5
Genotoxicity and Carcinogenicity

The carcinogenic potential of beryllium and beryllium compounds has been assessed by various agencies in the last decade. The International Agency for Research on Cancer (IARC 1993) classifies beryllium and beryllium compounds as carcinogenic in humans, the U.S. Environmental Protection Agency (EPA 1998a,b) considers them probable human carcinogens, and the National Toxicology Program (NTP 1999, 2005) lists them as reasonably expected to be carcinogens. As noted in Chapter 1, EPA has performed a dose-response analysis of the cancer data to estimate an inhalation unit risk of 2.4 × 10−3 per µg/m3. This chapter examines the literature used in the previous assessments, more recent reviews, and relevant new studies. Information on the genotoxic potential of beryllium and beryllium compounds is presented first, and then the literature on carcinogenicity, including the epidemiologic literature and animal bioassays, is reviewed.

GENOTOXICITY

Compounds of beryllium have tested positive in nearly 50% of the genotoxicity studies conducted without exogenous metabolic activation but were nongenotoxic in most bacterial tests. Beryllium chloride, beryllium nitrate, beryllium sulfate, and beryllium oxide have been shown to be nongenotoxic in the Ames plate incorporation assay and assays with Escherichia coli pol A, E. coli WP-2 uvr A, and Saccharomyces cerevisiae (Table 5-1) (reviewed in EPA 1998a,b; ATSDR 2002; Gordon and Bowser 2003). Beryllium sulfate also did not induce unscheduled DNA synthesis in primary rat hepatocytes (Williams et al. 1982, 1989), was not mutagenic when injected intraperitoneally into adult mice in a Salmonella typhimurium host-mediated assay (Simmon et al. 1979), and failed to increase the incidence of micronucleated polychromatic erythro-



The National Academies | 500 Fifth St. N.W. | Washington, D.C. 20001
Copyright © National Academy of Sciences. All rights reserved.
Terms of Use and Privacy Statement



Below are the first 10 and last 10 pages of uncorrected machine-read text (when available) of this chapter, followed by the top 30 algorithmically extracted key phrases from the chapter as a whole.
Intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text on the opening pages of each chapter. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.

Do not use for reproduction, copying, pasting, or reading; exclusively for search engines.

OCR for page 106
5 Genotoxicity and Carcinogenicity The carcinogenic potential of beryllium and beryllium compounds has been assessed by various agencies in the last decade. The International Agency for Research on Cancer (IARC 1993) classifies beryllium and beryllium com- pounds as carcinogenic in humans, the U.S. Environmental Protection Agency (EPA 1998a,b) considers them probable human carcinogens, and the National Toxicology Program (NTP 1999, 2005) lists them as reasonably expected to be carcinogens. As noted in Chapter 1, EPA has performed a dose-response analy- sis of the cancer data to estimate an inhalation unit risk of 2.4 ×10-3 per µg/m3. This chapter examines the literature used in the previous assessments, more re- cent reviews, and relevant new studies. Information on the genotoxic potential of beryllium and beryllium compounds is presented first, and then the literature on carcinogenicity, including the epidemiologic literature and animal bioassays, is reviewed. GENOTOXICITY Compounds of beryllium have tested positive in nearly 50% of the genotoxicity studies conducted without exogenous metabolic activation but were nongenotoxic in most bacterial tests. Beryllium chloride, beryllium nitrate, beryllium sulfate, and beryllium oxide have been shown to be nongenotoxic in the Ames plate incorporation assay and assays with Escherichia coli pol A, E. coli WP-2 uvr A, and Saccharomyces cerevisiae (Table 5-1) (reviewed in EPA 1998a,b; ATSDR 2002; Gordon and Bowser 2003). Beryllium sulfate also did not induce unscheduled DNA synthesis in primary rat hepatocytes (Williams et al. 1982, 1989), was not mutagenic when injected intraperitoneally into adult mice in a Salmonella typhimurium host-mediated assay (Simmon et al. 1979), and failed to increase the incidence of micronucleated polychromatic erythro- 106

OCR for page 106
TABLE 5-1 Genotoxicity Studies of Beryllium Compounds Without Assay or End Point Species Compound With Activation Activation Reference In vitro Plate incorporation assay Salmonella typhimurium Negative Negative Rosenkranz and Poirier 1979; BeSO4 Simmon 1979a; Simmon et al. 1979; Dunkel et al. 1984; Arlauskas et al. 1985; Ashby et al. 1990 S. typhimurium Negative Negative Arlauskas et al. 1985; Kuroda et Be(NO3)2 al. 1991 S. typhimurium BeO Negative Negative Kuroda et al. 1991 S. typhimurium Negative Negative Kuroda et al. 1991 BeCl2 Eschericia coli WP-2 uvrA Negative Negative Dunkel et al. 1984 BeSO4 Rec assay Bacillus subtilis — Positive Kada et al. 1980; Kanematsu et BeSO4 al. 1980 B. subtilis — Positive Kuroda et al. 1991 BeCl2 B. subtilis — Negative Nishioka 1975 BeCl2 B. subtilis — Positive Kuroda et al. 1991 Be(N03)2 B. subtilis BeO — Negative Kuroda et al. 1991 E. coli — Positive Dylevoi 1990 BeSO4 DNA modification E. coli pol A+/A− — Negative Rosenkranz and Poirier 1979 BeSO4 Bioluminescence test Photobacterium fischeri — Positive Ulitzur and Barak 1988 BeCl2 Recombogenic activity Saccharomyces cerevisiae Negative Negative Simmon 1979b BeSO4 Host-mediated assay S. cerevisiae — Negative Simmon et al. 1979 BeSO4 S. typhimurium — Negative Simmon et al. 1979 BeSO4 (Continued) 107

