D₂ Dopamine Receptor Gene Polymorphisms among African-Americans and Mexican-Americans: A Lung Cancer Case-Control Study¹

African-Americans in the United States are at higher risk than are Caucasians of developing and dying from lung cancer (1). At the same time, although African-American smokers are more likely to try to quit, they have a lower success rate than do Caucasian smokers (2). Recent family, twin, and molecular genetic studies provide compelling evidence of a role for genetic factors governing such smoking behavior (3, 4, 5, 6, 7, 8, 9, 10, 11). For example, it has been reported that the concordance rates for smoking, not smoking, and quitting are higher for monozygotic twins than for dizygotic twins. Specifically, the concordance rate for smoking in 82 pairs of identical twins reared apart was 79% (12). In addition, a meta-analysis of data from eight studies revealed an estimated heritability rate of 60% for smoking (12).

There is substantial evidence that nicotine is the dependence-producing component of tobacco (13). Once delivered to the brain,nicotine exerts psychoactive effects that affect mood and cognitive function. Specifically, nicotine intake is associated with enhanced pleasure, improved task performance and memory, reduced anxiety and tension, and weight control. It also relieves the effects of nicotine withdrawal (14). Nicotine acts through the dopamine reward pathway for its reinforcing effect. It has been hypothesized that persons with a functional deficit in the dopamine reward pathway may be more prone to drug addiction, including nicotine dependence (15).

There are two primary families of dopamine receptors,D₁ and D₂, which are differentiated by their ligand specificity, central nervous system distribution, physiological actions, and their stimulatory(D₁) or inhibitory (D₂)effects on G-protein second-messenger systems (16). In 1989, Grandy et al. (17) reported that the human D₂ dopamine receptor gene is located on chromosome 11q. Since then, several polymorphisms on this gene have been identified. In particular, studies conducted in the early 1990s examined associations between a polymorphism in the TaqI A allele (A1 and A2) and substance abuse, including tobacco use (15, 18). More recently, investigators have also explored associations between the TaqI B allele(B1 and B2) and smoking (9). D₂ dopamine receptor polymorphisms have been associated with obesity (19), neuropsychiatric disorders (20), and substance abuse, including alcoholism (16, 21, 22, 23), cocaine use (24), polysubstance use (18), and tobacco use (9, 15, 25).

Several studies have shown that the A1 allele is associated with a reduced number of dopamine binding sites in the brain (16, 26, 27, 28). It has been hypothesized that such persons who have a reduced dopamine D₂ receptor density have a deficit in their reward system and experience an enhanced reward when exposed to dopaminergic agents, thereby making them more prone to nicotine addiction (15). Results from a previous study in our laboratory support this hypothesis, in that persons who exhibited certain D₂ polymorphisms (A1 or B1) were more likely to have ever smoked, started smoking at an earlier age, and attempted to quit fewer times compared with persons without the polymorphism (9). Other investigators have reported that the A1 and B1 alleles are more prevalent among current and former smokers than among never smokers (15).

No studies have evaluated these polymorphisms in minority populations. Because African-Americans suffer disproportionately higher incidences of lung cancer and cardiovascular disease, it is important that these issues be explored. The primary objective of this study of African-Americans and Mexican-Americans was to test the hypothesis that ever smokers in ethnic groups other than Caucasians are more likely than never smokers in these ethnic groups to exhibit the DRD2³alleles associated with tobacco use (A1 and B1). Furthermore, because of a predicted higher prevalence of smokers in a family attributable to the patterns of inheritance of the genotypes associated with tobacco use, we also anticipated that individuals with these at-risk DRD2 alleles would be more likely to report a family history of smoking-related cancers. Because other inherited genetic variants may interact with smoking on cancer risk, we also explored whether this association differs between cancer patients and control subjects.

The study population consisted of a subset derived from a case-control study of lung cancer among minority populations. The case subjects included 140 African-American or Mexican-American lung cancer patients who had newly diagnosed, histologically confirmed lung cancer. There were no age or cancer stage restrictions for inclusion. Subjects were identified through The University of Texas M. D. Anderson Cancer Center and community and Veterans Affairs hospitals in Houston and San Antonio, Texas, during 1990 to 1994.

