Associations between Catalase Phenotype and Genotype: Modification by Epidemiologic Factors

Oxidative stress may play a significant role in the risk of chronic diseases (1-4). Although reactive oxygen species can cause oxidative damage to cellular macromolecules, such as DNA and lipids, multiple antioxidant defenses can neutralize reactive oxygen species (5-8). Variability in endogenous and exogenous pro-oxidants and antioxidants will determine levels of oxidative stress in the organism. Enzymes responsible for neutralizing reactive oxygen species endogenously include catalase, superoxide dismutase, and glutathione peroxidase as well as the glutathione S-transferases (1, 9). Fruits and vegetables are rich in vitamin C, vitamin E, selenium, and phytochemicals, such as flavonoids, which are potent exogenous antioxidants. It is not clear how these endogenous and exogenous modulators of oxidative stress interact in vivo; it is possible that endogenous and exogenous factors may act synergistically or that high levels of dietary antioxidants derived from consumption of fruit and vegetables may down-regulate endogenous antioxidant enzymes through feedback mechanisms.

Catalase is a heme enzyme that has a major role in controlling H₂O₂ concentrations in human cells, by converting H₂O₂ into H₂O and O₂. Working together with other antioxidant enzymes, catalase plays an integral role in the primary defense against oxidative stress. Catalase enzymatic activity was induced by exposure to H₂O₂ in hamster tracheal epithelial cells (10) and in human retinal pigment epithelial cells (11), indicating a key role for catalase in antioxidant defense and its inducibility in response to oxidative stress. A common polymorphism in the promoter region of the CAT gene consists of a C → T substitution at position −262 in the 5′ region (12), which is thought to result in reduced activity. Reduction in catalase activity may play a role in host response to oxidative stress and, indeed, variant CAT T alleles have been associated with increased risk of hypertension (13, 14) and vitiligo (15), both conditions that are related to oxidative stress. We previously showed that women with the common CAT CC genotype had a 17% reduction in risk of breast cancer compared with those with at least one T allele (CT and TT genotypes) in the Long Island Breast Cancer Study Project, a large population based case-control study (1,008 cases and 1,056 controls; ref. 16). We also observed a statistically significant multiplicative interaction between genotype and fruit consumption (P = 0.02) among women who did not use antioxidant supplements, with lowest risk observed for women who were high fruit consumers and had CC genotypes (odds ratio, 0.59; 95% confidence interval, 0.38-0.89). Although our preliminary study with specimens from healthy volunteers (n = 18) indicated that the polymorphism was associated with reduced catalase activity (16), the sample size was small, and the results need to be replicated.

Thus, we sought to confirm our findings in a much larger sample size and to evaluate if relationships between CAT genotype and phenotype were modified by demographic characteristics, fruit and vegetable consumption, and other lifestyle factors.

For these analyses, we used samples and data from healthy men and women who participated in case-control studies of breast and prostate cancer conducted in Arkansas. A detailed description of the study design, participant recruitment, and methodology has been reported previously (17). Control participants were identified from the Arkansas state driver's license/identity card records for those under 65 years of age and the Health Care Finance Administration lists for those 65 years and older and were matched to cases on race, age (within 10 years), sex, and county of residence. Informed consent from each participant was obtained according to an institutional review board–approved protocol. We calculated sample size to be able to detect a difference in catalase activity of 20% SD, with 80% power and 0.05 α level, and thus included 420 control participants (50% male and 50% female).

Erythrocyte Catalase Activity Assay. Stored erythrocytes from blood samples collected from study participants were assayed for catalase activity (18) using a commercial kit from OxisResearch (Oxis Research, Portland, OR), with modifications to accommodate a microtiter plate format. The erythrocytes were washed thrice with ice-cold 0.9% NaCl, centrifuging between each wash. The cell pellet was then resuspended in 4 volumes of ice-cold water, incubated on ice for 10 minutes to lyse the cells, and centrifuged at 3,000 × g for 10 minutes at 4°C. For the assay, the lysate was diluted 1:400 with sample diluent before use. Triplicates of 30 μL of diluted erythrocyte lysate were delivered to a 96-well microplate. The samples were incubated with 10 mmol/L H₂O₂ at room temperature for exactly 1 minute. The reaction was stopped with the provided stop reagent, and an aliquot was removed and added to a microplate containing horseradish peroxidase/Chromagen reagent and allowed to develop for 10 minutes. Absorbance was read at 520 nm on a SpectraMax UV-VIS microplate reader (Molecular Devices, Sunnyvale, CA), and activity units were assigned based on a calibration curve of known amounts of H₂O₂. Multiple aliquots of a control erythrocyte sample were made and processed identically to the study samples to assess variability of the assay. The reproducibility of the assay was high, with a coefficient of variation of 8.7%. Activity was normalized to hemoglobin content in the sample.

