The effect of aerosolized bacterial lysate on experimentally induced Mannheimia haemolytica pneumonia in calves

Abstract

Bovine respiratory disease (BRD) often occurs during specific periods of increased susceptibility when stress, viral infection, or reduced air quality are thought to suppress respiratory defences. The innate immune system is rapidly responsive and broadly protective and could be a target for preventing BRD during these periods of increased susceptibility. This study tested the hypothesis that stimulation of pulmonary innate immune responses by aerosol delivery of a lysate of killed Escherichia coli and Staphylococcus aureus bacteria would protect calves against Mannheimia haemolytica pneumonia. Ten clean-catch colostrum-deprived Holstein calves were randomly assigned to receive either aerosolized bacterial lysate or saline 24 hours before M. haemolytica challenge. Effects of this treatment on clinical, hematologic, microbiologic, and pathologic outcomes were assessed. Compared to controls, lysate-treated calves had lower serum haptoglobin and blood leukocyte and neutrophil concentrations following M. haemolytica challenge. There were no differences in temperature, heart and respiratory rates, clinical scores, ultrasound lesions, or number of M. haemolytica in the nasal cavity or lung. Thus, treatment with bacterial lysate prior to M. haemolytica challenge appeared to ameliorate early measures of inflammation but did not provide sufficient protection to substantially alter the course of disease.

Introduction

Bovine respiratory disease (BRD) is an economically important disease in beef, dairy, and veal calves. Metaphylaxis with antibiotics is commonly used to prevent BRD but alternative control strategies are needed to reduce the use of antimicrobials in food production. Suppression of immune defences by stress and viral infection has been considered the mechanism by which risk factors contribute to BRD (1). Stimulation of innate immune responses during high-risk periods can rapidly induce broad-spectrum protection against pathogens. In mouse experiments, stimulation of pulmonary innate immune responses prior to pathogen challenge reduced respiratory disease severity (2–5). For example, aerosolization of non-typeable Haemophilus influenzae before pathogen challenge protected against fatal bacterial challenge, with maximum protection in the first 24 h (3). Thus, we investigated the effects of stimulating an innate immune response on the subsequent development of experimentally induced M. haemolytica pneumonia. Previous experiments in calves characterized systemic and pulmonary responses following aerosolization of the lysate of Staphylococcus aureus and Escherichia coli bacteria used in the current study, including increased rectal temperatures, plasma acute phase proteins, recruitment of neutrophils into bronchioles and alveoli, and changes in bronchoalveolar lavage fluid cytokines and other proteins related to platelet activation and complement pathways (6). We hypothesized that stimulation of innate immune responses using this aerosolized bacterial lysate would protect calves against Mannheimia haemolytica pneumonia.

Materials and methods

Study design

Ten clean-catch colostrum-deprived calves were treated by aerosol with either a lysate of killed E. coli and S. aureus or saline, 24 h prior to aerosol M. haemolytica challenge (Supplemental Figure S1). Measured outcomes included the proportion of diseased lung, clinical parameters, clinical disease severity scores, M. haemolytica numbers in lung and nasal cavity, and blood leukocyte and acute phase protein concentrations.

Animals and aerosol challenge

Use of animals was approved by the Animal Care Committee of the University of Guelph. Ten Holstein bull calves were obtained at birth with no contact with the dam, did not receive colostrum, and were raised in controlled-access rooms in groups of 2 to 3. Calves were fed evaporated milk and water for the first 48 h, then milk replacer, and offered hay and unmedicated calf starter from 1 wk of age (7).

At 4 wk of age, experiments were conducted in age-matched pairs, with 1 calf randomly assigned to receive aerosolized bacterial lysate (n = 5) or phosphate-buffered saline (PBS, n = 5). Bacterial lysate consisted of S. aureus and E. coli [equal numbers of colony-forming units-(CFU)-equivalents] that were killed at 65°C (confirmed by lack of growth on agar), centrifuged, and sonicated (6). Gram-positive and Gram-negative bacteria were used for broad stimulation of an innate immune response; M. haemolytica was not used to avoid effects of a specific acquired immune response. On Day 0, all calves received an aerosol of 10⁹ CFU-equivalents of bacterial lysate in 10 mL PBS (innate immune stimulant) or 10 mL of PBS alone (control) as described (6).

