SLC7A7 gene mutation analysis was performed by Dr. Ginafranco Sebastio in Naples, Italy. ABCA3, SP-B and SP-C genes were sequenced and analyzed by Ambrey Genetics.

All buffer salts and reagents were obtained from Sigma unless otherwise stated. SP-D was isolated from BALF of an adult PAP patient as described previously 4041. Recombinant GM-CSF expressed in yeast (Leukine) was produced by Berlex (Seattle, WA). All cell culture media were obtained from Invitrogen (Carlsbad, CA).

Therapeutic WLL was performed in the LPI patient or idiopathic PAP patient, and the BALF samples were colleted in 100 ml vials. Diagnostic BALF sample was also obtained from another airway disease patient. BALF was filtered with nylon mesh and centrifuged at 200 × g for 10 min at 4°C. The pellets containing the cells were then washed twice with Hank's Buffered Salt Solution (HBSS). The supernatant was centrifuged again at 10,000 × g for 40 min at 4°C to collect surfactant lipid pellets, which were stored at -80°C until further analyses.

BCA^® Protein Assay kit from Pierce (Rockford, IL) was used for measuring the protein concentrations in the pre-filtered BALF from each collection vial according to the manufacturer's instructions. SP-D concentration levels and degradation were assessed in the BALF from the LPI patient by Western blotting using anti-human SP-D rabbit sera as described previously 42. An aliquot of the BALF from each collectin tube was immediately mixed with 0.2% (w/v) Trypan Blue solution and counted using a Hemocytometer at 20× or 40× normal magnification. The amount of cholesterol present in the BALF samples were assessed using the Cholesterol Assay Kit (Cayman Chemical Co., Ann Arbor, MI).

Cytospin preparations were made using 50 and 100 μl BALF, cells were air dried and subjected to Hemacolor^® staining (EMD Chemicals, Inc, Gibbstown, NJ) for histological analysis. Other cytospin preparations were fixed with 2% (v/v) paraformaldehyde (PFA) in PBS for 1 h at 4°C, and were subjected to staining with Oil Red O (Polysciences, Inc, Washington, PA) for lipid analysis of foamy macrophage cells. Oil Red O stock solution was prepared by dissolving 0.5% (w/v) Oil Red O powder in isopropanol at 37°C. This stock solution was then diluted to 60% (v/v) with dH₂O to make the working solution. Cytospin preparation of cells were rinsed with 60% (v/v) isopropanol, and stained with freshly prepared Oil Red O working solution for 15 min and counterstained with Hemacolor Solution 3 for nuclei. Cells were imaged using Leitz Laborlux D microscope with Leica IM 50 image manager software.

Cell pellets and surfactant lipid pellets obtained from 200 × g and 10,000 × g centrifugation of BALF, respectively, were fixed for at least 18 h at 4°C with 2% (v/v) paraformaldehyde (PFA) and 4% (v/v) glutaraldehyde (Electron Microscopy Sciences, Fort Washington, PA) in a buffer containing 50 mM HEPES (pH 7.4), 150 mM NaCl, and 1 mM CaCl₂. Samples were then processed, stained with 2% (w/v) uranyl acetate and analyzed by transmission electron microscopy (JEOL TEM 1011) with AMT Image Capture Engine software at 2,000-40,000 times normal magnification.

Cell pellets obtained from 200 × g centrifugation of the BALF were immediately re-suspended in Dulbecco's Modified Eagle's Medium (DMEM) supplemented with 10% (v/v) heat-inactivated Fetal Bovine Serum (FBS) with 100 U/ml penicillin, 100 μg/ml streptomycin, and 2.5 μg/ml Fungezone. Cells were then allowed to adhere to chamber slides or 24- or 48-well culture plates for 1 h, and non-adherent cells were washed. Cells were cultured in the same media, and subsequently, the media was changed to Macrophage Serum-Free Media (MSFM), a specific media available for culturing macrophage cells (Invitrogen), and does not require supplementation with serum, prior to the experiments. The Jurkat T cell line was maintained in RPMI containing 10% (v/v) FBS.

