Effects of beta-carotene supplementation on molecular markers of lung carcinogenesis in male smokers

Abstract

Two primary prevention trials unexpectedly showed adverse effects of supplemental beta-carotene on lung cancer incidence in cigarette smokers. To elucidate the molecular mechanisms that might underlie these effects, we studied the immunohistochemical expression of cytochrome P450 1A1, 1A2, and 2E1, retinoic acid receptor beta, activated protein-1 elements, cyclin D1, and Ki67 in lung tumors and, when available, adjacent normal tissues obtained from incident cases in the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study. Archival lung tissue was available from 52 men randomized to receive 20 mg of beta-carotene per day and 30 men randomized to the placebo arm, all of whom were diagnosed with incident non-small-cell lung carcinoma during the course of the trial and subsequently underwent radical pulmonary resection. In normal-appearing bronchial epithelium, positive staining for cyclin D1 was observed in 23% of cases in the beta-carotene group and 0% of cases in the placebo group (based on only 3 of 13 versus 0 of 11 cases staining positively, however; P = 0.04), with no differences in expression noted in lung tumor tissue (P = 0.48). There were no statistically significant differences in Ki67 expression in normal or cancerous lung tissue between intervention groups, although a small increase in staining in tumors was noted among cases in the beta-carotene versus placebo group (88% versus 71% of cases stained positive, respectively; P = 0.13). Contrary to expectation, beta-carotene supplementation had no apparent effect on retinoic acid receptor-beta expression. These findings suggest that male smokers supplemented with beta-carotene may have had an increased risk of lung cancer due to aberrant cell growth, although our results are based on a relatively small number of cases and require confirmation in other completed trials of beta-carotene supplementation.

Figure 1

Immunohistochemical localization of cyclin D1 in A) normal bronchial epithelium, placebo group, B) normal bronchial epithelium, β-carotene group, C) lung tumor cells, placebo group, and D) lung tumor cells, β-carotene group. In Panel A, a bronchial lumen filled with mucus and lined by tall columnar ciliated epithelium is shown. Cyclin D1 immunoreactivity is essentially absent, with the exception of rare scattered nuclei. Panel B shows numerous positive-staining nuclei within the basal layer and in regions of pseudo-stratification of the columnar ciliated bronchiolar-type epithelium. Tumor cells in Panel C and Panel D demonstrate frequent strong nuclear positivity for cyclin D1 along with some diffuse cytoplasmic staining. In contrast, stromal cells and chronic inflammatory cells show rare cyclin D1 staining in these latter panels. All photomicrographs are 20× magnification.

Figure 2

Immunohistochemical localization of Ki67 in A) normal bronchial epithelium, placebo group, B) normal bronchial epithelium, β-carotene group, C) lung tumor cells, placebo group, and D) lung tumor cells, β-carotene group. In Panels A and B, nuclear positivity for Ki67 is essentially absent in columnar ciliated bronchiolar-type epithelium. In Panel C, only scattered malignant cells show nuclear staining for Ki67, whereas numerous positive-staining nuclei are apparent in Panel D. Rare scattered inflammatory cells are also positive for Ki67 in Panel D. All photomicrographs are 20× magnification.