Interaction Between Serotonin Transporter Gene Variation and RearingCondition in Alcohol Preference and Consumption in Female Primates FREE

Alcoholism is a relapsing, lifetime illness that is notoriously difficultto treat. Although it is a complex disorder, with multiple subtypes and clinicalpictures, one defining feature of alcoholism is the compulsive use of ethanol,often in the face of negative social and psychological consequences. Bothpositive and negative reinforcement are thought to be critically involvedin the transition from casual alcohol use to compulsive alcohol-seeking behavior.¹ While the positive reinforcing effects of alcoholare essential to the initiation and early maintenance of intake, recent studiessuggest that alcohol-seeking behavior related to alleviation of symptoms duringabstinence (negative reinforcement) is equally, if not more, effective inmaintaining alcohol use.² Therefore, when consideringrisk factors for alcohol dependence, it is important to consider not onlysystems that are involved in alcohol reward but those activated followingalcohol exposure, during acute withdrawal and abstinence.

Among the systems involved both in positive reinforcement of alcohol-inducedreward and negative reinforcement of alcohol withdrawal is the limbic-hypothalamic-pituitary-adrenal(LHPA) axis. Acutely, exposure to ethanol increases hypothalamic release ofcorticotropin-releasing hormone (CRH) and arginine vasopressin.^{3 - 5} Thisis followed by an increase in pro-opiomelanocortin synthesis in both the adenohypophysisand the intermediate lobe of the pituitary gland. Posttranslational cleavageof the pro-opiomelanocortin precursor produces multiple peptides, including β-endorphin(β-EP) and corticotropin, the former peptide producing some of the rewardingand reinforcing effects of alcohol.^{6 - 7} Corticotropinis released into the peripheral circulation to stimulate the synthesis andrelease of glucocorticoids, which are known to potentiate the positive reinforcingeffects of drugs of abuse, from the adrenal cortex. Increases in CRH releasealso occur in forebrain structures during acute exposures to alcohol and forup to 12 hours of withdrawal,⁸ and releaseof CRH in the central nucleus of the amygdala is thought to contribute toanxiety associated with acute alcohol withdrawal.^{9 - 10}

The serotonin system is also involved both in reinforcement of alcoholintake and symptoms of withdrawal, and a number of studies have shown thatthe serotonin system modulates CRH release and that there are reciprocal interactinginfluences between the LHPA axis and central serotonin activity.^{11 - 12} Whileserotonin release following consumption of alcohol is involved in activationof reward pathways, neuroadaptive diminutions in release following alcoholexposure can lead to pain, dysphoria, and depression.¹³ Muchfocus in the area of alcohol research has been on the serotonin system. Serotoninis one of the key neurotransmitters to be released in response to alcohol.¹⁴ In rodents, alcohol preference is associated withdecreased levels of serotonin and its major metabolite, 5-hydroxyindoleaceticacid (5-HIAA), in relevant areas of the brain as well as a decrease in thenumber of serotonergic neurons in the raphe nuclei.¹⁵ Amongalcohol-preferring rats, this serotonin deficit is maintained even followingadministration of ethanol.¹⁶ In humans, lowcerebrospinal fluid levels of 5-HIAA are associated with alcohol-seeking behaviorand alcoholism.^{17 - 18} This hasbeen observed among nonhuman primates as well.^{19 - 22}

Variations in many of the genes that encode receptors, enzymes, andtransporters for factors release in response to alcohol have also been studiedin association with alcoholism. A common insertion-deletion polymorphism inthe promoter region for the serotonin transporter gene alters in vitro genetranscription,²³ in vitro transporter availability,²⁴ and in vivo serotonin transporter density.²⁵ There have been several associations of this polymorphismto behaviors and traits that relate to excessive alcohol intake, for exampleanxiety,^{23 ,26} as well as to certainsubtypes of alcoholism.^{27 - 29} Orthologousto the functional polymorphism in the serotonin transporter gene promoterin humans, a 21–base pair (bp) length variant in the transcriptionalcontrol region of the serotonin transporter gene of the macaque, rh5-HTTLPR,³⁰ has been shown to altertranscriptional efficiency,³¹ resulting indecreased serotonin transporter messenger RNA levels in brains of macaqueswith the l/s genotype.³² Theserotonin transporter gene promoter also contains a glucocorticoid responseelement,³³ making it responsive to stress-inducedlevels of corticosteroids, a phenomenon that is particularly evident amongboth human and rhesus carriers of the s allele.^{32 ,34}

