Part IV: CASE STUDIES: Using Human Genome Epidemiology Information to Improve Health Chapter 22

PAGE 9 of 13

View Table of Contents

“The findings and conclusions in this book are those of the author(s) and do not
necessarily represent the views of the funding agency.”
These chapters were published with modifications by Oxford University Press (2004)

Human Genome Epidemiology: A Scientific Foundation for Using Genetic Information to Improve Health and Prevent Disease

Immunogenetic Factors in Chronic Beryllium Disease

Erin C. McCanlies, Michael E. Andrew, Ainsley Weston


Tables | Figures | References


Individuals who are exposed to beryllium dust or fumes are at risk of developing a granulomatous lung disease called chronic beryllium disease (CBD). Epidemiologic investigations with an integrated genetic component have linked the development of CBD with the Human Leukocyte Antigen (HLA)-DPB1 gene [1-6]. This case study will focus on the role of HLA-DPB1 in CBD and the practical application of genetic information for the benefit of all members of the beryllium industry workforce.

Beryllium (Be), atomic number 4 on the periodic table of the elements, is extracted from either beryl ore or bertrandite. It is extremely light and, although brittle, stiffer than steel. Beryllium is fused with copper, nickel or aluminum to form highly stable flexible alloys that also have tensile strength. These properties make beryllium an ideal element for numerous technological applications and in high demand (Table 22-1). In 2000, more than 500 metric tons of beryllium were produced world-wide for commercial distribution [7]. Currently it is not known how many individuals have been exposed to beryllium; though estimates as high as 800,000 have been suggested [8]. Exposure to beryllium primarily occurs among workers in beryllium manufacturing plants in which metal fabrication and pressing of beryllia ceramics occurs. For this reason, the beryllium worker population is most often afflicted with CBD.

Though CBD is the principal form of beryllium disease seen today, prior to 1946 acute beryllium disease (ABD) was also a problem. ABD was the result of short, extremely high doses of beryllium. Although removal from exposure and treatment often lead to resolution, in its most severe form ABD was fatal. Of those cases that were not fatal, approximately 17% progressed to CBD [9-11].

CBD was first described among workers in the fluorescent light industry [12]. Soon after, it was also recognized in beryllium industry workers. In the U.S. in 1949, to reduce exposure experienced by beryllium workers, the Atomic Energy Commission, predecessor of the Department of Energy (DOE) proposed a beryllium exposure limit of 2 µg/m3 averaged over an 8 hour work period [13]. This restriction largely eliminated cases of ABD, but despite overall reduction in workplace exposure CBD continues to occur [13]. This has resulted in a series of investigations aimed at better understanding the natural history of disease, routes of exposure, and the immunogenetics of CBD [1-6,14-29].

In the 1950s, a beryllium case registry was established in the U.S. to study the pathology and natural history of CBD. Case registry information and cross-sectional surveillance of beryllium worker populations have been utilized to better define the mechanism of CBD. In conjunction with this research, laboratory based research determined that exposure to beryllium triggered a cell-mediated, type IV delayed hypersensitivity reaction resulting in the proliferation of beryllium-specific T lymphocytes [16,17]. Based on an understanding of this mechanism, a specific, cellular hypersensitivity response to beryllium challenge was demonstrated in vitro. This formed the basis for the development of a beryllium lymphocyte proliferation test (BeLPT) [18, 19]. Utilization of the BeLPT in cross-sectional beryllium worker surveys identified workers who were beryllium sensitized, but who did not have impaired lung function or overt clinical symptoms of CBD [20-25]. This suggested that CBD progressed from a subclinical state in which individuals first became sensitized to beryllium prior to developing CBD. It is not yet known if all beryllium sensitized individuals will eventually develop CBD.

Currently, the BeLPT is used to determine if beryllium workers have become immunologically sensitized to beryllium. Due to inter-laboratory and intra-laboratory variation, some persons only test positive at one laboratory or the other [20, 23-29]. For this reason, a diagnosis of beryllium sensitization is dependent on repeated positive BeLPT responses [20-29].

Early detection and diagnosis of CBD is dependent on physical screening of workers that includes the BeLPT. Workers who test positive in successive BeLPT tests are referred for bronchoalveolar lavage to assess the immune status of lymphocytes (BAL-LPT) in the lung and bronchial biopsy to determine if granuloma formation has occurred in the lungs. Radiographic changes consistent with CBD have also been used to diagnose CBD [20, 23-29].

Epidemiology of Chronic Beryllium Disease

Cross-sectional surveys have found that between 1% to 12% of beryllium workers are sensitized to beryllium. Of these, 36% to 100% were found to have CBD [20, 23-28]. The variation in these rates may be due to at least four different factors: the form of beryllium the worker is exposed to; the magnitude or duration of beryllium exposure; temporal variation in immune system response; differing susceptibilities. As previously discussed, rates of sensitization are also dependent on BeLPT laboratory variation and whether split sampling is conducted. In the case of CBD, prior to the manifestation of physical symptoms, missed granulomas in bronchial biopsies contribute to variable rates of CBD.

Cross-sectional epidemiologic studies that evaluated the prevalence of sensitization and CBD in relation to work processes and beryllium exposure measurements found that risk of sensitization and CBD was dependent on having performed certain jobs or tasks [23, 25]. In 1992, a study that evaluated 136 beryllium ceramics workers found that eight workers were beryllium sensitized. Of these, six had granulomatous lung disease (CBD). When the specific jobs and tasks performed by these workers were evaluated, machinists were significantly more likely to be beryllium sensitized than the other employees (14.3% vs 1.2%; OR=14.3, p =0.003) [25].

