Abstract
Naphthalene is a ubiquitous environmental contaminant that results in dose-dependent and tissue-, species-, and cell-selective necrosis of murine Clara cells upon exposure. Naphthalene is metabolized by CYP2F to a 1,2-epoxide, the first and obligate step in events leading to cytotoxicity. The studies reported here examine the relationship between levels of transcript (mRNA) and CYP2F protein in the respiratory tract of rodents with tissue susceptibility to injury. In both mice and rats, the lung contains more CYP2F transcript than liver; levels in kidney were undetectable. Mice expressed 4- and 8-fold greater CYP2F transcript in lung and liver tissue, respectively, than rats. Quantitative immunoblot blot analysis of CYP2F in airway subcompartments revealed mice to have 30- (minor daughters/terminal bronchioles), 20- (major daughter), 40- (trachea), and 6- (parenchyma) fold higher levels of CYP2F protein than rats. Within the lungs of both rodent species, the highest CYP2F expression was found in the distal airways. The kidney contained undetectable amounts of CYP2F; multiple immunoreactive bands in liver precluded quantification. The olfactory epithelium contains the greatest amount of cytochrome P450 protein of all tissues studied in the rat, consistent with the observed pattern of in vivo injury. Overall, these studies in rodents demonstrate a strong association between CYP2F expression levels and susceptibility to naphthalene-induced cytotoxicity. Of all primate tissues studied, only the nasal ethmoturbinates contain quantifiable amounts of CYP2F, roughly 10- and 20-fold less than the corresponding tissues in rats and mice, respectively. These results suggest that rhesus macaques may be refractory to naphthalene-induced pulmonary injury.
The environmental toxicant naphthalene can be found in ambient air, ground water, sidestream and mainstream cigarette smoke, and byproducts from the combustion of diesel fuel (Agency for Toxic Substances and Disease Registry, 1995). The toxicological assessment of naphthalene has been the focus of extensive research, and the results of the National Toxicology Program 2-year inhalation studies in mice and rats have found naphthalene to be associated with neoplasia in both species. The inhalation study in mice demonstrated slightly increased incidences of alveolar/bronchiolar adenomas only in females at the highest dose (30 ppm, 156 mg/m3) (National Toxicology Program, 1992). In rats, however, neoplasms were confined to the nasal respiratory epithelium, and the incidences were dose-dependent (National Toxicology Program, 2000). Previous work has shown that inhalation of naphthalene vapor at levels of 2 ppm (10.5 mg/m3) and above leads to necrosis of murine pulmonary bronchiolar epithelial cells (West et al., 2001). This exposure concentration is considerably less than the current Occupational Health & Safety Administration exposure standard of 10 ppm (52 mg/m3) for an 8-h time-weighted average. These and other data were employed in the recent International Agency for Research on Cancer classification of naphthalene as a possible human carcinogen (International Agency for Research on Cancer, 2002). Given the extent to which naphthalene is found in the environment, the likelihood of human exposure, and the body of evidence demonstrating its toxicity in animal models, understanding the mechanism of toxic injury is essential. Elucidating the factors most clearly associated with susceptibility to injury in animal models will allow far better assessments of human risk to be developed.
The ability of the lung to metabolize several protoxicants has been well established. Compounds such as 4-ipomeanol, trichloroethylene, 1-nitronaphthalene, 3-methylindole, and naphthalene undergo metabolic activation by monooxygenases (P450), leading to the injury of lung epithelial cells (reviewed in Gram, 1997). Within the heterogeneous population of more than 40 cell types in the lung, the nonciliated bronchiolar epithelial (Clara) cells are especially sensitive to injury, and this sensitivity correlates with the localization of P450 enzymes in this cell type. However, substantial differences have been observed in the sensitivity of this cell population across species. For example, parenteral administration of naphthalene and trichloroethylene resulted in Clara cell necrosis in the lungs of mice but not rats (Plopper, 1993). In rats, injury was confined to the nasal olfactory epithelium after parenteral administration of naphthalene at doses substantially less than the LD50. These observations could be explained by relative differences in species-specific P450 catalytic activities and/or the relative cellular P450 expression levels of mice and rats.
