Document YjgxzaXdDjNNMgdeG2g0nYYbK

Asbestos: Scientific Developments and Implications for Public Policy B. T. Mossman, J. Bignon, M. Corn, A. Seaton, J. B. L. Gee Physico-Chemical Characteristics of Asbestos Asbestos is a commercial term for a group of fibrous minerals often associated with the development of pulmo "Asbestos" is a broad commercial term for a group of naturally nary interstitial fibrosis (asbcstosis), lung cancer, and occurring hydrated silicates that crystallize in a fibrous habit. The malignant mesothelioma in occupationally exposed indi legal definition of a fiber as promulgated by the EPA and other U.S. viduals. The pathogenicity of different forms of asbestos regulatory agencies is one that possesses a 3:1 aspect ratio. varies--long, thin amphibole fibers are most pathogenic, However, this definition has been criticized by mineralogists (4). particularly in the induction of mesothelioma. Available Asbestos fibers in ores are not respirable until released and made data do not support die concept that low-level exposure airborne during mining and processing. The family of asbestos to asbestos is a health hazard in buildings and schools. minerals can be subdivided into serpentine and amphibole fibers The concentration of asbestos fibers in air, type of asbes (Fig. 1). Chrysoule, which accounts for over 90% of the world's tos, and size of fibers must be considered in evaluation of production of asbestos, is the most common fibrous serpentine, potential health risks. whereas the amphiboles, a chemically diverse group of less industri ally important minerals, include the fibrous minerals and crocidolice, amosite, anthophvilite asbestos, actinolite asbestos, and tremolite asbestos. Tremolite, actinolite and anthophyllite, which occur in sbestos engenders both fear and panic in U.S. both fibrous and nonfibrous forms, have been only rarely mined for Asociety. Observation chat asbestos-containing materials use as commercial asbestos. Both the fibrous and nonfibrous forms (ACM) have been used in schools, buildings, and hospitals, of these amphibole minerals are sometimes found as contaminants and the Asbestos Hazard Emergency Response Act (AHERAo)f, caommercial deposits of chrysorile, talc, vermiculire, and other mandate from the Environmental Protection Agency (EPA) mthinaet rals (4). The nonfibrous forms of crocidolite and amosite are requires inspection of the nation's public and private schools for referred to as riebeckice and grunerite, respectively. asbestos, have resulted in the explosive growth ofasbestos identifica The various types of asbestos fibers differ in their chemical tion and removal companies. By EPA estimates, extension of EPA composition, morphology, and durability. Therefore, the biologic requirements to approximately 733,000 public and'commercial effects should be considered individually for each fiber type. Identifi buildings containing asbestos will cost SS3 billion, discounted at cation of specific types of asbestos in air samples requires sophisti 10% over 30 years.(I). Because of uncertainties regarding the cated technology such as transmission electron microscopy, x-ray amount of asbestos and its condition in these buildings, estimates diffraction, or energy dispersive x-ray spectroscopy. The rod-like for removal of asbestos are as high as SI00 to S1S0 billion (2). amphiboles appear to penetrate the peripheral lung more readily Asbestos was shown to cause asbcstosis at the turn ofthe century. Its association with the causation of lung and pleural tumors in asbestos miners and wcukets was demonstrated in the 1950s and 1960s, respectively (J). An important issue is whether these diseases arc also hazards to the general population exposed to airborne levels of asbestos in schools and other buildings. Does available evidence support the concept thac asbestos causes disease in die nonoccupational environment? What are the mechanisms of asbestos-induced fibcogenesis and carcinogenesis? Most importantly, have recent data been adequately considered in formulating policies in the United States for regulation and banning of asbestos? In this article, we summarize recent developments and discuss their implications for Table 1. Summary statistics for average airborne fiber concentrations in U.S. schools (88) and buildings (89). The data used in the calculation of each statistic are the average concentrations (expressed as number offibers greater than S pm in length per cubic centimeter of air) in a building (for indoor samples) or the concentration outside each building [for outdoor samples (89)]. By visual inspection, category 1 buildings contained no asbestoscontaining material (ACM), category 2 buildings contained ACM in primar ily good condition, and buildings in category 3 showed at least one area of significantly damaged ACM. In the study on public buildings, 387 indoor and 48 outdoor air samples were evaluated. No asbestos fibers were detected in 83% of the 387 samples. The sample size is given in parentheses below each heading. public policy. B. T. Matanui a a member of the Department of Pnhoiocy of the Unvcninr of Vomom, College of Mcdxmc, Burfcnoon, VT 0540S. |. Brgnon it director of the Unite IN5ERM de Recherche tor b Siopedidogie a h Tauoaiogic Putaonairc a Houle, Octeil, Fiance. M. Com n difeaor of the Dtvtrion of Enviicnmcnal Hakb Engineering. School of Hygiene and Public Health, Iohm Hopkins Unheruty, Bafciniarc, MD 2120S. A. Seaton b dmetnrof the Imriaicc of Oonmarional'Medidne, Edinburgh. Scotland EH8 9SU.4. B. L. Gee ia a director ofthe Wmdusicr Chat CSnic and member of the Ditmemmi of Internal Medicine, Yale Unmntty School of Methane,' New Haven, CT 06510. Statistic Schools (71) Outdoor air (48) Median Mean SD 0.00024* 0.00053 0.00000 0.00039 0.00096 Public buildings Category 1 (6) Category 2(6) Category 3(37) 0.00010 0.00099 0.00198 0.00040 0.000S9 0.00052 0.000S8 0.00073 0.00072 50th percentJc - 0.0004S; 90th percentile - 0.00083. 296 SCIENCE, VOL. 247 369054 0343 Fig. 1. Unification and morphology of .isix-Moc fibers. The inserted photographs ire wanning. electron micrographs of Union Imemationalcconrre le Cancer refcreiK-e samples of chrysotilc A (left) and inx'idolnc asbestos (right) (field ofview is - 12 M-m in both photographs). The amphiNilcs arc depicted in order of their reljnvc industrial importance. Serpentine Chrysolite M96SI410(OHIS Asbestos Amohibole Crocrdohte Na2lFe3T3iFeJT3Si80,,i0Hi? Amosile (Fe.MgljSijOjjlOHlj AntbopbyiMe Tremolite Actinolile IMg.Fel7Si,OjjlOHIj Ca2Mg5Si,OjjlOHIj CajlMg.Fe)sSi,Ojj(OH)j than chrysotilc fibers, which arc curly, can occur in bundles, and can lx: intercepted at airway bifurcations (Fig. 1). The chemical makeup of each fiber type is complex, and fibers may consist of a variety of trace metals and organic compounds acquired in ores or during processing. Asbestos is attractive to industry because of its resistance to heat and chemicals, high tensile strength, and lower cost compared to exhibited a striking increase in lung cancers with duration of man-made materials. Although use of spray-on asbestos as a fire exposure when compared to Canadian chrysotilc miners and millers. proofing material or insulation has been banned in this country, as Solvents and oils in textile production might act as cocarrinogcns in well as in several European countries, asbestos is incorporated the development of these lung rumors (7). currently into cement construction materials (roofing, shingles, and