Previous studies in our laboratory have demonstrated an effect of serotonintransporter gene promoter variation on cerebrospinal fluid levels of 5-HIAA,³¹ the level of response to alcohol,³⁵ andLHPA-axis responses to stress³⁶ but only amonganimals reared in peer-only groups, a model for early-life stress. Both peer-rearingand the s allele are associated with increased glucocorticoidreceptor messenger RNA expression in rhesusbrain,³² suggestingthat these animals would be more sensitive to glucocorticoids and thereforethat they may also experience more of the positive reinforcing propertiesof alcohol. Reports from our laboratory indicate that peer-reared (PR) animalsconsume more alcohol as young adults^{22 ,37} andthat PR female rhesus macaques have exaggerated LHPA-axis responses to alcohol.³⁸ We have also found that low cerebrospinal fluid levelsof 5-HIAA, known to be associated with alcohol-seeking behavior in humansand animal models, are observed among PR animals with the l/s genotype.³¹ Our findings are interestingin light of those from studies in humans, which demonstrate that 5-HTTLPRmoderates the influence of life stress on depression, a risk factor for alcoholism.³⁹

Because their environments can be controlled, use of the macaque modelpermits investigation of independent influences as well as potential interactionsbetween serotonin system–related genes, maternal deprivation, and stressin the etiology of alcohol consumption. What follows is a study focusing onfemale drinking patterns in adolescent macaques given simultaneous accessto alcoholic and nonalcoholic aspartame-sweetened solutions for 1 hour a dayfor a period of 6 weeks. Both preference for alcohol and changes in patternsof alcohol consumption with repeated exposures were determined. Since serotonin, β-EP,and corticosteroids are involved in the reinforcement of alcohol intake andbecause peer-rearing and serotonin transporter gene promoter variation havebeen associated with alcohol consumption in human and animal populations,we predicted that rh5-HTTLPR genotype and rearing experience would interactto influence preference for alcohol in female rhesus macaques. As noted, theserotonin system and hypothalamic-pituitary-adrenal axis interact centrally.Because PR females have been shown to have higher levels of LHPA-axis activationfollowing administration of alcohol and since dysregulated release of serotoninand CRH are related to symptoms of anxiety during alcohol withdrawal, we alsowanted to determine the effects of these variables on the maintenance of voluntaryalcohol intake across the course of a 6-week alcohol consumption paradigm.

Animals were either PR or mother-reared (MR). The former condition deprivesanimals of parental input and the opportunity to learn appropriate socialbehaviors and context during early development and is a model for early-lifestress, whereas the latter is meant to approximate natural conditions. Theserearing conditions have been described in detail elsewhere.^{21 ,37 ,39} Briefly,MR animals (n = 18) were reared for the first 6 months of life withtheir mothers and fathers in social groups comprising 8 to 12 adult females(about half of whom had same-aged infants) and 2 adult males. Peer-rearedanimals (n = 14) were separated from their mothers at birth andhand-reared in a neonatal nursery for the first 37 days of life. For the first14 days, they were housed in an incubator and hand-fed. From day 14 untilday 37, they were placed alone in a nursery cage and provided a blanket anda terry cloth–covered, rocking surrogate. A bottle from which the infantswould feed was fixed to the surrogate. At 37 days of age, they were placedin a cage with 3 other age-mates with whom they had continuous contact. Thus,like MR monkeys, PR monkeys had daily opportunity to interact socially andpractice social skills but did so in the absence of adult role models. Asa consequence, PR monkeys were emotionally unstable and exhibited impairedsocial skills.^{20 - 21 ,39}

When the animals reached an average age of 7 months, MR and PR animalswere placed together to form a larger social group. This occurred after themonkeys were subjected to social separation in a paradigm consisting of 4one-week social separations.³⁹ To form thesegroups, MR and PR animals were placed together in a larger cage, forming apermanent social group. On average, groups were composed of mean ± SD59% ± 5% females and 41% ± 4% males. Withinthe groups studied, mean ± SD 56% ± 10%of the animals were MR and 44% ± 10% were PR. All animalslived in their respective social groups throughout the study and receivedidentical treatment. Protocols for the use of experimental animals were approvedby the institutional animal care and use committee of the National Institutesof Alcoholism and Alcohol Abuse, Poolesville, Md.