Naphthalene is metabolized to naphthalene 1,2-epoxide, a reactive intermediate, by the P450 system. Both the 1S,2R- and 1R,2S-stereoisomers are produced. This biotransformation is the first and obligate step in the cascade of events leading to toxic injury. In preparations of microsomes or airway explants from lungs of mice, naphthalene epoxide formation occurs with a high degree of stereoselectivity, with ratios of 1R,2S- to 1S,2R-epoxide of 10:1 and >25:1, respectively (Buckpitt et al., 1992, 1995). In rats, both the absolute rate and degree of stereoselectivity of epoxidation are significantly less than in mice. Previous work has shown that the presence of CYP2F2 in murine Clara cells correlates with the stereoselectivity and rate of naphthalene metabolism and the site of injury (Buckpitt et al., 1995). Coincubation of an inhibitory CYP2F2 antibody dramatically decreases the formation of the 1R,2S-naphthalene epoxide in murine lung microsomal incubations (Nagata et al., 1990). Using recombinant CYP2F2 (rCYP2F2) expressed in insect cells, Shultz et al. (1999) found naphthalene to be rapidly metabolized with a very high degree of stereoselectivity for the 1R,2S configuration, strongly suggesting the generation of this metabolite to be an important determinant of susceptibility.
This study addresses the relationship between the amount of transcript (mRNA), protein expression levels, and the catalytic activity of a reactive metabolite-generating enzyme (CYP2F) with in vitro metabolism and, ultimately, the extent and localization of cytotoxic injury observed in vivo. Accordingly, we have shown that species susceptibility to naphthalene toxicity is associated with clear differences in relative cellular expression levels of the CYP2F protein.
Materials and Methods
Rodents. Adult male Sprague-Dawley rats (225–250 g) and male Swiss-Webster mice (25–30 g) were purchased from Charles River Laboratories (Wilmington, MA). The animals were housed in an Association for Assessment and Accreditation of Laboratory Animal Care-accredited facility in cage racks supplied with high-efficiency particulate-filtered air and inert bedding. Animals had free access to food (Purina Rodent Chow; Purina, St. Louis, MO) and water, with a 12-h light/dark cycle for at least 7 days before use. Rodents were killed with an overdose of sodium pentobarbital administered i.p.
Northern Blot Analysis. Total RNA was isolated from lung, liver, and kidney of four mice and four rats using TRIzol reagent (Invitrogen, Carlsbad, CA). RNA from rat (20 μg) and mouse (5 μg) was separated on 1% agarose gels and transferred to positively charged nylon membranes according to the manufacturer's instructions (NorthernMax-Gly; Ambion, Austin, TX). cDNA probes were generated by RT-PCR of total RNA from rat lung. The CYP2F4 probe [1409 bp (26..1434), GenBank accession no. AF017393] was 93% homologous to CYP2F2; the rat GAPDH probe [849 bp (152..1000), GenBank accession no. M17701] was 95% homologous to the mouse transcript. Probes were 33P-labeled using a DECAprime II DNA labeling kit (Ambion), and blots were hybridized at a concentration of 2.5 × 106 cpm/ml overnight at 42°C in hybridization buffer (5× standard saline/phosphate/EDTA, 50% formamide, 0.5% SDS, 4× Denhardt's solution, 100 μg/ml salmon sperm DNA). Following stringent washing with standard saline citrate, blots were developed with storage phosphor screens (Amersham Biosciences Inc., Piscataway, NJ) scanned with a Typhoon 8600 (Amersham Biosciences Inc.), and bands were quantified using ImageQuant 5.1 software (Amersham Biosciences Inc.).
Immunoblotting of Rat and Mouse Tissues. Airway segments were microdissected from lungs according to methods previously described in detail (Plopper et al., 1991). Briefly, animals were killed, and tracheas were cannulated. Lungs were infused with 1% low-melting point agarose (SeaPlaque; FMC Bioproducts, Rockland, ME) in Waymouth's medium containing protease inhibitors (Cocktail Set III; Calbiochem, San Diego, CA) and then immediately cooled on ice in Waymouth's medium. Kidneys and liver were removed and snap-frozen in liquid nitrogen and stored at -80°C until processed. Airways were blunt-dissected under a dissecting microscope to obtain distal trachea, major and minor daughter bronchi, terminal bronchioles, and parenchyma. The collected tissue samples were snap-frozen in liquid nitrogen and stored at -80°C. Nasal compartments were microdissected as previously described (Fanucchi et al., 1999). The head of each rodent was removed from the carcass, the lower jaw and skin were removed, and the head was split in half along the medial suture. The septum, maxilloturbinates, nasoturbinates, and ethmoturbinates were removed by careful blunt dissection.
