Nasal Toxicity and Dosimetry of Inhaled Xenobiotics: Implications for Human Health Sponsors Agency for Toxic Substances and Disease Registry Basic Acrylic Monomer Manufacturers Association DuPont Haskell Laboratory European Centre for Ecotoxicology and Toxicology of Chemicals Methacrylate Producers Association National Institute of Environmental Health Sciences National Institute for Occupational Safety and Health Rohm and Haas Company Society of the Plastics Industry, Inc. U.S. Environmental Protection Agency Organizing Committee F. Miller, CIIT, chair G. Boorman, NIEHS C. Frederick, Rohm and Haas Co. J. Haartz, NIOSH J. Harkema, Inhalation Toxicology Research Institute J. Kimbell, CIIT N. Krivanek, DuPont R. McClellan, CIIT K. Morgan, CIIT J. Morris, University of Connecticutt P. Scherer, University of Pennsylvania C. Shoaf, U.S. EPA

A joint meeting between industry, policy makers, and academic investigators to discuss the implications for human health of nasal toxicity and dosimetry of inhaled xenobiotics was held in Durham, North Carolina, 20-22 September 1993. The conference was hosted by the Chemical Industry Institute of Toxicology (CIIT).

The nose is usually the first contact in the respiratory tract for many airborne chemicals of environmental and occupational concern. There is increasing interest in nasal toxicity and dosimetry of inhaled chemicals. This interest stems not only from the basic toxicological issues but also from the implications of the research for assessing potential human risks. The broad goals of the meeting were to assess the state of knowledge concerning nasal toxicity in humans and animals, to examine issues encountered in using animal data to assess human risk, and to identify areas of research that will improve our ability to make interspecies comparisons of nasal toxicity and dosimetry of inhaled chemicals in the future. Some of the more specific goals of the meeting were to discuss the toxic responses observed in the nasal passages and determine which lesions should be considered significant. The questions of severity of lesions at different levels and of adaptive response versus an adverse lesion were also addressed. The state of the art in dosimetry modeling was presented to toxicologists with the focus of discussion on what new data are needed to move modeling to the next level in understanding and extrapolating nasal toxicity responses to humans.

A major issue for toxicology studies is extrapolation of results from the animal species to humans and from higher exposure concentrations often used in animal studies to human exposures at concentrations several orders of magnitude lower. It was obvious from the presentations at this conference that this complex issue of extrapolation is even more difficult for the nasal passages, which differ in structure and function in rodents from those in humans. Most animals are obligate nose breathers, and patterns of breathing are also different among species. The conference ended with specific case studies used for setting human exposure limits. The cases were specifically chosen from different agencies to highlight the differences and similarities used by policy makers dealing with varying exposure populations (i.e., occupational exposures versus environmental exposures). The focus was on what important issues have surfaced in establishing exposure limits thus far.

With increasing knowledge and scientific complexity and with decreasing research dollars, it is crucial to work together to make sound scientific decisions to protect public health, as well as to make better and safer new products and technologies to continue to improve our lives. Extrapolation is estimating what is not readily observed from that which can be directly observed. While the primary concern is overt disease, with improved scientific technology, much smaller alterations can now be detected in test animals, but when do these alterations represent an adverse effect? Understanding the mechanism will help to separate an adverse effect from an adaptive response. In the future, with limited resources, policy makers will need to work closer with industry and academic toxicologists to determine what data are needed for sound policy decisions.

The mathematical approach has been used for modeling nasal cavity exposure. The advantages of the modeling came from simulating complex behavior of the flow of the gases through the nasal passage and allowing better visualization of complex processes of where the chemical or toxic metabolite is deposited. Mathematical modeling can help provide a basis for extrapolating from short- to long-term exposure, from high to low concentrations, and between different species with widely varying nasal structures and air flow patterns. The availability of supporting toxicology data is crucial in the validation and further development of any mathematical model.

