PAUL KOTIN and HANS L. FALK Carcinogenesis Studies Branch, National Cancer Institute, Public Health Service, Bethesda, Md.

The worldwide increase in the incidence of lung cancer has provided the epidemiologist and the experimentalist in the laboratory with a unique opportunity to jointly study factors concerned with the pathogenesis of this disease. The real and, at least until recently, progressive increase of lung cancer is etiologically associated with the contamination of our respiratory environment by carcinogenic agents and other environmental substances which serve to facilitate the action of these agents. The specific epidemiological pattern of lung cancer incidence is probably determined by a combination of these exogenous environmental agents as well as endogenous host factors. The major emphasis of this presentation will be directed toward exogenous factors, including polluted urban atmosphere, cigarette smoke, and viral infections. It must be emphasized that evidence for an exclusive environmental source or factor acting in the pathogenesis of lung cancer is lacking. Therefore, a discussion of air pollution and lung cancer independent of its relation to other environmental factors is unrealistic. However, sufficient epidemiological, clinical-pathological, and experimental data exist to incriminate polluted urban atmosphere as an environmental source pathogenetically related to the development of lung cancer. Since the chemistry of polluted urban air and that of cigarette smoke are remarkably similar, it is possible that much of what will be presented as being pertinent to the former will also be applicable to the latter.

Prior to discussing polluted air in relation to the other environmental factors concerned with the pathogenesis of lung cancer, it will be helpful to review the data which warrant its incrimination:

(1) Carcinogenic agents have been identified and quantitated in the polluted air of es

sentially all cities in which they have been sought.

(2) Chemical compounds with known tumor-promoting properties similarly have been identified and quantitated in polluted urban air.

(3) The stability and survival of carcinogenic hydrocarbons in the atmosphere are compatible with inhalation and a postulated biological effect in those exposed.

(4) Carcinogenic agents as well as noncarcinogenic respiratory epithelial irritants occur in the atmosphere in a physical state compatible with host entry and tracheobronchial deposition in exposed populations.

(5) Alteration in function and structure of the respiratory epithelium of representative mammalian species has been demonstrated following exposure to a broad spectrum of these environmental irritants. The resulting changes appear to facilitate the biological action of carcinogenic agents.

(6) Bioassay by skin painting and subcutaneous injection techniques has established the carcinogenic properties of compounds identified in and extracted from polluted air. Exposure of both susceptible and resistant strains of mice to aerosols of synthetically reproduced polluted urban air has resulted in the production of lung tumors in both strains. A synthesis of the findings described suggests that the carcinogenic properties of polluted urban atmosphere provide at least two indispensable links in the pathogenesis of lung cancer. The first and most obvious is the environmental presence and the host entry of agents proved experimentally to be carcinogenic and epidemiologically associated with

increased liability to the development of lung cancer. The second factor relates to the occurrence in the atmosphere of host-modifying factors which, by virtue of their effect on the ciliated mucussecreting epithelium of the tracheobronchial tree, facilitate the deposition and abnormal retention of particulate matter in the lungs. Elution of the carcinogenic polycyclic aromatic hydrocarbons (PAH), 3,4-benzpyrene, and 3,4-benzfluoranthene, by host proteins from soot particles is thereby facilitated. A significant increase in the local concentration of liberated carcinogens results. Atmospheric irritants may, in addition, periodically and intermittently cause denudation of the superficial epithelium so that the basal cell layer is directly apposed to the carcinogenic stimulus. We regard this periodic epithelial desquamation followed by regeneration in the presence of a carcinogenic stimulus as providing a favorable environment for subsequent abnormal growth.

The experimental evidence cited is especially meaningful by virtue of its quantitative compatibility with the pattern of diseases seen in human populations at risk. There exist, however, numerous apparent inconsistencies when quantitative extrapolation of the experimental data is attempted. These include limited dose response relationships, relatively low attack rate in those exposed, and deceleration in the rate of increase in incidence at a time when latent period cohort increases of known and suspected carcinogens were accelerating at a maximum. We have instituted several studies in an attempt to explain or elucidate the basis for these inconsistencies. A few of these will now be described in detail and several others mentioned.

Anticarcinogenesis. It has long been recognized that chemically related compounds may act upon one another in an additive, synergistic, or inhibitory manner. It has been observed in various fields of biochemistry that a biologically active compoundfor example, a vitamin-may be prevented from displaying its effect in the presence of a closely related derivative (anti-vitamin). The latter probably plays the part of a competitor for intracellular receptors. Along the line of earlier investigators, we have been studying the effect of closely related PAH upon one another when administered simultaneously or at varying time intervals. Experimentally, it has been possible to demonstrate that even the most potent carcinogens can have their effect reduced or totally obliterated in the presence of proper proportions of inhibitors. This has immedi

ate practical significance in that in all environmental sources of carcinogenic PAH, industrial effluents and automobile and diesel exhaust as well as cigarette smoke, a broad spectrum of noncarcinogenic and weakly carcinogenic hydrocarbons are formed. Experimentally these latter compounds unequivocally exert an inhibiting effect, and, under actual environmental conditions, they exert, in all probability, a modifying or inhibiting effect on the total carcinogenic potency of the milieu in which they exist.

Multiple exposure.—It has been experimentally and clinically shown that tissues reacting to proliferative stimuli are at increased risk to neoplasia when simultaneously or subsequently exposed to carcinogenic stimuli. Experimental attempts to induce squamous-cell cancer of the lung by the exposure of mice to aerosols of carcinogenic hydrocarbons alone or to influenza virus alone have been uniformly unsuccessful. However, in a series of experiments just completed by us, exposure to these environmental agents in combination has resulted in the production of the specific human type pulmonary neoplasm.

