Gene–Environment Interplay in Adolescent Drinking Behavior

Richard J. Rose, Ph.D., and Danielle M. Dick, Ph.D.

Many people begin to consume alcohol and establish drinking patterns during adolescence, making this an important developmental period for alcohol researchers to study. Both drinking initiation and establishment of drinking patterns are influenced to varying degrees by genetic as well as environmental factors. Using twin studies conducted in Finland and other countries, researchers have analyzed the specific genetic and environmental influences as well as the gene–environment interactions that shape drinking behavior in adolescence. These studies indicate that drinking initiation is determined primarily by environmental influences, whereas the establishment of drinking patterns is determined mostly by genetic factors, which themselves are subject to moderation by the environment. KEY WORDS: adolescent; young adult; age of onset; AOD (alcohol and other drug) use initiation; AOD use pattern; genetic theory of AODU (alcohol and other drug use); hereditary factors; environmental factors; hereditary vs. environmental factors; social environmental risk and protective factors; family risk and protective factors; longitudinal study; twin study; monozygotic twin; dizygotic twin; the FinnTwin studies; Finland; Collaborative Study on the Genetics of Alcoholism (COGA)

he years between early adolescence it also is necessary to study the factors also exists in people without a family and young adulthood are a cru­ influencing drinking initiation. history. Tcial period for alcohol researchers Another reason to focus on early But what is the relevance of this to study, because drinking typically is adolescent drinking comes from research association? In most cases, it is hazardous initiated during adolescence, and by evidence indicating that people who to draw a causal connection based on young adulthood, individual differences begin drinking at an earlier age have a such observations. Thus, it is possible in established drinking patterns are evi­ heightened risk for negative consequences dent. The earliest stages of alcohol use of drinking, such as alcohol abuse or RICHARD J. ROSE, PH.D., is a professor merit particular attention because alcohol dependence. For example, emeritus in the Department of Psychology, research has suggested that the initia­ interviews of adults consistently and Indiana University, Bloomington, Indiana. tion of alcohol use must be distin­ convincingly have shown a strong asso­ guished from the frequency, quantity, ciation between an early initiation of DANIELLE M. DICK, PH.D., is an assistant and density (i.e., the number of drinks drinking and later alcohol­related prob­ professor in the Departments of consumed in a certain length of time) lems: People who report drinking as Psychiatry and Psychology, Washington of alcohol use once drinking is initiated. early as age 13 or 14 are much more University, St. Louis, Missouri. Consequently, analyses in adults with likely to exhibit alcohol­related behavior already established drinking patterns can problems in adulthood than people who The FinnTwin studies are supported by explain only part of the mechanisms began drinking at a later age (Grant National Institute on Alcohol Abuse and and contributing factors underlying and Dawson 1997). The association of Alcoholism grants AA–00145, AA–08315, the development of alcohol abuse and drinking problems with earlier age of AA–09203, and AA–12502; by the dependence. To obtain a more compre­ initiation is particularly strong in people Academy of Finland; and by the Yrjö hensive explanation of these processes, with a family history of alcoholism, but Jahnsson Foundation.

