Tobacco Documents Online

I~UNDAMENTAL AND APPLIED TOXICOLOGY 37, 3"7--~3 (199"7) ART~Ct£ NO. FA972297 Lung Tissue Responses and Sites of Particle Retention Differ between Rats and Cynomoigus Monkeys Exposed Chronically to Diesel Exhaust and Coal Dust Kristen J. Nikula, Kelly J. Avila, .William C. Griffith, and Joe L. Mauderly In/wda~on Toxicology Research Institute, P.O. Box ~890, Albuquerque, New Mexico 87185 Received October 14, 1996; Lung Tissue Responses and Sites of Particle Retention Differ between Rats and Cynomolgus Monkeys Exposed Chronically to Diesel Exhaust and Coal Dust. Nikula, IC J., Avila, IL J., Griflith, W.C., and Mauderly, J.L. (1997). Fundam. Appl. Toxlcol. 37, 37-53. Several chronic inhalation bioassays of poorly soluble, nonfl- brous particles have resulted in an increased incidence of lung tumor~ in rats, no into'ease in lung tumors in Syrian hamsmm, and inconsistent results in mice. These results have raised concerns that rats may be morn prone than other spedes to develop persistent pulmonary epithelial hyperplasla, metaplasla, and tumors in re- sponse to the accumulation of inhaled particles. In addition, patti- tie deposition and the rate of particle clearance from the lung differ between rats and primates, as does the anatomy of the centri- acinar region. For these reasons, the usefulness of pulmonary car." cinogenicity data from rats exposed to high concentrations of par- titles for quantitatively predicting lung cancer risk in humans exposed to much lower environmental or occupational concentnv tions has been questioned. The purpose of this investigation was to directly compare the anatomical patterns of particle retention and the lung tissue responses of rats and monkeys exposed chroni- cally to high occupational concentrations of poorly soluble parti- cles Lung sections Rem male cynomolgns monkeys and F344 rats exposed 7 hr/day, 5 days/week for 24 months to filte~d ambient air, diesel exhaust (2 mg soot/mS), coal dust (2 rag respirable partic. ulate material/mS), or diesel exhaust and coal dust combined (1 mg soot and 1 mg respirable coal dnst/m3) were examined histo- pathologically. The relative volume density of particulate material and the volume percentage of the total particulate material in defined pulmonary compartments were determined morphometri- tally to assess the relative amount and the anatomic distribution of retained particulate material. In all goups, relatively more particulate material was retained in monkey than in rat lungs. After adjustment for differences between rat and monkey controls, the coal dust- and the combined diesel exhaust and coal dnst- exposed monkeys retained more particulate material than the coal dust- and the combined diesel exhaust and coal dust-exposed rats, respectively. There was no significant difference in the relative amount of retained particulate material between diesel exhaust. exposed monkeys and rats. Within each species, the sites of particle retention and lung tissue responses were the same for diesel soot, coal dust, and the combined material. Rats retained a greater 37 ~ J~nmwy 24, 1997 portion of the particulate material in lumens of alveolar ducts and alveoH than monkeys. Conversely, monkeys retained a greater portion of the particulate material in the interstitium than rats. Rats, but not monkeys, had significant alveolar epithelial hyper- plastic, inflanm.~a, tory, and septal fibrotic responses to the retained particles. These results suggest that intrapnimonary particle reten- tion patterns and tissue reactions in rats may not be predictive of retmtion patterns and tissue responses in primates exposed to poorly soluble partieles at concentrations representing high occu- pational exposures. © 1~ s~,cT.~. Several chronic inhalation bioassays of diesel exhaust in rats, Syrian hamsters, and mice have been conducted (results reviewed by Mauderly, 1995). These bioassays have consis- tently shown that diesel exhaust, inhaled chronically at high concentrations, causes increased incidences of lung minors in mrs. Exposures resulting in an increased incidence of lung tumors in rats also caused progressive soot accumulation, primarily within alveolar macrophages, alveolar epithelial hyperplasia, chronic-active inflammation, focal fibrosis, and epithelial metaplasia. None of the studies using Syrian ham- stets has demonstrated a diesel exhaust-related increase in lung tumors, although soot accumulated in hamster lungs. Mild bronchiolar-alveolar hyperplasia, which was much less severe than in comparably exposed rats, has been ob- served in hamsters that chronically inhaled diesel exhaust (Heinrich et aL, 1986). Inconsistent results have been oh- mined in studies using mice. Increased incidences of lung tumors occurred in some female groups of Swain A and Sencar mice, strains sensitive to chemical carcinogenesis, but results have been generally negative in other _swains. Lung tumors were not increased in CD-1 mice exposed under conditions carcinogenic to mrs (Mauderly er aL, 1996). These mice progressively accumulated soot, primarily in al- veolar macrophages, and septal fibrosis and bronchiolization of alveolar ducts were observed in areas of soot accumula- tion. However, the amounts of inflammation and epithelial hype~plasia were less in mice than in rots (Henderson et al., 1988: l~Iauderly er al., 1996). An interspecies difference in response, with rats but not 0272-0590/97 $25.00 Copyright ~) 1997 by the Society of Toxicology. All fights of reproduction in any form reserved.