Diffuse malignant mesothelioma is a fatal tumor arising from cement pipes), friction materials (brake linings and clutch pads), mesothelial cells or underlying mesenchymal cells in the pleura, jointing and gaskets, asphalt coats and sealants, and other similar pericardium, and peritoneum (8). The time between diagnosis and products. As a result of these applications, an estimated 20% of initial occupational exposure to asbestos commonly exceeds 30 buildings including hospitals, schools, and other public and private years. Smoking evidendy docs not enhance risk of mesothelioma in structures contain ACM (1). Asbestos in buildings does not sponta asbestos workers (J). Although mesotheliomas are extremely rare neously shed fibers, but physical damage to ACM by decay, malignancies, that is, only 1648 were recorded from 1973 to 1984 renovation, or demolition can cause release of airborne fibers (5). in one survey covering approximately 10% of the U.S. population (9), they may account for as much as 18% of the proportional mortality in crocidolite workers (10). Mesotheliomas also have been Diseases Associated with Occupational Exposure to Asbestos observed after household exposure of family members of asbestos wrorkers and in individuals living in close proximity to asbestos mines (11). Although mesothelioma has been considered by some as Occupational exposure to asbestos can cause four types of disor a disease pathognomonic of exposure to asbestos, approximately 20 ders: asbestosis; lung cancer; mesotheliomas of the pleura, pericardi to 30% of mesotheliomas occur in the general population in adults um, and peritoneum; and benign changes in the pleura (3). Asbcsto- not exposed occupationally to asbestos (12). Mesotheliomas are sis, a pulmonary interstitial fibrosis with excessive deposition of rarely found in children. collagen, caused progressive lung stiffening, impaired gas exchange, Diagnosis of mesotheliomas is a challenge as the tumor may disability, and death in many workers exposed before the enforce resemble metastascs of other rumor types occurring in the pleura or ment of occupational standards. Lung cancers, that is, tumors peritoneum and assume a wfide variety of microscopic appearances. arising in tracheobronchial epithelial or alveolar epithelial cells, have Thus, death certificates may either underestimate (because these occurred in asbestos workers in most cases 20 or more years after tumors are attributed to cancers of the gastrointestinal tract and their first exposure to asbestos. In general, lung cancels have been other organs) (13) or overestimate the incidence of mesotheliomas. found in asbestos workers who are smokers and only rarely in In France, mesotheliomas are overestimated by a factor of 3 on nonsmokers (tf). A number of epidemiologic studies have indicated death certificates in comparison to the mesocheliofna registry (14). that the relation between the development of lung cancers and A number of benign pleural changes that rarely cause functional cumulative exposure to asbestos is approximately linear, but wide impairment have been observed in asbestos workers. These include variations in slope of the line occur apparently related to fiber type pleural effusions, pleural fibrosis, pleural plaques, that is, accumula and industrial usage (7). Death rates from lung cancers in asbestos tions of acellular collagen on the diaphragm and chest wall, and workers, as measured by standard mortality ratios (SMRs), the pscudocumors or infoldings of the lung often associated with observed mortality of a.cohort divided by the mortality of a control . plaques. These pleural changes may reflect exposure to asbestos but population, arc lowest in chrysotilc miners and workers manufac have no demonstrated relation to the development ofmesothelioma. turing friction rtiatcrials..in contrast, lung cancer deaths are higher Tumors of the gastrointestinal tract, larynx, arid other organs in those mining and working with amphiboie asbestos. Textile including the kidney, ovary, pancreas, pericardium, eye,, and lym workers in a South Carolina plant in which chrysotilc was used phatic system, have been reported in some Cohorts of asbestos >9 JANUARY 1990 ARTICLES 29} 369054 0344 workers (13, IS). In general, the enhanced SMRs for these tumors arc not sransncallv distinguishable from normal SMRs and have nor been confirmed in most cohorts. Both laryngeal and gastrointestinal rumors have other etiologies such as smoking, alcohol, diet, and intestinal polyposis that confound the interpretation of epidemio logic data. The Amphibole Hypothesis The association of mesothelioma with asbestos exposure was first described in 1960 in the northwest Cape area of South Africa where long, thin crocidolite fibers were mined (16). Since then, an increased incidence of mesothelioma has been reported in a number of occupational settings including factories char presumably used only chrysotile. Within the past decade, sophisticated technology has allowed examination of the types of fibers in the lung tissue of these workers. Results revealed that many chrysotile-exposed work ers showed an appreciable lung burden of amphibole fibers, which were used for brief periods in the workplace (17). The persistence of amphibolcs in human lungs may be attributed to their increased ability to penetrate the peripheral lung, lack of clearance, or durability. In contrast, chrysotile has been found post-mortem in smaller amounts than expected in the lungs ofasbestos workers (18). It disappears with time most likely because magnesium and silica are leached from the fibers (19). Recently, the lung content of asbestos and nonasbestos fibers has been compared in diagnosed cases of mesothelioma, lung cancer, and cardiovascular disease (controls) from the western coast of France, a region containing shipyards (20). The number of amphibole fibers (crocidolite and amosite) was significantly higher in lungs from mesothelioma patients, whereas numbers of chrysotile and nonasbestos fibers were similar in all groups. These data suggest that the lung burden of chrysotile and nonasbestos fibers bears no relation to the occurrence of these cancers. Several recent studies indicate that the risk of pleural mesothelio ma is lower where chrysotile is used without admixture or contami nation by amphiboles (21). For example, a gradation of death rates from mesothelioma has been observed in both male and female asbestos-exposed cohorts. Mesothelioma has been responsible for approximately 6 to 8% of the proportional mortality in men working with mixtures containing crocidolite or amphibole (crocid olite or amosite) in comparison to less than 1% of the proportional mortality in men working with chrysotile (10). In female cohorts, the proportional mortality from mesothelioma was highest for amphibole exposure (10.6%) and lowest for chrysotile exposure (0.2%). Thus, these data suggest that amphiboles are the major cause of mesotheliomas in asbestos workers. Chrysotile miners and millers in Quebec who were supposedly exposed only to chrysotile have developed few mesotheliomas (22). However, leeem fiber analyses on the lungs of both these workers and chrysotile factory workers showed the presence of tremolite (23). This amphibole in the fibrous form has been implicated as the causative agent of mesotheliomas and lung cancers in miners exposed to vermiculite heavily contaminated with tremolite (24). Although tremolite composes less than 1% of the