from several research areas includingan increase in proinflammatory cytokines in blood andbrain, autoantibodies to numerous antigens, and HLAassociations.

Autism HLA associations have been observed across theentire HLA region. For example, in the class I region HLA-A2 has been associated with autism by at least two researchgroups [156, 157]. In the class II region several researchershave reported autism associations with the DRβ1∗04 allele[93, 99, 100]. Strong associations also appear in the classIII region where the C4B null allele has been associatedwith autism with relative risks of 4.3 [94] and 4.6 [97],and an odds ratio of 6.3 [103]. The HLA-associated risk isthe highest for autism (19.8) when combining two ancestralhaplotypes (44.1 and 62.1). Both of these haplotypes haveHLA-A2 and DRβ1∗0401 as well as other genetic similarities;however, these two alleles cannot account for all of the 19.8risk. Compared to other genetic associations with autism, theHLA associations may be more important than realized, asthey have the highest genetically associated risk, that we areaware of for autism. For example the MET gene, one of themost studies genetic regions in autism, has a relative risk of2.27 [17].

It is our premise that some of the autism missinginheritance may be hidden in the HLA region, both inclassical HLA alleles and nonclassical HLA genes, as seen inschizophrenia [31]. For example, the HLA class III regioncontains clusters of genes such as TNF-α, HSP70, C4A/C4B,and NF-κBIL1 that are seminal in cellular function andare also associated with numerous autoimmune diseases(Table 3).

Acknowledgment

This report was supported in part by NIH Grant(RO1ES016669) and the Utah Autism Foundation (SaltLake City, Utah).

References

[1] L. Kanner, “Autistic disturbances of affective contact,” TheNervous Child, vol. 2, no. 2, pp. 217–250, 1943.

[2] F. R. Volkmar, M. State, and A. Klin, “Autism and autismspectrum disorders: diagnostic issues for the coming decade,”Journal of Child Psychology and Psychiatry and Allied Disci-plines, vol. 50, no. 1-2, pp. 108–115, 2009.

[3] Centers for Disease Control and Prevention (CDC), “Preva-lence of autism spectrum disorders-Autism and develop-mental disabilities monitoring network, United States, 2006,”CDC—Morbidity and Mortality Weekly Report, vol. 58, no.10, pp. 1–20, 2009.

[4] B. M. Altevogt, S. L. Hanson, and A. I. Leshner, “Autism andthe environment: challenges and opportunities for research,”Pediatrics, vol. 121, no. 6, pp. 1225–1229, 2008.

[5] S. Chess, “Autism in children with congenital rubella,”Journal of Autism and Childhood Schizophrenia, vol. 1, no. 1,pp. 33–47, 1971.

[6] S. Folstein and M. Rutter, “Infantile autism: a genetic studyof 21 twin pairs,” Journal of Child Psychology and Psychiatryand Allied Disciplines, vol. 18, no. 4, pp. 297–321, 1977.

[7] S. Steffenburg, C. Gillberg, L. Hellgren et al., “A twinstudy of autism in Denmark, Finland, Iceland, Norway andSweden,” Journal of Child Psychology and Psychiatry andAllied Disciplines, vol. 30, no. 3, pp. 405–416, 1989.

[8] A. Bailey, A. le Couteur, I. Gottesman et al., “Autism as astrongly genetic disorder: evidence from a British twin study,”Psychological Medicine, vol. 25, no. 1, pp. 63–77, 1995.

[9] P. Bolton, H. Macdonald, A. Pickles et al., “A case-controlfamily history study of autism,” Journal of Child Psychologyand Psychiatry and Allied Disciplines, vol. 35, no. 5, pp. 877–900, 1994.

[10] E. R. Ritvo, L. B. Jorde, A. Mason-Brothers et al., “TheUCLA-University of Utah epidemiologic survey of autism:recurrence risk estimates and genetic counseling,” AmericanJournal of Psychiatry, vol. 146, no. 8, pp. 1032–1036, 1989.

[11] S. Ozonoff, G. S. Young, A. Carter et al., “Recurrence riskfor autism spectrum disorders: a baby siblings researchconsortium study,” Pediatrics, vol. 128, no. 3, pp. e488–e495,2011.

[12] J. Hallmayer, S. Cleveland, A. Torres et al., “Genetic heritabil-ity and shared environmental factors among twin pairs withautism,” Archives of General Psychiatry, vol. 68, no. 11, pp.1095–1102, 2011.

[13] C. M. Freitag, “The genetics of autistic disorders and itsclinical relevance: a review of the literature,” MolecularPsychiatry, vol. 12, no. 1, pp. 2–22, 2007.

[14] M. Rutter, “Genetic studies of autism: from the 1970s intothe millennium,” Journal of Abnormal Child Psychology, vol.28, no. 1, pp. 3–14, 2000.

[15] M. W. State, “The genetics of child psychiatric disorders:focus on autism and tourette syndrome,” Neuron, vol. 68, no.2, pp. 254–269, 2010.

[16] D. B. Campbell, J. S. Sutcliffe, P. J. Ebert et al., “A geneticvariant that disrupts MET transcription is associated withautism,” Proceedings of the National Academy of Sciences of theUnited States of America, vol. 103, no. 45, pp. 16834–16839,2006.

[17] M. C. Judson, K. L. Eagleson, and P. Levitt, “A new synapticplayer leading to autism risk: met receptor tyrosine kinase,”Journal of Neurodevelopmental Disorders, vol. 3, no. 3, pp.282–292, 2011.


[18] J. T. Glessner, K. Wang, G. Cai et al., “Autism genome-wide copy number variation reveals ubiquitin and neuronalgenes,” Nature, vol. 459, no. 7246, pp. 569–573, 2009.

[19] M. S. Ching, Y. Shen, W. H. Tan et al., “Deletions of NRXN1(neurexin-1) predispose to a wide spectrum of developmen-tal disorders,” American Journal of Medical Genetics B, vol.153, no. 4, pp. 937–947, 2010.

[20] K. A. Strauss, E. G. Puffenberger, M. J. Huentelman et al.,“Recessive symptomatic focal epilepsy and mutant contactin-associated protein-like 2,” The New England Journal ofMedicine, vol. 354, no. 13, pp. 1370–1377, 2006.

[21] M. Alarcon, B. S. Abrahams, J. L. Stone et al., “Linkage, asso-ciation, and gene-expression analyses identify CNTNAP2 asan autism-susceptibility gene,” American Journal of HumanGenetics, vol. 82, no. 1, pp. 150–159, 2008.

[22] D. E. Arking, D. J. Cutler, C. W. Brune et al., “A commongenetic variant in the neurexin superfamily member CNT-NAP2 increases familial risk of autism,” American Journal ofHuman Genetics, vol. 82, no. 1, pp. 160–164, 2008.

[23] B. Bakkaloglu, B. J. O’Roak, A. Louvi et al., “Molecularcytogenetic analysis and resequencing of contactin associatedprotein-like 2 in autism spectrum disorders,” AmericanJournal of Human Genetics, vol. 82, no. 1, pp. 165–173, 2008.

[24] T. A. Manolio, F. S. Collins, N. J. Cox et al., “Finding themissing heritability of complex diseases,” Nature, vol. 461, no.7265, pp. 747–753, 2009.

[25] D. B. Goldstein, “Common genetic variation and humantraits,” The New England Journal of Medicine, vol. 360, no. 17,pp. 1696–1698, 2009.

[26] J. McClellan and M. C. King, “Genetic heterogeneity inhuman disease,” Cell, vol. 141, no. 2, pp. 210–217, 2010.

[27] J. P. Gregg, L. Lit, C. A. Baron et al., “Gene expression changesin children with autism,” Genomics, vol. 91, no. 1, pp. 22–29,2008.

[28] A. M. Enstrom, L. Lit, C. E. Onore et al., “Altered geneexpression and function of peripheral blood natural killercells in children with autism,” Brain, Behavior, and Immunity,vol. 23, no. 1, pp. 124–133, 2009.

[29] J. T. Morgan, G. Chana, C. A. Pardo et al., “Microglialactivation and increased microglial density observed in thedorsolateral prefrontal cortex in autism,” Biological Psychia-try, vol. 68, no. 4, pp. 368–376, 2010.

[30] I. Voineagu, X. Wang, P. Johnston et al., “Transcriptomicanalysis of autistic brain reveals convergent molecularpathology,” Nature, vol. 474, no. 7351, pp. 380–384, 2011.

[31] H. Stefansson, R. A. Ophoff, S. Steinberg et al., “Commonvariants conferring risk of schizophrenia,” Nature, vol. 460,no. 7256, pp. 744–747, 2009.

[32] J. Croonenberghs, E. Bosmans, D. Deboutte, G. Kenis, andM. Maes, “Activation of the inflammatory response system inautism,” Neuropsychobiology, vol. 45, no. 1, pp. 1–6, 2002.

[33] P. Ashwood, P. Krakowiak, I. Hertz-Picciotto, R. Hansen, I. N.Pessah, and J. van de Water, “Associations of impaired behav-iors with elevated plasma chemokines in autism spectrumdisorders,” Journal of Neuroimmunology, vol. 232, no. 1-2, pp.196–199, 2011.

[34] D. L. Vargas, C. Nascimbene, C. Krishnan, A. W. Zim-merman, and C. A. Pardo, “Neuroglial activation andneuroinflammation in the brain of patients with autism,”Annals of Neurology, vol. 57, no. 1, pp. 67–81, 2005.

[35] X. Li, A. Chauhan, A. M. Sheikh et al., “Elevated immuneresponse in the brain of autistic patients,” Journal of Neuroim-munology, vol. 207, no. 1-2, pp. 111–116, 2009.

[36] U. S. Naik, C. Gangadharan, K. Abbagani, B. Nagalla, N.Dasari, and S. K. Manna, “A study of nuclear transcriptionfactor-κ B in childhood autism,” PLoS ONE, vol. 6, no. 5, p.e19488, 2011.

[37] A. M. Young, E. Campbell, S. Lynch, J. Suckling, and S.J. Powis, “Aberrant NF-κB expression in autism spectrumcondition: a mechanism for neuroinflammation,” Frontiers inPsychiatry, vol. 2, no. 27, pp. 1–8, 2011.

[38] V. K. Singh, R. P. Warren, J. D. Odell, W. L. Warren, and P.Cole, “Antibodies to myelin basic protein in children withautistic behavior,” Brain, Behavior, and Immunity, vol. 7, no.1, pp. 97–103, 1993.

[39] A. Rosenspire, W. Yoo, S. Menard, and A. R. Torres,“Autism spectrum disorders are associated with an elevatedautoantibody response to tissue transglutaminase-2,” AutismResearch, vol. 4, no. 4, pp. 242–249, 2011.

[40] P. Goines, L. Haapanen, R. Boyce et al., “Autoantibodies tocerebellum in children with autism associate with behavior,”Brain, Behavior, and Immunity, vol. 25, no. 3, pp. 514–523,2011.

[41] S. Wills, C. C. Rossi, J. Bennett et al., “Further charac-terization of autoantibodies to GABAergic neurons in thecentral nervous system produced by a subset of children withautism,” Molecular Autism, vol. 2, p. 5, 2011.

[42] M. Gonzalez-Gronow, M. Cuchacovich, R. Francos et al.,“Antibodies against the voltage-dependent anion channel(VDAC) and its protective ligand hexokinase-I in childrenwith autism,” Journal of Neuroimmunology, vol. 227, no. 1-2,pp. 153–161, 2010.

[43] B. Zhang, A. Angelidou, K. D. Alysandratos et al., “Mito-chondrial DNA and anti-mitochondrial antibodies in serumof autistic children,” Journal of Neuroinflammation, vol. 7, p.80, 2010.

[44] G. A. Mostafa and N. Kitchener, “Serum anti-nuclearantibodies as a marker of autoimmunity in Egyptian autisticchildren,” Pediatric Neurology, vol. 40, no. 2, pp. 107–112,2009.

[45] M. Cabanlit, S. Wills, P. Goines, P. Ashwood, and J. vande Water, “Brain-specific autoantibodies in the plasma ofsubjects with autistic spectrum disorder,” Annals of the NewYork Academy of Sciences, vol. 1107, pp. 92–103, 2007.


[47] A. M. Connolly, M. Chez, E. M. Streif et al., “Brain-derivedneurotrophic factor and autoantibodies to neural antigens insera of children with autistic spectrum disorders, Landau-Kleffner syndrome, and epilepsy,” Biological Psychiatry, vol.59, no. 4, pp. 354–363, 2006.

[48] M. Evers, C. Cunningham-Rundles, and E. Hollander, “Heatshock protein 90 antibodies in autism,” Molecular Psychiatry,vol. 7, supplement 2, pp. S26–S28, 2002.

[49] C. M. Brickman and Y. Shoenfeld, “The mosaic of autoim-munity,” Scandinavian Journal of Clinical and LaboratoryInvestigation, vol. 61, no. 235, pp. 3–15, 2001.

[50] A. Vojdani, A. W. Campbell, E. Anyanwu, A. Kashanian,K. Bock, and E. Vojdani, “Antibodies to neuron-specificantigens in children with autism: possible cross-reaction withencephalitogenic proteins from milk, Chlamydia pneumo-niae and Streptococcus group A,” Journal of Neuroimmunol-ogy, vol. 129, no. 1-2, pp. 168–177, 2002.


[51] V. K. Singh, R. Warren, R. Averett, and M. Ghaziuddin,“Circulating autoantibodies to neuronal and glial filamentproteins in autism,” Pediatric Neurology, vol. 17, no. 1, pp.88–90, 1997.

[52] H. Jyonouchi, S. Sun, and H. Le, “Proinflammatory andregulatory cytokine production associated with innate andadaptive immune responses in children with autism spec-trum disorders and developmental regression,” Journal ofNeuroimmunology, vol. 120, no. 1-2, pp. 170–179, 2001.

[53] V. M. Bashina, I. A. Kozlova, and T. P. Kliushnik, “Anelevation in the level of autoantibodies to nerve-growthfactor in the blood serum of schizophrenic children,” ZhurnalNevrologii i Psikhiatrii Imeni S.S. Korsakova, vol. 97, no. 1, pp.47–51, 1997.

[54] A. M. Connolly, M. G. Chez, A. Pestronk, S. T. Arnold,S. Mehta, and R. K. Deul, “Serum autoantibodies to brainin Landau-Kleffner variant, autism, and other neurologicdisorders,” The Journal of Pediatrics, vol. 134, no. 5, pp. 607–613, 1999.

[55] R. D. Todd and R. D. Ciaranello, “Demonstration of inter-and intraspecies differences in serotonin binding sites byantibodies from an autistic child,” Proceedings of the NationalAcademy of Sciences of the United States of America, vol. 82,no. 2, pp. 612–616, 1985.

[56] J. Croonenberghs, A. Wauters, K. Devreese et al., “Increasedserum albumin, γ globulin, immunoglobulin IgG, and IgG2and IgG4 in autism,” Psychological Medicine, vol. 32, no. 8,pp. 1457–1463, 2002.

[57] H. S. Singer, C. M. Morris, C. D. Gause, P. K. Gillin, S.Crawford, and A. W. Zimmerman, “Antibodies against fetalbrain in sera of mothers with autistic children,” Journal ofNeuroimmunology, vol. 194, no. 1-2, pp. 165–172, 2008.