38 NIKULA ET AL. mice showing an increased incidence of lung tumors, has been shown for a number of relatively insoluble particles including talc (NTP, 1993), carbon black (Heinrich et al., 1992), titanium dioxide (Heinrich et al., 1992), silica (re- viewed by Saf-fiotti, 1995), nickel subsulfide (NTP, 1994a), cadmium sulfate (Heinrich et a/., 1989; Glasser et al., 1990), cadmium sulfide (Heinrich et al., 1989; Glasser et al., 1990), and cadmium chloride (Heinrich et al., 1989; Glasser et al., 1990). Much of our knowledge of the response of human lungs to heavy particle loading comes from coal miners. They accumulate specific lung burdens of coal dust that are in the range of specific lung burdens associated with particle- induced carcinogenicity in rats (summarized in Mauderly, 1994), yet coal dust exposure alone does not significantly increase the risk for lung cancer in humans (Merchant et al., 1986). These results have raised concerns that rats may be more prone than other species to develop persistent pulmo- nary epithelial hyperplasia, metaplasia, and tumors in re- spouse to inhaled particles. The rate of particle clearance from the alveolar region differs among species. Rats and mice clear particles from the lung relatively quickly, whereas monkeys and humans clear particles more slowly (Snipes, 1989). Clearance in all these species can be described by two-phase kinetics com- prised of a faster and slower component; the half-time for the faster component is 25 to 30 days for all these species. In rats and mice, 90% of the total clearance falls into the faster component, and the half-time of the slower component is approximately 200 to 250 days. In monkeys and humans, only 20 to 30% of the total clearance falls into the faster component, and the half-time of the slower component is 600 to 700 days. These differences in the rate of particle clearance suggest that the mechanisms of clearance and/or the sites of particle retention differ between the faster-clear- ing and slower-clearing species. Anatomical differences between the faster-clearing and slower-clearing species could affect particle deposition, re- tention, and clearance. The functional unit of the lung, the acinus, is composed of the terminal bronchiole and the air spaces it supplies (Schreider and Raabe, 1981). Because mice and rats lack respiratory bronchioles, they have simple acini. Macaque monkeys and humans have similar numbers of generations of respiratory bronchioles between the termi- nal bronchiole and alveolar ducts (Phalen and Oldham, 1983; Tyler, 1983), and they have larger alveoli and alveolar ducts than rats (Mercer and Crape, 1988). Therefore, monkeys and humans have more complex, larger acini then rats. The amount of interstitial connective tissue in the lung also dif- fers, with small rodents having less and primates more. A greater portion of the pulmonary parenchyma (compq.sed of air in alveoli and ducts, capillary blood, and septal tissue) is composed of septal tissue in rhesus monkeys and humans than in rats (Pinkerton et al., 1982; Kapanci et al., 196"9, 1972; Crape et al., 1982). Lastly,. rats have thin pleura, rela- tively few pleural lymphatics, and no interlobular connective tissue, while humans have thick pleura, relatively abundant pleural lymphatics, and abundant interlobular connective tis- sue (McLaughlin et al., 1961, 1966; Leak and Jamuar, 1983). Nonhuman primates generally have pleura that are classified as thin, but which are thicker and have more lymphatics than rat pleura; they have scant interlobular connective tissue (McLanghlin et al., 1961; Tyler and Julian, 1992). The scientific and regulatory communities are currently debating the usefulness of pulmonary carcinogenicity data from rats exposed to high concentrations of particles for predicting lung cancer risk in humans exposed tO much lower environmental or occupational concentrations. For example, Snipes (1996) noted that although substantial pulmonary burdens of dust have been reported for monkeys and humans, altered pulmonary clearance, the defining feature of lung overload in rats, has not been demonstrated in the larger species. He predicted that chronic inhalation of dusts would result in different dust accumulation patterns in the lungs of monkeys or humans than in rats, and that the larger species would not exhibit manifestations of lung overload. Schultz (1996) stressed the importance of distinguishing between nonspecific pathologic effects of dust overload from more dust-specific effects seen at lower lung burdens. He con- eluded that a minimal overload concentration may be desir- able in some inhalation bioassays, but that rat-specific histo- pathology produced under these conditions would likely have little relevance to humans. Information relevant to these issues could be gained by comparing the responses of rat and nonhuman primate lungs to particles at exposure concentrations within the range of occupational exposure. Tissues available from the Lewis et al. (1989) 2-year bioassay of F344 rats and cynomolgus monkeys exposed at 2 mg respirable particulate material/m~ to diesel exhaust, coal dust, or a combination of diesel ex- haust and coal dust afforded the opportunity to make this comparison. The diesel exhaust exposure concentration was just below that which has proven carcinogenic (3.5 mg/m~) in rats under similar exposure conditions (Mauderly et al., 1987). The coal dust concentration was at the current permis- sible airborne concentration in underground coal mines in the U.S. (2 mg respirable particulate/ms) which has been shown by an extensive database to be noncarcinogenic in miners. Lewis and co-workers (1989) provided excellent documentation of this well-conducted study, and they exam- ined many parameters with an emphasis toward detection of possible synergism between diesel exhaust and coal dust exposure. Interspecies comparisons were not emphasized. The study did not include grading of lesions by the same pathologist using the same terminology and scale in rats and monkeys, nor did it include morphometric analyses of amounts of retained particulate material or sites of particle retention.