asbestos dust in the Quebec mines and mills, the relative ratio of tremolite to chrysotile fibers in the lungs of Canadian miners and millers is related directly to their risk of developing mesothelioma (25). For the reasons above, the few mesotheliomas observed in Canadian chrysotile workers appear to be attributable to fibrous tremolite, an observation compatible with other evidence that amphiboles are the most pathogenic asbesriform minerals. Likewise, recent data on London asbestos factory workers show that the 296 * seventy of asbestosis and carcinoma ot the lung s as well as mesothe lioma) correlates with the lung burden of crocidolite and amosite asbestos and that the proportions of chrysotile and nonasbestos fibers are decreased in comparison to marched control patients (261. A British cohort exposed since 1970 to chrysotile at airborne levels not exceeding 0.5 to 1.0 fiber per cubic centimeter in the manufac ture of friction materials showed no excess of deaths from lung cancer, ocher asbestos-related rumors, or chronic respiratory disease (27). These and ocher data (7, 17, 21, 28) suggest that amphibolcs are more potent than chrysotile in the induction of fibrotic lung disease and associated lung cancers. Experimental Models of Asbestos-Induced Lung Disease Several studies have shown that mesotheliomas are induced in a dosage-dependent fashion after intrapleural and intraperitoneai in jection of asbestos and other asbestos-like fibers into rodents (29). Chrysotile was as carcinogenic as the amphiboles by these routes of administration. However, differences have been observed between the carcinogenicity of fibrous and nonfibrous materials. For exam ple, in one study, fibrous tremolite was carcinogenic after intra pleural injection, whereas nonfibrous tremolite was noncarcinogcnic at identical concentrations (JO). Although the natural route of exposure to fibers by inhalacion was circumvented in these experi ments, they were useful in indicating chat fibers longer than 8 pun and less than 0.25 pm in diameter have the most marked carcino genic potential, that is, the "Stanton hypothesis." These data have been supported by the results of inhalation studies in rats in which short (5 pm in length) arid long fiber preparations of amosite and chrysotile asbestos have been compared (31). In contrast to the batches of amosite and chrysotile asbestos containing many long fibers, short fibers of amosite produced neither asbestosis nor pulmonary tumors. Short chrysotile produced a small amount of asbestosis and malignancies, but these were attributed to contamina tion of the short chrysotile preparation by longer fibers. Fewer long than short fibers of both types were present in the lungs ofall rats at the termination of exposure, but, regardless of size, fewer chrysotile fibers remained in the lung. These results support the observations that chrysotile fibers, in comparison to amphibole fibers, are cleared more rapidly from human lungs (17). This phenomenon and limited alveolar penetration of curly chrysotile bundles (rather than their inherent absence of carcinogenicity) may account for the apparent lack of association of chrysotile fibers with the development of mesothelioma in human cohorts. The exorbitant costs of inhalation experiments with animals preclude long-term studies to determine the carcinogenic potential of asbestos at low-level exposures. The development of malignancies in rodents approaches the 2- to 3-year life-span of these animals (32), a period of too brief to reflect the consequences of the long term solubility of chrysotile in the human lung. Mechanisms of Asbestos-Induced Inflammation and Fibrogenesis Both epidemiologic and experimental data support the concept o a threshold for chrysocile-induced pulmonary fibrosis. In a sheet model of asbestosis, inflammation and histopathologic evidence c disease were npt observed after less than 100'mg of chrysotile wet injected into the trachea of the sheep (33). After brief, intent inhalation of chrysotile, the sheep .accumulated alveolar macn phages (AMs) at areas ofdeposition offibers (4). These cell types at SCIENCE, VOL. 24 360054 0345 V viewed as "effector" cells of disease as they produce a mixrure of fibroblast growth factors, chemoncric factors, and fibronectin. Prosraglandins, plasminogen activator, a heat-stable factor sitmljr to platelct-denved growth factor (PDGF), lysosomal enzvmcs, and active oxygen metabolites, one or more of which may cause proliferation or functional impairment of neighboring epithelial cells and fibroblasts in che lung, were released after exposure of AMs to asbestos in vitro (35). These substances might mediate both acute and chronic inflammatory reactions in man and animals after inhalation of asbestos. In support of this hypothesis, AM-derived growth factor (AMDGF), PDGF, superoxide (Of), and H2O2 were spontaneously released from AMs recovered bv bronchoalveolar lavage from patients with asbestosis (36). Similarly, AMs lavaged from both mice and sheep chat had an earlier intratracheal injection of chrysotile released enhanced amounts of a growth factor that stimulated proliferation of a human embryonic lung cell line (WI38) (37). In one study, AMs from both normal individuals and patients with idiopathic pulmonary fibrosis expressed a 4.2-kilobasc messen ger RNA complementaiy.ro c-sis, a proto-oncogene coding for the B chain of PDGF (38). The amounts were approximately fourfold higher from AMs of patients with pulmonary fibrosis. Because PDGF is mitogenic to mesenchymal cells, which possess functional PDGF receptors, elevated levels of PDGF in lung tissue or fluids could induce lung fibroblasts to divide or to produce exorbitant amounts of collagen, the hallmark of the fibtotic lesion. Quiescent human mesochelial cells also undergo DNA synthesis after exposure to PDGF and a broad spectrum of other growth factors (39). Within the past few years, several laboratories have focused on active oxygen species (AOS) as causative agents of both asbestosis and asbestos-related malignancies. Increased amounts of superoxide (Of ) have been produced after rodent AMs were exposed in vitro to long asbestos fibers, whereas generation was minimal after shorter fibers and nonfibrous particles were introduced (40). Smaller fibers and particles are incorporated into phagolysosomes by AMs, whereas longer fibers arc incompletely phagocytosed, a process liberating more AOS. The observation that exogenous administration of scavengers of AOS prevents asbestos-induced cell death to cultures of tracheal epithelial cells and lung fibroblasts (41) suggests that AOS are intimately related to asbestos toxicity even in the absence of AMs. Fibers may induce generation of AOS after phagocytosis or by extracellular mechanisms. For example, recent studies with asbestos in cell-free systems have demonstrated by electron spin resonance' that chrysodle, crocidolite, and amosite generate AOS in the presence of H3O2 or physiological saline (42). Under these circum stances, Fc2* on the surface of the fiber appears to drive a modified Haber-Weiss (Fenton) taction that results in production of the toxic hydroxyl radical (OH') from HjOj and Of. These reactions result in lipid peraxidatioa, which is prevented by incubation of asbestos with the iron chelator, desferoxamine (43). At high concentrations, AOS are cytotoxic to cells of the respira tory tract, but at low concentrations they induce functional changes in rodent lung