INTRODUCTION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

The aspiration of foreign material, such as milk, food, gastric contents, or saliva, into the tracheobronchial tree is a common cause of acute and chronic lung disease in children. Predisposing anatomic and functional disorders include uncoordinated swallowing from neurological dysfunction or muscular weakness, vocal cord paralysis, laryngeal cleft, tracheoesophageal fistula, vascular ring, achalasia and gastroesophageal reflux (GER) (1). Aspiration can produce a variety of lung disorders including acute or recurrent pneumonia, bronchiectasis, and interstitial lung disease (1). As a result, significant loss of lung function and even respiratory failure may occur (2).

While the potential importance of aspiration in the pathogenesis of lung disorders is widely recognized, only recently have clinical and laboratory studies investigated the association between aspiration syndromes and lung injury. Fan and coworkers (7) detected aspiration in 14% of children presenting with chronic diffuse pulmonary infiltrates. In animal models, recurrent aspiration of milk produced alterations of neural control mechanisms of the airway smooth muscle, thus suggesting a potential mechanism involved in aspiration-induced airway dysfunction (8). These observations coupled with clinical experience demonstrate that aspiration represents an important pediatric problem that requires accurate diagnosis. Unfortunately, currently available diagnostic tests, such as barium swallow with videofluoroscopy and gastroesophageal scintigraphy (milk scan), are neither sensitive nor specific enough to establish a diagnosis of aspiration (9, 10).

The lipid-laden macrophage (LLM) index obtained by Oil Red O (ORO) staining of bronchoalveolar lavage (BAL) fluid has been widely used as a marker of pulmonary aspiration (1, 11, 12). However, clinical and laboratory studies have demonstrated the lack of specificity of ORO in the diagnosis of aspiration syndromes (13). Many lung constituents, such as cell membranes and surfactant, contain lipids. Lung injury that causes cell destruction leads to the release of these lipids and formation of LLMs (16, 20). As a result, the LLM index is not a specific marker for aspiration, but rather a nonspecific marker of lung injury.

In this study, we evaluated whether immunocytochemical staining of milk proteins (-lactalbumin and -lactoglobulin) within alveolar macrophages improves the ability to detect aspiration in a murine model of milk aspiration. In addition, we compared the sensitivity of immunocytochemistry (IC) and ORO after single and repeated aspiration of milk. We found that immunocytochemical staining of milk proteins provides a greater sensitivity and specificity for the diagnosis of milk aspiration when compared with ORO staining.

	METHODS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Experimental Animals

Female BALB/c mice (8 wk of age), obtained from Harlan Sprague Dawley (Indianapolis, IN), were used for all studies. All of the procedures used in this study were approved by the Animal Care and Use Committee of the University of Texas Health Science Center (Houston, TX) and conformed to National Institutes of Health (NIH, Bethesda, MD) guidelines.

Human milk and cow milk were used as the aspiration medium. After sedation with ketamine and xylazine, 50 µl of milk was placed into the nose of quietly breathing mice while the mouth was closed. This technique, a modification of methods previously described by our laboratory (8), allows for aspiration of milk or saline into the lungs with strong respiratory effort as evidenced by the coughing produced by inoculation of either material. The procedure did not lead to awakening of the mice, and the coughing resolved within 30-60 s after completion of the nasal injection of milk or saline. This protocol was employed for single and repeated aspiration. The latter was accomplished by injecting 50 µl of milk daily for 4 d. Control animals aspirated normal saline in an identical manner and schedule. At various time points after the last aspiration events (see below) lung lavage was obtained for analysis.

In selected experiments, the specificity of immunocytochemistry was also evaluated in other models of lung injury: respiratory syncytial virus (RSV) and Mycoplasma pulmonis (MP) infection. This was accomplished by staining alveolar macrophages (AMs) obtained from animals infected with 10⁷ plaque-forming units of human RSV and MP as previously described (21, 22).