Twenty-four hours later, all calves were challenged by aerosol with M. haemolytica (7). Mannheimia haemolytica (isolate B158, genotype 2, serotype 1) was grown to log phase, quantified based on OD₆₀₀, washed in PBS, and kept on ice (7). Purity was confirmed by Gram stain and MALDI-TOF mass spectroscopy (Animal Health Laboratory). The targeted challenge dose was 1 × 10¹⁰ CFU in 10 mL; the actual challenge dose was later determined by colony counts on blood agar (Supplemental Table S1).

Clinical monitoring and antemortem sampling

Clinical data were collected at baseline and at least every 12 h including rectal temperature, heart and respiratory rates, clinical score [Supplemental Table S2; (7)]; spontaneous and induced coughing were also recorded. Blood was sampled daily, and complete blood (cell) counts and serum haptoglobin and plasma fibrinogen concentrations were measured (Animal Health Laboratory). Baseline serum antibody to M. haemolytica leukotoxin was measured as described (7).

Daily systematic thoracic ultrasound examination was performed (Ibex Pro, E. I. Medical, Loveland, Colorado, USA), frequency of 6.2 MHz, depth set at 6 cm, with 18 dB gain near 31 dB, far 35 dB (8). The number of regions in each intercostal space (dorsal, middle, and ventral) with consolidated lung lesions were recorded, excluding dorsal and middle regions subjacent to the scapula. The daily ultrasound score was the total number of regions with consolidation (maximum score: 34).

Bilateral deep nasal swabs were collected at baseline and once daily and stored in transport medium; growth of M. haemolytica in aerobic cultures was subjectively scored (1+, 2+, 3+ or 4+; Animal Health Laboratory). Bronchoalveolar lavage fluid was collected from 4/5 saline-treated and 4/5 lysate-treated calves immediately prior to administration of M. haemolytica. Total and differential cell counts and concentrations of cytokines were measured in bronchoalveolar lavage fluid as described (6).

Study endpoints and postmortem assessments

On Day 4, all surviving calves were euthanized. Calves were euthanized before this endpoint if they developed dyspnea (score = 3), reduced strength (score ≥ 3), depression (score ≥ 3), or anorexia at 2 consecutive feedings (Supplemental Table S2). Gross lung lesions were evaluated by a single blinded researcher, as described (7). The percentage of lungs with pneumonia was estimated by visual inspection and palpation and confirmed by histologic examination of left and right cranioventral, caudoventral, and caudodorsal lung lobes. Lung:heart weight ratios were measured. Using photographs of the lateral surfaces of the left and right lungs, the proportion of lung consolidation (surface area) was quantified by image analysis (ImageJ software). Samples were collected from the left and right cranioventral, caudoventral, and caudodorsal lungs and accessory lobe; for each of these areas, the sampling was standardized to harvest lesions when present, or normal lung when no lesions were present. Formalin-fixed tissues were routinely processed and stained with hematoxylin and eosin (H&E), and a blinded evaluator assigned histopathologic scores (0 = absent; 1 = present, but rare; or 2 = present, moderate to severe) for bronchopneumonia, oat cells (leukocytes with streaming nuclear material), coagulative necrosis, pleuritis, and fibrin. For quantitative bacterial culture, serial dilutions were prepared from 100 mg of homogenized lung, plated on blood agar, incubated at 37°C for 48 h, and the number of CFU/g of lung tissue was determined.

Statistical analysis

Paired evaluation of differences between baseline values and single point-in-time outcomes (including postmortem parameters) were examined using paired t-tests and Chi-squared tests (GraphPad, Prism 8). Simple linear regression was used to correlate various outcomes. Data were tested for normality of distribution using the Shapiro-Wilk and Kolmogorov-Smirnov tests and transformed if indicated. Associations were considered significant at P-values ≤ 0.05.

For evaluation of repeated measures, mixed effects multivariate models were built and analyzed using STATA 15.1 (StataCorp, College Station, Texas, USA). Models included a random intercept for individual animals to account for repeated measurements taken from the same animal, and for pairs of animals to account for the effects of challenge group. Statistical analysis was performed on time points following M. haemolytica challenge up to 2 d after challenge. Statistical analysis of Day 3 data was not possible due to non-random missing data resulting from euthanasia of severely affected animals. Mixed-model linear regression was used to evaluate outcomes with repeated measures. Haptoglobin, fibrinogen, blood leukocytes, and blood neutrophil data were log-transformed. In addition to treatment, the day relative to challenge, baseline parameters, and interaction between treatment and day were included, with stepwise removal of non-significant variables (P > 0.05). Exclusion or inclusion of baseline data in the models did not affect the statistical significance of treatment.