[58] L. A. Croen, D. Braunschweig, L. Haapanen et al., “Maternalmid-pregnancy autoantibodies to fetal brain protein: theearly markers for autism study,” Biological Psychiatry, vol. 64,no. 7, pp. 583–588, 2008.

[59] K. A. Hogquist, T. A. Baldwin, and S. C. Jameson, “Cen-tral tolerance: learning self-control in the thymus,” NatureReviews Immunology, vol. 5, no. 10, pp. 772–782, 2005.

[60] J. Sprent, “Central tolerance of T cells,” International Reviewsof Immunology, vol. 13, no. 2, pp. 95–105, 1996.

[61] A. Basten, R. Brink, P. Peake et al., “Self tolerance in the B-cellrepertoire,” Immunological Reviews, no. 122, pp. 5–19, 1991.

[62] S. Baumann, A. Krueger, S. Kirchhoff, and P. H. Kram-mer, “Regulation of T cell apoptosis during the immuneresponse,” Current Molecular Medicine, vol. 2, no. 3, pp. 257–272, 2002.

[63] Q. Leng and Z. Bentwich, “Beyond self and nonself: fuzzyrecognition of the immune system,” Scandinavian Journal ofImmunology, vol. 56, no. 3, pp. 224–232, 2002.

[64] E. Meffre and J. E. Salmon, “Autoantibody selection and pro-duction in early human life,” Journal of Clinical Investigation,vol. 117, no. 3, pp. 598–601, 2007.

[65] F. J. Quintana and H. L. Weiner, “Understanding natural andpathological autoimmunity,” Journal of Neuroimmunology,vol. 174, no. 1-2, pp. 1–2, 2006.

[66] R. P. Warren, N. C. Margaretten, N. C. Pace, and A. Foster,“Immune abnormalities in patients with autism,” Journal ofAutism and Developmental Disorders, vol. 16, no. 2, pp. 189–197, 1986.

[67] D. R. Denney, B. W. Frei, and G. R. Gaffney, “Lymphocytesubsets and interleukin-2 receptors in autistic children,”

Journal of Autism and Developmental Disorders, vol. 26, no.1, pp. 87–97, 1996.

[68] P. Ashwood, P. Krakowiak, I. Hertz-Picciotto, R. Hansen, I.N. Pessah, and J. van de Water, “Altered T cell responses inchildren with autism,” Brain, Behavior, and Immunity, vol.25, no. 5, pp. 840–849, 2011.

[69] R. P. Warren, A. Foster, and N. C. Margaretten, “Reducednatural killer cell activity in autism,” Journal of the AmericanAcademy of Child and Adolescent Psychiatry, vol. 26, no. 3, pp.333–335, 1987.

[70] A. Vojdani, E. Mumper, D. Granpeesheh et al., “Low naturalkiller cell cytotoxic activity in autism: the role of glutathione,IL-2 and IL-15,” Journal of Neuroimmunology, vol. 205, no.1-2, pp. 148–154, 2008.

[71] T. L. Sweeten, D. J. Posey, and C. J. McDougle, “Highblood monocyte counts and neopterin levels in children withautistic disorder,” American Journal of Psychiatry, vol. 160, no.9, pp. 1691–1693, 2003.

[72] A. M. Enstrom, C. E. Onore, J. A. van de Water, and P.Ashwood, “Differential monocyte responses to TLR ligandsin children with autism spectrum disorders,” Brain, Behavior,and Immunity, vol. 24, no. 1, pp. 64–71, 2010.

[73] A. M. Comi, A. W. Zimmerman, V. H. Frye, P. A. Law, andJ. N. Peeden, “Familial clustering of autoimmune disordersand evaluation of medical risk factors in autism,” Journal ofChild Neurology, vol. 14, no. 6, pp. 388–394, 1999.

[74] T. L. Sweeten, S. L. Bowyer, D. J. Posey, G. M. Halberstadt,and C. J. McDougle, “Increased prevalence of familialautoimmunity in probands with pervasive developmentaldisorders,” Pediatrics, vol. 112, no. 5, pp. e420–e424, 2003.

[75] L. A. Croen, J. K. Grether, C. K. Yoshida, R. Odouli, and J. V.van de Water, “Maternal autoimmune diseases, asthma andallergies, and childhood autism spectrum disorders: a case-control study,” Archives of Pediatrics and Adolescent Medicine,vol. 159, no. 2, pp. 151–157, 2005.

[76] H. O. Atladottir, M. G. Pedersen, P. Thorsen et al., “Associ-ation of family history of autoimmune diseases and autismspectrum disorders,” Pediatrics, vol. 124, no. 2, pp. 687–694,2009.

[77] R. Horton, R. Gibson, P. Coggill et al., “Variation analysisand gene annotation of eight MHC haplotypes: the MHChaplotype project,” Immunogenetics, vol. 60, no. 1, pp. 1–18,2008.

[78] L. A. Knapp, “The ABCs of MHC,” Evolutionary Anthropol-ogy, vol. 14, no. 1, pp. 28–37, 2005.

[79] A. Ziegler, H. Kentenich, and B. Uchanska-Ziegler, “Femalechoice and the MHC,” Trends in Immunology, vol. 26, no. 9,pp. 496–502, 2005.

[80] B. G. Xiao and H. Link, “Immune regulation within thecentral nervous system,” Journal of the Neurological Sciences,vol. 157, no. 1, pp. 1–12, 1998.

[81] G. S. Huh, L. M. Boulanger, H. Du, P. A. Riquelme, T. M.Brotz, and C. J. Shatz, “Functional requirement for class IMHC in CNS development and plasticity,” Science, vol. 290,no. 5499, pp. 2155–2159, 2000.

[82] L. M. Boulanger and C. J. Shatz, “Immune signalling inneural development, synaptic plasticity and disease,” NatureReviews Neuroscience, vol. 5, no. 7, pp. 521–531, 2004.

[83] S. Cullheim and S. Thams, “The microglial networks of thebrain and their role in neuronal network plasticity afterlesion,” Brain Research Reviews, vol. 55, no. 1, pp. 89–96,2007.


[84] M. Ohtsuka, H. Inoko, J. K. Kulski, and S. Yoshimura,“Major histocompatibility complex (Mhc) class Ib geneduplications, organization and expression patterns in mousestrain C57BL/6,” BMC Genomics, vol. 9, p. 178, 2008.

[85] S. L. Bailey, P. A. Carpentier, E. J. McMahon, W. S. Begolka,and S. D. Miller, “Innate and adaptive immune responses ofthe central nervous system,” Critical Reviews in Immunology,vol. 26, no. 2, pp. 149–188, 2006.

[86] T. Shiina, H. Inoko, and J. K. Kulski, “An update of the HLAgenomic region, locus information and disease associations,”Tissue Antigens, vol. 64, no. 6, pp. 631–649, 2004.

[87] T. Shiina, K. Hosomichi, H. Inoko, and J. K. Kulski, “TheHLA genomic loci map: expression, interaction, diversity anddisease,” Journal of Human Genetics, vol. 54, no. 1, pp. 15–39,2009.

[88] C. Vandiedonck, M. S. Taylor, H. E. Lockstone et al.,“Pervasive haplotypic variation in the spliceo-transcriptomeof the human major histocompatibility complex,” GenomeResearch, vol. 21, no. 7, pp. 1042–1054, 2011.

[89] G. Candore, I. C. Campagna, I. Cuppari, D. di Carlo, C.Mineo, and C. Caruso, “Genetic control of immune responsein carriers of the 8.1 ancestral haplotype,” Annals of the NewYork Academy of Sciences, vol. 1110, pp. 151–158, 2007.

[90] E. G. Stubbs and R. E. Magenis, “HLA and autism,” Journalof Autism and Developmental Disorders, vol. 10, no. 1, pp. 15–19, 1980.

[91] R. P. Warren, V. K. Singh, P. Cole et al., “Increased frequencyof the null allele at the complement C4b locus in autism,”Clinical and Experimental Immunology, vol. 83, no. 3, pp.438–440, 1991.

[92] W. W. Daniels, R. P. Warren, J. D. Odell et al., “Increasedfrequency of the extended or ancestral haplotype B44-SC30-DR4 in autism,” Neuropsychobiology, vol. 32, no. 3, pp. 120–123, 1995.

[93] A. R. Torres, A. Maciulis, E. G. Stubbs, A. Cutler, andD. Odell, “The transmission disequilibrium test suggeststhat HLA-DR4 and DR13 are linked to autism spectrumdisorder,” Human Immunology, vol. 63, no. 4, pp. 311–316,2002.

[94] D. Odell, A. Maciulis, A. Cutler et al., “Confirmation ofthe association of the C4B null allelle in autism,” HumanImmunology, vol. 66, no. 2, pp. 140–145, 2005.

[95] F. R. Guerini, E. Bolognesi, S. Manca et al., “Family-based transmission analysis of HLA genetic markers inSardinian children with autistic spectrum disorders,” HumanImmunology, vol. 70, no. 3, pp. 184–190, 2009.

[96] F. R. Guerini, E. Bolognesi, M. Chiappedi et al., “HLApolymorphisms in Italian children with autism spectrumdisorders: results of a family based linkage study,” Journal ofNeuroimmunology, vol. 230, no. 1-2, pp. 135–142, 2011.

[97] R. P. Warren, J. D. Odell, W. L. Warren et al., “Strongassociation of the third hypervariable region of HLA-DR β1with autism,” Journal of Neuroimmunology, vol. 67, no. 2, pp.97–102, 1996.

[98] D. E. de Almeida, S. Ling, and J. Holoshitz, “New insightsinto the functional role of the rheumatoid arthritis sharedepitope,” FEBS Letters, vol. 585, no. 23, pp. 3619–3626, 2011.

[99] L. C. Lee, A. A. Zachary, M. S. Leffell et al., “HLA-DR4 infamilies with autism,” Pediatric Neurology, vol. 35, no. 5, pp.303–307, 2006.

[100] W. G. Johnson, S. Buyske, A. E. Mars et al., “HLA-DR4 as arisk allele for autism acting in mothers of probands possibly

during pregnancy,” Archives of Pediatrics and AdolescentMedicine, vol. 163, no. 6, pp. 542–546, 2009.

[101] Y. L. Chien, Y. Y. Wu, C. H. Chen et al., “Association of HLA-DRB1 alleles and neuropsychological function in autism,”Psychiatric Genetics, vol. 22, no. 1, pp. 46–49, 2012.

[102] R. P. Warren, R. A. Burger, D. Odell, A. R. Torres, and W.L. Warren, “Decreased plasma concentrations of the C4Bcomplement protein in autism,” Archives of Pediatrics andAdolescent Medicine, vol. 148, no. 2, pp. 180–183, 1994.

[103] G. A. Mostafa and A. A. Shehab, “The link of C4B nullallele to autism and to a family history of autoimmunity inEgyptian autistic children,” Journal of Neuroimmunology, vol.223, no. 1-2, pp. 115–119, 2010.

[104] L. F. Barcellos, S. L. May, P. P. Ramsay et al., “High-densitySNP screening of the major histocompatibility complex insystemic lupus erythematosus demonstrates strong evidencefor independent susceptibility regions,” PLoS Genetics, vol. 5,no. 10, Article ID e1000696, 2009.

[105] R. Kurata, H. Nakaoka, A. Tajima et al., “TRIM39 and RNF39are associated with Behcet’s disease independently of HLA-B∗51 and -A∗26,” Biochemical and Biophysical ResearchCommunications, vol. 401, no. 4, pp. 533–537, 2010.

[106] Y. Allanore, M. Saad, P. Dieude et al., “Genome-wide scanidentifies TNIP1, PSORS1C1, and RHOB as novel risk locifor systemic sclerosis,” PLoS Genetics, vol. 7, no. 7, Article IDe1002091, 2011.

[107] S. Gonzalez, J. Martinez-Borra, J. S. del Rıo et al., “TheOTF3 gene polymorphism confers susceptibility to psoriasisindependent of the association of HLA-Cw∗0602,” Journal ofInvestigative Dermatology, vol. 115, no. 5, pp. 824–828, 2000.

[108] K. Asumalahti, C. Veal, T. Laitinen et al., “Coding haplotypeanalysis supports HCR as the putative susceptibility genefor psoriasis at the MHC PSORS1 locus,” Human MolecularGenetics, vol. 11, no. 5, pp. 589–597, 2002.

[109] S. J. Holm, L. M. Carlen, L. Mallbris, M. Stahle-Backdahl,and K. P. O’Brien, “Polymorphisms in the SEEK1 and SPR1genes on 6p21.3 associate with psoriasis in the Swedishpopulation,” Experimental Dermatology, vol. 12, no. 4, pp.435–444, 2003.

[110] S. J. Holm, F. Sanchez, L. M. Carlen, L. Mallbris, M. Stahle,and K. P. O’Brien, “HLA-Cw∗0602 associates more stronglyto psoriasis in the Swedish population than variants of thenovel 6p21.3 gene PSORS1C3,” Acta Dermato-Venereologica,vol. 85, no. 1, pp. 2–8, 2005.

[111] S. Orru, E. Giuressi, C. Carcassi, M. Casula, and L.Contu, “Mapping of the major psoriasis-susceptibility locus(PSORS1) in a 70-Kb interval around the corneodesmosingene (CDSN),” American Journal of Human Genetics, vol. 76,no. 1, pp. 164–171, 2005.

[112] G. Gambelunghe, A. Brozzetti, M. Ghaderi, P. Candeloro,C. Tortoioli, and A. Falorni, “MICA gene polymorphism inthe pathogenesis of type 1 diabetes,” Annals of the New YorkAcademy of Sciences, vol. 1110, pp. 92–98, 2007.

[113] A. Falorni, A. Brozzetti, D. L. Torre, C. Tortoioli, and G.Gambelunghe, “Association of genetic polymorphisms andautoimmune Addison’s disease,” Expert Review of ClinicalImmunology, vol. 4, no. 4, pp. 441–456, 2008.

[114] G. Gambelunghe, R. Gerli, E. B. Bocci et al., “Contributionof MHC class I chain-related A (MICA) gene polymorphismto genetic susceptibility for systemic lupus erythematosus,”Rheumatology, vol. 44, no. 3, pp. 287–292, 2005.

[115] A. Gnjec, K. L. D’Costa, S. M. Laws et al., “Association ofalleles carried at TNFA -850 and BAT1 -22 with Alzheimer’sdisease,” Journal of Neuroinflammation, vol. 5, p. 36, 2008.


[116] S. Limou, S. le Clerc, C. Coulonges et al., “Genomewideassociation study of an AIDS-nonprogression cohort empha-sizes the role played by HLA genes (ANRS genomewideassociation study 02),” Journal of Infectious Diseases, vol. 199,no. 3, pp. 419–426, 2009.

[117] J. Castiblanco and J. M. Anaya, “The IκBL gene polymor-phism influences risk of acquiring systemic lupus erythe-matosus and Sjogren’s syndrome,” Human Immunology, vol.69, no. 1, pp. 45–51, 2008.

[118] G. Tamiya, M. Shinya, T. Imanishi et al., “Whole genomeassociation study of rheumatoid arthritis using 27039microsatellites,” Human Molecular Genetics, vol. 14, no. 16,pp. 2305–2321, 2005.