222 Alcohol Research & Health Gene–Environment Interplay in Adolescent Drinking Behavior

that drinking initiation at an earlier age which the people live—that is, there interplay across development for a vari­ is causally related to greater vulnerabil­ are gene–environment interactions. ety of behaviors. Therefore, these ity to alcohol’s cumulative effects. A person’s susceptibility to develop­ research designs also are particularly Alternatively, early onset of drinking and ing alcohol dependence and other alcohol­ useful in understanding the develop­ a greater likelihood of later problems related problems manifests as an unfold­ ment of drinking behaviors. The most may simply be correlated—that is, they ing developmental process that involves widely used design is a longitudinal/ are independent signs or symptoms of both genetic and environmental factors developmental study of child or adoles­ people who are at elevated risk of alco­ and is continuously modulated by gene– cent twins and their parents. Several hol dependence. In the latter case, pre­ environment interplay. This concept research groups in the United States vention efforts designed to promote of a complex developmental process is and other countries are engaged in delays in age of drinking initiation supported by fundamental findings in developmental behavior­genetic research would do little to reduce the prevalence recent behavior­genetic research that of this kind, and many of these groups of adult alcoholism—an issue that is have profoundly altered the earlier view explore risk factors for alcoholism as important to resolve. One approach to that people are passive players who part of their research aims. This article doing this is to conduct longitudinal mechanically mirror their environmental focuses on two longitudinal studies of studies that follow, over several years, experience. Instead, recent research young Finnish twins and their parents genetically informative samples, such findings indicate the existence of exten­ known as the FinnTwin studies. as samples of twins who differ in their sive gene–environment correlations, Finland is particularly suitable for age at onset of drinking. whereby people are active agents who conducting longitudinal twin studies This article reviews recent findings help create the experiences to which because of the availability of compre­ from two ongoing longitudinal twin they are exposed, selectively seek envi­ hensive birth records, the ease of locat­ studies conducted in Finland that elu­ ronments to match their specific dispo­ ing twins as they grow up, and the high cidate the contributions of genetic and sitions, and modify their environments level of compliance of organizations environmental influences to initiation through their actions. In short, new and individuals with research requests. of drinking and later trajectories of research suggests that constant gene­by­ To acquire a large sample of twins that is representative of the entire pop­ alcohol use and abuse. Data from these environment interplay is involved in the development of alcoholism risk ulation, it is almost imperative that population­based FinnTwin studies help over time. investigators have access to a compre­ investigators address questions about To understand the factors affecting hensive birth registry for the region or factors contributing to various stages of the development of alcoholism, it is country under investigation. However, alcohol use. The resulting research has essential to appreciate this continuing many countries, including the United underscored the importance of the interaction of genetic and environmen­ States, do not have such registries. concept of dynamic gene–environment tal influences (e.g., how a child’s behav­ Some researchers in the United States interactions. ioral dispositions interact with that have attempted to obtain comprehen­ child’s household environment). Simi­ sive data sets by identifying all twin larly, it is important to acknowledge students enrolled in a large public The General Role of Genetic that the household environment does school system or by systematically and Environmental Factors not stand alone, but is “nested” within accessing birth records within a State in Alcohol­Related neighborhoods, schools, communities, and then searching for the current Behavior and cultures. Accordingly, the ways addresses of the twins so identified. in which a child expresses his or her In the Nordic countries, in contrast, Both genetic and environmental factors internal dispositions are modulated by the entire population is enrolled at contribute to the different stages in a the adult models, parenting practices, birth into a centralized population reg­ person’s drinking history. The factors peer interactions, access to alcohol, istry. In Finland, every newborn citizen that influence individual differences in and school and neighborhood environ­ is given a unique personal identifica­ whether, and at what age, drinking is ments to which the child is exposed. tion number that includes the person’s initiated are different in kind from the All of these factors either enhance or date of birth and a link to the biologi­ factors that influence individual varia­ reduce the risk for alcoholism that cal mother. Accordingly, researchers tions in patterns of drinking, once the results from the child’s genetic makeup. who have approval to access the registry acute subjective effects of alcohol inges­ can identify all people born on the tion are experienced. Moreover, it is same day to the same mother, thereby clear from numerous studies that The FINNTWINStudies identifying the nation’s entire popula­ although genetic factors have some tion of twins and other multiples. influence on drinking behavior, the Longitudinal studies using genetically Moreover, Finnish law requires all citi­ effects of genetic differences between informative samples enable researchers zens to report all residential moves in a people can be enhanced or reduced by to assess the magnitude of genetic and timely manner, which means that fami­ variations in the environments in environmental influences and their lies of twins can be readily located and,

Vol. 28, No. 4, 2004/2005 223 in principle, followed up without loss. trait or behavior are identical because by the results from FinnTwin12–17, Finally, for cultural and historic reasons, the co­twins share all their genes. In which addressed this issue. Nearly two­ compliance with requests for research DZ twins, however, the genetic influ­ thirds (64 percent) of the twins in this participation is unusually high in the ences can differ because the co­twins study reported never having used alco­ Nordic countries. This high compliance share only about 50 percent of their hol at age 14, whereas 36 percent rate and the requirement to report resi­ genes. With respect to environmental reported that they had begun drinking dential moves make it possible to con­ influences, there should be no substan­ alcohol at this age. Comparisons of the duct followup assessments with large tial differences between MZ and DZ MZ and DZ twin pairs in the study numbers of subjects, which otherwise twins because, in childhood and ado­ determined that the environmental fac­ presents a particular challenge when lescence, all or nearly all twin pairs live tors shared by co­twins played the researching drinking behavior and its with one another and with the same largest role in influencing abstinence/ consequences (Kaprio et al. 2002). biological parent or parents. Therefore, drinking at age 14. These factors— both MZ and DZ twins will share equally which included the familial environment