Metabolism of carcinogenic hydrocarbons.—The basis for variations in response to environmental carcinogenic hydrocarbons is still an enigma. In an attempt to distinguish the susceptible from the nonsusceptible in an exposed population, we have been studying the metabolism of PAH. Experimentally it can be shown that interference with detoxification of PAH in the hepatobiliary system results in a delayed clearance of these compounds from the site of administration. Accompanying this there is a qualitative and quantitative alteration in the profile of metabolites recoverable from the bile. In parallel bioassay experiments, transient mild hepatic injury at the time of carcinogenic administration resulted in a significant increase in tumor yield in mice, when compared with controls, following administration of carcinogenic PAH. While the role of hepatic function may be but one of the many factors associated with responses to carcinogenic PAH, it is our belief that this factor is in all probability operable in humans and thereby contributes to differences in susceptibility.

Other factors almost certainly responsible for the epidemiological pattern of human lung cancer include: (1) the action of promoting agents, which though in and of themselves are noncarcinogenic, frequently determine the time of appearance and rate of incidence of malignant neoplasms; (2) the

presence of concomitant or antecedent pulmonary disease, which also appears to be significant since chronic bronchitis and tuberculosis are apparently associated with increased risk to pulmonary cancer; and (3) occupational factors, which in certain instances have been shown to be positively related to an increased risk to the development of lung cancer, and may have a more universal application than heretofore has been recognized.

From these limited examples, it is apparent that the hazard can be only partially defined by exclusive quantitation of carcinogenic agents in polluted air. The data also strongly suggest that even minute doses of carcinogens must be regarded as hazardous, even when dose response criteria in experimental species may suggest the absence of a hazard. Conversely, it must be emphasized that larger concentrations, though wholly undesirable, of course, may under certain circumstances represent minimal hazard.

This presentation of specific laboratory data has been included primarily to qualitatively denote the existence of a hazard in polluted air. Quantitative discussion, independent of the consideration of the significance of other environmental hazards, is difficult if not impossible.

Bronchogenic carcinoma represents one of the current critical problems in the field of pulmonary disease. Laboratory investigation can contribute

much information to the ultimate solution of this problem. In the physical science area, finite analytical data are possible. In the biological realm, strong supporting data can be secured despite the fact that experimental investigations are necessarily limited to nonhuman animal species. It is necessary, of course, to remember certain deficiencies inherent in the biological studies. Choice of species, selection of appropriate animal strain, duration of exposure, concentration of test material, and routes of administration are all variables that modify the extrapolation of experimental data. from other animals to man. Despite these limitations, past experience has shown a high index of meaningfulness of animal experiments for the human species. The broad spectrum of agents carcinogenic for visceral organs in experimental animals and apparently for man should make one proceed with caution in attempting to attribute absolute dominance of any one agent or source over another.

Epidemiologically, a reduction in lung cancer incidence may be properly anticipated as a result of reducing the concentration of carcinogenic agents in polluted air or any of the environmental sites discussed. It is our belief, however, that the reduction would be of a low order of magnitude in the absence of the removal of the remaining sources of irritants and carcinogenic agents from the respiratory environment.

The contamination of urban air by pollutants of various types and from various sources represents a problem which merits our full attention. The effect of such pollution can be immediate or extend over a period of years. It is the latter effect and its possible relationship to lung cancer that is under discussion here.

In studying air pollution, it is important that we include all of those factors that contribute to it, such as components resulting from combustion, industrial waste, road dust, and radiation. These factors must be studied, not only by themselves but also in terms of their interaction. Such a study requires the utilization of various phases of scientific research, including epidemiological and laboratory investigations.

We have recently reviewed the epidemiological evidence linking air pollution to lung cancer (1). We regard it as established that the urban population has a somewhat higher rate of lung cancer than does the rural population as shown by the study of Hammond and Horn (fig. 1) (2). Other evidence pointing to an "urban factor" as recently presented by Haenszel and his associates indicates that individuals who were born in rural areas but who lived later in metropolitan areas had a higher rate of lung cancer than those who lived in metropolitan areas all their lives. The authors present no explanation for this finding (3). The higher lung cancer rates of the urban population apply primarily to cigarette smokers. The recent survey

by Haenszel et al. also indicates that for nonsmokers the difference in rates in urban and rural areas is "trivial" (2). A major problem of the epidemiologist in evaluating the factor of air pollution is that the correlation between cigarette smoking and lung cancer is of such an order of magnitude as to overshadow other factors (fig. 2). In evaluating the basis for the "urban factor," we must, in addition to air pollution as such, consider the more common occurrence in cities of occupations that have higher-than-average rates of lung cancer, the tendency of city populations to smoke more cigarettes than the rural population, and the fact that cancer in general is reported somewhat more commonly in cities (1).

It is apparent that factors of different natures, both exogenous and endogenous, may be involved in the induction of lung cancer. The epidemiologist should consider all these factors as far as epidemiologic techniques permit. Viral and bacterial infections, for instance, may also play a role here. Recent evidence from our own studies suggests that chronic bronchitis influences the induction of lung cancer in man. Laboratory evidence suggests a similar role for viruses in the experimental animal (4). It is conceivable that the "urban factor" may, therefore, also be related to certain lung diseases which may occur more commonly in cities. In this connection, studies on the possible influence of climatic factors as well as population

680024-63- -11