Study animals were 32 young adolescent, female rhesus macaques, obtainedfrom 4 birth-year cohorts, ranging in age from 3.3 to 3.7 years (mean age,3.4 years) at the initiation of the alcohol self-administration study. Animalswere allowed to freely consume an aspartame-sweetened 8.4% (volume-to-volumeratio) alcohol solution for 1 hour a day, 4 days a week for 2 weeks.^{20 - 21 ,37} Briefly, this standardizedmethod consisted of 3 phases: (1) Spout training. The animals were trainedfor an hour a day across a 1-week period to drink from nipple-like spoutsthat dispensed aspartame-sweetened water. This phase lasted 5 days, at whichpoint all animals consumed more than 50 mL of the vehicle. (2) Initial alcoholexposure. This phase was designed to assure that all animals experienced thepharmacological effects of alcohol before beginning the experimental phaseof the study. To begin this phase, the color of the sweetened vehicle waschanged, and alcohol was added to the vehicle to produce an 8.4% volume-volumeratio alcohol solution. During the initial alcohol exposure phase, animalswere given free access to the alcohol solution for an hour each day. Eachof the animals included in this phase of the study fulfilled a preestablishedcriterion of consuming more than 0.67 g/kg body weight of the alcohol solutionon 2 or more occasions. Once all animals met the criterion, a second bottlecontaining sweetened vehicle was added. Thereafter, both the nonalcoholicand 8.4% alcoholic sweetened solutions were available in addition to normaldrinking water for an hour each day. No special methods, such as deprivationof food or water, were used to induce drinking, and animals established stableconsumption patterns within 2 weeks. (3) Experimental period. During the 6-weekexperimental phase, alcohol and vehicle were dispensed 4 days a week (Monday-Thursday)from 1300 to 1400 while the animals were in their home-cage environment. Animals’weights ranged from 4.2 to 9.3 kg (mean ± SD weight, 5.8 ± 1.2kg).

Using standard extraction methods, DNA was isolated from whole bloodand collected from the femoral vein after the animals had been given ketamineanesthesia (15 mg/kg, intramuscular). The serotonin transporter gene promoterregion (rh-5HTTLPR) was amplified from 25 ng of genomic DNA with flankingoligonucleotide primers (stpr5, 5′-GGCGTTGCCGCTCTGAATGC; intl, 5′-CAGGGGAGATCCTGGGAGGG)in 15-μL reactions using Platinum Taq and thePCRX Enhancer System kit (Invitrogen, Carlsbad, Calif), according to the manufacturer’sprotocol. Amplifications were performed on a thermocycler (9700) (Perkin-Elmer,Fremont, Calif) with 1 cycle at 96°C for 5 minutes followed by 30 cyclesof 94°C for 15 seconds, 60°C for 15 seconds, 72°C for 30 seconds,and a final 3-minute extension at 72°C. Amplicons were separated by electrophoresison a 10% polyacrylamide gel, and the short (s) (398-bp)and long (l) (419-bp) alleles of the rh5-HTTLPR wereidentified by direct visualization following ethidium-bromide staining.

To assess the effects of rh5-HTTLPR genotype and rearing condition onalcohol preference, 2-way analysis of variance was conducted. Animals wereassigned nominal independent variables according to rearing condition (MRor PR) and rh5-HTTLPR genotype (l/l or l/s), and the influence of these variables on alcohol preference wasdetermined. Average volumes (mL) of alcoholic and nonalcoholic solutions consumedduring the course of the drinking study were calculated, and alcohol preferencewas determined by dividing the volume of 8.4% alcohol solution consumed bythe total volume of solution consumed and multiplying by 100 ([millilitersof alcoholic solution/milliliters of alcoholic solution + millilitersof nonalcoholic solution] × 100).

To determine the influence of serotonin transporter gene variation andearly-life stress on maintenance of alcohol consumption across the courseof the drinking paradigm, weekly means of alcohol consumed were determined,and the influences of rh5-HTTLPR genotype and rearing condition on weeklyaverages of alcohol consumption in grams per kilogram were analyzed usingrepeated-measures analysis of variance. There were 14 MR l/l, 4 MR l/s, 10 PR l/l, and 4 PR l/s females in the study. Sinceanimals with the xl/l and s/s genotypewere rare, they were excluded from all analyses. Alcohol consumption rateswere available for 6 animals for only 4 or 5 weeks, and vehicle consumptionrates were unavailable for 1 animal. All analyses were performed using StatViewStatistical software (SAS Institute, Cary, NC). Criterion for significancewas set at P< .05.