Hemoglobin Assay. Stored erythrocytes were assayed for hemoglobin content. The erythrocyte pellet was resuspended in 50 volumes of ice-cold water, incubated on ice for 10 minutes to lyse the cells, and centrifuged at 3,000 × g for 10 minutes at 4°C. For the assay, triplicates of 20 μL of erythrocyte lysate were delivered to a 96-well microplate. The samples were incubated with Drabkin's reagent at room temperature for exactly 15 minutes. Absorbance was read at 540 nm on a SpectraMax UV-VIS microplate reader, and hemoglobin concentration was assigned based on a calibration curve of known amounts of Randox Cyanmethemoglobin Standard Set. Multiple aliquots of a control hemoglobin sample were made and processed identically to the study samples to assess variability of the assay.

CAT Genotype Determination. Genomic DNA was extracted from lymphocytes using a commercial kit (Qiagen, Inc., Valencia, CA). The CAT −262C → T polymorphism (rs#1001179) was determined by direct sequencing on the Beckman CEQ genetic analysis system as described previously (16), using primers 5′- ACGTTGGATGTCTGGCCCAGCAATTGGAGAG-3′ and 5′-ACGTTGGATGAGGATGCTGATAACCGGGAG-3′. All genotyping results were reviewed manually for quality control. Controls for genotype and two nontemplate controls were included on each plate. Laboratory personnel1 were blinded to catalase activity level.

Exposure Assessment. Data on demographic characteristics, diet, and other lifestyle factors were obtained in an in-person interview, using the Block food frequency questionnaire to assess dietary intake 12 months before the interview. The instrument queried intake of seven fruits and fruit juices and 11 vegetables (excluding French fries). Use of multiple vitamins (i.e., one-a-day, therapeutic, and Stress tab type) and single supplements of vitamin A, β-carotene, vitamin E, calcium or Dolomite, vitamin C, and iron were also included in the food frequency questionnaire. For food items and supplements, data were collected on frequency of consumption and use; and for food items, responses for portion sizes were relative to a stated average serving size with options of small, medium, or large. Frequency and portion size data were translated to servings per weeks.

All analyses were conducted using SAS version 8.2 (SAS Institute, Inc., Cary, NC). We analyzed the association between CAT genotype and phenotype using linear regression models as implemented by SAS PROC GLM. Catalase activities were natural logarithm transformed to approximate normal distributions. Least-square geometric mean activity and 95% confidence intervals within each genotype were calculated, with adjustment for age at the time of interview (continuous), sex, race (Caucasian and African American), smoking status, fruit and vegetable consumption, and vitamin supplement use. Percentage differences and Ps were calculated for the mean difference associated with the variant genotype (CT or TT genotype) compared with the common genotype (CC genotype).

To evaluate whether associations between CAT genotype and activity were modified by demographic, diet, and lifestyle factors, joint categories of CAT genotype and those factors were used in the multivariate model. Percentage differences and Ps were calculated for the mean difference associated with the variant genotype (CT or TT genotype) compared with the common genotype (CC genotype), within each demographic, diet, or lifestyle strata. For test for trend, we used the medians of the categories for fruit and vegetable intake modeled as continuous variables. Multiplicative interactions were tested by adding cross-product terms in the model. All statistical tests were two sided.

Population characteristics and their relationships with catalase activity are shown in Table 1. There were no significant differences in the geometric mean of catalase activity by age, race, sex, current smoking status, fruit and vegetable consumption, and vitamin supplement use. However, geometric means for catalase activity were somewhat lower among females than males, current smokers than never or former smokers, and higher consumers of fruits and vegetables than lower consumers. Dietary data and racial information were only available on a portion of the population [n = 392 (93%) and n = 381 (91%), respectively]; however, there were no differences in the characteristics of the study population between the entire sample with genotype and phenotype data and those who with missing data (data not shown).

Associations between CAT genotypes and catalase activity are shown in Table 2. Frequencies of the CAT T allele were 17% and 5% for Caucasians and African Americans, respectively, and genotype distributions followed Hardy-Weinberg equilibrium (P = 0.10 for Caucasian and P = 0.13 for African American). Allele frequencies were similar to those reported in national databases (19).