Design of the FinnTwinStudies important aspects of their household and the nonfamilial environments of experience, peers, neighborhood, school, peers, schools, and neighborhoods— The two FinnTwin studies differ in and other environmental factors that accounted for 76 percent of the varia­ the age groups they study and in their may enhance or reduce the risk for early tion in drinking initiation in both boys followup intervals. One study, Finn­ alcohol use and abuse. and girls (Rose et al. 2001b). Twin16–25, includes all twins born in Based on these premises regarding One relevant question in this con­ five consecutive birth cohorts, 1975 genetic and environmental influences, the text is whether the same factors influ­ through 1979, who were enrolled in following expectations can be articulated: ence drinking initiation in 14­year­old the study when they reached age 16. boys and girls. Although the magnitude Systematic followup surveys of these • If genetic factors are primarily impor­ of importance of environmental factors twins were completed at ages 17 and tant in the development of a mea­ is the same in boys and girls, it is possi­ 18.5 and at one point when the twins sured trait or behavior, MZ co­twins ble that different environmental factors were between ages 22 and 25. The sec­ will resemble one another about may be acting in each gender. Twin ond study, FinnTwin12–17, identified twice as much as do DZ co­twins. studies allow researchers to estimate the all twins in the five consecutive birth extent to which the same or different cohorts 1983 through 1987 who were • If genetic factors are of only modest environmental factors influence boys enrolled at age 11 or 12, with followup importance and shared environmen­ and girls. In the FinnTwin12–17 study, assessments at ages 14 and 17. tal influences are predominant, the 77 percent of the environmental factors Each FinnTwin study includes about two kinds of twins will show com­ that contribute to the twin siblings’ 2,500 pairs of twins, with equal propor­ parable resemblance for the trait or similarity in age of drinking initiation tions of brother–sister, brother–brother, behavior studied, because both share are the same for adolescent males and and sister–sister twin pairs. Among the the environmental influences rele­ females (i.e., the correlation between same­sex pairs, about half are identical, vant to that outcome. environmental factors was estimated at or monozygotic (MZ) twins and half 0.77). Thus, although many environ­ are fraternal, or dizygotic (DZ) twins. The following sections summarize mental influences impacting drinking MZ twins develop when a single egg is some of the findings of the FinnTwin initiation are the same for girls and fertilized by a single sperm, and the studies regarding the contributions of boys, some unidentified environmental resulting embryo divides shortly after various genetic and environmental fac­ influences are specific to each gender. fertilization to produce two individuals tors to a variety of drinking­related Although the FinnTwin data pro­ who share all their genes. DZ twins, in behaviors. In general, the results from vided little evidence of genetic effects contrast, develop when two egg cells the FinnTwin studies have been consis­ on drinking initiation by age 14, some are independently fertilized by separate tent with ongoing research with adoles­ gender differences were apparent in sperm at about the same time (or in cent twins in other European countries that respect as well. Genetic effects close temporal proximity), resulting in as well as in the United States. were found to have a modest influence the development of two individuals on girls (accounting for 18 percent of who, like ordinary siblings, share, on the variance in girls’ drinking at age average, half of their genes. The Initiation of 14) but no influence on boys’ drinking The fundamental approach in twin Alcohol Use at the same age. Perhaps this difference study research is to compare the extent reflects the fact that girls are more to which MZ twins resemble each other Twin studies consistently have demon­ developmentally mature at age 14 than in relation to a specific trait or behavior strated that initiation of alcohol use is boys. By this age, most girls already with the resemblance between DZ largely influenced by environmental have undergone significant pubertal twins for the same trait or behavior. In rather than genetic factors (Hopfer et development, whereas most boys are MZ twins, the genetic influences on that al. 2003), a conclusion that is supported just beginning to experience pubertal

