• E-mail
• Print
• Comment
• Font Size
• Digg
• del.icio.us
• Discuss article

Carcinogenicity Studies of Diesel Engine Exhausts in Laboratory Animals: A Review of Past Studies and a Discussion of Future Research Needs

Posted on: Sunday, 31 July 2005, 03:00 CDT

Diesel engines play a vital role in world economy, especially in transportation. Exhaust from traditional diesel engines using high- sulfur fuel contains high concentrations of respirable carbonaceous particles with absorbed organic compounds. Recognition that some of these compounds are mutagenic has raised concern for the cancer- causing potential of diesel exhaust exposure. Extensive research addressing this issue has been conducted during the last three decades. This critical review is offered to facilitate an updated assessment of the carcinogenicity of diesel exhaust and to provide a rationale for future animal research of new diesel technology. Life- span bioassays in rats, mice, and Syrian hamsters demonstrated that chronic inhalation of high concentrations of diesel exhaust caused lung tumors in rats but not in mice or Syrian hamsters. In 1989, the International Agency for Research on Cancer (IARC) characterized the rat findings as "sufficient evidence of animal carcinogenicity," and, with "limited" evidence from epidemiological studies, classified diesel exhaust Category 2A, a "probable human carcinogen." Subsequent research has shown that similar chronic high concentration exposure to particulate matter generally considered innocuous (such as carbon black and titanium dioxide) also caused lung tumors in rats. Thus, in 2002, the U.S. Environmental Protection Agency (EPA) concluded that the findings in the rats should not be used to characterize the cancer hazard or quantify the cancer risk of diesel exhaust. Concurrent with the conduct of the health effects studies, progressively more stringent standards have been promulgated for diesel exhaust particles and NO^sub x^. Engine manufacturers have responded with new technology diesel (improved engines, fuel injection, fuels, lubricants, and exhaust treatments) to meet the standards. This review concludes with an outline of research to evaluate the health effects of the new technology, research that is consistent with recommendations included in the U.S. EPA 2002 health assessment document. When this research has been completed, it will be appropriate for IARC to evaluate the potential cancer hazard of the new technology diesel.

Keywords Animal Bioassay, Cancer Diesel, Engine Exhaust, Inhalation Exposure, Eung Cancer

INTRODUCTION

Scope of This Review

In this review, we are concerned with the evidence from laboratory animal inhalation studies of the potential carcinogenicity of exhausts from combustion engines, primarily those fueled by diesel, but also those fueled by gasoline, or liquid or compressed natural gas (LNG or CNG). (A key to the acronyms used in this article is provided in Appendix 1) Traditional diesel engine exhaust (TDE) has been studied much more extensively than exhaust from engines fueled by gasoline or other fuels. Hence, the majority of the present review deals with TDE. As used in this review, TDE refers to the exhaust from diesel technology typical of the engines, fuel injection systems, fuels, and lubricants in use prior to 1988, when exhaust emissions were not regulated.

The International Agency for Research on Cancer (IARC) in 1988 reviewed the results of research on TDE and placed whole diesel exhaust in Category 2A, "probably carcinogenic to humans." The IARC category was based on a finding of (a) "sufficient animal evidence" of carcinogenicity of diesel exhaust, (b)"limited human evidence," and (c) supporting evidence from an array of biochemical and cellular studies (IARC, 1989). As will be reviewed, it is now generally accepted that the positive rat findings represent a species-specific, protracted high exposure concentration effect not relevant to the evaluation of human carcinogenicity. This article recommends that IARC reevaluate its 1988 classification of TDE based on the new mechanistic evidence.

Based on concerns for the health effects of diesel exhaust, substantial and continuing effort has been directed to reducing emissions, especially of particulate matter and NOx. New technology diesel has remarkably lower exhaust particle emissions compared to the older technology. The new technology diesel exhaust (NTDE) is just beginning to be evaluated as to its potential to cause adverse health effects. In this review, the term NTDE refers to the diesel exhaust from new integrated systems (engines, fuel injection systems, low-sulfur fuels, lubricants, and exhaust after treatment) intended for introduction to meet stringent new U.S. Environmental Protection Agency (EPA) standards for particulate matter and nitrous oxides (NOx) emission standards to be effective in 2007 (U.S. EPA, 2002). In the present article, we outline an approach for conducting new inhalation studies on NTDE. When the new research on NTDE is completed, we recommend that IARC conduct an independent evaluation of the potential carcinogenicity of the new technology that is separate from the evaluation of TDE.

This article includes the following topics in subsequent sections: historical perspectives, pre-1988 studies, post-1988 studies, studies of NTDE, carcinogenicity of other types of engine exhaust, and recommendations for future studies and evaluations. The present review is limited to the potential of engine exhausts (diesel and gasoline, as well as other types of engines) to cause pulmonary tumors and other effects that may lead to or are closely associated with pulmonary tumorigenicity, including inflammation, interstitial fibrosis, hyperplasia, and mutagenicity. Effects relating to asthma, hypersensitivity, and other health issues are beyond the scope of this article.

Historical Development of Internal Combustion Engines

Internal combustion engines were developed just over a century ago. The spark ignition engine fueled by gasoline was invented in 1876. Rudolph Diesel invented the diesel engine in 1893 using a concept of high combustion-chamber pressure to burn hydrocarbon fuel in the absence of a spark ignition source. Both types of engines soon gained popularity, leading to a revolution in the transportation sector and other areas of industry requiring compact power sources. Applications of the two types of internal combustion engines eventually split, especially post World War II: Gasoline- fueled spark ignition engines were used extensively to power passenger vehicles; diesel engines soon dominated the heavy-duty truck market and became the power source of choice for modern railroad locomotives and for offroad applications such as agricultural and construction activities.

The fuel efficiency of diesel engines indirectly had major impact on the course of diesel engine use. In the 1970s, shortages of petroleum resulted in great enthusiasm for increased use of diesel engines in passenger cars. In the United States, there were predictions that within a short period of time, a significant percentage of passenger vehicles would be diesel powered (NRC/NAS, 1981,1982,1983; Watson et al., 1988). Similar predictions were made for Europe and other markets, where diesel-fueled passenger vehicles were already in wide use. The predictions of increased use of diesel engines stimulated research to evaluate the potential health hazards of diesel exhaust, with special attention directed to the particulate-matter (PM) components.

Over the years, advances in diesel technology have resulted in progressive reductions in diesel engine emissions (Figures 1 and 2) (U.S. EPA, 2002). By 1988, advances in diesel technology had resulted in a 40% reduction in PM emissions compared to the earlier unregulated engines (Figure 1). In 1994, further advances in diesel technology (e.g., electronic engine controls) resulted in a 90% reduction in PM emissions. Diesel engines have undergone similar reductions in nitrogen oxides (NOx) emissions (Figure 2). In 2007 much more stringent engine emissions standards promulgated by the U.S. EPA and the California Air Resources Board (CARB) will go into effect, which will require reductions in PM and NOx emissions by approximately 99% from the pre-1980 levels (CaIEPA, 1998; U.S. EPA 2002).