[119] K. Reich, U. Huffmeier, I. R. Konig et al., “TNF polymor-phisms in psoriasis: association of psoriatic arthritis withthe promoter polymorphism TNF∗-857 independent of thePSORS1 risk allele,” Arthritis and Rheumatism, vol. 56, no. 6,pp. 2056–2064, 2007.

[120] L. C. Oliveira, G. Porta, M. L. C. Marin, P. L. Bittencourt, J.Kalil, and A. C. Goldberg, “Autoimmune hepatitis, HLA andextended haplotypes,” Autoimmunity Reviews, vol. 10, no. 4,pp. 189–193, 2011.

[121] R. Kieszko, P. Krawczyk, S. Chocholska, A. Dmoszynska, andJ. Milanowski, “TNF-α and TNF-β gene polymorphisms inPolish patients with sarcoidosis. Connection with the suscep-tibility and prognosis,” Sarcoidosis Vasculitis and Diffuse LungDiseases, vol. 27, no. 2, pp. 131–137, 2010.

[122] M. C. Eike, M. Olsson, D. E. Undlien et al., “Genetic variantsof the HLA-A, HLA-B and AIF1 loci show independentassociations with type 1 diabetes in Norwegian families,”Genes and Immunity, vol. 10, no. 2, pp. 141–150, 2009.

[123] H. Cwiklinska, M. P. Mycko, B. Szymanska, M. Matysiak, andK. W. Selmaj, “Aberrant stress-induced Hsp70 expression inimmune cells in multiple sclerosis,” Journal of NeuroscienceResearch, vol. 88, no. 14, pp. 3102–3110, 2010.

[124] M. C. Pickering and M. J. Walport, “Links between com-plement abnormalities and systemic lupus erythematosus,”Rheumatology, vol. 39, no. 2, pp. 133–141, 2000.

[125] D. Franciotta, M. Cuccia, E. Dondi, G. Piccolo, and V. Cosi,“Polymorphic markers in MHC class II/III region: a studyon Italian patients with myasthenia gravis,” Journal of theNeurological Sciences, vol. 190, no. 1-2, pp. 11–16, 2001.

[126] F. Jenhani, R. Bardi, Y. Gorgi, K. Ayed, and M. Jeddi, “C4polymorphism in multiplex families with insulin dependentdiabetes in the Tunisian population: standard C4 typingmethods and RFLP analysis,” Journal of Autoimmunity, vol.5, no. 2, pp. 149–160, 1992.

[127] M. M. Fernando, C. R. Stevens, P. C. Sabeti et al., “Identifi-cation of two independent risk factors for lupus within theMHC in United Kingdom families,” PLoS Genetics, vol. 3, no.11, p. e192, 2007.

[128] O. Gorlova, J. E. Martin, B. Rueda et al., “Identification ofnovel genetic markers associated with clinical phenotypesof systemic sclerosis through a genome-wide associationstrategy,” PLoS Genetics, vol. 7, no. 7, Article ID e1002178,2011.

[129] S. Pathan, R. E. Gowdy, R. Cooney et al., “Confirmationof the novel association at the BTNL2 locus with ulcerativecolitis,” Tissue Antigens, vol. 74, no. 4, pp. 322–329, 2009.

[130] N. Milman, C. B. Svendsen, F. C. Nielsen, and T. V. O.Hansen, “The BTNL2 A allele variant is frequent in Danish

patients with sarcoidosis,” Clinical Respiratory Journal, vol. 5,no. 2, pp. 105–111, 2011.

[131] C. W. Pyo, S. S. Hur, Y. K. Kim, T. Y. Kim, and T. G. Kim,“Association of TAP and HLA-DM genes with psoriasis inKoreans,” Journal of Investigative Dermatology, vol. 120, no.4, pp. 616–622, 2003.

[132] A. Camarena, A. Aquino-Galvez, R. Falfan-Valencia et al.,“PSMB8 (LMP7) but not PSMB9 (LMP2) gene polymor-phisms are associated to pigeon breeder’s hypersensitivitypneumonitis,” Respiratory Medicine, vol. 104, no. 6, pp. 889–894, 2010.

[133] U. Kramer, T. Illig, T. Grune, J. Krutmann, and C. Esser,“Strong associations of psoriasis with antigen processingLMP and transport genes TAP differ by gender and pheno-type,” Genes and Immunity, vol. 8, no. 6, pp. 513–517, 2007.

[134] C. B. Casp, J. X. She, and W. T. McCormack, “Genesof the LMP/TAP cluster are associated with the humanautoimmune disease vitiligo,” Genes and Immunity, vol. 4, no.7, pp. 492–499, 2003.

[135] M. L. Sanchez, K. Katsumata, T. Atsumi et al., “Association ofHLA-DM polymorphism with the production of antiphos-pholipid antibodies,” Annals of the Rheumatic Diseases, vol.63, no. 12, pp. 1645–1648, 2004.

[136] J. Morel, F. Roch-Bras, N. Molinari, J. Sany, J. F. Eliaou, andB. Combe, “HLA-DMA∗0103 and HLA-DMB∗0104 alleles asnovel prognostic factors in rheumatoid arthritis,” Annals ofthe Rheumatic Diseases, vol. 63, no. 12, pp. 1581–1586, 2004.

[137] J. Morel, C. S. Cda, O. Avinens, J. Sany, B. Combe, and J.F. Eliaou, “Polymorphism of HLA-DMA and DMB allelesin patients with systemic lupus erythematosus,” Journal ofRheumatology, vol. 30, no. 7, pp. 1485–1490, 2003.

[138] Y. M. Sang, C. Yan, C. Zhu, G. C. Ni, and Y. M. Hu,“Association of human leukocyte antigen non-classical geneswith type 1 diabetes,” Chinese Journal of Pediatrics, vol. 41,no. 4, pp. 260–263, 2003.

[139] X. Chen and P. E. Jensen, “MHC class II antigen presentationand immunological abnormalities due to deficiency of MHCclass II and its associated genes,” Experimental and MolecularPathology, vol. 85, no. 1, pp. 40–44, 2008.

[140] J. I. Satoh, M. Nakanishi, F. Koike et al., “Microarray analysisidentifies an aberrant expression of apoptosis and DNAdamage-regulatory genes in multiple sclerosis,” Neurobiologyof Disease, vol. 18, no. 3, pp. 537–550, 2005.

[141] S. L. Montgomery and W. J. Bowers, “Tumor necrosis factor-α and the roles it plays in homeostatic and degenerativeprocesses within the central nervous system,” Journal ofNeuroimmune Pharmacology. In press.

[142] M. K. McCoy and M. G. Tansey, “TNF signaling inhibitionin the CNS: implications for normal brain function andneurodegenerative disease,” Journal of Neuroinflammation,vol. 5, article 45, 2008.

[143] I. G. Ovsyannikova, R. A. Vierkant, V. S. Pankratz, R. M.Jacobson, and G. A. Poland, “Extended LTA, TNF, LST1and hla gene haplotypes and their association with rubellavaccine-induced immunity,” PLoS ONE, vol. 5, no. 7, ArticleID e11806, 2010.

[144] U. S. Naik, C. Gangadharan, K. Abbagani, B. Nagalla, N.Dasari, and S. K. Manna, “A study of nuclear transcriptionfactor-κ B in childhood autism,” PLoS ONE, vol. 6, no. 5,Article ID e19488, 2011.

[145] A. M. Young, E. Campbell, S. Lynch, J. Suckling, and S.J. Powis, “Aberrant NF-κB expression in autism spectrumcondition: a mechanism for neuroinflammation,” FrontPsychiatry, vol. 2, no. 27, pp. 1–8, 2011.


[146] C. E. Foster, M. Colonna, and P. D. Sun, “Crystal structure ofthe human natural killer (NK) cell activating receptor NKp46reveals structural relationship to other leukocyte receptorcomplex immunoreceptors,” Journal of Biological Chemistry,vol. 278, no. 46, pp. 46081–46086, 2003.

[147] G. Turturici, G. Sconzo, and F. Geraci, “Hsp70 and itsmolecular role in nervous system diseases,” BiochemistryResearch International, vol. 2011, Article ID 618127, 18 pages,2011.

[148] J. I. Kakimura, Y. Kitamura, K. Takata et al., “Microglialactivation and amyloid-β clearance induced by exogenousheat-shock proteins,” The FASEB Journal, vol. 16, no. 6, pp.601–603, 2002.

[149] M. P. Mycko, H. Cwiklinska, A. Walczak, C. Libert, C.S. Raine, and K. W. Selmaj, “A heat shock protein gene(Hsp70.1) is critically involved in the generation of theimmune response to myelin antigen,” European Journal ofImmunology, vol. 38, no. 7, pp. 1999–2013, 2008.

[150] D. G. Millar, K. M. Garza, B. Odermatt et al., “Hsp70promotes antigen-presenting cell function and converts T-cell tolerance to autoimmunity in vivo,” Nature Medicine, vol.9, no. 12, pp. 1469–1476, 2003.

[151] A. Pawlik, M. Kurzawski, T. Szczepanik et al., “Associationof allograft inflammatory factor-1 gene polymorphism withrheumatoid arthritis,” Tissue Antigens, vol. 72, no. 2, pp. 171–175, 2008.

[152] T. Hou, H. Macmillan, Z. Chen et al., “An insertion mutant inDQA1∗0501 restore susceptibility to HLA-DM: implicationsfor disease associations,” Journal of Immunology, vol. 187, no.5, pp. 2442–2452, 2011.

[153] F. Deshaies, D. A. Diallo, J. S. Fortin et al., “Evidence for ahuman leucocyte antigen-DM-induced structural change inhuman leucocyte antigen-DOβ,” Immunology, vol. 127, no. 3,pp. 408–417, 2009.

[154] P. S. Ramos, C. D. Langefeld, L. A. Bera, P. M. Gaffney, J.A. Noble, and K. L. Moser, “Variation in the ATP-bindingcassette transporter 2 gene is a separate risk factor forsystemic lupus erythematosus within the MHC,” Genes andImmunity, vol. 10, no. 4, pp. 350–355, 2009.

[155] H. A. Engstrom, S. Ohlson, E. G. Stubbs et al., “Decreasedexpression of CD95 (FAS/APO-1) on CD4+ T-lymphocytesfrom participants with autism,” Journal of Developmental andPhysical Disabilities, vol. 15, no. 2, pp. 155–163, 2003.

[156] P. Ferrante, M. Saresella, F. R. Guerini, M. Marzorati, M.C. Musetti, and A. G. Cazzullo, “Significant association ofHLA A2-DR11 with CD4 naive decrease in autistic children,”Biomedicine and Pharmacotherapy, vol. 57, no. 8, pp. 372–374, 2003.

[157] A. R. Torres, T. L. Sweeten, A. Cutler et al., “The associationand linkage of the HLA-A2 class I allele with autism,” HumanImmunology, vol. 67, no. 4-5, pp. 346–351, 2006.


Research Article

Intracellular and Extracellular Redox Status and Free RadicalGeneration in Primary Immune Cells from Children with Autism

Shannon Rose, Stepan Melnyk, Timothy A. Trusty, Oleksandra Pavliv,Lisa Seidel, Jingyun Li, Todd Nick, and S. Jill James

Department of Pediatrics, Arkansas Children’s Hospital Research Institute, University of Arkansas for Medical Sciences,Little Rock, AR 72202, USA

Correspondence should be addressed to S. Jill James, [email protected]

Received 30 July 2011; Revised 12 August 2011; Accepted 12 September 2011


Copyright © 2012 Shannon Rose et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

The modulation of the redox microenvironment is an important regulator of immune cell activation and proliferation. Toinvestigate immune cell redox status in autism we quantified the intracellular glutathione redox couple (GSH/GSSG) in restingperipheral blood mononuclear cells (PBMCs), activated monocytes and CD4 T cells and the extracellular cysteine/cystine redoxcouple in the plasma from 43 children with autism and 41 age-matched control children. Resting PBMCs and activated monocytesfrom children with autism exhibited significantly higher oxidized glutathione (GSSG) and percent oxidized glutathione equivalentsand decreased glutathione redox status (GSH/GSSG). In activated CD4 T cells from children with autism, the percent oxidizedglutathione equivalents were similarly increased, and GSH and GSH/GSSG were decreased. In the plasma, both glutathioneand cysteine redox ratios were decreased in autistic compared to control children. Consistent with decreased intracellular andextracellular redox status, generation of free radicals was significantly elevated in lymphocytes from the autistic children. Thesedata indicate primary immune cells from autistic children have a more oxidized intracellular and extracellular microenvironmentand a deficit in glutathione-mediated redox/antioxidant capacity compared to control children. These results suggest that the lossof glutathione redox homeostasis and chronic oxidative stress may contribute to immune dysregulation in autism.

1. Introduction

Autism is a behaviorally defined neurodevelopmental disor-der that usually presents in early childhood and is charac-terized by significant impairments in social interaction andcommunication and by abnormal repetitive hyper-focusedbehaviors. The prevalence of autism spectrum disorders hasincreased more than 10-fold in the last two decades, nowaffecting one in 110 US children, yet the etiology of thesedisorders remains elusive [1]. Glutathione depletion and oxi-dative stress have been implicated in the pathology of numer-ous neurobehavioral disorders including schizophrenia [2],bipolar disorder [3], and Alzheimer’s disease [4]. Accumu-lating evidence suggests that redox imbalance and oxidativestress may also contribute to autism pathophysiology. Mul-tiple biomarkers of oxidative stress have been identified inblood samples from children with autism [5–12]. Our grouphas reported a decrease in concentrations of glutathione

(GSH) and several of its metabolic precursors, an increasein oxidized glutathione disulfide (GSSG), and a decrease inglutathione redox ratio (GSH/GSSG) in case-control evalua-tions of plasma and lymphoblastoid cell lines derived fromchildren with autism [13–16]. Recently, several interactivepolymorphisms in enzymes regulating glutathione synthesiswere found to be more prevalent in children with autismsuggesting that the glutathione deficit and predisposition tooxidative stress may be genetically based in some children[17].

Oxidative stress occurs when cellular antioxidant defensemechanisms fail to counterbalance endogenous ROS produc-tion and/or exogenous prooxidant environmental exposures.Glutathione (γ-L-glutamyl-L-cysteinylglycine) is a tripeptidethat functions as the major intracellular antioxidant andredox buffer against macromolecular oxidative damage. Theglutathione thiol/disulfide redox couple (GSH/GSSG) is thepredominant mechanism for maintaining the intracellular


microenvironment in a highly reduced state that is essentialfor antioxidant/detoxification capacity, redox enzyme regula-tion, cell cycle progression, and transcription of antioxidantresponse elements (ARE) [18–23]. Subtle variation in therelative concentrations of reduced and oxidized glutathioneprovides a dynamic redox signaling mechanism that regu-lates these vital cellular processes [24–27]. For example, inboth CNS precursor cells and naıve immune cells, intracel-lular glutathione redox status is the primary determinantmodulating the cellular decision to undergo cell cycle arrest,differentiation, or proliferation [27]. A reducing intracellularenvironment is required for proliferation, while a moreoxidized microenvironment favors cell cycle arrest and dif-ferentiation. A chronic deficit in the GSH/GSSG redox ratiois considered to be a reliable indicator of oxidative stress andincreased vulnerability to oxidative damage from prooxidantenvironmental exposures [28, 29].

