Beta-Carotene and lung cancer promotion in heavy smokers--a plausible relationship?

Initial findings from two major clinical trials, the BetaCarotene and Retinol Efficacy Trial (CARET) in the United States (7) and the Alpha-Tocopherol, Beta-Carotene Cancer Prevention Study (ATBC Study) in Finland (2), unexpectedly showed that supplemental p-carotene alone or in combination with retinol appeared to increase the risk of lung cancer, particularly in current smokers. CARET also reported that the combination of retinol and p-carotene increased the risk of lung cancer in asbestos workers, most of whom were current or former smokers. In contrast, the U.S. Physicians' Health Study (3) reported that supplemental P-carotene did not increase the risk of lung or other cancers, even in smokers. Finally, a consistent body of epidemiologic data (4) shows that increased consumption of P-carotene-rich foods and higher blood levels of p-carotene are associated with a reduced risk of lung cancer. No plausible explanation has been offered for why P-carotene increased the risk of lung cancers in the ATBC Study (2) and in CARET (7) but did not increase the risk of lung cancer in the Physicians' Health Study (3) and failed to increase the risk of cancers at other organ sites in these and other trials (2J^). Also unresolved is why P-carotene should increase the risk of lung cancer in current but not former smokers (7). In this issue of the Journal, ATBC Study investigators (6) and CARET investigators (7) present additional analyses from these important trials. Subgroup analyses by Albanes et al. (6) indicate that the adverse effects of supplemental p-carotene in the ATBC Study are restricted to persons who smoked one pack of cigarettes or more per day or who drank above-average levels of alcohol. The new finding that lung cancer risk was not elevated by supplemental p-carotene in moderate intensity smokers in the ATBC Study extends the related finding from CARET that former smokers who received the supplements were not at increased risk. CARET investigators similarly examined their data for evidence of an alcohol effect; results suggested an interaction with alcohol (7), although the lack of a dose-response effect reduces confidence that this interaction is real. Other intriguing findings emerge, such as the finding in both trials that large-cell cancers were increased to a greater degree following p-carotene supplementation than other histologic types (6,7). Given the new and older findings from these and other pV carotene trials, what can be surmised about mechanisms of action, causality, and generalizability? The aggregate data suggest that concurrent cigarette smoke exposure (/,7) of relatively high intensity (6) is necessary for a promotional effect of supplemental P-carotene on lung cancer to occur. This hypothesis is consistent with a direct interaction between cigarette smoke and P-carotene. The gas phase of cigarette smoke is highly oxidative, possessing an array of free radicals (8,9). Handelman et al. (10) exposed human plasma to the gas phase of cigarette smoke to assess effects on carotenoids, tocopherols, and retinol. Exposure of plasma to room air did not degrade these micronutrients. In contrast, exposure of plasma to cigarette smoke led to the destruction of carotenoids and a-tocopherol in the plasma, despite their physical location in lipoproteins and presumed protection by endogenous antioxidants. Oxidation of other plasma lipids was minimal, suggesting selective oxidative susceptibility of carotenoids and a-tocopherol. Thus, it is reasonable to propose that P-carotene in the lungs of heavy smokers undergoes oxidative attack. While researchers have not yet identified oxidation products of P-carotene induced by tobacco smoke, products of P-carotene induced by other oxidants have been characterized (11-13). The chemical structures of many of these products (i.e., epoxides) suggest that they would be unstable under conditions of oxidative stress, further contributing to oxidation. This hypothesis is consistent with observations in vitro that P-carotene can have pro-oxidant activity, particularly under conditions of ambient (14) or higher (15,16) oxygen tensions. Conditions in the lungs of heavy smokers may tip the P-carotene antioxidant-pro-oxidant balance toward a prooxidant state. Agents that induce a cellular pro-oxidant state are known to act as tumor promoters (17). Auto-oxidation of P-carotene in vitro is dose dependent (15,16). This dose dependence may explain why lung cancer risk was not elevated among p-carotene-supplemented smokers in the Physicians' Health Study—blood levels of P-carotene in this study (1.2 ng/mL) were considerably lower than those achieved in subjects receiving the supplements in the ATBC Study and in CARET (median = 3.0 and 2.1 \ig/mL, respectively) (1-3). As detailed by Albanes et al. (6) in this issue of the Journal, participants with the highest blood P-carotene levels following supplementation in the ATBC Study were not at greater risk than participants with lower levels. These lower