To comply with the 2007 emissions standards, the diesel industry has developed new technology diesel (NTD) engines, which consist of an integrated system of advanced engine design, ultra-low-sulfur fuel, specialized lubricants, and a catalyzed particulate trap filter (International, 2002). Emissions from in-use diesel engines (roughly those manufactured after 1994) can be dramatically reduced by retrofitting the engine with the new catalyzed particulate filters and running them on ultra-low-sulfur fuel. Prototype new technology diesel (NTD) engines have been developed to comply with the 2007 regulations and to be introduced into the marketplace by 2007 (NTD engines are currently available for limited marketplaces and applications such as in California school buses). New technology diesel emissions (NTDE) are as clean as or cleaner than emissions from engines fueled by either gasoline or natural gas (see School Bus Emission Study, later). The advances in diesel technology for 2007 are far-reaching to the extent that NTDE does not resemble TDE- especially not pre-1988 DE, which was studied in all published chronic inhalation studies of DE. Thus, as we elaborate later, NTDE should be e\valuated separately from TDE for potential tumorigenicity and other health effects.

FIG. 1. Reductions in diesel participate matter emissions in the United States. U.S. EPA standards for particulate emissions from heavy duty diesel trucks (t) or urban buses (ub), calculated as grams particulate matter emitted per brake-horsepower-hour (g/b hp- h) and adjusted relative to pre-1998 unregulated engine emissions. From U.S. EPA Health Assessment Document for Diesel Engine Exhaust, May 2002 (Table 2:4, p. 2-16).

FIG. 2. Reductions in diesel nitrogen oxide (NO^sub x^) emissions in the United States. U.S. EPA standards for NO^sub x^ emissions from heavy duty diesel engines, calculated as grams particulate matter emitted per brake-horsepower-hour (g/b hp-h). From U.S. EPA Health Assessment Document for Diesel Engine Exhaust, May 2002 (Table 2:4, p. 2:16).

Composition of Engine Emissions

Prior to commencing a review of the carcinogenicity research, a brief summary of the composition of traditional engine emissions used in past studies is needed. An excellent overview of combustion engine emissions is provided in Johnson (1988). We return to this topic in a later section, which discusses a study that compared the composition of emissions from NTD (also referred to as low-emitting diesel, traditional diesel, and natural gas engines).

Whole diesel exhaust includes three basic components: elemental carbon particles in respirable clusters that are less than 1 m in diameter (Cheng et al., 1984); organic matter adsorbed onto the surface of the carbon particles, which is readily extractable with organic solvents; and a mixture of gas and vapor phases that include volatile organic compounds (VOC) (Scheutzle, 1983; Scheutzle et al., 1985; Scheutzle and Frazier, 1986; Scheutzle and Lewtas, 1986; Johnson, 1988). Both the extractable organic matter and the VOC components include known bacterial and mammalian cell mulagcns (Huisingh el al., 1978; Clark et al., 1981, 1984; Claxton, 1983; Brooks et al., 1984; Lewis et al., 1986).

TDE is much higher in particulate matter than emissions from gasoline, CNG, or NTD. Gasoline engine exhaust (GE) is higher in carbon monoxide and heat content than DE, which limits the concentration of fresh GE to which laboratory animals can be exposed without causing adverse effects from these two factors.

The composition and concentrations of various components of engine exhaust are affected by a number of factors, among which are the type, age, and condition of the engine, conditions under which it is operated (high or low load, warm or cold operating conditions, etc.), the type of fuel and fuel additives, and exhaust filters or cleaners. In this review, no effort has been made to address the great variability that undoubtedly existed in the chemical compositions of the engine exhausts to which the laboratory animals were exposed in past studies. Even within a category of engines, such as diesel or gasoline, the chemical and physical variability of the exhausts studied in the past was likely considerable.

Historical Development of Research on Carcinogenicity of Engine Exhausts

The potential for combustion engine exhaust to produce carcinogenic responses was not a major topic of investigation prior to attention being focused on diesel exhaust. Primary concern for engine exhaust emissions had focused on noncancer effects. To the extent that there was concern that combustion engines emissions might cause cancer, it was based on the presence in engine exhaust of numerous organic compounds, including specifically polynuclear aromatic hydrocarbons.

The concern that diesel exhaust might produce health effects was prompted by the respirable size (less than 1 μ,m diameter) of the carbonaceous particles with absorbed organic compounds. One of the new tools used in the new research initiative was the Ames test, a bacterial mutagenicity assay developed by Professor Bruce Ames and his colleagues at the University of California-Berkeley (Ames et al., 1975). The assay used different tester strains of mutated Salmonella to evaluate the potential for test agents to back mutate the bacteria. In the case of diesel exhaust, the test materials were organic compounds extracted from collected diesel exhaust particles using powerful organic solvents. In 1978, Huisingh and colleagues, working in a U.S. EPA laboratory, found that extracts of diesel exhaust particles were mutagenic (Huisingh et al., 1978). The observation of Huisingh et al. that extracts of diesel exhaust were mutagenic was not unexpected, recognizing that in 1955 Kotin and colleagues had reported that extracts of diesel exhaust particles contained polycyclic aromatic hydrocarbons (PAHs), which when painted on mouse skin produced skin cancer (Kotin et al, 1955), presumably because the particles contained mutagenic constituents. As an aside, Kotin and colleagues had earlier observed that extracts of particles collected from Los Angeles city air and gasoline engine exhaust contained PAHs, which when painted on mouse skin produced skin cancer (Kotin et al., 1954a, 1954b).

This finding led to an internal U.S. EPA communication to treat diesel exhaust as a carcinogen (U.S. EPA, 1977). This finding was subsequently noted in the New York Times (1977), ensuring that the observation became widely known. Within a short period of time an international research effort involving both the public and private sector was mounted to evaluate the carcinogenicity of diesel exhaust. Indeed, the scope of the research program on diesel exhaust may exceed that directed to any single agent other than radiation. A committee of the National Research Council (NRC) was convened to evaluate the early evidence on the health impacts of increased use of diesel technology (NRC/NAS, 1981). The results of the research program were reported in three major symposia (Pepelko et al., 1980; Lewtas, 1982; Ishinishi et al., 1986a) and hundreds of peer- reviewed publications. A related symposium proceedings on particle overload in the rat lung and lung cancer (Mauderly and McCunney, 1996) will also be of interest to readers. Ultimately, most of the research findings would be reviewed in three documents: one prepared by the Health Effects Institute (HEI, 1995), the second by the California Environmental Protection Agency (CalEPA, 1998), and the third by the U.S. EPA (2002). Other comprehensive reviews of the state of knowledge of the possible health effects of diesel exhaust have been published at various time points (McClellan, 1987; Watson et al., 1988; HEI, 1995; WHO, 1996; CaIEPA, 1998; Mauderly, 2000).

In the United States, most of the research was carried out with support from two government agencies, the U.S. EPA and what was to become the U.S. Department of Energy, and two major automotive companies, General Motors and Ford. The research program of the Health Effects Institute, a public/private partnership created in 1980 and sponsored jointly by the U.S. EPA and the manufacturers of combustion engines, from its origin to the present time has been substantially influenced by the concern over diesel engine exhaust.