The control group was a convenience sample accrued from community centers, churches, cancer-screening programs, and employee groups. The group consisted of 222 healthy volunteers who were frequency-matched to the cases by age (±5 years), sex, and ethnicity (2:1 for Mexican-Americans and 1:1 for African-Americans). Because we recruited Mexican-American cases from Houston, as well as from Galveston and San Antonio, we matched only very loosely on residence, i.e.,(Houston/Galveston versus San Antonio). The controls were recruited from 37 community locations, of which two were prostate cancer screening programs.

Epidemiological data were collected by personal interview. After informed consent was obtained, a structured interview lasting ∼45 min was conducted by trained bilingual interviewers. Data were collected on sociodemographic characteristics, recent and prior tobacco use, other lifestyle habits, and family history of cancer. At the completion of the interview, blood was drawn into heparinized tubes for cytogenetic and molecular genetic analyses.

The subjects were genotyped for the TaqI A and TaqI B sites of the DRD2 gene using methods described previously (9). Genomic DNA was extracted from the blood sample and used as a template for the PCR. Primers 5′-CCG TCG ACC CTT CCT GAG TGT CAT CA-3′ and 5′-CCG TCG ACG GCT GGC CAA GTT GTC TA-3′ were used to amplify a 310-bp fragment spanning the polymorphic TaqI A site of the DRD2 gene. Primers 5′-GAT ACC CAC TTC AGG AAG TC-3′ and 5′-GAT GTG TAG GAA TTA GCC AGG-3′ were used to amplify a 459-bp fragment spanning the polymorphic TaqI B site of the DRD2 gene. PCR was performed in 30-μl reaction mixtures containing 1.5 mmMgCl₂, 0.2 mmdeoxynucleotide triphosphates, 0.5 μm primers,1 μg of template DNA, and 1.5 units of Taq polymerase with a PCR buffer consisting of 20 mm Tris-HCl (pH 8.4) and 50 mm KCl (Boehringer Mannheim, Indianapolis,IN). After an initial denaturation at 94°C for 4 min, the DNA was amplified during 35 cycles of 30 s at 94°C, 30 s at 58°C,and 30 s at 72°C, followed by a final extension step of 5 min at 72°C. A portion of the PCR product (20 μl) was digested with 5 units of TaqI for 22 h at 65°C to reveal the DRD2 TaqI A polymorphism and digested with 5 units of TaqI for 5 h at 65°C to reveal the DRD2 TaqI B polymorphism. Twenty μl of the PCR digestion products were then resolved on a 3% agarose gel (5 V/cm) containing ethidium bromide. There were three DRD2 TaqI A genotypes:(a) the predominant homozygote, A2A2, indicated by two fragments, 180 and 130 bp; (b) the heterozygote, A1A2, revealed by three fragments, 310, 180, and 130 bp; and(c) the rare homozygote, A1A1, shown by the uncleaved 310-bp fragment. There were also three DRD2TaqI B genotypes: (a) the predominant homozygote, B2B2, indicated by two fragments, 267 and 192 bp; (b) the heterozygote, B1B2, revealed by three fragments, 459, 267, and 192 bp; and (c) the rare homozygote, B1B1, shown by the uncleaved 459-bp fragment.

The DRD2 genotype data were merged with the interview data. Ever smokers were defined as individuals who had smoked >100 cigarettes in their lifetime. Former smokers were defined as those who had quit at least 1 year before the interview. Latency was defined as the length of time that had elapsed since the cessation of smoking. Pack-years were calculated using the average number of cigarettes smoked/day and the number of years smoked. Subjects who reported one or more cases of lung, oral cavity, pharyngeal, esophageal, pleural,cervical, urinary bladder, kidney, or head and neck cancers among first-degree relatives were coded as having a positive family history of smoking-related cancers.

Bivariate relationships were assessed using χ²tests of independence. Fisher’s exact tests were used to test independence when cell sizes were small. To adjust for relevant covariates, the relationship between genotype and a family history of cancer among first-degree relatives was computed using standard logistic regression procedures.