Jurkat T cells were first labelled with 2 μl/ml DiI membrane dye (Invitrogen) in MSFM for 4 min at 37°C, followed by three washes with HBSS. Apoptosis was induced by UVC irradiation for 30 s using UV Stratalinker 2400 (120 J/cm²). Cells were then placed in the humidified incubator at 37°C for 2 h in the absence or presence of SP-D (5 μg/ml).

Cells were incubated first in MSFM overnight prior to the experiment. Cells were then treated with SP-D (1 μg/ml) for 4 h, and GM-CSF (10 ng/ml) for 30 min prior to the experiment. Alexa-647 conjugated BSA was added to the culture for 30 min, at which point the cells were fixed with 2% (v/v) PFA for 10 min at room temperature, followed by fresh 2% (v/v) PFA for 1 h at 4°C. DiI-labelled apoptotic Jurkat T cell uptake assay was performed in the same manner as BSA using the Jurkat cells 2 h after the induction of apoptosis (1:5 of apoptotic cells:macrophages).

Images for the uptake assays, and granuloma assay were captured using Zeiss Axiovert 200 M fluorescence microscope with Volocity software. Measurement of cell area, color (label-specific gray levels) and size was performed using Volocity software.

Cells were incubated in the presence or absence of SP-D (1 μg/ml) and/or GM-CSF (10 ng/ml) for 29 days ex vivo. Culture media was replaced with a fresh media with the addition of appropriate reagents (SP-D and/or GM-CSF) every 3-5 days. After 28 days, the granulomas started to dislodge from the substratum.

A two-tailed student's t-test was performed using Microsoft Excel on Oil Red O positive cell counting and apoptotic Jurkat cell uptake analysis. Protein uptake assay, and cell shape and size were analyzed using Dunnett's multiple mean comparison tests using JMP statistical analysis software (Version 5.0; SAS Institute Inc.). A p-value was set at 0.05 for statistical significance. Non-linear regression analysis and comparisons of regression lines by ANOVA and F-test was performed using GraphPad PRISM statistical analysis software (Version 4.0; GraphPad Software).

A Canadian boy with Dutch ancestry was diagnosed with LPI at 9 months of age. His symptoms included feeding difficulty, vomiting, hepatomegaly and failure to thrive. Symptoms improved after initiating dietary therapy with protein restriction and citrulline supplementation. However, a progressive productive cough developed by age 18 months. Chest X-rays showed an interstitial pneumonitis pattern and bronchiolar lavage fluid confirmed the presence of Periodic acid-Schiff (PAS) stain positive staining alveolar macrophages consistent with the diagnosis of PAP. The patient, however, did not have mutations in ABCA3, SP-B, or SP-C. Despite appropriate dietary therapy and treatment of gastro-esophageal reflux, the boy's respiratory symptoms progressed and he developed a persistent oxygen need. A trial of systemic steroid therapy (2 mg/kg body weight) was ineffective. At 2 1/2 years of age referral was made to our pediatric respiratory medicine service. Since WLL is associated with significant morbidity and technically difficult in small children, a three-month trial of inhaled GM-CSF (250 μg once daily delivered through a pari-star nebulizer with a mouthpiece) was instituted. The boy's respiratory deterioration continued. Therefore, a WLL procedure was performed after the discontinuation of GM-CSF treatment. After the procedure, there seemed to be some initial clinical improvement, but the respiratory symptoms relapsed within 3 to 5 days. His disease progressed to the point of respiratory failure requiring intubation and mechanical ventilation despite two further lung lavages. A lung biopsy only confirmed the diagnosis of PAP and interstitial pneumonitis secondary to LPI. The presence of cholesterol granulomas in the airways was also confirmed by the biopsy. The boy was unresponsive to further courses of high dose pulse intravenous corticosteroid therapy and eventually died.

Diagnostic BALF samples from a patient who had bronchitis of unknown etiology and no specific airways disease were used as a control. In some experiments, we also used BALF samples obtained from adult idiopathic PAP patient.

To determine the alterations in the proteins present in the patient's airways, we analyzed the lavage samples by BCA assay. Large quantities of proteins were present in the patient's airways, and the protein content gradually decreased in successive collection aliquots (Figure 1A). Compared to the control BALF, the protein concentration in the airways of the LPI patient was over 100-fold higher in the first lavage (Figure 1B). Although repeated BAL procedures across different procedural days (BAL2, 3, and 4) and different lobes reduced the protein concentration, they remained over 30-fold higher compared to control (Figure 1B). In addition to increased protein concentration, large numbers of dead/dying cells were present in these samples as revealed by Trypan blue exclusion assays (Figure 1C). On average, approximately 40% dying cells were present in these LPI BALF samples.