224 Alcohol Research & Health Gene–Environment Interplay in Adolescent Drinking Behavior

changes. Girls’ more advanced matura­ Despite these differences, the two resemblance of unrelated classmates of tion may foster associations with older sets of symptoms are significantly cor­ twin siblings to one another and to friends and peers that offer greater access related (0.50), meaning that the CD their classmate twins. to alcohol, thereby creating opportuni­ symptoms were more common among The FinnTwin12–17 study under­ ties for girls to express drinking­related twins with a diagnosis of alcohol took such an approach (Rose et al. genetic predispositions at an earlier age. dependence and, similarly, the symp­ 2003). For this analysis, researchers toms of alcohol dependence were more linked each twin with a classmate con­ common among twins diagnosed with trol subject of the same gender who Symptoms of Alcohol CD. However, because the twin data had the same year of birth, lived in the Dependence at Age 14 produced no evidence that genetic fac­ same neighborhood, and attended the tors contribute to alcohol dependence same school and classroom as the twin. It is uncommon for 14­year­olds to symptoms at this age, the association of As discussed earlier, behavioral exhibit symptoms of alcohol depen­ alcohol symptoms with CD symptoms resemblance of co­twins results from dence, and at this age such symptoms must be attributed entirely to environ­ their shared genetic makeup and from are not attributable to genetic effects. mental factors that influence both con­ the household experiences they share To explore the issue of alcohol depen­ ditions (Rose et al. 2004). One inference with each other and with their parents, dence at this early age in more detail, from these results is that CD is an early as well as experiences they share with researchers administered a structured manifestation of a genetic predisposi­ peers outside the home. In contrast, face­to­face interview to a sample of more tion that later contributes to the devel­ any behavioral resemblance between than 1,800 twins from the FinnTwin opment of alcohol­related problems and the twins and their classmates (who are 12–17 study. Only 12.4 percent of the alcohol dependence. Therefore, efforts genetic strangers and share no house­ 14­year­old twins in this sample reported to identify adolescents at risk for devel­ hold experiences with the twins being any symptoms of alcohol dependence, oping alcoholism and to create oppor­ studied) must result entirely from expe­ and only 1 percent met the criteria for tunities for targeted interventions should riences shared in school and neighbor­ a diagnosis of alcohol dependence. More­ focus on identifying adolescents exhibit­ hood. Accordingly, this research approach over, more girls than boys reported ing symptoms of CD. Other research provides a direct estimate of the role of alcohol dependence symptoms or met designs, such as studies of adopted chil­ environmental influences associated diagnostic criteria; this gender differ­ dren and their biological and adoptive with schools and neighborhoods. ence may be attributable to the fact families, also may yield important In the FinnTwin12–17 study, that, at this age, girls are more likely information on the genetic relationship behavior of the twins and the classmate than boys to be drinking. between CD and alcohol dependence. control subjects was assessed when the For comparison, the investigators children were 11 or 12 years old by also determined the frequency of symp­ means of a questionnaire administered toms of conduct disorder (CD) in the Identifying Specific during school hours under research same sample of twins. This analysis Environmental Influences supervision. The questionnaire included yielded several results that differed sharply on Drinking three drinking­related items illustrating from those obtained regarding alcohol nonfamilial environmental effects: dependence symptoms (Rose et al. 2004): The environmental factors shared by siblings, and assessed in twin studies • “Have you ever drunk alcohol?” • Symptoms of CD were common such as FinnTwin, include not only among the adolescent twins: 44 household environmental characteristics • “Have you ever seen any of your percent of the sample reported one (such as family structure and status, friends drunk?” or more symptoms of CD, and 12 parenting practices, and home atmo­ percent met diagnostic criteria for sphere) but also environmental influences • “Have you ever drunk alcohol with the condition. exerted through schools, neighborhoods, your friends without adults around?” and communities. For some behaviors • CD symptoms were more common and at some ages, the nonfamilial expe­ The prevalence of affirmative answers among boys, who accounted for riences that children share with their to these questions differed widely, as 65 percent of all twins diagnosed peers at school and in their neighbor­ would be expected among children at with CD. hoods may be more salient than the this age. For example, more than two­ household influences they share with thirds of the boys and more than half • Genetic factors played a much different their families in their homes. Drinking of the girls admitted to having had a role in CD than in alcohol depen­ in early adolescence may be one such drink of alcohol—which is not uncom­ dence. Whereas genetic factors had behavior. To evaluate this hypothesis, mon in Finnish family culture, where a negligible influence on alcohol researchers can assess the contribution children may have a drink at holiday dependence, they had a substantial of nonfamilial environments to chil­ time and with parental knowledge. In influence on symptoms of CD. dren’s behavior by comparing the contrast, only 18 percent of boys and

Vol. 28, No. 4, 2004/2005 225 11 percent of girls reported having seen accordingly. Findings from the FinnTwin percent of twins at age 16 to only friends drunk, and only 7 percent of 16–25 study illustrate this process. about 10 percent of twins at age 18.5. the boys and girls reported ever having One component of Conversely, the proportion of twins drunk alcohol with their friends with­ FinnTwin16–25 was an assessment of who reported drinking at least twice out adults around. the twins at ages 16, 17, and 18.5. monthly more than doubled over the The investigators determined the Over this 30­month period of followup, 30­month period, and at age 18.5 more relative influences of genetic and specific the number of abstinent participants than 60 percent of twin boys and girls environmental factors on these three decreased dramatically, from about 25 drank that often (Rose et al. 2001a). variables based on the relative similarities of MZ and DZ twins and their classmate control subjects (see figure 1). They found that genetic effects accounted for only 40% 8 percent of the individual differences in whether the children had begun to 30% use alcohol, but they accounted for 30 20% percent of the variance in alcohol use without adult supervision (Rose et al. 10% 2003). Conversely, familial effects were 0%

very influential in initiation of use, Variance Familial � School � Personal � accounting for nearly 40 percent of the Genetics Environment Environment Environment variance. They had little influence, how­ Factors Influencing Drinking Behavior: ever, on whether these children had ever Have you ever drunk alcohol?