Kidney and liver tissue were homogenized on ice using a rotor/stator (VirTishear; VirTis, Gardiner, NY) in phosphate buffer (100 mM, pH 7.4, with protease inhibitors). The crude homogenates were centrifuged for 20 min at 10,000g (4°C), and the postmitochondrial supernatants were used for immunoblotting. To each sample of lung or nasal tissue, 100 μl of 2× treatment buffer (0.125 M Tris-Cl, pH 6.8, 4% SDS, 20% glycerol, 10% 2-mercaptoethanol) was added. Samples were heated to 70°C for 10 min, sonicated twice for 20 s, placed at 70°C for an additional 10 min, then centrifuged at 15,000g for 30 min at ambient temperature. The supernatant was used for immunoblotting. The limited quantities of respiratory protein available necessitated the use of a fluorescent assay with increased sensitivity for determining total protein concentrations (NanoOrange; Molecular Probes, Eugene, OR). The increased sensitivity allowed for adequate sample dilution to minimize interference with the assay from the buffer constituents.
Proteins were separated by SDS-polyacrylamide gel electrophoresis using 1-mm 10% Bis-Tris gels (Invitrogen). Gels were wet-transferred (Mighty Small Transphor; Amersham Biosciences Inc.) to a polyvinylidene difluoride membrane (Immobilon-Psq; Millipore Corporation, Billerica, MA) in transfer buffer (25 mM Tris, 192 mM glycine, 20% methanol, 0.05% SDS) and then incubated in 5% powdered milk at 4°C overnight (Bio-Rad, Hercules, CA). The CYP2F isoforms were detected using rabbit anti-CYP2F2 (Nagata et al., 1990) diluted 1:20,000 and chemiluminescent detection according to the manufacturer's directions (ECL Plus; Amersham Biosciences Inc.) with the following modifications. The goat anti-rabbit secondary antibody was diluted 1:100,000. After the final TBST rinse, membranes were placed in a resealable bag (Kapak Corporation, Minneapolis, MN), and 1 ml of ECL substrate was applied directly to the blot and then placed in the dark. Blots were analyzed 2 h later with a Storm 860 fluorescence scanner (Amersham Biosciences Inc.) in blue mode (≈450-nm light-emitting diode) with a photomultiplier tube voltage of 850 mV and a resolution of 100 μm. Bands were quantified using ImageQuant 5.1 (Amersham Biosciences Inc.).
Prior to the determination of relative CYP2F protein expression between species, the ability of the rabbit anti-CYP2F2 polyclonal antibody to recognize both CYP2F2 and CYP2F4 was compared semiquantitatively. Similar amounts of both recombinant CYP2F2 (Shultz et al., 1999) and CYP2F4 (R. M. Baldwin, unpublished results) protein were electrophoresed on two identical gels, one of which was stained with SYPRO Ruby (Molecular Probes) and the other immunoblotted as described above. The SYPRO Ruby-stained gel was scanned on a Typhoon 8600 scanner using a 532-nm excitation wavelength, a 610BP30 emission filter, a photomultiplier voltage of 600 mV, and a resolution of 50 μm. Quantitations were based on the assumption that fluorescent signal areas were proportional to the amount of recombinant protein present in each lane. By determining the relative amount of recombinant protein per nanogram of cell lysate protein for the two expressed CYP2F isoforms, immunoblot signal intensities could then be normalized for differences in expression levels and compared. Relative fluorescent areas versus nanogram of insect cell lysate protein for each CYP2F isoforms were plotted for both the SYPRO Ruby-stained gel and the immunoblot (Fig. 2, A and B). For an equally cross-reactive antibody, the relationship of RFU per nanogram of protein (plot slope) between the cell lysates containing either of the two recombinant proteins would be the same on plots of both the immunoblot and SYPRO Ruby-stained gel.
To compare CYP2F protein expression level across samples from multiple blots, four protein standard solutions were made, using either pooled mouse terminal bronchiole tissue or recombinant CYP2F4 that were diluted 2-, 4-, and 8-fold. These four standards were run with every immunoblot, and the relative fluorescent units of each band were plotted versus dilution factor. From the linear regression analysis of these points, band intensities could readily be normalized across membranes. Additionally, this method facilitated the assessment of appropriate sample dilution and recognition of aberrant experiments.