• CHRONIC EXPOSURE TO DIESEL EXHAUST AND COAL DUST 39 The purpose of this investigation was to use materials from the Lewis et aL study to make direct quantitative com- parisons of the patterns of particle retention and the lung tissue responses of rats and monkeys exposed chronically to diesel exhaust, coal dust, or diesel exhaust combined with coal dust. The hypothesis was that particles would be re- rained differently in lungs of chronically exposed rats and monkeys, and that the tissue responses would also differ. MATERIALS AND METHODS Animal txposurts, ¢xpo~wt matarials, amt ~ ~p~ ~ I~g fi~ e~n~ were ~m ~e ~wis et ~ ~1989) 2-y~ bi~y of c~omol~ ~ys ~d ~ m~ ex~ w fil~ condifion~ ~biem ~r ~ cun~is (C) or to ~lu~ whole ~1 e~¢ ~ a ~ p~cle co~en~on of 2 m~m~ (DE), ~ c~ d~z ~oli~ in ~ ~ a ~ ~pi~ie p~cle concen~on of 2 m~m~ (CD), ~ to a commo- tion of I m~m~ ~e~l ~t wi~ ~e s~ g~ or va~r ~me cun~- d~t (DECD). MMe cynomolgm (M~a f~c~u~) mo~ys (4-~ ~ wh~ o~n~) ~ ~e ~ fete ~ ~ (8-10 w~ old w~ ~ ~ e~ w~ ~ by b~ng No. 2 ~1 ~! cun~ng ~ng. fic~o~ of 80% < 10 ~m ~d 50% < 5 ~m ~d ~li~ ~ing a Wfi~ ~o~ in ~e ~ ~ DE~ ch~ w~ 4.98 • 0.82 (~ ~ SD) ~d 3.23 ~ 0.~ m~m3, ~vely. ~sp~le ~cle con~n~o~ ~d DECD ch~, ~vely (~m ~m ~wis er ~, 1989). ~er~e end of ~e ~mon~ ex~s~ w~ ob~ ~m ~e U.S. N~o~ Ins~mw ~ ~cop~on~ SffeW ~d H~ (~OSH). L~g s~o~ ~m 14 C, 14 CD, 15 D~ ~d 15 DE~ mo~ys w~ ex~n~ ~e 1~ h~ ~n infl~ via in~h~ ~afion of b~e~ fo~ ~ 20- ~ cm hy~s~c ~ss~. ~r fixation, ~o ~fio~ of ~h 1o~ (o~ ~m ~e pmx~ ~on ~d one ~m ~e ~s~ ~on) h~ ~n em~- (H&E). ~e pmxi~ ~d ~s~ s~uns ~m ~e [e~ apic~-c~, leh diap~c, ~d fi~t ~ap~g~c lo~s we~ e~n~ by H~t ~- cmscopy. ~ [o~s we~ chosen ~a~e ~ey s~pled ~e c~ ~d Cau~ lung ~d ~e right ~d le~ si~s ~d we~ most consistently av~le; • e~ we~ only two c~s wbe~ one s~fion w~ ~ssing. Lungs of eight ~e ~ from e~h ex~s~ ~ap h~ ~en infla~ via in~che~ instillation of 2% ~ovsky's fixative in 0.2 M ~ c~ylate buffer at 20 cm by,static p~ssu~. A~er fixation, a ~sv~e tissue block (approximately 4 mm thick) from the mid-left lobe and from the mid-right diaphragmatic lobe had been embedded in paraffin, sectioned at 5 ~m. and stained with H&E. A second slide was generated from each block by cutting a section from the opposite (distal) face and staining with H~E. These four sections were examined by light microscopy, rn order to ~ the number of rat lungs examined to match the monkeys, lungs from seven additional male rats per exposure group were randomly selected. These lungs had been intratracheally instilled with 10% buffered formalin in sire until full inspirational volume was reached. The trachea was clamped and the lungs immersed in formalin. One section each from the left lung, right apical, right cardiac, and right diaphragmatic lobes had been cut and stained with H&E. These sections were examined by light microscopy. A standard lesion terminology developed previously (Nikula et al., 1995) was used, and histopathologic findings were entered into a computer dam base (PathTox; Xybion Medical Systems, Cedar Knolls, NJ). Lesions were scored as present or absent. If present, the severity of each lesion was graded on a scale of slight to marked, indicating the approximate fraction of the lung or structure judged to be involved (dight = I-2%, minimal = 3-I0%. mild = 11-24%, moderate = 25-50%, and marked =. 51-I00%). A locally intense lesion was scored as slight if only a small portion of the lung was affected. In an effort to be consistent across expos.ure groups and species, one monkey from each exposure group was examined followed by one rat in each exposure group, then the cycle was repeated. The same lesion terminology and grading scale were used for both species. Morphom~Uy, Lungs that had been inflated at a constant pressure of 20-25 cm (monkeys) or 20 cm (rats) hydrostatic pressure were used for mmphometry. Lungs from eight rats per exposure group were available, and the same four sections used for histopathology from each animal were examined. To examine lungs from an equal number of monkeys, lungs with debris and pigment due to pulmonary acariasis that was scored as mild or above were excluded from selection, and then eight lungs from each expo- sure group were randomly selected. The same proximal and distal sections fiom the left apical-cardiac, left diaphragmatic, and right diaphragmatic lobes as used for histopathology were examined. The distal po~on of the right diaphragmatic lobe was missing in one case, so the distal portion of the right apical lobe was substituted. The point-counting method of planimetw (Elias and Hyde, 1983) was used to estimate the relative volume densi~ of particulate material in the lung sections and the volume percentage of the tomi particulate material in defined anatomic compamnents of the lung. These compartments are listed and defined in Table 1. A systematic, random sampling scheme was used to record an average of 102 and 2.38 lung fields in monkeys and rats, respectively. The relative scarcity of particles in the rat versus monkey lungs necessitated greater sampling of the rat lungs to record a sufficient number of points hitting particulate material for statistical analysis. Digi- tized images of the lung were captured using a Sony ICX 038AK color camera interfaced to an Olympus BK?.-~ microscope and a Macintosh Quedra 950 computer. The images were captured using a 40× microscope objective and projected onto the computer monitor screen at a final magni- fication of 1280×. A 64-point grid (Stereology Toolbox, Davis, CA) was superimposed over each image, and the number of points hitting particulate material the location of each point, and the number of points hitting the lung section but not hitting particulate material were recorded. The relative volume density expressed as a percentage is a volume ratio of p~ticulate material to lung and was calculated for each animal from the number Of points hitting particulate material divided by the total number of test points × 100%. These data are estimates of the relative amount of retained paniculate material in the lung. The volume percentage of the total particulate material in a defined anatomic compartment is a volume ratio of particulate material in a compartment to total particulate material, and was calculated for each animal from the number of points hitting particulate material in a defined compartment divided by the total number of points hitting pat~icolate material × 100%. These dam are estimates of the ana- tomic distribution of the retained particulate material. ~0 0

