INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Animal models have been developed to attain greater understanding of occupational sensitizers that cause occupational asthma (1). Toluene diisocyanate (TDI) is a low molecular weight chemical, well known as occupational sensitizer (2). Recently, we developed an animal model of immunization to TDI in the guinea pig (3). In this model, we found that intradermal injections of TDI induced a specific antibody response, and an inflammatory process in both central and peripheral airways. The key cells were CD4+ T cells, eosinophils, and mast cells, and the greatest increase in inflammatory cells occurred 6 h after challenge. The administration of anti-IL-5 antibody inhibited the TDI-induced eosinophil influx, suggesting that IL-5 was an important cytokine in the development of eosinophilia in this model. A role for tachykinins in TDI- induced airway hyperresponsiveness and airway inflammation has been suggested by in vivo and in vitro animal studies (4). There are several lines of evidence that neurogenic inflammation may be important in animal models in response to irritants or chemical sensitizers (7), especially in rodents where tachykinin effects are more pronounced. Bozic and coworkers (10) have recently published data that demonstrated a critical role for the release of substance P and interaction with NK-1R in immune-complex-mediated injury in the mouse lung. These data heightened interest in studying the tachykinins as being critical in other inflammatory models such as asthma.

In this study, we immunized animals to TDI and investigated whether immunization and challenge with TDI caused an antibody response, an inflammatory process, and changes in neuropeptides (substance P and calcitonin gene-related peptide) in the guinea pig central airways.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Design, Immunization, and TDI Challenge

Male Dunkin Hartley guinea pigs (Rodentia Laboratories, Italy) weighing 250 to 300 g were used. They received three weekly intradermal injections of 50 µl of 100% TDI (2,4 to 2,6 isomers; ratio, 80:20) or saline into each of two dorsal sites. Seven days after the third injection of TDI or saline, animals were exposed to TDI in a glass chamber (30 L). TDI was generated by blowing air over the surface of TDI in a 5-ml glass wash bottle; airflow was maintained at 2 L/min. TDI was monitored continuously by MDA Model 7005 isocyanate detection equipment (MDA Scientific Inc., Glenview, IL). The exposure was continued for 15 min and was kept between 5 and 20 ppb. In this study 24 animals were used. Nonimmunized animals included two groups of animals: the first group (n = 6) received intradermal injections of saline and was not challenged with TDI; the second group (n = 6) received injections of saline and was challenged with TDI. Immunized animals (n = 12) received intradermal injections of TDI; one group was challenged with TDI (n = 6) and one group (n = 6) was not challenged with TDI. Animals were killed 6 h after the end of exposure to TDI. All guinea pigs were killed by intraperitoneal injection of pentobarbital sodium (100 mg/kg). All animals survived exposure to TDI.

Serum Antibody Titer

Blood samples were collected from the jugular vein immediately after the administration of the lethal dose of pentobarbital sodium, and the serum was obtained to measure the titer of TDI-specific antibodies. TDI-specific IgG and IgG₁ were measured in the serum samples by enzyme-linked immunosorbent assay (ELISA). Results were expressed as geometric mean and GSEM.

Tissue Processing

Immediately after blood sampling, the chest was opened and the lungs were excised. One lung from each animal was inflated with 4% formaldehyde via the main bronchus and fixed for 4 h at 20 cm H₂O. Lung blocks were embedded in paraffin. Sections 5 µm thick were stained with hematoxylin-eosin to quantify eosinophil infiltration. Cell infiltration was studied in the wall of central intrapulmonary airways (segmental bronchi with diameters of about 1,000 µm). Cells were counted in a 50-µm-thick layer of the submucosa and expressed as number of cell/mm².