Rhesus Macaque Immunomapping. Through a collaborative effort, samples of respiratory tissue from young adult rhesus macaques used as untreated controls from studies conducted at the University of California, National Primate Research Center (Davis, CA) were obtained over a series of several months. In all cases, the tissues obtained during necropsy were placed temporarily on ice until snap-frozen in liquid nitrogen and stored at -80°C. Airway segments (trachea, proximal and medial conducting airways, respiratory bronchioles, and parenchyma) and nasal subcompartments (septum, maxilloturbinates, nasoturbinates, and ethmoturbinates) were processed identically to the rodent tissues described above.
Electrophoresis and Immunoblotting of Ethmoturbinate Microsomes from Mice, Rats, and Monkeys. Ethmoturbinate microsomes from mice, rats, and rhesus macaques were prepared by differential centrifugation according to established methodology (Schenkman and Jansson, 1999). To allow for sufficient amounts of microsomal protein, ethmoturbinate tissue was pooled from several animals of each species [mice (8), rat (4), monkey (4)]. Using two identical medium format (140 × 1.5 mm) 7% Tris-glycine polyacrylamide gels (Duracryl; Genomic Solutions, Ann Arbor, MI), ethmoturbinate microsomal proteins were electrophoresed in duplicate. In a fashion analogous to the antibody cross-reactivity experiment, the protein in one gel was stained (Coomassie Blue in this instance) and the other immunoblotted as described above, with one addition. After being electroblotted, the nylon membrane was briefly rinsed with water and then stained with Coomassie Blue (0.1% Brilliant Blue R250, 1% glacial acetic acid, 50% methanol). After adequate destaining in 50% methanol, proteins could be visualized and assessed for adequate resolution and protein transfer. The location of the molecular weight markers were etched in pencil on the nylon membrane. Following a quick methanol rinse (≈30 s) to remove the bulk of remaining stain, the membrane was rinsed twice with TBST (5 min) and placed in 5% powdered milk-TBST. Immunodetection was then performed as described above.
Identification of Proteins by Peptide Mass Fingerprinting. To accurately obtain the Coomassie-stained bands corresponding to the immunoreactive bands observed on the immunoblot of the duplicate gel, each sample of ethmoturbinate microsomal protein was electrophoresed with a molecular weight marker in both adjacent lanes (Fig. 8). Three protein standard bands were then used to align each immunoreactive band observed in the immunoblot with the Coomassie-stained gel. To assist in this alignment, a digital image of the scanned immunoblot was printed on a plastic transparency and placed beneath the Coomassie-stained gel. Each band was excised with a clean surgical blade.
Excised protein bands were washed thoroughly four times with Milli-Q (Millipore Corporation) water. The protein band was diced into approximately 1-mm squares and dried in a vacuum centrifuge. Proteins were reduced and alkylated according to an established procedure (Shevchenko et al., 1996). Briefly, proteins were reduced with 10 mM dithiothreitol in 100 mM NH4HCO3 (pH 8, 55°C) for 1 h and then alkylated with 55 mM iodoacetamide in 100 mM NH4HCO3 for 45 min in the dark at room temperature. Excess reagent was removed, and gel pieces were washed with 100 mM NH4HCO3 and partially dehydrated with acetonitrile; complete dehydration was then done in a vacuum centrifuge. Finally, proteins were digested in 50 mM NH4HCO3 containing sequencing-grade, modified trypsin (Promega, Madison, WI) at a final trypsin concentration range of 10 to 25 ng/μl (37°C) for 17 h. Peptides were extracted once each with 0.1% trifluoroacetic acid and then 5% formic acid in 50% acetonitrile. The volume was then decreased to 15 μl for mass spectrometric analysis by using the vacuum centrifuge.