In 1993-1994 cross-sectional surveillance was conducted at a beryllium manufacturing plant in which pure metal, oxides, and alloys are produced. Prior to 1980, this plant had also manufactured beryllium ceramics, but this process had since moved to another facility [23]. Fifty-nine workers of 646 participating employees were identified as being beryllium sensitized. Forty-seven of these individuals underwent clinical evaluation that included bronchoscopy. Twenty-four cases of CBD were identified [27]. When specific areas, jobs, and tasks were evaluated, individuals who had ever worked in the area referred to as the pebble plant were at the greatest risk of CBD (OR= 23.5; 95% CI=4.4-125.5) followed by those who had ever worked in ceramics (OR=4.4; 95% CI=1.8-10.5). Employees who worked in metal production were more likely to be beryllium sensitized compared to other workers (7.3% vs 1.3%), and had an increased prevalence of CBD (19.2%) [27].

Air sample measurements taken at the ceramics plant in 1992 indicated that compared to other processes, machining had higher general area and breathing zone beryllium measurements (p=0.0001). These results demonstrated a quantitative relationship between process-related beryllium exposure and the risk of beryllium sensitization [23]. However, air sample measurements and rates of sensitization and CBD have not always correlated. For example, even though rates of CBD were highest in individuals who had ever worked in the pebbles plant, the median average beryllium air exposure since 1984 was estimated to be 1.3 µg/m3, not significantly different than other areas [27]. Further, a follow-up study of the 1992 ceramics plant found that the overall decline in beryllium exposures was not matched by a decline in the prevalence of sensitization or CBD [28]. These results are consistent with previous findings in which the number of individuals with CBD did not reflect the level of beryllium exposure [27, 30]. This discrepancy is even more apparent in light of documented cases of CBD in workers who had very little apparent beryllium exposure (e.g. administrative workers), as well as family and community clusters of cases among individuals who had never worked in the beryllium industry [30]. These results are important because they raised concerns about how beryllium exposure is assessed [31]. They are also suggestive of the immunogenetic nature of CBD.

Gene Associations with CBD

In vitro studies conducted to evaluate the immunogenetic nature of CBD demonstrated that anti-HLA-DP antibodies blocked beryllium-stimulated T lymphocyte proliferation [32]. This suggested that MHC class II antigen-bearing cells were involved in a beryllium-specific T-lymphocyte mediated (type II) response in the development of CBD. Thus, genetic variants of the HLA
(-DP, -DQ, and -DR) loci were evaluated as potential risk factors for CBD. A map of chromosome 6p12.3 shows the relative locations of the HLA genes (Figure 22-1) [33,34].

Heteroduplex analysis, allele-specific polymerase chain reaction (PCR), restriction fragment length polymorphism (RFLP), oligonucleotide hybridization and direct and indirect sequencing of PCR products have all been used to investigate the HLA-gene locus. HLA-DP variants have been identified as being associated with sensitization and CBD.

The HLA-DP molecule is composed of two chains, alpha (A1) and beta (B1)1 . The A1 locus lies proximal to the B1 locus and the gene products form a heterodimer. There are fewer known human HLA-DPA1 variants than B1 variants (20 vs. 103), and it appears that there are a limited number of A/B haplotypes [35]. Structurally, the alpha and beta chains form a groove in the DP molecule that plays an important function in immunologic processes. Initial research focused on the HLA-DPB1, but more recent investigations have also evaluated the role of HLA-DPA1, alone and in combination with HLA-DPB1 in the risk of sensitization and CBD [1-6].

Richeldi et al., 1993 used oligonucleotide hybridization techniques to evaluate HLA-DPB1 haplotypes characterized by polymorphisms in codons 36, 55 – 57, and 65 – 69 in beryllium workers with and without CBD (Table 22-2) [1]. The results of this initial study indicated that CBD cases were more likely to have HLA-DPB1 alleles coding for aspartic acid (D) and glutamic acid (E) in positions 55 and 56, respectively, compared to the controls who where more likely to code for an alanine (A) in those positions (79% vs 41%; OR=5.4, 95% CI=1.7-17.6). Furthermore, alleles characterized by a codon for glutamic acid residue at position 69 (E69) in the amino acid sequence, also occurred more often in individuals with CBD than in those without (97% vs 27%; OR=85.3, 95% CI=10.9-3,578.0) (Table 22-2). No significant association was seen between CBD and the polymorphic codon 36. Allele specific analysis implicated an association between inheritance of the common HLA-DPB1*0201, glutamic acid 69 containing allele and CBD (OR=4.3, 95% CI=1.1-20.8). Conversely, HLA-DPB1*0401, which does not code for glutamic acid at position 69, occurred less often in cases compared to controls (14% vs 48%; OR=0.2, 95% CI=0.05-0.7).

Wang et al.,1999 investigated the presence and absence of both HLA-DPB1 and HLA-DPA1 alleles in beryllium exposed individuals with (n = 20) and without (n = 75) CBD [3]. This study was important for two reasons: it verified the association between HLA-DPB1E69 and CBD; and it evaluated allele specific relations including the effect of homozygosity versus heterozygosity and disease status. Although there are some methodological concerns regarding the geographic origin of the cases and controls [3], this study was technically superior to Richeldi et al.,1993, 1997 [1, 2] in its use of allele-specific dideoxy-chain termination-DNA sequencing to characterize HLA alleles.