In the extracellular plasma compartment, the cysteine/cystine (thiol/disulfide) redox couple independently providesthe ambient redox environment for circulating immune cells.The ambient extracellular cysteine/cystine redox potentialhas been shown to be more oxidized than the intracellularGSH/GSSG redox potential and is independently regulated[30]. Dynamic shifts in the plasma cysteine/cystine redoxpotential alter the redox status of cysteine moieties in cellsurface proteins to induce conformational changes in proteinstructure that can reversibly alter function [31, 32]. For ex-ample, under oxidizing extracellular conditions, redox-sensi-tive cysteine residues in the catalytic core of protein tyrosinephosphatases become oxidized and reversibly inactivateenzyme activity depending on the ambient cysteine/cystineredox potential [31, 33, 34]. Extracellular cysteine/cystineredox status is emerging as an important new signal trans-duction mechanism that can induce posttranslational alter-ations in downstream redox-sensitive proteins including avariety of enzymes, transcription factors, receptors, adhesionmolecules, and membrane signaling proteins resulting in thedynamic modulation of their activity and function [32, 35,36].

Recent studies have revealed numerous immunologicabnormalities among children with autism including alter-ations in immune cell proportions [37–40] and shifts inhelper T-cell subpopulations after mitogenic stimulation[41, 42]. Peripheral blood mononuclear cells (PBMCs) fromindividuals with autism have been shown to produce higherlevels of proinflammatory cytokines and abnormal levels ofregulatory cytokines compared to control PBMCs at baselineand upon mitogenic stimulation [43–46]. Taken together,the immunological studies suggest a role for a dysregulatedimmune system in autism that potentially could be related toa deficit in glutathione-mediated antioxidant capacity and anoxidized microenvironment in immune cells. To investigatethis possibility, we examined whether primary immune cells(PBMCs) from children with autism exhibit decreased in-tracellular glutathione redox capacity compared to PBMCsfrom age-matched control children and whether a moreoxidized intracellular and extracellular microenvironmentis associated with increased production of oxidizing intra-cellular free radicals. Because immune cells from children

with autism have been shown to have abnormal responsesto stimulation, we also elected to challenge the PBMCs withimmune activators known to promote oxidative stress andmeasure the resulting intracellular glutathione redox statusin activated isolated monocytes and T cells.

2. Subjects and Methods

2.1. Participants. This investigation was conducted on asubset of children from the autism IMAGE (IntegratedMetabolic and Genomic Endeavor) study at Arkansas Chil-dren’s Hospital Research Institute (ACHRI) that has recruit-ed over 162 case and control families to date. The IMAGEcohort for this study consisted of 43 children diagnosedwith autistic disorder and 41 unaffected control children(16 of which were unaffected siblings). The autism casefamilies were recruited locally after referral to the Universityof Arkansas for Medical Sciences (UAMS), Dennis Devel-opmental Center and diagnosed by trained developmentalpediatricians. Children aged 3 to 10 with a diagnosis ofautistic disorder as defined by the Diagnostic and StatisticalManual of Mental Disorders, Fourth Edition (DSM-IV 299.0),the Autism Diagnostic Observation Schedule (ADOS),and/or the Childhood Autism Rating Scales (CARS >30)were enrolled. Children diagnosed with other conditions onthe autism spectrum or rare genetic diseases associated withsymptoms of autism were excluded from the study. Childrenwith chronic seizure disorders, recent infection, and high-dose vitamin or mineral supplements exceeding the RDAwere also excluded because these conditions are potentialconfounders that could affect redox status. Unaffected sib-lings and unrelated, neurotypical children aged 3 to 10 withno medical history of behavioral or neurologic abnormalitiesby parent report made up the comparison group. Theprotocol was approved by the Institutional Review Board atUAMS, and all parents signed informed consent.

2.2. Materials. Culture flasks, plates, and pipettes were ob-tained from Corning Life Sciences (Lowell, Mass, USA).RPMI 1640, penicillin/streptomycin, Dulbecco’s phosphate-buffered saline (PBS), fetal bovine serum (FBS), and glu-tamine were purchased from Life Technologies (Carlsbad,Calif, USA). Carboxy-H2DCFDA (6-carboxy-2′,7′-dichlo-rodihydrofluorescein diacetate, diacetoxymethyl ester) wasobtained from Molecular Probes (Carlsbad, Calif, USA).Human Monocyte Isolation Kit II and Human CD4 Tcell Isolation Kit II were purchased from Miltenyi Biotec(Bergisch-Gladbach, Germany). Histopaque-1077 and allother chemicals were obtained from Sigma-Aldrich (St.Louis, Mo, USA).

2.3. Isolation of PBMCs and Stimulation of Monocytes andCD4 T Cells. Fasting blood samples (≤20 mL) were collectedbefore 9:00 AM into EDTA-Vacutainer tubes and immedi-ately chilled on ice before centrifuging at 1300×g for 10 minat 4◦C. Aliquots of plasma were stored at −80◦C in cryostattubes until extraction and HPLC quantification. PBMCswere isolated by centrifugation over Histopaque-1077. Red


blood cells were lysed using a brief (15 s) incubation with1 mL ice-cold water. Approximately, 30 × 106 PBMCs wereresuspended in RPMI 1640 medium (supplemented with10% FBS, 1% penicillin/streptomycin, and 2 mM glutamine)at a density of 106 cells/mL. Note that because we were unableto obtain 20 mL blood volume from every child, it was notpossible to isolate and analyze monocytes and CD4 T cellsfor all participants. For monocyte stimulation, PBMCs weretreated with 0.1 μg/mL lipopolysaccharide (LPS); for T-cellstimulation, PBMCs were treated with 10 ng/mL phorbol 12-myristate 13-acetate (PMA) and 1 μg/mL ionomycin. Cellswere placed in a humidified 5% CO2 incubator at 37◦Cfor 4 hr. Stimulated monocytes and CD4 T cells were thenisolated by negative selection using magnetic cell labeling asdescribed by the manufacturer (Miltenyi Biotec, Bergisch-Gladbach, Germany). Using flow cytometry, we determinedthat ≥75% of isolated monocytes are positive for CD14 andthat ≥87% of isolated CD4 T cells are positive for CD4. ForHPLC quantification of GSH and GSSG, approximately 2 ×106 unstimulated (resting) PBMCs, stimulated monocytes,or stimulated CD4 T cells were pelleted, snap frozen on dryice, and stored at −80◦C.

2.4. Cell Extraction and HPLC Quantification of IntracellularGlutathione and Plasma Cysteine Redox Status. The storageinterval at −80◦C before extraction was consistently between1-2 weeks after blood draw and cell isolation to minimizepotential metabolite interconversion. The methodologicaldetails for intracellular and extracellular GSH extractionand HPLC elution and electrochemical detection have beendescribed previously [15, 16], and metabolite detection doesnot require derivatization. Although most GSSG is presentas a mixed disulfide with other thiols including cysteine,our measurements detect only the free GSSG in plasma.Glutathione and cysteine concentrations were calculatedfrom peak areas of standard calibration curves using HPLCsoftware. Intracellular results are expressed as nanomolesper milligram of protein using the BCA Protein Assay Kit(Pierce, Rockford, Ill, USA), and plasma results are expressedas micromoles per liter.

2.5. Measurement of Intracellular Free Radicals. Carboxy-H2DCFDA (DCF) is a membrane-permeable ROS/RNS-sen-sitive probe that remains nonfluorescent until oxidized byintracellular free radicals. The intensity of DCF fluorescenceis directly proportional to the level of free radical oxidation.Approximately, 106 PBMCs were resuspended in 1 mL RPMI1640 medium supplemented with 10% FBS, 1% penicil-lin/streptomycin, and 2 mM glutamine and stained in thedark for 20 min with 1 μM DCF at 37◦C. Stained cells werewashed and resuspended in PBS and analyzed immediatelyon a Partec CyFlow flow cytometer (Gorlitz, Germany) using488 nm excitation wavelength with 530/30 nm (FL1) emis-sion filter. For each analysis, the fluorescence properties of10000 cells were collected, and the data were analyzed usingthe FCS Express software (De Novo Software, Los Angeles,Calif, USA). Intracellular free radical levels are expressed asmedian fluorescence intensity (MFI) of subject sample DCF

Table 1: Demographics of study population.

Case childrenn = 43

Control childrenn = 41

Age; mean (SD) 5.42 (1.98) 6.16 (2.29)

Male; n (%) 36 (84) 20 (49)

White; n (%) 38 (88.4) 31 (75.6)

Asian; n (%) 2 (4.65) 0 (0)

African American; n (%) 2 (4.65) 8 (19.5)

Hispanic; n (%) 1 (2.3) 2 (4.9)

OTC multivitamin use; n (%) 17 (39.5) 8 (19.5)

fluorescence normalized to DCF fluorescence of a standardPBMC preparation. As an internal control, the standardPBMC preparation was isolated from a 100 mL blood samplefrom an unaffected healthy adult volunteer, aliquoted andfrozen at −180◦C in 90% FBS/10% DMSO. An aliquot ofthe standard PBMC preparation was stained and analyzedwith each subject sample. Evaluation of oxidizing free radicalproduction was possible only in those case and unrelatedcontrol samples for which sufficient (∼20 mL) blood volumewas obtained.

2.6. Statistical Analysis. Within the control group, 16 of the41 unaffected control children were case siblings. There were27 additional case children without a sibling and 25 addi-tional unrelated control children comprising the total case-control cohort of 84 children. To down-weight the impact ofoutliers, three metabolites observations were curtailed at theextremes of the distributions for PBMC GSH, PBMC GSSG,and Monocytes GSH/GSSG (see footnote in Table 2). Thesibling data are correlated resulting in a combined sampleof correlated and uncorrelated data; thus, the assumption ofall data being independent is not satisfied for the standardtwo-sample t-test. To make use of all data from dependentand independent observations, we used the corrected Z-testproposed by Looney and Jones [47]. This statistical approachprovides adequate control of Type 1 errors and has morepower than a standard Student’s t-test. Because the DCF datacompared cases and unrelated controls (without siblings)the standard Student’s t-test was used with significance setat 0.05. Nonparametric intercorrelations (Spearman correla-tion coefficients) between age and gender and the 7 outcomevariables, GSH, GSSG, GSH/GSSG, % oxidized glutathione,cysteine, cystine, and cysteine/cystine were determined withthe significance level set at 0.05. Data was analyzed using SAS9.2 software (SAS Institute Inc, Cary, NC, USA).

3. Results

3.1. Demographics of Study Population. Table 1 lists thedemographics of the study population. The only majordifferences between cases and controls are that the controlgroup was composed of a greater proportion of females andAfrican Americans, whereas the case group had a greater pro-portion of Asian subjects. Over-the-counter multivitaminsupplement use was higher among cases (39.5%) compared


Table 2: Intracellular glutathione redox status in resting PBMCs and activated monocytes and CD4 T cells.

MetaboliteCase children Control children Corrected Z-test

n Mean ± SD n Mean ± SD Difference(95% CI)

P value

Resting PBMCs

GSH (nmol/mg protein) 43 25.45± 8.16 41 23.35± 6.38 2.09 (−1.09, 5.29) 0.19

GSSG (nmol/mg protein) 43 0.90± 0.3 41 0.66± 0.23 0.24 (0.13, 0.35) <0.001

GSH/GSSG 43 29.58± 9.04 41 37.58± 10.89 −7.99 (−12.51, −3.48) <0.001

Oxidized GSH (%) 43 0.07± 0.02 41 0.05± 0.01 0.02 (0.0075, 0.024) <0.001

Activated monocytes

GSH (nmol/mg protein) 18 7.73± 3.16 20 8.55± 2.5 −0.82 (−2.02, 0.38) 0.18

GSSG (nmol/mg protein) 18 0.62± 0.24 20 0.47± 0.17 0.14 (0.03, 0.25) 0.01

GSH/GSSG 18 13.31± 7.26 20 19.30± 6.35 −5.98 (−9.99, −1.97) 0.003

Oxidized GSH (%) 18 0.14± 0.05 20 0.10± 0.03 0.04 (0.02, 0.07) <0.001

Activated CD4 T cells

GSH (nmol/mg protein) 18 6.82± 3.0 19 10.16± 3.74 −3.33 (−5.24, −1.42) <0.001

GSSG (nmol/mg protein) 18 0.68± 0.29 19 0.63± 0.24 0.05 (−0.11, 0.22) 0.51

GSH/GSSG 18 10.47± 4.19 19 17.49± 6.95 −7.02 (−10.17, −3.87) <0.001

Oxidized GSH (%) 18 0.17± 0.05 19 0.11± 0.05 0.05 (0.03, 0.08) <0.001

GSH: glutathione; GSSG: oxidized glutathione disulfide; oxidized GSH: (%)2GSSG/(GSH+2GSSG); curtailment: PBMC GSH >45 set = 45 (n = 1); PBMCGSSG >1.75 set = 1.75 (n = 1); Monocytes GSH/GSSG >35 set = 35 (n = 1).

to controls (19.5%); however, the glutathione redox statuswas statistically unaffected by vitamin use (data not shown).

3.2. Decreased Intracellular Glutathione Redox Status inAutism. Table 2 presents the relative intracellular concen-trations of GSH, GSSG, the glutathione redox ratio, andthe percentage of oxidized glutathione equivalents in resting(unstim-ulated) PBMCs and in isolated stimulated mon-ocytes and CD4 T cells from children with autism andage-matched control children. The percent oxidized gluta-thione is expressed in absolute glutathione equivalents as2GSSG/(GSH+2GSSG). Relative to controls, the intracellularconcentration of GSSG and the percent oxidized glutathionewere significantly increased (∼40%), and the GSH/GSSGratio decreased (∼21%) in PBMCs from children withautism (P < 0.001). After stimulation with LPS, mono-cytes from children with autism also exhibited significantlydecreased GSH/GSSG (∼31%, P = 0.003), increased GSSGconcentration (∼32%, P = 0.01), and 40% higher percentoxidized glutathione (P < 0.001). In mitogen-stimulatedCD4 T cells from children with autism, the intracellular GSHconcentration was ∼33% lower, the GSH/GSSG was ∼40%lower (P < 0.001), and the percent oxidized glutathionewas ∼55% higher than in stimulated CD4 T cells fromcontrol children (<0.001). As expected, activation with LPSand PMA both resulted in decreased intracellular GSHlevels and GSH/GSSG in isolated monocytes and CD4 Tcells compared to resting (unstimulated) PBMCs. Uponstimulation, there was a greater decrease in intracellularGSH and GSH/GSSG in both CD4 T cells and monocytesfrom children with autism compared to control children.Neither age nor gender was significantly correlated with any

of the outcome measures. The protein content per 106 cellsdid not differ between case and control children (data notshown).

3.3. Decreased Extracellular Glutathione and Cysteine RedoxStatus in Autism. Table 3 presents the relative concentrationsof GSH, GSSG, GSH/GSSG, % oxidized GSH, cysteine,cystine, and the cysteine/cystine redox ratio in the extracel-lular plasma compartment. Children with autism exhibiteda significantly decreased extracellular concentration of GSH(∼21%) and GSH/GSSG (∼54%) and increased concentra-tion of GSSG and the percent oxidized glutathione (52%and 82%, resp., P < 0.001). Figures 1(a) and 1(b) com-pare GSH/GSSG and % oxidized glutathione equivalents,respectively, in plasma, T cells, and monocytes from case andcontrol children and graphically demonstrates the consistentdecrease in both extracellular and intracellular glutathioneredox status among the case children.