OCR for page 106
TABLE 5-1 Genotoxicity Studies of Beryllium Compounds 108 Without Assay or End Point Species Compound With Activation Activation Reference Chromosomal aberration Swine lymphocytes — Positive Vegni-Talluri and Guiggiani 1967 BeCl2 Syrian hamster embryo cells — Positive Larramendy et al. 1981 BeSO4 Human lymphocytes — Positive Larramendy et al. 1981 BeSO4 Chinese hamster ovary cells — Negative Brooks et al. 1989 BeSO4 Cytogenetic assay Chinese hamster lung cells Negative Negative Ashby et al. 1990 BeSO4 Sister-chromatid Chinese hamster V79 cells — Positive Kuroda et al. 1991 BeCl2 exchange assay Chinese hamster V79 cells — Positive Kuroda et al. 1991 Be(NO3)2 Chinese hamster V79 cells BeO — Negative Kuroda et al. 1991 Syrian hamster embryo cells — Positive Larramendy et al. 1981 BeSO4 Human lymphocytes — Positive Larramendy et al. 1981 BeSO4 Human lymphocytes — Negative Andersen 1983 BeSO4 — Negative Andersen 1983 Mouse macrophage P388D1 BeSO4 cells DNA repair Rat hepatocytes — Negative Williams et al. 1982, 1989 BeSO4 Transformation assay Syrian hamster embryo cells — Positive DiPaolo and Casto 1979 BeSO4 Rat respiratory epithelial Rocket-exhaust — Mixed Steele et al. 1989 cells residue Rat respiratory epithelial BeO (low-fired) — Positive Steele et al. 1989 cells Rat respiratory epithelial BeO (high-fired) — Mixed Steele et al. 1989 cells BALB/c-3T3 cells — Positive Keshava et al. 2001 BeSO4

OCR for page 106
DNA damage Rat respiratory epithelial Rocket- exhaust — Mixed Steele et al. 1989 cells residue Rat respiratory epithelial BeO (low-fired) — Positive Steele et al. 1989 cells Rat respiratory epithelial BeO (high-fired) — Mixed Steele et al. 1989 cells Mutation of HGPRT gene Chinese hamster ovary — Positive Hsie et al. 1979 BeSO4 K1-BH4 cells Chinese hamster V79 cells — Positive Miyaki et al. 1979 BeCl2 Mutation of lacI gene E. coli — Positive Zakour and Glickman 1984 BeCl2 Mutation of K-ras gene Rat lung tumors Be — Weak positive Nickell-Brady et al. 1994 Mutation of p53 Rat lung tumors Be — Negative Nickell-Brady et al. 1994 gene Mutation of c-raf-1 gene Rat lung tumors Be — Negative Nickell-Brady et al. 1994 In vivo Transformation assay Syrian hamsters (embryo — Positive DiPaolo and Casto 1979 BeSO4 cells evaluated after maternal exposure) Micronucleus assay CBA mice — Negative Ashby et al. 1990 BeSO4 Abbreviations: Be, beryllium; BeCl2, beryllium chloride; Be(NO3)2, beryllium nitrate; BeO, beryllium oxide; BeSO4, beryllium sulfate. Source: Adapted from ATSDR 2002. 109

OCR for page 106
110 Managing Health Effects of Beryllium Exposure cytes in bone marrow (Ashby et al.1990). Lung tumors in F344/N rats treated with beryllium sulfate did not have mutations of the p53 or c-raf-1 gene, but weak mutations were detected in the K-ras gene (Nickell-Brady et al. 1994). Positive genotoxic results have been reported with beryllium sulfate in the Bacillus subtilis rec assay (Kada et al. 1980; Kanematsu et al. 1980) and the E. coli rec assay (Dylevoi 1990), with beryllium nitrate in the B. subtilis rec assay (Kuroda et al. 1991), and with beryllium chloride in the B. subtilis rec assay with spores (Kuroda et al. 1991), the E. coli forward-mutation assay (Zakour and Glickman 1984), and the Photobacterium fischeri assay (Ulitzur and Barak 1988). Gene mutations have been observed in mammalian cells cultured with beryllium chloride (Vegni-Talluri and Guiggiani 1967; Hsie et al. 1979; Miyaki et al. 1979) and beryllium sulfate (Larramendy et al. 1981; Brooks et al. 1989); beryllium nitrate has resulted in clastogenic alterations (Kuroda et al. 1991). Overall, mutation and chromosomal-aberration assays of beryllium compounds have yielded somewhat contradictory results. Although the bacterial assays have been largely negative, the mammalian test systems exposed to beryllium com- pounds have shown evidence of mutations, chromosomal aberrations, and cell transformations. However, results of the mammalian test systems have also been inconsistent. The chemical form of beryllium does not appear to determine genotoxic results (Table 5-1). Further studies would confirm the mutagenic or genotoxic properties of the various beryllium compounds. CARCINOGENICITY Epidemiologic Studies Several studies and reviews on cancer in relation to human exposure to be- ryllium are available. Two worker cohorts involved in beryllium extraction, pro- duction, and fabrication have been extensively studied and have been the pri- mary basis of conclusions drawn on cancer in humans. One cohort was in Lorain, Ohio, and the other in Reading, Pennsylvania. The original study (Mancuso 1979) reported a lung-cancer standardized mortality ratio (SMR) in the two plants combined of 1.42 (95% confidence interval [CI], 1.1-1.8). The study involved 1,222 workers at the Ohio plant and 2,044 workers at the Penn- sylvania plant who had been employed for at least 3 months during 1942-1948. No analysis by job title or by exposure category was performed, and the excess– lung-cancer finding was limited to workers who were employed for less than 5 years. Worker exposure was often at high concentration. For example, a study at the Ohio plant in 1947-1948 by the U.S. Atomic Energy Commission measured beryllium at concentrations ranging from 411 µg/m3 in the mixing area to 43,300 µg/m3 in the breathing zone of alloy operators (Zielinski 1961). Control limits at U.S. plants were introduced in 1949 (Wagoner et al. 1980). Mancuso (1980) reanalyzed the same two cohorts but expanded the period of employment of the study cohorts to 1937-1948 and used workers at a rayon plant for com-