fibroblasts that may be critical to the pathogenesis of asbestos-induced fibnxic lung disease. For example, after addition of xanthine and xanthine oxidase (a chemical generating system producing Of), rat lung fibroblasts in vitro produced increased amounts of cell-associated collagen in a pattern similar to that observed after their exposure to crocidolite asbestos (40, 44). In an inhalation model of rapid-onset asbestosis, osmotic pumps- contain ing polyethylene glycol (PEG)-conjugated catalase, the enzyme ^avenging H2Q2, were implanted subcutaneously into rats before they were exposed tb crocidolite -for 20 days (45). This procedure boosted levels of catalase in the sera and lungs of these animals and ameliorated both the inflammation and the severity and extent of fibroric lesions that normally develop after inhalation of asbestos. This study was the first successful experimental approach to che prevention of asbestos-associated lung disease. Moreover, the results support the concept of a cause and effect relation between AOS and the development of asbestosis. Mechanisms of Asbestos-Induced Carcinogenesis Carcinogenesis is a multistage process that classically has been described in two stages (46). The "initiation" stage corresponds to a heritable generic change (point mutation) induced in a cell by a carcinogenic substance. It is followed by the "promotion" stage, a series of events in which the initiated cell undergoes proliferative and genotypic changes conferring the malignant phenotype. During the past few years, the identification of a number of protooncogenes has resulted in a new understanding of the successive genetic events involved in the process of malignant transformation. Increased expression of these genes may cause the production of growth factors or growth-factor receptors. Loss ofother genes (antioncogenes) also appears to contribute to the carcinogenic process. These findings indicate that the distinction between genetic and epigenetic events in carcinogenesis is not simple, especially because chromosomal rearrangements or deletions associated with point mutation and activation or loss of genes can happen at any stage in the process of cell transformation. Whether the multistage model is directly applicable to asbestosassociated carcinogenesis is unclear. Unlike most carcinogens, asbes tos does not cause base substitution and frameshift mutations in bacterial-mutation assays (47). Of die 23 agents designated as group 1 human carcinogens by the International Agency for Research on Cancer (IARQ, only asbestos and conjugated estrogens were nongenotoxic as defined by both the Ames test and rodent bonemarrow assays for detection of chromosomal aberrations or micronucleated erythrocytes (48). Although asbestos was weakly muta genic in Chinese hamster lung cells (49), it was not mutagenic in liver epithelial cells or in Syrian hamster embryo (SHE) fibroblasts (50). Asbestos did not cause morphologic transformation of C3H 10T1/2 cells (51), but transformed both BAJLB/c#3T3 and SHE fibroblasts (52). Glass fibers and nonfibrous silica (albeit at much higher concentrations) also were active in the SHE bioassay. In this system, longer, thinner fibers were mote potent in the induction of transformation and chromosomal anomalies, an observation consist ent with the increased malignant potential ofthese fibers in compar ison to shorter fibers or particles after their administration intrapleurally, intraperitoneally, or by inhalation to rodents (29, 30). In these and other in vitro studies, the biologic effects of fiber types have been assessed comparatively on a mass (milligrams of fibers per dish) rather than a numerical (numbers offibers ofa given size per dish) basis. Cytotoxicity and cytogenetic effects of chryso dle, crocidolite, and erioahe (an aluminosilicate fiber) recendy were compared in Chinese hamster hmg fibroblasts (V79 ceils) (S3). Numbers of chrysodle fibers required to produce cytotoxic or cytoge netic changes were several orders ofmagnitudes higher in comparison to crionitc, the most potent fiber, or crocidolite, a fiber ofintermediate potency. These results are consistent with the higher rumorigenk potential of crionitc in rodent inhalation experiments (54). In some studies, asbestos appears to augment the mutagenic and carcinogenic effects of chemical carcinogens and radiadon. For example, both crocidolite and chiysodk increased the frequency of' mutation and transformation in rodent epithelial cells and fibro blasts exposed to benzo(]pyrene (BaP) (50) and radiation or radon 19 JANUARY 1990 ARTICLES 297 369054 0346 alpha particles (57). However, synergistic effects of asbestos and BaP were not observed in two studies with SHE and rat mesotheliai cells, respectively (52, 55). The particulate nature of asbestos and its capacity to bind nucleic acids has prompted transfection studies in which asbestos was used as a vehicle for introducing DNA or &NA into a number ofcell lines (56). Under these circumstances, asbestos was intermediate in rank in comparison to a number of ocher insoluble facilitators including calcium phosphate, talc, and kaolin, none of which have been associated with the induction of cancer. After addition to human or rat mesotheliai cells, both chrysotilc (in rats) (57) and amosite (in humans) (58) have caused ancuploidy and altered growth characteristics after repeated passaging. Injection pf rat mesotheliai cells into nude mice after a single exposure to chrysotilc did not cause tumors in animals, but multiple exposures (36 times) to chrysotile and repeated passaging resulted in tumorigenic cell populations (57). In contrast, human mesotheliai cells displaying chromosomal abnormalities and growth alterations after duplicate exposures to cytotoxic concentrations of amosite were not tumorigcnic in nude mice (58). Asbestos promoted the proliferation of mesotheliai cells both in organ cultures of human mesothelium exposed to asbestos in vitro and in mice given intrapicntonca] injections of asbestos (59). Asbestos fibers come in contact with the chromosomes of rat mesotheliai cells (60) and the mitotic apparatus of V79 (53) and SHE (67) cells in vitro. These interactions might induce chromo somal misaggregacion or abnormalities. Several investigators have examined chromosomal aberrations in human mesotheliomas, but changes appeared inconsistent from tumor to tumor. The most common abnormalities involved inversions, translocations, and de letions of chromosomes 1, 3, 7, 9, 17, and 22 (62). Constitutively enhanced expression ofthe PDGF-B gene, the proto-oncogene c-sis, was observed in human mesothelioma cell lines when compared to normal human mesotheliai ceils (65). In comparison to human mesotheliai cells, human bronchial epithelial cells in vitro are relatively resistant to asbestos. In one study, concentrations of chrysotilc, crocidolite, or amosite asbestos approximately ten times as high as that required for mesotheliai cells were required to achieve a comparable increase in toxicity (as measured by a 50% decrease in colony-forming efficiency of human bronchial epithelial cells) (64). In another study, aneuploidy was not increased significandy over a range of concentrations of either crocidolite or chrysotile asbestos (65). This latter observation and the demonstration chat insertion of asbestos into rat tracheal grafts can cause the development of carcinomas following insertion of subcarcinogcnic amounts of the polycyclic aromatic