BERYLLIUM

In 1998-2000, NIOSH and the beryllium producer conducted surveys of current workers at the company’s three main facilities: the primary production plant, which produces beryllium metal, various alloys, and beryllium oxide powder (Schuler 2011); the ceramics plant, which produces beryllium oxide ceramics from oxide powder ( Henneberger 2001 ); and the copper-beryllium alloy plant, where further processing of rod and wire and strip materials takes place ( Schuler 2005 ). In addition, NIOSH surveyed former workers who had participated in earlier surveys at the primary production plant in 1993-94 ( Kreiss 1997 ) and the ceramics plant in 1992 ( Kreiss 1996 ). Results from these surveys have demonstrated that engineering controls put in place at the highest risk operations in the mid- to late 1990s were not sufficient to reduce risk of sensitization and CBD ( Henneberger 2001 , Schuler 2012); that sensitization and CBD can occur under conditions of low beryllium exposure ( Henneberger 2001 , Schuler 2005 ); and that the true risk to beryllium-exposed workers has been underestimated by reliance on results from cross-sectional surveys (Schuler 2008).

Evaluation of exposure-response relations in these populations is ongoing. Earlier efforts at these facilities and elsewhere demonstrated no consistent exposure-response relationship. However, process-related risks have been identified in the absence of exposure-response (Kreiss 1996, Kreiss 1997, Henneberger 2001, Schuler 2005, Kreiss 2007), suggesting that predictive exposure factors do exist but may not have been adequately characterized. Possible explanations may include: inclusion of irrelevant exposure due to uncertainty in ascertaining time of onset of sensitization or CBD in cross-sectional epidemiologic studies, use of airborne exposure metrics that do not address biological relevance rather than metrics that consider lung deposition (McCawley 2001, Stefaniak 2003 , Stefaniak 2011 , Virji 2011 ), and lack of consideration of alternate pathways of exposure ( Day 2007 ), among others. We have created individual exposure estimates for one group of workers from extensive databases of full-shift personal lapel samples of total mass concentration of airborne beryllium (Virji 2011, Park 2012). In addition to total mass, we are investigating the role of particle size, chemical form of beryllium, material solubility ( Stefaniak 2006 , Stefaniak 2011 ), and potential dermal exposure ( Day 2006 ) in risk for sensitization and CBD.

Investigators in the Respiratory Health Division and the Health Effects Laboratory are currently working on four research projects that were funded by the National Occupational Research Agenda (listed below).

In response to the results of multiple epidemiologic surveys at the company's facilities, and especially in light of the failure of targeted engineering controls to reduce sensitization and CBD, the company began to implement a comprehensive preventive program in 2000-01. This program was designed to keep beryllium out of the lungs, off the skin, and at the work process, among other points, and emphasized a well-trained and prepared workforce. To that point, no preventive programs had demonstrated effectiveness in preventing beryllium sensitization and CBD among beryllium-exposed workers. NIOSH and the company evaluated the early effectiveness of this program by comparing medical surveillance data from groups of workers hired after the preventive program began to similar groups of workers hired just prior to the program. Early results were promising, showing substantial reductions in sensitization at the beryllium ceramics plant (Cummings 2007), the alloy finishing plant (Thomas 2009), and the primary production plant (Bailey 2010). The summary paper, over a follow up period of nine years and including many former workers who had left employment, documented continued efficacy since the beginning of the preventive program (Thomas 2013). The early years of the program were less protective of workers than the later years, consistent with the gradual implementation of this cultural health and safety change. Not all beryllium sensitization and CBD were prevented in workers hired after the beginning of the preventive program, but no worker developed CBD after leaving employment in that follow up. In 2016, NIOSH began working on a project to evaluate prevalence of sensitization and CBD in workers of a new pebble plant that began operations in late 2011.