The dynamic balance between the reduced and oxidizedforms of glutathione can also be expressed as the redoxpotential or reducing power of the GSH/GSSG redox couple(Eh) and can be calculated from the Nernst equation, Eh =E0 + RT/nF ln[disulfide]/([thiol 1] ∗ [thiol 2]), where E0

is the standard potential for the glutathione redox couple(−264 mV), R is the gas constant (8.314 J/◦Kmol), T is theabsolute temperature of analytical measurement (25◦C =298◦K), n is 2 for the number of electrons transferred,and F is Faraday’s constant (96,485 coulomb/mol) [48].The calculated Eh value for the GSH pool in the childrenwith autism is −116 mV, which is 12 mV more oxidizedthan in the control children, with an Eh value of −128 mV(Table 3).


Table 3: Extracellular (plasma) glutathione and cysteine redox status.

MetaboliteCase children Control children Corrected Z-test

n Mean ± SD n Mean ± SD Difference(95% CI)

P value

Plasma

GSH (μM) 38 1.58± 0.23 41 1.99± 0.22 −0.41 (−0.50, −0.31) <0.001

GSSG (μM) 38 0.20± 0.06 41 0.13± 0.04 0.07 (0.05, 0.09) <0.001

GSH/GSSG 38 8.24± 2.20 41 17.14± 5.54 −8.73 (−10.52, −6.94) <0.001

Oxidized GSH (%) 38 0.20± 0.05 41 0.11± 0.03 0.09 (0.07, 0.10) <0.001

Eh for GSH −116 mV −128 mV

Cysteine (μM) 41 21.7± 4.88 41 21.43± 4.08 0.13 (−1.88, 2.14) 0.90

Cystine (μM) 41 29.2± 10.6 41 19.26± 4.8 9.73 (6.25, 13.2) <0.001

Cysteine/Cystine 41 0.79± 0.18 41 1.14± 0.18 −0.33 (−0.41, −0.26) <0.001

Eh for Cysteine −106 mV −111 mV

GSH: glutathione; GSSG: oxidized glutathione disulfide; Eh: steady-state redox potential; Eh for GSH: −264 mV +(30 mV) ∗ log([GSSG]/[GSH]2); Eh forcysteine: −250 mV + (30 mV)∗ log([CySSCy]/[Cys]2).

0

5

10

15

20

25

30

GSH

/GSS

G

ControlCase

Plasma Activated monocytes Activated T cells

∗

∗

∗

(a)

ControlCase

Plasma Activated monocytes Activated T cells0

0.05

0.1

0.15

0.2

0.25

0.3O

xidi

zed

GSH

(%)

∗∗

∗

(b)

Figure 1: Intracellular and extracellular glutathione redox imbalance in autism. (a) presents the GSH/GSSG in plasma, isolated activatedmonocytes, and CD4 T cells from case and control children; (b) presents the % oxidized glutathione equivalents. Both extracellular andintracellular glutathione redox status are consistently significantly decreased among the case children (∗P < 0.01).

The concentration of cystine, the oxidized form ofcysteine, was significantly elevated (∼52%), while the cys-teine/cystine redox ratio was significantly decreased (∼31%)in plasma from children with autism (P < 0.001). The Ehvalue for the cysteine pool can also be calculated from theNernst equation (see above) where the E0 for cysteine is equalto −250 mV [30]. The calculated Eh value for the cysteinepool in children with autism is −106 mV, or 5 mV moreoxidized than the control Eh value of −111 mV (Table 3).

3.4. Elevated Free Radical Production in Autism. The level ofintracellular free radicals was measured in available restingPBMCs from children with autism (n = 15) and unaffectedcontrol children (n = 16) using DCF, an ROS/RNS-sensitivefluorescent probe. Monocytes and lymphocytes were gatedbased on light scatter properties (size and density) and ana-lyzed separately. Figure 2 presents the median fluorescenceintensity (MFI) of lymphocytes from children with autism

and unaffected control children (normalized to MFI of thestandard PBMC preparation). Gated lymphocytes from chil-dren with autism exhibited a significantly higher mean levelof intracellular free radicals compared to lymphocytes fromcontrol children (P < 0.05). No differences in free radicalproduction were observed in gated monocytes from case andcontrol children. Intracellular free radical production wasnot correlated with age or gender in this cohort.

4. Discussion

Oxidative stress is generally defined as an imbalance betweenoxidant production and endogenous antioxidant defensemechanisms and can be clinically defined in humans by adecrease in the redox status of GSH/GSSG and cysteine/cys-tine thiol/disulfide redox couples [49]. The relative equi-librium between reduced and oxidized sulfhydryl groupsdefines the ambient redox state. Low glutathione redox status


has been associated with the pathophysiology of several neu-robehavioral disorders including schizophrenia [2, 50], bipo-lar disorder [3], alcoholism [51], HIV [52], and Alzheimer’sdisease [53]. This is the first study to evaluate intracellularglutathione-mediated antioxidant/redox capacity in primarycells from children with autism as well as the extracellularplasma cysteine/cystine redox status. Because these two redoxsystems are compartmentalized and independently regulat-ed, evaluation of both redox couples provides a completepicture of the primary immune cell microenvironment inchildren with autism. Supporting and extending our previ-ous findings of decreased plasma and lymphoblastoid cellGSH/GSSG, we now report that both primary immune cellGSH/GSSG and plasma cysteine/cystine redox couples aresimilarly compromised resulting in a more oxidized immunecell microenvironment in children with autism compared tocontrol children.

Recent evidence supports the notion that subtle fluctu-ations in ambient redox status may provide an importantregulatory mechanism that can dynamically modulate im-mune cell function. Activation and proliferation of T cellsrequire a reducing intracellular microenvironment, whereasa more oxidized environment promotes cell cycle arrest andblunted responsiveness to immune stimulation [54–57]. Forexample, a mechanism involving extracellular redox mod-ulation by regulatory T cells (Tregs) was recently elucidatedby Yan et al. [35]. Tregs were shown to inhibit the release ofcysteine into the immune synapse between dendritic cellsand naıve T cells, which effectively reduces GSH levels in Tcells by eliminating the rate-limiting amino acid for GSHsynthesis. A high ratio of reduced to oxidized glutathione isrequired for cell cycle progression from G1 to S phase andinduction of the T-cell proliferative response [55]. Thus, themore oxidized GSH/GSSG redox state of the intracellularglutathione pool in PBMCs and in activated CD4 T cellsobserved in children with autism (Table 2) would suggesta hyporesponsive phenotype that is less conducive to T-cellactivation and proliferation. Consistent with this hypothesis,several recent studies have documented abnormalities inthe adaptive immune response in children with autism[44, 58].

A glutathione deficit in T cells has been shown to nega-tively affect the adaptive immune response and T-cell pro-liferation by reducing IL-2 receptor turnover and IL-2-dependent DNA synthesis [59, 60]. In monocytes, an oxi-dized intracellular environment has been shown to alter thecytokine profile and skew the Th1 and Th2 balance [61, 62].Studies in mice have demonstrated that the intracellular GSHcontent of antigen presenting cells (APCs) reversibly altersthe Th1 and Th2 cytokine response pattern [61]. Specifically,a GSH deficit reduced Th1-associated IFN-γ production andexaggerated Th2-associated IL-4 production. Restoration ofGSH restored the Th1 cytokine response and normalizedthe Th2 response. Consistent with these observations, twoindependent studies have reported that helper T-cell sub-populations in PBMCs from children with autism are shiftedtowards T helper 2 (Th2) dominance [41, 42]. Further, adecrease in T-cell IL-2 receptor expression has been reported

∗

Intr

acel

lula

rfr

eera

dica

ls(n

orm

aliz

edto

stan

dard

)

Control Case

1.5

1

0.5

0

Figure 2: Intracellular Free Radicals are Elevated in Lymphocytesfrom Children with Autism. Intracellular free radicals were mea-sured in freshly isolated PBMC from children with autism andunaffected control children using 1 uM DCF. Presented is medianfluorescent intensity (MFI) of the gated lymphocyte populationfrom subject samples normalized to MFI of a standard PBMCpreparation also treated with 1 uM DCF and analyzed with eachsubject sample. Lymphocytes from children with autism exhibiteda significantly higher mean level of intracellular free radicals thancontrols (P = 0.04). Control median (95% CI) = 0.576 (0.551–0.640); case median (95% CI) = 0.689 (0.561–1.086).

to be associated with decreased proliferative response aftermitogen stimulation in children with autism [58].

The more oxidized GSH/GSSG redox status in plasmaand primary immune cells in children with autism (Figure 1)may offer a mechanistic explanation for the abnormal adap-tive immune response previously reported in these children.When intracellular oxidative stress exceeds glutathione redoxcapacity, cells export GSSG into the plasma as a mechanismto restore internal redox homeostasis [49, 63]. The elevatedGSSG concentrations in PBMCs (Table 2) suggest thatthe GSSG export mechanism and intracellular antioxidantcapacity were not sufficient to maintain intracellular redoxhomeostasis and that redox imbalance was chronic in thesechildren. The association between a more oxidized immunecell microenvironment and an abnormal adaptive immuneresponse warrants continued investigation especially in lightof the potential reversibility of immune dysfunction withtargeted treatment to restore redox homeostasis [15].

The calculated Eh values for the extracellular GSH andcysteine pools (Table 3) in our control population differsomewhat from previously published values. In adults, theplasma glutathione Eh is more reduced at around −137 mV,and the plasma cysteine redox couple is more oxidized at−80 mV [30, 48]. These discrepancies may reflect method-ological differences in sample preparation in that our electro-chemical detection does not require derivatization for detec-tion. It is also possible that children (age 3–10 years) mayhave less reducing capacity than previously reported in adults(age 25–35 years) [48]. Nonetheless, our calculated Eh valuesare consistent with previous reports that plasma cysteine Eh(−111 mV) is more oxidized than that of GSH (−128 mv).


Mean intracellular free radical production was higher inprimary lymphocytes from children with autism relative tolymphocytes from age-matched control children (Figure 2)and was driven by a subset of 5 (33.3%) children whoselymphocytes exhibited especially high levels of free radicals.Mitochondria are the primary producers and targets ofintracellular free radicals, and mitochondrial dysfunction hasbeen postulated to be a contributing factor in the pathogen-esis of autism and numerous other neurological disorders[64–67]. In a lymphoblastoid cell model, we previously dem-onstrated that the GSH/GSSG redox ratio in mitochondriawas significantly lower in autism compared to control cellsand was associated with a significantly lower mitochondrialmembrane potential after nitrosative stress [16]. It is wellestablished that mitochondria are highly concentrated inpresynaptic terminals and that loss of redox control can neg-atively affect the efficiency of neurotransmission and synap-tic plasticity [68, 69]. Similarly, mitochondrial localizationand redox signaling at the immunological synapse betweenlymphocytes and antigen presenting cells are required forimmune activation, and excessive ROS can interrupt thesesignaling pathways [70–72]. A recent study of mitochondrialdefects in lymphocytes from children with autism founddecreased complex I activity and overreplication of and de-letions in mitochondrial DNA compared to control lym-phocytes [73]. Based on this evidence, it is plausible to hy-pothesize that mitochondria may be the source of theincreased levels of lymphocyte free radicals observed in thesubset of autistic children presented in Figure 2. Consistentwith this hypothesis, a recent meta-analysis estimated thatmitochondria dysfunction may affect up to 30% of childrenwith autism [64]. Based on this evidence, further study ofmitochondrial function and redox status in lymphocytesfrom children with autism is warranted.

Relevant to our observations, two recent papers haverevealed that an oxidized extracellular cysteine/cystine redoxstatus can initiate a redox signaling cascade that stimulatesintracellular mitochondrial ROS production as a mechanismto initiate an inflammatory immune response [74, 75]. Thesignal transduction from the extracellular to intracellularcompartments occurs through oxidative modification ofredox-reactive cysteines on cell surface proteins. Exposedcysteine sulfhydryl groups on proteins can be reversibly oxi-dized to sulfenic acid or disulfide bonds resulting in alteredprotein structure and function that initiate downstreamredox signaling cascades [33, 76, 77]. In an elegant seriesof experiments, Imhoff and Hansen demonstrated that mi-tochondrial ROS production was significantly increased incells incubated under extracellular oxidized cysteine/cystineredox conditions [74]. The stimulated intracellular ROSproduction resulted in the expression of Nrf-2, the transcrip-tion factor responsible for initiation of the inflammatoryresponse. Treatment to block the availability of cell surfacecysteine thiol groups abrogated mitochondrial ROS produc-tion and Nrf-2 expression. Go et al. confirmed and extendedthese observations by demonstrating that treatment to main-tain mitochondrial redox status abrogated ROS productionin the presence of oxidized extracellular cysteine/cystine[75]. Although the precise mechanism for the oxidative

cysteine/cystine-dependent signaling for mitochondrial ROSproduction is not yet clear; the authors provide evidence ofa possible link to changes in the redox state of cytoskeletalproteins that could be functionally linked to the mitochon-drial membrane. Other studies have demonstrated that anoxidized plasma cysteine/cystine redox potential is associatedwith proinflammatory conditions [78, 79] and can be mod-ulated by diet [80, 81]. These observations support the pos-sibility that the oxidized plasma cysteine/cystine in childrenwith autism may be functionally related to the increase inlymphocyte free radical production observed and contributeto immune cell abnormalities in these children.

In summary, we show for the first time that both theextracellular and intracellular immune cell compartmentsare more oxidized in children with autism compared to age-matched unaffected control children. Randomized clinicaltrials will be needed to determine whether treatment tonormalize plasma and intracellular redox status will improveimmune cell function and possibly the health and behaviorin children with autism.

Conflict of Interests

The authors declare no conflict of interests.

Acknowledgments

The authors would like to express their gratitude to thefamilies in Arkansas affected by autism whose participa-tion made this study possible. They also acknowledge theinvaluable help of the nurses and clinicians at the DennisDevelopmental Center for referral and evaluation. Thisresearch was supported, in part, with funding from theNational Institute of Child Health and Development (RO1HD051873; SJJ), the Department of Defense (AS073218P1;SJJ), and by grants from the Arkansas Children’s Hospitaland Arkansas Biosciences Institute (SJJ).

References

[1] CDC, “Prevalence of autism spectrum disorders—autism anddevelopmental disabilities monitoring network, United States,2006,” Morbidity and Mortality Weekly Report, vol. 58, no. 10,pp. 1–20, 2009.

[2] J. K. Yao, S. Leonard, and R. Reddy, “Altered glutathioneredox state in schizophrenia,” Disease Markers, vol. 22, no. 1-2,pp. 83–93, 2006.

[3] A. C. Andreazza, M. Kauer-Sant’Anna, B. N. Frey et al.,“Oxidative stress markers in bipolar disorder: a meta-analysis,”Journal of Affective Disorders, vol. 111, no. 2-3, pp. 135–144,2008.