Tryptic peptides were analyzed with a Bruker Biflex III MALDI-TOF mass spectrometer (Bruker-Franzen Analytik, Bremen, Germany) equipped with a pulsed N2 laser (337 nm), a delayed extraction ion source, and a reflectron. Fractions of tryptic peptides were desalted using C18 ZipTips (Millipore Corporation). Peptides were eluted from the ZipTip with 0.1% trifluoroacetic acid-acetonitrile (50:50). Next, 0.5 μl of the peptide aliquot (eluant) was mixed with an equal volume of matrix solution [a saturated solution of α-cyano-4-hydroxycinnamic acid in 0.1% trifluoroacetic acid-acetonitrile (50:50)] and applied to the target. The mass spectra were acquired in the reflectron mode. Due to the number and intensity of peaks in each mass spectrum, trypsin autolysis peaks (842.5 and 2211.1 Da) were obscured. Instead, internal mass calibration was performed by spiking samples with three peptide internal standards of masses 842.51, 1046.54, and 3148.47 Da. This procedure typically results in mass accuracies of 50 ppm or better. Measured monoisotopic masses of tryptic peptides were used as inputs to search the corresponding species' databases [SwissProt 9/19/03 release and NCBI (National Center for Biotechnology Information) 9/1/03] using both the Profound (http://129.85.19.192) and EXPASY (Expert Protein Analysis System) (http://us.expasy.org) search engines. The constraints of a 50 ppm mass accuracy, a molecular weight range from 50 to 62 kDa (≈56 kDa ± 10%), and zero-missed tryptic cleavage sites were used for all searches. The Profound search engine was used iteratively; protein matches were accepted based on the highest probability followed by subtraction of the “matched” peptides and repeating the process with the remaining peptides. Given the high-sequence homologies between P450 enzymes from humans and the rhesus macaque and the limited protein sequence data presently available for the rhesus macaque, all P450 isoforms identified in the rhesus macaque correspond to matches with tryptic fragments from human P450 isoforms.
Results
Northern Blot Analysis of CYP2F Transcript in Rodents. Prior to comparing the relative CYP2F transcript levels in lung, liver, and kidney of mice and rats (Fig. 1), the expression patterns of two commonly used “housekeeping” genes, β-actin and GAPDH, were compared across species and tissue types in a subset of samples. For a given tissue type, similar GAPDH signal per microgram of total RNA ratios were observed regardless of species. Ratios varied more than 5-fold between tissue types (data not shown). Because the β-actin and CYP2F transcripts differed only by approximately 100 bp, CYP2F expression levels were normalized to GAPDH expression. Initially, several cDNA probes of differing lengths were compared for specificity and sensitivity to differences in sequence homology across species. The 1409 bp CYP2F4 probe and 849 bp rat GAPDH probe were found to be the least affected by differences in hybridization affinities across species and demonstrated acceptable specificity. Negligible differences were observed when Northern blot analysis was repeated using analogous probes made from mouse genes (data not shown).
The highest signal intensities for CYP2F transcript were observed in mouse lung and to a lesser extent in mouse liver (Fig. 1). A similar trend was observed in the rat, which, compared with the mouse, expressed 4- and 8-fold less CYP2F transcript in lung and liver tissue, respectively. Negligible amounts of CYP2F transcript were detected in the kidney of either species.
Species Differences in CYP2F Protein by Immunoblot Analysis. Using recombinant proteins as standards, the polyclonal anti-CYP2F2 antibody was estimated to have comparable affinities for rCYP2F2 and rCYP2F4 (Fig. 2). The system of ECL and fluorescent scanning was reproducible and linear over a 50-fold range, and all samples were diluted appropriately to concentrations within the linear concentration range. No signal was observed on immunoblots prepared from the kidney of either rat or mouse. In contrast, multiple poorly resolved but reproducible immunoreactive protein bands precluded the comparison of hepatic CYP2F protein levels across species (Fig. 3).
A representative immunoblot of minor daughter airway proteins from mice and rats along with the four protein standards shows that the antibody recognizes a single 50- to 52-kDa protein in mouse airway when relatively small amounts of protein (1.5 μg) were electrophoresed (Fig. 4). In comparison, several proteins of lower molecular weight, in addition to CYP2F4, were observed in rat airways when 20-fold higher amounts of protein were loaded onto the gel. However, within all areas of lung tissue studied, a single immunoreactive protein with an apparent molecular weight identical to rCYP2F4 was observed within the molecular weight range of 40 to 60 kDa.
Immunoblot analysis of proteins from nasal tissues revealed a significantly smaller differential between CYP2F expression levels in mice and rats. Additionally, in many instances, two immunoreactive bands having apparent molecular weights consistent with potential cross-reactivity with other cytochrome P450 enzymes were observed (Fig. 5). Although the proteins from rat nasal tissues demonstrated this “double-banding” pattern to a lesser extent than did corresponding samples from mice, the ethmoturbinates of both species yielded the most distinctive “double-banding” of all nasal areas studied. No additional immunoreactive bands were observed outside the molecular weight ranges shown in Fig. 5. The relative CYP2F expression levels of all respiratory tissues studied in mice and rats are summarized in Fig. 6.