There was a significant main effect of serotonin transporter gene variationon alcohol preference (Figure 1) (F_1,31 = 5; P< .04). Alcoholpreference for female macaques with the l/s genotype(mean ± SEM alcohol preference, 28% ± 6%)was higher than that of l/l females (mean ± SEMalcohol preference, 22% ± 7%; Fisher probable least-squaresdifference [PLSD], P< .05). In addition,there was an interaction between serotonin transporter genotype and rearingcondition (Figure 1) (F_1,31 = 5.3; P< .03). Post hoc analysis demonstrated that PR l/s animals (mean ± SEM alcohol preference,38% ± 9%) consumed more alcohol than did PR l/l (mean ± SEM alcohol preference, 25% ± 16%;Fisher PLSD, P< .005) or MR l/l animals (mean ± SEM alcohol preference, 28% ± 6%;Fisher PLSD, P< .03).

Using repeated-measures analysis of variance, we found a significanteffect of the number of weeks of exposure to alcohol on levels of consumption(Figure 2) (F_5,95 = 17; P< .001). Post hoc analyses demonstrated significantdifferences in the levels of consumption between all weeks of study, withthe exception of the graded comparisons between weeks 1 and 2, 2 and 3, and3 and 4 (Fisher PLSD, P<.001). Mean ± SEMlevels of consumption for all animals studied were 0.14 ± 0.03g/kg per hour (range, 0-0.63 g/kg per hour) during the first week of the studyand 0.67 ± 0.12 g/kg per hour (range, 0-2.2 g/kg per hour)during the last week of the study. There was an interaction between weeksof exposure to alcohol and rearing condition such that PR animals increasedtheir consumption more than MR animals across the course of the 6-week alcoholconsumption paradigm (Figure 2) (F_5,95 = 4; P<.006). In contrast,there were no effects of rearing or genotype on vehicle consumption acrossthe course of the study. During the first week of testing, neither alcohol(MR l/l, mean ± SEM week 1 consumption,0.24 ± 0.06 g/kg per hour; MR l/s,mean ± SEM week 1 consumption, 0.17 ± 0.05g/kg per hour; PR l/l, mean ± SEMweek 1 consumption, 0.15 ± 0.10 g/kg per hour; PR l/l, mean ± SEM week 1 consumption, 0.30 ± 0.17g/kg per hour) nor vehicle (MR l/l, mean ± SEMweek 1 consumption, 25 ± 5 mL/kg per hour; MR l/s, mean ± SEM week 1 consumption, 24 ± 5mL/kg per hour; PR l/l, mean ± SEMweek 1 consumption, 19 ± 8 mL/kg per hour; PR l/s, mean ± SEM week 1 consumption, 17 ± 4mL/kg per hour) consumption differed significantly among the 4 groups of study.Mother-reared animals exhibited approximately a 2-fold increase in their consumptionof vehicle (MR l/l, mean ± SEM week6 consumption, 39 ± 10 mL/kg per hour; MR l/s, mean ± SEM week 6 consumption, 45 ± 16mL/kg per hour) and alcohol by the last week of testing (MR l/l, mean ± SEM week 6 consumption, 0.54 ± 0.19g/kg per hour and MR l/s, mean ± SEMweek 6 consumption, 0.48 ± 0.19 g/kg per hour). Conversely,PR monkeys, which exhibited no increase in vehicle consumption across thecourse of the study, (PR l/l, mean ± SEMweek 6 consumption, 21 ± 7 mL/kg per hour; PR l/s, mean ± SEM week 6 consumption, 13 ± 3mL/kg per hour), increased their alcohol intake by 5-fold (PR l/l, mean ± SEM week 6 consumption, 0.91 ± 0.2g/kg per hour and PR l/s mean ± SEMweek 6 consumption, 1.4 ± 0.16 g/kg per hour).

