Fetal programming by maternal stress :

REVIEW

Fetal

programming

by

maternal

stress:

Insights

from

a

conflict

perspective

Marco

Del

Giudice

DepartmentofPsychology,UniversityofTurin,ViaPo14,10123Torino,Italy

Received17March2012;receivedinrevisedform21May2012;accepted22May2012

Contents

1. Introduction... 1615

1.1. Behavioraloutcomesofprenatalstressasadaptiveresponses ... 1615

1.1.1. Thecomplexrelationbetweenstresshormonesandenvironmentalstates... 1616

1.1.2. Prenatalstressinhumans... 1616

1.2. Conflictinfetalprogramming... 1616

1.2.1. Prenatalconflicts... 1617

1.2.2. Prenatalstressasaconflictarena ... 1617

2. Theoreticalinsights... 1618

2.1. Whatistheconflictabout?... 1618

2.1.1. Prenatalprogrammingofpostnatalplasticity... 1618

2.2. Fetaltactics,maternaltactics,andconflictoutcomes... 1619 KEYWORDS

11b-HSD; Cortisol;

Fetalprogramming; Genomicimprinting; HPAaxis;

Maternaleffects; Parent-offspringconflict; Placentalhormones; Pregnancy;

Prenatalconflict; Progesterone; Stress

Summary Maternalstressduringpregnancyhaspervasiveeffectsontheoffspring’sphysiology andbehavior,includingthedevelopmentofanxious,reactivetemperamentandincreasedstress responsivity.Theseoutcomescanbeseenastheresultofadaptivedevelopmentalplasticity: maternalstresshormonescarryusefulinformationaboutthestateoftheexternalworld,which canbeusedbythedevelopingfetustomatchitsphenotypetothepredictedenvironment.This account,however,neglectstheinherentconflictofinterestbetweenmotherandfetusaboutthe outcomesoffetalprogramming.Theaimofthispaperistoextendtheadaptivemodelofprenatal stressbyframingmother-fetusinteractionsinanevolutionaryconflictperspective.Inthepaper,I showhowaconflictperspectiveprovidesmanynewinsightsinthefunctionsandmechanismsof fetalprogramming,withparticularemphasisonhumanpregnancy.Ithentakeadvantageofthose insightstomakesenseofsomepuzzlingfeaturesofmaternalandfetalphysiologyandgenerate novelempiricalpredictions.

#2012ElsevierLtd.Allrightsreserved.

E-mailaddress:marco.delgiudice@unito.it.

Availableonlineatwww.sciencedirect.com

j o ur nal h o m ep a ge:w ww.e ls e vi e r.c o m /l o c at e/ p s y ne ue n

0306-4530/$—seefrontmatter#2012ElsevierLtd.Allrightsreserved.

3. Physiologicalinsights... 1620

3.1. Maternalstressandfetalprogramminginhumans ... 1620

3.1.1. Pregnancy-relatedchangesinstressphysiology ... 1620

3.1.2. Hormonalmediatorsoffetalprogramming... 1621

3.2. Physiologicalconflictsinfetalprogramming... 1622

3.2.1. Placentalfilteringmechanisms... 1622

3.2.2. Fetalmanipulationofmaternalphysiology... 1622

3.3. Aroleforimprintedgenes? ... 1623

3.3.1. Imprintedgenesandfetalprogramming... 1624

4. Conclusions ... 1624

4.1. Summaryofthepaper ... 1624

4.2. Furtherimplications... 1625

Acknowledgement... 1625

References... 1625

1.

Introduction

Fromconceptiontodelivery,mammalianfetusesareexposed toacontinualstreamofchemicalsignalscarriedbymaternal blood.Amongthesesignals,aprominentroleisplayedbythe hormones associated with the stress response. A growing amount of evidence from human and nonhuman studies showsthatmaternal stressduringpregnancy exerts perva-sive, long-lasting effects on the development ofthe fetal nervoussystem—and,ultimately,ontheoffspring’s physiol-ogy and behavior.In psychology andmedicine, the classic approach to maternalstress hasbeen to treat it as a dis-ruptiveinfluenceonfetaldevelopmentandanetriskfactor forfuturepathology(e.g.,VandenBerghetal.,2005; Wein-stock,2005;Mennesetal.,2006).Inmarkedcontrastwith thisview,thelastdecadehaswitnessedtheemergenceand consolidationofanalternativemodel,basedonevolutionary biology, in which fetal programmingby maternal stress is seenasanevolvedadaptiveprocess(e.g.,Matthews,2002; Kaiser andSachser,2005,2009; Kapooretal.,2006; Talge etal.,2007;Glover,2011;PluessandBelsky,2011;Sandman etal.,2012).

Atthecoreoftheadaptivemodelistheideathatmaternal stresshormonescarryusefulinformationaboutthestateof theexternalworld—forexampleitssafetyand predictabil-ity,thepresenceofthreats,andsoforth—thatisotherwise inaccessibletothefetus.Thedevelopingfetuscanusethis informationasa‘‘forecast’’oftheenvironmentalconditions it will eventually face after birth, and start adjusting its physiological andbehavioral profileto match the require-ments oftheworlditwillprobably encounter.Inthis per-spective, fetal programming is an instance of adaptive developmentalplasticity(West-Eberhard, 2003).More spe-cifically,itcanbeconsideredapredictiveadaptiveresponse, asthefetusmakesuseofpresentcuestoentrainthe devel-opmentofalternativephenotypesthatwillbecomeadaptive inthefuture(Hinde,1986;Belskyetal.,1991;Batesonetal., 2004;GluckmanandHanson,2004;Gluckmanetal.,2005). The adaptive model of prenatal stress is a powerful application of evolutionary theory to early development. Byuncoveringtheultimatelogicofphysiologicalprocesses, ithelps makesenseof abroadarrayofempiricalfindings. Nevertheless,acrucialpieceisstillmissingfromthepuzzle: theinherentconflictofinterestbetweenmotherandfetus

abouttheoutcomesoffetal programming.Theaimof this paperistoextend theadaptivemodelbyframing mother-fetusinteractionsinanevolutionaryconflictperspective.In theremainderofthissectionIwillbrieflyreviewthe beha-vioralandneurobiologicaloutcomesofprenatalstress, intro-duce the theory of parent-offspring conflict, and discuss some conflictual aspects of mother-fetus interactions. In Sections 2 and 3, I will show how a conflict perspective providesmanynewinsightsinthefunctionsandmechanisms offetal programming. Iwillthen take advantageof those insightstomakesenseofsomepuzzlingfeaturesofmaternal andfetalphysiology, andgenerate novelempirical predic-tions.

1.1. Behavioral outcomesofprenatalstressas adaptiveresponses

Ageneralpattern,observedinratsandnonhumanprimates alike,is that the offspring of prenatally stressed mothers tendtoshowincreasedanxiety-likebehaviorsandreduced attentionalspan.Theyalsoshowhigherbasalactivityofthe hypothalamic-pituitary-adrenal(HPA)axis,aswellas poten-tiatedandprolongedHPAresponsestostressors(reviewedin

VandenBerghetal.,2005;Talgeetal.,2007;Cirullietal., 2009; Flinn et al., 2011). While the behavioraleffects of prenatalstressmayvaryacrossspeciesandbetweenmales andfemales, thebig picture shows a consistenttendency towardhyper-responsivity.Asanexampleofamorecomplex pattern,theoffspringofguineapigsexposedtosocial stres-sorsduringpregnancyshowfemalemasculinizationandmale infantilization(i.e.,displayofbehavioralpatternstypicalof veryyoungmales).Thesebehavioraloutcomescanbe adap-tivebecauseofthedifferentimplicationsofsocialinstability forthemalesandfemalesofthisspecies(KaiserandSachser, 2005,2009).

etal.,1991; Ellisetal.,2009).Lifehistoryconceptshave beenexplicitly invokedto explain theeffectsofpostnatal care(andtheassociatedexposureto stress)on the devel-opmentofstressresponsivity,bothinrats(Cameronetal., 2005,2009)andinhumans(DelGiudiceetal.,2011;seealso

Cameronetal.,2005;Meaney, 2007).Intandemwith pat-terns ofpostnatal care,prenatal stressmaycontribute to potentiatetheorganism’sdefensiveresponsestothreat,and adjustits reproductiveandmatingtacticstofunction opti-mallyinadangerousenvironment.Ofcourse,behavioraland physiological adaptations to danger can be expected to imposesomecostsontheorganism;however,whatittakes foratraittobebiologicallyadaptiveisthat(onaverage)its fitnessbenefitsoutweighthecosts.Indeed,massivelycostly traitscanbefavoredbynaturalselectionifnotexpressing themwouldbeeven moredetrimentaltofitness(seee.g.,

FrankenhuisandDelGiudice,2012).

1.1.1. Thecomplexrelationbetweenstresshormones

andenvironmentalstates

Astandardassumptioninthefetalprogrammingliteratureis thathigherlevelsofstresshormonessignalamoredangerous and/orunpredictableenvironment(e.g.,Kapooretal.,2006; Talgeetal.,2007;Glover,2011;Sandmanetal.,2012).This assumption, however, may be only partly correct. Fetal exposuredepends on the combined effect of two factors: theamountofstressorsintheenvironmentandthemother’s responsivitytostressors.Studiesofhumansandrodentsshow thatbothdangerousenvironmentsandverysafeonestendto promotethedevelopmentofhighHPAandautonomic respon-sivity,givingriseto a U-shaped relationbetween environ-mentalqualityandstressresponsivity(e.g.,Ellisetal.,2005; Macrı`etal.,2007,2009; Gunnaretal.,2009;Del Giudice etal.,2012;forreviewsseeEllis etal.,2006;DelGiudice etal.,2011).Thelikelyadaptivefunctionofhighresponsivity in‘‘good’’environmentsistoincreasetheorganism’s open-nesstoopportunitiesandresources,forexampleby enhan-cingsocial learning andaffiliation(see Del Giudiceetal., 2011).As a consequence,mothers raised in safe and pre-dictableenvironmentsshouldrespondstronglytomild stres-sors,thus producing comparatively highamounts ofstress hormoneseven in absence of significant danger or uncer-tainty.Furtherimplicationsofanonlinearrelationbetween environmentalstatesandstressresponsivitywillbediscussed inSection3.