There was a dose-response reduction in catalase activity by genotypes, with geometric means of 115.4 units/mg hemoglobin for those with CC genotypes, 82.1 units/mg hemoglobin for CT genotypes, and 73.5 units/mg hemoglobin for TT genotypes [% activity difference = −28.8% (P = 0.002) for CT genotype versus CC genotype; % activity difference = −36.3% (P = 0.02) for TT genotype versus CC genotype]. Associations between genotypes and activity were most apparent among Caucasians, with a 33.7% difference in activity between those with CC genotypes and those with CT and TT genotypes combined. Among African Americans, there was little difference in activity by genotype (−2.6%), although there were only 10 individuals with variant alleles.. There also seemed to be differential relationships by sex. Comparing those with any T alleles to those with CC genotypes, there seemed to be somewhat greater reductions in activity among women than men (% activity difference = −35.7% for women and −24.6% for men). Relationships between CAT genotype and phenotype did not vary by smoking status. There were also no activity differences by CAT genotype among age groups (data not shown).

Catalase activities were consistently higher in individuals with CAT CC genotype compared with those with at least one variant T allele, regardless of dietary fruit and vegetable consumption (Table 3; Fig. 1). However, genotype/phenotype relationships were modified by fruit and vegetable consumption. Activity difference by CAT genotype was more pronounced among those in the highest tertile of fruit and vegetable consumption [% activity difference = −35.0% (P = 0.003) for the highest tertile; % activity difference = −21.8% (P = 0.16) for the lowest tertile]. Among those with at least one variant T allele, there was a notable reduction in activity levels by each increasing tertile of fruit and vegetable consumption. Those with T alleles who were in the lowest tertile of consumption had a geometric mean of 93.0 units/mg hemoglobin; levels were lower for those in the intermediate group (78.0 units/mg) and even lower in the high consumption tertile (74.2 units/mg). However, this trend was not statistically significant (P = 0.09). Among individuals with CC genotype, catalase activity remained unchanged by tertiles of fruit and vegetable consumption. Overall, the lowest catalase activity levels were among those with CT or TT genotypes and the highest consumption of fruits and vegetables.

Because use of antioxidant vitamin supplements is another exogenous source of antioxidants, we also evaluated the effect of supplement use on associations between genotype and enzyme activity. The relationship between CAT genotype and activity did not differ between supplement users and nonsupplement users (data not shown). However, associations among CAT genotypes, catalase activity, and fruit and vegetable consumption seemed to be strongest among those who did not use vitamin supplements. Among non-vitamin supplement users, we observed a marginally significant multiplicative interaction (P = 0.07), with greatest activity difference by CAT genotype observed for those who were in the higher two tertiles of fruit and vegetable consumption (tertile 1: −6.3% activity difference; mean = 131.8; 95% confidence interval, 104.3-166.7 and mean = 123.5; 95% confidence interval, 101.4-150.3 for CC genotype and for CT + TT genotype, respectively; tertiles 2 and 3 combined: −54.6% activity difference; mean = 122.7; 95% confidence interval, 92.1-123.7 and mean = 55.7; 95% confidence interval, 63.1-98.4 for CC genotype and for CT + TT genotype, respectively). Among supplement users, however, there was no significant multiplicative interaction between fruit and vegetable consumption and CAT genotype (P = 0.73).

In this large cross-sectional study, we found that CAT genotype was associated with catalase activity, with individuals with variant CT and TT genotypes having significantly lower activity than those with common CC genotypes. Associations between genotype and phenotype were only apparent among Caucasians, with no differences in activity noted by genotype among African Americans, and relationships seemed somewhat stronger for women than for men. Although smoking did not seem to affect associations, there was modification of relationships by consumption of fruits and vegetables, with catalase activity decreasing with increasing intake, only among those with CT or TT genotypes. These findings support our a priori hypothesis that CT and TT genotypes are associated with lower catalase activity and provide a biological basis for epidemiologic studies showing associations between CAT genotype and risk of diseases related to oxidative stress (13, 14, 16, 20).

In a small pilot study (n = 18), we had previously reported that women with variant CT and TT genotypes had significantly lower catalase activity compared with those with CC genotypes (P = 0.007; ref. 16). We had also found that women with common CC genotypes had a 17% reduction in breast cancer risk in the Long Island Breast Cancer Study Project data compared with those with at least one T allele, based on the hypothesis that oxidative stress plays a role in breast carcinogenesis and that higher catalase activity reduces oxidative stress (16). Ahsan et al. also showed that arsenic exposure resulted in hyperkeratosis to a much greater extent (4-fold) among those with CAT TT genotypes than among those with CC genotypes (20). The TT genotype was also associated with higher blood pressure compared with those with CC genotype, and those with CT genotype had intermediate risk (13). Thus, our findings of genotype/phenotype relationships support the results from epidemiologic studies, providing a functional basis for the investigation of the −262C → T polymorphism in relation to risk of conditions associated with oxidative stress.