Results

Baseline parameters

Prior to administration of the bacterial lysate, there were no significant differences between treatment groups in any evaluated parameters. For antibodies to M. haemolytica leukotoxin, all calves were within the range previously determined to be negative (Supplemental Table S1) (7) and enzyme-linked immunosorbent assay (ELISA) values were similar between groups (P = 0.74). Prior to challenge, none of the nasal swab cultures yielded M. haemolytica, and calves had no evidence of clinical disease or lung consolidation on ultrasound examination.

Response to aerosolized bacterial lysate

Aerosol delivery of bacterial lysate or saline did not result in clinical signs. One day after aerosolization (before M. haemolytica challenge), mean rectal temperatures were slightly higher in lysate-treated versus control calves [mean: 39.5; 95% confidence interval (CI): 39.1, 39.8 versus 39.0°C (95% CI: 38.6, 39.5); P = 0.025]. Lysate-treated calves had higher concentrations of plasma fibrinogen compared to saline-treated animals (P = 0.043) at 1 d after aerosolization of lysate or saline compared to baseline. At this time, serum haptoglobin concentrations and blood leukocyte counts were not significantly different between groups. In bronchoalveolar lavage fluid at 24 h after aerosolization of lysate or saline, calves receiving bacterial lysate had higher percentage of neutrophils compared to calves receiving saline [20.0% (95%CI: 9.4, 30.6) versus 7.3 % (95% CI: 0, 17.1); P = 0.031], higher IL-8 concentrations [20.3 pg/L (95% CI 1.7, 14.0) versus 7.8 pg/L (95% CI: 0, 40.7); P = 0.11], and lower IL-10 concentrations [3.5 pg/L (95% CI: 2.6. 4.3) versus 6.6 pg/L (95% CI: 1.0, 12.2); P = 0.13] (Figure 1). Other measured cytokines were generally below the limit of quantification: IL-17 (< 2.44 pg/mL), IFN-α (< 9.8 pg/mL), IL-1β (< 9.8 pg/mL), IL-2 (< 39.1 pg/mL), IL-6 (< 156.3 pg/mL), TNF-α (< 39.1 pg/mL), and IFN-γ (< 2.4 pg/mL).

Figure 1.

Leukocytes and cytokines in bronchoalveolar lavage fluid (BALF) of control and lysate-treated calves 24 h after calves were aerosolized with a lysate of killed E. coli and S. aureus bacteria or with saline (control). Samples were collected prior to M. haemolytica challenge. The data show the mean and individual-animal data points; n = 4 per group. A — The percentage of neutrophils, macrophages, and lymphocytes. *P ≤ 0.05; **P ≤ 0.01. B — Interleukin-8. C — Interleukin-10.

Responses to Mannheimia haemolytica challenge

After M. haemolytica challenge, clinical signs of cough, depression, increased respiratory effort and rate, fever, inappetence, and recumbency increased over time (P < 0.003) (Figure 2, Table I, Supplemental Figures S2–S3). Calves given bacterial lysate prior to M. haemolytica challenge had numerically lower clinical scores than those receiving saline, but the difference was not significant (P = 0.272, Figure 2).

Figure 2.

Changes in clinical scores over time in control and lysate-treated calves, before and after challenge with M. haemolytica. Calves received aerosolized bacterial lysate or saline at time −24 h (arrow A). All calves received aerosolized M. haemolytica at time 0 h (arrow M). There were 5 surviving calves at 60 and 72 h after challenge and these time points were excluded from the statistical analysis. A — Clinical scores were assigned based on the sum of individual assigned scores for demeanor (0–4), strength (0–4), appetite (0–3), respiratory effort (0–3) for a maximum score of 14. Clinical scores (mean and interquartile range) increased following M. haemolytica challenge (P < 0.001, ANOVA, n = 10), but were not different in calves administered bacterial lysate or saline. B — Body temperature increased following M. haemolytica challenge with elevations apparent by 12 h (P < 0.001). There were no differences between calves that received bacterial lysate and control calves (mean ± SEM). C — Respiratory rate (mean ± SEM). D — Heart rate (mean ± SEM).