Whilethesefindingschallengethestandardassumptions ofprogrammingmodels,theydonotimplythatthesignals reachingthefetuslookexactlythesameinsafeand danger-ousenvironments.Eveninpresenceofacurvilinearrelation between environmental quality and stress responsivity, responsivemothersinsafeenvironmentsshouldexperience infrequentandshort-livedbouts of HPA/autonomic activa-tion, whereas mothers in truly dangerous environments shouldexperience(andtransmit)asteadierandmorechronic patternofphysiologicalactivity(DelGiudiceetal.,2011). Thus,thefetusmaystillbeabletodiscriminatebetweenthe two scenariosby separating the long-term andshort-term components of hormonal signals. Indeed, there is some evidencefromrodentstudiesthattheplacentalmechanisms responsibleforfilteringoutmaternalcortisolrespond differ-ently to acute versus repeated stressors (Welberg et al., 2005).

1.1.2. Prenatalstressinhumans

In human children, the outcomes of prolonged, intense maternal stress during pregnancy include higher anxiety and fearfulness, temperamental difficulty, impulsivity, reduced executive functions, impaired attention, higher aggression and risk-taking, and increased basal activity andresponsivityoftheHPAaxis(reviewedinGloveretal., 2009b; Glover, 2011; Pluess and Belsky, 2011). Vigilance, impulsivity,andcompetitiverisk-takingcanallimprove fit-nessinadangerous,unpredictableworld(Talgeetal.,2007; Glover,2011;PluessandBelsky,2011);asnotedabove (Sec-tion1.1.1),highstressresponsivitycanalsobeadaptivein particularly safe and protected environments (Boyce and Ellis,2005;Ellisetal.,2006;DelGiudiceetal.,2011).While most empirical studiesin humans lack controls for shared geneticeffectsinmothersandchildren,agenetically infor-mativestudybyRiceetal.(2010)confirmedthatthe asso-ciation between maternal stress during pregnancy and children’sanxietyandantisocialbehaviorisenvironmentally mediated.

Otherstudieshavedetectedassociationsbetweensevere stressandincreasedriskforautism,schizophrenia, attention-hyperactivity disorder, and impairedcognitive ability (e.g.,

Bergman et al., 2007; Khashanet al., 2008; Kinneyet al., 2008).These outcomesaremoredifficultto interpretinan adaptive light, and may simply reflect the damaging side effectsofearlystress.Glover(2011)recentlyspeculatedthat autism,cognitiveimpairment,andattention-hyperactivity dis-ordermaybringadaptivebenefitstosocialgroupsbyincreasing cognitive diversity;however, suchgroup-selection scenarios areproblematicfromanevolutionarystandpoint,andwillneed additionalunpackingbeforetheycanbeacceptedasplausible (seeWestetal.,2007;GardnerandGrafen,2009).Finally,any evolvedmechanismcanloseitsadaptivefunction—oreven become maladaptive— if theenvironment undergoesrapid changes;accordingly,sometheoristshavesuggestedthatfetal programmingbymaternalstressmaynolongerbeadaptivein modernenvironments(e.g.,Kapooretal.,2006;Talgeetal., 2007;CottrellandSeckl,2009).Thiscaveatdoesnotcontradict theviewthattheevolvedfunctionoffetalprogrammingisto improvethefitnessprospectsofthefetus,byhelpingmatchits phenotypetothecharacteristicsoftheexternalenvironment (Sandmanetal.,2012).

1.2. Conflictinfetalprogramming

consequencesof behavioral and physiologicalprocesses. As such,theydonotimplyconsciousdeliberationortheexplicit representation of goals. Likewise, any organism (including those without a nervous system) canbe usefullydescribed

as if it were following a certain‘‘strategy’’ (or ‘‘tactic’’), irrespectiveoftheexactnatureoftheprocessesinvolved.

Theevolutionarylogicunderlyingtheconflictofinterest betweenparentsandtheiroffspringisexplainedby parent-offspringconflicttheory(Trivers,1974;Parker,1985;Parker etal.,2002;Royleetal.,2004.SeeSchlomeretal.,2011,for anon-mathematicaloverview).Thekeymessageof parent-offspring conflict theory is that, in most organisms, the geneticinterestsofparentsandoffspringareonlypartially overlapping.Wheneveragiventraitorbehaviorresultsina cost to theparent anda benefit to the offspring(or vice versa), parent and offspring are expected to ‘‘disagree’’ aboutthe optimallevelofexpressionof thattrait. Stated otherwise,thelevelofatraitthatmaximizestheparent’s fitness will differ from the level that maximizes the off-spring’sfitness,resultinginabiologicalconflictofinterest aboutthetrait/behaviorinquestion.

Thelogicofparent-offspringconflictiseasytoillustratein the case of parental investment (e.g., food provision). A parenthasthesamedegreeofrelatednesstoallitsoffspring (i.e., a gene in the parent has a 50% chance of being transmittedtoeachoffspring),andtherefore—allelsebeing equal—willmaximizeitsfitnessbyinvestingequallyineach of them. However, an offspringis moreclosely related to

itself thanto itssiblings;ageneinanoffspringhasonlya probability,usually50%orless,ofbeingpresentintheother offspring (present or future) that compete for parental resources.Thus,naturalselectionwillfavorthoseoffspring who increasetheir share of resources above the parental optimum. In pregnancy, for example, there is an intrinsic conflictaboutfetal growth(seeSchlomer etal.,2011):in ordertomaximizeitsownbiologicalfitness,thefetusshould trytoextractmorenutrientsfromthemother(andgrowata fasterrate)thanisoptimalforher,whilethemothershould trytoreducetheflowofnutrientstothefetus,keeping it belowitsoptimalgrowthrate.

1.2.1. Prenatalconflicts

In a landmarkarticle, Haig (1993) examinedmother-fetus interactions in the light of parent-offspring conflict. By adopting thiscounterintuitive perspective, hewas ableto explain anumberofpuzzlingfactsabout theregulationof maternal physiology during pregnancy. For example, the hormoneinsulinreducesthebloodconcentrationofglucose. Inhumanmothers,insulinlevelsrisedramaticallyduringthe thirdtrimester,butatthesametimemothersdevelopstrong insulinresistance.Insulinresistancecanhavesevereadverse effectsforthemother, bothduringpregnancyandlaterin life.Theparadoxicalcombinationofinsulinhyperproduction andinsulinresistancedoesnotmakesenseuntilonerealizes that insulin resistance is actually induced by the fetus, through secretion of placental hormones, as a way to increaseglucoseconcentrationinthebloodflow.Themost plausiblecandidatehormonesareplacentallactogen(hPL,as suggested by Haig, 1993) and placental growth hormone (hPGH; see Haig,2008). Increasedinsulin secretion, then, canbeseenas a(costly)maternalcountermeasureagainst fetalmanipulation.

Anothercrucial conflict inpregnancy concerns the reg-ulation of maternal blood pressure. Cardiac output rises dramaticallyearlyinpregnancyandremainselevateduntil thethirdtrimester;atthesametime, vasodilationoccurs, loweringarterialresistance duringthefirstandsecond tri-mester.Theneteffectisadropinbloodpressureinearlyand mid-pregnancy,withpressuretypicallyrisingagainduringthe thirdtrimester.Vasodilationappearstoputanunnecessary strain on the mother’s heart; an efficient system would increasebloodsupplytotheplacentabyincreasingarterial resistance(vasoconstriction).Infact,maternalvasodilation can be explained as a defense of mothers against fetal releaseoffactorsthatinduceconstrictionofmaternal ves-sels.Sinceplacentalarterieshaveextremelylowresistance, increasing maternal blood pressure benefits the fetus by directlyincreasing placentalflow.The sideeffects of this physiologicaltug-of-warincludetheoccasionalriskof fetal-inducedmaternalhypertension(seeHaig,1993,2007).

The examples just reviewed aptly illustrate four key featuresofmother-fetusconflict.Firstofall,thefetus(or, moreprecisely,the fetoplacental unit)is an activeplayer rather than a passive recipient of maternal ‘‘decisions.’’ Second,theescalationofconflictoftenresultsinthe evolu-tionofconvoluted,costly,andapparentlywasteful physio-logicalmechanisms.Thefunctionofsuchmechanismscannot beproperlyunderstoodoutsideof anexplicitconflict per-spective.The intensity ofthecoevolutionary ‘‘arms race’’ betweenmothersandfetusesisreflected intheamazingly rapidevolutionofplacentalgenesinrodentsandprimates (Haig,2008;Houetal.,2009;Chuongetal.,2010).Third,the interplay between maternal and fetal manipulation and countermeasurescanmaskthetrueintensityofthe under-lying conflict: despite the strong competition between motherandfetusintheregulationofglucoseconcentration, theactual concentration remainswithin relativelynarrow limitsmostofthetimes.Finally,fetalattemptsat manipula-tionofmaternalphysiologytypicallyoccurthrough produc-tionofplacentalhormones(Haig,1996).

1.2.2. Prenatalstressasaconflictarena

Maternal stressduringpregnancy increasesanxiety, impul-sivity,andHPAresponsivityintheoffspring,withcascading effectson a wide range of traits and behaviors including exploration,dispersalfromthenatalenvironment, intrasex-ualcompetition,andsoforth(seeMeaney,2007;DelGiudice etal.,2011;PluessandBelsky,2011).Ifthese traits/beha-viorsdifferentiallyaffectthefitnessofmotherandoffspring, themotherwillbenefitbydistortingthesignalssheissending to the fetus, so as to move its developmental trajectory closerto herownoptimum.This isthecrucialdilemma of prenatal stress:the fetusis goingto receivea mixtureof usefulinformation andmanipulative signals, andneeds to striketherightbalancebetweenthebenefitofinformation andthecostofmanipulation.Thegeneralproblemhasbeen recognizedforsometimebyevolutionarybiologistsstudying maternaleffects(e.g.,LoveandWilliams,2008;Uller and Pen,2011).

programming.Nepomnaschyandcolleagues (2006;seealso

Flinn et al., 2011) applied a conflict perspective to the abortogeniceffectsofmaternalstressinhumans.Sincestress increasesthelikelihoodofmiscarriageduringthefirstweeks of pregnancy, the mother’s stress response is potentially dangerousfor thefetus.Nepomnaschy andcolleagues sug-gestedthat,over time, a‘‘defensivefetus’’activelygains controlof its own gestation by reducingthe efficiency of maternalabortivemechanisms,sothatafterthefirstweeks of pregnancy even severe stressors are unable to induce miscarriage.Despitethesepromisingleadsintheliterature, theconflictsofinterestinvolvedintheregulationofprenatal stresshaveremained virtuallyunexplored.Thegoalofthe followingsectionsistomakeafirststepinthisdirection.