In this beryllium exposed population the odds of disease in the presence of HLA-DPB1E69 was estimated to be 22.9 (95% CI = 4.8 – 108.2). Thus, in beryllium industry workers with CBD, 19/20 or 95% were found to carry at least one HLA-DPB1E69 variant (or putative disease allele) compared to 34/75 (45%) without disease. The one individual with CBD who did not carry at least one HLA-DPB1E69 variant was found to be homozygous K69/K69 (i.e. lysine homozygote, no putative disease alleles). Although the numbers were small, the data also suggested that individuals homozygous for HLA-DPB1E69 were at an increased risk of disease compared to heterozygous individuals. Among the 19 individuals with CBD who carried at least one HLA-DPB1E69 variant, 6 (~32%) were homozygous for HLA-DPB1E69 compared to only 1 of the 34 individuals without CBD (1.3%) (OR = 15.2; 95% CI = 15.2 – 721.0).

Wang et al.,1999 also evaluated the distribution of alleles in HLA-DPB1E69 individuals with and without CBD. These data suggested that variants in positions other than 69 also had a bearing on CBD risk (Figure 22-1) [3]. Alleles coding for amino acids valine, histidine or tyrosine, and leucine (V, H/Y, L) at positions 8, 9 and 11 were found to occur more often in individuals with CBD than in those without (79% vs. 29%; OR = 9.0, 95% CI = 2.6 – 31.6). Alleles coding for aspartic acid, glutamic acid, alanine, and valine (D, E, A, V) at positions 84 – 87 were also found to occur more frequently in individuals with CBD than alleles coding for glycine, glycine, proline, methionine (G, G, P, M) (84% vs. 35%; OR = 9.8, 95% CI = 2.6 – 36.6). Data for absolute characterization of genotypes with detail pertaining to these positions are not presented and cannot be deduced from the literature reports [1-3]. However, analysis of the allele specific data suggests a hierarchy with respect to the risk associated with specific alleles coding for E69. The lowest odds appears to be associated with HLA-DPB1*0201/2 (OR ~ 15, 95% CI ~ 3 – 85). The highest odds were associated with non-HLA-DPB1*0201/2 alleles (e.g. HLA-DPB1*1901 < HLA-DPB1*1301 < HLA-DPB1*0901 = HLA-DPB1*1001 < HLA-DPB1*0601 < HLA-DPB1*1701 (OR ~ 246, 95% CI ~ 38 – 1594)). In the case of individual alleles, a small sample size results in extremely large overlapping confidence intervals. A consensus has not yet been reached in the literature concerning the relative risk potency of the HLA-DPB1*0201/2 versus the rarer non-*0201/2 HLA-DPB1E69 alleles. The main differences between these alleles lie in codons 9 and 84 – 87. This observation forms the basis of a potentially important hypothesis that will be difficult to address because of small numbers and the problem of multiple comparisons.

Based on the HLA-DPB1 allele specific information, Wang et al. evaluated the distribution of HLA-DPA1 alleles in HLA-DPB1*0201 and non-*0201/2 HLA-DPB1E69 carriers. Among the HLA-DPB1*0201 individuals 29 of 30 (7/8 CBD vs 22/22 controls) were found to have at least one HLA-DPA1*0103 allele [3]. In contrast, HLA-DPA1*0201 occurred more often in the non-*0201/2 HLA-DPB1E69 carriers (14/16 CBD vs 12/13 controls). Although these results are preliminary, they indicate that HLA-DPA1 may also play a role in CBD risk and warrant further study.

Wang et al., 2001 extended their earlier work to include 25 beryllium-sensitized individuals (BeLPT positive without CBD) [4]. They then evaluated the frequency of HLA-DPB1E69 in individuals with CBD, beryllium sensitization, and beryllium exposed individuals without either beryllium sensitization or CBD. Interestingly, the frequency of both the high risk HLA-DPB1E69 and non-*0201/2 HLA-DPB1E69 alleles among the sensitized, consistently fell between the frequency observed in the CBD cases and controls. For example, 24% of the beryllium sensitized individuals were homozygous HLA-DPB1E69 compared to 30% of the individuals with CBD and 3% of the controls (p<0.001). Similarly, sensitized individuals were more likely to have at least one non-*0201/2 HLA-DPB1E69 allele compared to the controls (52% vs 13%; p< 0.001), but this occurred less often than in those with CBD (52% vs 80%; not significant). Of the non-*0201/2 HLA-DPB1E69 alleles examined, HLA-DPB1*1701 occurred most often. Among the beryllium sensitized and individuals with CBD, 16% and 30% respectively, were HLA-DPB1*1701 positive compared to only 2% of the control group (p< 0.01).

These results are of interest for two reasons. First they extended the work by Richeldi et al [1,2]. Second, it was the first study to evaluate the role of HLA-DPB1*0201 and non-*0201/2 HLA-DPB1E69 alleles in individuals who were beryllium sensitized. However, due to the small sample size and potential problems with the composition of the CBD cases, sensitized, and control groups these results must be viewed with caution. For example, 10 of the sensitized did not have signs of respiratory impairment, but none were clinically evaluated for granulomatous lung disease, and 2 of the sensitized individuals in the most recent study were not known to have been occupationally exposed to beryllium. Any of these factors might affect the observed rates.

Saltini and collegues (2001) recently conducted a study that analyzed the presence and absence of specific HLA-DPB1 alleles in 22 individuals with CBD, 23 individuals with beryllium sensitivity (BeLPT positive without CBD), and 93 control samples [5]. They reported an association between HLA-DPB1E69 and disease (OR=3.7, 95% CI=1.4-10.0), but not with sensitization (OR=0.9, 95% CI=0.3-2.2). They also stated that an increased frequency of *0201 HLA-DPB1E69 alleles was associated with CBD, however, this did not appear to be significant [5].