[4] W. R. Markesbery, “Oxidative stress hypothesis in Alzheimer’sdisease,” Free Radical Biology and Medicine, vol. 23, no. 1,pp. 134–147, 1997.

[5] G. A. Mostafa, E. S. El-Hadidi, D. H. Hewedi, and M. M.Abdou, “Oxidative stress in Egyptian children with autism:relation to autoimmunity,” Journal of Neuroimmunology,vol. 219, no. 1-2, pp. 114–118, 2010.

[6] N. A. Meguid, A. A. Dardir, E. R. Abdel-Raouf, and A. Hashish,“Evaluation of oxidative stress in autism: defective antioxidant


enzymes and increased lipid peroxidation,” Biological TraceElement Research, vol. 143, no. 1, pp. 58–65, 2011.

[7] A. Chauhan, V. Chauhan, W. T. Brown, and I. Cohen,“Oxidative stress in autism: increased lipid peroxidation andreduced serum levels of ceruloplasmin and transferrin—theantioxidant proteins,” Life Sciences, vol. 75, no. 21, pp. 2539–2549, 2004.

[8] S. S. Zoroglu, F. Armutcu, S. Ozen et al., “Increased oxidativestress and altered activities of erythrocyte free radical scaveng-ing enzymes in autism,” European Archives of Psychiatry andClinical Neuroscience, vol. 254, no. 3, pp. 143–147, 2004.

[9] S. Sogut, S. S. Zoroglu, H. Ozyurt et al., “Changes in nitricoxide levels and antioxidant enzyme activities may have a rolein the pathophysiological mechanisms involved in autism,”Clinica Chimica Acta, vol. 331, no. 1-2, pp. 111–117, 2003.

[10] T. L. Sweeten, D. J. Posey, S. Shankar, and C. J. McDougle,“High nitric oxide production in autistic disorder: a possiblerole for interferon-γ,” Biological Psychiatry, vol. 55, no. 4,pp. 434–437, 2004.

[11] M. A. Junaid, D. Kowal, M. Barua, P. S. Pullarkat, S. S. Brooks,and R. K. Pullarkat, “Proteomic studies identified a singlenucleotide polymorphism in glyoxalase I as autism suscepti-bility factor,” American Journal of Medical Genetics, vol. 131,no. 1, pp. 11–17, 2004.

[12] M. Boso, E. Emanuele, P. Minoretti et al., “Alterations ofcirculating endogenous secretory RAGE and S100A9 levelsindicating dysfunction of the AGE-RAGE axis in autism,” Neu-roscience Letters, vol. 410, no. 3, pp. 169–173, 2006.

[13] S. J. James, P. Cutler, S. Melnyk et al., “Metabolic biomarkers ofincreased oxidative stress and impaired methylation capacityin children with autism,” American Journal of Clinical Nutri-tion, vol. 80, no. 6, pp. 1611–1617, 2004.

[14] S. J. James, S. Melnyk, S. Jernigan et al., “Metabolic endophe-notype and related genotypes are associated with oxidativestress in children with autism,” American Journal of MedicalGenetics, Part B, vol. 141, no. 8, pp. 947–956, 2006.

[15] S. J. James, S. Melnyk, G. Fuchs et al., “Efficacy of methylcobal-amin and folinic acid treatment on glutathione redox status inchildren with autism,” American Journal of Clinical Nutrition,vol. 89, no. 1, pp. 425–430, 2009.

[16] S. J. James, S. Rose, S. Melnyk et al., “Cellular and mito-chondrial glutathione redox imbalance in lymphoblastoid cellsderived from children with autism,” FASEB Journal, vol. 23,no. 8, pp. 2374–2383, 2009.

[17] K. Bowers, Q. Li, J. Bressler, D. Avramopoulos, C. Newschaffer,and M. D. Fallin, “Glutathione pathway gene variation andrisk of autism spectrum disorders,” Journal of Neurodevelop-mental Disorders, vol. 3, no. 2, pp. 132–143, 2011.

[18] S. Biswas, A. S. Chida, and I. Rahman, “Redox modificationsof protein-thiols: emerging roles in cell signaling,” BiochemicalPharmacology, vol. 71, no. 5, pp. 551–564, 2006.

[19] G. Filomeni, G. Rotilio, and M. R. Ciriolo, “Cell signallingand the glutathione redox system,” Biochemical Pharmacology,vol. 64, no. 5-6, pp. 1057–1064, 2002.

[20] M. Fratelli, L. O. Goodwin, U. A. Ørom et al., “Gene expres-sion profiling reveals a signaling role of glutathione in redoxregulation,” Proceedings of the National Academy of Sciences ofthe United States of America, vol. 102, no. 39, pp. 13998–14003,2005.

[21] A. Pastore, G. Federici, E. Bertini, and F. Piemonte, “Analysis ofglutathione: implication in redox and detoxification,” ClinicaChimica Acta, vol. 333, no. 1-2, pp. 19–39, 2003.

[22] Y. W. Kwon, H. Masutani, H. Nakamura, Y. Ishii, and J. Yodoi,“Redox regulation of cell growth and cell death,” BiologicalChemistry, vol. 384, no. 7, pp. 991–996, 2003.

[23] F. Q. Schafer and G. R. Buettner, “Redox environment of thecell as viewed through the redox state of the glutathione disul-fide/glutathione couple,” Free Radical Biology and Medicine,vol. 30, no. 11, pp. 1191–1212, 2001.

[24] H. M. Lander, A. J. Mllbank, J. M. Tauras et al., “Redox regula-tion of cell signalling,” Nature, vol. 381, no. 6581, pp. 380–381,1996.

[25] P. Ghezzi, “Regulation of protein function by glutathionyla-tion,” Free Radical Research, vol. 39, no. 6, pp. 573–580, 2005.

[26] D. A. Dickinson and H. J. Forman, “Glutathione in defenseand signaling: lessons from a small thiol,” Annals of the NewYork Academy of Sciences, vol. 973, pp. 488–504, 2002.

[27] M. Noble, J. Smith, J. Power, and M. Mayer-Proschel, “Redoxstate as a central modulator of precursor cell function,” Annalsof the New York Academy of Sciences, vol. 991, pp. 251–271,2003.

[28] D. P. Jones, “Extracellular redox state: refining the definition ofoxidative stress in aging,” Rejuvenation Research, vol. 9, no. 2,pp. 169–181, 2006.

[29] M. Asensi, J. Sastre, F. V. Pallardo et al., “Ratio of reduced tooxidized glutathione as indicator of oxidative stress status andDNA damage,” Methods in Enzymology, vol. 299, pp. 267–276,1999.

[30] S. S. Iyer, C. J. Accardi, T. R. Ziegler et al., “Cysteine redoxpotential determines pro-inflammatory IL-1β levels,” PLoSOne, vol. 4, no. 3, Article ID e5017, 2009.

[31] D. Barford, “The role of cysteine residues as redox-sensitiveregulatory switches,” Current Opinion in Structural Biology,vol. 14, no. 6, pp. 679–686, 2004.

[32] D. P. Jones, Y. M. Go, C. L. Anderson, T. R. Ziegler, J. M.Kinkade Jr., and W. G. Kirlin, “Cysteine/cystine couple is anewly recognized node in the circuitry for biologic redox sig-naling and control,” FASEB Journal, vol. 18, no. 11, pp. 1246–1248, 2004.

[33] D. Spadaro, B. W. Yun, S. H. Spoel, C. Chu, Y. Q. Wang, andG. J. Loake, “The redox switch: dynamic regulation of proteinfunction by cysteine modifications,” Physiologia Plantarum,vol. 138, no. 4, pp. 360–371, 2010.

[34] J. Den Hertog, A. Groen, and T. van der Wijk, “Redox regu-lation of protein-tyrosine phosphatases,” Archives of Biochem-istry and Biophysics, vol. 434, no. 1, pp. 11–15, 2005.

[35] Z. Yan, S. K. Garg, J. Kipnis, and R. Banerjee, “Extracellularredox modulation by regulatory T cells,” Nature ChemicalBiology, vol. 5, no. 10, pp. 721–723, 2009.

[36] C. E. Cooper, R. P. Patel, P. S. Brookes, and V. M. Darley-Usmar, “Nanotransducers in cellular redox signaling: mod-ification of thiols by reactive oxygen and nitrogen species,”Trends in Biochemical Sciences, vol. 27, no. 10, pp. 489–492,2002.

[37] R. P. Warren, L. J. Yonk, R. A. Burger et al., “Deficiency ofsuppressor-inducer (CD4+CD45RA+) T cells in autism,” Im-munological Investigations, vol. 19, no. 3, pp. 245–251, 1990.

[38] L. J. Yonk, R. P. Warren, R. A. Burger et al., “CD4+ helper Tcell depression in autism,” Immunology Letters, vol. 25, no. 4,pp. 341–346, 1990.

[39] D. R. Denney, B. W. Frei, and G. R. Gaffney, “Lymphocytesubsets and interleukin-2 receptors in autistic children,”Journal of Autism and Developmental Disorders, vol. 26, no. 1,pp. 87–97, 1996.


[40] T. L. Sweeten, D. J. Posey, and C. J. McDougle, “High bloodmonocyte counts and neopterin levels in children with autisticdisorder,” American Journal of Psychiatry, vol. 160, no. 9,pp. 1691–1693, 2003.

[41] S. Gupta, S. Aggarwal, B. Rashanravan, and T. Lee, “Th1-and Th2-like cytokines in CD4+ and CD8+ T cells in autism,”Journal of Neuroimmunology, vol. 85, no. 1, pp. 106–109, 1998.

[42] C. A. Molloy, A. L. Morrow, J. Meinzen-Derr et al., “Elevatedcytokine levels in children with autism spectrum disorder,”Journal of Neuroimmunology, vol. 172, no. 1-2, pp. 198–205,2006.

[43] P. Ashwood and A. J. Wakefield, “Immune activation of pe-ripheral blood and mucosal CD3+ lymphocyte cytokine pro-files in children with autism and gastrointestinal symptoms,”Journal of Neuroimmunology, vol. 173, no. 1-2, pp. 126–134,2006.

[44] H. Jyonouchi, S. Sun, and H. Le, “Proinflammatory andregulatory cytokine production associated with innate andadaptive immune responses in children with autism spectrumdisorders and developmental regression,” Journal of Neuroim-munology, vol. 120, no. 1-2, pp. 170–179, 2001.

[45] P. Ashwood, A. Enstrom, P. Krakowiak et al., “Decreasedtransforming growth factor beta1 in autism: a potential linkbetween immune dysregulation and impairment in clinicalbehavioral outcomes,” Journal of Neuroimmunology, vol. 204,no. 1-2, pp. 149–153, 2008.

[46] K. Okada, K. Hashimoto, Y. Iwata et al., “Decreased serumlevels of transforming growth factor-β1 in patients withautism,” Progress in Neuro-Psychopharmacology and BiologicalPsychiatry, vol. 31, no. 1, pp. 187–190, 2007.

[47] S. W. Looney and P. W. Jones, “A method for comparingtwo normal means using combined samples of correlated anduncorrelated data,” Statistics in Medicine, vol. 22, no. 9,pp. 1601–1610, 2003.

[48] D. P. Jones, J. L. Carlson, V. C. Mody, J. Cai, M. J. Lynn, andP. Sternberg, “Redox state of glutathione in human plasma,”Free Radical Biology and Medicine, vol. 28, no. 4, pp. 625–635,2000.

[49] Y. M. Go and D. P. Jones, “Redox compartmentalization ineukaryotic cells,” Biochimica et Biophysica Acta, vol. 1780,no. 11, pp. 1273–1290, 2008.

[50] R. Gysin, R. Kraftsik, J. Sandell et al., “Impaired glutathionesynthesis in schizophrenia: convergent genetic and functionalevidence,” Proceedings of the National Academy of Sciences ofthe United States of America, vol. 104, no. 42, pp. 16621–16626,2007.

[51] M. Y. Yeh, E. L. Burnham, M. Moss, and L. A. Brown, “Chronicalcoholism alters systemic and pulmonary glutathione redoxstatus,” American Journal of Respiratory and Critical Care Med-icine, vol. 176, no. 3, pp. 270–276, 2007.

[52] H. Nakamura, H. Masutani, and J. Yodoi, “Redox imbalanceand its control in HIV infection,” Antioxidants and RedoxSignaling, vol. 4, no. 3, pp. 455–464, 2002.

[53] C. Cecchi, S. Latorraca, S. Sorbi et al., “Gluthatione level isaltered in lymphoblasts from patients with familial Alzheim-er’s disease,” Neuroscience Letters, vol. 275, no. 2, pp. 152–154,1999.

[54] A. Larbi, J. Kempf, and G. Pawelec, “Oxidative stress modula-tion and T cell activation,” Experimental Gerontology, vol. 42,no. 9, pp. 852–858, 2007.

[55] J. P. Messina and D. A. Lawrence, “Cell cycle progression ofglutathione-depleted human peripheral blood mononuclearcells is inhibited at S phase,” Journal of Immunology, vol. 143,no. 6, pp. 1974–1981, 1989.

[56] M. Klemke, G. H. Wabnitz, F. Funke, B. Funk, H. Kirchgessner,and Y. Samstag, “Oxidation of cofilin mediates T cell hypore-sponsiveness under oxidative stress conditions,” Immunity,vol. 29, no. 3, pp. 404–413, 2008.

[57] T. Ando, K. Mimura, C. C. Johansson et al., “Transductionwith the antioxidant enzyme catalase protects human T cellsagainst oxidative stress,” Journal of Immunology, vol. 181,no. 12, pp. 8382–8390, 2008.

[58] P. Ashwood, P. Krakowiak, I. Hertz-Picciotto, R. Hansen, I.N. Pessah, and J. Van de Water, “Altered T cell responses inchildren with autism,” Brain, Behavior, and Immunity, vol. 25,no. 5, pp. 840–849, 2011.

[59] C. M. Liang, N. Lee, D. Cattell, and S. M. Liang, “Glutathioneregulates interleukin-2 activity on cytotoxic T-cells,” Journal ofBiological Chemistry, vol. 264, no. 23, pp. 13519–13523, 1989.

[60] H. Gmunder, S. Roth, H. P. Eck, H. Gallas, S. Mihm, andW. Droge, “Interleukin-2 mRNA expression, lymphokine pro-duction and DNA synthesis in glutathione-depleted T cells,”Cellular Immunology, vol. 130, no. 2, pp. 520–528, 1990.

[61] J. D. Peterson, L. A. Herzenberg, K. Vasquez, and C.Waltenbaugh, “Glutathione levels in antigen-presenting cellsmodulate Th1 versus Th2 response patterns,” Proceedings ofthe National Academy of Sciences of the United States of Ameri-ca, vol. 95, no. 6, pp. 3071–3076, 1998.

[62] Y. Murata, T. Shimamura, and J. Hamuro, “The polarization ofTh1/Th2 balance is dependent on the intracellular thiol redoxstatus of macrophages due to the distinctive cytokine produc-tion,” International Immunology, vol. 14, no. 2, pp. 201–212,2002.

[63] L. Eklow, H. Thor, and S. Orrenius, “Formation and efflux ofglutathione disulfide studied in isolated rat hepatocytes,” FEBSLetters, vol. 127, no. 1, pp. 125–128, 1981.