NIKULA TABLE 1 Morphometri¢ Compartments for Particle Counts" Point hitting particle in (1-9): 1. Pleura Including associated connective tissue and lymphatics 2. Bronchus-associated lymphoid tissue or intrapulmonary lymph node 3. Bmnchovascular interstitium of conducting airways Including airway walls, peribmncbovascular connective tissue, and associated vessels, including lymphatics 4. Lumen of conducting airway 5. Interstitium of respiratory bronchinle (monkey) Including respiratory bmnchinlar wall, peribronchiular and vascular connective tissue, and associated vessels, including lymphatics. Not including septa of alveolar outpockctings 6. Lumen of respiratory bronchiole (monkey) 7. Perivascular interstitium Perivascular adventitia and lymphatic capillaries surrounding pulmonary arterioles, veins, and vanules not associated with conducting airways or respiratory bronchidies. Includes interiobular connective tissue. 8. Sepmm of alveolar duct or alveolus Including septa of alveolar outpocketings of respiratory bmnchioles in monkeys 9. Lumen of alveolar duct or alveolus I0. Point hitting lung section but not hitting a panicle "For each field, points hitting th~ lung section (test points) were counted in one of 10 categories. Thus. total points hitting'the lung section, total points hitting particles, and points hi.rig pm~icles in each defined lung compartment could be calculated. Although the lungs used for morphomew/from both species were per- fused under constant pressure at similar hydrostatic Im~sures, slight differ- enccs in pressure, shrinkage due to fixing and processing, and sampling strategy could have affected the comparative interspecies results for the relative volume densities of particulate material in the lungs, i.e., volume of particulate nmteriai/volume of lung. However, these factors would not have affected the location of the particulate material, i.e., the percentage of mud particulate material in defined anatomic compartments; thus, they would not have affected the comparisons of particle location across species or within species by panicle type. Also, because the fixing, processing, and sampling factors were the same within each species~ they would not have affected the comparisons of relative volume densities by particle type within each species. The volume fraction of the lung occupied by the various compartruents, as represented in these lung sections, was determined so that values ob~fined from these lung sections could be compared with normal values to assess the lung inflation and sampling. A systematic, random sampfing scheme was used to record an average of 36 fields in the monkeys and rats. Using the same imaging system described above, images were captured using a 10× objective and projected onto the screen at a final magnification of 320×. A 42-point grid was superimposed over each image, and the number oF points hitting each anatomic compartment of the lung section was re- corded. The volume fraction of the lung occupied by specific compartments was calculated for each animal from the number of points hitting the com- partment divided by the toud number of test points. Because of the anatomic difference between rat and monkey lungs and because the diesel and coal particles differ morphologically, it was not possible m evaluate the sections while blinded to species and treatment group. However, because grid points were evaluated, and these were chosen randomly, the morphometry should be free from bias. In addition, morphometric and histopathologic evaluations were conducted by different people, thus providing another guard against bias. SUe/st/ca/ams/ys#~. The crimrion for smtisticai significance was set at p < 0.05 for all analyses. The pathology severity scores were analyzed using polychotomous logistic regression to estimate the prevalence of lesious with a particular score. This is a mchnique for analyzing ordered category data with more than two outcomes, and the spacing between the categories does not have to be equal. In this technique, the cumulative disu'/bution probabilities of the scores are modeled, i.e., the probability of a score and all lower scores. A linear model of the logit of the cumulative pmhabilities was estimated. Other explanatory vm-iahies were then tested in the analysis to de~rmiue if they were significantly related to the prevalence of the lesions being seored. The simplest model of parallel regressions for scores was used (McCullagh and Nelder, 1983). Using this analysis, it was possible to determine whether exposure m pmlicles or a type of particle was a significant explanatory variable and whether there were any significant differences between species. An odds ratio and its 95% confidence interval were estimated for each factor. The odds ratio summarizes the effect of ~ factor for all severity scores. Odds ratios greater than 1 indicate an inc~ssed response above the comparison group, values less than 1 indicate a decreased response, and a value of 1 indicates no change from the comparison group. The statistical significance of the odds ratio is at the p a 0.05 level when the 95% confidence interval does not contain the value of 1. In some cases, it was not possible to estimate an odds ratio because all or almost all of the controls had no response, and