seen any of their friends drunk, which Observed

was more closely linked to genetic to 40% influences and the effects of the child’s personal experiences (i.e., experiences 30% not shared with the twin sibling). Finally, 20% experiences shared with school class­ mates contributed significantly to each Contribute 10% of these alcohol­related experiences. 0% Familial � School � Personal � Thus, this expanded twin–classmate Genetics Environment Environment Environment control design has identified important Factors Factors Influencing Drinking Behavior:

influences on drinking behavior that the originate both within and outside the Have you ever seen friends drunk? family environment. These findings 40% document the varying influences of a Much person’s genetic dispositions on different 30% aspects of drinking behavior in early How 20% adolescence. 10% 0% Drinking Patterns Across Familial � School � Personal � Adolescence Genetics Environment Environment Environment Factors Influencing Drinking Behavior: The relative contributions of genetic Have you ever drunk with your friends without adults? and environmental factors to the drink­ ing behavior of young people change during adolescence. As previously Figure 1 The relative contributions of genetic factors, familial environment, school described, familial and environmental environment, and personal environment to drinking­related experiences influences predominate in their impact of Finnish twins ages 11 and 12. Influences from the familial environment on initiation of drinking behavior. are shared by twin siblings but not shared with their unrelated classmates; When people move from early experi­ school­based neighborhood effects are shared fully by twin siblings and their classmates. Data were obtained from 631 same­sex twin pairs (i.e., mental alcohol use to more established 1,262 twins) and a matched, same­sex classroom control subject for patterns of use, however, the importance each twin. of individual genetic predispositions increases greatly, and the importance of SOURCE: Rose et al. 2003. common environmental factors decreases

226 Alcohol Research & Health Gene–Environment Interplay in Adolescent Drinking Behavior

Interestingly, the influence of genetic Gene–Environment tion in drinking patterns among rural factors also increased across this age Interaction youth, but for only one­third of the range. Genetic factors accounted for variance in youth living in urban set­ only 33 percent of the variation in drink­ Although the contribution of genetic tings. This finding has been replicated ing frequency at age 16, but 50 percent influences to alcohol use increases dur­ recently in a twin study conducted in of the variation by age 18 (Rose et al. ing adolescence, environmental factors Minnesota (Legrand et al. 2003). 2001a). At the same time, common do not become irrelevant but play an Subsequent analyses have attempted environmental factors became less important role in moderating the influ­ to identify the processes underlying this important, accounting for 37 percent ence of genetic predispositions. For phenomenon. What characteristics of of the variance at age 16, but only 14 example, findings from the FinnTwin urban settings allow for greater expression percent by age 18 (see figure 2). This studies demonstrated that urban or of genetic effects? Why do environ­ shift in the relative influence of envi­ rural residency has a significant impact mental influences exert a larger influ­ ronmental factors is even more dramatic on the relative contributions of genetic ence in rural settings? To explore these when one includes the data on drink­ and other environmental factors which questions, researchers have expanded ing initiation from 14­year­old twins affect adolescent drinking. Genetic the classic twin model to incorporate in the FinnTwin12–17 study. Taken influences on drinking frequency were more continuous measures of the envi­ together, these findings indicate that much stronger among twins living in ronment and examine a series of socio­ familial and community factors appear urban settings, where they accounted regional variables that they hypothesized to have their greatest impact on adoles­ for 34 percent of the variance in drinking to be more closely related to the inter­ cents’ decisions to initiate alcohol use at age 16, than among twins residing in action effect (Dick et al. 2001). and on their early experimental drinking; rural Finland, where they accounted Finland offers an opportunity to as drinking patterns develop, differentiate, for 18 percent of the variance in drink­ explore such questions because the and stabilize across adolescence, how­ ing (Rose et al. 2001a). Conversely, country is divided into municipalities— ever, genetic influences on drinking common environmental effects accounted local government units that levy taxes patterns become increasingly important. for more than 50 percent of the varia­ and assume responsibility for educa­ tion, health care, and social services. For these municipalities, whether in the metropolitan Helsinki area or in