The other lung was distended with 3 ml of 1% paraformaldehyde in phosphate-buffered saline (PBS) 0.01 M at pH 7.4 via the main bronchus and fixed at 4° C for 6 h. Then the lung was transferred to PBS containing 0.45 M sucrose and 0.01% sodium azide for at least 12 h at 4° C. For each animal, three cryostat lung blocks were prepared by immersion in liquid nitrogen. Frozen sections (10 µm thick) were taken up on poly-L-lysine-coated microscope slides and stained with rabbit antisera, utilizing the avidin-biotinylated complex (ABC) peroxidase method. Antiserum to PGP 9.5 (Ultraclone, Isle of Wight, UK) (11) and antisera raised to the neuropeptides CGRP and SP (raised in rabbits at the Royal Postgraduate Medical School, Hammersmith Hospital, London, UK) (12) were used. The antiserum to SP was used as a marker of nerve fibers containing tachykinins, since it cross-reacts with other tachykinins. One section per block was stained for each antigen. In brief, endogenous peroxidase activity was blocked by immersing slides in 0.03% hydrogen peroxide in methanol for 1 h. After washing in PBS (3 × 10 min), nonspecific binding was blocked by a 30-min incubation in 3% normal swine serum in PBS containing 0.05% bovine serum albumin (BSA) and 0.1% sodium azide (BSA solution). Sections were incubated overnight at 4° C with primary antisera in BSA solution at predetermined optimal dilutions (PGP 9.5, 1:20,000; SP, 1:4,000; CGRP, 1:5,000). Negative controls were performed by omission of the primary antibody. Moreover, preabsorption with a specific antigen was performed for antisera to neuropeptides. After washing in PBS (3 × 10 min), the sections were incubated successively for 30 min with biotinylated swine antirabbit immunoglobulin (E431; DAKO A/S, Glostrup, Denmark) and streptavidin and biotinylated horseradish peroxidase reagent (K377; DAKO) for 60 min. The sections were rinsed first in PBS (3 × 10 min), then in acetate buffer (0.1 M at pH 6.0). The peroxidase activity was revealed using the nickel enhancement method (13). Sections were dehydrated, mounted in Pertex, and examined with a light microscope (Olympus BH-2; Olympus Corp., London, UK).

Quantitative and Statistical Analysis

Cell infiltration was studied in the wall of central intrapulmonary airways (segmental bronchi with a diameter of about 1,000 µm). Bronchi with diameters of 1,000 µm were obtained in 10 nonimmunized and in nine immunized animals. Cells were counted in a 50-µm-thick layer of the submucosa and expressed as cells/mm².

Quantitative analysis of immunoreactive nerve fibers was performed in central intrapulmonary airways with a diameter of about 1,000 µm, using the method described by Cowen (14). Images of appropriate airways were captured by an operator who was unaware of the identity of the specimens, and they were digitized by a Seescan Symphony image analysis system (Seescan, Cambridge, UK) via a video camera (Sony XC77CE, Milan, Italy) attached to a microscope (Model BH-2; Olympus). Digital images were 512 × 512 pixels (picture points), each pixel represented by a grey value from 0 to 255 equivalent to the intensity of the light in the corresponding area of the microscopic image. The area of the airway wall (Aw) was measured in the image by interactive delineation. The boundaries comprised by the luminal (surface of the epithelium) and the adventitial surfaces were traced using the computer mouse. The number of pixels contained within the boundaries was converted to the tissue area using a factor determined by calibrating the image analyzer with a series of area standards viewed with the same microscope and objective used for the study.

The area of immunostained nerves (An) within the airway wall boundaries was determined by the same method but using a different technique for delineation: interactive thresholding. For this, the image was segmented to separate nerves from non-nerve tissue by selecting a threshold grey value that is intermediate between the intensities of each. The image analyzer highlighted the pixels having a grey value smaller than that of the threshold in an artificial color, altered until all the nerves were selected, and their area was then calculated.

The results were expressed as the area density of nerves, i.e., An/ Aw. For each peptide, five fields per bronchus (three bronchi per animal) were measured. Each reported value is mean ± SEM. The Mann-Whitney U test was used to compare differences between unchallenged and challenged nonimmunized animals, between unchallenged and challenged immunized animals, and between unchallenged nonimmunized and unchallenged immunized animals. Correlation coefficients were calculated using Spearman's rank method. Probability values < 0.05 were considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

IgG and IgG₁ Titers

Specific antibodies to TDI were observed only in animals that received injections of TDI. In immunized animals, the IgG and IgG₁ titers were, respectively, 23,800 (1.90) and 9,060 (1.38), whereas in immunized TDI-challenged animals they were 27,959 (1.59) and 8,360 (2.75) (geometric mean and GSEM).