Tracheotomy and Lung Lavage

Under anesthesia with intraperitoneal ketamine (200 mg/kg) and xylazine (10 mg/kg), the trachea was isolated and cannulated with a 22-gauge plastic catheter. Lung lavage was performed with four aliquots of 0.5 ml of phosphate-buffered saline (PBS) solution at 24, 48, and 72 h after the last aspiration event. Typical cell recoveries from murine lung lavage yield approximately 500,000 cells per mouse, 95% AMs. Immediately after cell isolation, specimens were placed in sterile cytofunnels (Shandon, Pittsburgh, PA) and then centrifuged at 1,000 rpm for 5 min onto positively charged glass slides (Superfrost/plus; Fisher Scientific, Pittsburgh, PA), using a Cytospin 2 centrifuge (Shandon). The slides were then used for individual stainings as described below. For immunocytochemical staining, slides were quickly fixed in acetone for 15 s and then preserved at 20° C.

Oil Red O Staining

Oil Red O staining was performed as previously described (23). Slides were fixed with formaldehyde vapor for 5 min and then placed in ORO solution for 20 min, rinsed with distilled water for 5 min and counterstained with Mayer's hematoxylin (Sigma, St. Louis, MO) for 5 min. The slides were then coverslipped with glycerin and evaluated by light microscopy (Zeiss, Oberkochen, Germany). One hundred consecutive macrophages were counted and results expressed in terms of percent positive cells (mean ± SE).

Immunocytochemical Staining for -Lactalbumin and -Lactoglobulin

Slides were fixed in acetone and then soaked in 3% H₂O₂ (Sigma) in methanol solution. After application of 0.05% Triton (Sigma) in PBS, slides were incubated with blocking antibody (Vector Laboratories, Burlingame, CA). Primary antibodies against -lactalbumin (rabbit anti-human -lactalbumin; DAKO, Carpinteria, CA) and -lactoglobulin (rabbit anti-human -lactoglobulin, Bethyl Laboratories, Montgomery, TX) were applied at the dilutions recommended by the manufacturers. The slides were then incubated with biotinylated anti-rabbit immunoglobulins (Vector Laboratories, Burlingame, CA) and then exposed to avidin-biotin-peroxidase complex (Vector Laboratories). After counterstaining with Mayer's hematoxylin, slides were coverslipped with glycerin and examined under a light microscope.

The following negative controls were used for our studies: (1) cells obtained after aspiration of normal saline; (2) cells obtained after aspiration of milk and stained without the primary antibody; and (3) cells incubated with isotype-matched rabbit antibody (Vector Laboratories). Results were expressed in terms of percent positive cells (mean ± SE).

Statistical Analysis

All results were expressed as means ± SE, and n equals the number of experimental observations. The nonparametric Wilcoxon test was used to assess differences between IC and ORO staining at three time points (24). A p value < 0.05 was considered significant.

	RESULTS

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

After single and repeated aspiration of human milk, a large number of alveolar macrophages displayed a positive staining for -lactalbumin (ALA). Representative fields of AMs obtained by lung lavage 24 h after a single aspiration of milk are shown in Figure 1. A distinct cytosolic immunoreactivity for ALA was detected within the majority of macrophages (Figure 1B). Similar results were obtained by immunocytochemical staining for -lactoglobulin (BLG) in AMs after a single episode of cow milk aspiration. A strong immunoreactivity for BLG was detected in the form of cytosolic granules within the macrophages (Figure 1C). All negative controls used for these studies (see above) did not display immunoreactivity for ALA or BLG (Figure 1A). In addition, no immunoreactivity for milk proteins was found in AMs obtained from mice infected with RSV and MP.

View larger version (15K): [in this window] [in a new window]   Figure 1.   Immunocytochemistry for -lactalbumin and -lactoglobulin in mouse alveolar macrophages. Representative fields of alveolar macrophages obtained by BAL 24 h after aspiration of normal saline (A), human milk (B), or cow milk (C ). A distinct cytosolic immunoreactivity for -lactalbumin (B) and -lactoglobulin (C ) is present within alveolar macrophages.