hydrocarbon, dimcthyiboi2o[d]anthncene (66), suggest that asbestos is a pro moter in the development of lung cancers. In support of this concept, both crocidolite and chrysotilc asbestos induced a number of biochemical and proliferative alterations in both rodent and human tracheal epithdiai cdf and organ cultures that were similar to those observed in mouse skin that had been treated with the tumor promoter 12-0-tendecanoylphocbot-I3-acetate (TPA) (67). The repertoire of these asbestos-associated proliferative changes, which were masked in tracheal epithelial cells in a high-serum containing medium or by addition of transforming growth factor-type B (TGF-Pi) (63), included enhanced incorporation of JH-thymidinc, increases in colony-forming efficiency, and the development of squamous metaplasia, that is, conversion of differentiated mucoci liary cells to keratinizing cells resembling- epidermis.' Induction of ornithine decarboxylase (ODC),, a rate-limiting enzyme in the biosynthesis of polyamines that is increased in mouse skin after exposure to TPA but not after addition of nontumor promoting phorbol derivatives, also occurred in a dosage-dependent fashion in tracheal epithelial cells exposed to long, thin, asbestos and glass fibers. In contrast, nonfibrous particles and shorter fibers did not increase ODC activity at similar concentrations (69). Until quite recendy, it was unclear how asbestos triggered proliferation in tracheal epithelial cells. However, several pieces of data suggest chac mechanisms ofcell signaling by asbestos are similar to those observed with TPA, a soluble tumor promoter that binds dirccdy to protein kinase C (PKC), a calcium- and phospholipiddependent enzyme that activates a limb of the phosphoinosiode signal-transduction pathway (70). Mitogenic concentrations of crocidolitc asbestos caused increased accumuladon of diacylglycerol in tracheal epithelial cells (77) and subsequent aedvation of PKC (72), presumably by activation of membrane phospholipases. The in creased production of inositol tris- and tetrakisphosphates appeared responsible for the generation of diacylglycerol, which preceded increased cell division. Abrogation of crocidolite-induced ODC activity in. tracheal epithelial cells by inhibitors of PKC and calcium channel antagonists (69) suggests that PKC is related causally to asbestos-associated cell proliferation. Mechanisms other than tumor promotion by asbestos also may explain interactions between smoking and asbestos; these mecha nisms could be important in the development of lung cancers in asbestos workers (7, 75). For example, smoking impaired clearance of amosite asbestos from rodent lungs and increased retention of fibers in airway epithelial cells (74). Both cigarette smoke and asbestos induced AOS in a synergistic fashion in vitro and damaged isolated bacteriophage DNA (75). AOS liberated from asbestos fibers also catalyzed the oxidation of 6-hydroxybenzo[a} pyrene to a more mutagenic and carcinogenic radical (76). Because crocidolite and chrysotile asbestos adsorbed BaP and acted as vehicles to increase both uptake of these lipophilic carcinogens and formation of DNA adducts in tracheal epithelial cells (77), fibers might facilitate the initiation of lung tumors by BaP. Public Policy The available experimental and epidemiological data indicate that both fiber type and size are important determinants of the pathoge nicity of asbestos. Although asbestos has caused disease in the workplace (78) and such occurrence has resulted in calls for regula tions to protect workers (79), recent epidemiologic data are concor dant with the suggestion that exposure to chrysotile at current occupational standards does not increase the risk of asbestosassodated diseases (77, 27, 27, 28). Unlike most other countries, particularly in the European community, which have more stringent requirements for regulation and importation of amphiboles, federal policy in the United States does not differentiate between different types of asbestos. Does airborne asbestos present a risk to the health of individuals in schools and other buildings? The available data do not indicate that asbestos-associated malignancies or functional impairment will occur as a result of exposure to most airborne concentrations of asbestos in buildings.. First and foremost, the levels of airborne asbestos in buildings, even with damaged ACM, are magnitudes lower than concentrations in the unregulated workplace in the past and approximately 1/100 of the permissible exposure of 0.2 fibers per cubic centimeter of air in the U.S. workplace (80). Before the enforcement ofoccupational standards, workplace concentrations of 100 or more fibers per cubic centimeter of air were not uncommon (87). In contrast, surveys ofasbestos in schools and public buildings show that the mean airborne concentrations are several thousand fold lower (Table 1). With few exceptions, the type of asbestos fiber found predominantly in buildings is chrysotilc. Accumulating evi- 298 SCIENCE, VOL. 14' 369054 0347 itcncc indicates that this asbestos type is probablv not associated with the occurrence ot mesotheliomas at low levels of exposure. For example, recent analyses on the fiber concentrations in lungs of asbestos workers showed that chrysotilc workers with mesothelioma had 400 times the median lung fiber burden in comparison to workers exposed to amphiboles. Data indicate that mesotheliomas in chrysotilc workers appeared at lung burdens comparable to that required for the development of asbestosis, a Hicnaa- associated with occupational exposure to asbestos in the past unregulated workplace {82). Transmission electron microscopy of air samples is essential for the identification and quantitation of finer asbestos fibers. In the United States and United Kingdom, the direct transmission electron microscopy method is advocated to determine airborne asbestos fiber concentrations in buildings. In France, the indirect transmis sion electron microscopy technique is used, and concentrations are expressed on a mass (milligram) basis. The limit for detection of fibers by phase-contrast microscopy is approximately 0.01 fibers per cubic centimeter of air, a concentration higher than that reported in most schools and buildings (Table 1). Moreover, phase-contrast microscopy cannot be used to identify types of fibers (asbestos or nonasbestos) or to detea fibers less than 0.5 p.m in diameter, twice the diameter offibers associated with the greatest biological activity and induaion of tumors in rodents (diameters 0.25 pm, that is, Stanton fibers) (29, JO). Such long, thin asbestos fibers are rarely found in air samples of buildings (81, 83). As shown in Table 1, fiber concentrations from recent studies in buildings are comparable to levels in outdoor air, a point surely relevant to assessing the health risks of asbestos in buildings. Airborne concentrations of asbestos in buildings reported in the 1970s were somewhat higher, presumably because of earlier, less sophisticated sampling and analytical tech niques. Recent epidemiologic studies of deaths from mesothelioma in the general population also suggest that risk from asbestos in buildings is miniscule (9, 10, 84). In comparison to lung cancers (an average of 130,000 cases per year in the United States, largely attributed to smoking), an estimated 1,500 cases of mesothelioma per year occur in the U.S. population (5). The data on death rates from pleural or peritoneal mesotheliomas over