Addicted to Nicotine
A National Research Forum

Section II: Nicotine-Individual Risk Factors for Initiation
Richard R. Clayton, Ph.D., Chair

THE GENETIC EPIDEMIOLOGY OF SMOKING

Kenneth S. Kendler, M.D.
Medical College of Virginia
Virginia Commonwealth University

Introduction

The goal of this presentation is to review the use of genetically informative samples to clarify the causes of variation in the use of tobacco products and in nicotine dependence. While a number of studies in rodents have suggested genetic influences on brain nicotine receptors and the development of tolerance to nicotine, this presentation focuses entirely on data from humans.

What We Know

• To What Extent Is Smoking Heritable? Eight studies, published between 1958 and 1995 from seven different countries, have reported some measure of concordance for "smoking status" in samples of monozygotic (MZ) and dizygotic (DZ) twin pairs. These studies together contain 891 MZ pairs, of whom 660 or 74.1±1.5 percent are concordant for smoking status. The parallel figures for DZ twins are 1,072 pairs, of whom 602 or 56.2±1.5 percent are concordant.

  There are several reasons to be skeptical of these results. Several measures were used, although most studies examined either current or lifetime regular smoking. Given the importance of peer group exposure in smoking initiation and the evidence that MZ twins socialize more closely in adolescence than do DZ twins, the MZ/DZ difference could be partly a result of environmental effects. This issue can be partly addressed by four small sample studies of MZ twins reared apart. Of the 108 pairs, concordance for smoking status was seen in 82 or 75.9±4.0 percent, a figure similar to that seen in MZ twins reared together. In addition, three investigations examined whether the greater similarity for smoking in reared together MZ versus DZ twins could be the result of environmental effects and concluded that it could not be.

  More recently, correlational and heritability analyses of lifetime regular smoking in adult MZ and DZ twins have been reported in five samples containing a total of 4,719 MZ and 3,147 same-sex DZ pairs. The weighted correlation in liability to lifetime regular smoking was +0.80 in the MZ pairs and +0.53 in the DZ pairs. These results suggest that the heritability of liability to regular smoking is likely 50 to 60 percent. Of note, twin resemblance also appears to be influenced by the family environment, which appears to contribute approximately 25 percent of the variance in liability. Finally, one adoption study of cigarette smoking reported significant correlations for cigarette consumption between biological parents and their offspring (+0.21 ) and biological siblings (+0.11 to +0.30), but not between adoptive parents and their adoptive offspring (-0.02) or adoptive siblings (+0.05).

  These data show that tobacco consumption, especially in the form of cigarettes, is substantially influenced by genetic factors. Some familial resemblance for smoking behavior is also mediated through familial environmental factors.

• Is Nicotine Dependence (ND) Heritable? Several twin studies have examined characteristics that indirectly reflect ND, particularly smoking persistence (i.e., current versus former smoker) and number of cigarettes consumed. Among twin pairs where both members have used cigarettes, these studies uniformly find greater resemblance in MZ versus DZ pairs for persistence or for number of cigarettes consumed. Heritability estimates for these proxy measures of ND are mostly in the range of +0.50 to +0.70. In female-female twin pairs from the Virginia Twin Registry, ND was directly assessed utilizing items from the widely used Fagerstrom Tolerance Questionnaire and from DSM-III-R, and attempts were made to clarify the relationship between the risk factors for smoking initiation (SI) and ND. Significant commonality was found in the risk factors for SI and for ND. However, there also was evidence for familial factors (which are probably at least in part genetic) that specifically influence the risk for ND given prior SI. Results suggest that "new genes" are expressed by exposure to smoking that influence the probability of development of ND. Estimate of the heritability of ND in this sample was about 70 percent, of which about one-third was due to genetic risk factors expressed subsequent to SI.

  The available information strongly suggests that the pathway to ND is complex and involves multiple genetic and environmental risk factors. Substantial evidence indicates that smoking behavior is significantly influenced by genetic factors, although not to the exclusion of a range of social environmental factors. Although considerably less is known about the role of genetic factors in the etiology of ND per se, accumulating evidence also suggests that genetic factors play an important role here as well. Efforts at prevention and treatment of both SI and subsequent ND may need to take in account the large individual differences in vulnerability that are, in part, genetically determined.

What We Need To Know More About

• Which Specific Genes Influence Smoking Behavior and ND? In the past several years, the gene-finding methods of human molecular genetics have begun to be applied to a range of complex, genetically influenced human traits, including ND. These studies are still in their infancy and have produced no definitive or replicated results. Given the complexity of this trait, quite large sample sizes may be required. A number of plausible "candidate genes" for ND are under investigation, including enzymes that metabolize nicotine, nicotinic receptor subunits, and dopamine receptors that may modulate the hedonic response to nicotine. It is likely, but by no means certain, that important advances will be made in this line of research in the upcoming decade. The existence of such studies that examine SI is uncertain.

• How Do Genetic Risk Factors Influence Risk for Smoking and ND? Do these risk factors largely operate on broad domains of personality or on specific biological systems (e.g., nicotine metabolism, brain nicotinic or dopamine receptors)? Might there be a dynamic interaction between genetic and environmental risk factors, such as temperamental traits influencing the selection of peer groups and lovers, that in turn affects the risk for smoking initiation?

• Are Genetic Risk Factors for Smoking the Same for Males and Females? Some studies show lower correlations in smoking behavior among opposite-gender versus same-gender fraternal twin pairs, suggesting distinct familial influences on smoking in men and women.

• How Specific Are the Genetic Risk Factors for Smoking and ND? To what extent are liability genes for smoking or ND unique to tobacco products versus shared with other forms of psychoactive substance use? Do the genetic risk factors for smoking or ND also predispose a person to psychiatric disorders such as major depression or anxiety states?

Recommended Reading

Heath, A.C.; Cates, R.; Martin, N.G.; Meyer, J.; Hewitt, J.K.; Neale, M.C.; and Eaves, L.J. Genetic contribution to risk of smoking initiation: Comparisons across birth cohorts and across cultures. J Subst Abuse 5:221-246, 1993.

Heath, A.C., and Madden, P.A.F. Genetic influences on smoking behavior. In: Turner, J.R.; Cardon, L.R.; and Hewitt, J.K., eds. Behavior Genetic Approaches in Behavioral Medicine. New York: Plenum Press, 1995. pp. 45-66.

Kendler, K.S.; Neale, M.C.; MacLean, C.J.; Heath, A.C.; Eaves, L.J.; and Kessler, R.C. Smoking and major depression: A causal analysis. Arch Gen Psychiatry 50:36-43, 1993.

Swan, G.E., and Carmelli, D. Behavior genetic investigations of cigarette smoking and related issues. In: Noble, E.P., and Blum, K., eds. Handbook of Psychiatric Genetics. Boca Raton: CRC Press, 1997. pp.379-398.