The U.S. EPA conducted most of its research in-house and focused on the use of a matrix approach to evaluate the carcinogenicity of diesel exhaust (Albert, 1994). This approach made substantial use of observations of the mutagenicity and shortterm predictors of carcinogenicity for extracts from diesel exhaust particles from both light- and heavy-duty engines, gasoline engine exhaust particles, and extracts of related known human carcinogens (cigarette smoke, roofing tar, and coke oven emissions) to estimate the carcinogenicity of diesel engine exhaust (Cuddihy et al., 1982; Albert, 1983). The animal studies conducted by the U.S. EPA have been reviewed elsewhere (Pepelko andPeirano, 1983).

The Department of Energy research program was carried out largely by the Eovelace organization in Albuquerque, NM. The Lovelace program was broad-based and included studies using a 5.7-L diesel engine produced by General Motors. The program included the evaluation of the chemical composition of DE, the disposition of inhaled diesel exhaust particles, and endpoints ranging from mutagenicity to life-span cancer bioassays to functional impairment. The General Motors research program was likewise very broad in scope and made substantial use of studies in a number of laboratory animal species exposed to emissions from a 5.7-L diesel engine. The Ford Motor Company research program emphasized evaluation of the chemical characteristics of diesel exhaust and associations with mutagenicity.

In Europe, research was carried out largely with support from diesel engine manufacturers. At the Fraunhofer Institute in Germany, life-span cancer bioassays using a 1.6-L diesel engine and a 1.66-L gasoline engine were conducted in rats, Syrian hamsters, and mice, with support from the Volkswagen and Daimler-Benz Companies (Heinrich et al., 1986b, 1995). Battelle Memorial Institute (Geneva, Switzerland), with support from the European Automobile Manufacturers Association, also conducted major studies in which rats and Syrian hamsters were exposed to exhaust from diesel and gasoline engines (Brightwell et al., 1986, 1989). The Battelle studies, patterned after the Fraunhofer studies, included exposures to both whole DE and DE from which the particles were removed.

In Japan, an extensive array of studies was conducted at the Japanese Automobile Research Institute (JARI). The JARI animal exposure facility was built specifically for the conduct of these studies and included participation from many Japanese academic scientists. Studies were also conducted in Japan at other institutions. The JARI studies used two different diesel engines, a light-duty (1.8 L) and a heavy-duty (11 L), and included studies with both rats and mice (Ishinishi et al., 1986b; Iwai et al., 1986, 1997).

Historical D\evelopment of Carcinogen Classification Schemes

To provide a contextual setting for considering the results of the animal bioassays of carcinogenicity, we first review the various schemes used to classify agents or processes with regard to human carcinogenic potential.

The potential for exposure to specific chemicals and emissions from various technologies to produce adverse health effects has been recognized for some time. Early attention focused on functional disorders of organ systems such as respiratory or neurological disease. After World War II, increased attention began to be given to cancer. This attention was spurred in part by a steady increase in the incidence of lung cancer, which is now known to have been due almost exclusively to an increase in cigarette smoking that began decades earlier. At that time, a few industrial processes, a few specific chemicals, and radiation had been identified as carcinogenic to humans largely by epidemiological methods. Research to identify other potential carcinogens, such as that of Kotin and colleagues noted earlier, was just beginning to be conducted. Most importantly, epidemiological evidence of the linkage between cigarette smoking and lung cancer had not yet emerged.

Along with concern for cancer came increased research to identify methods for the detection of carcinogens. This included the development of animal bioassay methods with administration of test agents via ingestion, inhalation, instillation, implantation, or skin painting. Research was also conducted to develop methods for identifying mutagens based on the assumption that mutagenesis was an obligatory step in the overall process of carcinogenesis (Ames et al., 1975).

Public and scientific concern for cancer led to a call for government action at the local, state, national, and international level. One obvious need was for a system to interpret the available scientific evidence and classify agents or processes as to their potential to cause cancer in humans. In 1969, a new international organization, the International Agency for Research on Cancer (IARC), an agency of the World Health Organization, took the lead in developing a carcinogen classification scheme. The IARC scheme is shown in Table 1 (for more detailed explanation of the IARC carcinogenicity classification process, see the "Preamble to Evaluation" at the beginning of any IARC monograph, e.g., Vol. 84; IARC, 2004). It gives the greatest weight to epidemiological findings, since such evidence is clearly the most relevant to identifying human carcinogens. In the absence of adequate positive human cancer findings, weight is given to observations in laboratory animals. Finally, consideration is given to supporting data such as mutagenicity or metabolic properties of an agent that might give insight into the agent's carcinogenic potential.

In 1992, IARC modified its classification scheme to explicitly recognize that mechanistic information could be used to either upgrade or downgrade the classification of an agent (for example, upgrade from 2a to 1 or downgrade, from 2b to 3) (IARC, 1992). The first instance in which IARC used mechanistic data to downgrade or upgrade the classification of an agent involved ethylene oxide (IARC, 1994). In this case, IARC determined that the epidemiological evidence for carcinogenicity of ethylene oxide was limited and the animal data were sufficient evidence of carcinogenicity. Pre-1992, this lead to classification of ethylene oxide in Category 2a, a probable human carcinogen. Based on mechanistic data, IARC in 1993 reclassified ethylene oxide as Category 1, a human carcinogen.

In another example of an IARC downgrade based on mechanism, IARC in 2001 reevaluated the carcinogenicity of synthetic vitreous fibers (SVFs, e.g., fiber glass and its relatives; IARC, 2001). Previously, IARC had designated all SVF wools as 2B, possibly carcinogenic. However, in 2001, IARC decided to divide the SVF wools into two groups, biodurable and biosoluble. The 2B designation was retained for durable ultra-thin specialty glass fibers, such as 475 microfiber. The biosoluble SVFs (e.g., standard insulation glass wool) were designated as IARC category 3, not classifiable as to carcinogenicity. The IARC decision to divide SVFs into two categories with respect to carcinogenicity classification (category 3 for biosoluble vs. category 2B for biodurable) was significantly influenced by the mechanistic data, showing the importance of fiber biodurability in producing a carcinogenic response.

The U.S. Environmental Protection Agency (EPA) initially adopted a carcinogen classification scheme patterned after the IARC scheme (U.S. EPA, 1986). Indeed, the original U.S. EPA scheme was virtually identical to the IARC scheme except that the U.S. EPA used an alphanumeric classification while IARC used a numeric-alpha scheme. Almost immediately after releasing its original guidelines, the U.S. EPA began revising its guidelines for carcinogen risk assessment and the associated classification scheme. The major changes have been directed at increasing the use of "mechanistic" data (i.e., that might clarify how the agent causes its carcinogenic effect at the molecular, cellular, and/or organ level), providing alternative methods of extrapolating cancer findings from observations in humans or laboratory animals to much lower likely levels of human exposure, and adoption of a more flexible "narrative" approach (i.e., classification is supported by short descriptive text statements rather than by using a few simple categories as in the IARC scheme) (Table 1) for describing the conclusions of the carcinogen classification process (U.S. EPA, 1999).

TABLE 1

Carcinogen evaluation scheme, International Agency for Research on Cancer (IARC)

A third carcinogen classification scheme has been used in the United States by the National Toxicology Program (NTP) in preparing its periodic reports on carcinogens. The NTP scheme determines whether an agent is a human carcinogen or reasonably anticipated to be a human carcinogen; or after evaluation the NTP may determine that the agent should not be listed in either of these categories. Like the IARC and U.S. EPA schemes, the NTP modified its process in 1996 to facilitate the use of mechanistic data (NTP, 1996). Mechanistic data can now be used to either upgrade or downgrade the classification of an agent. For example, it is now possible in the NTP scheme, based on mechanistic information, to classify an agent as a human carcinogen in the absence of actual cancer findings in humans or laboratory animals.