The complexity of the nose continued to be highlighted. For example, variations in the thickness of the mucous layer, the pH of the layer, plus mucous and airflow patterns result in different patterns of toxicity. There are multiple models for estimating tissue buildup of toxic materials, and each model has disadvantages and advantages depending on the question asked. The model is used to visualize "hot spots" in the nose where toxicity may be expected. By using plastic models and fluid flow to simulate air flow in the nasal passages, it was interesting to note that rodents and primates have different flow patterns, but computer models correlated well with the flow of colored dyes in aqueous models. An enlarged physical model (20 times larger than actual size) and a finite element computer model of the human nose was constructed based on CAT scans of patients. This work compares well with the modeling efforts in both rodent and nonhuman primate noses. For this field to progress, toxicology data are needed to refine the models. Further study of the boundary conditions at the air/mucous interface and the mucous/tissue/blood interface in the nasal cavity is needed.

Nasal uptake of vapor was measured in a variety of laboratory animal species. The data revealed that uptake of acetone, a nonreactive, nonmetabolized vapor, was more efficient in the rat (Sprague-Dawley, F344) or mouse (B6C3F₁) nose than in that of the hamster or guinea pig. These studies suggest that nasal tissue perfusion rate, which may vary between species, influences uptake of nonreactive, nonmetabolized vapors. Uptake of directly reactive and/or metabolized vapors displayed differing and variable species or strain differences, suggesting that species differences in metabolic potential may be important and should be considered when extrapolating animal nasal toxicity data to humans.

Based on ¹⁸O isotope studies, about 14% of inhaled ozone is retained in the rat's head. ¹⁸O uptake in both humans and rats correlates with ozone toxicity. Thus, this model accurately predicts ozone toxicity across species and model systems. The toxic responses in animals and risk assessment of ethyl acrylate, a polymer used in latex paint and in a wide variety of other products, was discussed. Nasal toxicity is an important toxic effect of ethyl acrylate. Air flow, blood flow in the nasal cavity, and glutathione levels in the different nasal tissues are important in determining the location and severity of nasal toxicity in animal models. Formaldehyde dosimetry predictions for site-specific toxicity were presented, which correlated well with the occurrence of neoplasms in the rat nasal passage. DNA protein cross-linking was shown to be a good biological marker of potential toxicity and eventual tumor response for this chemical. Cell proliferation was also increased in the areas where toxicity was found. The toxicity, cross-linking, and cell proliferation were all found at 6 and 15 ppm, but not at lower concentrations, and this nonlinearity has implications for risk assessment at low formaldehyde concentrations.

The second day of the conference focused on the nasal lesions found in humans and animals. The distribution of nasal lesion by the type of epithelium involved was discussed. For most species there are four distinct epithelial cell types: 1) squamous epithelium that lines the nasal vestibule, 2) transitional (cuboidal, nonciliated) epithelium that occurs between the squamous and respiratory epithelium, especially on the lateral wall, 3) respiratory epithelium that covers the majority of the nasal passage, and 4) olfactory epithelium that occurs dorsally and caudally over the ethmoid turbinate. Based on the airflow patterns, the primary location for lesions of the respiratory epithelium is the mid-ventral septum and the middle turbinate, which correlates well with predictions by computer models. It was stressed that it is crucial for the toxicological pathologist to use care in recording lesions. The diagnosis "hyperplasia, nasal cavity" is not helpful for the policy maker and is meaningless for scientists trying to advance modeling systems to predict nasal toxicity. Thus, lesions need to be recorded carefully by site, severity, level of nasal cavity, type of epithelium involved, etc.