all of the exposed animals had at least a slight response. In these cases, became the conuols were the denominator and had 0 response (or nearly 0), the odds ratio could nn~ be estimated. Because the relative volume density of particles was estimated based upon counting the number of grid points falling on panicles, the dam were analyzed using Poisson regression fo~ which it is assumed that the variability in the numbe~ of panicles counted follows a Poisson distribution. In this study the~ were additional sources of variation among anita., especially monkeys, beyond flmse due to smnpiing. This was accounted for in the analysis by using a scaling fac~. greatm- than 1, to multiply the sampling variance. Also. because sfightly different numbers of grid points were exam- ined for each animal, an off~ based upon the log of the number of grid points was included in the Poisson regression (McCullagh and Nelder, 1983). The estimates of the volume percentages of the toud particulate material in the lumens of alveolar ducts and alveoli and in interstitial compartments were based upon the number of points b/uing panicles in these compart: ments divided by the total number of points hitting particles for each animal. These dam were assumed to be binomially disu-ibuted, and logistic regres- sion was used for the statistical mmlysis. Again, there were additional sources of variation among animals beyond those due to sampling, and this was accounted for in the mmlysis by using a scaling factor, greater than 1, to multiply the sampling vmiance (McCullagh and Nelder, 1983). Histopathology General Descriptive Overview The lungs of diesel exhaust-exposed (DE), coal dust-ex- posed (CD), and diesel exhaust combined with coal dust- exposed (DECD) rats exhibited the same histopathology. The incidences and severity of some specific components of the pulmonary response to particles differed only slightly. Overall, the particles were observed mainly within mukifo- cal collections of alveolar macrophages. Most commonly

CHRONIC EXPOSURE TO DIESEL EXHAUST AND COAL DUST 41 these macrophage aggregates were located within centriaci- nat alveoli (Fig. IA), but the alveoli immediately adjacent to the pleura was another common location (Fig. 1B). A lesser portion of the retained particles was located in the interstitium. The characteristic tissue response to densely aggregated alveolar macrophages was alveolar epithelial hy- perplasia. Other responses, in descending order of incidence, were particle-associated inflammation, a local septal fibrotic reaction, and alveolar proteinosis. The frequency of these lesions was directly correlated with the amount of particulate material observed within aggregated alveolar macrophages. The DE rats, wh!ch appeared to have the most aggregates of heavily particle-laden alveolar macrophages, had slightly greater incidences of these associated responses. Most of the control monkey lungs contained particulate material, and similar nonsoot, noncoal-dust, particulate ma- terial was present along with the soot and coal dust in ex- posed monkey lungs. The nonsoot, noncoal-dust, particulate material, which was generally lighter colored, gray to golden, and more loosely packed than the diesel soot or coal dust in exposed monkeys, consisted of endogenous pig- ments, materials inhaled and retained over the lifetime of the animal, and, in some of the monkeys, debris from pulmonary mites. The incidences (one-third of the monkeys were af- fected) and severity of pulmonary acariasis and the associ- ated focal chronic eosinophilic bronchiolids were the same across all exposure groups. This bronchiolitis, which af- fected respiratory bronchioles and alveolar ducts, was char- acterized by eosinophilic, chronic granulomatous inflamma- tion, presence of mites or mite debris, and frequendy con- comitant thickening of brunchiolar or ductular walls and adjacent septa due to smooth muscle hyperplasia, increased collagen, and slight alveolar epithelial hyperplasia. When these characteristic mite-induced findings occurred together, the single diagnostic term "'chronic eosinophilic granuloma- tous bronchiofitis" was used instead of scoring each subpart of the lesion separately. At low magnification, particulate material was more ap- parent in the monkey lungs than in the rat lungs. Unlike the rats, the DE monkey lungs appeared to contain less particu- late material than the CD monkeys. The predominant sites ofparticle retention and the characteristic tissue responses differed between the rats and monkeys, although they were similar in DE, CD, and DECD monkeys. The retained pani- cles had a muldfocal distribution in the monkey lungs, but, in contrast to the rat, more of the particulate material was located in the interstitium than in the alveoli. The most com- mon locations of particle-laden macrophages were (1) within the lumens of alveoli at the level of first- and second-genera- tion alveolar ducts (Fig. 1C), (2) within.the alveolar septa (Figs. 1C and 1D), (3) within the interstitium of respiratory bronchioles (Fig. 1D), (4) within the adventitia and lym- phatic capillaries surrounding arterioles and veins within the pulmonary parenchyma, and (5) in the pleura. The interstitial particulate material did not seem to elicit a tissue response. The portion of the particulate material within intralumenal collections of alveolar macrophages was smaller than that in rats, and the aggregates of particle-laden macrophages elicited much less of a tissue response in the monkeys than Lesions, Severity Scores, and Statistical Analyses Tables 2 and 3 show the incidences and average severity scores of each category of lesion. These lesions, the compar- ative histopathological findings, and results of the statistical analyses are described below. Alveolar macrophage hyperpla~ia. Alveolar macro- phage hyperplasia in particle-exposed rats consisted of in- creased numbers