OCR for page 106
111 Genotoxicity and Carcinogenicity parison purposes. The comparison between the two types of industrial workers found a significant relative lung-cancer SMR of 1.40 for the beryllium-worker cohort. Wagoner et al. (1980) expanded the cohort in the Pennsylvania plant to in- clude workers employed during 1941-1967. The group of 3,055 workers was found to have a lung-cancer SMR of 1.25 (95% CI, 0.9-1.7). When the analysis was adjusted for latency, there was a significant SMR of 1.68 in the group that had a latency of 25 years or longer, but there was no relationship with duration of employment. The results of a 1968 medical survey of smoking histories of the workers showed that smoking increased the cancer risk in beryllium workers by 14%. However, if the working population’s risk is compared with lung-cancer mortality in the county where the plant was instead of using the U.S. rates, the SMR is underestimated by 19% (Wagoner et al. 1980). The National Institute for Occupational Safety and Health (NIOSH) con- ducted a retrospective cohort mortality study of seven beryllium-production fa- cilities that included the Pennsylvania and Ohio cohorts previously studied by Mancuso and Wagoner. Ward et al. (1992) developed a cohort of 9,225 male workers who had worked for at least 2 days during 1940-1969 and were fol- lowed through 1988. The lung-cancer SMR was 1.26 (95% CI, 1.12-1.42) on the basis of 280 lung-cancer deaths. The researchers also observed an SMR for nonmalignant respiratory disease of 1.48 (95% CI, 1.21-1.80). Ward et al. (1992) reported that SMR increased with latent period: there was a significant SMR of 1.46 for a latent period greater than 30 years in the workers at the com- bined seven plants. The lung-cancer SMR was a significant 1.42 in those hired before 1950 and was less than l in those hired during 1960 through 1969. IARC (1993) has provided a detailed description and critique of the cohort studies. It pointed out that the risk of lung cancer was consistently higher in plants in which there was excess mortality from nonmalignant respiratory dis- ease. IARC also concluded that the association between lung-cancer risk and beryllium exposure did not appear to be confounded by smoking. A second line of investigation is embodied in the Beryllium Case Registry, which was established in 1952 to follow the clinical aspects of and complica- tions in beryllium-related diseases, including chronic beryllium disease (CBD) and acute beryllium-related pneumonitis. The data were analyzed first by Infante et al. (1980) and more recently by Steenland and Ward (1991). In the Steenland and Ward study, the cohort consisted of 689 people who entered the registry during 1952-1980 and were followed through 1988. The researchers reported an SMR of 2.00 (95% CI, 1.33-2.89) on the basis of 28 observed lung-cancer deaths. The lung-cancer SMR was greater in people who had acute beryllium pneumonitis (SMR, 2.32) than in those who had CBD (SMR, 1.57); the former was statistically significant. IARC (1993) concluded that the studies of cases in the Beryllium Case Registry provided indirect evidence that beryllium, rather than smoking, explained the increase in lung cancer on the grounds of the as- sumption that people with acute pneumonitis were unlikely to smoke more than workers with CBD.