The different epithelial types in the nasal airways were discussed. Regional responses to inhaled toxicants could be quite different. For example, short-term ozone exposure induces cell replication in transitional epithelium with little proliferative response in other nasal epithelia. The acute effects (2 weeks) of ozone at low concentrations (0.12 ppm) were restricted to the anterior portion of the nasal passages, but chronic exposures (20 months, 5 days/ week, 6 hr/day) caused alterations in both anterior and posterior aspects of the nasal airways. In rats at 0.12 ppm, there was little or no effect, while at 0.5 and 1 ppm clear effects could be found in F344/N rats after 20 of months exposure. Ozone induces an increase in the mucous goblet cells in the transitional epithelium in both rats and nonhuman primates. This raises the question of whether this is an adaptive response or whether it should be considered an adverse response to ozone exposure. These structural alterations have been correlated with functional alterations (i.e., mucostasis).

Toxicity to the olfactory mucosa was the next topic. This epithelium has metabolic capacities and is susceptible to toxicity, especially from compounds that are metabolized to a toxic metabolite. Some chemicals given systemically can cause olfactory mucosa toxicity. For example, phenacetin given orally to rats will result in microcysts in the olfactory epithelium. Methyl bromide is another compound that given orally can impair the olfactory senses. Dibasic esters caused toxicity of the sustentacular cell of the olfactory epithelium. Female rats were more sensitive to this effect than males.

Chemical toxicity in the accessory nasal structures was also presented. The nasal passages contain numerous glands and structures that are often overlooked in toxicity studies. The glands of the nasal passage include Bowman's glands, Steno's glands, septal glands, and turbinate glands. The vomeronasal organ is important in the sense of smell and is related to sexual activity. There are numerous examples of toxicity; in some cases, toxicity may be more severe in the glands than in other tissues. For example, with exposure to the tobacco-specific nitrosoamine 4-(N-methyl-N-nitrosamino)-1-(3-pyridyl)-1-butanone, there is severe toxicity in Steno's glands.

The olfactory epithelium was described as a "window to the brain" and noted that neurological changes for Alzheimer patients can be found in nasal biopsies of the olfactory epithelium. This was discovered subsequent to the clinical observation that Alzheimer patients quickly lose their sense of smell.

An overview of nasal lesions found in clinical practice showed that the diversity of lesions and different types of exposure patterns are broad. Patients are often exposed at home and at the workplace. Some occupations, such as furniture making or wood working, are associated with an increased incidence of nasal tumors in the UK and other countries.

The third day dealt with implications of chemically induced nasal toxicity for setting exposure limit values. Both toxicologists and scientists at government agencies recognize that there are ranges of human responses, that there are complex dose- response relationships, that lesions may vary in their severity, persistence, and homology, and that some lesions found in animal studies may not represent adverse responses in humans. Policy makers need to play a more active role in deciding the type and level of data that are needed for sound decisions. It was suggested that policy makers help decide what improvements in our knowledge base during the next decade would provide a sounder scientific basis for regulatory decisions.

Exposure limit regulations include a number of uncertainty factors in the calculations. The number of uncertainty factors may be reduced if there are more and better scientific studies; uncertainty factors that remain will be better defined. Human data are preferred for setting limits but are often not available. There were some questions about the possible redundancy of different regulatory agencies working on setting limits for the same chemical with different criteria and uncertainty factors. Some differences are related to the particular application, i.e., higher but shorter workplace exposure as opposed to lower environmental exposures that may extend for a life-time.

There is increasing concern about chemically sensitive populations and exposures to multiple chemicals when setting exposure limits. Some exposures to multiple chemicals are voluntary (e.g., smoking). The endpoints from these exposures vary widely in incidence and severity; nasal discharge may occur at a high incidence but is not severe, whereas nasal cancer is a rare (but important) event.

The approaches that various agencies use in assessing human health risks and establishing exposure limits were discussed. Two examples of inhalation reference concentrations that have nasal endpoints were presented.

In summary, modelers and toxicologists need to be in touch with the policy makers to be sure that their data and research direction will be helpful for making regulatory decisions. Policy makers are now more responsive to the scientific community and have access to better science upon which to base policy decisions.