Immunoblot Analysis of CYP2F in Monkey Lung and Nasal Epithelium. Immunoblot analysis of proteins from all airway levels of the rhesus macaque tested at the highest protein loads (30 μg) failed to detect any trace of an immunoreactive protein using the rabbit anti-CYP2F2 polyclonal antibody (data not shown). Similar results were observed in the nasal and maxilloturbinates and in two of the three septal samples. Interestingly, there were readily detectable and quantifiable levels of a single immunoreactive protein with an apparent molecular mass identical to rCYP2F4 in the protein samples from ethmoturbinates (Fig. 7). Using mean values, the amount of immunoreactive protein in the ethmoturbinate of the rhesus macaque was 10- and 20-fold less than from the ethmoturbinates of rats and mice, respectively. The limit of detection for all monkey samples tested was a factor of 4 below the values observed in the ethmoturbinates.
Peptide Mass Fingerprinting of Ethmoturbinate Microsomal Proteins.Figure 8 shows the SDS-polyacrylamide gel electrophoresis separation of proteins stained with Coomassie Blue from the ethmoturbinates of mouse, rat, and monkey aligned with the immunoblot of the duplicate gel. The two immunoreactive bands in the mouse and rat aligned well with two intensely stained areas on the Coomassie-stained gel, whereas the single band observed in monkey aligned with an area of diffuse Coomassie staining. To determine whether the immunoblotting results properly represented the quantity of CYP2F in the sample or whether cross-reactivity with other CYP2 proteins led to the observed signal, bands on the gel marked with lines in Fig. 8 were excised and digested with trypsin for peptide mass fingerprinting as described above under Materials and Methods. Table 1 lists all cytochromes P450 with more than five unique matches from the 60 to 100 peptide fragments detected in each gel slice by MALDI-TOF mass spectrometry. For both mice and rats, the predominant P450 proteins observed in the upper immunoreactive band were CYP2A and CYP2G1. The lower band was found to contain CYP2F with some signal associated with CYP2A. Monkey ethmoturbinate microsomes contained CYP2A, and the presence of CYP2J2 and CYP2F are strongly suggested (listed in Table 1 as tentative identifications).
Discussion
Previous work demonstrating a strong correlation between CYP2F2 expression in mouse airways and naphthalene toxicity suggests insensitivity of the rat may result from differences in the catalytic activity of the rat ortholog, CYP2F expression levels, or a combination of both. Although the rat is insensitive to naphthalene-induced pulmonary injury, rat nasal epithelium is a target after either long-term inhalation (National Toxicology Program, 2000) or single-dose parenteral administration (Plopper et al., 1992). However, unlike lung, where previous immunohistochemical studies have provided qualitative information on CYP2F expression patterns in mice and rats, little is known about CYP2F expression levels in nasal epithelium. The data presented here demonstrate significant differences in CYP2F transcript and protein levels within the respiratory tract of these two rodent species. Extension of these studies to the rhesus macaque reveals an even greater differential in CYP2F expression.
Comparison of CYP2F transcript levels from mouse and rat lung revealed trends consistent with the species and tissue selectivity of naphthalene-induced injury in vivo. Both species displayed higher levels of transcript in lung compared with liver, and murine tissue contained greater amounts than the corresponding rat tissue. Similar results were observed by Nhamburo et al. (1990), who reported greater amounts of CYP2F transcript in rat lung than liver. Compared with the 4-fold rodent species difference observed in the current studies, Ritter et al. (1991) reported a dramatic 50-fold higher level of CYP2F mRNA in mouse versus rat lung and a 5- to 10-fold greater amount of transcript in rat liver versus rat lung.
The 4-fold difference between mouse and rat CYP2F transcript levels in whole lung observed in the current work are consistent with the differential observed in immunoblots of parenchymal tissue. Since conducting airway epithelial cells account for only 5% of the total cell population in rat lung (Mercer et al., 1994), and alveolar cells make up 74% (Stone et al., 1992), distal sections of whole lung are likely most similar to parenchymal tissue.