OCR for page 106
112 Managing Health Effects of Beryllium Exposure With respect to other cancer end points, Carpenter et al. (1988) conducted a nested case-control study of cancers of the central nervous system in workers at facilities in Oak Ridge, Tennessee. There were 72 male and 17 female deaths due to central-nervous-system cancer. Using job titles, the investigators consid- ered the potential exposure to each of 26 chemicals, including beryllium. There was a weak association with exposure to beryllium (odds ratio, 1.5; 95% CI, 0.6- 3.9). The authors concluded that their study did not support the hypothesis that occupational exposure to the chemicals they studied increased the risk of cancer of the central nervous system appreciably. IARC (1993) noted that the risk of cancer of the central nervous system increased with longer duration of employ- ment in jobs with greater exposure to beryllium. On the basis of the studies described above, IARC concluded that there is sufficient evidence of the carcinogenicity of beryllium and beryllium compounds in humans. That conclusion was based on the cohort studies, which showed • A large number of lung-cancer cases with a stable estimate of the SMR. • Consistency among locations. • A greater excess of lung cancer among workers hired before 1950, when exposure was greater. • The highest lung-cancer risk at the plant that had the greatest propor- tion of acute beryllium-pneumonitis cases in the Beryllium Case Registry. • High lung-cancer risks at plants with the greatest risk of pneumoconio- sis and other respiratory diseases. • A greater lung-cancer risk in the Beryllium Case Registry cohort. • Increasing risk with longer latency. IARC pointed out the following limitations: • Absence of individual exposure measurements. • Relatively low excess lung-cancer risk. • Absence of any mention of exposure to other lung carcinogens in the workplace. A series of letters and papers issued after the IARC report raised concerns about and objections to the basis of its conclusions. Some raised concerns about the IARC procedures, the information available to the IARC working group, and possible conflicts of interest (Kotin 1994a,b; Vainio and Kleihues 1994). Others questioned the validity of the Ward et al. study. Questions were raised about the dataset used to estimate background lung-cancer rates, how to combine data from multiple plants, and how to adjust for cigarette-smoking (MacMahon 1994; Levy et al. 2002). Levy et al. (2002) have reported that alternative adjustments and comparisons to address those issues left no statistical association between beryllium exposure in the workers and lung cancer.

OCR for page 106
113 Genotoxicity and Carcinogenicity Since the IARC evaluation in 1993, there have been two additional stud- ies. Sanderson et al. (2001b) conducted a nested case-control study of plant workers at the Reading, Pennsylvania, facility. The cohort of 3,569 male work- ers was the same as the cohort in the 1992 Ward et al. study. The lung-cancer cases numbered 142 on the basis of a followup of the cohort through 1992, and each was age- and race-matched to five controls. In addition to assessment of beryllium exposure, the potential for confounding by smoking was evaluated. The cases had lower lifetime exposure to beryllium. However, when a 10-year lag and a 20-year lag were applied, the exposure metrics were higher in cases. Furthermore, significant positive trends in the log of exposure metrics were ob- served, and the authors concluded that smoking did not confound the exposure- response analysis. Methodologic concerns have been raised about the Sanderson et al. study (Deubner et al. 2001c, 2007; Levy et al. 2007). The principal objection con- cerned the potential for confounding by birth year or age at hire, neither of which had been explicitly considered by Sanderson et al. In response, Schu- bauer-Berigan et al. (2008) reanalyzed the Sanderson et al. study and investi- gated potential confounding and effect modification by birth year (which was highly correlated with age at hire). They also assessed the sensitivity of the ex- posure-risk association to a small but potentially important methodologic choice: using a small value as a “start” when taking the logarithm of exposure metrics to avoid taking the logarithm of zero. The Schubauer-Berigan et al. re- analysis confirmed a significant association between beryllium exposure and lung-cancer risk, although the exposure metric and time lag that revealed the strongest evidence differed from those in the original study. Sanderson et al. reported an association between cumulative exposure with a latency of 20 years, whereas Schubauer-Berigan et al. found the beryllium–lung-cancer association when using average exposure with a 10-year latency. Changing the “start” value that was used in lagging exposure metrics did not substantially affect the results. Brown et al. (2004) conducted a nested case-control study of lung cancer in a cohort of plutonium-exposed workers at Rocky Flats. The main focus of the study was on the risk posed by plutonium, but an attempt was also made to as- sess risks associated with asbestos, hexavalent chromium, nickel, and beryllium. The 120 cases of primary lung cancer identified from death certificates and a tumor registry were matched to 720 controls. There was evidence of increased lung-cancer risk with increasing plutonium dose. Beryllium exposure was esti- mated with a job-exposure matrix, but no details were provided in the paper. The authors reported that cumulative exposure to beryllium was not “signifi- cantly associated” with lung-cancer risk, but no details or results were presented. Their paper provides only limited evidence bearing on the question of beryllium carcinogenicity in that no quantitative results were presented. NIOSH is conducting a new retrospective cohort study of the principal U.S. beryllium-production facilities, including a detailed exposure reconstruc- tion. The study should provide considerably stronger findings on human lung- cancer risk than the existing studies.