Table I.

Prevalence of clinical signs in control and lysate-treated calves at different times relative to challenge.

	Day −1 to Day 0		Day 0 to Day 1		Day 1 to Day 2
	Control	Lysate	Control	Lysate	Control	Lysate
Cough	0	0	4	4	2	3
Depression	0	0	3	5	4	3
Loss of appetite	0	0	1	2	3	1
Increased respiratory effort	0	0	4	5	4	4
Decreased strength	0	0	2	2	4	2

Calves were aerosolized with either bacterial lysate (n = 5) or saline (n = 5) at Day −1, and then challenged with aerosolized M. haemolytica at Day 0. The data show the number of calves showing each of the clinical signs, in each time period.

Euthanasia prior to the planned endpoint was required for 3/5 control calves and 1/5 lysate-treated calves because of marked depression, dyspnea, reluctance to rise, inappetence, and fever. In addition, one clinically normal saline-treated calf was euthanized at the same time as its ill lysate-treated penmate because of welfare concern for individual housing.

Rectal temperatures increased following challenge with M. haemolytica and remained elevated until euthanasia (Figure 2, Supplemental Figures S2–S3), and were not different between control and lysate-treated calves (P > 0.05). Treatment with bacterial lysate did not affect either respiratory or heart rate, as evaluated using multivariate mixed model linear regression with pair and group as random effects (Figure 2, Supplemental Figures S2–S3). All calves had ultrasonographic evidence of pulmonary consolidation in ≥ 1 lung regions by 1 d after M. haemolytica inoculation (Supplemental Figure S4). These lesions persisted until euthanasia and correlated with grossly evident pneumonia. Additional areas of consolidation appeared in some calves at later time points (2 to 3 d after M. haemolytica challenge), and some areas of consolidation expanded over time. There was no difference in ultrasound scores between treatment groups.

From 1 to 2 d after challenge with M. haemolytica, total blood leukocyte and blood neutrophil concentrations increased over time (P = 0.001; Figure 3, Table II, Supplemental Figure S5–6). Calves receiving bacterial lysate before M. haemolytica challenge had significantly lower total leukocyte (P = 0.001) and blood neutrophil concentrations (P < 0.001) at 1 and 2 d after challenge. Serum haptoglobin and plasma fibrinogen concentrations increased over time after M. haemolytica challenge. Calves treated with bacterial lysate had lower haptoglobin concentrations following M. haemolytica challenge compared to controls (P = 0.002; Figure 3, Table II, Supplemental Figure S7–8).

Figure 3.

Changes in blood leukocytes and neutrophils and serum haptoglobin and fibrinogen in control and lysate-treated calves, before and after challenge with M. haemolytica. Calves received aerosolized bacterial lysate or saline at day −1 (arrow A). All calves received aerosolized M. haemolytica at time 0 (arrow M). Calves had daily blood collection for evaluation of: A — total blood leukocytes; B — blood neutrophils; C — serum haptoglobin; and D — plasma fibrinogen (mean ± SEM). Significant differences between lysate and control calves are shown (*P ≤ 0.05; **P ≤ 0.01). There were 5 surviving calves at 3 d post-challenge and this time point was excluded from the statistical analysis.

Table II.

Results of multivariate mixed model analysis of acute phase protein and hematologic values in clean-catch colostrum-deprived calves following aerosol challenge with Mannheimia haemolytica.

Outcome		β coefficient of effect	95% CI		P-value
Serum haptoglobin (log-transformed)
Treatment	Controla	referent
	Lysate	−0.705	−1.147	−0.264	0.002
Time after challenge	Day 1b	referent
	Day 2	0.468	0.093	0.844	0.015
Baseline	Day 0	−0.460	−0.806	−0.113	0.009
Plasma fibrinogen (log-transformed)
Treatment	Control	referent
	Lysate	0.397	0.090	0.704	0.011
Time after challenge	Day 1	referent
	Day 2	0.268	0.185	0.350	< 0.001
Interaction	Treatment* time	−0.139	−0.257	−0.020	0.022
Blood leukocyte count (log-transformed)
Treatment	Control	referent
	Lysate	−0.360	−0.572	−0.147	0.001
Time after challenge	Day 1	referent
	Day 2	0.161	0.053	0.269	0.004
Baseline	Day 0	1.025	0.521	1.529	0
Blood neutrophil count (log-transformed)
Treatment	Control	referent
	Lysate	−0.690	−1.011	−0.369	< 0.001
Time after challenge	Day 1	referent
	Day 2	−0.299	−0.465	−0.132	< 0.001
Baseline	Day 0	0.251	0.117	0.384	< 0.001
Blood lymphocyte count
Treatment	Control	referent
	Lysate	0.538	−0.090	1.166	0.093
Time after challenge	Day 1	referent
	Day 2	−0.852	−1.480	−0.224	0.008