80 sparsely populated areas of Finnish Genetic Influences Lapland and the Åland Islands, a wealth 70 of sociodemographic information is Common Environmental � available that may be relevant to risk 60 Influences for adolescent drinking. 50 To study the influence of community­ level factors on adolescent drinking Variance 40 outcomes, Dick and colleagues (2001) 30 linked this sociodemographic informa­

Percent tion to each twin pair’s residency code. 20 By doing so, the investigators were able 10 to identify characteristics of the com­ munities in which each twin pair lived 0 that could be related to risk. The specific 14 16 17 18.5 sociodemographic variables assessed to Age (Years) characterize the urban/rural interaction effect included the following:

• Percentage of young adults, ages Figure 2 The changing influence of genetic and common environmental factors on drinking behavior during adolescence. Data for age 14 illustrate influ­ 20–24, in a municipality. ences on initiation of drinking and were taken from the FinnTwin12–17 study. Data for all other ages illustrate influences on drinking frequency • Neighborhood instability, as indi­ and were taken from the FinnTwin16–25study. Genetic influences at cated by the percentage of migration age 14 are significant only for girls; boys show no evidence of genetic into and out of a given region. influences on drinking initiation at this age. • Per capita expenditure on alcohol in SOURCE: Rose et al. 2001a,b. each region, relative to the mean amount spent in the entire country.

Vol. 28, No. 4, 2004/2005 227 The researchers found that each of potential moderating role of environ­ of Alcoholism (COGA). COGA these variables had dramatic effects on mental factors, such as parenting prac­ researchers recently published reports the genetic and environmental influences tices and parental models (D’Onofrio of several genes associated with alcohol on drinking behavior, with a more than et al. 2003). dependence in adults (Edenberg et al. fivefold difference in the magnitude of 2004; Dick et al. 2004a; Wang et al. genetic effects between environmental 2004), and some of these findings already extremes. Thus, communities charac­ Molecular Genetic have been replicated by other investiga­ terized by more young adult role models, Advances in Understanding tors (Covault et al. 2004; Lappalainen greater social mobility, and higher alcohol Drinking Patterns et al. 2005; Luo et al. 2005). sales allowed for increased expression The next challenge will be to expand of genetic dispositions that contribute Advances in genetic analysis and in the these findings to children and adoles­ to individual differences in adolescent technology used to determine a per­ cents to better understand how specific drinking. On the other hand, communi­ son’s genetic makeup (i.e., genotype) predisposing genes influence trajecto­ ties with more social stability created an now make it possible to move beyond ries of drinking and other behavioral environment in which common envi­ estimates of unmeasured (i.e., latent) problems across development. The ronmental effects, within families and sources of genetic variance. Researchers identification of specific genes influenc­ within communities, assumed greater have begun to identify specific genes ing alcohol use and abuse will enable importance. involved in complex behavioral out­ researchers to begin to decipher how Analyses of other twin data sets also comes such as alcohol use. the effects of these genes are modified have demonstrated the importance of Using a candidate gene approach, by specific environmental influences. environmental factors in moderating which examines the influence of spe­ Elucidating the influence of specific genetic and environmental influences cific genes suspected of playing a role genes on various aspects of drinking, on drinking patterns. In Australian in the behavior under investigation, the actions of these genes, and their twin data, marital status moderated the two studies recently reported associations interactions with specific environmen­ importance of genetic effects on alcohol between specific gene variants and tal risk factors should dramatically consumption in women (Heath et al. drinking behavior among college stu­ enhance our understanding of the 1989)—that is, living in a marriage­ dents. The first study investigated the development of alcohol use and abuse. like relationship reduced the impact of role of variation in the gene encoding a genetic influences on drinking. In this protein called the serotonin transporter. study, genetic factors accounted for only This study found that in a sample of Conclusions and Future half as much variance in drinking among White students, those with a particular Outlook married women (31 percent) as among version of this gene engaged in binge unmarried women (60 percent) (Heath drinking more often, drank to intoxica­ Both genetic and environmental factors et al. 1989). tion more often, and consumed more contribute to the development of pat­ In another study, on the effect of alcoholic drinks per drinking occasion terns of alcohol use during adolescence. religiosity on alcohol use among women, than students with other variants of the The relative importance of genetic and genetic factors played a significant role gene (Herman et al. 2003). environmental factors, however, varies in women who had not had a religious Another study focused on the gene for different phases of alcohol use and upbringing, dwarfing the influence of that encodes aldehyde dehydrogenase different stages of development. Environ­ shared environment. By contrast, there (ALDH), an enzyme involved in the mental influences originating both was no evidence of genetic influence in breakdown of alcohol in the body. This within the family and in the child’s women with a religious upbringing, study reported that Asian American school and neighborhood greatly influ­ but shared environment had a substan­ college students who carried a particular ence the initiation of alcohol use. Once tial impact (Koopmans et al. 1999). version of the ALDH gene which results drinking has been initiated, however, Lastly, researchers recently have begun in less efficient alcohol breakdown were genetic influences assume increasing to explore the potential moderating less likely to be regular drinkers and importance across development, as role of parenting factors on adolescent engage in binge­drinking episodes; they individual differences in patterns of alcohol use. Preliminary analyses have also reported a lower number of maxi­ drinking emerge. The degree to which provided strong evidence that changes mum drinks consumed in a 24­hour genetic effects contribute to adolescent in the relevance of genetic and environ­ period than Asian students with other drinking patterns also is subject to mental factors are related to differences ALDH variants (Wall et al. 2001). moderation by the environment, with in parental monitoring, involvement, Such studies are complemented by factors such as religiosity, marital status, and support (Dick et al. 2005). large­scale efforts to identify genes that and regional residency all contributing Complementary twin designs, such contribute to alcoholism. One of these to the relative importance of genetic and as studies of the offspring of adult twins, efforts, funded by the National Institute environmental effects on alcohol use. also provide rich opportunities to study on Alcohol Abuse and Alcoholism, is Other recent efforts have been gene–environment interaction and the the Collaborative Study on the Genetics directed toward identifying specific