IARC Cancer Classification of Engine Exhausts

In 1988, IARC convened an international panel of independent experts to review all published data relating to the potential carcinogenicity of engine exhausts. At that time, only diesel and gasoline engines were of concern; engines using other types of fuels, such as methanol or compressed natural gas, were not yet sufficiently developed, and no data were available on the possible health effects of these types of exhausts. After reviewing human, animal, and in vitro research findings, the panel classified diesel engine exhaust (DE) as Group 2A, probably carcinogenic to humans, and gasoline engine exhaust (GE) as Group 2B, possibly carcinogenic to humans (IARC, 1989). For DE (whole exhaust), IARC found the evidence for carcinogenicity to be sufficient in experimental animals but limited in humans. However, for GE (whole exhaust), IARC found the evidence for carcinogenicity to be in inadequate in both animals and humans.

U.S. Cancer Classification of Diesel Exhaust

In the early 1990s, The U.S. EPA initiated a review of the research findings on DE carcinogenicity and other health effects. The review was completed and published in 2002 (U.S. EPA, 2002). In the 2002 document (Chap. 7.5.6, p. 7-141), the U.S. EPA concluded that diesel exhaust is a likely human carcinogen. The accompanying narrative statement is provided in Table 2. The U.S. EPA has not evaluated the evidence for GE as a potential human carcinogen.

The U.S. National Toxicology Program first listed DE as "reasonably anticipated to be a human carcinogen" in 2000 in its ninth Report on Carcinogens and reaffirmed this finding in 2002 in its tenth Report on Carcinogens (NTP, 2000, 2002). The NTP has not evaluated the evidence for GE as a potential human carcinogen.

TABLE 2

Characterization of overall weight of evidence for carcinogenicity of diesel engine exhaust, U.S. Environmental Protection Agency

Chronic Inhalation Bioassays

For laboratory studies of the potential health effects of aerosols, the traditional gold standard is the animal inhalation bioassay, in which rodents or other mammals are exposed by inhalation to the same or similar airborne substance of concern for human exposure (NTP, 2002). In a typical chronic inhalation bioassay, 50 or more laboratory rodents (rats, mice, hamsters, or guinea pigs) of each gender are exposed to an aerosol in a manner that simulates human occupational exposure (i.e., for 68 h/day, 5 days/week). Aerosol concentrations are higher than those in human occupational settings (to increase the possibility of seeing a low- level positive effect) but below that which would cause a nonspecific effect from overload of defense mechanisms. If possible, nose-only exposure (with rats or mice, using specialized exposure chambers) is superior to whole-body exposure because it ensures delivery of the agent to the inhaled air stream. Ideally, because carcinogenic effects are often latent, animals should be exposed for lifetime. During the course of the study, animals not assigned to the cancer bioassay are sacrificed at predetermined times and evaluated for respiratory pathology and, if studying particulate aerosols, lung burden of inhaled particles. The use of inhalation exposure is advantageous in that it takes into account the respirability of the test mat\erial as well as the defenses of the respiratory system in the whole animal.

The IARC 1988 finding of "sufficient" animal evidence for carcinogenicity of DE rested primarily on a series of chronic inhalation studies in rats exposed to pre-1988 TDE (engines manufactured prior to emissions regulations) (Table 3). Only a few chronic studies have been conducted on DE since then. The designs for the pre-1988 chronic inhalation studies of DE conducted in a half-dozen different laboratories around the world had some common features as well as important differences. Even though engines and fuel types varied, the experimental exposure concentrations were consistently reported as mass concentration of particulate matter (mg/m^sup 3^). Aerodynamic diameters of particles, when reported, suggest limited variability in this parameter. Dilution factors of emission to air were also generally reported for both DE and GE studies (Tables 3-6).

All of the chronic inhalation studies reported here used whole- body exposure to fresh engine exhaust diluted with ambient room air. However, exposure regimes were extremely variable in terms of hours per day, days per week, and months of duration. The studies did not consistently report the light-dark cycle used and the relationship of the light cycle to exposure times.

The exposure chambers varied in size from a fraction of a cubic meter to walk-in rooms. In some studies, animals were individually housed, while in other studies animals were group housed. Some studies included concurrent exposure of two or more species. Some studies included comparison of the effects of engine exhausts with the effects of other particles such as titanium dioxide, coal dust, or carbon black. In addition to pulmonary pathology and histopathology, some studies also evaluated multiple endpoints: for example, inflammatory indicators in bronchoalveolar lavage fluid; lung function; lung burden of carbonaceous particles; lung clearance of carbonaceous particles and/or radiolabeled reference particles; and gross and histopathological evaluations. None of the studies reported numbers of particles per cubic centimeter of air. To facilitate comparison of studies, some authors have used a normalized weekly exposure metric (mg/m3 * h/week) or the average exposure concentration per week (normalized weekly exposure divided by 168 h/week, expressed as g/m^sup 3^).

Studies Using Noninhalation Routes of Exposure

Animal studies using nonphysiological routes of exposure, such as intratracheal instillation (IT) or intrapulmonary implantation, should be clearly distinguished from those that use inhalation exposure. Studies in which animals are exposed to extracted components of engine exhaust by IT have limited value for assessing the human health risks of engine exhausts for several reasons. In the case of engine exhausts, a primary limitation of these studies is the relevance of the test substance, which must be collected as a solid and a vapor from the whole exhaust, recombined, then suspended or dissolved in a delivery vehicle. Thus, the test substance would not be fresh and it would be chemically and physically different from the fresh, whole, airborne exhaust, to which people would be exposed. Another major concern is that the IT process bypasses normal respiratory defense mechanisms. Furthermore, instillation can deliver large particles or clusters of particles to the lower lung that would not become airborne or would not be able to travel to the lower lung in the inspired air stream. Also, IT can result in particles being nonuniformly distributed in the lung, resulting in some lung regions having very high doses of particles that could overload the natural defenses. Some have argued that injection studies may provide useful information on the mechanisms of toxicity of particles; however, we have concerns as to the relevance of the mechanisms to inhalation exposure. With these caveats, some IT studies have also been included here.

TABLE 3

Chronic inhalation studies in rats with traditional diesel engine exhaust

TABLE 3

Chronic inhalation studies in rats with traditional diesel engine exhaust

TABLE 4

Chronic inhalation studies in various species with traditional diesel engine exhaust

TABLE 5

Chronic inhalation studies in animals with gasoline engine exhaust

However, intraperitoneal injection studies, in which the test substance is injected directly into the abdominal cavity of the laboratory animal, are not included in this article because, in addition to the problems cited already for IT, the intraperitoneal injection also does not involve the appropriate target organs (the respiratory tract) and affords little or no opportunity for macrophage-mediated clearance, which is a major protective mechanism in the lung.