Using the anti-CYP2F2 antibody, comparison of airway subcompartments across species showed that mice had 30- (minor and terminal bronchi), 20- (major daughter), 40- (trachea), and 6- (parenchyma) fold higher levels of immunoreactive CYP2F than rats (Fig. 6). These trends agree with data demonstrating higher rates of naphthalene metabolism in airways of mice than rats (Buckpitt et al., 1995). The distribution of CYP2F within the respiratory tract is similar in both species. In mice, the highest CYP2F expression was found in distal airways, with trachea, more proximal airways, and parenchyma having 86, 36, and 4% of the expression found in terminal bronchioles, respectively. This distribution is consistent with previous immunohistochemical studies in mice (Buckpitt et al., 1995). The minimal anti-CYP2F2 staining observed in previous studies of rat airways can be attributed to limited sensitivity since lung areas with the greatest CYP2F abundance were 30-fold less than the corresponding mouse tissue. Although immunohistochemical studies only allow for the approximation of differences in CYP2F expression, cells containing an immunoreactive epitope could be visualized within the unique population of cell types in each lung subcompartment. Although the present methodology is considerably more sensitive and quantitative than immunohistochemical approaches, perceived differences in the CYP2F signal between species are potentially attributable to both increased CYP2F expression per cell and differences in the percentage of CYP2F expressing cells within the region sampled. For example, although the trachea was not found to express large amounts of CYP2F in either rats or mice, the greatest relative difference across species was observed in this location. Clara cells make up a larger percentage of the tracheal epithelial cell population in mice than in rats, likely contributing to the observed differential in CYP2F expression (Plopper, 1993).
The observation of two immunoreactive bands in the rodent olfactory epithelium necessitated further studies to determine the location of CYP2F and to characterize other cross-reacting isoforms. Peptide mass fingerprinting of ethmoturbinate proteins corresponding to the two immunoreactive bands yielded fragment matches consistent with CYP2A, CYP2G1, and CYP2F. The upper (54-kDa) immunoreactive band contained the largest number of peptide fragments from CYP2A, and the only positive identification for CYP2G1. CYP2F was found exclusively in the lower (≈50-kDa) band along with fragments corresponding to CYP2A. Fewer fragments for CYP2A were observed in the lower band compared with the upper band. These findings agree with previous studies suggesting that CYP2A5 and CYP2G1 isoforms account for up to 35% of olfactory mucosal P450 content in mice (Gu et al., 1998). The sequence homologies of CYP2A5 and CYP2G1 with CYP2F2 (52 and 55%, respectively) suggest that cross-reactivity is likely. Although the present methodology has some limitations, it allows for the determination of the predominant nasal P450 isoforms and provides a basis for quantifying the appropriate immunoreactive band associated with CYP2F. However, the number of peptide fragments generated from numerous proteins in each gel slice precludes the development of a complete picture of the content of nasal P450 isoforms. Only three P450 isoforms present in rodent olfactory epithelium (reviewed in Thornton-Manning and Dahl, 1997a) were identified.
The greatest amount of CYP2F was observed in the ethmoturbinates and maxilloturbinates of both rodent species. Similarly, immunohistochemical studies have localized CYP2A3 in both olfactory and respiratory epithelia of the rat (Thornton-Manning et al., 1997b). In comparison, CYP2G1 protein is present exclusively in the olfactory epithelium of both mouse (Hua et al., 1997) and rat (Zupko et al., 1991).
Within each rodent species, the comparison of relative CYP2F protein expression levels in the nasal epithelium and tracheobronchial airways revealed expression patterns consistent with both in vitro metabolism studies and cell injury observed in vivo. In contrast to the lung where mouse/rat differences in CYP2F expression are dramatic, values observed in the nasal epithelia of mice and rats were more similar, making the explanation of tissue susceptibility less obvious. Although observation of 2-fold greater amounts of CYP2F protein in murine nasal epithelia is consistent with previous metabolism studies (Buckpitt et al., 1995), it fails to explain the higher susceptibility of rat nasal olfactory epithelium compared with mouse. The reason for the lack of correlation between catalytic activities and levels of CYP2F protein with the susceptibilities observed in the olfactory epithelium of these rodent species is unclear. Potential explanations include 1) more naphthalene (parent compound) reaches the olfactory epithelium of the rat, 2) other P450 enzymes are important in naphthalene activation in the rat, or 3) differences in olfactory detoxification mechanisms (e.g., glutathione levels) exist across these species.