OCR for page 106
114 Managing Health Effects of Beryllium Exposure Animal Studies This section focuses on studies of inhalation exposure to beryllium and its compounds and the later development of neoplasms in laboratory animals (see Table 5-2). Lung neoplasms have been found in rats and monkeys exposed to beryllium compounds by inhalation. Albino Sherman and Wistar rats (male and female) were exposed by inha- lation to an aqueous aerosol of beryllium sulfate tetrahydrate (which contained beryllium at 35.7 µg/m3) 8 hours/day, 5.5 days/week, for 6 months (Schepers et al. 1957). The rats were observed for 18 months after exposure. Lung neoplasms (18 adenomas, five squamous-cell carcinomas, 11 papillary adenocarcinomas, and seven alveolar-cell adenocarcinomas) were observed in the treated rats but not in the control rats. A study by Vorwald and Reeves (1959) reported the development of lung neoplasms in Sherman rats (number and sex not reported) exposed by inhalation to beryllium sulfate at 6 and 54.7 µg/m3 6 hours/day, 5 days/week, for up to 18 months. The neoplasms observed were primarily adenomas and squamous-cell cancers. A study by Reeves et al. (1967) exposed male and female Sprague-Dawley rats to beryllium sulfate at 34.25 µg/m3 7 hours/day, 5 days/week. The mean particle size of the beryllium sulfate aerosol was 0.118 µm. Exposure lasted up to 72 weeks. After 13 months of exposure, all the exposed rats had developed alveolar adenocarcinomas; the control rats had no lung neoplasms. The neopla- sia was preceded by a proliferative response that progressed from hyperplasia to neoplasia. In another study in which particle size was calibrated, Charles River CD rats were exposed to beryllium sulfate at 35.16 µg /m3, with a mean particle size of 0.21 µm, 35 hours/week (Reeves and Deitch 1969). Exposure duration was 800, 1,600, and 2,400 hours. The lung- tumor incidence in young rats exposed for 3 months (86%, 19 of 22 rats) was the same as that in older rats exposed for 18 months (87%, 13 of 15 rats). However, older rats that were exposed to beryl- lium sulfate for 3 months had fewer lung neoplasms than rats that were exposed when they were younger. The pulmonary neoplasms were typically observed after a latency of 9 months. Preneoplastic lesions were described as epithelial hyperplasia at 1 month, metaplasia at 5-6 months, and anaplasia at 7-8 months. Male and female rhesus monkeys (Macaca mulatta) were exposed to be- ryllium sulfate at 35 µg/m3 6 hours/day, 5 days/week (Vorwald 1968). Exposure was often interrupted for considerable periods to prevent the monkeys from de- veloping acute beryllium pneumonitis (four monkeys died of acute chemical pneumonitis during the first 2 months of the study). The longest exposure was for a total of 4,070 hours, and most of the exposure periods occurred during the first 4.5 years of the study. A 6-month exposure occurred 2.5 years after the ini- tial 4.5-year exposure period. The authors reported that pulmonary anaplastic carcinomas (adenomatous and epidermoid patterns) were observed in eight of 12

OCR for page 106
TABLE 5-2 Inhalation-Carcinogenicity Studies of Beryllium Reference Species Route Dose Findings Acute exposure Sanders et al. 1978 Rat Inhalation 1.0-91 µg of Be from BeO (single alveolar Alveolar half-life of Be in lungs, 325 days; 1 of deposition); particle size, 1.10 ± 0.17 µm 184 rats had lung tumors after 2 years (GSD, 2.17 ± 0.17 µm) Groth et al. 1980 Rat Intratracheal Be at 0.5, 2.5 mg/m3 as passivated metal Lung adenomas, adenocarcinomas found in 2 of (Be-Cr), alloys (Al, Cu, Ni, Cu-Co); 21, 9 of 16 treated with Be; 7 of 20. 9 of 26 particle size, 1-2 µm treated with Be-Cr; 1 of 21, 4 of 24 treated with Be-Al, respectively; no tumors with other alloys Litvinov et al. 1983 Rat Intratracheal BeO at 0.036, 0.36, 3.6, 18 mg/kg (low- Malignant epithelial lung tumors after low-fired and high-fired) BeO, 0 of 76, 0 of 84, 2 of 77, 2 of 103, respectively; after high-fired BeO, 3 of 69, 7 of 81, 18 of 79, 8 of 26, respectively Nickell-Brady Rat Inhalation Be at 410, 500, 830, 980 mg/m3 (single 64% of rats developed lung tumors (primarily et al. 1994 exposure; lung burdens, 110, 40, 360, 430 adenocarcinomas) after 14 months µg); particle size, 1.4 µm (GSD, 1.9 µm) Short-term and subchronic exposure Vorwald and Reeves Rat Intratracheal 4.5 mg of Be as BeO, 0.1071 mg of Be as Lung tumors began to appear after 8 months; 1959 percentage of rats affected not specified BeSO4 (three injections over 3 weeks) Ishinishi et al. 1980 Rat Intratracheal 1 mg of BeO (low-fired) (once a week for 1 adenocarcinoma, 1 squamous-cell carcinoma, 15 weeks) 4 adenomas Chronic exposure Dutra et al. 1951 Rabbit Inhalation BeO at 1, 6, 30 µg/L (5 hours/day, 5 6 of 9 rabbits developed osteosarcomas after days/week, 9-13 months); particle size, 1 year 0.285 µm (mean), 0.11-1.25 µm (range) (Continued) 115