the past 10 to 20 years indicate that mesotheliomas are increasing in males over 65 years of age who have a past occupational history' ofexposure to asbestos (84). By contrast, death rates from mesothelioma in females of all ages have declined slightly or remained constant. These results support the concept that asbestos in buildings is not an important risk faaor, as one would expea increased mesotheliomas in both males and females in this case. A recent French study did not show increased risks ofasbestosassociated malignancies, pleural plaques, or functional impairment of the lung (effects dearly present in asbestos workers) in persons exposed for 10 years to airborne asbestos in buildings (86). Al though this survey is still in progress, no mesotheliomas have been observed to date among approximately 15,000 permanent occu pants. Although the validity of extrapolating from high to low dose levels has never been confirmed empirically in the evaluation of asbestos, calculated lifetime risks from mesotheliomas and lung cancers attributable to asbestos in schools and other buildings have appeared in recent years (85, 87). The linear dose-response equations in these models have been used with the assumption that there is no threshold for disease, a hypothesis which is open to question. Moreover, the range of estimated risks varies from study to study. With the exception of one analysis (85), differences between the pathogenic potential of chrysctiles-and amphiboles have noc been considered in these assessments, and the importance of fiber size has been ignored. Regardless, examination of combined data from Table 2. Estimates of nsk from asbestos exposure in schools m comparison toother risks in U.S. jocictv Data Irom sn published risk estmurcs (87) m which total deaths (lung cancer and mesotheliomas) attributable to asbestos exposure over a lifetime were estimated per 1 million students exposed to 0.00024 fibers per eubic centimeter air (the mean airborne concentration in schools. Tabic l) for five school yean, beginning at age 10 Estimates indicate that the annual rate ts 0.005 to 0.093 deaths per million students for an average life expectancy of 75 years. Modified with permission from Weill and Hughes (90). Cause Asbestos exposure in schools Whooping cough vaccination (1970 to 1980) Aircraft accidents (1979) High school football (1970 to 1980) Drowning (ages 5 to 14) Motor vehicle accident, pedestrian (ages 5 to 14) Home accidents (ages 1 to 14) Long-term smoking Annual rate (deaths per million) 0.005 to 0.093 1 to 6 6 10 27 32 60 1200 published risk estimates shows char risks of asbestos-related coral deaths (both lung cancers and mesotheliomas) due ro exposure in schools are magnitudes lower chan commonplace risks in modernday society (Tabic 2). The AHERA ruling of 1986 brought asbestos co che attention of the U.S. public and instilled fears in parents that their children would contraa asbestos-related malignancies because of high levels of airborne asbestos fibers in schools. Panic has been fueled by unsupported concepts such as the "one fiber theory," which main tains that one fiber of inhaled asbestos will cause cancer. As a result of public pressure, asbestos often is removed haphazardly from schools and public buildings even though most damaged ACM is in boiler rooms and ocher areas which'arc inaccessible to students or residents (I). The removal of previously undamaged or encapsulated asbestos can lead to increases in airborne concentrations of fibers in buildings, sometimes for months afterwards (83), and can result in problems with safe removal and disposal. Asbestos abatement also has led to che exposure of a large new cohort of relatively young asbestos removal workers. While these people should be proreaed by careful regulation of the circumstances of removal, they are often exposed under suboprimal working conditions. As a result of the AHERA ruling, public and private schools are required ro inspea for asbestos and inform parents if ACM are present. Although the law does not require or set standards for the removal of asbestos, schools, often with little expert advice, must submit a management plan detailing how they will deal with damaged asbestos and can be fined a maximum of S5000 per day for lack of compliance ro deadlines. The EPA has recommended bulk sampling of ACM to determine the presence of asbestos and visual inspection to determine the course of aaion, rather than measure ment ofairborne levels offibers--data that arc far more important in determining the need, if any, for removal of ACM. The available data and comparative risk assessments (Table 2) indicate that chrysotilc asbestos, the type of fiber found predomi nantly in U.S. schools and buildings, is not a health risk in the nonoccupational environment. Clearly, the asbestos panic in the U.S. must be curtailed, especially because unwarranted and poorly controlled asbestos abatement results in unnecessary risks to young removal workers who may develop asbestos-related cancers in later decades. The extensive removal of asbestos has occurred less fre quently in Europe. Prevention (especially in adolescents) of tobacco smoking, the principal cause of lung cancer in the general population, is both a more promising and rational approach ro eliminating lung rumors 19 January 1990 ARTICLES Z99 369054 0348 chan asbestos abatement. Even acknowledging that brief, intense exposures to asbestos might occur in custodians and service workers in buildings with severely damaged ACM, worker education and building maintenance will prove far more effective in risk prevention for these workers. REFERENCES AND NOTES 1. Rcpon to the Congress. Slurfp of Ajbeuot-Contmming Motrrioh m PiMit Bvtldivgt (U.S. Environmental Protection Agency, Washington, DC, February 1988), p. 5. 2. M. Com, paper presented at the 22nd International Congress on Occupational Health, Sidney, Australia, September. 1986. 3. B. T. Moesman and |. B. L. Get. N. Engl. J. Mrf 320, 1721 (1989). 4. H. C W. Skinner, M. Rocs, C Frondel, Eds., Askeiwt and Ofor Fihoom Mouriob (Oxford Umv Press, New York. 1988). 5. M. Com, Am. Ind. Hyg. Auot. J. 47, SIS (1986). 6. R. Saracci, Epid. Rrv. 9, 175 (1987). 7. J. C. McDonald and A. D. McDonald, in Asbestos-Reload Moligvmey, K. H. Amman and ). Aisner, Eds. (Grunc and Stratton, Orlando, 1987), pp. 57-79. 8. I Chretien, I Bignon, A. Hindi, Eds., The Plrvm m Hoaltk and Diteoa (Dekker, New York, 1985). 9. R. R. CormeUy, ft- Spirtas, M- H. Meyen, C L. Petty, J. F. Fraumeni, J. Nod. Cancer Inst. 79, 31 (1987). 10. A. D. McDonald and ). C McDonald, in Asbrstos-Rrlsetd MoHgtmvey, K- H. Amman and J. Aisner, Eds. (Grunc and Stratton, Orlando, 1987), pp. 31-55. 11. H. A Anderson, R. Lilis, S. M. Daum, A. S. Fishbein, L J. Sdikoff, Am. N. Y Aend. So. 271, 311 (1976). 12. A Hirsh a al.. Am. J. Ind. Med. 3, 413 (1982). 13. R. Doll and J. Peto, in Asbesm-Retoied Motignmey, K. H. Amman and J.' Aisner, Eds. (Grunc and Stration, Orlando, 1987), pp. 81-96. 14. J. Bignon, P. Scbasben, L. DuMenza, H. Payan, Am. N.Y. Aend. Sri. 330, 455 (1979). 15. D. A Eddman, fir. J. Ind. Med. 45, 75 (1988); C K. Qian and J. B. L Gee, J. Ocmp. Mrd. 30, 23 (1988). 16. I. C Wagner, C A Sleggs. P. Marchand, fir. J. Ind. Med. 17, 260 (1960); G. K. Slui^Oaner, Am. N. Y. Aend. So. 132, 215 (1965). 