^aSaline-treated (control) calves were the referent category in each analysis.

^bDay 1 following M. haemolytica challenge was the referent category.

Final model coefficients (β), 95% confidence intervals (CI), and P-values are shown for statistically significant (P ≤ 0.05) variables of interest.

Mannheimia haemolytica was not isolated from nasal swabs prior to challenge. After challenge, M. haemolytica was isolated from all calves including 22/26 nasal swabs from lysate-treated calves and 23/24 from control calves (Figure 4).

Figure 4.

Nasal carriage of M. haemolytica in control and lysate-treated calves, before and after challenge. Five pairs of calves received aerosolized bacterial lysate or saline at Day −1. All calves received aerosolized M. haemolytica at Day 0. The number of M. haemolytica colonies isolated by aerobic culture of left and right nasal swabs were semi-quantified as 0, 1+, 2+, 3+, 4+, and the individual data points show the sum of these scores (maximum score = 8; median ± interquartile range).

Postmortem findings

Mannheimia haemolytica was isolated from all calves, and from 54/68 (79%) lung lobes tested. The average number of bacteria (CFU/g) was not significantly affected by lysate treatment (P = 0.72; Figure 5). There was a strong correlation in the number of bacteria isolated from the different lung lobes tested from the same animal (P < 0.001), but the number of bacteria did not consistently differ by lung lobe (P = 0.159).

Figure 5.

Postmortem findings in 5 pairs of calves that received aerosolized bacterial lysate or saline 1 d before aerosolized M. haemolytica challenge. A — Percentage of diseased lung based on image analysis (individual-animal data with matching of calf pairs). B — Mean M. haemolytica concentrations for all lung lobes (individual-animal data with matching of calf pairs). C — M. haemolytica concentrations by lung lobe (mean ± SEM; n = 5 per group).

All calves had gross lesions of bronchopneumonia, with a lobular pattern of well-demarcated red-purple firm lesions affecting 1 to 43% of the lung (Figure 5, Supplemental Figure S9). The percentage of diseased lung tended to be lower in lysate-treated versus saline-treated calves [2.5% (95% CI: 0, 5.5) versus 19% (95% CI: 0, 41.6); P = 0.055]. Fibrinous pleuritis was present in 1/5 lysate-treated calves and 4/5 control calves. Lung:heart weight ratios were not significantly different between groups (P = 0.51). The extent of pneumonia was positively correlated with M. haemolytica concentration in lung tissue (R² = 0.74, P = 0.001). Visually estimated proportions of diseased lung were strongly correlated with digital image analysis estimates (R² = 0.82; P = 0.0003) and with log-transformed lung:heart weight ratios (R² = 0.611, P = 0.008) (Supplemental Figure S10).

All calves (10/10) had histologic lesions of bronchopneumonia, 5/10 had oat cells (dead neutrophils with streaming of chromatin), and 4/10 had focal coagulative necrosis (Table III). There were no differences in severity or nature of the histologic lesions between lysate-treated and control animals (Supplemental Figure S11).

Table III.

Histopathologic findings in control and lysate-treated calf pairs following Mannheimia haemolytica challenge.

Treatment	Pair	Broncho-pneumonia	Oat cells	Coagulative necrosis	Pleuritis	Fibrin
Control	1	2	2	2	2	2
	2	1	0	0	1	0
	3	1	0	0	2	0
	4	2	2	2	2	2
	5	2	2	2	2	2
Lysate	1	2	0	0	0	2
	2	1	0	0	0	0
	3	2	0	0	0	2
	4	2	2	2	2	2
	5	2	2	0	0	2

The 10 calves were assigned a score by a blinded evaluator for the severity of histologic lesions in the lung. For each category, severity scores were assigned as follows: 0 = absent; 1 = present, but rare; or 2 = present, moderate to severe.