Epidemiology

The U.S. EPA Health Assessment Document (HAD) (U.S. EPA, 2002) contains the most authoritative, recent review of the human epidemiology evidence for DE carcinogenicity. The epidemiology evidence basically consists of three occupational studies: railroad workers (Garschick et al., 1987, 1988; Woskie et al., 1988a, 1988b; Vermaetal., 1999), truck drivers (Steenland et al., 1990, 1992, 1998; Zaebst et al., 1991; Whittaker et al., 1999) and underground miners (Slavin, 2001). In the HAD, the U.S. EPA based its health risk assessments of TDE primarily on the studies of railroad workers and truck drivers. However, both sets of data have serious weaknesses that make them unsuitable for cancer risk assessment: The studies lack contemporaneous measurements of exposures to TDE; reconstruction of exposure history of workers was problematic; and other exposures (e.g., gasoline exhaust, cigarette smoke) were not adequately taken into account (Bunn et al., 2004). It should be noted that there has not been and still isn't an accepted marker for measuring occupational exposure to DE. In the third occupational exposure, underground miners, for whom TDE exposure is much higher than for railroad workers and truck drivers, there was no increase in lung cancer. These concerns prompted the U.S. EPA, to conclude that, while TDE was a "likely" carcinogen, a unit risk value or range of risk cannot be calculated from existing data and that the risk could be zero (U.S. EPA, 2002).

Thus, if there is a causal role for DE in the lung cancer risks for railroad workers and truckers, then the lung cancer incidence for underground miners should be very high, given the more than 10- fold higher level of TDE exposure for miners compared with truckers and railroad workers. However, underground miners do not show a lung cancer excess associated with TDE; in fact, no clear difference can be seen in lung cancer rates between highly exposed underground miners and lightly exposed aboveground miners. We therefore conclude that the published epidemiology data on traditional diesel exhaust does not provide persuasive evidence of carcinogenicity in humans.

CARCINOGENIClTY STUDIES OF TRADITIONAL DIESEL ENGINE EXHAUST

Pre-1988 Studies

The 1988 IARC classification of DE as Group 2A, probable human carcinogen, was based to a large extent on animal studies, which the IARC panel deemed "sufficient" evidence of carcinogenicity in animals (IARC, 1989). The foundation of the animal evidence was a series of animal chronic inhalation studies in which rodents and other mammals were exposed by inhalation (whole-body) to DE for at least 1 year and typically for lifetime. The numbers of chronic inhalation studies of DE reviewed by IARC included: rats, 7; mice, 5; hamsters, 3; monkeys, 1; cats 1 (Tables 3-5). Additionally, IARC considered as supporting animal evidence of DE carcinogenicity several subchronic inhalation studies and several studies in which rodents were exposed to suspensions of DE soot by intratracheal instillation or intrapulmonary implantation. A summary of each of these studies follows.

Chronic Inhalation Studies in Rats

Eight rat chronic inhalation studies of DE were reported between 1981 and 1988, each from a different laboratory (Table 3). Five of these studies were positive for lung tumors. Three studies were negative for tumorigenicity and/or were questionable due to technical limitations.

Positive Rat Studies

The five pre-1988 positive rat chronic inhalation studies include two from Japan (Ishinishi et al., 1986a, 1986b; Iwai et al., 1986; Ishihara, 1988), one from the Fraunhofer Institute (Heinrich et al, 1986a), one from Battelle-Europe (Brightwell et al., 1986, 1989), and one from the United States (Mauderly et al., 1986, 1987; Lovelace Institute). Findings from all five studies were reported at the International Symposium held in Japan (Ishinishi et al., 1986a), and more detailed results were published later (Table 3).

In the primary Japanese study, groups of approximately 120 male and female F344 rats were exposed to exhausts from lightduty (1.8 L) or heavy-duty (11.0 L) diesel engines at one of a range of concentrations (in mg/m^sup 3^: 0.11-2.3 for light-duty; 0.5-3.7 for heavy-duty) for 16 h/day, 6 days/week, for up to 30 months (Ishinishi et al., 1986b; Ishihara, 1988). The incidence of lung tumors (adenomas and carcinomas) was significantly increased (8/ 124; 6.5%) only in the group exposed to the highest concentration of heavy-duty diesel exhaust (3.7 mg/m^sup 3^). The second highest heavy-duty exposure (1.84 mg/m^sup 3^) also had an elevated incidence that was not statistically significant (4/123; 3.3%). Controls exposed to filtered air had a tumor incidence of 1/123 (0.8%). The IARC Working Group noted that "although this incidence was not statistically different from that in the controls, it suggested an overall positive response for the two highest exposure levels." The Working Group further noted that the light duty DE groups also showed nonsignificant elevations in tumor incidences, but their control group had a higher incidence than that of the heavy-duty control group (3.3% v 0.\1%) (IARC, 1989). In view of the current summation of historical background lung tumor incidences in F344 rats (up to 4%), these elevations do not seem striking.

In another study conducted in Japan, Iwai et al. exposed 24 F344 female rats 8 h/day, 6 days/week, for 24 months to 5 mg/m^sup 3^ DE or to an equivalent dilution of filtered DE (no particles; from a 2.4-L engine) or to air (Iwai et al., 1986). Rats were then maintained for 6 additional months of recovery. Tumor incidences in surviving rats were significantly elevated only in rats exposed to whole DE. Tumor incidences for DE, filtered DE, or air were 9/19 (42%), 0/16 (0%), or 1/22 (4.5%), respectively.

In the German study, 96 Wistar female rats were exposed for 19 h/ day, 5 days/week, for 2| years, to 4.2 mg/m3 DE or to an equivalent dilution of filtered DE (Heinrich et al., 1982, 1986b, 1986c). Only the rats exposed to whole DE showed an increased incidence of pulmonary tumors: 9/95 (9.5%) adenomas/carcinomas plus an additional 8/95 (8.4%) benign keratinizing cysts. In contrast, tumor incidences were 0/96 for air-exposed controls and 0/92 for filtered DE. Furthermore, only the whole DE rats had significant bronchoalveolar hyperplasia (99%) and metaplasia (62%), reduced lung clearance of particles, reduced lung function, and increased inflammatory parameters in the bronchoalveolar lavage fluid (BALF; 8- to 10-fold increase in lactose dehydrogenase, alkaline phosphate, total protein, and protease, as well as increased cell counts).

The study conducted at the Battelle Institute in Geneva, Switzerland (Brightwell et al., 1986, 1989) exposed 72 male and 72 female Fischer 344 rats (and hamsters, described later) to DE 16 h/ day, 5 days/week, for 2 years to DE at 0.7, 2.2, or 6.6 mg/m^sup 3^ or to filtered DE at equivalent concentrations. The rats were maintained without further exposure for an additional 6month postexposure period. Filtered DE and low-concentration whole DE did not result in elevated incidences of tumors. The mid and high concentrations resulted in significantly increased lung tumor incidences, which were markedly higher in females than in males (96% and 44%, respectively, for the high dose).

The major study conducted in the United States at Lovelace Institute (Albuquerque, NM) exposed 230 male and female F344 rats (as well as mice, reported in a later section of this article) 7 h/ day, 5 days/week, for up to 30 months to 0.35, 3.5, or 7.1 mg/m^sup 3^ DE from a 5.7-L engine (Mauderly et al., 1986; Wolff et al., 1986, 1987,1989; Mauderlyetal., 1987; Henderson et al., 1988). The DE paniculate contained 12% extractable organic matter by mass. Survival and body weight were unaffected by exposure.