OCR for page 106
TABLE 5-2 Continued 116 Reference Species Route Dose Findings Schepers et al. 1957 Rat Inhalation Lung-cancer rates higher in exposed rats than in Be at 35.7 µg/m3 as BeSO4 (8 hours/day, 5.5 days/week, up to 6 months) controls Vorwald and Reeves Rat Inhalation Lung tumors began to appear after 9 months; Be at 0.0547 mg/m3 as BeSO4, at 0.006 1959 mg/m3 as BeO (6 hours/day, 5 days/week, percentage of rats affected not specified, but later various durations up to 18 months) report (Vorwald et al. 1966) described incidence of cancer as “almost 100%” in “large number” of surviving rats Reeves et al. 1967 Rat Inhalation All rats developed lung tumors Be at 34.25 µg/m3 (mean) as BeSO4 (7 hours/day, 5 days/week, 72 weeks); (adenocarcinomas) by 13 months particle size, 0.118 µm Vorwald 1968 Monkey Inhalation 8 of 12 monkeys had pulmonary anaplastic BeSO4 at 35 µg/m3 (6 hours/day, 5 days/week, various interruptions, variable carcinomas (adenomatous, epidermoid patterns); durations up to 4,070 hours) first tumor observed after 3,241 hours of exposure Reeves and Rat Inhalation 19 of 22 young rats, 13 of 15 older rats BeSO4 at 35.16 µg/m3 (35 hours/week for Deitch 1969 800, 1,600, 2,400 hours); particle size, developed lung tumors after 3 and 18 months, 0.21 µm (mean) respectively; at 3 months, older rats had fewer lung neoplasms than younger rats Wagner et al. 1969 Rat, hamster, Inhalation Bertrandite dust at 15 mg/m3 (Be at 210 18 of 19 rats exposed to beryl ore had lung squirrel µg/m3), beryl ore at 15 mg/m3 (Be at 620 tumors (bronchial alveolar-cell tumors, monkey µg/m3) (6 hours/day, 5 days/week, up to 23 adenomas, adenocarcinomas, epidermoid months); particle size, bertrandite, 0.27 µm tumors); no increased incidence of tumors in rats (mean); beryl ore, 0.64 µm (mean) from dust or in other species from either compound Litvinov et al. 1975 Rat Inhalation Lung tumors found in treatment groups BeF2 at 0.04, 0.4 mg/m3, BeCl2 at 0.02, 0.2 mg/m3 (1 hour/day, 5 days/week, 4 months)

OCR for page 106
Litvinov et al. 1984 Rat Inhalation Malignant lung tumors found in 3 of 44, 4 of 39, BeO, BeCl2 at 0.8, 4, 30, 400 µg/m3 (1 hour/day, 5 days/week, 4 months) 6 of 26, 8 of 21 in BeO group and 1 of 44, 2 of 42, 8 of 24, 11 of 19 in BeCl2 group, respectively Finch et al. 1998b Mouse Inhalation 15, 60 µg of Be (60 µg achieved over 3 4 of 28 p53+/- mice in high-dose group developed (p53+/- days); particle size, 1.4 µm (mean) (GSD, lung tumors; no lung tumors in low-dose group knockout 1.8 µm) or in p53+/- mice and p53+/+ wild- type) Mode-of-action studies Skilleter et al. 1991 Rat hepatic In vitro 50 µM BeSO4 BeSO4 inhibited cell division during G1 phase of BL9L cells cell cycle, but expression of c-myc was maintained in serum-stimulated cells Nickell-Brady Rat Inhalation Be at 410, 500, 830, 980 mg/m3 (single Analysis of p53 and c-raf-1 genes in neoplasms et al. 1994 exposure; lung burdens, 40, 110, 360, 430 did not indicate genetic alterations; weak µg); particle size, 1.4 µm (GSD, 1.9 µm) evidence of mutation of K-ras gene Swafford et al. 1997 Rat primary Aberrant methylation status of p16INK4a , leading lung tumors, to loss of expression cell lines from Nickell-Brady et al. (1994) study Joseph et al. 2001 Mouse Transformed Analyses of gene expression indicate that cell BeSO4 at 50-200 µg/mL BALB/c- cells injected transformation and tumorigenesis are associated 3T3 cells into nude mice, with upregulated expression of genes related to cell lines derived cancer (such as c-fos, c-jun, c-myc, R-ras) and from resulting downregulated expression of genes involved in tumors DNA synthesis, repair, recombination (such as MCM4, MCM5, PMS2, Rad23, DNA ligase I) (Continued) 117

OCR for page 106
TABLE 5-2 Continued 118 Reference Species Route Dose Findings Keshava et al. 2001 Mouse In vitro BeSO4 at 50-200 µg/mL Results show that cells transformed by BeSO4 BALB/c-3T3 are potentially tumorigenic; transformation cells might involve gene amplification of K-ras, c- jun; some transformed cells have neoplastic potential because of genomic instability Belinsky et al. 2002 Rat Inhalation 40, 110, 360, 430 µg of Be (single dose; Tumors induced in part through inactivation of mean lung burdens) p16 and ER genes Misra et al. 2002 Mouse In vitro Phosphorylation increased kinases MEK1, 1-5 nM BeF2 peritoneal ERK1, p38 MAPK, JNK; increases also seen in macrophages NF-κB, CREB transcription factors, c-fos, c-myc Abbreviations: Al, aluminum; Be, beryllium; BeCl2, beryllium chloride; BeF2, beryllium fluoride; BeO, beryllium oxide; BeSO4, beryllium sulfate; Co, cobalt; Cr, chromium; Cu, copper; GSD, geometric standard deviation; Ni, nickel.