[64] D. A. Rossignol and R. E. Frye, “Mitochondrial dysfunctionin autism spectrum disorders: a systematic review and meta-analysis,” Molecular Psychiatry. In press.

[65] A. H. Schapira, J. M. Cooper, D. Dexter, P. Jenner, J. B. Clark,and C. D. Marsden, “Mitochondrial complex I deficiency inParkinson’s disease,” The Lancet, vol. 1, no. 8649, p. 1269, 1989.

[66] A. Maruszak and C. Zekanowski, “Mitochondrial dysfunctionand Alzheimer’s disease,” Progress in NeuroPsychopharmacolo-gy and Biological Psychiatry, vol. 35, no. 2, pp. 320–330, 2011.

[67] H. B. Clay, S. Sillivan, and C. Konradi, “Mitochondrial dys-function and pathology in bipolar disorder and schizophre-nia,” International Journal of Developmental Neuroscience,vol. 29, no. 3, pp. 311–324, 2011.

[68] D. J. Keating, “Mitochondrial dysfunction, oxidative stress,regulation of exocytosis and their relevance to neurodegen-erative diseases,” Journal of Neurochemistry, vol. 104, no. 2,pp. 298–305, 2008.

[69] Z. Li, K. I. Okamoto, Y. Hayashi, and M. Sheng, “Theimportance of dendritic mitochondria in the morphogenesisand plasticity of spines and synapses,” Cell, vol. 119, no. 6,pp. 873–887, 2004.

[70] A. Quintana, C. Schwindling, A. S. Wenning et al., “T cell ac-tivation requires mitochondrial translocation to the immuno-logical synapse,” Proceedings of the National Academy of Scien-ces of the United States of America, vol. 104, no. 36, pp. 14418–14423, 2007.

[71] G. Pani, R. Colavitti, S. Borrello, and T. Galeotti, “Redox reg-ulation of lymphocyte signaling,” IUBMB Life, vol. 49, no. 5,pp. 381–389, 2000.

[72] T. M. Buttke and P. A. Sandstrom, “Redox regulation of pro-grammed cell death in lymphocytes,” Free Radical Research,vol. 22, no. 5, pp. 389–397, 1995.


[73] C. Giulivi, Y. F. Zhang, A. Omanska-Klusek et al., “Mitochon-drial dysfunction in autism,” Journal of the American MedicalAssociation, vol. 304, no. 21, pp. 2389–2396, 2010.

[74] B. R. Imhoff and J. M. Hansen, “Extracellular redox statusregulates Nrf2 activation through mitochondrial reactive oxy-gen species,” Biochemical Journal, vol. 424, no. 3, pp. 491–500,2009.

[75] Y. M. Go, H. Park, M. Koval et al., “A key role for mitochondriain endothelial signaling by plasma cysteine/cystine redoxpotential,” Free Radical Biology and Medicine, vol. 48, no. 2,pp. 275–283, 2010.

[76] L. I. Leichert and U. Jakob, “Protein thiol modificationsvisualized in vivo,” PLoS Biology, vol. 2, no. 11, Article ID e333,2004.

[77] P. Ghezzi, “Oxidoreduction of protein thiols in redox regula-tion,” Biochemical Society Transactions, vol. 33, no. 6, pp. 1378–1381, 2005.

[78] S. S. Iyer, D. P. Jones, K. L. Brigham, and M. Rojas, “Oxidationof plasma cysteine/cystine redox state in endotoxin-inducedlung injury,” American Journal of Respiratory Cell and Molecu-lar Biology, vol. 40, no. 1, pp. 90–98, 2009.

[79] S. K. Garg, V. Vitvitsky, R. Albin, and R. Banerjee, “Astrocyticredox remodeling by amyloid beta peptide,” Antioxidants andRedox Signaling, vol. 14, no. 12, pp. 2385–2397, 2011.

[80] S. E. Moriarty-Craige, J. Adkison, M. Lynn et al., “Antioxidantsupplements prevent oxidation of cysteine/cystine redox inpatients with age-related macular degeneration,” AmericanJournal of Ophthalmology, vol. 140, no. 6, pp. 1020–1026, 2005.

[81] D. P. Jones, Y. Park, N. Gletsu-Miller et al., “Dietary sulfuramino acid effects on fasting plasma cysteine/cystine redoxpotential in humans,” Nutrition, vol. 27, no. 2, pp. 199–205,2011.


Clinical Study

High Complement Factor I Activity in the Plasma of Childrenwith Autism Spectrum Disorders

Naghi Momeni,1 Lars Brudin,2 Fatemeh Behnia,3 Berit Nordstrom,4

Ali Yosefi-Oudarji,5 Bengt Sivberg,4 Mohammad T. Joghataei,5 and Bengt L. Persson1

1 School of Natural Sciences, Linnaeus University, 39182 Kalmar, Sweden2 Department of Clinical Physiology, Kalmar County Hospital, 39185 Kalmar, Sweden3 Department of Occupational Therapy, University of Social Welfare and Rehabilitation Sciences, Tehran, Iran4 Department of Health Sciences, Autism Research, Faculty of Medicine, Lund University, Box 157, 22100 Lund, Sweden5 Cellular and Molecular Research Centre, Tehran University of Medical Sciences (TUMS), Tehran, Iran

Correspondence should be addressed to Bengt Sivberg, [email protected]

Received 17 June 2011; Revised 22 August 2011; Accepted 22 August 2011

Academic Editor: Judy Van de Water

Copyright © 2012 Naghi Momeni et al. This is an open access article distributed under the Creative Commons Attribution License,which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

Autism spectrum disorders (ASDs) are neurodevelopmental and behavioural syndromes affecting social orientation, behaviour,and communication that can be classified as developmental disorders. ASD is also associated with immune system abnormality. Im-mune system abnormalities may be caused partly by complement system factor I deficiency. Complement factor I is a serine pro-tease present in human plasma that is involved in the degradation of complement protein C3b, which is a major opsonin of thecomplement system. Deficiency in factor I activity is associated with an increased incidence of infections in humans. In this paper,we show that the mean level of factor I activity in the ASD group is significantly higher than in the control group of typically de-veloped and healthy children, suggesting that high activity of complement factor I might have an impact on the development ofASD.

1. Introduction

Autism spectrum disorders (ASDs) are characterized by im-pairments in social interaction, communication, and repeti-tive or restricted patterns of interests, or behaviours, and areclassified as developmental disorders in DSM-IV [1]. ASD isabout 4-5 times more prevalent in boys than in girls. Theratio is estimated to range from 5.5 : 1.4 to 16.8 : 4.0 [2].Recent research clearly indicates that the underlying causesof autism are neurobiological disorders and combinations ofdifferent factors, such as environmental and genetic factors,and abnormality in the communication between neurons,probably associated with an abnormal set of neuropeptidesin the brain [3–9].

The symptoms of ASD have been linked with elevatedplasma levels of serotonin [10, 11] and opioid [12], abnormallevels of melatonin [13], altered levels of activity of the serineprotease prolyl endopeptidase [14], and infectious and im-munological factors such as abnormalities of T cells, B cells,natural killer (NK) cells, and of the complement system [15–21].

The complement system comprises a group of proteinswhich, when activated, provide one of the first lines ofdefence by promoting lysis and the removal of invadingmicrobes. Activation of the complement system in responseto an infection or foreign antigen is achieved via three com-plement pathways, the classical pathway, which is activatedby antigen-antibody complexes, the lectin pathway, which isactivated by the interaction of microbial carbohydrates withmannose-binding proteins in the plasma and tissue fluids,and the alternative complement pathway, which is activatedby C3b binding to microbial surfaces and to antibody mole-cules. All of the three pathways converge with the activationof the central C3 component. This leads to a final commonpathway with assembly of the C5–C9 components to forma cell surface transmembrane pore (membrane attack com-plex) [22, 23]. It has been shown by comparison with healthycontrol children that several differentially expressed proteinsare related to the complement system in children with ASD[22]. The alternative pathway consists of six proteins: C3, fac-tor B, factor D, factor H, factor I, and properdin. The plasma


glycoprotein factor I (C3b/C4b inactivator) is a serine pro-tease that acts as a regulator of the complement C3 cascade.Factor I has a molecular weight of about 88 kDa, consists oftwo disulfide-linked polypeptide chains (50 kDa and 38 kDa,resp.), and is synthesized as a single-chain precursor in theliver [24, 25]. Factor I cleaves C3b and C4b in a reaction,where fI is dependent on various cofactors, such as factorH, C4b-binding protein CR1 and membrane cofactor protein(MCP) [26]. Factor I-mediated cleavage of the α chain of C3bliberates 3 fragments with molecular weights of 68 kDa,43 kDa, and 2 kDa. Degradation of C3b by fI abrogates theaction of this protein in the C3 pathway [27]. ComplementC3b is the major opsonin of the complement system whichfacilitates the phagocytosis process by coating antigens (eachof the phagocytes expresses a complement receptor such asCR1, CR3, or CR4 that binds C3b, C4b, or C3bi) [28, 29].Factor I deficiency can be conferred by a C3 deficiency, sincethis also increases susceptibility to pyogenic infections byNeisseria meningitides, Haemophilus influenza, and Strep-tococcus pneumonia and increases the incidence of immunecomplex diseases due to impaired complement-mediatedfunction [30]. Immune system abnormalities have been asso-ciated with autism [15–20], and it has been suggested thatchildren with ASD might have an increased incidence of bac-terial inflammation [31]. Immunological aspects of the earlyonset of autism have recently highlighted the fact that im-mune dysfunction may occur in some children with autism[31, 32].

Having previously discovered altered levels of the serineprotease prolyl endopeptidase in children with ASD [14], theaim of this study was to investigate if an association existsbetween serine protease fI deficiency and the development ofASD.

2. Materials and Methods

2.1. Participants. Thirty children with ASD and thirty typicalcontrol children participated in this study. The ASD groupcomprised 23 boys and 7 girls with a mean age of 4.5 years(age range 3–9 years). The control group comprised 13 boysand 17 girls, mean age 6.0 years (age range 3–12 years),(Table 1).

Children in the ASD group were recruited from the Au-tism Rehabilitation Centre at the University of Social Welfareand Rehabilitation Sciences in Tehran, Iran. After havingobtained informed consent from the parents, blood sampleswere collected. All children with ASD were examined by cli-nical specialists on autism. A child psychiatrist and a childneurologist or child psychologist examined all of the chil-dren. All consultants agreed on the diagnosis of autism ac-cording to the DSM-IV criteria [1]. However, diagnostic pro-cedures applied in Europe and in the US/Canada using theautism diagnostic observation schedule (ADOS) [33] and theAutism Diagnostic Interview—Revised [34] were not used inthe diagnostic process applied in Iran. This shortcoming wascompensated for by the extensive clinical experience by thechild neurologist/child psychiatrist who was familiar with thecore behaviours in autism stated by the American Academy

of Pediatrics in its Embargo from 2007 [35]. The controlgroup consisted of typically developed and healthy childrenshowing no signs of neurodevelopmental disorders who wererecruited from the same area as the children with ASD.Children who had any kind of infection/infectious diseasewithin two weeks prior to the time of examination wereexcluded from this study.

The study was approved (MT/1247) by the ethics com-mittee of the Iran University of Medical Sciences, Tehran.

2.2. Procedure

2.2.1. Blood Sample Collection. Blood samples were collectedby a paediatric nurse, and those from the children diagnosedwith autism were collected under the supervision of a childpsychiatrist with special training in the field of childhoodpsychosis. Venous blood was collected into 3 mL EDTAtubes (Vacutainer System; Becton-Dickinson Inc., Plymouth,UK), and plasma was separated immediately thereafter bycentrifugation at 1,300 g for 10 min at 4◦C. Thereafter, aninhibitors cocktail (30 μL per 1 mL plasma) was added to theresultant plasma sample (cocktail inhibitor solution: 2.0 MTris, 0.9 M Na-EDTA, 0.2 M Benzamidine, 92 μM E-64, and48 μM Pepstatin; Sigma, St. Louis, Mich, USA). The plasmawas stored at −80◦C.

2.2.2. Assay. Methods based on the hydrolysis of fluorogenicsubstrates have previously been described by Tsiftoglou andSim [36] and Gupta et al. [15]. The following assay procedurewas found to be optimal for assaying fI activity in theplasma. 20 μL of plasma was incubated with 80 μL of buffer(100 mM phosphate buffer, pH 7.5, containing 1 mM EDTA,1 mM DTT and 1 mM sodium azide) for 10 min at 37◦C toreach thermal equilibrium. 100 μL of substrate solution(200 μM Boc-Asp(OBz)-Pro-Arg-7-amino-4-methylcouma-rin; Bachem, Bubendorf, Switzerland) in 25 mM phosphatebuffer, pH 7.4, was then added, and the mixture was incu-bated at 37◦C for 60 min (see Figure 1). The reaction wasinhibited by the addition of 1 mL of 1.5 M acetic acid, and therelease of 7-amino-4-methylcoumarin was measured byspectrofluorometer (Hitachi-f 2000; λex: 360 nm; λem:440 nm; slit width: 2.5). All measurements were carried outrandomized and in duplicate. Background fluorescence inthe assay was monitored by the use of plasma in the absenceof substrate and was subtracted from values obtained in thepresence of substrate.

2.3. Data Analysis and Statistics. Plasma fI activity was log-normally distributed, and logarithmic values were, therefore,used when analysing differences between the ASD group andthe control group. To adjust for age (dichotomized using themedian value, 5 years) and gender, factorial ANOVA wasused. A P value < 0.05 was considered statistically significant.Statistica 8.0 (StatSoft c©, Tulsa, Okla, USA) was used. Intra-and interassay variability of the plasma fI activity was expres-sed as the standard error of a single determination (Smethod),using the formula first proposed by Dahlberg [37]

Smethod = √( ∑di

2

(2n)

), (1)


Table 1: Age/y, gender, and medication of the participants.

Parameter ASD (n = 30) Controls (n = 30) P value∗

Age

Mean (SD) 4.8 (1.7) 6.1 (2.3)

Median (range) 4.5 (3–9) 6 (3–12) 0.033

≤5 years 21 14

>5 years 9 16 0.115

GenderMales 23 13

Females 7 17 0.017

MedicationNo specific medication 8 30 —

Risperdal alone or in combination 18 0 —

Ritalin alone or in combination 4 0 —∗

Difference between ASD and controls. Mann-Whitney U-test for age and Fisher’s exact test for age category and gender.

3000

2500

2000

1500

1000

500

01 2 3

Time (hours)

Flu

ores

cen

cein

ten

sity

un

it

Factor I activity and substrate stability

Figure 1: Fluorescence intensity of release of 7-amino-4-methylcoumarin as a function of plasma incubation period (mean ± SD;1 h (974 ± 44.2), 2h (1995 ± 45.2) and 3 h (2374 ± 1.2)).

where di is the difference between the i : th paired measure-ment and n is the number of differences. The Smethod was ex-pressed as the coefficient of variation (%).

3. Results

There was significantly higher activity of plasma fI in thechildren with ASD (geometric mean (95% confidence limit):523 (154–1776) when compared with the control group: 361(135–967; ANOVA P = 0.015, adjusted for age and gender;Figure 2).