Pulmonary tumor incidences for controls and the low, medium, and high concentrations of DE were 0.9%, 1.3%, 2.8%, and 7.9%, respectively (combined incidences for adenoma, adenocarcinoma, and squamous-cell carcinoma). The mid and high (3.5 and 7.1 mg/m^sup 3^) exposures, but not the low exposure, resulted in tumor incidences that were significantly elevated over controls. Both the mid- and high-exposure rats also had benign squamous cysts (0.9% and 4.9%, respectively), dose-dependent interstitial lung fibrosis, and proliferative lung disease.

The Lovelace study reported a significant relationship between tumor prevalence and both exposure concentration and soot lung burden. In the mid- and high-exposure rats, two pulmonary inflammatory indicators in the bronchoalveolar lung fluid (BALF) were elevated and correlated with the presence of pulmonary fibrosis (Henderson et al., 1988). In contrast, BALF factors and histology in rats exposed to the lowest DE concentration (0.35 mg/m^sup 3^) did not differ from the air control rats. Notable in this study is the observation of a threshold of DE exposure-between 0.35 mg/m^sup 3^, the no-observed-adverse effect level (NOAEL), and 3.5 mg/m3-below which carcinogenicity, inflammation, and fibrosis were not observed in the rat. The authors point out, "This threshold was well above environmentally relevant levels of diesel exhaust but may be in the range of some occupational exposures" (Henderson et al., 1988). As an aside, if more exposure levels had been studied between 0.35 and 3.5 mg/m^sup 3^, a higher NOAEL might have been observed.

In the Lovelace rat study, the investigators concluded that a high degree of caution should be exercised in extrapolating the lung tumor findings from rats exposed to high levels of DE to people exposed to lower levels of DE (Mauderly et al., 1986). They noted that the lung burden data suggested an overload of the lung clearance mechanisms at the medium and high exposure levels. For example, the authors point out, after 24 months of exposure, the ratio of soot in the lung (mg) to soot exposure concentration (mg/ m^sup 3^) was 1.7 for the low exposure but 3.3 and 3.0 for the medium and high exposures, respectively.

The concern over lung clearance impairment resulting from chronic inhalation of high levels of particles of low toxicity (i.e., DE) in the Lovelace study was further examined by Wolff et al. (1986, 1987, 1989). Rats (F344 male and female) for this study were exposed concurrently with the animals in the carcinogenicity study to DE at 0, 0.35, 3.5, or 7 mg/m^sup 3^ (control, low, medium, or high; C, L, M, H) for 7 h/day, 5 days/week, for up to 30 m. Lung burdens were evaluated at various time points up to 24 m. Clearance of each of two inhaled non-DE radiolabeled particles (^sup 67^Ga^sub 2^O^sup - ^^sub 3^ or ^sup 134^Cs-fused aluminosilicate particles) was determined after various DE exposure periods up to 24 months. Depositions of radiolabeled particles were similar in C, L, M, and H groups at the various DE-exposure time points. Low exposure (0.35 mg/ m^sup 3^) resulted in very low DE burden and did not significantly impair clearance of radiolabeled tracer particles (Figures 3 and 4). However, M and H exposures resulted in lung burdens of DE soot that were greater than expected (disproportionately greater than burdens with low exposure) and did impair lung clearance of tracer particles (Figures 3 and 4).

FIG. 3. Lung clearance of radiolabeled particles after 24 months of exposure to diesel exhaust at low (0.35 mg/m^sup 3^), medium (3.5 mg/m^sup 3^), or high (7.0 mg/m^sup 3^) concentrations. From Wolff et al. (1987).

For rats exposed to the medium and high concentrations of DE, lung burden simulation models fit the actual lung burdens of diesel soot only if the models assumed that clearance of DE soot was normal for several months and then completely stopped thereafter (Figure 4). Such a model was supported by the finding that clearance of radioactive tracer particles was impaired only after several months exposure to mid and high concentrations of DE, when lung burdens of DE soot were high. In contrast, exposure to particles of high toxicity (nonradioactive Ga^sub 2^O^sub 3^) impaired clearance of tracer particles after a relatively brief exposure, when lung burdens were still low. The authors concluded that "Particle clearance impairment fi.e., lung overload] should be considered both in the design of chronic exposures of laboratory animals to inhaled particles and in extrapolating the results to people" (Wolff et al., 1989). Thus, in an animal inhalation study, to avoid a nonspecific particle effect (i.e., an effect that would result from a very high lung burden of any inert particles), exposure concentrations must be kept below the level that overloads and impairs normal lung- clearance mechanisms.

FIG. 4. Lung burden of carbonaceous material in lungs of rats chronically exposed to diesel exhaust. Adapted from Wolff et al. (1987).

The findings on high exposure concentrations of particles altering lung clearance were noted by National Toxicology Program (NTP) personnel. Recognizing the need for guidance in conducting future inhalation studies with airborne particles, the NTP convened a panel of experts to review the available data and publish a guidance document for selecting aerosol exposure concentrations (Lewis et al., 1989).

Oberdorster expanded on the concepts in the NTP guidance and identified the maximum lung dose that would not impair lung clearance as the maximum tolerated dose (MTD) (Oberdorster, 1997). The exposure concentration of insoluble particulates that resulted in the MTD was viewed as the highest appropriate exposure level for an animal cancer bioassay. The mid- and high-DE exposures in the Lovelace study apparently exceeded the MTD. However, it is to be noted that these views were not completely developed until after the 1988 IARC decision.

Negative Rat Studies

In studies conducted at Battelle Northwest Laboratories in the U.S. groups of 6 male Wistar rats were exposed to 8 mg/m^sup 3^ DE, 7 mg/m^sup 3^ coal dust (CD), 15 mg/m^sup 3^ CD, 8 mg/m^sup 3^ DE plus 6 mg/m^sup 3^ CD (DE/CD, to simulate exposure in underground coal mining), or clean air, for 6 h/day, 7 days/week, for 20 m (Karagianes et al., 1981). DE soot particles were considerably smaller than CD particles; mean mass aerodynamic diameters were 0.7 m (95% respirable) for DE and 2.1 m (50% respirable) for CD. The authors reported negative tumor results. Two lung tumors (adenomas) were observed, one in 6 rats exposed to DE and one in 6 rats exposed to DE/CD. The authors stated that the lung tumors "may be exposure- related, although their occurrence was not statistically significant in this study." The authors further noted that, in spite of "ample gross and microscopic evidence that soot and coal dust were deposited in the rats' lungs," there were no significant effects on body weight, mortality, or hematologic parameters in any of the exposed groups. However, lung lesions found in all particleexposed groups (but not in controls) were similar to those found in human coal workers with pneumoconiosis: particle deposition, macrophage accumulation, alveolar-cel\l hypertrophy, interstitial fibrosis, and emphysema. The interpretation of this study is limited by the small numbers of rats exposed for the full 20 months (6 per group) and no postexposure recovery time to take into account latent tumor development.