17. I C Wagner. G. Bcnv, F. D. Pooler, fir Med. J. MS,603 (1982); M. J. Gardner, P. D. Winter, B. Fawn. C A Powell. Be J. Ind. Mrd. 43,726 (1986); A Churg, Otttt 93,621 (1988); A M. Linger and R. P. Nolan, in Naa-Drniyanmd Exponor to Mtnerol Fibers. J. Bignon. ]. ro, R- Saracci, Eds. (liuemaiional Agency for Research on Cancer, Lyon, 1989), pp. 330-335. 18. J. C Wagner a a/.. Am. Otmp. Hyg. 26, 423 (1982). ___ 19. M. C Jaunnd, J. Bignon, P. Scbasocn, J. Goni, Swum. Res. 14, 245 (1977); A Morgan and A Holmes, ibid. 39, 475 (1986). 20. A Gaudidiet a al., ibid. 32 (suppL), 213 (1988). 21. ). Dement, R. L. Harris, M. 1. Symons, C M. Shy, Am. J. Ind. Med. 4, 421 (1983); M. Finkdstem, Am. Rev. Respir. Dis. 129, 754 (1984); ). M. Hughes a a/., fir. J. Ind. Med. 66.161 (1987); A D. McDonald, J. S. Fry, A J. WooOey, J. C McDonald, Aid. 40,368 (1983); ibid., p. 361; C G. OMsonandC. Hagsredt, Aid. 42, 397 (198S). 22. A Churg, Osesi 93. 621 (1988). 23. P. Scbasocn.J. C McDonald, A. D. McDonald, B. Case, IL Hanky, fir. J. Ind. Med. 46, 180 (1989). 24. J. C McDonald a al , ibid. 43. 436 (1986). 25. A Churg, B. Wiggs, L Depaoli. B. Karripc, B. Stevens, Am. Rev. Respee. Dis. 130, 1042 (1984). 26. )- C Wagner, M. L Ncwhouse. B. Cotrin, C E. R- RoasiitT. D. M. Grifida, fir. J. Ind. Med. 45, 305 (1988). 27. M. L Ncwhouse aid 5. R. Sullivan, Aid. 46, 176 (1989). 28. G. Betty and M. L. NcwhouK, Aid. 40,1 (1983);H.F.ThaniM,I. L Bcniainan,P. C Ehtood, P. M. Su.niMi, fir. J. bed. MM. 39, 273 (1982). 29. ) M. G. Davis, in Pi.....of Mt Mm ana.) Cdhpiam m Deem Mmrnrmg Tedonifte ond Sasmggy (Mon kaemabonal haocininn. Mtannrshurg. 1985), pp. 25-35; F. PoqmdK. H. Ftnfaida. Nmn wA ylai 59,318 (1972); M. F. Scamon and C Wiensk,. J. Gw 48,797 (1972); M. C lamd, J. Fkury, C. Monchau, M. Ncfaa, J. Ijpini. J. Nod. Cmtee km. 79,797 (1987). 30. W. E. Srnah, D. D. Hubert. H. J. Sobd, E. Marquee, in Chan and Desmse. R_ Lena and J. M. Danas, E8, (Fahoaa, Pali Foot 1L, 1979), pp. 335-339. 31. ). M. a Diva a d,, fit. J. Exp. MM. 67,415 (1986);J. M. a Davit in Noso- Oempneinnnl Expmm a Mmrrol Fbers, J. Bignon, J. Mo, R. Sanaa, Eds (Imemjsicnal Agency far Research on Cancer, Lyon, 1989), pp. 33-45. 32. J. C Wagner, G. Berry, J. W. Skidmore, V. Timbre*, fr. J. Cam 29, 252 (1974). 33. R. Begin, S. Maae, M. Rofa-Ptazorraki. M. Bocair, G. Drapeau, in Asheaas Taxidty, G. L Fisher and M. A Gaiks, Eds. (Dekker, New York, 1987), pp. 87- 107. 34. A R. Brody. 1. M. Hi*. B. Adkins. R. W. O'Connor. Am. Rev. Rem*. Dis. 123, 670(1981). 35. B. T. Moesman and R. Begin, Eds., Efnts ofM'mesod Dusts m Cells (North Adanbc Treaty Organization, Advanced Science Institute Scries H, troL 30) (Sprurgu- Vcrkg, Berlin. 1989). 36. W. N. Rom. p. B. Bmennan. S. 1. Romani, A Canon, R. G. Crystal, Aai. Hrv. Remit. Dis. 136, 1424 (1987). . 37. 1. Lemaire, H. Bctadotsi, 5. Masae, 5. Grondm, Am. J. MW. 122, 205 (1986). 38. ). |. Momci a al.. J. dm. fares. 78, 61 (1986). 39 E W Gabndson tt ol. hASEll J 1. 3717 (1988 i. E M. Lavrek. A N A Jonwri, L. L. Moore, B I. Gcrwm, | F Laxhncr./ turu CWI Drv 24. 10*7 ( IV8S'i 40. K. Hansen and B. T. Mossman, Canter Res 47, 1681m 1987), H T Mruaman. K. Hansen, J. P. Marsh. M. E Brew. J. PctrusLa. m Lxpoiurr >, Mineral fibers. J Bignon. ). Po. R. Saracci. Eds. -(IntcmationaJ A^cncv tor Research on Cancer. Lvon, 1989), pp. 81-92. 41. B. T Moisman, >. P Marsh, M. A. Shares. Lstb Invru. 54, 204 (1986): M. A Sharos, I. P. Marsh, B. T Mossman, Envirw kn 44, 103 (1987). 42. S. A. Weioman and P. Graccffa, Anti Bnv*rwi Rtcpkys 228. 373 (1984). R Zaima, L Bonneau. M. C. Jaurand. J. Gurgnard. H. Pezcrat, Can J Chrm. 65, 2338 (1987). 43. S. A. Weioman and A. B. Weubcrg, Btothrm. J. 225, 259 (198$). 44. B. T. Mouman n at.. Cbm 89, )60 (1986). 45. B. T. Mosman n al., J. Fret RaJ. B*ot \M. 2, 335 (1986); B. T. Mossman rt al.. Am. Rev. Respit Dis., in press. 46. I. Bcrenbium, Cancer kn 1, 44 (194)). 47. M. Chamberlain and E. M.Tirmy, Muui Res. 43,159 (1977); W. G. Light and E. T. Wei, in TV in Vina Efietu of Mineral Dot, R. C. Brown. I. P. GormJcv, M. Chambertain, R. Davies, Eds. (Academic Press. London, 1980), pp. 139-145. 48. M. D. Shelby. Mmat Res. 204, 3 (1988). 49. S. L. Huang, ibid. 68. 26$ (1979). 50. B. Ron. C Tong. S. Tebny, G. M. Wilburn, Envirm. Res 31. 100 (1983); M. Oshinum, T. W. Hcsrcrtxrg, T. Tsuout, J. C Barrett, Cancer Res. 44, 5017 (1984);). A. DiPaoks, A. } DeMarmis, J. Doniger, Pharmacology 27, 65 (1983). 51. R. C Brown, A. Poole, G. T. A. Fleming, Cmtcer Lea. 18, 221 (1988); T. K. Hei, C R. Geard, R. S. Osmak, M. Tnvisano, Be. J. Cancer 52. 591 (1985); T. K. Hei. in Ejects of Mineral Dusts m Cells, B. T. Mossman and R. Begin, Eds. (North Adamic Treaty Organization, Advanced Science Institute Sena H, vo). 30) (Springer-Veriag. Berlin, 1989). 52. T. W. Hesterbcrg and J. C Barrett, Cancer Res 44, 2170 (1984); Y. P. Lu, G Lasne, R. Lowy, I. Chouroulinkov. Mutagenesis 3, 355 (1988); S. O. Mikaben, E RjvedaL, T. Sanner, Canmogenesu 9, 891 (1988). 53. L. D. Pakkar, B. M. Mom, D. L. Coffin, Environ. Res. 46, 142 (1988); L. D. Pakkar,). F. Evrc, B. M. Most, D. L. Coffin, CimMfnieu 8, 553 (1987). 54. ). C Wagner W., Be. J. Cancer 51. 727 (1985). 55. M. J. Pacaour, J. Bignon, M. C Jaurand, Cnnwogcntsu 6, 523 (1985). 56. J. D. Appel, T. M. Fasy, D. S. Kohtz, J. D. Koho. E. M. Johnson, Peoc. Nod. Acad. Sci. U.S. A. 85, 7670 (1988); G. R. Dubes and L R. Mack, h Vara Cell. Dev. Biel. 24, 175 (1988). 57. M. G Jaurand, L. JCheuang, L. Magnc, /. Bignon, Muuc. Res. 169, 141 (1986); Sc. Ebovte <r al., in Ejects of Mineral Dusts on Cells, B. T. Mouman and R. Begin, Eds. (North Adantic Treaty Organization, Advanced Science Institute Sena) (Spiingci'Veriag, Berlin, in press). 58. J. F. Lechner n at., Prac. Natl. Acad. Sci. U.S. A. 82, 3884 (1985). 59. K.T. Rajan, J. G Wagner, P. H. Evans, Name 238, 346 (1973); P. A. MoaJli, J. L. McDonald, L A. Goodgbck, A. B. Kane, Am. J. Pathol. 128, 426 (1987). 60. N. S. Wang ft al., AM. J. Pathol. 126. 343 (1987). 61. T. W. Hocerbcrg and J. G Barrett, Carcinogenesis 6, 473 (1985). 62. N. G Popocu, A. P. Chahinian, J. A. DiPaolo, Cancer Res. 48, 142 (1988); Z. Gibas n al.. Cancer Genet. Cytofenet. 20, 190 (1986); M. Tiainen, L. Tammikhto, JC Mamon, 5. Knuuola, ibid. 33, 251 (1988). 63. B. Gametal.. Canter Res. 47,6180 (1987); M. A. VcmeJ, A. Hugemajer, M. J. Bouts, T. H. van der Kwast, H. C Haagsteden, Oncogene 2, 601 (1988). 64. A. Haugen et al., Ira. J. Comer 30, 265 (1982). 65. Y. JCodama, G J. Boreiko, S. G Maness, T. W. Hesterberg, in preparation. 66. D. G Topping and P. Nctteshcim, J. Nad. Cancer last. 65, 627 (1980). 67. B. T. Moasnan, G. S. Cameron, L. Yom, in Canter: A Comprehensive Survey: Cancer of ffir Respteauey Tract, Predispomtg Factors, M. J. Moss, D. G. ICau/marv, J. M. Siegfried, V. E Sceek, S. Ncsnow, Eds. (Raven, New York, 1985), pp. 217-230. 68. A. M. Sfffco and B. T. Mouman, Cancer Res. 49, 2743 (1989). 69. J. P. Manh and B. T. Mouman, ibid. 48, 709 (1988). 70. Y. Nkhasdsa, Science 233, 305 (J986). 71. A. M. Sedas, M. Cabot. B. T. Mouman, in preparation. 72. M. Pcderiact, J. P. Manh, B. T. Mouman, in preparation. 73. E G Hammond, I. J. Sdikoff, H. Sodman, Am. N.Y. Acad. Sci. 330, 473 (1979); J. G McDonald, Chest 78 (tuppL), 374 (1980); G. Bcny, M. L Newhouse, P. Antonis, Be. J. ind. Med. 42, 12 (1985). 74. D. McFaddm, J. L Wright, B. Wiggs, A. Churg, Aw. Rev. Rasp*. Dis. 133, 372 (1986); D. McFadden, J. L. Wright, B. Wiggs, A. Chuig, Aw. f. Pathol. 123, 9S (19*6). 75. J. H. Jackaon et d.% J. dm. Invest. 80, 1090 (1987). 76. P. Gracefa and $. A. Weioman, Airib. Biochem. Bitphyi. 257, 481 (1987). 77. A. Eattntwt, B. T. Mouman, E Bicanack, Cmcer Res. 43, 1251 (1983). 