Rossman et al., 2002 published information on the genetics of beryllium sensitization and CBD. The study population consisted of 137 individuals who had been referred to a tertiary referral hospital for clinical evaluation of CBD. Fifty-five of the participants had a positive BeLPT and were designated at having beryllium hypersensitivity (BH). On clinical exam 25/55 were determined to have CBD and 30/55 were defined as having beryllium hypersensitivity without clinical disease (BHWCD). The control group consisted of 82 beryllium exposed individuals who had been evaluated for CBD at the hospital. None had BH or positive BeLPT results, though 10 had abnormal chest radiographs. HLA-DPB1 genotyping was conducted on all the samples and the frequency of alleles compared across the groups with and without disease [6]. HLA-DQB1 and HLA-DRB1 were evaluated, but not in conjunction with HLA-DPB1, and so will not be discussed here.

HLA-DPB1E69 occurred more often in BHWCD (90%) and individuals with CBD (84%) than those without disease (48%). The highest odds of disease was associated with BHWCD (OR=9.9; 95% CI 2.8-35.3). When the frequency of HLA-DPB1E69 in individuals with BHWCD was compared to the frequency in individuals with CBD there was no significant difference. When specific HLA-DPB1E69 alleles were evaluated HLA-DPB1*0601 and HLA-DPB1*1301 occurred more often in BHWCD individuals than controls (p< 0.05). These however did not remain significant after they were corrected for the number of alleles evaluated.

The presence of lysine at position 11 (HLA-DPB1L11) and aspartic acid at position 55 (HLA-DPB1D55) were significantly associated with BH. HLA-DPB1L11 was also associated with BHWCD. However, Rossman et al., 2002 also notes that this association only remained significant in the presence of HLA-DPB1E69. Further, there was no difference between the frequency of either the HLA-DPB1-E69 -L11 or the HLA-DPB1-E69 -D55 in the individuals with CBD or BHWCD [6]. Rossman et al., 2002 concluded that HLA-DPB1E69 was the most important epitope in the development of BH, but could not be used to predict whether someone would develop CBD.

While all of the studies conducted agree that HLA-DPB1E69 is associated with CBD they differ in the relative importance placed on the role of the HLA-DPB1*0201 alloforms in CBD [1, 3-6]. Further, it is of interest that while Wang et al., 2001 [4] and Rossman et al., 2002 [6] showed a relationship between HLA-DPB1E69 and beryllium sensitization, Saltini et al., 2001 did not report this relationship [5]. This discrepancy might be the result of the different methods used to determine HLA-haplotypes, or differences in the populations under study. Future studies will formally address the differences observed across these studies, and studies with larger sample sizes might help to elucidate specific high risk HLA-DP alleles or other high risk genes.

Population Frequencies of HLA-DPB1

A numbers of studies have examined the frequency of HLA-DPB1E69 in various racial/ethnic groups [35-54]. The data in the literature are variable in the amount of haplotype detail given, however, typing for 36 populations that discriminated between*0201/2 HLA-DPB1E69 alleles and non-*0201/2 HLA-DPB1E69 alleles is summarized in Figure 22-2 . These data indicate that some populations have a complement of E69-alleles that form the majority of HLA-DPB1 haplotypes, while others have very low frequencies (in some populations undetectable). Moreover in some populations the *0201 family is more frequent than the non*0201, while in others the non*0201 alleles are more common (Figure 2). These data are consistent with recently published finding using an RFLP, where the E69 carrier frequencies among Caucasians, African-Americans, Hispanics and Chinese fell in the range 0.33 – 0.59 [55].


Richeldi et al., 1997 evaluated the risk of CBD in the presence of both a high risk job and HLA-DPB1E69 [2]. As beryllium machining had been found to confer the highest job related risk, machining and inheritance of the HLA-DPB1E69 were evaluated. Using logistic regression, the odds of CBD associated with HLA-DPB1E69, independent of a machining job history was estimated to be 11.8 (95% CI = 1.3-108.8). The odds of disease associated with machining alone was 10.1 (95% CI = 1.1-93.7) (Table 22-3). Based on these results they reported that genetic and job factors had at least an additive effect for risk of beryllium disease in the industrial environment. However, due to small numbers they were unable to statistically verify this in the regression model. We have included an additional summary of disease prevalence by HLA-DPB1E69 and machining job history for this study (Table 22-4). While it was not possible to estimate odds ratios referenced to the lowest risk group because of zero observed cases, it is clear from looking at the prevalence estimates and confidence intervals that the presence of both HLA-DPB1E69 and machining job history account for a remarkable proportion of cases.

Saltini et al., 2001 utilized a series of two by two tables extracted from a larger two-by-four table to evaluate the risk of beryllium sensitization or CBD in the presence of either one, or a combination of the genes: HLA-DPB1E69; tumor necrosis factor (TNF)-?-308*2; and HLA-DRR74 (2) . HLA-DRR74 was found to be independently associated with sensitization (OR=4.0, 95% CI=1.5-10.1), but not with CBD (OR=0.9, 95% CI=0.3-2.6) [5]. An association was also reported between BeLPT positive individuals (sensitized and CBD) and TNF-α-308*2 (OR=7.8, 95%C I=3.2-19.1).