Given the strong association between CYP2F expression levels and susceptibility to naphthalene-induced cytotoxicity demonstrated in rodents, studies were extended to the rhesus macaque, a species that has a similar pulmonary physiology and morphology to the human. Of all the tissues tested in the rhesus macaque, only the ethmoturbinates yielded a CYP2F signal. In contrast with the two distinct bands observed in rodents (Fig. 8), ethmoturbinate microsomal proteins from the rhesus macaque contained a single diffuse immunoreactive band. To confirm the identity of the immunoreactive ethmoturbinate protein(s), peptide mass fingerprinting was utilized. An analysis of the wide area around the single immunoreactive band yielded tryptic fragments matching CYP2A13/2A6 and potentially CYP2J2 and CYP2F. The compliment of P450s identified here along with the absence of CYP2G1 is nearly identical to the human olfactory epithelium (Gu et al., 2000; Su et al., 2000). Although the peptide mass fingerprinting methodology employed is not considered quantitative, our observation of more CYP2A13-specific peptide fragments compared with CYP2A6 is consistent with quantitative RT-PCR results demonstrating CYP2A13 transcript to be more abundant than CYP2A6 in the human olfactory epithelium (Chen et al., 2003).
Quantification of the immunoreactive band observed in monkey ethmoturbinate tissue as CYP2F was based on 1) an apparent molecular weight identical to rCYP2F standards, 2) peptide mass fingerprinting demonstrating CYP2A3 and CYP2A5 to have larger apparent molecular weights than rCYP2F standards, 3) identification of two unique tryptic fragments consistent with a CYP2F isoform, and 4) identification of a monkey ethmoturbinate CYP2F transcript (mRNA) using RT-PCR (R. M. Baldwin, unpublished results). Assuming a reasonably cross-reactive antibody, a conservative immunoblot limit-of-detection estimate provides that all areas devoid of a CYP2F signal have more than a 40- or 160-fold differential in CYP2F expression compared with rat ethmoturbinates or mouse terminal bronchioles. Even if CYP2F is entirely absent from this primate species and the signal observed in ethmoturbinates is due to cross-reactivity with CYP2A, the conclusion that there are dramatic rodent/primate differences in CYP2F expression remains unchanged.
Lung microsomal incubations have demonstrated the rhesus macaque to metabolize naphthalene at a rate roughly 100- and 10-fold less than mice and rats (Buckpitt et al., 1992). Those data and the data presented here demonstrating no detectable CYP2F in any of the lung subcompartments tested suggest the rhesus macaque to be refractory to naphthalene-induced pulmonary toxicity. Even though assessment of nasal susceptibility is less clear, the magnitude of the rhesus-to-rodent differential in olfactory CYP2F expression would also suggest a lack of susceptibility for the monkey.
The low rates of naphthalene metabolism observed in human lung microsomes (Buckpitt and Bahnson, 1986) and in the rhesus macaque (Buckpitt et al., 1992) suggest that rodents do not accurately predict human pulmonary response to naphthalene exposure. Ongoing kinetic studies using rodent and rhesus nasal microsomes and isolated airway segments of the rhesus are designed to provide additional comparative data for extrapolating the potential toxicity of naphthalene in humans. Ultimately, unequivocal determination of primate susceptibility to naphthalene will require in vivo testing and the ascertainment of pulmonary and nasal injury.
Footnotes
-
This work was supported by National Institute of Environmental Health Sciences (NIEHS) Grants ES08408, ES04311, ES06700, ES04699, and National Center for Research Resources Grant RR00169. The University of California at Davis is an NIEHS Center (05707), and support for core facilities used in this work is gratefully acknowledged.
-
Article, publication date, and citation information can be found at http://jpet.aspetjournals.org.
-
DOI: 10.1124/jpet.103.062901.
-
ABBREVIATIONS: P450, cytochrome(s) P450; RT-PCR, reverse transcription-polymerase chain reaction; bp, base pair(s); GAPDH, glyceraldehyde-3-phosphate dehydrogenase; ECL, enhanced chemiluminescence; TBST, Tris-buffered saline/Tween 20; RFU, relative fluorescence unit; MALDI-TOF, matrix-assisted laser desorption ionization/time of flight.
- Received November 11, 2003.
- Accepted December 22, 2003.
- The American Society for Pharmacology and Experimental Therapeutics