In the 1950s, Curtis and co-workers demonstrated that beryllium-naïve research participants could be sensitized to beryllium by skin patch testing. This experimental information was not incorporated into industrial hygiene practice over succeeding decades until the major beryllium company recognized that lowering of beryllium air concentrations had not resulted in prevention of beryllium sensitization early in employment. In support of the company’s intention to protect skin from beryllium exposure in the comprehensive beryllium protection program initiated in 2000, previous NIOSH researcher Sally Tinkle demonstrated that particles less than one micrometer in diameter could enter intact human skin with mechanical motion and that mice could be sensitized to beryllium salts by application on their ear skin (Tinkle 2003). Studies to estimate relative historical skin contact in various processes by beryllium content of cotton gloves over vinyl gloves were conducted by Greg Day for use in epidemiologic studies of exposure-response (Day 2006). This work was continued by Jenna Armstrong to evaluate correlations among skin, surface, and air concentrations of beryllium (Armstrong 2014). Contaminated surfaces, like tools and equipment, represent potential sources for both inhalation and skin exposures. Settled dust on surfaces can be re-suspended in air and/or transferred to workers’ clothing and skin. Consideration of migration of beryllium within facilities can focus exposure measurement and control efforts in the riskiest areas (Day 2007; Armstrong 2014; Armstrong 2014 abstract ).

The historical confusion about exposure-response relations for beryllium might be a reflection of differing properties of beryllium particles and compounds. Aleks Stefaniak and coworkers have investigated the influence of particle size, chemistry, surface area, and solubility in biological fluids, all of which influence bioavailability (Stefaniak 2003; Stefaniak 2004; Day 2005; Stefaniak 2005; Stefaniak 2007; Stefaniak 2008; Stefaniak 2008; Virji 2011; Stefaniak 2012). Small particles can get deeper into the lung than large particles. Characterization of particle size by process-related risk showed that respirable beryllium concentrations were highest for the high risk processes in oxide production and pebbles plant (Virji 2011).

The chemistry of beryllium metal, oxide, ore, and alloys has historically been associated with different workplace risks. An example is the difference in toxicity of low-fired and high-fired beryllium oxide in ceramics with respect to chronic beryllium disease. NIOSH researchers conceptualized this potential differing chemistry as depending on surface area and solubility. Dr. Stefaniak addressed surface area of different beryllium materials by morphometrically examining surface area and measuring it (Stefaniak 2003 ). He evaluated solubility by creating artificial lung alveolar macrophage phagolysosomal fluid (Stefaniak 2005; Stefaniak 2011a), artificial sweat (Stefaniak 2011b), and artificial airway epithelial lining fluid (Stefaniak 2012), With these tools, he and coworkers examined process materials (Stefaniak 2008), mineral ores (Duling 2012 ), and a variety of metals for dermal bioaccessibility (Stefaniak 2014). With beryllium as a paradigm, Dr. Stefaniak has contributed to better understanding of the lung hazards of skin exposure in the work environment (Stefaniak 2011c ).

The immune response reflected in sensitization is likely related to the ease of beryllium ions being recognized by immune cells, i.e. T lymphocytes. The disease CBD is likely related to persistence of beryllium ions in tissue from poorly soluble forms of beryllium. In fact, the explanation of the two different lung diseases historically associated with beryllium, acute beryllium disease and chronic beryllium disease, lies in the high solubility and excretion of beryllium fluoride salts in the production of beryllium metal, resulting in a reversible acute lung disease when the beryllium is eliminated from the body (Cummings 2009). In contrast, chronic beryllium disease occurs in work settings in which exposure to poorly soluble and persistent beryllium compounds predominate, such as beryllium metal, oxide, and hydroxide.  Some of these physicochemical characteristics have been linked with exposure data and health outcome; others are still being prepared for publication. Although this interdisciplinary work of industrial hygienists, physical scientists, and epidemiologists has had application in the preventive program introduced by the primary producer, proposed regulation has not differentiated among beryllium compounds or characteristics. Nevertheless, worker and management understanding of this NIOSH research can benefit common sense approaches to worker risk.