Our study population was composed of 43 Mexican-American cases, 97 African-American cases, 111 Mexican-American controls, and 111 African-American controls. Study population characteristics are presented in Table 1 by ethnicity and case-control status. The two groups did not differ significantly by sex or age. Among case subjects, 88% of Mexican-Americans and 97% of African-Americans were ever smokers; of control subjects, 56% of Mexican-Americans and 65% of African-Americans were smokers (both Ps, <0.001). The mean pack-years, number of cigarettes smoked/day, and years smoked were significantly higher in cases than in controls for both ethnic groups(all Ps, <0.001).

Table 2 summarizes the distribution of the DRD2 genotypes by ethnicity and case-control status. The DRD2 TaqI B genotype distribution was strongly associated with ethnicity in both cases and controls. Specifically, the common genotype(B2B2) was significantly less prevalent among Mexican-American control subjects than among African-American control subjects (42.3% compared with 66.1%). The B1 allele frequencies were 0.36 for Mexican-American control subjects and 0.19 for African-American control subjects; a similar pattern was evident in the case subjects. The TaqI A genotypes did not,however, differ significantly by ethnicity. There also were no statistically significant differences in the distribution of the TaqI A or TaqI BDRD2 genotypes between cases and controls of either ethnic group (all Ps, >0.4; data not shown). Among controls, no deviations from the Hardy-Weinberg equilibium were detected for either the TaqI A or TaqI Bgenotypes.

We observed significant correlations between the DRD2genotype and various smoking-related variables among control subjects(Table 3). Specifically, among Mexican-American controls with the A1A1 genotype, 20.0% were never smokers and 60.0% were current smokers, compared with 53.1% and 28.1%, respectively, for those with the A2A2 genotype. A similar trend was observed for Mexican-American control subjects with the B1B1 genotype(21.4%, never smokers; 64.3%, current smokers). In contrast, among Mexican-American control subjects with the B2B2 genotype,53.7% were never smokers and 19.5% were current smokers. In Mexican-Americans, the A1 frequencies for never smokers, former smokers, and current smokers were 0.35, 0.44, and 0.50. Aχ ² test for linear trend found increasing A1 frequency for these three groups, respectively (P =0.06). The B1 frequencies for never smokers, former smokers, and current smokers were 0.28, 0.30, and 0.52. Theχ ² test for linear trend showed a strong trend(P = 0.005) in increasing B1 frequency for these three groups. These patterns were not observed among African-American subjects (data not shown).

As Table 3 also shows, persons with the A1A1 genotype were heavier smokers (30.8 pack-years and 21.9 cigarettes per day) than were subjects with the A2A2 genotype (18.6 pack-years and 12.6 cigarettes/day). Because the pattern for amount of smoking was similar for Mexican-American and African-American controls, we present the data for the two ethnic groups combined. There was also evidence of a gene-dose effect, i.e., as the number of A1allele copies increased, the number of pack-years and cigarettes/day reported increased correspondingly. In addition, individuals with the A1A1 genotype had started smoking at earlier mean ages(15.6 ± 5.0 years) than subjects with the A2A2genotype (18.8 ± 6.9 years). Subjects with the B1genotypes also showed a greater smoking intensity than those with the B2 genotype. There were no statistically significant associations between genotype and smoking-related phenotypes among case subjects, although we did observe a modest association(P < 0.10) between the A1 genotype and the mean number of cigarettes smoked/day (data not shown).