Although catalase activity normalized to hemoglobin content (units/mg hemoglobin) in erythrocytes was significantly influenced by genotype, we were unable to determine if this relationship extrapolates to catalase activity in other tissues. We were also unable to determine if this genetic polymorphism was related to altered expression levels of the protein. However, a transcription factor binding site search (21) indicates that the −262 C allele is located in close proximity to several binding sites for transcription factors, such as Sp1, and could potentially influence rates of transcription. Forsberg et al. (12) previously showed that the T allele was associated with greater catalase protein levels in blood than the C allele (n = 29, males). These results are in contrast to our findings for erythrocyte catalase activity, and it could be due to stability of the enzyme according to sample preparation methods or to the absence of females in the previous study. Nonetheless, the evidence that associations between T alleles and reduced activity were shown in two different study populations, and that studies have consistently noted associations between T variant allele and increased risks of oxidative stress related conditions, strengthen our confidence in these findings.

The activity differences by CAT genotype were more pronounced among Caucasians, whereas relationships were not observed among African Americans. Zhou et al. (14) reported that having at least one T allele was associated with an increase in risk of hypertension only among Caucasians, not among African Americans, which is consistent with these findings regarding racial difference by catalase activity. Our previous results of reduced breast cancer risk among women with CAT CC genotype was based on a population that consisted of 90% Caucasian women (16), which are also consistent with the associations by race. Nonetheless, the lack of association between genotypes and catalase activity could be due to chance due to the relatively small number of African Americans in this study. If additional studies replicate these findings, further research will be required to investigate the biological mechanisms responsible for these racial differences.

There were notable decreases in catalase activity with increasing fruit and vegetable consumption particularly for women with variant T allele. Because H₂O₂ concentrations positively regulate CAT expression (11), and because high consumption of fruits and vegetables, rich sources of exogenous antioxidants, may reduce H₂O₂ level in vivo, it is possible that high fruit and vegetable consumers have low CAT expression and thus have low catalase activity. Although it is not clear why the inverse trend was particularly noted among those with variant T alleles, it is possible that the effects of fruits and vegetables may be more important for those with low catalase activity (i.e., at least one T allele), and that individuals with high catalase activity (i.e., CC genotype) are less susceptible to dietary effects. It is also possible that if antioxidant status is high from dietary exposures, feedback inhibition could decrease the expression levels of catalase. It is also consistent with the view that higher antioxidant intake did reduce levels of oxidative stress biomarkers among susceptible populations (i.e., overweight smokers) but did not reduce markers among healthy populations (22). Nonetheless, despite lowest catalase activity among higher consumers with T alleles, it is interesting to note women with T alleles in our prior study did not have a higher risk of breast cancer than lower consumers with common CC genotypes (16).

Due to the cross-sectional nature of this study, we are only able to capture catalase activity at one point in time, which is unlikely to be representative of lifetime activity, although there is no evidence that catalase activity varies by age in our data. In addition, because we did not measure catalase expression, we are not able to directly compare Forsberg's results (12) with our findings. Although it would be valuable to assess differences in oxidative damage in participants by CAT genotype, we are unable to do that in this population due to the limitations of sample acquisition and processing and specimen availability. However, findings from this study provide a basis for future research to determine biological mechanisms by which CAT genotype may reduce total body oxidative stress and thus influence cancer etiology and other oxidative stress–related conditions.

In conclusion, this study provides an example of a gene-environment (diet) interaction in which catalase phenotype is notably affected by the individual genetic background and dietary practices. Moreover, because catalase activity is closely linked to endogenous antioxidant activity, these data provide a plausible biological explanation for our previous findings of the association between the CAT −262CC genotypes and the reduced risk for breast cancer and other disease risk. Importantly, the finding that associations between genotype and activity are modified by demographic and lifestyle factors (race, gender, and dietary habits) has implications for molecular epidemiology studies. When polymorphisms are evaluated in relation to disease risk, there is little accounting for the biochemical processes that underlie these associations. The demonstration that phenotype related to genotype is modified, in particular, by dietary habits, could lend support to the notion that some of the inconsistencies in findings from molecular epidemiologic studies could be due to differences in the populations studied and underlying characteristics that are not accounted for that mediate the relationship between genetic polymorphisms and the actual in vivo phenotypes. To advance the field of molecular epidemiology, studies such as these may be useful for elucidating the complex relationships among genetic polymorphisms, phenotypic effects, and cancer risk.

We thank Drs. Neli Ulrich and Rachel Sparks (Fred Hutchinson Cancer Center) and Mr. Steve Jacob (Roswell Park Cancer Institute) for their insightful input.