Histopathologic Examination

The number of mononuclear cells in the submucosa of central airways was similar in nonimmunized and immunized animals (869.8 ± 133.4 and 719.8 ± 102.4 cells/mm², respectively). Instead, the numbers of eosinophils were significantly different between the two groups (37.2 ± 1.4 and 46.8 ± 2.6 cells/mm², respectively, p < 0.01). Challenge with TDI did not modify the numbers of mononuclear cells (869.8 ± 133.4 and 713.4 ± 45.0 cells/mm², respectively) and eosinophils (37.2 ± 1.4 and 37.2 ± 5.8 cells/mm², respectively) in nonimmunized animals, whereas in immunized animals it caused a further increase in eosinophils (from 48.6 ± 2.6 to 64.4 ± 7.5 cells/mm², p < 0.05). In the immunized and TDI-challenged animals, we found that, in comparison with nonimmunized animals, the number of eosinophils in the airways (Table 1, Figure 1) was significantly elevated (64.4 ± 7.5 and 37.2 ± 1.4 cells/mm², respectively, p < 0.01).

View larger version (116K): [in this window] [in a new window]   View larger version (109K): [in this window] [in a new window]   Figure 1.   Eosinophils (arrowheads) in the airways with diameters of about 1,000 µm of immunized TDI-challenged (a) and nonimmunized (b) guinea pigs at 6 h after exposure to TDI (hematoxylin-eosin stain; light microscopy; original magnification: ×400).

Immunohistochemistry Examination

Airway wall. The bronchiolar wall area was similar in all the groups examined (Table 2).

PGP 9.5. PGP 9.5-immunoreactive nerve fibers were abundant and they were seen in the epithelium, smooth muscle, and around blood vessels (Figure 2a and b). The quantitative analysis is shown in Figure 3. TDI exposure did not change the overall innervation in both nonimmunized and immunized animals. The density of PGP 9.5-immunostained nerve fibers was also similar between nonimmunized and immunized animals (0.0119 ± 0.008 and 0.0098 ± 0.0019 An/Aw, respectively), but it was significantly different between nonimmunized and immunized TDI-challenged animals (0.0119 ± 0.008 and 0.00683 ± 0.0006 An/Aw, respectively, p < 0.05). When all the animals were considered together, the number of eosinophils in the submucosa was inversely correlated with the density of PGP 9.5-positive nerves (Rho = 0.514; p < 0.05) (Figure 4).

View larger version (112K): [in this window] [in a new window]   View larger version (115K): [in this window] [in a new window]   Figure 2.   Immunostaining of airways with diameters of about 1,000 µm for protein gene product 9.5 (PGP 9.5) in immunized TDI-challenged (a) and nonimmunized (b) guinea pigs. Original magnification: ×100.

View larger version (9K): [in this window] [in a new window]   Figure 3.   Individual counts for PGP 9.5-nerve fiber density in airways with diameters of about 1,000 µm of nonimmunized (NI), nonimmunized TDI-challenged (NI + TDI), immunized (I), and immunized TDI-challenged (I + TDI) animals. The results are expressed as the area density of the nerves, i.e., area of immunoreactive nerves (An)/area of the airway wall (Aw). Horizontal bars represent median values.

View larger version (12K): [in this window] [in a new window]   Figure 4.   Relationship between the number of eosinophils in the submucosa of central airways and the density of PGP 9.5-, SP-, and CGRP-immunoreactive nerves. Open circles = nonimmunized animals; closed circles = immunized TDI-challenged animals. Rho and p values correspond to Spearman's rank correlation coefficient.

Tks (SP). TK-immunoreactive nerve fibers were seen in the epithelium, smooth muscle and around blood vessels, and they were not different between nonimmunized and immunized animals (0.0040 ± 0.0004 and 0.0029 ± 0.0003 An/Aw, respectively). TDI exposure significantly decreased the density of TK-positive nerves in nonimmunized animals (p < 0.05), whereas in immunized animals, there was a trend for the density of TK-positive nerves to be decreased (p < 0.08) (Figure 5). Immunized TDI-challenged animals had a significant decrease (p < 0.05) in the density of TK-positive nerves, as compared with nonimmunized animals (0.0018 ± 0.0004 and 0.0040 ± 0.0004 An/Aw, respectively, p < 0.05).

View larger version (11K): [in this window] [in a new window]   Figure 5.   Individual counts for TK-nerve fiber density in airways with diameters of about 1,000 µm of nonimmunized (NI), nonimmunized TDI-challenged (NI + TDI), immunized (I), and immunized TDI-challenged (I + TDI) animals. The results are expressed as the area density of the nerves, i.e., area of immunoreactive nerves (An)/area of the airway wall (Aw). Horizontal bars represent median values.

When all the animals were considered together, the number of eosinophils in the submucosa was inversely correlated with the density of TK-positive nerves (Rho = 0.547, p < 0.05) (Figure 4).