No statistically significant interactions were found withregard to gender and age, and no significant associations werefound between fI activity and age or gender (ANOVA; P =0.25 for gender and 0.42 for the two age groups, Figure 2). Inthe ASD group, some children with severe autism were under

medication with Risperdal to reduce hyperactivity and vio-lent behavior, and a few were under medication with Ritalinto improve attention (Table 1). It would have been ethicallyquestionable to discontinue medication with the purposeof controlling the experimental design. We have correlatedthe data shown in Figure 2(a) with the type of medicationthe children in the ASD group were receiving. Although wedid not see any clear correlation between medication anddistribution for the scatter plot data, it cannot be excludedthat some differences in the pattern may be influenced bymedication, as has been previously discussed [22].

The values were statistically significantly higher in thechildren with ASD, and there was a weak association withgender. No statistically significant differences were found,however, between the age groups.

The methodological intra-assay error was small, 0.5%.The interassay methodological error was 13%. We founda significantly higher complement factor I enzyme activityin children with ASD compared to the control group ofaround the same age. This is, as far as we know, the firstreport regarding dysfunction of fI activity in children withASD. Although not statistically significant, males tended toexhibit higher fI activity than females, and the differencebetween the control group and the ASD group was moreconvincing amongst the younger children, as shown inFigure 2. Due to fI’s role as a regulating factor in the comple–ment system pathway, an fI abnormality could play a rolein the onset of ASD. A defect in this pathway makes the-individual more vulnerable to various inflammations. Somereports [21, 24–26] provide increasing evidence of a con-nection between immunological abnormality and humandisease. ASD numbers among the types of disorders thatare associated with immune system abnormalities [28, 38].Our results are consistent with a recent proteomic study ofserum from ASD children, where a significant differentialexpression was shown for proteins related to the complementsystem [22].

4. Discussion

The aetiology of ASD is still largely unknown despite the factthat many factors such as genetic, environmental, immuno-logical, and neurological aspects are thought to influence


3000

2000

1000

0Control ASD

Participants

Risperdal

Ritalin

Combination

Others

Flu

ores

cen

cein

ten

sity

Non medicated(Risperdal and Ritalin)

(a)

7

6.6

6.2

5.8

5.4

Females

Males

Age ≤ 5 years Age > 5 years

Females

Males

Control ASD

Participants

Control ASD

Participants

log

(flu

ores

cen

cein

ten

sity

)

(b)

Figure 2: (a) Complement factor I activity in EDTA plasma from children with ASD (n = 30) and healthy control children (n = 30). Ascatter plot of factor I activity for each individual is shown. Samples from ASD children who were not under medication at the time of theinvestigation, those under medication with Risperdal, Ritalin, a combination thereof, or other medications such as antipsychotics (thiorida-zine) or anticonvulsants (fenobarbital and sodium valproate) are shown. (b) Mean values and standard errors of the complement factor Iactivity in EDTA plasma are shown for the different age groups and genders for both children with ASD and the healthy control group. Inthe ASD group; age ≤ 5 years: males (n = 17), females (n = 4) and age > 5 years: males (n = 6), females (n = 3). In the healthy controlgroup; age ≤ 5 years: males (n = 4), females (n = 10) and age > 5 years: males (n = 9), females (n = 7).

the development of ASD [39]. A dysregulated immune res-ponse has been reported among children with ASD. Pro-tein products of immune activation, such as cytokines, canbe linked to core features of ASD, such as difficulties to reg-ulate affect, sleep, nutritional uptake, and can also affectbehaviour and social communication [39]. This study high-lights an elevated level of factor I that may contribute to adysregulation of the immune response associated with thecomplement system. Recent research proposes an extensivecommunication between the immune and nervous systems,affecting both the development of the central nervous system[32] in the promotion of health as well as disease. Earlydamage during critical periods in neurodevelopment of thefoetus has in a mouse model study been shown to affectcognitive development. It is reasonable to suggest thatcomplement factor I might contribute to ASD, while changesin the complement system may predispose the mother orfoetus to infections during development, possibly causingresultant abnormalities in brain development [22]. Also, thefrequency of autoimmune diseases has been reported to besignificantly higher in families with a child with ASD [40]than in families with children with typical development.

The pathophysiology of ASD is poorly understood. Chil-dren with ASD are prone to recurrent viral and bacterial in-flammations. There are also some reports of immune sys-tem abnormalities in children with ASD [15–20]. An associa-tion between ASD and immune system abnormalities toge-ther with the vulnerability of the ASD group with regard toinflammatory processes may indicate an impaired mecha-nism in this system. It is known that phagocytosis is animportant part of the body’s defence mechanism. Thismechanism requires an active complement system and

a functioning C3b protein. C3b is degraded by fI, and abnor-mal fI activity might cause an abnormal and uncontrolleddegradation of C3b protein, resulting in the loss of thephagocytosis function for this particular protein (a functionwhich partly protects the body from invasion of foreignorganisms). Altered levels of other serine protease activities,such as that of proline endopeptidase (PEP), have also beenfound in a group of children with ASD when compared to acontrol group [14]. The results of the present study, togetherwith our previous findings on altered levels of PEP activity,may indicate a connection between the onset of ASD andserine protease dysfunction. The higher plasma fI activityobserved in the male group as compared to that of thefemale group (Figure 2) is paralleled by a higher occurrenceof ASD in male children. Also, the higher plasma fI activity inchildren younger than six years of age may indicate that theinflammatory process is more active in younger ages or thatwe may be dealing with two subgroups of children with ASDwith different onsets of the disease.

5. Conclusions

The preliminary findings of this study together with ourprevious report [14] suggest that there may be an associa-tion between abnormal serine protease activity and the deve-lopment of ASD. Further research is needed, however, toestablish a possible role of serine proteases in the aetiologyof ASD.

Conflict of Interests

The authors declare that they have no conflict of interest.


Acknowledgments

The authors wish to thank Dr. Mohammad A. Karbasian andall of the staff at the Robat Karim Medical Diagnostic Labora-tory and the staff at the Cellular and Molecular ResearchCentre, TUMS, Tehran, for their assistance in this study.

References

[1] American Psychiatric Association, “Disorders usually firstdiagnosed in infancy, childhood, or adolescence,” in Diagnosticand Statistical Manual of Mental Disorders, American Psychi-atric Association, Arlington, Va, USA, 4th edition, 1994.

[2] “Prevalence of autism spectrum disorders—Autism andDevelopmental Disabilities Monitoring Network SurveillanceYear 2006 Principal Investigators, Prevalence of autism spec-trum disorders—autism and developmental disabilities mon-itoring network, United States, 2006,” MMWR SurveillanceSummaries, vol. 58, no. 10, pp. 1–20, 2009.

[3] M. L. Bauman and T. L. Kemper, “Neuropathology of infantileautism,” Molecular Psychiatry, vol. 7, no. 2, pp. S12–S13, 2002.

[4] K. B. Nelson, J. K. Grether, L. A. Croen et al., “Neuropeptidesand neurotrophins in neonatal blood of children with autismor mental retardation,” Annals of Neurology, vol. 49, no. 5, pp.597–606, 2001.

[5] S. H. Fatemi, J. M. Stary, A. R. Halt, and G. R. Realmuto, “Dys-regulation of Reelin and Bcl-2 Proteins in Autistic Cerebel-lum,” Journal of Autism and Developmental Disorders, vol. 31,no. 6, pp. 529–535, 2001.

[6] M. L. Bauman, “Microscopic neuroanatomic abnormalities inautism,” Pediatrics, vol. 87, no. 5, pp. 791–796, 1991.

[7] G. J. Blatt, C. M. Fitzgerald, J. T. Guptill, A. B. Booker, T. L.Kemper, and M. L. Bauman, “Density and distribution of hip-pocampal neurotransmitter receptors in autism: An Autorad-iographic Study,” Journal of Autism and Developmental Disor-ders, vol. 31, no. 6, pp. 537–543, 2001.

[8] L. Y. Tsai, R. R. Crowe, S. R. Patil, J. Murray, and J. Quinn,“Search for DNA markers in two autistic males with the fragileX syndrome,” Journal of Autism and Developmental Disorders,vol. 18, no. 4, pp. 681–685, 1988.

[9] E. London, “The environment as an etiologic factor in autism:a new direction for research,” Environmental Health Perspec-tives, vol. 108, no. 3, pp. 401–404, 2000.

[10] S. Tordjman, L. Gutknecht, M. Carlier et al., “Role of the sero-tonin transporter gene in the behavioral expression of autism,”Molecular Psychiatry, vol. 6, no. 4, pp. 434–439, 2001.

[11] V. K. Singh, E. A. Singh, and R. P. Warren, “Hyperserotonine-mia and serotonin receptor antibodies in children with autismbut not mental retardation,” Biological Psychiatry, vol. 41, no.6, pp. 753–755, 1997.

[12] S. I. Deutsch, “Rationale for the administration of opiate anta-gonists in treating infantile autism,” American Journal of Men-tal Deficiency, vol. 90, no. 6, pp. 631–635, 1986.

[13] J. Melke, H. Goubran Botros, P. Chaste et al., “Abnormal mela-tonin synthesis in autism spectrum disorders,” Molecular Psy-chiatry, vol. 13, no. 1, pp. 90–98, 2008.

[14] N. Momeni, B. M. Nordstrom, V. Horstmann, H. Avarsaji, andB. V. Sivberg, “Alterations of prolyl endopeptidase activity inthe plasma of children with autistic spectrum disorders,” BMCPsychiatry, vol. 5, no. 27, article 27, pp. 1–6, 2005.

[15] S. Gupta, S. Aggarwal, and C. Heads, “Dysregulated immunesystem in children with autism—beneficial effects of intra-venous immune globulin on autistic characteristics,” Journal

of Autism and Developmental Disorders, vol. 26, no. 4, pp. 439–452, 1996.

[16] G. Delgiudice-Asch and E. Hollander, “Altered immune fun-ction in autism,” CNS Spectrums, vol. 2, no. 5, pp. 61–68, 1997.

[17] R. P. Warren and A. N. C. Foster Margaretten, “Reduced nat-ural killer cell activity in autism,” Journal of the American Aca-demy of Child and Adolescent Psychiatry, vol. 26, no. 3, pp. 333–335, 1987.

[18] R. P. Warren, L. J. Yonk, R. A. Burger et al., “Deficiency of sup-pressor-inducer (CD4+CD45RA+) T cells in autism,” Immun-ological Investigations, vol. 19, no. 3, pp. 245–251, 1990.

[19] L. J. Yonk, R. P. Warren, R. A. Burger et al., “CD4+ helper T celldepression in autism,” Immunology Letters, vol. 25, no. 4, pp.341–346, 1990.

[20] R. P. Warren, N. C. Margaretten, N. C. Pace, and A. Foster,“Immune abnormalities in patients with autism,” Journal ofAutism and Developmental Disorders, vol. 16, no. 2, pp. 189–197, 1986.

[21] G. Goldberger, M. A. Arnaout, D. Aden, R. Kay, M. Rits, andH. R. Colten, “Biosynthesis and postsynthetic processing ofhuman C3b/C4b inactivator (factor I) in three hepatoma celllines,” Journal of Biological Chemistry, vol. 259, no. 10, pp.6492–6497, 1984.

[22] B. A. Corbett, A. B. Kantor, H. Schulman et al., “A proteomicstudy of serum from children with autism showing differentialexpression of apolipoproteins and complement proteins,”Molecular Psychiatry, vol. 12, no. 3, pp. 292–306, 2007.

[23] J. Parkin and B. Cohen, “An overview of the immune system,”Lancet, vol. 357, no. 9270, pp. 1777–1789, 2001.

[24] G. Goldberger, G. A. P. Bruns, M. Rits, M. D. Edge, andD. J. Kwiatkowski, “Human complement factor I: analysis ofcDNA-derived primary structure and assignment of its geneto chromosome 4,” Journal of Biological Chemistry, vol. 262,no. 21, pp. 10065–10071, 1987.

[25] R. B. Sim, A. J. Day, B. E. Moffatt, and M. Fontaine, “Com-plement factor I and cofactors in control of complement sys-tem convertase enzymes,” Methods in Enzymology, vol. 223, pp.13–35, 1993.

[26] C. J. Soames and R. B. Sim, “Interactions between humancomplement components factor H, factor I and C3b,” Bio-chemical Journal, vol. 326, no. 2, pp. 553–561, 1997.

[27] B. P. Morgan and M. J. Walport, “Complement deficiency anddisease,” Immunology Today, vol. 12, no. 9, pp. 301–306, 1991.

[28] M. M. Frank and L. F. Fries, “The role of complement ininflammation and phagocytosis,” Immunology Today, vol. 12,no. 9, pp. 322–326, 1991.

[29] J. Kuby, Immunology, W. H. Freeman and Company, NewYork, NY, USA, 3rd edition, 1997.

[30] T. J. Vyse, P. J. Spath, K. A. Davies et al., “Hereditary comple-ment factor I deficiency,” Quarterly Journal of Medicine, vol.87, no. 7, pp. 385–401, 1994.

[31] P. Ashwood, B. A. Cobett, A. Kantor, H. Schulman, J. Vande Water, and D. G. Amaral, “In search of cellular immuno-phenotypes in the blood of children with autism,” Public Libr-ary of Science One, vol. 6, no. 5, Article ID e19299, 2011.

[32] P. A. Garay and A. K. McAllister, “Novel roles for immunemolecules in neural development: implications for neurode-velopmental disorders,” Frontiers in Synaptic Neuroscience, vol.2, pp. 1–16, 2010.

[33] C. Lord, M. Rutter, P. C. Dilavore, and S. Risi, Autism Diagnos-tic Observation Schedule (ADOS), Western Psychological Ser-vices, Los Angeles, Calif, USA, WPS edition, 1999.

[34] M. Rutter, A. Lecouteur, and C. Lord, Autism DiagnosticInterview—Revised (ADI-R), Western Psychological Services,Los Angeles, Calif, USA, WPS edition, 2003.


[35] Embargo Policy, “Identification and evaluation of childrenwith autism spectrum disorders,” The American Academy ofPediatrics, vol. 120, no. 5, pp. 1183–1215, 2007.

[36] S. A. Tsiftsoglou and R. B. Sim, “Human complement factor Idoes not require cofactors for cleavage of synthetic substrates,”Journal of Immunology, vol. 173, no. 1, pp. 367–375, 2004.

[37] G. Dahlberg, Statistical Methods for Medical and BiologicalStudents, George Allen and Unwin, London, UK, 2nd edition,1940.

[38] E. Korvatska, J. Van de Water, T. F. Anders, and M. E. Gersh-win, “Genetic and immunologic considerations in autism,”Neurobiology of Disease, vol. 9, no. 2, pp. 107–125, 2002.

[39] P. Ashwood, S. Wills, and J. Van De Water, “The immuneresponse in autism: a new frontier for autism research,” Journalof Leukocyte Biology, vol. 80, no. 1, pp. 1–15, 2006.

[40] G. A. Mostafa and A. A. Shehab, “The link of C4B null allele toautism and to a family history of autoimmunity in Egyptianautistic children,” Journal of Neuroimmunology, vol. 223, no.1-2, pp. 115–119, 2010.

Documents

Autism: Where Genetics Meets the Immune Systemdownloads.hindawi.com/journals/specialissues/908965.pdf · 2019. 8. 7. · healing in the skin, in the gastrointestinal, and respiratory