2.

Theoretical

insights

2.1. Whatistheconflictabout?

Fromaconflictperspective,thefirst questiontoaskabout prenatalstressis,domaternalandfetal interestsdiverge? Andifso,why?Theoutcomesofprenatalstress(Section1.1) aresovariedthatparent-offspringconflictmightpotentially ariseaboutanyoneofthem.Whilemanyoftheseoutcomes can be seen as manifestations of underlying reproductive strategies,toolittleisknownabouttheeffectsofoffspring’s reproductive strategies on parental fitness to formulate plausiblehypothesesinthisrespect.However,thatofspecific behavioraloutcomesmaynotbethemostproductivelevelof analysis.Instead,prenatalstresscanbeexpectedtoreliably elicitparent-offspringconflictbecauseofitscrucialrolein thedevelopmentofpostnatalplasticity.

Postnatal plasticityistheextent towhichthepostnatal environmentcanshapeormodifytheoffspring’sphenotype. Moreformally, the developmentalreactionnorms of more plasticindividualshaveasteeperslopeandcoverabroader phenotypicrange(seeViaetal.,1995;West-Eberhard,2003; StampsandGroothuis,2010).Inhumandevelopmental psy-chology,postnatal plasticity hasbeen characterized (with somewhat different theoretical implications) as biological sensitivitytocontext(Boyce andEllis,2005) and suscept-ibility to environmental influences (Belsky, 1997, 2005; Belsky et al., 2007; Ellis et al., 2011). In species with prolongedmaternalcare, themother constitutes a funda-mentalpartofthepostnatalenvironment;andbydefinition, highpostnatalplasticityimpliesincreasedsusceptibility to theeffectsofmaternalbehavior(includingfeeding,caring, protection,teaching,andso forth).Therefore,themother wouldbenefit ifshewereable—by whatevermeans—to increasetheoffspring’s susceptibility to her own behavior beyond the offspring’s optimum, as this would give her increased leverage in subsequent instances of parent-off-springconflict.Conversely,andallotherthingsbeingequal, theoffspringshouldavoidbecomingtooplasticand suscep-tibleto maternalinfluence.Inrecentyears, abundant evi-dencehasaccumulatedshowingthatstressresponsivityand negativeemotionalitydramaticallyincreaseearlyplasticity in humans, and the same may plausibly apply to other mammals;asaconsequence,earlypostnatalplasticitycan bemodulatedbythelevelofstressexperiencedbythefetus (Belsky and Pluess, 2009; Pluess and Belsky, 2011), thus

setting the stage for pervasive parent-offspring conflict aboutfetalprogramming.

Insummary, ifprenatal stresscontributesto determine postnatalplasticityintheoffspring,motherandfetuscanbe expectedtohaveconflictingoptimaabouttheleveloffetal exposuretostresshormones.Specifically,themothershould favor higher exposure than is optimal for the fetus. This statementshouldbeinterpretedcarefully,asthepredicted conflictisaboutrelativeexposurelevels.Forexample,both motherandfetusmaybenefitfromhighfetalexposureina dangerousenvironment(Section1.1);however,theoptimal exposurelevelofthefetus—whilehighinabsoluteterms—is expected to be somewhat lower thanthat of themother, resultinginsomedegreeofmother-fetusconflict.

2.1.1. Prenatalprogrammingofpostnatalplasticity

There is remarkable individual variation in infants’ and children’ssusceptibility toenvironmental influences.Some children are barely affected by experiences in their first yearsoflife,whileothersrespondstronglytotheconditions they encounter early on (e.g., by developing aggressive, antisocialbehaviorsinresponsetoharshandabusive parent-ing).Infact,highlyplasticchildrenshowenhancedresponses tobothpositiveandnegativefeaturesoftheenvironment; for example,thesame childrenwho wouldbecomehighly aggressivewhenraisedinharshfamiliestendtobecomeeven

lessaggressivethanotherchildrenwhenrearedinpositive, safe environments.Thebidirectionalnature ofplasticityis well captured by the phrase ‘‘for better and for worse’’ (Belsky etal., 2007). While much remains to be learned, substantial progresshasrecently been madeinour under-standing of what makes infants andchildren more or less susceptible to environmental effects, including those of parentalbehavior.Atthegenotypiclevel,ithasbeenpossible toidentifyanumberofalleles(ingenessuchasDAT1,DRD2,

DRD4,5HTT,andMAOA)thatcontributetoincreased plasti-city.Thecombinedeffectofallelicvariationinthosegenes ondevelopmentalreactionnormshasturnedouttobe sub-stantial(Belskyetal.,2009;BelskyandBeaver,2010).Atthe phenotypiclevel,ithasbecomeapparentthathighlyplastic infantsandchildrenarecharacterizedbynegative emotion-ality,temperamentaldifficulty,andhighresponsivityofthe HPAaxis(e.g.,Boyceetal.,1995,2006;Kochanskaetal., 2007;Obradovicetal.,2010;O’Nealetal.,2010;reviewedin

Belsky, 2005; Belsky and Pluess, 2009; Ellis et al., 2011; PluessandBelsky,2011).Indeed,oneofthemainfunctions ofthestressresponsesystemistomodulatetheorganism’s openness totheenvironment, which—atleastinthefirst yearsoflife—includessusceptibilitytothebehaviorofone’s parents (Ellis etal., 2006; Del Giudice etal., 2011). It is probablynocoincidencethatplasticity-relatedgenestendto be involved in serotonergic and dopaminergic pathways, whichinturninteractdeeplyandreciprocallywiththestress responsesystem(Porteretal.,2004;vanGoozenetal.,2007; Gotlibetal.,2008;Alexanderetal.,2011).

rearing (Stevens et al., 2009). Temperamentally reactive mice also show larger stress-induced changes in behavior (Veenemaetal.,2004).InastudyofratsbyFrancisetal. (1999), the biological offspring of dams who engaged in reducedmaternalcare(i.e.,damswithafearfulandreactive phenotype;seeCameronetal.,2005;Meaney,2007)became more fearful than other rats when reared by low-caring mothers, but less fearful when reared by high-caring mothers. This suggests thatthe geneticand/or epigenetic factorspasseddownbyanxiousdamsincreasetheplasticity ofpupsandtheirresponsetovariationinrearingconditions. Consistentwiththishypothesis,theoffspringoflow-caring mothers showed enhanced behavioral responses to an enriched environment compared with the offspring of high-caringmothers(ChampagneandMeaney,2007).Finally, thegenesthatregulateplasticityinprimates(Stevensetal., 2009)androdents(Sachseretal.,2011)arefoundin stress-relatedserotonergicanddopaminergicpathways,thesame pathwaysthatregulate plasticityinhumans. Forexample, knockoutmicelackingexpressionoftheserotonin transpor-ter gene (5HTT) are characterized by exaggerated HPA/ sympathetic reactivity. As expected, they are also more developmentallyplastic:theygrowupmoreanxious follow-ingpoormaternalcare(Carolaetal.,2008;seealsoHeiming etal.,2009,2011),aswellasinresponsetomildearlylife stress(Carrolletal.,2007).

Insummary,stressresponsivityandemotionalreactivity arekeypredictorsofearlypostnatalplasticity,inhumansand (mostlikely) othermammals as well.More responsive off-springaremoresusceptibletotheeffectsofawholerangeof (generallyadaptive)maternalbehaviors,whichcontributeto shapetheirdevelopmentaltrajectories.Sinceprenatalstress increasesHPAresponsivityandemotionalreactivity, postna-tal plasticity canbe programmed by prenatal exposureto stress, especially in those offspring who carry an already susceptible genotype(Belsky andPluess, 2009; Pluessand Belsky,2011;Pluessetal.,2011).Therefore,theprediction canbemadethatmothersshouldtrytoamplifythe physio-logicaleffectsofprenatalstress,whilefetusesshouldtryto reducethem.

The resulting conflict should be especially intense in species with extended postnatal care, in which a mother cangreatlybenefitfromenhancingheroffspring’splasticity. This, of course, does not rule out the existence of other sourcesofconflictconcerningspecificoffspringtraits(e.g., competitiverisk-taking,reproductivetiming,ortheriskfor schizophrenia). However, in species with sufficiently long periodsofpostnatalcare,amotherhasampleopportunity toshapeheroffspring’sbehaviorafterbirth;thisshouldlimit the comparative benefits of directly programmingspecific behavioralfeaturesduringgestation,andincreasethe ben-efitsofenhancingoffspringplasticityacrosstheboard.

Even ifparent-offspring conflict is an inescapable out-comeofevolutionarydynamics,theintensityofconflictcan beincreasedorreducedbyvariationinecologicalconditions. Specifically,conflictismaximizedwhen(a)genetic related-nessbetweensiblingsislow(i.e.,thereisahighprobability that one’s siblings are in fact half-siblings), and (b) the interdependence between the reproductive success of mothersandfathersislow(Trivers,1974;LesselsandParker, 1999;seeSchlomeretal.,2011).Theseconditionsareboth met in mating systems characterized by promiscuity and

unstablepairrelationships.Anintriguingimplicationisthat theintensityofconflictaboutpostnatalplasticitymayvary dependingonthelevelofstressexperiencedbythemother.If stressful,dangerousenvironmentsarealsocharacterizedby unstablepair-bonds andpromiscuous mating (asin human societies;seeBelsky etal.,1991;GangestadandSimpson, 2000;Quinlan andQuinlan, 2007;Ellis et al.,2009), then fetusesshouldtendtoputupastrongerresistanceto mater-nalprogrammingwhentheyareexposedtochronicallyhigh levelsofstresshormones.Conversely,whenhormonalsignals areconsistentwithasafe, protectedenvironment, fetuses shouldacceptmaternalprogrammingtoagreaterextent.

To conclude this section,it is important to stress that parent-offspringconflictstandsonabackgroundofextensive mother-offspringcooperation.Iamnotarguingthatconflict dominatesprenatalinteractions,northatthesignalscarried bystresshormonesareentirelymanipulativeanddevoidof useful information; in other words, the perspective pre-sentedhereisanextensionofthestandardadaptivemodel, notanalternativetoit.Themotherbenefits(uptoapoint) fromprovidingthefetuswithcorrectinformationaboutthe state ofthe environment, and thefetus benefits (upto a point)fromlettingmaternalsignalsshapeitsdevelopmental trajectory.Theintricacyofmother-fetusinteractionsstems fromthisdelicatebalanceofcooperationandconflict(Haig, 1993).