OCR for page 106
119 Genotoxicity and Carcinogenicity monkeys; the first tumor was in a monkey that had been exposed for 3,241 hours. The neoplasms metastasized to mediastinal lymph nodes and other areas of the body. Lung tumors were observed in male white random-bred rats exposed to beryllium fluoride (at 0.4 and 0.04 mg/m3) or beryllium chloride (at 0.2 and 0.02 mg/m3) 1 hour/day, 5 days/week, for 4 months (Litvinov et al. 1975). The first neoplasms were observed after 16 months in rats exposed to beryllium fluoride at 0.4 mg/m3 and beryllium chloride at 0.2 mg/m3. Neoplasms also developed in the lungs of rats exposed at the lower concentrations but not in the lungs of the control rats. Litvinov et al. (1984) exposed female albino rats to beryllium oxide or be- ryllium chloride at 0.8, 4, 30, and 400 µg/m3 1 hour/day, 5 days/week, for 4 months. Malignant lung neoplasms developed in a dose-related manner after exposure to either beryllium oxide or beryllium chloride but were not found in the controls. The carcinogenicity of two beryllium ores, bertrandite and beryl, was evaluated in male squirrel monkeys (Saimiri sciurea), male Charles River-CD rats, male Greenacres Controlled Flora (GA) rats, and male Golden Syrian ham- sters (Wagner et al. 1969). The rats and hamsters were exposed to bertrandite or beryl at 15 mg/m3 6 hours/day, 5 days/week, for 17 months, and the monkeys were exposed for 23 months. Beryllium from bertrandite was present in the test atmospheres at 210 µg/m3 and from beryl at 620 µg/m3; the geometric mean diameters of the particles were 0.27 µm and 0.64 µm, respectively. In the beryl- exposed rats, squamous metaplasia or small epidermoid tumors were identified in the lungs of five of 11 rats killed after 12 months of exposure and in 18 of 19 rats after 17 months of exposure. Eighteen of the rats had bronchiolar alveolar- cell tumors, nine had adenocarcinomas, seven had adenomas, and four had epi- dermoid tumors. Although granulomatous lesions were observed in the ber- trandite-exposed rats, no neoplasms were identified in the rats exposed for 6, 12, or 17 months. Neither neoplasms nor granulomas developed in the control rats. Atypical proliferations were observed in the lungs of hamsters 12 months after exposure to bertrandite or beryl. The lesions were reported to be larger and more adenomatous in the beryl group after 17 months. No pulmonary lesions occurred in the control hamsters. No tumors were observed in either the ber- trandite- or beryl-treated monkeys. The carcinogenicity of beryllium metal has also been investigated. In one study, male and female F344/N rats received a single, nose-only exposure to a beryllium-metal aerosol at 500 mg/m3 for 8 minutes, at 410 mg/m3 for 30 minutes, at 830 mg/m3 for 48 minutes, or at 980 mg/m3 for 39 minutes (Nickell-Brady et al. 1994). The latent period for development of neoplasms was about 14 months; tu- mor incidence was 64% over the lifetime of the rats. Most of the neoplasms were adenocarcinomas, although multiple tumor types were observed. In another study of beryllium metals (Groth et al. 1980), lung adenomas and adenocarcinomas were observed in nine of 16 female Wistar rats that re- ceived a single intratracheal instillation of 2.5 mg of beryllium metal, nine of 26

OCR for page 106
120 Managing Health Effects of Beryllium Exposure rats treated with 2.5 mg of passivated beryllium metal, and four of 24 rats given 2.5 mg of beryllium-aluminum alloy. No neoplasms were observed in the lungs of the controls. Pulmonary neoplasms developed in inbred albino rats that were given sin- gle intratracheal deposits of beryllium oxide (fired at high and low temperatures) at 0.036, 0.36, 3.6, or 18 mg/kg (Litvinov et al. 1983). The neoplasms were ade- nomas, adenocarcinomas, and squamous-cell carcinomas. Wistar rats received intratracheal instillations of 1 mg of beryllium oxide (low-fired) once a week for 15 weeks (Ishinishi et al. 1980). An adenocarci- noma, a squamous-cell carcinoma, and four adenomas were observed in the lungs of 30 beryllium-treated rats and no neoplasms in the 16 controls. The mechanism by which beryllium induces carcinogenesis remains un- known. Analysis of p53 and c-raf-1 genes in neoplasms did not indicate genetic alterations, although there was weak evidence of mutations of the k-ras gene (Table 5-2) (Nickell-Brady et al. 1994). It has been suggested that cell transfor- mation and tumorigenesis are associated with upregulated expression of c-fos, c- jun, c-myc, and R-ras genes and downregulation of such genes as MCM4, MCM5, PMS2, Rad23, and DNA ligase I (Joseph et al. 2001). Others have re- ported that transformation by beryllium might involve amplification of k-ras and c-jun (Keshava et al. 2001). Peritoneal macrophages exposed to beryllium showed increased phosphorylation for several kinases and increases in NF-κB, CREB transcription factors, c-fos, and c-myc (Misra et al. 2002). Additional studies are warranted to improve understanding of the mechanisms by which beryllium induces pulmonary neoplasms in animals. In summary, beryllium compounds have induced pulmonary neoplasms in seven strains of rats and the rhesus monkey. Forms of beryllium that have in- duced neoplasia in animals are beryllium sulfate, beryllium fluoride, beryllium chloride, beryllium oxide, the beryllium ore beryl, and beryllium metals. The types of pulmonary neoplasms reported after exposure included adenomas, ade- nocarcinomas, papillary adenocarcinomas, alveolar-cell adenocarcinomas, bron- chiolar alveolar-cell tumors, epidermoid tumors, and squamous-cell carcinomas. Lung tumors have also been induced by beryllium in p53+/- knockout mice. Results of genotoxicity studies show no consistency in type of beryllium tested or genotoxic assay performed. Cancer Risk Estimates EPA (1998a,b) classifies beryllium as a likely human carcinogen on the basis of epidemiologic studies that found increases in lung cancer and support- ing evidence from animal studies that beryllium induces lung cancer in rats and monkeys. For carcinogens, EPA calculates an inhalation unit risk, which is the upper-bound excess lifetime cancer risk estimated to result from continuous exposure to an agent at a concentration of 1 µg/m3. For beryllium, the unit risk is estimated to be 2.4 × 10-3. The cancer dose-response assessment for that esti-