228 Alcohol Research & Health Gene–Environment Interplay in Adolescent Drinking Behavior

genes contributing to variation in drink­ by a polymorphism in the 5-HTT gene. Science iors: Studies on Finnish twins and twin families. ing patterns. As scientists begin to 301:386–389, 2003. PMID: 12869766 Twin Research 5:358–365, 2002. PMID: 12537860 understand more about the nature of COVAULT, J.; GELERNTER, J.; HESSELBROCK, V.; ET KENDLER, K.S., AND EAVES, L.J. Models for the gene–environment interactions, molec­ AL. Allelic and haplotypic association of GABRA2 joint effect of genotype and environment on liabil­ ular genetic studies can be designed with alcohol dependence. American Journal of ity to psychiatric illness. AJP: The American Journal Medical Genetics (Neuropsychiatric Genetics) of Psychiatry 143:279–289, 1986. PMID: 3953861 that take this emerging information 129B:104–109, 2004. PMID: 15274050 into account (Kendler and Eaves 1986). KOOPMANS, J.R.; SLUTSKE, W.S.; VAN BAAL, DICK, D.M.; ROSE, R.J.; VIKEN, R.J.; ET AL. Exploring The potential of the molecular genetics G.C.M.; AND BOOMSMA, D.I. The influence of gene–environment interactions: Socioregional mod­ religion on alcohol use initiation: Evidence for approach has been illustrated by a eration of alcohol use. Journal of Abnormal genotype x environment interaction. Behavior recent study that found an association Psychology 110:625–632, 2001. PMID: 11727951 Genetics 29:445–453, 1999. PMID: 10857249 between a variant of the serotonin DICK, D.M.; EDENBERG, H.J.; XUEI, X.; ET AL. transporter gene and major depression, LAPPALAINEN, J.; KRUPITSKY, E.; REMIZOV, M.; ET Association of GABRG3 with alcohol dependence. AL. Association between alcoholism and gamma- but only in people who had experi­ Alcoholism: Clinical and Experimental Research amino butyric acid alpha2 receptor subtype in a enced stressful life events (Caspi et al. 28:4–9, 2004a. PMID: 14745296 Russian population. Alcoholism: Clinical and 2003). Large-scale collaborative pro­ DICK, D.M.; PURCELL, S.; VIKEN, R.J.; ETAL. Experimental Research 29:493–498, 2005. PMID: jects have begun to identify genes Identifying environmental influences on adolescent 15834213 substance use in the Finnish Twin studies. Behavior involved in alcohol dependence and LEGRAND, L.; MCGUE, M.; KEYES, M.; ET AL. Do Genetics 34:636–637, 2004b. related behaviors (Dick et al. 2004a; rural environments constrain genetic expression of Edenberg et al. 2004; Wang et al. 2004). DICK, D.M.; ROSE, R.J.; AND KAPRIO, J. The next externalizing psychopathology? Behavior Genetics Once researchers have identified challenge for psychiatric genetics: Characterizing 33:710, 2003. the risk associated with identified genes. Annals of specific genes associated with complex Clinical Psychiatry Special Issue: Remi Cadoret LUO, X.; KRANZLER, H.R.; ZUO, L.; ET AL. disorders such as alcoholism, they can Festschrift, 2005. CHRM2 gene predisposes to alcohol dependence, begin to address questions regarding drug dependence and affective disorders: Results D’ONOFRIO, B.M.; TURKHEIMER, E.N.; EAVES, from an extended case-control structured associa­ developmental effects or gene–environ­ L.J.; ET AL. The role of the children of twins design tion study. Human Molecular Genetics 14:2421– ment interactions associated with these in elucidating causal relations between parent char­ 2434, 2005. PMID: 16000316 genes (Dick et al. 2004b). Such ques­ acteristics and child outcomes. Journal of Child tions include whether genetic variants Psychology