2.2. Fetaltactics, maternaltactics, andconflict outcomes

Providedthatmother andfetushave conflictinggoals con-cerningfetalexposuretostresshormones,whattacticsdo they have at their disposal? From the perspective of the fetus,anobvioustacticis tofilter outmaternal hormones beforetheyreachitsbrain.Ofcourse,completelyfiltering outmaternalsignalswoulddeprivethefetusofuseful infor-mation, so the filtering must be only partial. A possible maternalcountermeasuretothistacticistoproducelarger quantitiesof stresshormones,even if thismaycarry con-siderablephysiologicalcosts(seeMcEwen,1998;McEwenand Wingfield,2003;Ganzeletal.,2010).Anotherpossible coun-termeasure(andalesscostlyone)istodirectlyinterferewith fetalfilteringmechanisms.Filteringisapassivetactic,but thefetuscanalsoengageinactiveformsofmanipulation.For example,thefetuscouldtrytoreducethemother’sstress responsivity,soastodampenthehormonalpeaksthatfollow astressfulevent.Inturn,themothercouldtrytoresistfetal manipulation, either by interfering with fetal signals (a cheaperoption)or byenhancing thereactivity ofherown stressresponsesystem(acostlieroption).

Alloftheabovetacticsandcountermeasuresconcernthe

intensityofmaternalsignals,butsignalaccuracyisanother parameterthatmightbemanipulatedbythemotherand/or thefetus.In principle,themothermight notjustbecome more(orless)responsivetostress,butalsomoreuniformly responsive to different types of stressful events. This, of course,would make the signal less informative about the actualstateoftheenvironment.

thecurrentstateoftheenvironmentandusesthis informa-tiontogenerateasignal;theoffspringdoesthesame,and thetwosignalsjointlydeterminetheoffspring’sphenotype. Theresultingphenotypeaffectsthefitnessofbothmother andoffspring; however,the optimal phenotype in a given environmentisnotthesameforthetwoactors,whichsets thestageforparent-offspringconflict.Themothercanuse hersignalstomanipulatetheoffspring’sphenotype,andthe fetuscanrespondbyfilteringmaternalsignals(i.e., discount-ingthembysomeamount).

Inbrief,modelresultsshowthat:(a)ifsignalsare cost-free,andtheoffspringisconstrainedinitsabilityto filter maternal signals, the resulting phenotype will match the mother’s optimum. (b) If signals are cost-free, and the filtering mechanism can evolve without constraints, the resulting phenotype will match the offspring’s optimum, andacoevolutionaryarmsracewillensuebetweenthegenes expressed by the mother and the offspring; (c) this will happeneveniftheoffspringhasnoindependentinformation abouttheenvironment,becausematernalsignalsstillcarry useful information. However, (d) if signals are costly, the outcomeoftheevolutionaryprocesswilloftendepartfrom both optima;and finally,(e) both parties benefit from an accurateassessmentoftheenvironmentbythemother.

Whatdotheseresultssayaboutconflictinfetal program-ming?Therelevantscenarioisoneinwhichstresshormones functionascostlysignals,andthefetushasnoindependent informationabouttheexternalenvironment.Firstofall,the modelsuggeststhatfetuseswilltendtoevolvemechanisms thatfiltermaternalhormones(and/orreducetheirlevelby othermeans),becauseafetusequippedwithsuch mechan-ismscandramaticallyshifttheconflictoutcometowardits own optimum. Overevolutionary time, the resulting arms racebetweenmaternalandfetalgenesmayleadtomassive productionofstresshormonesbythemothercoupledwith nearlycompletefilteringofmaternalhormonesbythefetus (seeUllerandPen,2011).Second,becauseofthe physiolo-gicalcostsofstresshormones,theconflictoutcomecanbe expectedtodepartinsignificantwaysfromboththe mater-nalandthefetal optimum.Inotherwords,conflictsabout

programmingprobablyhaveno ‘‘winner’’inastrictsense. Finally,tacticsthatreducetheaccuracyofmaternalsignals (for example by making maternal physiology uniformly responsivetostressfulandnon-stressfulevents)canberuled out,astheyareultimatelydetrimentaltobothmotherand fetus.The range ofplausible maternalandfetal tacticsin fetalprogrammingissummarizedinFigure1.

3.

Physiological

insights

Thanks to the insights gained in the previoussection, the mechanisms of prenatal stress can be examined with an updated set of conceptual tools. While the focus of this sectionisonhumanpregnancy,thesameprinciplescanbe appliedtothephysiologyofotherspecies.Itshouldbenoted that species differences involve not only the details of physiologicalfunctioning, but also the details ofhow par-ent-offspring conflict is played out — as, for example, in animalsthatgivebirthtolittersinsteadofsingletons(seefor exampleHaig,2008).

3.1. Maternalstressandfetal programmingin humans

3.1.1. Pregnancy-relatedchangesinstressphysiology

mostoftheCRHproducedbytheplacentaisinactivatedby CRH-bindingproteins(CRH-BP)inthemother’sblood( Smir-nakiandMagiakou,2006),suggestingthatplacentalCRHand maternalCRH-BPmaybeinvolvedinaconflictaboutfetal nutrition. In the placenta, 50—90% of maternal cortisol is inactivatedbytheenzyme11b-hydroxysteroid dehydrogen-ase type 2 (11b-HSD2) before it reaches fetal circulation (Section3.2.1);as a result,cortisol concentration infetal bloodismuchlowerthaninmaternalblood(typicallyaround 10—20%ofmaternallevels;seeMurphyetal.,2006).

Pregnancy-relatedchangesalsooccurinthesympathetic nervous system. The available dataindicate thatthe epi-nephrine(E)responsetostressorsbecomesblunted,whereas the norepinephrine (NE) response seems to remain intact (Russelletal.,2008).Accordingly,fetusesatdifferentstages of gestation show real-time cardiac responses to induced maternalstress;interestingly,fetalresponsesappeartobe even stronger in late pregnancy than at mid-pregnancy, despite theincreasingly bluntedresponse of the maternal sympatheticsystem(DiPietroetal.,2003).

Becauseofthecombinedeffectofhypercortisolismand bluntedHPAresponsivity,cortisollevelsinpregnancycanbe expected to be only weakly correlated with the mother’s subjective distress. Empirical findings are inconsistent (O’Donnelletal.,2009),withsomestudiesfindingapositive correlation between perceived stress and cortisol levels (e.g.,Wadhwaetal.,2001;Diegoetal.,2006;Rothenberger etal.,2011)andothers findingnosuch correlation(Sarkar etal.,2006;Davisetal.,2010;Kaasenetal.,2011).Noneof these studies measured the relation between perceived stressand cortisol responsivity;however,cortisol peaks in responsetostressors(andtheirpatternofoccurrenceover time) carrymoreinformation aboutthe environment than thebasalcortisollevel,andcanthusbeexpectedtobemore directlyinvolved infetal programming.The availabledata indicatethat,despitephysiologicalblunting,HPAresponses can be elicited even in late pregnancy (de Weerth and Buitelaar, 2005; Nierop et al., 2006; de Weerth et al., 2007; Nierop et al., 2008). At a more fundamental level, the relation between environmental conditions and HPA responsivityisnonlinear (Section1.1.1);testingfor simple linearcorrelationsbetweendistressandindicesofHPA func-tioningmayprovideincompleteormisleadinganswers.

3.1.2. Hormonalmediatorsoffetalprogramming

Theinformationtransferbetweenmotherandfetusinvolves theactionofmultiplehormonalsystems.Onelikelymediator offetalprogrammingismaternalcortisol,andextensivedata fromnonhumananimals supportthisprediction(Matthews, 2002). Cortisol secretion entails substantial physiological costs, and peaks in response to unpredictable and/or uncontrollable events. For these reasons, cortisol carries high-quality,reliableinformationabouttheseverityof exter-nalstressors(seeDelGiudiceetal.,2011;Koolhaasetal., 2011),makingitanidealsourceofinformationforthefetus. Maternal cortisol levels havebeen shownto predict beha-vioralreactivityandHPAfunctioning innewbornsand chil-dren(Guttelingetal.,2004,2005;Davisetal.,2010).Further indirectsupportforakeyroleofcortisolcomesfromstudies of licorice exposure in pregnancy. Glycyrrhizin (the main sweet-tasting compound in licorice roots) inhibits the enzyme 11b-HSD2, thus reducing the effectiveness of

placentalfiltering andexposing the fetus to an increased influxofcortisolfrommaternalcirculation.Predictably,the childrenofmotherswhoconsumedhighamountsoflicorice duringpregnancyshowelevateddiurnalcortisolandhigher HPAresponsivitytostressors(Ra¨ikko¨nen etal.,2010).

Theroleofcortisolinhumanfetalprogramminghasbeen questionedbysomeauthors.Davisetal.(2010)arguedthat cortisolisunlikelytomediatetheeffectsofmaternalstress, becauseoftheweak andinconsistentassociationbetween self-reporteddistressandcortisollevelsinpregnantwomen. However,therelationbetweenenvironmentalstressandHPA functioning may not be as straightforward as is often assumed;moreover,thephysiologicalchangesofpregnancy makecorrelationsmoredifficultto detect (Section3.1.1). Otherresearchers(Talgeetal.,2007;Brunton,2010)argued that the reduced HPA responsivity of the mother andthe filteringactionof11b-HSD2severelylimitfetalexposureto cortisolpeaks,thusmakingcortisol problematicasa med-iatoroffetalprogramming.Thiscriticismcanbeaddressedin twoways.First,asnotedbyFlinnetal.(2011),thedifference incortisolconcentrationbetweenmaternalandfetal circu-lation is so large that even small fluctuations in maternal cortisol can exert significant effects on fetal physiology. Second,andmoreimportant,inthefollowingsectionsIwill showthatbothcortisolfilteringandHPAdampeningcanbe understoodasadaptivefetal tactics.Seen inthis perspec-tive,theproblemevaporates:itispreciselybecauseofthe crucialroleofcortisolinfetalprogrammingthatthefetushas evolvedmeanstolimititsownexposuretomaternalcortisol peaks.Inother words,Isurmisethatthemechanismsthat limittheprogrammingeffectivenessofmaternalcortisolare not‘‘bugs’’butadaptive,evolvedfeaturesoffetal physiol-ogy.