To further evaluate the sensitivity of the IC staining for ALA in our murine model, lung cells were stained at various time points after aspiration of human milk. The number of cells positive for ALA staining at 2, 4, 6, 12, 24, 48, 72, and 96 h is shown in Figure 2. Immunocytochemical staining for ALA was detected within AMs beginning at 2 h. Approximately 40% of the AMs had positive immunoreactivity 96 h after the aspiration event.

View larger version (8K): [in this window] [in a new window]   Figure 2.   Time course of percent positive cells stained by immunocytochemistry for -lactalbumin in mouse alveolar macrophages. After a single aspiration of milk, a large number of alveolar macrophages displayed a positive immunoreactivity beginning at 2 h. Approximately 40% of the cells had positive immunoreactivity by 96 h after the aspiration of milk.

To compare the sensitivity of ORO and IC for ALA, both stainings were performed on cells recovered from the same animals after single and repeated aspiration at various time points after the last aspiration event. After single aspiration, the percentage of positive AMs, using the IC method for ALA, was significantly higher when compared with ORO: 86.6 ± 3.7% (n = 5), 74.4 ± 3.4% (n = 5), and 56.8 ± 3.7% (n = 5) versus 3.0 ± 0.8% (n = 5), 1.6 ± 0.2% (n = 5), and 1.0 ± 0.3% (n = 5) at 24, 48, and 72 h, respectively (Figure 3, p < 0.05). After repeated aspiration (Figure 4), the percent positive cells for IC and ORO were as follows: 92.4 ± 2.4% (n = 5), 83.0 ± 5.3% (n = 5), and 68.4 ± 5.9% (n = 5) versus 8.0 ± 1.7% (n = 5), 5.2 ± 1.6% (n = 5), and 3.4 ± 1.0% (n = 5) at 24, 48, and 72 h, respectively (p < 0.05).

View larger version (9K): [in this window] [in a new window]   Figure 3.   Immunocytochemistry for -lactalbumin and Oil Red O staining of mouse alveolar macrophages after a single aspiration of milk. Results are shown as percent positive cells (means ± SE) for ORO (open columns) and -lactalbumin (solid columns). The percentage of alveolar macrophages positive for -lactalbumin is significantly higher when compared with ORO at all time points studied. *Values significantly greater (p < 0.05) than those obtained for ORO.

View larger version (10K): [in this window] [in a new window]   Figure 4.   Immunocytochemistry for -lactalbumin and Oil Red O staining of mouse alveolar macrophages after recurrent aspiration of milk. Results are shown as percent positive cells (means ± SE) for ORO (open columns) and -lactalbumin (solid columns). The percentage of alveolar macrophages positive for -lactalbumin is significantly higher when compared with ORO at all time points studied. *Values significantly greater (p < 0.05) than those obtained for ORO.

	DISCUSSION

TOP ABSTRACT INTRODUCTION METHODS RESULTS DISCUSSION REFERENCES

Commonly used methods to detect aspiration of foreign material into the respiratory tract lack sensitivity and specificity. Although barium swallow with videofluoroscopy and gastroesophageal scintigraphy (milk scan) can diagnose pulmonary aspiration, their sensitivity is low and negative studies do not rule out aspiration (9, 10, 25). The identification of LLMs in BAL fluid is perhaps the most common diagnostic method used for pulmonary aspiration. Early work by Pinkerton (4) showed that fat-filled macrophages in the bronchial secretions of patients with pneumonia were indicative of aspiration of milk or oily substances. Further studies by Williams and Freeman (26) confirmed the presence of fat-filled macrophages in tracheal aspirates of infants with chronic pulmonary problems and swallowing dysfunction. More recently, refinements in bronchoscopy and BAL techniques have led to wider application of LLM detection in BAL fluid for the diagnosis of aspiration (11, 12, 27, 28).