CGRP. The distribution of CGRP-immunoreactive nerves was similar to that described for SP. Nonimmunized and immunized animals had a similar CGRP-nerve density (0.0018 ± 0.0001 and 0.0017 ± 0.0002 An/Aw, respectively). Although TDI exposure did not modify the density of CGRP-positive nerves in noimmunized animals, there was a significant decrease in the density of CGRP-immunoreactive nerves after TDI challenge in immunized animals (p < 0.05) (Figure 6). Moreover, in immunized TDI-challenged animals, there was a significant decrease in the density of CGRP-positive nerves as compared with nonimmunized unchallenged animals (0.0009 ± 0.0001 and 0.0018 ± 0.0001 An/Aw, respectively, p < 0.01). There was a negative correlation between airway inflammation, assessed as eosinophil infiltration, and CGRP nerve density (Rho = 0.633, p < 0.01) in immunized TDI-challenged animals. Likewise, the correlation was present when all animals were considered together (Rho = 0.682, p < 0.01 (Figure 4).

View larger version (10K): [in this window] [in a new window]   Figure 6.   Individual counts for CGRP-nerve fiber density in airways with diameters of about 1,000 µm of nonimmunized (NI), nonimmunized TDI-challenged (NI + TDI), immunized (I), and immunized TDI-challenged (I + TDI) animals. The results are expressed as the area density of the nerves, i.e., area of immunoreactive nerves (An)/area of the airway wall (Aw). Horizontal bars represent median values.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

This study has demonstrated that immunization and challenge with TDI result in a decrease in the density of nerves immunoreactive for SP and CGRP in guinea-pig airways. The overall innervation in the lung tissue was similar in nonimmunized and immunized animals, as suggested by the unchanged immunoreactivity for PGP 9.5, but it was decreased in immunized TDI-challenged as compared with nonimmunized animals. Moreover, intradermal injections of pure TDI induced a specific antibody response and an inflammatory process, represented by an increased number of eosinophils in the submucosa of central airways. The density of PGP 9.5-, CGRP-, and SP-immunoreactive nerves was negatively correlated with airway inflammation, as assessed by eosinophil cell infiltration.

These findings confirmed the validity of our model, to study both immunopathologic and nerve changes induced by exposure to TDI in the guinea pig airways. The data collected in this study are consistent with an earlier report from our laboratory. This study showed that the eosinophil is a key cell in this model: eosinophils were increased in central and peripheral airways of immunized compared with nonimmunized animals, and interleukin-5 (IL-5) was essential for the activation of eosinophils (3). Studies using other models of inflammation have also shown that IL-5 is a pivotal component of the inflammatory pathways that involve eosinophils and the execution of effector functions (15). Eosinophil infiltration of bronchial mucosa is a common finding in bronchial asthma. An important role for the eosinophil has been described in patients who died during an asthma attack; 40% of airway nerve bundles had eosinophils associated with them: eosinophils and extracellular major basic protein were identified in close association with nerve fibers in the lamina propria and in the smooth muscle (16). A close proximity between mast cells and sensory nerves in the airways has also been described, suggesting a critical role for neuropeptides in the regulation of the function of inflammatory and immune cells (17). The different techniques used in this study to quantify eosinophils and nerve density did not allow to show whether eosinophils were associated with airway nerve bundles. However, we have previously shown that, in this model, eosinophilia was an important component of the airway inflammatory process (3), and there is ongoing evidence that eosinophils play a critical role in the pathophysiology of asthma (18). In different experimental preparations, we provided evidence that TDI activates the efferent function of capsaicin-sensitive sensory nerves (19), and that it is likely that the action of TDI on sensory nerves is indirect through the release of prostanoids (22). Recently, a murine model to investigate TDI-induced asthma has been reported (23). In this model, an important role for sensory nerves and neuropeptides in the induction of tracheal hyperreactivity after exposure to TDI has been found (24). Sensory nerves seem to be important targets in both in vitro and in vivo models of TDI-induced asthma in several species. Furthermore, decreased neutral endopeptidase activity occurs after exposure to TDI, resulting in exaggerated neurogenic inflammation and suggesting an important role in the airway inflammatory process observed after TDI exposure (25). Sensory neuropeptides mediate contraction of the airway smooth muscle, mucus secretion, and plasma exudation, which are features of bronchial asthma (26). Moreover, neuropeptides released in the airways are able to stimulate inflammatory and immune cells. There is increasing evidence that the immune and nervous systems are functionally linked and involved in a bidirectional communication. Stimuli such as pollutants, viruses, and antigens are not detected directly by the nervous system, but they may be transmitted indirectly by mediators. On the other hand, the nervous system detects inflammatory stimuli, the signal may be transmitted by mediators that are recognized by receptors on leukocytes. The links between the two systems are the neuropeptides. They can increase the expression of adhesion molecules on vascular endothelium and stimulate human T- and B-lymphocytes, monocytes, eosinophils, and fibroblasts (27- 30). Moreover, tachykinins may activate alveolar macrophages to release inflammatory cytokines such as interleukins 1 and 6 (IL-1, IL-6) and tumor necrosis factor-alpha (31). Neurogenic inflammation may contribute to bronchial asthma, and the peptides released from sensory nerves may be considered proinflammatory. Indeed, it has been shown that neuropeptides such as substance P and CGRP play a significant role in eosinophil infiltration by priming cells in allergic inflammation (32). Two different mechanisms have been proposed in eosinophil activation: receptor-mediated and nonspecific nonreceptor-mediated activation (31). In fact, substance P was able to induce eosinophil peroxidase from eosinophils that was not inhibited by a SP receptor antagonist, supporting the existence of a nonspecific activation of eosinophils. These cells express the gene for SP and store the immunoreactive peptide (33), and SP increases the expression of surface receptors on eosinophils (34), supporting the hypothesis that this peptide may be involved in priming or tuning of eosinophil exocytosis.