Inaddition to cortisol, a number of stress-related hor-monesandneurotransmittershavebeenproposedaspossible mediators of fetal programming. The main candidates are maternalandplacentalCRH(deWeerthandBuitelaar,2005; Flinn et al., 2011); adrenocorticotropic hormone (ACTH;

synergisticallyinfetal programming; ifso, theinteraction betweenthetwohormonesmightprovemoreimportantand predictivethantheirindividualeffects.

3.2. Physiological conflictsinfetal programming

3.2.1. Placentalfilteringmechanisms

AsoutlinedinSection3.1,theplacentaldehydrogenase11b -HSD2inactivates50—90%ofmaternalcortisolbyconverting ittocortisone.Theexpressionof11b-HSD2intheplacenta increases linearly during pregnancy, and is stimulated by cortisol(van Beek etal.,2004; Schoof etal., 2001). Itis commonly assumed that 11b-HSD2 serves to protect the fetus from excessive cortisol exposure and the adverse effectsofprenatalprogramming(e.g.,BruntonandRussell, 2011).Itismoredifficulttoexplainwhythedecidua(i.e., thematernaltissueindirectcontactwiththeplacenta)does not express 11b-HSD2, but rather high amounts of the enzyme11b-HSD1(11b-hydroxysteroiddehydrogenasetype 1).Inlivetissue,11b-HSD1functionsmainlyasareductase: it converts cortisone to cortisol, thus counteracting the effectsof11b-HSD2(Krozowskietal.,1999;Seckl,2006). Inotherwords,fetaltissueinactivatescortisoltocortisone, whilematernaltissueconvertscortisonebacktocortisolat theinterfacewiththeplacenta.Thisapparentlyparadoxical mechanismmakesperfectsensefromthevantagepointof parent-offspringconflict,showingtheinterplaybetweena fetaltactic(i.e.,cortisolfiltering)andamaternal counter-measure (i.e., interference with placental filtering; see

Figure2).Inadditionto11b-HSD2,theplacentaexpresses severalotherenzymesthat metabolizecortisoland corti-sone, including 3a/3b-HSD, 5b-reductase, and 20b-HSD (Pasqualini,2005).

Placental filtering of catecholamines (NE in particular) canalsobeseenasafetaldefensefrommaternal manipula-tion.Indeed,theexpressionlevelsoftheNETgene(SLC6A2) andthe11b-HSD2gene(HSD11B2)arepositivelycorrelated (Ponderetal.,2011).BymediatingNEuptake,NET substan-tiallydecreasesNEconcentrationinfetalcirculation(Ponder etal.,2011).Accordingly, wecanpredicttheexistenceof maternal mechanisms that counteract the effectsof NET, resultinginincreasedNElevelsinfetalblood.

3.2.2. Fetalmanipulationofmaternalphysiology

During pregnancy,thefetoplacentalunitengagesinactive manipulationofmaternalphysiology,includingthemother’s HPAaxis(Figure2).AsnotedinSection3.1,placentalCRH contributestoelevatethebasallevelofmaternalcortisol. However, conflictabout programmingshould also leadthe fetus to reduce the mother’s responsivity to stressors. Indeed, during pregnancy the maternal HPA axis becomes hyporesponsive,largelyasaresultoffetalmanipulation.The mainphysiologicalagentsoffetalmanipulationare proges-terone (P4), its neuroactive metabolites pregnenolone (PREG)andallopregnanolone(ALLO),andhumanplacental lactogen (hPL), a hormone with high affinityfor prolactin receptors.

Atthebeginningofgestation,P4isproducedbythecorpus luteumfrommaternalLDLcholesterol,followingstimulation by placentalgonadotropin(hCG). Direct placental produc-tion of P4 starts at about 8 weeks; over the course of pregnancy,P4serumlevelsincreasebyabouttwentytimes (Luisi et al., 2000; Smirnaki and Magiakou, 2006). In the classicview,maternalLDLcholesterolisthemainsubstrate for placentalP4 production(e.g.,Tuckey,2005);however, recentdiscoveriespointtoanadditionalmetabolicpathway

for P4 production, onethat isindependent from maternal substrates. PlacentalCRH stimulatessecretion of sulfated steroidsbythefetalzoneofthefetaladrenalgland;adrenal steroidsareconvertedtoP4intheplacenta,andP4isthen releasedintomaternalcirculation(Hilletal.,2010a,b).This pathwayeffectivelydecouplesplacentalP4productionfrom theavailabilityofmaternalsubstrates.Inaconflict perspec-tive, thisallows thefetus to produce high amounts of P4 whilecurtailingopportunitiesformaternalcountermeasures (e.g.,themothermightdown-regulatecholesterol produc-tioninorderto limitP4concentration).Itisinterestingto notethatP4reducestheactivityof11b-HSD2(Murphyetal., 2006),andmightthusinterferewithcortisolfiltering; how-ever,placentaltissueisabletodegradeP4viatheactionof 17b-hydroxysteroiddehydrogenasestype5and7(17b-HSD5 and17b-HSD7;seeHilletal.,2010b).

Inthe brain,P4 isconverted to PREG andALLO. These powerful neuroactive steroids exert anxiolytic effects via modulationoftheGABAAreceptor(especiallyathighdoses), withanefficacysimilartothatofbenzodiazepines. Further-more, they directly down-regulate the HPA response by inhibiting transcription of CRH and vasopressin (AVP), a processmediatedbyopioidsignaling(reviewedinDiaz Brin-tonetal., 2008;Brunton andRussell, 2011; Wirth,2011). WhileP4mayhaveminoranxiolyticeffectsviaintracellular progesteronereceptors(Wirth,2011),ithastobeconverted toALLOandPREGinordertoexertmaximalinfluenceonthe mother’sHPAaxis.Intriguingly,thebrain’sabilitytoconvert P4toneuroactivesteroidssuchasALLOismarkedlyincreased inpregnancybecauseofenhanced5a-reductaseactivity;the likelysourcesofincreased5a-reductaseactivityare prolac-tinanditsplacentalanaloguehPL(Brunton,2010).Besides stimulatingP4conversiontoALLO,hPLcansuppresstheHPA responsebybindingtoprolactinreceptorsinthe hypothala-mus(NumanandWoodside,2010).Insummary,theplacental hormones P4 and hPL appear to act synergistically in the maternalbrain,resultinginloweredanxietyanddampened reactivityoftheHPAaxis.

Inaconflictperspective,fetalmanipulationbyP4andhPL raises the question of maternal countermeasures. We can predicttheexistenceofphysiologicalmechanisms counter-actingtheactionofP4,ALLO,PREG,andhPLinthemother’s brain.Possiblecountermeasuresmightinvolve down-regula-tion of the prolactin receptor (PRL-R) and of the GABAA receptor, as well as enhanced metabolism of neuroactive steroids.The opioidsystem isanother plausible targetfor maternaltactics.

Giventheextensiveevidenceforfetalmanipulationofthe mother’sHPAaxis,itisreasonabletohypothesizethatthe fetus may also attempt to regulate maternal secretion of catecholamines.Ontheonehand,GABA-mediatedinhibition (stimulatedbyALLOandPREG)dampensnotonlytheHPAaxis butalso thesympatheticsystem (e.g.,Ulrich-Lai and Her-man,2009).Ontheotherhand,thefetusmayalsobeableto specificallytarget themother’ssympatheticphysiology. As notedinSection3.1.1,sympatheticsuppressionduring preg-nancyappearstobeselective,withabluntedEresponsebut intact NE reactivity. This finding may reflect a complex interplay between maternal and fetal tactics, with the motherultimatelyhavingthe‘‘upperhand’’inNEsignaling (forexamplebecausesuppressingtheNEresponsewouldbe toodangerousforthemother,thefetus,orboth).

3.3. A roleforimprintedgenes?

Nested within the biological conflict betweenmother and fetus,adifferentkind ofconflicttakesplacebetweenthe maternalandpaternalhalvesofthefetalgenome.Thisform ofintragenomicconflictisplayedoutbyimprintedgenes.A geneislabeledasimprintedifitsexpressionisconditionalon theparentoforigin,i.e.,onwhetheritwascontainedinan ovumorasperm.Forexample,theIGF2gene(seebelow)is paternally expressed in humans, and its homologues are paternallyexpressedinvariousmammalsincludingrats,mice andsheep(seeHaig,2004).Thismeansthatalthoughevery individualinheritstwocopiesofthegene,onlythepaternally derived copy is actually expressed, while the maternally derived one is silenced (usually by DNA methylation and histone modification) andhas no effect on theorganism’s development.Imprinted genesare foundin mammals and angiosperm plants; complex imprinting patterns can be observedas,forexample,somegenesshowparent-specific expression only in specific tissues or only during certain phases of development (see Bartolomei and Tilghman, 1997;Wilkins,2008).

The kinship theory of genomic imprinting (Haig, 1997, 2004;WilkinsandHaig,2003)predictsthat,inmostspecies, paternallyexpressedgenesshould‘‘side’’withtheoffspring intheregulationofmaternalinvestment,andevolvesoasto increasethetransferofmaternalresourcestotheoffspring (forexample,by increasing fetal growthrate).Maternally expressed genes are expected to evolve in the opposite direction,thusinhibitingthetransferofmaternalresources totheoffspring.Theconflictbetweenmaternallyand pater-nally derived genes stems from asymmetric relatedness: whenever there is somedegree of multiple paternity,the genetic similarity between siblings born from the same motherislowerfromtheperspectiveofpaternalgenesthan from thatof maternal genes.Ifa givenmaternal allele is presentinoneofthesiblings,thereisa50%chance(assuming Mendelianinheritance)thatthesameallelewillbepresentin anyothersibling;whereas,owingtomultiplepaternity,the chancethat a paternally derived allele will bepresent in anothersiblingislessthan50%.Asaresult,paternalgenes areexpectedtovaluethefitnessofone’ssiblingslessthan maternal genes.The resulting tug-of-war between mater-nallyandpaternallyimprintedgenesisexpectedtoleadto costlymanifestationsofconflictatthephysiologicaland/or behaviorallevel.