78. I. SdifcoC E G Huunond, H. Sodman, Cmcer 46,2736 (1980). 79. J. B. L Gee and A. Boidtuys, N. Engl. J. Med. 285,1317 (1971). 80. In 1971, the Occupational Safety and Hcakh Adnunotracian (OSHA) passed kgisiabon restricting airborne asbestos in the workpber to five Sben greater than 5 inn in kngsh per cubic centimeter air over an 8-hour rime-woghred average. In 1986, this uindard was reduced to 02 fibers per cubic ccnriincttr air. 81. J. M. G. Davis and J. G McDcmald, Br J. hrd. Med. 4S, 505 (1988). 82. A. Churg and J. L Wright, in Mw Qmpmmtl Fvpanvt ra Mineral Fibres, J. Bignon, J. Pens, R- Saracds, Eds. (Inaemabonal Agency Cor Research on Cancer, Lyon, 1989), pp. 314-318. 83. G. J. Burden and S. A. M. T. Jatfrey, Am. Oenp. Hyg. 30,185 (1986); M. Com, in Praeetdmgs ofthe Syrapatmm m Heahb AtpatU ofExpoatee to Asbestos w Bnildmgt, J. D. Spengkr, H Odaynik, J. F. McCarthy, H. Lee, Eds. (Harvard Univ..Press, Cambridge, MA, in pres); D. G. Massey and G. Foumaer*Masacy, Hamah Med. J. 46, 153 (1987); R. N. Sawyer, A. N. Rohl, A. M. Lmgcr. Envirm. Res 36, 46 300 SCIENCE, VOL. 247 369054 0349 nwsv. 84 V KcnJcrsoo and P. Emcrlmc. J Shit! Caueer h,u 79. 31 (1987] $5 J R. Hughe* and H Weill Am. fry hespir Du 133, 5(1986) 86 5 Gmlier e< ai . Arrk. txrtr,wt Health 42, 303 {1987) 87. NjikrfaaJ Research Council, Comnunec on Nonoccupauonzi Health Risks of /febesutorm Ftbcn. .Asbrm^wm Hbers: Nometupotienal Health Xtsks (NaricituJ Acad cniv Press, Washington, DC. 1984); Report to the U.S. Consumer Product Safety Commission by the Chronic Hazard Advisory Panel on Asbestos (Consumer PrtxJoa Safety Commission. Washington, DC 1983); Airborne Atbettos Health Assessment Update (Environmental Protection Agency, Washington, DC, 1986); E. 0 Acheson and M. J. Gardner, Asbestos: The Control Limn joe Asbestos (Her Majesty's Stationery Office, London, 1983); Report on marten of health and safety arising from the use of asbestos in Ontario (Ontario Royal Commission. Ontario Ministry of the Attorney General Toronto, 1984); R. DoU and |. Pcto, Asbestos: tijfrnt <vr Health of Expmiirt to 4thn/r*> (Her Majesty's Stationery Office. London. 1985) 88- M Com, K. E. Cnimp. D McFcc. R Lee. in preparauon 89. Randle Columbus Division. Pncc Associates. Alliance Technologies Corporation. Energy Technology Consultants and Midwest Research Institute, Annum 4i6r*iti im l\,hli( (tPA Conn No 6b-02-*2<M, draft report for the Exposure Evaluation Division, 1989). 90. H. Weill and J. M. Hughes. Am Kn> Pnbtu Health 7, 171 (1986) 91 We thank L. Sabcns. V. Kcilehcr. and M. Bergeron. Research in the laboratory of B.T.M. has been supported by grants trom the National Cancer Institute (R01 CA33501). National institute of Environmental Health Sciences (R0! ES03878). National Heart, Lung, and Blood Insnrute (ROl HL39469 and SCOR grant 14212). and American Cancer Society (BC 415). We also thank H. C W. Skinner for her hdpfuJ comments. Priming and Human Memory Systems Endel Tulving and Daniel L. Schacter Priming is a nonconscious form of.human memory, which is concerned with perceptual identification of words and objects and which has only recently been recognized as separate from other forms of memory or memory systems. It is currently under intense experimen tal scrutiny. Evidence is converging for the proposition that priming is an expression of a perceptual representa tion system that operates at a prc-semantic level; it emerges early in development, and access to it lacks the kind of flexibility characteristic of other cognitive memo ry systems. Conceptual priming, however, seems to be based on the operations of semantic memory. and memory, one that is not procedural, semantic, or episodic. It has come to be known as priming (2). Its function is to improve identification of perceptual objects. Priming is a type of implicit memory; it does not involve explicit or conscious recollection of any previous experiences. It has affinities to both procedural and seman tic memory. Priming resembles procedural memory in chat it enhances perceptual skills. It also resembles semantic memory in that it involves cognitive representations of the world and expresses itself in cognition rather than behavior. The prototypical priming experiment consists of two stages, in the first (study) stage, the subject is presented with a stimulus object (target). Target stimuli may comprise words, line drawings of objects, drawings of faces, and the like. In the second (test) stage, which may follow the first after an interval chat can vary from seconds to months, the subject is given reduced perceptual informa tion about the object and asked to name or categorize it. Reduced emory was traditionally thought to be a unitary cues may consist of initial letters or graphcmic fragments of words, Mfaculty of the mind. Recently, however, many researchers partially obliterated words or figures, originally presented faces in a have adopted the hvpochcsis that memory consists of a mote highly schematized form, or tachistoscopic presentation of number of systems and subsystems with different operating cshtiamraucli. Priming is said to have been demonstrated if the probability teristics. The problem ofwhat these systems and their propertoiefs tahre, identification of the previously encountered targets is in and how they are related to one another, now occupies the center creased, or the latency of the identification response is reduced, in stage in research on memory. comparison with similar measures for nonstudied control items. The One broad, as yet tentative, organizational scheme distinguishes difference between performance on the target items and the nonstu procedural, semantic, and episodic memory (I). Procedural memory died items provides a measure of the magnitude of the priming underlies changes in skillful performance and appropriate respond effect. ing to stimuli; semantic memory has to do with acquisition and use Although priming and ocher kinds of implicit memory have been of factual knowledge in the broadest sense; and episodic memory reported from time to time, systematic attempts to explore it began enables people to remember personally experienced events. The about 10 years ago (3). One ofthe triggers for the study of priming domain ofprocedural memory is behavior, whereas that ofsemantic turned out to be experiments by Warrington and Weiskrantz (4) and episodic memory is cognition or thought. Cognitive memory showing that densely amnesic patients, who were severely impaired systems have the capability of modeling the external world--that is, in their ability to remember recently seen information, exhibited of storing representations of objects, events, and relations among near-normal learning when they were tested by methods that tapped them--whereas procedural memory does not have this capability. what we now know is priming. A second stimulus for the study of Evidence is accumulating about yet another category of learning priming lay in research concerned with the nature of and access to lexical representations (5). A third source of influence, was the *E. TuMng ii a univeaicy .professor at the University of Toronto, Toronto, Canada, MSS 1A1. D. L. Schacter is a professor of paychotogy at the University of Arizona, Tucson, AZ. $5721. growing interest in the classification of memory into distinctive categories such as episodic and semantic memory (6) and procedural and declarative memory (7). 19 JANUARY 1990 articles 301 369054 0350