of alveolar macrophages that contained die- sel soot or coal dust (Fig. IA). The.macrophages were multi- focally aggregated within small groups of alveoli, rather than evenly distributed among the alveoli. The most commonly affected alveoli were .along the first-generation alveolar ducts. The macrophages in these aggregates were heavily particle-laden, and the lumens of alveoli were sometimes obliterated by the macrophages. Peripheral to the focal accu- mulations of particle-laden macrophages, the number of macrophages per alveolus was markedly decreased. Widely scattered macrophages with particles were .disseminated throughout the lungs. The scattered macrophages were smaller and contained less particulate material than those in the focal accumulations. Alveolar macrophage hyperplasia in particle-exposed monkeys also consisted of increased numbers of alveolar macrophages that contained particulate material (Fig. 1C). Particle-laden macrophages were mosf commonly aggre- gated in alveoli adjacent to first- and second-generation alve- olar ducts. In monkeys, macrophages rarely obliterated alve- oli, which are larger in monkeys than in rats. As in the rats, scattered, smaller macrophages containing less particulate material were disseminated throughout the lungs. In both species, alveolar macrophage hyperplasia had scores of none to mild, and there was no significant differ- ence between species (p = 0.5). In both species, alveolar macrophage hyperplasia was significantly greater in particle- exposed than in control animals (p < 0.001). However, it was not possible to estimate the odds ratio because most of the controls did not exhibit alveolar macrophage hyperplasia, and all of the exposed animals had at least slight alveolar macrophage hyperplasia. DECD rats had a slightly lower response compared to DE and CD rats (p ffi 0.04) with an odds ratio (and 95% confidence interval) of 0.24 (0.055, 0.99), but there was no statistical d~fference in the response when the DE or the CD rats were compared separately to the DECD rats. There was no significant difference in alveolar 0

42

CHRONIC EXPOSURE TO DIESEL EXHAUST AND COAL DUST 43 TABLE 2 Incidences and Average Severity Scores of Particle-Associated or Possibly Particle-Associated" I-listopathologic Findings in Rats Diagnosis C DE CD DECI~ Number of rata examined Alveolar macrophage hyperplasia Alveolar epithelial hyperplasia Panicle-associated inflammation Septal fibrotic reaction Alveolar prot~innsis Particles in lumens of alveolar ducts and alveofi Pa~cles in insterstitium 15 15 15 15 ! (1)~ 15 (1.7) 15 fl.5) 15 (l.2) 1 (1) 15 (1.7) 14 (1.5) 15 (1.2) 0 (--) 10 (1.1) 7 (1.0) 7 (1.0) o (--) 7 (1.1) 4 (1.o) 4 (t.0) 0 (--) 4 (1.0) 2 (1.0) 3 (1.0) 0 (m) 15 0.7) 15 0.3) 15 (1.2) 0 (--) 15 (1.1) !5 (1.1) 15 (1.1) "Histopathologic findings whose incidence or severity suggested a possible particle association in any exposure group in either rats or monkeys. b Rats were exposed to filtered air as controls (C), diesel exhaust (DE), coal dust (CD), or a combination of diesel exhaust and coal dust (DECD) at target concentrations of 2 mg respirable particulate material/mL ~F'LrSt number is the number of rats with the findings. Number in parenthesis is the average severity score calculated as ,,sum of the severity scores . Lesions were scored as 1, slight; 2, minimal; 3, mild; 4, moderate; and 5, marked. number of mm with the finding macrophage hyperplasia among the DE, CD, or DECD mon- keys (p = 0,7). Epithelial hyperplasia. Alveolar epithelial hyperplasia was observed in the lungs of all particle-exposed rats, except one CD rot. Epithelialhypezplasia was multifocal and-con- sisted of an increased number of hypertrophic, cuboidal, alveolar epithelial cells lining alveolar sept~ Affected alve- oil contained particle-laden macrophages; most often, the alveolar lumens were densely packed with macrophages (Fig. 2A). Usually, the severity of the alveolar epithelial hyperplasia was proportional to the aggregation of particle- laden alveolar macrophages. Sometimes, the amount of alve- olar epithelial hyperplasia was disproportional to the amount of particulate material (Fig. 2B). One control rat exhibited a single focus of alveolar epithelial hyperplasia of unknown etiology; no particles or other alterations were associated with the lesion. Alveolar epithelial hyperplasia was less common in parti- cle-exposed monkeys than in rats. In one C and one DECD monkey with this lesion, the alveolar epithelial hyperplasia was focal and consisted of an increased number of hypertro- phic, cuboidal to columnar epithelial cells lining subpleural, TABLE 3 Incidences and Average Severity Scores of Particle-Associated or Possibly Particie-Associated" I-Ilstopathologic Findings in Monkeys Exposure group~ D/agnosls C DE CD DECD Number of monkeys examined Alveolar macrophage hypetvlasia Alveolar epithelial hyperplasia Particle.associated inflammation Septal fibrotic re~:don Alveolar proteinosis Particies in lumens of alveolar ducts and alveoli Particles in interstitium 14 15 14 15 2 (I.0)c 15 (1.2) 14 (1.4) 15 (1.5) 2 (1.5) 4 (1.5) 3 (i.0) 4 (2.0) t (l.O) 3 (I.0) 4 (1.0) 4 (1.2) 3 (1.3) 0 (--) 0 (--) ! (1.0) 0 (--) 0 (--) 0 (--) 0 (--) I (1.0) 15 (1.3) 14 (1.4) ~5 (1.5) 10 (1.3) 15 (1.9) 14 (2.0) 15 (2.1) ~ Histopathologic findings whose evidence or severity suggested a possible particle association in any exposure group in either rats oi" monkeys. ~ Monkeys were exposed m filtered air as controls (C), diesel exhaust (dE), coal dust (CD), or a combination of diesel exhaust and coal dust (DECD) at target concentrations of 2 mg respirable particulate materiaYmL ~First number is the number of monkeys with the findings. Number in parenthesis is the average severity score calculated as sum of the severity scores Lesions were scored as 1, slight; 2, minima/; 3, mild; 4, moderate; and 5, marked. number of monkeys with the finding."