Unsurprisingly,severalimprintedgenesareinvolvedinthe physiologyofprenatalconflicts.Forexample,theinsulin-like growthfactorIIgene(IGF2)ispaternallyexpressedinhumans andpromotesfetal growth;IGF2over-expressionresultsin overgrowthsymptomsandisassociatedwiththe Beckwith-Wiedemannsyndrome,aconditioninvolvingprenatal over-growthandenlargedplacenta.Anothergene,H19,is mater-nallyexpressedandhasoppositegrowth-inhibitoryeffects.

expressedIPL,whichishighlyexpressedintheplacentaand, ifinactivated, resultsinplacentalovergrowth.Many other imprinted genes are expressed during fetal development, althoughtheir functionandtheir relevance to the kinship theoryarepresentlylessclear(reviewedinHaig,2004).

3.3.1. Imprintedgenesandfetalprogramming

Given the involvement of imprinted genes in both fetal conflict and fetal programming (e.g., Heijmans et al., 2008),itisnaturalto askwhethertheymaybeimplicated inthephysiologyoffetal programmingby maternalstress, and(ifso)whatspecificrolestheymightplayinthe mechan-ismsreviewedabove.Unfortunately,theamountofrelevant evidenceisextremelylimited.Somefindingssuggestthatthe genecodingfortheb-subunitofhCGmaybeimprintedand paternallyexpressed(Goshenetal.,1994),andthatthehCG receptoronthematernalsidemaybeimprintedtoo(Allen et al., 2003). This is intriguing, since hCG stimulates P4 production in early pregnancy. However, hCG is simulta-neouslyinvolvedinanumberofotherphysiologicalprocesses relatedtoprenatalconflict(e.g.,preventionofspontaneous abortion,inductionofpregnancysickness;seeDelGiudice, 2007); thus, parent-specific expression of hCG may have moretodowiththeregulationofparentalinvestmentthan withstress-relatedprogramming.Thesameinterpretive pro-blemwouldapply to imprintedgenesin thehPL pathway, since hPL mayalso be involved in the regulation of fetal nutrition(Section1.1).

Whilethere isadearth ofrelevantempiricaldata,itis possible to make a preliminary theoretical point. If the theoryadvancedinthispaperiscorrect,thelikelihoodthat imprintedgenes are involvedin fetal programmingshould

decrease to the extent that both parents benefit from enhancedpostnatalplasticity.Whenthemotheristheonly beneficiaryofoffspringplasticity(asinthestandard mam-malian pattern of exclusive maternal care), paternally expressedgenescanbeexpectedtoreduceplasticity,thus favoringtheoffspring’sgeneticinterestattheexpenseofthe mother. This makes it more likely — though by no means certain—thatplacentalgenesinvolvedinfetalprogramming (e.g.,HSD11B2)willevolve apattern ofpaternally biased expression.Inthehumanspecies,however,bothparents(and theirkin)potentiallyinteractwiththeinfant/child,andboth canbenefitfromtheopportunitytoshapehis/herbehavior according to their own interest. This may dramatically reduce interparental conflict about prenatal exposure to stress hormones, and render the evolution of imprinted expressionpatternsconsiderablylesslikely.

Insummary,totheextentthatimprintedgenesareinvolved in prenatal stress physiology, there are grounds to expect adaptivespecies-specificpatternsofimprinting.Theevolution ofimprintingeffectsshould dependat leastinpartonthe relativeamountofcarebyfathers(and/orpaternalkin)and mothers (and/or maternal kin), as well as on the specific effectsoftheoffspring’sreproductivestrategyonmaternal andpaternalfitness(Sections1.1and2.1;seealsoDelGiudice etal., 2010;U´beda, 2008).The expressionpatternof11b -HSD2isaninterestingcaseinpoint.Inhumans,theHSD11B2

geneislocatedonchromosome16.Itspromoterregion com-prisesfourCpGislands(thereisonlyoneintherathomologue); theseCpGislandsarepotentialtargetsformethylation,which represses HSD11B2 transcription both in vivo and in vitro

(Alikhani-Koopaeietal.,2004;Marsitetal.,2012).Thisraises thepossibilitythatHSD11B2mayshowapatternof parent-specificexpression.However,astudybyMcTernanetal.(2001)

foundbiallelictranscriptionofHSD11B2inthreeoutofthree genetically informative placentas. Thus, the available evi-dencesuggeststhatHSD11B2isnotimprintedintheplacenta, whichisconsistentwiththehypothesisthatparentalconflict aboutplasticityisweakorabsentinourspecies.Still,much additionalresearchwillbeneededbeforefirmconclusionscan bedrawn.Theregulationof11b-HSD2expressiondependsona hostofothergenes;forexample,thep38mitogen-activated protein kinase (MAPK) critically up-regulates 11b-HSD2 expressionintheplacentabyalteringmRNAstability(Sharma etal.,2009).Thegenesinvolvedintheregulationof11b-HSD2 expressionmay eventuallyshow imprintingeffects, evenif

HSD11B2itselfdoesnot.Hopefully,theroleofimprintedgenes infetalprogrammingwillbecomeclearerasempiricalfindings accumulate.

4.

Conclusions

4.1. Summaryofthepaper

Inthispaper, Ihaveshownhowaconflictperspective gen-erates novel insights in the logic and physiology of fetal programmingbymaternalstress.Inthe adaptivemodelof programming,maternalstresshormonesprovideuseful infor-mationtothefetus,andthefetusmakesuseofthat informa-tion to adaptively adjust its developmental trajectory. A conflictperspectiveextendsthismodelbyhighlightingthe ways in which maternal and fetal interests diverge, and providesaframeworkformakingsenseofthemanyapparent paradoxesgeneratedbyconflictualdynamics.

countermeasuresagainsttheactionofP4and/orhPL;(c)the existence of a system of fetal manipulation specifically targeted at sympathetic catecholamine signaling, and of thecorrespondingmaternal countermeasures;(d) the pre-dictionthathigherlevelsofbiparentalcare(and/orcareby bothmaternalandpaternalkin)shoulddecreasethe involve-mentofimprintedgenesinthephysiologyofprenatalstress.

4.2. Furtherimplications

Whiletheconsequencesofmother-fetusconflictareusually constrainedbythereciprocalinteractionbetweenopposing physiologicalmechanisms,theremaybeoccasionsinwhich conflict imposes severe costs on one or both parties. For variousreasons(includinghighgenotypicsusceptibilityofthe fetus),thephysiologicalarmsracebetweenmotherandfetus maysometimesescalatebeyondcontrol, resultinginanet fitnesscostforthemother,thefetus,orboth.Forexample, prenatal conflicts in theregulation blood glucose (Section

1.2.1) sometimes end up causing gestational diabetes, a dangerousconditionthatincreasestheriskforlaterhealth problems,bothinthemotherandthefetus(e.g.,Haig,1993; Boney et al., 2005; Feig et al., 2008). This suggests the intriguingpossibilitythatthesevere,fitness-reducing pathol-ogiesassociatedwithprenatalstress—includingautismand schizophrenia—mayactuallyrepresentthesideeffectsof escalated conflict. In other words, the increased risk for severe psychopathology following maternal stress might not be a ‘‘design feature’’ of prenatal programming (as suggestedbyGlover,2011),butratheranoccasional mala-daptive consequence of the conflictual interplay between adaptivemechanismsinthemotherandfetus.Thisseemsa worthy topicoffutureresearch:abetterunderstandingof the sideeffects ofprenatal conflict wouldgreatly help in drawing the line between adaptive and maladaptive out-comesoffetalprogramming(Section1.1.2),andmay even-tuallyleadtodeviseeffectivepreventionstrategies.

Onamoregenerallevel,theconflictperspective devel-opedherehasfar-reachingimplicationsfortheoriesof devel-opmentalplasticity.Belsky(1997,2005;seealsoEllisetal., 2011) proposed that differential susceptibility to rearing (i.e., differential plasticity) is adaptive for both parents andchildren,becauselesssusceptiblechildrenarealsoless likelytosufferfromfitness-detrimentalparentalinfluences. Whileparentsattempttoshapetheiroffspring’sphenotype so asto matchfutureenvironmental conditions,their pre-dictionsinevitablycontain someerror,andsometimesturn outtobemistaken.Forexample,amothermayestimatethat theenvironment willremain safe andresource-richin the foreseeablefuture,andshapethebehaviorofheroffspring accordingly(e.g.,bymakingthemlessvigilantand aggres-sive).Ifthepredictionturnsouttobewrong,however,those offspringwho resistedparental influencewillenjoyhigher fitness than those who let themselves be shaped by the parent. A conflict perspective adds a layer of complexity tothisview,becauseitsuggeststhat—allelsebeingequal— reducedplasticityshouldbenefittheoffspringmorethanthe parent. Thus, as long as they can benefit from a plastic phenotypeintheiroffspring,parentsshouldengagein beha-vioral and physiological strategies that increase the off-spring’splasticity. Ofcourse,allelsemaynotbeequal;in particular,theintensityofconflictmaybereducedinsafe,

predictablecontexts(Section2.1),thusshiftingtheoffspring optimum toward higher levels of plasticity and favoring increased acceptance of parental influences. Still, an unknownproportionofbetween-individualvariancein post-natalplasticitymayultimatelybeexplainedbytheinterplay ofconflict-relatedmechanismsinprenataldevelopment.

Finally, when postnatal plasticity benefits one parent morethantheother,interparentalconflictmayresult,with oneparent(usuallythemother)promotingplasticityandthe other(usuallythefather)promotingthedevelopmentofa morefixedphenotype.Ofcourse,thebiologicalinterestsof anabsentfathermaybefurtheredbypaternalkin,aswellas by paternally expressed genes in the offspring’s genome (Section3.3).

The discovery that parent-offspring interactions are intrinsicallydrivenby conflictas wellas altruism (Trivers, 1974) has been a major propulsive force in evolutionary biology. After a long delay, this perspective is starting to makewayinpsychology(Schlomeretal.,2011).Inthefield ofprenatal development,however,thelogicof parent-off-spring conflict is still not widely appreciated, despite its remarkable explanatory and heuristic potential. That between the mother and the developing fetus is one of themostvital,intimate,andcomplexrelationshipsbetween twolivingbeings.Inordertofullyunderstandit,weneedto appreciatehowconflictandcooperationcangohandinhand, andlearntolookfortheirtracesinthedazzlingintricacyof physiologicalmechanisms.