The “nature vs nurture” controversy over the developmentof personality is long-standing.^{40 - 41} Thisdebate extended to the development of psychopathology and neuropsychiatricdisease as well.^{41 - 42} Today, itis widely accepted that neuropsychiatric disorders are complex traits, drivenby both environmental and genetic influence, and that, moreover, there arepotential gene-gene and gene × environment interactions that may beat the root of the development of some personality traits and neuropsychiatricdiseases.^{43 - 44} With the adventof modern molecular and statistical tools, we are now capable of refiningour approach so that specific gene × environment interactionsin the development of psychiatric disease can be revealed.

Alcohol abuse and alcoholism are among the psychiatric disorders thoughtto be influenced strongly by both genetic and environmental factors. Individualswho experience early-life psychosocial stressors, such as abuse or loss ofa parent, are at increased risk for anxiety and depression, known risk factorsfor alcoholism, and parental monitoring is known to modulate the risk foralcohol abuse in adolescents.⁴⁵ The 5-HTTLPR s allele has been demonstrated to be associated with anxiety,neuroticism, and associated traits in numerous studies,⁴³ andrecent studies have demonstrated carriers of the s alleleto be more susceptible to major depression in the face of repeated stress.³⁹ The obvious strength of animal studies is that environmentalfactors can be controlled, such that relative contributions or potential interactionsbetween genes and environment can be examined.⁴⁶ Moreover,the animal model allows for us to study mechanisms of these interactions,since molecular and environmental interactions can be closely studied as theyrelate to behaviors that are potential contributors to the susceptibility,pathogenesis, and progression of psychiatric illnesses.

Some studies have shown an association of the l/s serotonin transporter genotype with alcoholism.²⁸ However,other studies have produced negative results.⁴⁷ Thesecontradictory results could be attributable, in part, to variations in earlyexperience or in sex composition among different populations of study. Inthe present study, we report that there is an association between serotonintransporter gene promoter variation and ethanol preference in female rhesusmacaques. The effect of serotonin transporter genotype, however, is environmentallydependent. Perhaps the most interesting aspect of this study is that a genotypethought to confer risk for a wide variety of psychopathological illnessesand traits exerts its effect on alcohol preference and consumption only whenanimals are reared in an environment that models parental absence or neglect.This interaction is observed only in relation to alcohol consumption. Vehicleconsumption, on the other hand, is affected neither by rearing nor by rh5-HTTLPRgenotype. Although our limited sample size dictates that we interpret ourdata with caution, it appears as if maternal input provides a buffering function,such that MR animals with the l/s genotype, unlikePR l/s animals, are not at risk for high alcoholpreference or intake. This is reminiscent of the stress × biologicalrisk diathesis model that was proposed in early theories of psychopathology.What this particular study provides is an updated approach to understandingpsychopathology and its treatment, explaining the basis for the observationsthat 2 individuals from similarly impoverished environments, but with differentgenetic backgrounds, show disparate developmental outcomes. It is also consistentwith twin studies, which show that identical twins are not always concordantfor psychopathology. Finally, our findings are in agreement with those ofCaspi et al,⁴⁸ which demonstrate interactionsbetween serotonin transporter gene promoter variation and stress in the pathogenesisof depression.

Previous studies in our laboratory have shown that PR female macaqueshave augmented corticotropin responses to alcohol and that elevated corticotropinlevels persisted for weeks following discontinuation of the alcohol consumptionstudy.³² In the present study, we have demonstratedthat early experience and exposure to alcohol interact with one another, suchthat while there is no effect of rearing condition on alcohol consumptionduring initial exposures to alcohol, females subjected to stress early inlife, especially those who are carriers of the s allele,demonstrate progressive increases in their levels of alcohol consumption,achieving levels that would be expected to produce blood alcohol concentrationsin the range of 100 mg%-150 mg%.⁴⁹ This isin contrast to the maximal alcohol intake observed among MR females, whoseconsumption would be predicted to produce blood alcohol concentrations inthe range of 50 mg%, well lower than the legal level of intoxication. Corticosteroidsare thought to influence hepatic expression of alcohol dehydrogenase⁵⁰ and therefore could result in acceleration of thedevelopment of induced tolerance to alcohol with progressive exposures. Althoughin PR female macaques, differences in levels of corticotropin have been notedfollowing exposures to alcohol, and there does not appear to be a differentialeffect of alcohol on total cortisol levels among these animals. We thereforedo not think that this is a likely explanation for accelerated increases inrates of consumption among PR females.