The role of SP-D in human PAP is not well understood. Using multiple LPI PAP lavage samples from the patient, we first determined whether there were any alterations in SP-D. Western blots detected SP-D that was either modified or tightly associated with other components present in the airways (upward smear) as well degraded SP-D (low molecular weight bands) (Figure 1D). This suggested that the bioavailability of functional SP-D would be low in these airways. To further assess this point, we analyzed the surfactant lipids isolated from the supernatant obtained after removing the cell pellet. The surfactant lipid isolation was performed by a subsequent centrifugation of this supernatant at high-speed (10,000 × g). We first examined the surfactant lipid structure in detail. Lattice-like tubular myelin structure was detected in the surfactant lipid (Figure 1E, arrowhead). There were also many amorphous electron-dense structures present (Figure 1E, arrow). Notably, the PAP surfactant also contained unusual forms of tubular myelin with circular lattices (Figure 1F). This circular tubular myelin had the cross-sectional diameter of approximately 90 nm, with electron dense "spot" in its centre, a structure similar to that of SP-D mediated tubular myelin 43. These unusual structures indicated that SP-D is likely trapped inside these tubular structures. The presence of atypical circular tubular myelin and the cleaved/modified collectin suggested that the bioavailability of SP-D would be low in these airways.

To determine the cholesterol level in the surfactant, we measured the amount of cholesterol present in the LPI BALF. The LPI BALF contained over 60-fold higher amounts of cholesterol compared to the control samples (Figure 2A). This is particularly important as serum cholesterol levels in the LPI patient were found to be in the normal range (data not shown). We also detected many needle-like cholesterol crystals in the lipid pellet, ranging from 2 μm to over 30 μm in length (Figure 2B)

To characterize the different types of immune cells present in the airways, cytospin preparations were made from BALF, fixed and stained with Oil Red O, a dye that primarily stains neutral lipids, triglycerides and lipoproteins (Figure 2C). While few cells from bronchitis control BALF were Oil Red O positive, a significantly higher percentage of cells stained positive in the LPI BALF indicating the presence of lipid-laden foamy macrophage cells (> 20-fold; p < 0.05; Figure 2C). To further confirm the nature of these cells, cells were isolated from BALF samples and examined using electron microscopy. Large numbers of vacuolated foamy macrophages, other immune cells, small-sized apoptotic cells with condensed nuclei and cellular debris were readily detected (Figure 2D). Furthermore, foamy cells that contain cholesterol crystals within their cytoplasm were often observed (Figure 2E). Large cholesterol needles were also present outside the cells in these preparations. Collectively, these results demonstrated that the LPI patient had PAP with multiple alterations potentially impairing lung function.

We then sought to determine whether SP-D and/or GM-CSF would enhance the uptake of various components typically present in the PAP airways by LPI alveolar macrophages. To assess this, we tested whether the uptake of protein (Alexa-647 conjugated BSA) and apoptotic Jurkat cells could be increased in the presence of SP-D and/or GM-CSF (Figure 3). Among the cultured BAL cells, "healthy" cells with alveolar macrophage morphology readily adhered to the culture plates for long period of time. However, the foamy cells did not survive the culturing conditions, and did not show detectable uptake of labeled protein (Figure 3A); thus, these healthy cells were used for the subsequent experiments (Figure 3B-E). For this, BAL cells that were collected by centrifugation of BALF at low speed (200 × g) and plated in chamber slides were cultured. In the uptake assays, these cells were pre-incubated with SP-D (1 μg/ml) for 4 h, and/or with GM-CSF (10 ng/ml) for 30 min prior to feeding (Figure 3B and 3D). In general, qualitative examination showed that pre-incubation of cells with SP-D and GM-CSF enhanced the uptake of Alexa-647 BSA by the BAL cells. In order to quantify the uptake of Alexa-647 BSA, we manually counted different types of BSA uptake of randomly selected cells from the images (double blinded) using a sorting system. The counting categories consisted of "Cell Periphery", "Uniform Distribution", "Perinuclear" and "No Uptake". Of those categories, "Uniform Distribution" and "Perinuclear" categories were considered separate because the internalized materials are processed within the cells by fusing with lysosomes, which are primarily located in large numbers in the perinuclear regions of the cells 44. We inferred that the increase in the number of cells in "Perinuclear" category would be an increase in the processing rate of the internalized material. Compared with buffer control, co-incubation of cells with both SP-D and GM-CSF significantly increased the proportion of cells belonging to perinuclear category (p < 0.05; Figure 3B). Therefore, we infer that the presence of both SP-D and GM-CSF enhances the uptake/clearance of proteins by LPI macrophages.