Role

of

funding

source

Nofundingdeclared.

Conflict

of

interest

Noconflictdeclared.

Acknowledgment

I wish to thank David Haig, Tobias Uller, Bruce Ellis, Jay Belsky, andMichael Pluess for their useful comments and constructivecriticismonpreviousversionsofthemanuscript.

References

Alexander,N.,Osinsky,R.,Mueller,E.,Schmitz,A.,Guenthert,S., Kuepper,Y.,Hennig,J.,2011.Geneticvariantswithinthe dopa-minergicsysteminteracttomodulateendocrinestressreactivity andrecovery.Behav.BrainRes.216,53—58.

Alikhani-Koopaei,R., Fouladkou, F.,Frey, F.J., Frey, B.M., 2004. Epigeneticregulationof11b-hydroxysteroiddehydrogenasetype 2expression.J.Clin.Invest.114,1146—1157.

Allen,E.,Horvath,S.,Tong,F.,Kraft,P.,Spiteri,E.,Riggs,A.D., Marahrens,Y.,2003.Highconcentrations oflong interspersed nuclearelementsequencedistinguishmonoallelicallyexpressed genes.Proc.Natl.Acad.Sci.U.S.A.100,9940—9945.

Bartolomei, M.S., Tilghman, S.M., 1997. Genomic imprinting in mammals.Ann.Rev.Genet.31,493—525.

H.G.,Sultan,S.E., 2004.Developmental plasticityand human health.Nature430,419—421.

Belsky,J.,1997.Variationinsusceptibilitytorearinginfluences:an evolutionaryargument.Psychol.Inquiry8,182—186.

Belsky,J.,2005.Differentialsusceptibilitytorearinginfluences:an evolutionaryhypothesisandsomeevidence.In:Ellis,B., Bjork-lund,D.(Eds.),OriginsoftheSocialMind:Evolutionary Psychol-ogyandChildDevelopment.Guilford,NewYork, pp.139—163. Belsky,J.,Bakermans-Kranenburg,M.J.,vanIJzendoorn,M.H.,2007. Forbetterandforworse:differentialsusceptibilityto environ-mentalinfluences.Curr.Dir.Psychol.Sci.16,300—304. Belsky,J.,Beaver,K.M.,2010.Cumulative-geneticplasticity,

par-entingandadolescentself-regulation.J.ChildPsychol. Psychia-try52,619—626.

Belsky,J.,Jonassaint,C.,Pluess,M.,Stanton,M.,Brummett,B., Williams,R.,2009.Vulnerabilitygenesorplasticitygenes? Mol. Psychiatry14,746—754.

Belsky,J.,Pluess,M.,2009.Thenature(andnurture?)ofplasticityin earlyhumandevelopment. Perspect.Psychol.Sci.4,345—351. Belsky,J., Steinberg,L.,Draper,P., 1991.Childhoodexperience, interpersonaldevelopment,andreproductivestrategy:an evo-lutionarytheoryofsocialization.ChildDev.62,647—670. Bergman,K.,Sarkar,P.,O’Connor,T.G.,Modi,N.,Glover,V.,2007.

Maternalstressduringpregnancypredictscognitiveabilityand fearfulnessininfancy.J.Am.Acad.ChildAdol.Psychiatry46, 1454—1463.

Boney, C.M.,Verma, A., Tucker, R., Vohr, B.R.,2005. Metabolic syndromeinchildhood:associationwithbirthweight,maternal obesity,andgestationaldiabetesmellitus.Pediatrics115,e290— e296.

Boyce,W.T.,Alkon,A.,Tschann,J.M.,Chesney,M.A.,Alpert,B.S., 1995. Dimensions of psychobiologic reactivity: cardiovascular responses to laboratory stressors in preschool children. Ann. Behav.Med.17,315—323.

Boyce,W.T.,Ellis,B.J.,2005.Biologicalsensitivitytocontext:I.an evolutionary—developmentaltheoryoftheoriginsandfunctions ofstressreactivity.Dev.Psychopathol.17,271—301.

Boyce,W.T.,Essex,M.J.,Alkon,A.,Goldsmith,H.H.,Kraemer,H.C., Kupfer,D.J.,2006.Earlyfatherinvolvementmoderates biobe-havioralsusceptibilitytomentalhealthproblemsinmiddle child-hood.J.Am.Acad.ChildAdol.Psychiatr.45,1510—1520. Brunton,P.J.,2010.Resettingthedynamicrangeof

hypothalamic-pituitary-adrenal axis stress responses through pregnancy. J. Endocrinol.22,1198—1213.

Brunton,P.J.,Russell,J.A.,2011.Neuroendocrinecontrolof mater-nalstressresponsesandfetalprogrammingbystressinpregnancy. Prog.Neuro-Psychoph.35,1178—1191.

Cameron, N.M., Champagne,F.A., Parent, C., Fish, E.W., Ozak-Kuroda,K.,Meaney,M.J.,2005.Theprogrammingofindividual differencesindefensiveresponsesandreproductivestrategiesin theratthroughvariationsinmaternalcare.Neurosci.Biobehav. Rev.29,843—865.

Cameron,N.M.,Fish,E.W.,Meaney,M.J.,2009.Maternalinfluences onthesexualbehaviorandreproductivesuccessofthefemale rat.Horm.Behav.54,178—184.

Carola,V.,Frazzetto,G.,Pascucci,T.,Audero,E.,Puglisi-Allegra,S., Cabib, S., Lesch, K.P., Gross,C., 2008. Identifying molecular substrates in a mouse model of the serotonin transporter x environmentriskfactorforanxietyanddepression.Biol. Psychi-atry63,840—846.

Carroll,J.C.,Boyce-Rustay,J.M.,Millstein,R.,Yang,R.,Wiedholz, L.M.,Murphy,D.L.,Holmes,A.,2007.Effectsofmildearlylife stressonabnormalemotion-relatedbehaviorsin5-HTTknockout mice.Behav.Genet.37,214—222.

Champagne,F.A.,Meaney,M.J.,2007.Transgenerationaleffectsof socialenvironmentonvariationsinmaternalcareandbehavioral responsetonovelty.Behav.Neurosci.121,1353—1363.

Chuong,E.B.,Tong,W.,Hoekstra,H.E.,2010.Maternal—fetal con-flict:rapidlyevolvingproteinsintherodentplacenta.Mol.Biol. Evol.27,1221—1225.

Cirulli,F.,Francia,N.,Berry,A.,Aloe,L.,Alleva,E.,Suomi,S.J., 2009.Earlylifestressasariskfactorformentalhealth:roleof neurotrophinsfromrodentsto non-humanprimates.Neurosci. Biobehav.Rev.33,573—585.

Coppens,C.M.,deBoer,S.F.,Koolhaas,J.M.,2010.Copingstylesand behavioural flexibility: towards underlying mechanisms. Phil. Trans.R.Soc.B365,4021—4028.

Cottrell,E.C.,Seckl,J.R.,2009.Prenatalstress,glucocorticoidsand theprogrammingofadultdisease.Front.Behav.Neurosci.3,19. Davis,E.P.,Glynn,L.M.,Waffarn,F.,Sandman,C.A.,2010.Prenatal maternalstressprogramsinfantstressregulation.J.Child Psy-chol.Psychiatry52,119—129.

DelGiudice,M.,2007.Theevolutionarybiologyofcrypticpregnancy: a re-appraisal ofthe‘‘denied pregnancy’’phenomenon. Med. Hypotheses68,250—258.

Del Giudice, M., Angeleri,R.,Brizio, A.,Elena, M.R., 2010. The evolutionofautistic-likeandschizotypaltraits:Asexualselection hypothesis.Front.Psychol.1,41.

Del Giudice, M., Ellis, B.J., Shirtcliff,E.A., 2011. The adaptive calibration model of stress responsivity. Neurosci. Biobehav. Rev.35,1562—1592.

Del Giudice, M., Hinnant, J.B., Ellis, B.J., El-Sheikh, M., 2012. Adaptivepatternsofstressresponsivity:apreliminary investiga-tion.Dev.Psychol.48,775—790.

deWeerth,C.,Buitelaar,J.K.,2005.Physiologicalstressreactivityin humanpregnancy—areview.Neurosci.Biobehav.Rev.29,295— 312.

deWeerth,C.,Gispen-deWied,C.C.,Jansen,L.M.C.,Buitelaar,J.K., 2007.Cardiovascular and cortisolresponses toapsychological stressorduringpregnancy.ActaObstet.Gynecol.86,1181—1192. DiazBrinton,R.,Thompson,R.F.,Foy,M.R.,Baudry,M.,Wang,J.M., Finch,C.E.,Morgan,T.E.,Pike,C.J.,Mack,W.J.,Stanczyk,F.Z., Nilsen,J., 2008.Progesteronereceptors:formandfunctionin brain.Front.Neuroendocrinol.29,313—339.

Diego,M.A.,Jones,N.A.,Field,T.,Hernandez-Reif,M.,Schanberg, S.,Kuhn,C.,Gonzalez-Garcia,A.,2006.Maternalpsychological distress,prenatalcortisol,andfetalweight.Psychosom.Med.68, 747—753.

DiPietro,J.A.,Costigan,K.A.,Gurewitsch,E.D.,2003.Fetalresponse toinducedmaternalstress.EarlyHum.Dev.74,125—138. Ellis,B.J.,Boyce,W.T.,Belsky,J.,Bakermans-Kranenburg,M.J.,van

IJzendoorn,M.H.,2011.Differentialsusceptibilitytothe envi-ronment:anevolutionary—neurodevelopmentaltheory.Dev. Psy-chopathol.23,7—28.

Ellis,B.J.,Essex,M.J.,Boyce,W.T.,2005.Biologicalsensitivityto context:II.Empiricalexplorationsofan evolutionary—develop-mentaltheory.Dev.Psychopathol.17,303—328.

Ellis,B.J.,Figueredo,A.J.,Brumbach,B.H.,Schlomer,G.L.,2009. Theimpactofharshversusunpredictableenvironmentsonthe evolutionanddevelopmentoflifehistorystrategies.Hum.Nat. 20,204—268.

Ellis,B.J., Jackson,J.J., Boyce,W.T., 2006. Thestressresponse system: universality and adaptive individual differences.Dev. Rev.26,175—212.