Discussion

We hypothesized that stimulating innate immune responses before exposure to M. haemolytica could be an alternative to antibiotics for disease prevention. Aerosolization of bacterial lysate altered the concentrations of inflammatory cytokines in bronchoalveolar lavage fluid, recruited neutrophils and activated macrophages to the lungs, and led to evidence of complement, neutrophil and platelet activation (6). In the current study, aerosolized bacterial lysate induced an innate immune (inflammatory) response based on increased proportions of neutrophils in bronchoalveolar lavage fluid, increase in body temperature, and increase in plasma fibrinogen concentrations relative to control calves.

An aerosol M. haemolytica challenge was chosen to mimic the natural route of exposure, and bacteria delivered by aerosol may be more directly exposed to leukocytes and antimicrobial proteins induced by the bacterial lysate than with intrabronchial delivery of fluid containing bacteria. The aerosol challenge model required use of clean-catch colostrum-deprived calves raised in isolation to ensure that calves were free of M. haemolytica and did not have antibody against M. haemolytica (7). This was expected to reduce variability of responses, and all animals indeed had qualitatively similar clinical signs and responses after M. haemolytica challenge. Calves treated with bacterial lysate had evidence of partial protection against disease. Only 1 lysate-treated calf compared to 3 control calves were euthanized before the planned endpoint, and mean clinical scores tended to be lower in lysate-treated versus control calves at 1 and 2 d post-infection (dpi), although differences were not statistically significant. Lysate-treated calves had significantly lower serum haptoglobin concentrations than control calves at 2 dpi, and significantly lower blood neutrophil levels on 1 and 2 dpi, suggesting milder inflammatory disease. Similarly, postmortem lung lesions were milder in lysate-treated versus saline-treated calves with lower frequency of fibrinous pleuritis. Since consolidation detected by ultrasound examination remained stable or worsened over time, euthanasia of calves on Day 2 after M. haemolytica challenge before the planned endpoint may have under-estimated differences between study groups. Several responses to aerosolized bacterial lysate may have conferred the partial protection against M. haemolytica challenge, including induction of antimicrobial peptide gene expression, increased levels of IL-8 leading to neutrophil infiltration in alveoli and bronchioles, recruitment and activation of macrophages, activation of complement, and platelet activation (6).

The milder neutrophilia and lower serum haptoglobin concentrations suggest that lysate-treated calves had reduced induction of inflammatory responses, potentially due to fewer viable bacteria in the early stages after infection. However, if lysate administration did lead to a reduction in bacterial numbers soon after infection, it was not sufficient to clear enough bacteria to prevent pneumonia, nor to prevent the bacterial numbers from later increasing. Thus, even if there were partial reduction in M. haemolytica loads soon after challenge, this did not appear to confer a meaningful difference at the study’s endpoint.

Lysate-treated calves generally had milder clinical signs and less severe lung lesions despite similar or greater loads of M. haemolytica in the lung and nasal cavity. These findings could suggest that aerosolized bacterial lysate altered the inflammatory response to M. haemolytica challenge such as by modulating calves’ tolerance of infection (9). Additional studies are needed to investigate the mechanisms of these effects.

Calves with higher average M. haemolytica bacterial counts in the lung had a higher percentage of lung affected by consolidation. However, the other calves with less extensive consolidation also had numerous M. haemolytica isolated from the lungs. The number of M. haemolytica bacteria per gram of tissue did not vary by lung lobe, whereas consolidation was primarily cranioventral. These findings imply that the presence of bacteria is not the only determinant of why bronchopneumonia tends to affect the cranioventral part of the lung.

In conclusion, aerosolization of a lysate of E. coli and Staphylococcus aureus bacteria induced an innate immune response but provided only limited protection against subsequent M. haemolytica challenge. Specifically, lysate-treated calves had lower blood neutrophil numbers and serum haptoglobin concentrations as well as a tendency for lower clinical scores, increased survival to the planned end point, and less severe lung lesions at postmortem examination. However, the lysate did not consistently reduce the severity of clinical signs, lung lesions, or M. haemolytica numbers in nasal cavity or lung.