SP-D deficiency results in the accumulation of apoptotic cells, in vivo 30, and PAP airways contain increased number of dead cells (Figures. 1 and 2). Thus, we next asked whether SP-D and/or GM-CSF could enhance the clearance of apoptotic cells by the cells isolated from the LPI BALF (Figure 3C and 3E). These cells were incubated with apoptotic Jurkat T cells for 4 h in the presence/absence of SP-D and/or GM-CSF. SP-D was pre-incubated with apoptotic cells after the induction of apoptosis with UV irradiation. We manually quantified the images of cells as either "Uptake" where the cells have taken up apoptotic cell material, and "No Uptake". Manual counting of randomly selected images (double blinded) revealed that, like BSA uptake, the uptake of apoptotic cells/bodies was significantly increased when macrophages were incubated with SP-D coated apoptotic cells in the presence of GM-CSF (p < 0.05; Figure 3C).

We next asked whether SP-D and GM-CSF alter the cells isolated from BALF. Analysis of randomly selected images obtained from the feeding experiments showed that SP-D and GM-CSF affect the shape and size (as defined by the area occupied by each cell) of the cells (Figure 4). While control cells consisted of relatively uniform shapes and sizes (Figure 4A), the presence of SP-D promoted cell spreading (Figure 4B) and GM-CSF made cells more spherical/oval in shape (Figure 4C). Presence of both SP-D and GM-CSF generated mixed populations of both elongated and spherical cells (Figure 4D). Measurement of cell size also showed that GM-CSF, both in the presence and absence of SP-D, caused a significant decrease in cell size (p < 0.05; Figure 4E). Thus, these results show that SP-D may increase cell spreading, whereas GM-CSF maintains cells in a spherical shape.

During the course of our experiments, we serendipitously discovered that the BAL cells began to progressively extend long processes, especially those cultured ex vivo for longer than 7 days. In agreement with this phenomenon, the cell size increased with increase in number of days (Figure 4F). Because of the change in cell shape and size, we asked whether there was an effect on the ability of these cells to internalize BSA. Quantifying BSA uptake revealed that the proportion of cells participating in the uptake of BSA decreased over time, and was inversely related to cell size (Figure 4F).

Remarkably, when these BAL cells were cultured ex vivo for an even longer period of time (> 20 days), these cells began to form previously uncharacterized granulomatous structures (Figure 5). The formation of these structures was reproducible under different conditions. The granulomas formed ex vivo consisted of smaller spherical cells and elongated cells. These clearly defined cells were primarily found at the periphery of the granulomas (Figure 5A). Although the precise nature of these structures is unknown, given our observation that GM-CSF makes cells more spherical, we asked whether incubation of these cells with GM-CSF would reverse, or slow down the formation of elongated cells and attenuate or prevent granuloma formation. Surprisingly, presence of GM-CSF (10 ng/ml) dramatically increased the number of granulomatous structures present in the culture (Figure 5B). On the other hand, the control BAL cells only became slightly elongated. We performed a regression analysis on the quantification of granulomas formed in each culture and compared each condition using an F test (Figure 5B). The analysis revealed that conditions were significantly different from one another. Notably, culturing the cells in the presence of both SP-D (1 μg/ml) and GM-CSF (10 ng/ml) significantly reduced the formation of granulomas, compared to GM-CSF alone (p < 0.05; Figure 5B). Together, these results suggested that LPI BAL cells form granulomatous structures spontaneously, and GM-CSF enhances their formation. The data also suggest that SP-D can reduce GM-CSF mediated increase in granuloma formation (p < 0.05).