Feig,D.S.,Zinman,B.,Wang,X.,Hux,J.E.,2008.Riskof develop-mentofdiabetesmellitusafterdiagnosisofgestationaldiabetes. Can.Med.Assoc.J.179,229—234.

Flinn,M.V.,Nepomnaschy,P.A.,Muehlenbein,M.P.,Ponzi,D.,2011. Evolutionaryfunctionsofearlysocialmodulationof hypothalam-ic-pituitary-adrenalaxisdevelopmentinhumans.Neurosci. Bio-behav.Rev.35,1611—1629.

Frankenhuis,W.E.,DelGiudice,M.,2012.Whendoadaptive devel-opmentalmechanismsyieldmaladaptiveoutcomes? Dev. Psy-chol.48,628—642.

Gangestad, S.W., Simpson, J.A., 2000. The evolution of mating: trade-offsandstrategicpluralism.Behav.BrainSci.23,675—687. Ganzel,B.L.,Morris,P.A.,Wethington,E.,2010.Allostasisandthe humanbrain:integratingmodelsofstressfromthesocialandlife sciences.Psychol.Rev.117,134—174.

Gardner,A.,Grafen,A.,2009.Capturingthesuperorganism:aformal theoryofgroupadaptation.J.Evol.Biol.22,659—671. Giannakoulopoulos,X.,Teixeira,J.,Fisk,N.,Glover,V.,1999.Human

fetaland maternalnoradrenalineresponsestoinvasive proce-dures.Pediatr.Res.45,494—499.

Glover,V.,2011.Prenatalstressandtheoriginsofpsychopathology: an evolutionary perspective. J. Child Psychol. Psychiatry 52, 356—367.

Glover,V.,Bergman,K.,Sarkar,P.,O’Connor,T.G.,2009a. Associa-tionbetweenmaternalandamnioticfluidcortisolismoderated bymaternalanxiety.Psychoneuroendocrinology34,430—435. Glover,V.,O’Connor,T.G.,O’Donnell,K.,2009b.Prenatalstressand

theprogrammingoftheHPAaxis.Neurosci.Biobehav.Rev.35, 17—22.

Gluckman,P.D.,Hanson,M.A.,2004.Livingwiththepast:evolution, development,andpatternsofdisease.Science305,1733—1736. Gluckman,P.D.,Hanson,M.A.,Spencer,H.G.,2005.Predictive adap-tiveresponsesandhumanevolution.TrendsEcol.Evol.20,527— 533.

Goshen,R., Gonik, B., de Groot,N., Hochberg, A.A., 1994. The genomicbasisoftheb-subunitofhumanchorionicgonadotropin diversityintriploidy.Am.J.Obst.Gynecol.170,700.

Gotlib,I.H.,Joormann,J.,Minor,K.L.,Hallmayer,J.,2008.HPAaxis reactivity: a mechanism underlying the associations among 5-HTTLPR, stress, and depression. Biol. Psychiatry 63, 847—851.

Gunnar, M.R., Frenn, K., Wewerka, S.S., Van Ryzin, M.J., 2009. Moderateversussevereearlylifestress:associationswithstress reactivityandregulationin10-12-year-oldchildren. Psychoneur-oendocrinology34,62—75.

Gutteling, B.M., De Weerth, C., Buitelaar, J.K., 2004. Maternal prenatal stress and 4-6 year old children’s salivary cortisol concentrationspre-andpost-vaccination.Stress7,257—260. Gutteling,B.M.,deWeerth,C.,Buitelaar,J.K.,2005.Prenatalstress

andchildren’s cortisolreactiontothefirstdayofschool. Psy-choneuroendcrinology30,541—549.

Haig,D.,1993.Geneticconflictsinhumanpregnancy.Q.Rev.Biol. 68,495—532.

Haig,D.,1996.Placentalhormones,genomicimprinting,and mater-nal-fetalcommunication.J.Evol.Biol.9,357—380.

Haig,D.,1997.Parentalantagonism,relatednessasymmetries,and genomicimprinting.Proc.R.Soc.Lond.B264,1657—1662. Haig,D.,2004.Genomic imprintingandkinship: howgood isthe

evidence? Ann.Rev.Genet.38,553—585.

Haig,D.,2007.Puttingupresistance:maternal—fetalconflictoverthe control ofuteroplacental blood flow. In:Aird, W.C. (Ed.), En-dothelial Biomedicine. Cambridge University Press, New York, NY, pp.135—141.

Haig, D., 2008. Placental growth hormone-related proteins and prolactin-relatedproteins.Placenta29,s36—s41.

Heijmans, B.T., Tobi, E.W., Stein, A.D., Putter, H., Blauw, G.J., Susser,E.S.,Slagboom,P.E.,Lumey,L.H.,2008.Persistent epi-geneticdifferencesassociatedwithprenatalexposuretofamine inhumans.Proc.Natl.Acad.Sci.105,17047—17049.

Heiming,R.S.,Bodden,C.,Jansen,F.,Lewejohann,L.,Kaiser,S., Lesch,K.P.,Palme,R.,Sachser,N.,2011.Livinginadangerous worlddecreasesmaternalcare:astudyinserotonintransporter knockoutmice.Horm.Behav.60,397—407.

Heiming,R.S.,Jansen,F.,Lewejohann,L.,Kaiser,S.,Schmitt,A., Lesch,K.P.,Sachser,N.,2009.Livinginadangerousworld:the

shapingofbehavioralprofileby earlyenvironment and5-HTT genotype.Front.Behav.Neurosci.3,26.

Hill,M., Parı´zek,A.,Cibula,D., Kancheva,R.,Jira´sek,J.E., Jir-kovska´,M.,Velı´kova´,M.,Kuba´tova´,J.,Klı´mkova´,M.,Paskova´,A., Zizka, Z., Kancheva, L., Kazihnitkova´, H., Zamrazilova´, L., Sta´rka, L., 2010a. Steroidmetabolome in fetal and maternal bodyfluidsinhumanlatepregnancy.J.Ster.Biochem.,Mol.Biol. 122,114—132.

Hill,M.,Parı´zek,Jira´sek,J.E.,Jirkovska´, M.,Velı´kova´,M.,Duskova´, M., Klı´mkova´, M.,Paskova´,A.,Zizka,Z.,Germanova´, A.,Koucky,M., Kalousova´,M.,Sta´rka,L.,2010b.Ismaternalprogesteroneactually independentofthefetalsteroids? Physiol.Res.59,211—224. Hinde,R.A.,1986.Someimplicationsofevolutionarytheoryand

comparativedataforthestudyofhumanprosocialand aggres-sive behaviour. In: Olweus, D., Block, J., Radke-Yarrow,M. (Eds.), Development of Anti-social and Prosocial Behaviour. AcademicPress,Orlando, pp.13—32.

Hou,Z.,Romero,R.,Uddin,M.,Than,N.G.,Wildman,D.E.,2009. Adaptivehistoryofsinglecopygeneshighlyexpressedintheterm humanplacenta.Genomics93,33—41.

Kaasen,A.,Helbig,A.,Malt,U.F.,Godang,K.,Bollerslev,J.,Næs,T., Haugen,G.,2011.Therelationofpsychologicaldistressto sali-varyandserumcortisollevelsinpregnantwomenshortlyafterthe diagnosisofastructuralfetalanomaly.ActaObstetr.Gynecol. Scand.91,68—78.

Kaiser,S.,Sachser,N.,2005.Theeffectsofprenatalsocialstresson behaviour:mechanismsand function.Neurosci.Biobehav.Rev. 29,283—294.

Kaiser,S., Sachser, N.,2009. Effects ofprenatalsocial stresson offspringdevelopment:pathologyoradaptation? Curr.Direct. Psychol.Sci.18,118—121.

Kapoor,A.,Dunn,E.,Kostaki,A.,Andrews,M.H.,Matthews,S.G., 2006.Fetalprogrammingofhypothalamo-pituitary-adrenal func-tion:prenatalstressandglucocorticoids.J.Physiol.572,31—44. Khashan,A.S.,Abel,K.M., McNamee,R.,Pedersen, M.G., Webb, R.T.,Baker,P.N.,Kenny,L.C.,BoMortensen,P.,2008.Higherrisk ofoffspringschizophreniafollowingantenatalmaternalexposure tosevereadverselifeevents.Arch.Gen.Psychiatry65,146—152. Kinney,D.K.,Miller,A.M.,Crowley,D.J.,Huang,E.,Gerber,E.,2008. Autismprevalencefollowingprenatalexposuretohurricanesand tropicalstormsinLouisiana.J.AutismDev.Disord.38,481—488. Kochanska,G.,Aksan,N.,Joy,M.E.,2007.Children’sfearfulnessasa moderatorofparentinginearlysocialization: twolongitudinal studies.Dev.Psychol.43,222—237.

Koolhaas,J.M.,Bartolomucci,A.,Buwalda,B.,deBoer,S.F.,Flu¨gge, G.,Korte,S.M.,Meerlo,P.,Murison,R.,Olivier,B.,Palanza,P., Richter-Levin,G.,Sgoifo,A.,Steimer,T.,Stiedl,O.,vanDijk,G., Wo¨hr,M.,Fuchs,E.,2011.Stressrevisited:acriticalevaluationof thestressconcept.Neurosci.Biobehav.Rev.35,1291—1301. Krozowski,Z.,Li,K.X.Z., Koyama,K.,Smith,R.E., Obeyesekere,

V.R.,Stein-Oakley,A.,Sasano,H.,Coulter,C.,Cole,T.,Sheppard, K.E.,1999.ThetypeIandtypeII11b-hydroxysteroid dehydroge-naseenzymes.J.Ster.Biochem.Mol.Biol.60,391—401. Lessells,C.M.,Parker, G.A.,1999. Parent—offspringconflict:The

full-sib-half-sibfallacy.Proc.R.Soc.B266,1637—1643. Love,O.P.,Williams,T.D.,2008.Theadaptivevalueofstress-induced

phenotypes:effectsofmaternallyderivedcorticosteroneon sex-biasedinvestment,costofreproduction,andmaternalfitness. Am.Nat.172,e135—e149.

Luisi, S., Petraglia, F., Benedetto, C., Nappi, R.E., Berardi, F., Fadalti,M.,Reis,.M.,Luisi,M.,Genazzani,A.R.,2000.Serum allopregnanolone levels in pregnant women: changes during pregnancy, at delivery,and in hypertensive patients. J.Clin. Endocrinol.Metab.85,2429—2433.