Studies have shown that while alcohol acutely decreases glucocorticoidresponse element binding in the rat amygdala (which other studies suggestwould, ultimately, produce anxiolysis),⁵¹ serotoninreceptor antagonism prevents this from occurring.⁵² Thissuggests the potential for an interaction between the serotonin system, alcohol,and amygdalar reactivity during exposures to alcohol. Although, unlike rodents,primates are not thought to become sensitized to the positive reinforcingeffects of drugs of abuse, they may, like humans, become sensitized to alcohol’snegative reinforcing effects. Recently, we demonstrated there to be an interactionbetween rearing condition and rh5-HTTLPR genotype on LHPA-axis activationin female macaques.³⁶ One potential mechanismthat could explain the progressive increases in alcohol consumption observedin PR females (especially those carrying the s allele)is rapid sensitization to the negative reinforcing effects of alcohol. Ifit is the case that female PR l/s animals progressivelyincrease their alcohol intake in order to alleviate the negative symptoms(ie, anxiety or dysphoria) associated with alcohol exposure, then in humanpopulations, women with variation in the serotonin transporter gene promoterwho are also exposed to early-life stress may be particularly vulnerable toalcoholism. The influences of early experience and chronic stress are particularlyrelevant with regard to type I alcoholism,⁵³ whichis related to anxiety and is the more common type of alcoholism among women.It is also of interest that the s allele has, inseveral instances, been associated with suicide ideations among type 1 alcoholics,suggesting an interaction between alcohol intake and serotonin transportergene variation in the etiology of severe depressive symptoms.⁵⁴ Inaddition, there are known interactions between 5-HTTLPR and stress in theincidence of depression.³⁹ It may be that alcoholpreference in PR l/s females is reflective not onlyof altered serotonin release following exposure to alcohol but a predispositionto anxiety both independent of and in relation to alcohol consumption.

Traits characteristic of type 1 and type 2 alcoholism are thought torelate to dysregulated central nervous system serotonin functioning. To theextent that they generalize to humans, our findings suggest that the pathogenesisof alcohol dependence has its genesis, at least in part, in the interactinginfluence of early deleterious rearing experience and genetic factors. Thesimilarity of humans and rhesus monkeys in genetic variation of the serotonintransporter gene promoter region as well as serotonin-mediated behavioraldeficits suggest that the nonhuman primate model may have value for determiningwhether genetic variation may be used to identify or develop appropriate pharmacotherapiesfor the treatment of serotonin-related disorders, including alcoholism. Italso allows us to observe behavioral patterns, for example, patterns of alcoholconsumption during adolescence, that may lead to susceptibility, pathogenesis,and progression of alcohol-related disorders.

One major limitation in the treatment of addiction is the inabilityto restore the addicted brain to its preaddicted state. Early-life stresscan cause persistent changes in the neuroendocrine stress axis and serotoninsystem, both of which are implicated in alcohol-induced allostasis and allostaticload in the brain.⁵⁵ Since activation of theneuroendocrine stress axis and dysregulated serotonin neurotransmission arethought to be factors that predispose individuals to alcohol withdrawal, andtherefore depression, dysphoria, and relapse, it is possible that combinationtherapies that both regulate the serotonin system and prevent overactivityof the neuroendocrine stress axis would not only help to prevent progressionof alcoholism and related disorders but may also be effective in returningthe alcoholic brain to an earlier allostatic set point. By learning more aboutthe interactions between genes, early experience, and alcohol intake in thenonhuman primate, we may better be able to design combination therapies forpreventing and treating alcoholism.

Submitted for Publication: May1, 2003; final revision received April 16, 2004; accepted April 21, 2004.

Correspondence: Christina S. Barr, VMD,PhD, National Institutes of Health Animal Center, PO Box 529, Bldg 112, Poolesville,MD 20837 (cbarr@mail.nih.gov).

Funding/Support: This study was supported bythe National Institute of Child Health and Human Development and the NationalInstitute on Alcohol Abuse and Alcoholism Intramural Research Programs, Poolesville,Md, and grants SFB 581 and Le 629/4-2 from Deutsche Forschungsgemeinschaft,Bonn, Germany.

Additional Information: Drs Lesch, Suomi, Goldman,and Higley contributed equally to this study.

Acknowledgment: We thank Alan Dodson; AnneSponberg, BA; Ted King; Todd Graham, BA; Ruth Andrews, BS; Tami Gura, BA;and Kim Wojteczko, BS, for help and assistance in data collection and assays.