Table 4 shows the associations between self-reported smoking-related cancers among first-degree relatives and the DRD2 genotypes as a function of case-control status for the two ethnic groups combined. For Mexican-Americans, there were an average of 9.0 first-degree relatives/case and 8.0 first-degree relatives/control. For African-Americans, there were an average of 8.7 first-degree relatives/case and 9.5 first-degree relatives/control. After adjusting for age, sex, ethnicity, smoking status, and the number of first-degree relatives, case subjects with an A1 allele(A1A1 or A1A2 genotypes) proved to be 3.6 times(CI, 1.0–13.2) more likely to report a family history of smoking-related cancers among first-degree relatives than were case subjects without an A1 allele (A2A2 genotype). Specifically, of the case subjects with an A1 allele, 84.2%had a family history of smoking-related cancer in a first-degree relative compared with 15.8% of the individuals with the A2A2 genotype. In addition, case subjects with a B1 allele (B1B1 or B1B2) were 1.8 times (CI, 0.5–5.7) as likely to have a family history of smoking-related cancer than were case subjects without a B1allele (B2B2), although the difference in the frequency was not statistically significant. Among control subjects, there was no such relationship between A1 or B1 genotypes and a family history of smoking-related cancers. We also noted a nearly significant interaction between DRD2 A1 genotypes and case/control status for increased risk of smoking-related cancer among first-degree relatives. Case individuals with the DRD2 A1A1 or A1A2 genotypes were 3.7 (CI, 0.8–16.4) times more likely to have an affected first-degree relative than other groups. In addition, case individuals with the DRD2 B1B1 or B1B2 genotypes were 2.3 (CI, 0.6–8.7) times more likely to have an affected first-degree relative than other groups (data not shown).

There is a growing body of evidence implicating genetic factors as correlates of smoking behavior and of lung cancer risk, and the investigation reported here provides support for the hypothesized link between the dopamine receptor genes and smoking status in a minority population. In particular, we observed a difference in the frequency of the B1 risk allele by ethnicity for both case and control subjects, i.e., it was significantly less common among African-American controls than among Mexican-American controls; a relationship between the smoking status and genotype status for Mexican-American control sub- jects; relationships between the smoking intensity (pack-years and cigarettes smoked/day) and genotype status among controls (ethnic groups combined); and evidence that lung cancer patients of both ethnic groups with the A1 allele exhibited familial aggregation of smoking-related cancers.

To date, studies of the DRD2 genotypes and smoking primarily have focused on Caucasians. The study reported here is perhaps the first to provide insight into ethnic group-related differences in the DRD2 genotypes. Among Mexican-Americans, we observed a significant relationship between smoking status and the TaqI B genotypes; a similar, but nonsignificant, trend was apparent for the TaqI A genotypes. These findings confirm our previous observations in a study in a Caucasian population (9). In both instances, never smokers were less likely to have the risk alleles (A1 or B1) than were current smokers. Among African-Americans, however, we observed no relationship between smoking status and the TaqI A or TaqI B genotypes. The discordant findings by smoking status in the two ethnic groups could be attributed to the relatively greater importance of contributing factors such as psychosocial, environmental, or socioeconomic determining the complex behavior of smoking. In addition, there may be ethnic differences in nicotine pharmacokinetics and metabolism. It is plausible that such factors may be more important in determining the ever-smoking status among African-Americans than are the DRD2 genotypes. Furthermore, the B1 allele seems more predictive than the A1 allele in smoking status. The frequency of B1allele for African-Americans was less than half that for Mexican-Americans.

We also observed that among Mexican-American controls, the A1 allele frequency increased from 0.35 in never smokers to 0.44 for former smokers, and 0.50 for current smokers with aχ ² test for linear trend P of 0.06. The B1 allele frequencies were 0.28 for never smokers, 0.30 for former smokers, and 0.52 for current smokers with aχ ² test for linear trend P of 0.005. In African-American controls, the A1 allele frequency was 0.41 for never smokers, 0.36 for former smokers, and 0.34 for current smokers, and the B1 frequency was 0.24 for never smokers,0.16 for former smokers, and 0.17 for current smokers. In a previously published study (9), we identified A1 allele frequencies of 0.16 for never smokers, 0.22 for former smokers, and 0.23 for current smokers among Caucasian controls. The B1frequencies in these Caucasian controls were 0.03, 0.15, and 0.17,respectively. The B1 allelic frequencies that we observed in the current study were two times higher among Mexican-Americans (0.38 for cases and 0.36 for controls) than among African-Americans (0.16 for cases and 0.19 for controls). This point suggests a need to determine how, and the extent to which, these allelic differences translate into ethnic differences in the effects of nicotine, both behaviorally and biologically.