A study from Japan (Takemoto et al., 1986), also reported negative tumor results. In this study, 26 female F344 rats were exposed 4 h/day, 4 days/week, for life to 2-4 mg/m^sup 3^ DE with average paniculate diameter of 0.32 m. Because a number of rats were sacrificed for evaluations at early time points, a major limitation of this study was the small numbers of rats exposed 18 months or more. Other limitations were short daily and weekly exposure periods and no postexposure recovery period.

The third negative study was conducted in U.S. EPA facilities in Cincinnati, OH, with support from the National Institute of Occupational Safety and Health (NIOSH) (Lewis et al., 1986, 1989). They exposed more than 1000 male and female F344 rats to DE, coal dust (CD), or DE/CD (50:50) at 2 mg/m^sup 3^ for 7 h/day, 5 days/ week, 144 animals were exposed for up to 24 months. The 2 mg/m^sup 3^ DE was achieved by a 1:27 dilution of DE to air (DE from a 7.0-L engine). The 2-mg/m^sup 3^ exposure level and the combined DE/CD exposure were an effort to simulate permissible or recommended airborne concentrations for underground mining operations more realistically than previous studies. CD particles were much larger than those of DE; mass median diameters were 8.6 m for CD (50% of particles were <5 m) in contrast to 0.36 m for DE. At least 10 rats from each group were sacrificed after 3, 6, 12, and 24 months of exposure and evaluated for pathology and pulmonary histopathology. Other rats were evaluated for a variety of parameters.

Lewis et al. observed no increase in pulmonary tumor incidence in any of the particle-exposed groups; combined incidence of pulmonary adenomas and carcinomas was 4% for the air control group as wells as for the DE, CD, and DE/CD groups. No exposure-related effects were observed in body weight gain, lymphocyte chromosomal aberrations, bone-marrow micronucleus test, or lung function tests. Rats exposed to DE actually showed more rapid lung clearance of inhaled ^sup 59^Fe^sub 3^O^sub 4^ particles than control rats. In all particle- exposed rats, microscopic examination of tissues revealed similar black deposits throughout the lungs, on the pleural lung surfaces, and in the peribronchial lymphoid tissue, and collections of macrophages containing black pigmented particles associated with hyperplasia. The authors noted that these effects tended to be more pronounced with DE than with CD exposure, but quantitative data were not reported. Lung burdens of inhaled particles for the three particle-exposed groups were similar, with CD tending to be a little higher than DE or DE/CD. Lung burdens as percent of dry lung weight after 12 and 24 months of exposure, respectively, were: for DE, 0.52% and 0.91%; for CD, 0.91% and 1.15%; for DE/CD, 0.63% and 0.84%. The lack of a large increase in lung burden over time suggests that lung clearance mechanisms were still functioning (i.e., lungs were not overloaded) at these lung doses.

The mutagenicity of solvent extracts of DE paniculate and CD in bacteria were also evaluated in the NIOSH study. Solvent extracts of DE paniculate were significantly mutagenic in Salmonella typhimurium strains TA98 and TA100. Numbers of revenants in exposed cultures compared to control cultures were: for DE, 5- to 6-fold greater; for DE/CD, 3- to 4-fold greater; for CD, not significantly greater. Neither DE nor CD, separately or in combination, increased incidence of micronuclei or sister chromatid exchanges in rat bone-marrow cells or lymphocytes, respectively.

An additional long-term study of DE conducted at the Southwest Research Institute with support from GM is worthy of mention (although it was not included in the 1989 IARC review). A/J mice, guinea pigs, and F344 rats were exposed to DE from a 5.7-L engine at concentrations of 0.25, 0.75, or 1.50 mg/m^sup 3^ for 20 h/day, 5.5 days/week, for 9 or 15 months. Additional groups of rats were'similarly exposed for 15 months and then maintained for an additional 8 months in clean air. Mice had no increase in tumor incidence, and results with guinea pigs were not meaningful. In total, 5 bronchoalveolar carcinomas were observed in the rats, but only observed in groups exposed 15 months and maintained an additional 8 months. Tumor incidences in these rats for exposure levels of 0.25, 0.75, or 1.50 mg/m^sup 3^ were 1/30, 3/30, and 1/ 30, respectively. No tumors were observed in the 30 control rats maintained in clean air for 23 months. No tumors were observed in 180 rats exposed to diesel for either 9 or 15 months and then killed or 60 controls for these groups-a total of 270 rats without any pulmonary tumors. However, the tumor findings are not dose dependent and are not significantly increased over the historical background tumor incidence in aging F344 rats (0-7%) (Schreck et al., 1981; Kaplan et al., 1983; White et al., 1983).

Chronic Inhalation Studies in Mice

Prior to 1988, three laboratories investigated the tumorigenic potential of DE following chronic inhalation of high concentrations in five different strains of mice (Table 4). Tumorigenicity findings from these studies were somewhat variable; the overall indication was that DE inhaled chronically at high concentrations was not consistently tumorigenic in mice. It was not until after the IARC 1988 decision that an in-depth evaluation of findings in mice provided strong support for this conclusion (Mauderly et al., 1996; see next section, Post-1988 Studies).

The U.S. EPA conducted a series of studies in mice, in which male and female Strain A (and some SENCAR) mice were exposed 8 h/day, 7 days/week, for 9 to 11 months to 6 or 12 mg/m^sup 3^ from a 3.2-L engine (Pepelko, 1982;PepelkoandPeirano, 1983). In the first study, 400 male mice that were exposed to 6 mg/m^sup 3^ DE actually had a lower tumor incidence than control mice exposed to clean air. In the second study, 85 female mice similarly exposed to DE and sacrificed after 9 months showed a tumor incidence "slightly but not significantly greater" than that in clean-air controls. In a subsequent series of 4 studies, both male and female mice were exposed to DE at 12 mg/m^sup 3^ for 9 or 12 months with clearly negative results, as seen in Table 4. In each case, the DE mice had a lower tumor incidence than the air controls. Also noteworthy were several additional observations: Background lung tumor incidence was very high for Strain A (at least 24% in the air controls); tumorigenicity in the air controls doubled from 9 months to 12 months (24-29% at 9 months to 58% at 12 months); and injection with 5 mg urethane (a tumor initiator) also greatly increased the tumor rate in both control and DE mice. With such a high background tumor rate in this strain, the value of a tumor initiator seems dubious. Finally, any question of possible nonsignificant positive results among the female mice in the earlier study is completely negated by the last set of studies. (These studies were briefly mentioned but not reviewed in the IARC monograph [IARC, 1989].)

In Japan, concurrent with the rat study discussed earlier, a study was conducted with mice: 60ICR and 150 C57BL6N mice were exposed 4 h/day, 4 days/week, for life (up to 28 months) to 2-4 mg/ m^sup 3^ of DE (0.27-L engine) or to air (Takemoto et al., 1986). Accumulations of particles in the lung increased with exposure duration. The authors reported that "there was a tendency for the diesel-exposed mice to have higher incidences of lung tumors compared with non-exposed mice." Tumor incidences were: in ICR mice, 14/56 (25%) for DE versus 7/60 (12%) for controls; in C57 mice, 17/ 150 (11.3%) for DE versus 1/51 (2%) for controls. While the authors found the differences to be nonsignificant for both strains, the IARC Working Group found the elevated incidence among the C57 mice to be significant (IARC, 1989).