Gene-gene analyses identified an association between sensitization and the presence of TNF-?-308*2 and HLA-DRRR74 [5]. An association between the risk of sensitization in individuals who were HLA-DPB1E69 positive, HLA-DRRR74 negative was also reported. However, scrutiny of their tabulated data shows that sensitization was associated with HLA-DRR74 positive, HLA-DPB1E69 negative individuals. Interestingly, neither HLA-DPB1E69 nor HLA-DRRR74 alone, or in combination, were associated with CBD. This may be an effect of the construction of the 2-by-4 tables, resulting in small cell counts. TNF-α-308*2 was independently associated with CBD, but in the presence of HLA-DPB1E69, this risk was even greater. The risk associated with TNF-α-308*2 alone was estimated to be 4.6. In individuals with both a TNF-α-308*2 and HLA-DPB1E69 compared to those with neither, the odds of disease increased to 9.7. The extent to which these analyses may have been affected by the use of different laboratory methods to determine TNF-α-308*2, HLA-DPB1E69, and HLA-DRRR74 alleles is unknown. Although exposure to beryllium is a necessary component in the development of CBD, these results suggest that genes other than, or in conjunction with, HLA-DPB1E69 may play a role in the risk of both sensitization and disease.

Availability of HLA-DPB1 Gene Testing

One of the leading beryllium manufacturing facilities, in collaboration with a tertiary referral hospital, is offering prospective employees the opportunity to confidentially obtain HLA-DPB1E69 genotyping. Individuals applying for work at the plant are referred to the hospital which is responsible both for conducting the genetic analysis and counseling the prospective employees about their risk of beryllium sensitization and disease. The plant does not receive individual results, rather they receive a summary report including the number of individuals who requested the test and the number of individuals who accepted employment at the plant. This report is only conveyed after enough individuals have participated to prevent the plant from being able to identify individual participants.

The decision by the beryllium industry to offer applicants HLA-DPB1E69 genotyping through an independent third party was reinforced by recommendations both by current workers and the Beryllium Industry Scientific Advisory Committee. Their goal is two fold; first, it is to enable applicants to better be able to assess their risks associated with working with beryllium. Secondly, it is to potentially lower the risk of CBD among new hires.

Besides this program, as part of ongoing molecular epidemiologic research studies, experimental laboratory methods are being conducted to characterize HLA-DPB1 sequence motifs in individuals with CBD, beryllium sensitization and in individuals without CBD. Individuals who participate in these research studies can obtain their genetic information upon request.

Currently there are no commercial kits available, nor are there Clinical Laboratory Improvement Advisory Committee (CLIAC)-approved laboratories offering HLA-DPB1E69 genetic testing. Thus persons interested in obtaining their HLA-DPB1E69 genotype, but who are not seeking employment at the beryllium plant, nor participating in the epidemiologic research are required to seek out an independent laboratory capable of providing such a test. There are a number of potential problems associated with this. The cost of the test could be prohibitive and depending on the reliability of the laboratory, the results may or may not be accurate. Further, it is not evident that a test sought in this manner would include counseling. What the results mean and how this genetic information can be used to benefit, or negatively impact current, former, and prospective workers is of particular concern to beryllium researchers, the beryllium industry and to the individuals who obtain their genetic information. The benefits and risks are discussed below.

Potential Benefits and Risks Associated with Genetic Information

HLA-DPB1E69 genetic information is valuable for better understanding the molecular mechanism of CBD and beryllium sensitization. It is a necessary component in the development of animal models and may lead to better treatments and modes of intervention to prevent disease for all workers. In addition, for some individuals, knowing their HLA-DPB1E69 status may be important in assessing whether or not they want to work in the beryllium industry.

An animal model for beryllium sensitization or CBD is not available. For example, when mice are exposed to beryllium they can become sensitized but they do not develop granulomas like humans. A major difference between the mouse and human histocompatibility antigens is that the mouse does not have an HLA-DPB1 homologue [56]. Therefore, if the correct HLA-DP haplotypes could be introduced, it is possible that a mouse model could be developed that would become sensitized to beryllium and develop CBD. A transgenic mouse model susceptible to a disease condition similar to humans would be invaluable for studying, (a) the pathobiology of sensitization and CBD, (b) questions concerning dose and route of exposure, (c) the development of better diagnostic tools, and (d) the development of post-exposure intervention strategies.

Currently, treatment options for chronic beryllium disease are limited to anti-inflammatory and immunosuppressive agents (e.g., prednisone). Thus, treatment is generally palliative and can have serious adverse effects (i.e. electrolyte imbalance, diabetes, congestive heart failure, hypertension, and others) [57]. In the absence of complete abrogation of exposure, research on the genetic underpinnings of sensitization and disease may lead to better diagnostic tests, effective post-exposure interventions, and implementation of exposure limits that would protect all workers.

For individuals, knowing their HLA-DPB1E69 status is beneficial if they are interested in knowing more about their own risk. To help these individuals understand what their results might mean, and how the information may or may not benefit them, it is important that they receive extensive written and oral information describing the risks and benefits associated with receiving their genetic results. This includes information about potential insurance and employment discrimination, the predictive value for HLA-DPB1E69 , and the odds of disease associated with HLA-DPB1E69 .

The most obvious risks prospective, current, and former beryllium workers face is the potential for insurance and employment discrimination. There are no recorded cases to date in which beryllium workers have been refused health insurance or employment based on their genotype status. However, current laws vary in who is protected and the extent of protection, individuals who decide to obtain their genetic information must be made aware of this, and that insurance or employment discrimination is a potential risk.

Another issue prospective, current, and former beryllium workers must be made aware of is the predictive value associated with HLA-DPB1E69 and beryllium disease. The positive predictive value of a diagnostic test provides information about how well the presence of a positive diagnostic test outcome will detect the presence of disease. Typically the positive predictive value has been defined as the probability that an individual will have disease given that the diagnostic test is positive. It is a function of test sensitivity, test specificity and disease prevalence [58]. This definition of positive predictive value is cross-sectional in nature and does not involve information about disease incidence.

