Alcohol abuse, or alcoholism, can be defined as "a heterogeneous set of behaviours that includes any pattern of alcohol intake that causes medical and/or social complications" (Cloninger, Bohman and Sigvardsson, 1981).
It has long been noted that alcoholism "runs in families". However, "familial" is not necessarily the same as "hereditary" (Goodwin, 1987). Speaking a native language may also be "familial", but it is not hereditary.
There are different genetic types of alcoholism. Cloninger et al. (1981) distinguished two forms of alcoholism. One type was a milieu-limited or environment related type of alcoholism, associated with recurrent alcohol abuse, but without criminality in the biological parents. The other type was found to be highly heritable and was associated with criminality and necessary treatment for alcohol abuse in the biological parents.
These two types of alcoholism were later classified as Type I and Type II. Milieu-limited (Type I) alcoholism occurs in both men and women, has a later age of onset, is less severe, and is not often associated with social problems such as fighting and arrests. Male-limited (Type II) alcoholism occurs mainly in males, has an earlier age of onset, a more severe course, and more alcohol-related social problems (Devor and Cloninger, 1989).
Cloninger (1987) noted that the tendency to begin seeking alcohol was different to susceptibility of loss of control once drinking had begun, suggesting that there may be different genetic contributions in different types of alcoholism. However, most alcoholics cannot be classified as one type or the other, since many of them have features of each type.
It is highly unlikely that the same genes cause alcoholism in all families (Gordis et al., 1990). Alcoholism genes may show incomplete penetrance, and non-genetic forms of alcoholism may be present in the same families where genetically influenced alcoholism is present.
Genetic analysis of behaviour is complex because behaviour reflects both genetic and environmental influences. Behaviour is influenced by many genes, each having small effects (Plomin, 1990). Behavioural studies in humans are difficult because it is impossible to control all variables. While studies in animals allow for adequate environmental controls, no animal models for human behavioural problems (especially alcoholism) have been developed (Schuckit, 1984).
Clearly, any research into the genetics of alcoholism will be difficult due to the heterogeneity of the disorder, as well as the multitude of genetic and environmental influences involved.
Twin, family and adoption studies have all shown that alcoholism has a genetic component.
Twin studies allow the comparison of concordance rates between identical (monozygotic) and non-identical (dizygotic) twins (Saunders and Phillips, 1993). The concordance rate is the rate at which a trait expressed in one twin is also expressed in the other. If a condition has a genetic component, then the concordance rate is expected to be higher in monozygotic twins, who share all of their genes, than in dizygotic twins, who share only half of their genes. Twin studies can give an indication of the environmental and genetic components of phenotypic variance (Devor and Cloninger, 1989).
Studies on over 16,000 twin pairs have shown that concordance rates for alcoholism are higher among monozygotic twins than dizygotic twins (Loehlin, 1972; Hrubec and Omenn, 1981; and Pickens et al., 1991). Concordance rates for alcoholism in monozygotic twins show that environmental variables, in addition to genetic variables influence alcoholism in twins (Tabakoff and Hoffman, 1988).
Families are useful for genetic analysis because they tend to be more closely related in ethnic and biological makeup than populations, and so are less susceptible to stratification or admixture (Baron, 1993).
An alcoholic is more likely to have an alcoholic first degree relative than a non-alcoholic. Rates of alcoholism have been found to be higher in relatives of alcoholics than in relatives of non-alcoholics (Cotton, 1979; Winokur et al., 1970).
Another approach involves studying individuals who have been separated from their biological parents soon after birth and raised by non-related people (Goodwin et al., 1973). If genetic factors are involved in alcoholism, then the children of alcoholics should show a higher incidence of alcoholism than the children of non-alcoholics (Noble, 1992).
Adoptees who had an alcoholic biological parent but who were raised by non-alcoholics were found to be more likely to develop alcohol-related problems in adult life than similar adoptees whose biological parents were not alcoholics (Goodwin et al. 1973, 1974; Cadoret and Gath, 1978; Bohman, 1978; Cloninger, Bohman and Sigvardsson, 1981; and Cadoret et al., 1986).
Many studies have investigated electrophysiological and biochemical markers that may cause, or be associated with, susceptibility to alcoholism. Physiological markers for alcoholism are markers which must not be caused by alcoholism. A candidate marker for alcoholism must have been present before the development of alcoholism, and should remain present during long periods of abstinence. Studies of animal models have been useful in distinguishing potential markers. Markers of an increased risk for alcoholism may be directly involved in the etiology of alcoholism, or they may be factors that do not directly predispose alcoholism, but are still associated with the development of alcoholism in an individual (Tabakoff and Hoffman, 1988).