and Psychiatry 44:1130–1144, 2003. ROSE, R.J.; DICK, D.M.; VIKEN, R.J.; AND KAPRIO, PMID: 14626455 J. Gene–environment interaction in patterns of affect different behaviors at different adolescent drinking: Regional residency moderates developmental stages, whether particular EDENBERG, H.J.; DICK, D.M.; XUEI, X.; ET AL. longitudinal influences on alcohol use. Alcoholism: genetic variants are more or less impor­ Variations in GABRA2, encoding the α2 subunit Clinical and Experimental Research 25:637–643, of the GABA(A) receptor, are associated with alco­ 2001a. PMID: 11371711 tant at various ages, or whether the hol dependence and with brain oscillations. risk associated with a specific gene varies American Journal of Human Genetics 74:705–714, ROSE, R.J.; DICK, D.M.; VIKEN, R.J.; ET AL. according to environmental context. 2004. PMID: 15024690 Drinking or abstaining at age 14: A genetic epi­ demiological study. Alcoholism: Clinical and Integrating the knowledge obtained GRANT, B., AND DAWSON, D. Age at onset of alco­ from behavior-genetic studies with Experimental Research 25:1594–1604, 2001b. hol use and its association with DSM–IV alcohol PMID: 11707634 findings emerging from the areas of sta­ abuse and dependence: Results from the National tistical and molecular genetics promises Longitudinal Alcohol Epidemiological Survey. ROSE, R.J.; VIKEN, R.J.; DICK, D.M.; ET AL. It does to enhance our understanding of the Journal of Substance Abuse 9:103–110, 1997. take a village: Nonfamilial environments and children’s PMID: 9494942 behavior. Psychological Science 14:273–277, 2003. factors that contribute to drinking PMID: 12741753 among adolescents and young adults. HEATH, A.C.; JARDINE, R.; AND MARTIN, N.G. Interactive effects of genotype and social environ­ ROSE, R.J.; DICK, D.M.; VIKEN, R.J.; ET AL. ment on alcohol consumption in female twins. Genetic and environmental effects on conduct dis­ Journal of Studies on Alcohol 50:38–48, 1989. order, alcohol dependence symptoms, and their Acknowledgment PMID: 2927121 covariation at age 14. Alcoholism: Clinical and HERMAN, A.I.; PHILBECK, J.W.; VASILOPOULOS, Experimental Research 28:1541–1548, 2004. The FinnTwin studies are conducted N.L.; AND DEPETRILLO, P.B. Serotonin transporter PMID: 15597087 in collaboration with Professors Jaakko polymorphism and differences in alcohol consump­ WALL, T.L.; SHEA, S.; CHAN, K.K.; AND CARR, Kaprio and Markku Koskenvuo at the tion behaviour in a college student population. L.G. A genetic association with the development of University of Helsinki and Professor Alcohol and Alcoholism 38:446–449, 2003. PMID: alcohol and other substance use behavior in Asian 12915525 Lea Pulkkinnen at the University of Americans. Journal of Abnormal Psychology 110:173– Jyvaskylä, Finland. ■ HOPFER, C.; CROWLEY, T.J.; AND HEWITT, J.K. 178, 2001. PMID: 11261392 Review of twin and adoption studies of adolescent WANG, J.C.; HINRICHS, A.L.; STOCK, H.; ET AL. substance use. Journal of the American Academy of Evidence of common and specific genetic effects: Child & Adolescent Psychiatry 42:710–719, 2003. Association of the muscarinic acetylcholine receptor References PMID: 12921479 M2 (CHRM2) gene with alcohol dependence and CASPI, A.; SUGDEN, K.; MOFFITT, T.E.; ET AL. KAPRIO, J.; PULKKINEN, L.; AND ROSE, R.J. Genetic major depressive syndrome. Human Molecular Influence of life stress on depression: Moderation and environmental factors in health-related behav­ Genetics 13:1903–1911, 2004. PMID: 15229186

Vol. 28, No. 4, 2004/2005 229