Positive predictive value can also be defined in terms of the probability that individuals will develop disease subsequent to screening [58, 59]. This definition of positive predictive value is based on test sensitivity, test specificity and disease incidence rather than disease prevalence. This modified definition of positive predictive value is interpreted as the probability that an individual will develop the disease subsequent to screening given that they have a positive screening test.

While both definitions of positive predictive value have utility in disease screening and prevention it is essential to preserve the distinction between the cross-sectional and longitudinal definitions when interpreting estimates of positive predictive value. For situations where both disease prevalence and incidence are available then it may be useful to calculate both the cross-sectional and longitudinal positive predictive value. Since disease incidence data are not always readily available then the cross-sectional positive predictive value provides useful information about the probability that an individual has developed disease given a positive test result.

Alternatively, negative predictive value is defined longitudinally as the probability that an individual will not develop disease subsequent to screening given that a negative screening test has occurred. Cross-sectionally negative predictive value is defined as the probability that an individual does not have the disease of interest given a negative outcome on a given screening test.

With these definitions in mind we used the model developed by Khoury et al., 1985 [59] and cross-sectional data from Richeldi et al., 1997 [2] to estimate the positive and negative predictive value and the sensitivity and specificity of a genetic test based on determination of the supratypic marker HLA-DP?1E69 (Figure 22-3 ). This estimate spans a wide range HLA-DP?1E69 carrier frequencies, and assumes that risk is independent of other genetic risk factors that are not in disequilibrium with the HLA-DP?1 locus. It is based on a disease prevalence of 5% for beryllium industry workers and 15% for high exposure tasks (machinists and lappers). Figure 22-3 shows that regardless of the prevalence of disease (5% or 15%), as the frequency of HLA-DP?1 increases in a population, the positive predictive value decreases well below 50%. What this means, and what must be stressed to the workers is that although the odds of beryllium sensitization and disease is greater in individuals with HLA-DP?1E69, the absence of HLA-DP?1E69 does not protect them from beryllium sensitization and CBD [1-6].

Furthermore, while prospective employees might benefit from knowing their genotype (if it is confidentially provided) the utility of risk information for people already exposed to beryllium is less clear, since CBD risk continues even with exposure cessation. Currently it is not known whether risk can be lowered by leaving the industry or whether genetic characterization of sensitized or CBD cases has prognostic implications. These issues must also be discussed with prospective, current, and former beryllium workers, as do the potential risks associated with insurance and employment discrimination.


Occupational exposure to beryllium presents a clear risk of adverse health outcome. Despite the implementation of the 2 µg/m3 exposure limit and implementation of control technology (e.g. respirators), CBD and beryllium sensitization continues to be problematic [20, 23-28]. In addition to beryllium exposure, a strong association between certain genetic factors and the risk of disease has also been identified [1-6]. In light of the high odds of disease associated with specific genes it would seem prudent to implement a genetic testing program that would potentially reduce the numbers of high risk people from being exposed to beryllium. However, where this has proven effective for other diseases that have a strong genetic component, (e.g. hemochromotosis and HFE gene) [60], in the case of CBD, because of potential employment and insurance discrimination, the ethical, legal and social implications must be stressed.

Here also, we focused on the scientific issue of the positive predictive value of such a genetic test and have found that HLA-DPB1E69 in the beryllium worker population results in a low positive predictive value. However, because the association between HLA-DPB1E69 and CBD is unequivocal, if confidentially provided, prospective employees may find having their genetic information useful for deciding if they want to work in the beryllium industry. Despite the level of uncertainty, for some, knowing only that their risk was higher is sufficient, regardless of the predictive value. Similarly for current workers, though they have already been exposed, the realization that they are at an increased risk may also affect their decision about remaining in the industry.

Unfortunately the scientific evidence is not yet available to advise beryllium workers on a definite course of action based on their genotype. Rather, only the potential risks and benefits can be discussed. However, as integrated genetic and epidemiologic research studies continue, many of these issues may be resolved. Research will allow specific questions associated with the natural history of CBD to be addressed. The identification of other high risk genes, gene-exposure, and gene-gene interactions will also improve personal risk assessment and help determine if specific genes or alleles are more valuable as prognostic indicators. Furthermore, these and similar studies will be able to clarify the pathology of CBD, opening the door to more effective treatment and interventions.