Ethanol ingestion has an effect upon the central nervous system (CNS). Two measures of CNS function are electroencephalogram (EEG) readings which measure spontaneous electrical activity in the brain, and elicited activity via the evoked-potential (EP) or event-related potential (ERP), which measures the specific electrical response of the brain to external sensory stimuli (Devor and Cloninger, 1989; Begleiter and Porjesz, 1988; Tabakoff and Hoffman, 1988).
EEG patterns have been shown to be different in alcoholics and controls, and relatives of alcoholics have similar patterns, indicating that the differences between alcoholics and controls are not simply a consequence of alcohol abuse (Propping, Kruger and Mark, 1981). Monozygotic twins have been shown to have almost identical EEG responses to alcohol (Tabakoff and Hoffman, 1988).
Subjects at high risk for alcoholism can be differentiated from controls on the basis of their EEG alpha activity (Pollock et al., 1983). Alcoholic subjects had greater increases of slow alpha activity and greater decreases of fast alpha activity after alcohol intake than controls. The high risk subjects also showed greater decreases in mean alpha frequency after alcohol intake. The EEG findings suggested that subjects at high risk for alcoholism are physiologically more sensitive to alcohol than control subjects.
In a review by Begleiter and Porjesz (1988), children at high risk of alcoholism (those born to an alcoholic father) were found to differ from low risk children (children of non-alcoholics) in the pattern of their evoked potential (EP) waves.
Biochemical markers for alcoholism are gene products, or closely linked gene products, that predispose an individual to alcohol related problems (Tabakoff and Hoffman, 1988).
Genes that have been suggested for a genetic predisposition to alcoholism include the S allele of the C3 antigen (Hill et al., 1975), the CW3 HLA antigen (Shigeta et al., 1980), and serum prolactin (PRL) levels (Schuckit, Parker and Rossman, 1983). Corsico et al. (1988) found the B40 and DR4 HLA antigens had a much higher frequency in a population of alcoholics than in controls, while DR3 had a much lower frequency.
Monoamine oxidase (MAO) has also been suggested as a potential biochemical marker (Tabakoff and Hoffman, 1988). Monoamine oxidase is an enzyme involved in the metabolism of biogenic amines. Platelet monoamine oxidase (MAO) activity has been shown to separate Type I and Type II alcoholics (von Knorring et al., 1985). Platelet MAO activity was normal in Type I alcoholics and in healthy controls, but was lower in Type II alcoholics, suggesting a possible biological marker for alcoholism. However, in a study by Tabakoff et al. (1988), MAO activities were not significantly different in alcoholics than in control subjects. Instead they found that monoamine oxidase activity was correlated with the time since the last drink in alcoholics, but not in controls.
Research on the genetics of alcoholism is making use of molecular biological techniques. Linkage studies consider the cosegregation of a particular disorder with marker genes in families. Evidence that alcoholism is transmitted genetically may be shown by linkage between an alcoholism susceptibility gene (or genes) and a polymorphic marker. If linkage is demonstrated, then an alcohol susceptibility gene can be located to a region of a chromosome (Hill, Aston and Rabin, 1988).
Linkage analysis has been made easier with the use of restriction fragment length polymorphisms, or RFLPs. RFLPs are different sized fragments of DNA created by polymorphisms in the recognition sites of restriction endonucleases. Certain patterns of these fragments may be transmitted through families along with a disease gene, indicating the location of variant genes (Gordis et al., 1990).
There have been suggestions of a genetic linkage between alcoholism and the MNS blood group (Hill et al., 1988), and the esterase-D locus on chromosome 13q (Tanna et al., 1988). However, neither of these two linkages has been confirmed.
Association studies involve showing that the frequency of a particular allele differs in a population of affected individuals from that seen in a normal control population. While linkage refers to cosegregation of disease and a marker allele within families, association refers to population differences. Association studies can find genes with relatively small influences, whereas linkage studies find genes with moderate-to-large effects (Baron, 1993).
Association studies are less complicated than other forms of genetic analysis because DNA only needs to be obtained from affected individuals, and not from a large number of family members. Association studies are very specific. If an association is present, the disease-influencing gene will be found very quickly. However for this to occur, the candidate gene that is chosen must be very nearly exactly correct. If variation at a locus affects the phenotype under study, an association study will only be able to detect that effect if the variant allele is in strong linkage disequilibrium with the marker alleles being tested. This requires low mutation rates for both the markers and the allele variants, and a very low recombination frequency between the two sites (Gelernter et al., 1991).