More recent work, however, has suggested that LLMs can be detected in other pulmonary disorders such as pulmonary fat embolism seen in trauma patients (13) and patients with sickle cell disease (14). In addition, abundant LLMs are found in tracheal aspirates from patients who receive intravenous lipid infusion and systemic drugs (15, 29). Cohen and Cline (16) demonstrated the presence of LLMs behind totally obstructed bronchi and suggested that ingestion of lipids from degenerating epithelial cells and/or surfactant by alveolar macrophages leads to formation of LLMs.

Detection of LLMs, even in asymptomatic individuals or in patients with lung diseases other than aspiration, and biochemical studies of the fate of pulmonary surfactant in the lung, support the concept that LLMs are not specific for aspiration (11, 20). Using murine models of lung injury, our laboratory has confirmed these findings by detecting LLMs in the BAL of mice after infections with RSV and MP (21). Taken together, these observations demonstrate that LLMs can be detected in a variety of pulmonary disorders and should not be considered diagnostic for pulmonary aspiration.

In an attempt to improve the specificity of methods based on the detection of LLMs within alveolar macrophages, Corwin and Irwin (11) developed a semiquantitative LLM index. However, the low specificity of the index in this study suggests that the presence of LLMs in the BAL represents a nonspecific marker of parenchymal lung disease. In contrast, Colombo and Hallberg (12), using a similar semiquantitative index in a pediatric population, did not find any overlap between aspirators and nonaspirators and suggested that an LLM index > 90 was both sensitive and specific for chronic pulmonary aspiration in children. As a result, this test has been widely used in clinical practice. However, controversies regarding its specificity continue (17), and optimal interpretation of test results often requires clinical correlation (18, 19, 30).

Another important consideration in applying the LLM index and similar methods to the diagnosis of pulmonary aspiration is the timing of the test. As aspiration may be sporadic and intermittent, macrophages that have engulfed the aspirated material may have been cleared from the lungs by the time the BAL is performed. Using a rabbit model of milk aspiration, Colombo and coworkers (31) described a detailed relationship between time after aspiration and the LLM index. In this respect, the LLM index was found to be significantly increased within 6 h after aspiration and remained elevated for 48 h.

In the present study, we sought to identify proteins naturally found in milk (-lactalbumin and -lactoglobulin) as markers of pulmonary aspiration. Thus, a positive staining for milk proteins within alveolar macrophages provides conclusive evidence for a diagnosis of pulmonary aspiration. The large statistical difference between ORO staining of lipids and immunocytochemical staining of milk protein at all time points suggests that the latter technique is more sensitive in detecting milk aspiration in mice. Another important feature of this technique is the ability to detect ALA in approximately 40% of alveolar macrophages 96 h after a single aspiration event. Therefore, this method may detect sporadic and intermittent aspiration events.

In addition to high sensitivity, immunocytochemical staining for milk proteins is highly specific, as demonstrated by our experiments using a variety of negative controls. In preliminary studies, we found ALA to be specific for human milk without cross-reaction with cow ALA. In addition, AMs obtained from animals with other inflammatory lung disorders (RSV and MP pneumonia) did not display immunoreactivity for ALA or BLG, thus supporting the specificity of this method in our murine model.

In some infants and young children, aspiration may occur as a result of gastroesophageal reflux (28). As milk proteins retain their antigenicity when placed in an acid environment (32), ALA and BLG should still be detectable by immunocytochemical staining. However, the rate of phagocytosis and clearance by alveolar macrophages might be different after the conformational changes in protein tertiary structure have occurred. Therefore, detailed studies in animal models and humans are needed to investigate the role played by these and other variables in immunocytochemical staining of milk proteins within alveolar macrophages.

In summary, we have described a novel diagnostic method for the detection of milk aspiration in a murine model. Our findings demonstrate that immunocytochemical staining of milk proteins (-lactalbumin and -lactoglobulin) within alveolar macrophages represents a sensitive and specific method for the diagnosis of pulmonary aspiration. While additional work is needed to confirm its clinical application, this method should significantly enhance the ability to diagnose milk aspiration owing to its high sensitivity and specificity. In light of these observations, immunocytochemistry should be valuable in establishing causative links between aspiration syndromes and lung disorders.