OCR for page 106
121 Genotoxicity and Carcinogenicity mate was originally performed in 1987 and based on an occupational-exposure study (Wagoner et al. 1980). For its cancer-risk estimation, EPA had to assume an occupational exposure of 100 mg/m3 or 1,000 mg/m3. Dose-response assess- ments from animal studies yielded similar estimates of risk, but EPA considered epidemiologic data to be a better basis for estimating cancer risks. In 1998, EPA noted that new epidemiologic studies had been published but found that they shared the same limitations as the Wagoner et al. (1980) study in lacking indi- vidual exposure monitoring or job-history data to support a better quantitative dose-response assessment. However, EPA also noted that a NIOSH study that was being published appeared to have exposure data that might be suitable for quantitative estimation. Until those data were published, EPA recommended that its original unit risk of 2.4 × 10-3 be retained. On the basis of that value, EPA estimated that air concentrations of 0.04, 0.004, and 0.0004 µg/m3 would result in cancer risks of 1 × 10-4, 1 × 10-5, and 1 × 10-6, respectively. The NIOSH study was published in 2001 (Sanderson et al. 2001b); EPA is reassessing beryllium cancer risks. In reviewing the lung-cancer case-control study by Sanderson et al. (2001b) and the reanalyses by Levy et al. (2007) and Schubauer-Berigan et al. (2008), the committee questioned whether a quantitative cancer risk assessment could be carried out with these data. The key reasons for the question were the following: • Three dose metrics were used (cumulative exposure, average exposure, and maximum exposure). The increased odds ratios were associated only with average and maximum exposure, and this suggests that if beryllium induces lung cancer, it may be a high-dose effect. • The odds ratios were approximately constant over the quartiles except for the lowest quartile. That makes it problematic in fitting any dose-response model that is intended to estimate low-dose effects. CONCLUSIONS AND RECOMMENDATIONS There is evidence from controlled studies that exposure to beryllium can cause lung cancer in both sexes of rats, and one study reported lung tumors in monkeys. Epidemiologic studies have reported increases in lung-cancer risk in two worker cohorts exposed to beryllium. Those studies were instrumental in forming the basis of the current cancer classifications by such agencies as IARC, EPA, and NTP. New studies and reanalyses of data performed since those as- sessments have not added substantially to understanding of the carcinogenicity of beryllium or of a dose-response relationship between exposure to beryllium and the development of lung cancer. The committee agrees with the other agen- cies that the balance of the evidence supports a conclusion that beryllium is likely to be a human carcinogen.

OCR for page 106
122 Managing Health Effects of Beryllium Exposure The committee was charged with developing carcinogenic risk estimates for different magnitudes of inhalation exposure to beryllium. However, the committee judged that the available human and animal data are inadequate to support a dose-response analysis with low-dose extrapolation to current expo- sure level magnitudes. Several critical questions needed to characterize a dose- response relationship cannot be answered now. Those questions include • What physicochemical characteristics and particle sizes are associated with beryllium-induced cancer? • Is cancer risk driven by peak exposures, by cumulative exposures, or by other dose metrics? • How have earlier and later exposures of beryllium workers to other lung carcinogens affected disease incidence? • How have changes in workplace practices affected the ability to iden- tify dose-response relationships? • Is there an animal model in which beryllium induces lung cancer by the same mode of action as in humans? The committee concludes that a meaningful cancer dose-response assess- ment cannot be conducted until more information is available on existing or new worker cohorts: complete work history, possible exposure to other carcinogens, and better exposure histories. It may also be necessary to investigate mode of action further with animal studies if a suitable model can be identified. Such studies could help to elucidate the relative importance of peak vs cumulative exposures in cancer incidence. Furthermore, carcinogenic risk estimates would be of limited utility in light of the committee’s recommended approach to preventing beryllium dis- ease. The committee found that it is not possible to reliably identify an exposure magnitude at which there is no risk of sensitization and development of CBD and therefore recommended that the Air Force implement a surveillance and medical-monitoring program to keep exposure as low as feasible to prevent ad- verse health effects. Recommendations for designing such a program are pre- sented in Chapter 7.