~68898908

CHRONIC EXPOSURE TO DIESEL EXHAUST AND COAL DUST 45 TABLE 4 Distribution of Histopathologic Severity Scores for Alveolar Epithelial Hyperplasia among Groups of Rats and Monkeys Severity score" Rats Monkeys Exposure group# 0 I 2 3 4 0 I 2 3 4 C 14" I 0 0' 0 12 I 1 0 0 DE 0 7 6 1 1 12 2 2 0 0 CD I II 0 2 1 II 3 0 0 0 DECD 0 12 3 0 0 11 2 0 2 0 ~ Severity scores: 0, lesion not observed; I, slight; 2, minimal; 3, mild, and 4, moderate. t' Animals were exposed to filtered air as controls (C) or to diesel exhaust (DE), coal dust (CD), or a combination of diesel exhaust and coal dust (DECD) as described ~arlier. 15 rats examined in each group. 14 C, 15 DE, 14 CD, and 15 DECD monkeys were examined. " Number of animals with ~ach severity score. fibrotic alveolar septa (Figs. 2C and 2D). In one C, one DE, and two CD monkeys, slightly increased numbers of alveolar epithelial cells were seen in alveoli with a concomitant mixed inflammatory infiltrate including eosinophils or in alveoli with nearby, but not contiguous, lesions of chronic eosinophilic bronchiolitis. Although the l~sions in these alve- oil were not pathognomonic for pulmonary acariasis, it is likely that a reaction to mites caused or contributed to the lesions. In one DE and two DECD monkeys, small foci of alveolar epithelial hyperplasia occurred in alveolar ducts with smooth muscle hyperplasia or in subpleural alveoli in the absence of other lesions. In two DE, one CD, and one DECD monkey, foci of alveolar epithelial hyperplasia oc- curred in association with acute or chronic inflammation that was not eosinophilic. In these last four cases, particle-laden macrophages were located in the areas of the lesions. The epithelial hyperplasia c6nsisted of a slight increase in the number of type II cells, but, unlike the rats, most of the surface of the affected alveolus was still lined by type [ cells (Fig. 2E). Alveolar epithelial hyperplasia had scores of none to mod- erate in rats and none to mild in monkeys. Table 4 shows the distribution of severity scores among groups in both species. Rats had a significantly greater response to particle exposure than monkeys (p < 0.00 ! ) with an odds ratio (95% confidence interval) of 14 (5.5, 39). Alveolar epithelial hy- perplasia was significantly greater in particle-exposed than control rats (p < 0.001) with an odds ratio of 620 (36, 11,000), but there was no difference in the hyperplastic re- sponse due to type of particle exposure (p = 0.3). The mon- keys did not have a significant alveolar epithelial hyperplas- tic response to particle exposure (p = 0.4). Inflammation. Particle-associated inflammation was not common in rats. Most loci of aggregated particle-laden mac- rophages did not exhibit inflammation, and o.nly I0, 7, and 7 DE, CD, and DECD rats, respectively, exhibited this lesion. Particle-associated inflammation in rats ranged from low numbers of neutrophils in alveolar lumens and septa con- comitant with aggregation of particle-laden macrophages to small foci of chronic inflammation, which involved alveoli adjacent to the pleura and were characterized by degenerat- ing macrophages, cholesterol clefts, fibrosis, and low num- bers of neutrophils or mononuclear inflammatory cells (Fig. IB). In monkeys, possibly particle-associated inflammation was much less common than in rats. In one DE, one CD, and one DECD monkey, the inflammation appeared to be related more to pulmonary acariasis than inhaled particles based on the presence of eosinophils. In the remaining nine cases, including one control, where some loci of inflamma- tion appeared to be associated with particulate material, the inflammation was predominantly observed as increased sep- tal neutrophils and less often as a few neutrophils in alveolar lumens or as chronic interstitial inflammation. Focal lesions with cholesterol clefts, fibrosis, and inflammation like those in the particle-exposed rats did not occur in the monkeys. Particle-associated inflammation had scores of none to minimal in both species. Rats had a significantly greater inflammatory response to particles than monkeys (p = 0.02) with an odds ratio (95% confidence interval) of 2.8 ( 1.2, 6.7). Particle-exposed rats had significantly greater inflammation than C rats (p < 0.00l), but it was not possible to estimate an odds ratio because none of the controls had inflammation. There was no significant difference in the inflammatory re- sponse due to the type of particle exposure in rats (p = 0.5). In monkeys, there was no significant effect of particle exposure (p = 0. l). Septal fibrotic reaction. The septal fibrotic reaction in rats occurred as increased collagen within alveolar septa in ~3 0 FIG. 2. Comparison of alveolar epithelial hyperplasia in rats and monkeys. (A) DE rat. Hyperplastic, cuboidal alveolar epithelium (open arrows) lining septa of alveoli containing aggregates of soot-laden macrophages. :~240. (B) DE rat. Hyperplastic, hypertrophic alveolar epithelial cells (open arrows) in an area with less aggregation of particle-laden maerophages than in 2A: x240. Compare the amounts of particulate material and alveolar epithelial hyperplasia in these rats with the typical reaction to alveolar particulate material in monkeys illustrated in Fig. IC. (C and D) Control monkey (C) and DECD monkey (D). Hyperplastic alveolar epithelium lining fibrotic alveolar septa near the pleura. Note the lack of active inflammation and paniculate material. Both C and D, x240. (E) CD monkey. The number of alveolar type II epithelial cells (op*n arrows) is slightly increased surrounding the aggregated particle-laden macrophages. Note that the degree of.hyperplasia is much less than that in A or B. x240.