In the present study, both tachykinins and CGRP seem to play a role in the response of guinea pig airways to the exposure to TDI. This study does not establish the exact mechanism through which TDI produces neuropeptide depletion in the sensory neurons, and whether the loss of immunoreactivity is persistent or whether its magnitude increases over time. Moreover, this study does not rule out the possibility that TDI acts as an irritant that activates nociceptive neurons leading to their "degranulation" or that eosinophilia and the decreases in nerve densities may represent two unconnected processes that occur after immunization and TDI challenge. However, the relationship between airway inflammation (i.e., eosinophil cell infiltration) and nerve density suggests a link between the nervous and immune system and an interaction between neuropeptides and inflammatory cells. CGRP colocalizes with and is coreleased with substance P from sensory nerves. CGRP- and SP-containing nerves have been found to be markedly depleted in other animal models after treatment with either acrolein or cigarette smoke (8, 9). Both sensory and sympathetic fibers are depleted in the synovium of rheumatoid arthritis, and nerve fiber depletion was only seen in areas of inflammatory cell infiltration, indicating that a mixed lymphocyte and macrophage population of cells was necessary to cause depletion of the finely myelinated and unmyelinated neuropeptide-containing nerves (35). Recent studies have documented that the eosinophil is a key cell in many inflammatory disorders since these cells synthesize several proinflammatory and cytotoxic mediators such as basic proteins, hydrolytic enzymes, lipid mediators, cytokines, oxygen metabolites, and neuropeptides (36, 37). Moreover, in rats, it has been shown that CGRP is capable of causing eosinophilia in the lung in vivo, and it may contribute to airway inflammation (38). In our model, both the reduction in the number of nerves and the release of peptides may be involved since the immunoreactivity for PGP 9.5 (a general neural antigen) was reduced after immunization and TDI challenge. Our results support a link between eosinophils and neuropeptides. However, the release of neuropeptides might, in turn, be induced by mast cell degranulation (30) since we have shown that in this model both mast cells and eosinophils were increased in the submucosa of central airways of immunized TDI-challenged animals (3). Indeed, a reduction in substance P-like immunoreactivity has been documented in asthma (39), suggesting the depletion of neuropeptide stores from sensory nerves. Previous studies have shown that TDI induces decreased neutral endopeptidase (NEP) activity (25), which may contribute to the increased release of neuropeptides since NEP is an enzyme that degrades neuropeptides. Thus, these findings are consistent with the hypothesis that the neuroimmune interactions in the lung could also include sensitization of capsaicin-sensitive C-fiber afferents through inflammatory mediators.

In summary, our data show that exposure to TDI in the immunized guinea pig causes TDI-specific IgG antibodies, airway inflammation, and changes in nerve density that occur 6 h after TDI challenge. This model may shed further advances in our knowledge in the field of neuroimmunology and in our understanding about the mechanisms involved in isocyanate-induced asthma.