No natural animal gets CBD. It is important to have an animal model for laboratory experiments, since in order to validate dose-response information derived from job-exposure matrices and epidemiologic studies as well as being able to answer other basic toxicological questions. Genetic studies have led to the creation of an animal model where different human HLA-DP alleles have been inserted into mice; this is called a “transgenic” or “knock-in” model. The NIOSH model comprises three different strains of this genetically engineered mouse: low risk, where the transgene codes for a lysine residue at the 69th position of the B-chain; medium risk, where the transgene codes for a glutamic acid residue at the 69th position of the B-chain and glycine residues at positions 84 and 85; and high risk, where the transgene codes for glutamic acid at the 69th position of the B-chain and aspartic acid and glutamic acid residues at positions 84 and 85, respectively (Tarentino-Hutchinson et al. 2009 ). This mouse model has been used to compare: different types of beryllium (alloy versus metal versus oxide versus salts), different doses of beryllium, and routes of exposure (skin versus lung), although the results have not yet been published. In the meantime, researchers at the University of Colorado Health Sciences Center have used a medium risk mouse model that was independently derived to make breakthrough discoveries about the mechanism of beryllium recognition and sensitization (Mack et al. 2014). Beryllium ions bind to the binding groove of the macrophage without requiring binding to a hapten (carrier) serum protein. The high risk knock-in mouse molecule has a more electronegative surface of the binding groove that presumably promotes beryllium binding and stability. Any number of proteins can then bind to the beryllium lodged deep in the binding groove and thereby change the surface confirmation of the T lymphocyte. This changed confirmation is recognized as a foreign substance by the T cell, similar to what happens in auto-immune disease. Now that a basic immunologic mechanism has been identified, mouse models can be used to test interventions in the sensitized mice and new therapeutic approaches to chronic beryllium disease.

NIOSH has also conducted genetic analyses in research independent of the company.  At the initiation of the genetic research, NIOSH planned to verify the importance of the genetic super-allelic marker associated with CBD, which was a glutamic acid in the 69^th position of the HLA-DP1 molecule, one of the immunoresponse genes on chromosome 6. Indeed, we confirmed the importance of this glutamic acid marker which conferred a 3.3-fold risk of beryllium sensitization and a 9.4-fold risk of chronic beryllium disease in the beryllium worker population in three plants (McCanlies 2004; Weston 2005). After completing the analyses for glutamic acid in the 69^th position, NIOSH researchers developed more detailed methods for characterizing genotypes that shared this common glutamic acid marker. About 40% of the worker population had this genetic marker, but overall prevalence of beryllium sensitization and disease were low. With continuing analyses, we found that a subset of rare alleles with highly electronegative surfaces were much higher risk than other alleles with less electronegative surfaces, all of which had a glutamic acid in the 69^th position (Weston 2005 ; Snyder 2003 ; Snyder 2008). We contributed our deidentified genotype information to a multi-center research group at National Jewish Health in Denver, which substantiated our findings that rare alleles with negative 9 surface charge were associated with high risk of beryllium sensitization and disease (Silveira 2012).

Because the beryllium workers used in the case-control study conducted by Silveira and co-workers were derived from case series from several different workplaces, the odds ratio results by genotype were not readily applied in practice by either researchers, corporate medical directors, or beryllium workers. NIOSH was unique among the contributors to the multi-center case-control study in having population-based data on three workforces in the beryllium industry. With information on the majority beryllium-exposed workers in the three workforces, NIOSH calculated prevalence of beryllium sensitization and disease by individual genotype. This approach resulted in the observation that some genotypes resulted in more workers having beryllium health effects than other genotypes. For example, two-thirds of the workers who had inherited the genotype HLA-DP1*1701/*0201 had chronic beryllium disease at the time that their blood was drawn more than a decade earlier (Kreiss 2016).

Although the molecular biologists who designed the genetic analyses in 1999 were motivated to create a transgenic mouse model with human genes based on epidemiologic data, another motivation was to examine whether genetically more susceptible workers were more likely to develop beryllium sensitization and beryllium disease at lower air concentrations of beryllium in the workplace. Historical beryllium exposure estimates were eventually created for workers who began work since 1993, and these will be used in ongoing research to evaluate exposure-response in these workers, adjusted for genetic marker status.

Beryllium Research Highlights is a series of newsletters written for current and former workers in the beryllium production industry who participated in our beryllium research program with the primary beryllium producer in the U.S. These newsletters provide information to study participants about completed studies, current research findings, and upcoming activities. Though dated 2003-2008, they may have utility for communication to an industrial workforce.