46 NIKULA ET ~. ~..~ ~ ~,_. ;, FIG. 3. S~ptal fibrotic reaction in a DE rot. A~wheads indicate in- ~ collagen. Note the concomit~t aggregation of paaicle-laden macm- ~ag~, epithelial hy~l~ia (o~n a~w), and inflammato~ c¢11 infilt~te (~ ~ws). x2~. foei of alveolar macrophage aggregation, alveolar epithelial hyperplasia, and particle-associated inflammation (Fig. 3). The fibrotic reaction occurred only in the presence of all three of these other lesions. In monkeys, a septal fibrotic reaction was not associated with diesel soot or coal dust particles. In one C monkey, u~e reaction occurred near an area with chronic eosinophilic bronehiolitis. In the other th~:ee cases, two C and one DECD, it occurred as small scars or fibrotic sequelae of chronic inflammation of unknown etiology (Figs. 2C and 2D). Rats had a significantly greater septal fibrotic reaction to particles than monkeys (p = 0.006). In rats, this reaction was significantly greater in particle-exposed than control rats (p = 0.001), but it was not possible to estimate an odds ratio because none of the controls had a septal fibrotic reaction. There was no significant difference in the septal fibrotic reaction due to the type of particle exposure in rats (p = 0.I5). There was less septal fibrotic reaction in p _a~'ticle- exposed than control monkeys (p = 0.02) with an odds ratio (95% confidence interval) of 0.083 (0.008, 0.9). However, the incidence of this lesion (three controls and one exposed affected) was low. Alveolarproteinosis. Alveolar proteinosis occurred only in a small portion of particle-exposed rats and in single foci where the combined particle-associated lesions described above were most severe. There was no significant effect of exposure to particles (p = 0.5) and no difference between exposure groups (p = 0.8) in rats. Alveolar proteinosis did not occur in the monkeys. Site of particle retention. The relative distribution of the retained particles between the alveolar spaces and the interstitium was estimated during the histopathologic exami- nation by scoring the amount of observed particulate material in these compartments. Low magnification (40×) was used to scan the slides and assess the amount of particulate mate- rial; higher magnification (200 to 400×) was used as needed to determine particle location. For this scoring, particles in lymphatics and those in the connective tissue beneath the pleural mesothelium were grouped with those in the other interstitial tissues of the lung. In the rats, more of the material appeared to be in the alveolar lumens than in the interstitium (Figs. I A and 2A). The material in the lumens w.as primarily within aggregated alveolar macrophages as described above. The material in the interstitium was primarily in the septa of alveoli con- tainin~ aggregated macrophages (Fig. 4A). Other common sites for interstitial particles in rats were the connective tissue and lymphatics surrounding postcapillary venules and small veins in the parenehyma contiguous with foci of alveolar macrophage aggregation (Fig. 4A) and in the pleura adjacent to foci of alveolar maerophage aggregation (Fig. 4B). In monkeys, the retained particulate material was primar- ily in the interstitium. Retained partieles were most com- monly observed in the septa of alveolar ducts (Figs. I D and 5A), the pleura (Fig. 5B), the walls of respiratory bronchioles (Fig. 5C), and the connective tissue and lymphatics sur- rounding veins and arterioles in the pulmonary parenchyma (Fig. 5D). Unlike the rats, material was frequently located in interstitial tissues in the absence of particles within adjacent alveolar lumens. Particulate material in lumens of alveolar ducts and alveoli had scores of none to mild in both rats and monkeys, and there was no significant difference between species (p = 0.7). In both species, particles in lumens of alveolar ducts and alveoli increased significantly due to particle exposure (p < 0.001), but it was not possible to estimate an odds ratio because none of the controls had alveolar lumenal particles scored, except one C monkey. There was significantly less particulate material in lumens of alveolar ducts and alveoli in DECD rats than in DE rats (p -~ 0.007) with an odds ratio of 0.12 (0.022, 0.63) and nearly significantly less in the CD rats than in the DE rats (p = 0.0505) with an odds ratio of 0.23 (0.05, 1.I). There was no significant difference due to the type of particle exposure in monkeys (p = 0.4). Particulate material in the interstitium had scores of none to minimal in rats and none to mild in monkeys (Table 5). Particulate material in the interstitium was significantly greater in monkeys than rats (p < 0.001) with an odds ratio of 210 (42, i 100). In rats, there was significantly more inter- stitial particulate material in particle-exposed than in C rats