It was a somewhat unexpected finding that the relationships between DRD2 and several of the smoking phenotypes were observed among controls but not among lung cancer cases. For example, smoking status was strongly related to TaqI B1 allele frequencies among Mexican-American controls but not among Mexican-American cases. It is possible that factors other than DRD2 are stronger contributors to tobacco use as well as development of lung cancer. For example, the CYP2A6 gene,which has been shown to be related to nicotine metabolism and lung cancer susceptibility (29), may be a more influential factor for tobacco use among lung cancer patients.

There is a pressing demand for studies that elucidate the functional effects of genetic polymorphisms. For example, although there is some evidence that the presence of an A1 allele is associated with a decreased number of D₂ dopamine receptors in the brain (16), the mechanism by which the DRD2 A1 or B1 alleles increase susceptibility to tobacco use are not known. One possible mechanism is that this decreased density of receptors results in a deficiently functioning dopamine reward system, such that persons with an A1 allele experience an enhanced reward when exposed to dopaminergic agents, rendering them more vulnerable to nicotine dependence. However, until researchers can delineate the mechanisms by which genetic variants alter the biological and psychological effects of nicotine, we cannot fully understand the observed relationships between genes and smoking behavior.

This study is the first to examine the relationship between the DRD2 genotypes and a family cancer history. Specifically, we tested the hypothesis that the A1 and B1 alleles are risk factors for smoking and that if these polymorphisms segregate in families, the rates of smoking-related cancers among the first-degree relatives of probands with these genotypes would be elevated. However, our results supported this hypothesis only for cases with the A1 allele. In particular, the odds of a person with an A1 allele (A1A1 or A1A2 genotypes)having a positive family history was more than three times greater than that for persons with the A2A2 genotype. Although the risk estimate is large, its confidence intervals are wide. This hypothesis therefore needs to be examined further in larger studies. We also observed a similar but smaller trend for the B1 risk allele,but the difference seen for the B1 and B2 alleles was not statistically significant, suggesting that the A1allele is a more powerful risk factor for this phenotype than the B1 allele is. Regardless, a relationship between the DRD2 alleles and a family history of tobacco-related cancers was found only among case subjects, suggesting that a genetic susceptibility to tobacco exposure was also associated (or aggregated)with a predisposition to smoking. Among control subjects, on the other hand, persons with the dopamine receptor genes appear to be at an elevated risk for smoking, but this elevated smoking risk does not appear to translate into a familial risk of tobacco-related cancers. We also noted a nearly significant interaction between DRD2 A2genotypes and case/control status for increased risk of smoking-related cancer among first-degree relatives. A large study is needed to confirm this finding.

The present study was limited in that it examined a narrow range of phenotypes. If we are to gain a better understanding of a genetic predisposition to tobacco use and/or nicotine dependence, investigators must examine a more comprehensive range of phenotypes, such as measures of dependence, the rate of convergence on regular smoking, biological measures of nicotine pharmacokinetics, and receptor activity. Further drawbacks in our study were that the smoking-related phenotypes were assessed retrospectively in study participants, and all data were self-reported without subsequent verification. In addition, it is possible that the validity of the self-reported data may vary between cases and controls. The potential for recall bias regarding first-degree relatives with a tobacco-related cancer should be considered when interpreting the results of our study.

There is little debate that the dopamine reward system is involved in the development of nicotine addiction and patterns of tobacco use,although other neurotransmitters and factors such as nicotine metabolism also appear to play a role. The extent to which genetic makeup modulates susceptibility to tobacco use and the development of nicotine dependence is not entirely clear. Although there does not appear to be a single gene that can serve as a predictor of smoking behavior, it is possible that if the candidate genes are examined in combination, in the form of a genetic risk profile, researchers may be able to gain a clearer picture of the extent to which genetics plays a role in tobacco use.

Even if genetics does play an important role in tobacco use and nicotine addiction, it is not likely that knowledge of genetic risk alone could lead to the development of effective tobacco prevention strategies. Smoking is a complex behavior that also results from a host of environmental and psychological factors. Future research efforts should therefore focus on an integrative approach that incorporates psychological, social, cultural, behavioral, pharmacological, and genetic influences, as well as the interactions among these factors (30).