  1. Richeldi L, Sorrentino R, Saltini C. HLA-DPB1 glutamate 69: a genetic marker of beryllium disease.
    Science 1993;262:242-4.
  2. Richeldi L, Kreiss K, Mroz M, et al. Interaction of genetic and exposure factors in the prevalence of berylliosis. Am J Ind Med 1997;32:337-40.
  3. Wang Z, White PS, Petrovic M, et al. Differential susceptibilities to chronic beryllium disease contributed by different Glu69 HLA-DPB1 and -DPA1 alleles. J Immunol 1999;163:1647-53.
  4. Wang Z, Farris GM, Newman LS, et al. Beryllium sensitivity is linked to HLA-DP genotype. Toxicology 2001;165:27-38.
  5. Saltini C, Richeldi L, Losi M, et al. Major histocompatibility locus genetic markers of beryllium sensitization and disease. Eur Respir J 2001;18:677-84.
  6. Rossman M, Stubbs J, Lee CW, et al. Human leukocyte antigen class II amino acid epitopes. Susceptibility and progression markers for beryllium hypersensitivity. Am J Respir Crit Care Med 2002;165:788-794
  7. Cunningham L. Beryllium. In: US Geological Survey Minerals yearbook – 2000, pp12.1 – 12.5.
  8. Cullen MR, Kominsky JR, Rossman MD, Cheerniak MG, Rankin JA, Balmes JR, Kern JA, Daniele RP, Palmer L, Naegel GP, McManus K and Cruz R. Chronic beryllium disease in a precious metal refinery. Am Rev Resp Dis 1987;135: 201-208.
  9. Van Ordstrand HS, Chemical pneumonia in worker extracting beryllium oxide. The Cleveland Clinic Quarterly 1943;10.
  10. Van Ordstrand HS Hughes R, De Nardi JM, Beryllium poisoning. JAMA 1945;129: 1084.
  11. Hardy HL. Beryllium poisioning: lessons in control of man-made disease. N Engl J Med 1965; 273:1188-1199.
  12. Hardy HL, Tabershaw IR. Delayed chemical pnuemonitis occurring in workers exposed to beryllium compounds. J Ind Hyg Toxicol, 1946;28:197-211.
  13. Eisenbud M. Origin of the standards for control of beryllium disease (1947-1949). Environ Res 1982;27:79-88
  14. Tinkle SS, Antonini JM, Abrigo BA, et al. 2000. Particle penetration of the skin as a route of sensitization in occupational lung disease (abstract). Society of Toxicology Meetings 54:149
  15. Lademann, J., Weigmann, H., Rickmeyer, C., Barthelmes, H., Schaefer, H., Mueller, G., and Sterry, W. Penetration of titanium dioxide microparticles in a sunscreen formulation into the horny layer and the follicular orifice, Skin Pharmacol Appl Skin Physiol 1999; 12:247-56.
  16. Newman LS. Lloyd J. Daniloff E. The natural history of beryllium sensitization and chronic beryllium disease. Environl Health Perspec 1996;104S:937-943.
  17. Curtis GH, The diagnosis of beryllium disease with special reference to the patch test. Arch Ind. Health 1959;19:150-153.
  18. Hanifin JM. Epstein WL. Cline MJ. In vitro studies on granulomatous hypersensitivity to beryllium. Journal of Invest Dermatol 1970;55:284-288.
  19. Deodhar SD, Barna B, Van Ordstrand HS: A study of the immunologic aspects of chronic berylliosis. Chest 1973;63:309-313.
  20. Kreiss K, Newman LS, Mroz MM, et al. Screening blood test identifies subclinical beryllium disease. J Occup Med 1989;31:603-8.
  21. Williams WR, Williams WJ: Development of beryllium lymphocyte transformation tests in chronic beryllium disease. Int Arch Allergy Appl Immunol 1982;67: 175-180.
  22. Rossman MD, Kern JA, Elias JA, Proliferative response of bronchioalveolar lymphocytes to beryllium: a test for chronic beryllium disease. Ann Intern Med 1988;108: 687-693.
  23. Kreiss K, Wasserman S, Mroz MM, et al. Beryllium disease screening in the ceramics industry. Blood lymphocyte test performance and exposure-disease relations. J Occup Med 1993;35:267-74.
  24. Kreiss K, Mroz MM, Zhen B, et al. Epidemiology of beryllium sensitization and disease in nuclear workers. Am Rev Respir Dis 1993;148:985-91.
  25. Kreiss K, Mroz M, Newman LS, et al. Machining risk of beryllium disease and sensitization with median exposures below 2 micrograms/m3. Am J Ind Med 1996;30:16-25.
  26. Stange AW, Furman FJ, Hilmas DE. Rocky Flats Beryllium Health Surveillance. Environ Health Perspect 1996;104S:981-6.
  27. Kreiss K, Mroz MM, Zhen B, et al. Risks of beryllium disease related to work processes at a metal, alloy, and oxide production plant. Occup Environ Med 1997;54:605-12.
  28. Henneberger PK, Cumro D, Deubner DD, et al. Beryllium sensitization and disease among long-term and short-term workers in a beryllium ceramics plant. Int Arch Occup Environ Health 2001;74:167-76.
  29. Deubner DC, Goodman M, Iannuzzi J: Variability, predictive value, and uses of the beryllium blood lymphocyte proliferation test(BLPT): preliminary analysis of the ongoing workforce survey. Appl Occ Env Hyg 2001;16:521-526.
  30. Eisenbud M and Lisson J. Epidemiological aspects of beyllium induced non-malignant lung disease: A 30-year update. J Occ Med 1983;25:196-202
  31. Kent MS, Robins TG, Madl AK. Is total mass or mass of alveolar-deposited airborne particles of beryllium a better predictor of the prevalence of disease? A preliminary study of a beryllium processing facility. Appl Occ and Envir Hyg 2001;16:539-558.
  32. Saltini C. Weinstock K. Kirby M. Pinkston P, Crystal RG.: Maintenance of alveolitis in patients with chronic beryllium disease by beryllium specific helper T cells. N Engl J Med 1989;320:1103-1109.