In a study conducted at the Fraunhofer Institute in Germany, concurrent with the rat study discussed earlier, 192 female NMRI mice were exposed 19 h/day, 5 days/week, for 28 months to 4.2 mg/ m^sup 3^ whole DE or an equivalent amount of filtered DE (1:17 dilution of diesel to air)(similar exposure of rats and hamsters in the Fraunhofer study is reviewed in this article, earlier and later, respectively) (Heinrich et al., 1986b). Both whole and filtered DE resulted in similar elevated rates of total lung tumors (31-32%; adenomas, adenocarcinomas, and carcinomas) compared with 13% in concurrent air-exposed controls. Only the incidence of malignant tumors (adenocarcinomas and carcinomas) was significantly elevated in the DE-exposed mice: For air, filtered DE, and whole DE, malignant tumor rates were 2.4%, 11%, and 15%, respectively (adenocarcinomas and carcinomas). However, the IARC Working Group noted that previous control groups of NMRI mice in this laboratory had background tumor rates as high as 32%. Other exposure-related effects were observed only in mice exposed to whole DE: decreased body weight during the second year of exposure; bronchoalveolar hyperplasia (64% of animals exposed to whole DE, compared with 15% for filtered DE and 5% for controls); and interstitial fibrosis in 43% of animals (negligible for filtered DE and controls). The authors note that the exposure concentration of 4.2 mg/m^sup 3^ was achieved by diluting the tailpipe exhaust with air by a factor of 17 and that this concentration is about 250 times greater than that typical of urban street canyons. In regard to their findings on the carcinogenic effects of DE in mice, the authors note, "These findings, although significant, need confirmation."

Chronic Inhalation Studies With Hamsters

Three chronic inhalation studies of DE using Syrian golden h\amsters were reported prior to 1988; all were negative for tumorigenicity. Heinrich et al. exposed 48 female hamsters, 7 h/ day, 5 days/week, for lifetime to whole DE at 4 mg/m^sup 3^ or to filtered DE at 2 mg/m^sup 3^ and reported no pulmonary tumors (Heinrich et al., 1982). In a later study, Heinrich et al. exposed 52 male and female hamsters (concomitantly with rats and mice, described earlier) for 19 h/day, 5 days/week, for 24 months to 4.2 mg/m^sup 3^ whole or filtered DE (Heinrich et al., 1986b). No pulmonary tumors were observed in either the exposed groups or the air controls. The third study exposed 104 male and female hamsters (and rats, described earlier) for 16 h/day, 5 days/week, for 24 months to whole or filtered DE (or gasoline engine exhaust, described later) at 0.7, 2.2, or 6.6 mg/m^sup 3^ or to air (Brightwell et al., 1986, 1989). Additional groups of hamsters were pretreated with diethylnitrosamine (a tumor initiator) and then exposed to either DE or air. The authors reported no pulmonary tumors in any of the DE-exposed or air control hamsters, including those pretreated with nitrosamine. Thus, in the hamster inhalation test system, DE did not act as a cocarcinogen.

Chronic Inhalation Studies With Other Species

Two reports were published prior to 1988 of nonrodent species exposed to DE: cats (not reviewed in the IARC monograph) and cynomolgus monkeys (Table 4).

Studies conducted by the U.S. EPA (Pepelko, 1982; Plopper et al., 1983; Hyde et al., 1985; Moorman et al, 1985) exposed 25 adult male cats, 8 h/day, 5 days/week, for up to 2.3 years (124 weeks) to DE at 6 mg/m^sup 3^ (first 61 weeks) and then 12 mg/m^sup 3^ (subsequent 63 weeks). Lung function was tested at intervals during and after exposure. Some cats were sacrificed immediately after exposure and others were sacrificed after an additional postexposure period of 6 months, at each of these time points, lungs were examined for pathology and histopamology. No pulmonary tumors were reported in either the DE-exposed cats or the air controls; this was not an unexpected finding in view of the fairly short observation time relative to the lifespan of the species. No effects on lung function were observed after the first 61 weeks of exposure. However, after the subsequent 63 weeks of exposure, a number of effects were observed in the DEexposed cats: significantly reduced lung volumes and diffusing capacity (according to the authors, indicating a classic pattern of restrictive lung disease, suggesting interstitial pulmonary fibrosis); peribronchiolar fibrosis; increased lymphocytes and fibroblasts; numerous interstitial and alveolar macrophages laden with black particles; bronchiolar epithelial metaplasia with increased ciliated and basal cells; increased newly synthesized collagen. Following the 6-month postexposure period, the bronchiolar epithelium returned to normal. Lung function results following recovery were not reported. The authors concluded that "exposure to diesel exhaust produces changes in both epithelial and interstitial tissue compartments and that the focus of these lesions in peripheral lung is the centriacinar region where alveolar ducts join terminal conducting airways" (Plopper et al., 1983).

The NIOSH study exposed 15 male cynomolgus monkeys (concomitantly with rats, described earlier, and mice, discussed later in Post- 1988 Studies) for 7 h/day, 5 days/week, for 2 years to 2 mg/m^sup 3^ of DE, CD (coal dust), DE + CD (50:50), or air (Lewis et al., 1986, 1989). Monkeys were then necropsied and examined histologically. No tumors were observed in any of the monkeys. The absence of a tumor response, even if the DE were acting as a carcinogen, is not unexpected in view of the fairly short observation period relative to the life span of the species. In contrast to the cat studies already described, pulmonary function tests of the monkeys did not suggest restrictive lung disease but did suggest obstructive airway disease. In all three dust-exposed groups of monkeys, aggregates of black particles were observed in the distal airways and avleoli of the lungs. No signs of dustrelated fibrosis, emphysema, or inflammation were observed. Synergistic effects of DE and CD were not demonstrated.

Interpretation of Pre-1988 Rat Findings

During the 1970s and early 1980s, substantial research was conducted to evaluate the mechanisms by which DE might cause adverse health effects. Much of this research was reviewed in the three early diesel exhaust symposiums. A substantial portion of this research focused on understanding the nature of the mutagenic components of diesel exhaust particle extracts and their potential role in inducing cancer. This was exemplified by the paper of Lewtas and Williams (1986). A second line of research focused on understanding the role of the accumulation of particles in the lungs in the induction of lung cancer in the rats. A preview of the direction of this work was given by McClellan et al. (1982) and Vostal et al. (1982a, 1982b) at the second diesel exhaust symposium (Lewtas, 1982). At the third diesel exhaust symposium (Ishinishi et al., 1986a), several papers reported progress in understanding the disposition of inhaled diesel exhaust particles and especially the role of alterations in particle kinetics on induction of lung cancer in rats with exposures at high concentrations.

Wolff et al. (1986) summarized results from Lovelace (later to be published in Wolff et al., 1987) and General Motors (Strom, 1984) on the disposition of inhaled diesel exhaust particles. As exposures continued beyond 6 months, the lung burden of particles in rats exposed at the higher concentrations (3.5 and 7.0 mg/m^sup 3^) were disproportionately higher than burdens in the rats exposed at the lowest concentration (0.35 mg/m^sup 3^) (Figure 4). The fractional deposition of radiolabeled tracer particles was similar at all three exposure concentrations; however, the clearance of radiolabeled tracer

Source: Critical Reviews in Toxicology

More News in this Category