  33. International Human Genome Sequencing Consortium. Initial sequencing and analysis of the human genome.
    Nature 2001;409:860-921
  35. al-Daccak R, Wang FQ, Theophille D, Lethielleux P, Colombani J, Loiseau P. Gene polymorphism of HLA-DPB1 and DPA1 loci in caucasoid population: frequencies and DPB1-DPA1 associations. Hum Immunol 1991;31:277-285.
  36. Carippa A, Sands B, Forcione, D, et al. Analysis of MHC classII DP, DQ and DR alleles in Chrohn’s disease. Gut 1998;43:210-215
  37. Loudova M, Sramkova I, Cukrova V, Sieglova Z, Brdicka R. Frequencies of HLA-DRB1, -DQB1 and -DPB1 alleles in Czech population. Folia Biologica 1999;45:27-30.
  38. Lui Z, Lin J, Chen W, Jia Z, Pan D, Xu A. Sequence of complete exon 2 and partial intron 2 of HLA-DPB1*8001 allele. GenBank accession # AF336231. 2001.
  39. MacHulla HK, Schonermarck U, Schaaf A, Muller LP, Kloss C, Kruger J, Kunze G, Schonermarck G, Langner J. HLA-A, B, Cw and DRB1, DRB3/4/5, DQB1, DPB1 frequencies in German immunoglobulin A-deficient individuals. Scand J Immun 2000;52:207-211.
  40. Magzoub MM, Stephens HA, Sachs JA, Biro PA, Cutbush S, Wu Z, Bottazzo GF. HLA-DP polymorphism in Sudanese controls and patients with insulin-dependent diabetes mellitus. Tissue Antigens 1992; 40:64-68.
  41. Marsh SG. 2000. Nomenclature for factors of the HLA system, update September 2000, WHO Nomenclature Committee for Factors of the HLA System. Tissue Antigens 2000; 56:565-566.
  42. Marsh SG. 1998. Nomenclature for factors of the HLA system, update March 1998. WHO Nomenclature Committee for Factors of the HLA System. Tissue Antigens 1998; 51:681-683.
  43. May J, Mockenhaupt FP, Loliger CC, Ademowo GO, Falusi AG, Jenisch S, Dippmann K, Schnittger L, Kremsner PG, Bienzle U, Meyer CG. HLA DPA1/DPB1 genotype and haplotype frequencies, and linkage disequilibria in Nigeria, Liberia, and Gabon. Tissue Antigens 1998;52:199-207.
  44. McTernan CL, Mijovic CH, Cockram CS, Barnett AH. The nucleotide sequence of two new DP alleles, DPA1*02015 and DPB1*8401, identified in a Chinese subject. GenBank accession # AF077015, 2001.
  45. Moonsamy PV, Suraj VC, Bugawan TL, Saiki RK, Stoneking M, Roudier J, Magzoub MM, Hill AV, Begovich AB. Genetic diversity within the HLA class II region: ten new DPB1 alleles and their population distribution. Tissue Antigens 1992;40:153-157.
  46. Nishimaki K, Kawamura T, Inada H, Yagawa K, Nose Y, Nabeya N, Isshiki G, Tatsumi N, Niihira S. HLA DPB1*0201 gene confers disease susceptibility in Japanese with childhood onset type I diabetes, independent of HLA-DR and DQ genotypes. Diabetes Res Clin Pract 2000;47:49-55.
  47. Poulton KV, Kennedy LJ, Ross J, Thomson W, Mbanya JC, Ollier WE. A study of HLA-DPB1 phenotypes reveals DPB1*6301 in a rural population from Cameroon. Eur J Immunogenet 1998;25:375-377.
  48. Rani R, Sood A, Lazaro AM, Stastny P. Associations of MHC class II alleles with insulin-dependent diabetes mellitus (IDDM) in patients from North India. Hum Immunol 1999;60:524-531.
  49. Ravikumar M, Dheenadhayalan V, Rajaram K, Lakshmi SS, Kumaran PP, Paramasivan CN, Balakrishnan K, Pitchappan RM. Associations of HLA-DRB1, DQB1 and DPB1 alleles with pulmonary tuberculosis in south India. Tuber Lung Dis 1999;79:309-317.
  50. Rozmuller EH, van der Zwan AW, Voorter CE, Tilanus MG. DPB1*8501, a novel variant in the US Black population. Tissue Antigens 2000;56:282-284.
  51. Steiner LL, Wu J, Noreen HJ, Moehlenkamp C, Cavalli A, Davidson M, Johnson S, Winden T, Segall M, Begovich AB, Williams TM. Four new alleles identified in a study of 500 unrelated bone marrow donor-recipient pairs. Tissue Antigens 1999;53:201-206.
  52. Varney MD, Tait BD. Identification of a novel DPB1* allele (DPB1*6901) and the occurrence of HLA haplotypes that extend to DPB1. Tissue Antigens 2000;55:188-190.
  53. Voorter CE, Tilanus MG, van den Berg-Loonen EM. Two new HLA DPB1 alleles identified by sequence-based typing: DPB1*8201 and DPB1*8301. Tissue Antigens 2000; 56:560-562.
  54. Wang FQ, al-Daccak R, JulY, Lethielleux P, Charron D, Loiseau P, Colombani J. HLA-DP distribution in Shanghai Chinese–a study by polymerase chain reaction–restriction fragment length polymorphism. Hum Immunol 1992;33:129-132.
  55. Weston A, Ensey J, Kreiss K, Keshava C, McCanlies EC. Racial differences in prevalence of a supratypic HLA-genetic marker, immaterial to pre-employment testing for chronic beryllium disease. Am J Ind Med 2000;41:457465.
  56. Mouse Genome Informatics. The Jackson Laboratory
  57. Medical Economics Company. 1998. Physicians Desk Reference, 52nd Edition, Medical Economics Company Inc. pp1797-1799.
  58. Van Damme K, Casteleyn L, Heseltine E, et al. Individuals susceptibility and prevention of occupational diseases: Scientific and ethical issues. JOEM 1995; 37:91-99.
  59. Khoury MJ, Newill CA, Chase GA. Epidemiologic evaluation of screening for risk factors: application to genetic screening. Am J Pub Health 1985; 75:1204-1208.
  60. Hanson, EH, Imperatore G, Burke W. HFE gene and hereditary hemochormatosis: A HuGE review. Am J Epi;154:193-206

Address correspondence to Dr. Khoury at
Office of Genomics and Disease Prevention
Centers for Disease Control and Prevention
6 Executive Park, Mail Stop E-82
Atlanta, Georgia 30329