Pulmonary Surfactant Protein A Activates a Phosphatidylinositol 3-Kinase/Calcium Signal Transduction Pathway in Human Macrophages: Participation in the Up-Regulation of Mannose Receptor Activity¹

Inhaled particles and infectious agents represent a constant threat to the respiratory system. To defend itself, the respiratory tract has a complex system consisting of mechanical reflexes, physical barriers, and a diverse set of biochemical and cellular defenses. Acquired humoral and cellular immunity are of obvious importance in lung defense; however, it is becoming increasingly clear that components of innate immunity also play important roles in determining the response to and fate of inhaled particles. Pulmonary surfactant is a multimolecular complex, including proteins and lipids, that serves to reduce the surface tension of the alveoli, allowing expansion of the lung during inspiration. Recently, components of surfactant, specifically surfactant protein A (SP-A)³ and surfactant protein D (SP-D), have been shown to play important roles in the lung innate immune response (1, 2).

SP-A, a member of the collectin family, is the most abundant protein component (by weight) of human pulmonary surfactant (2). SP-A is produced as a 28- to 36-kDa (rat) or 35-kDa (human) polypeptide that contains a hydroxyproline-rich collagen-like domain, a carbohydrate recognition domain (CRD), and one (human) or two (rat) N-linked oligosaccharide attachment sites at Asn¹ in the N terminus (rat) and Asn¹⁸⁷ (human and rat) in the CRD (3). In the lung microenvironment, SP-A monomers assemble into an octadecamer, a structure that provides high valency-binding sites for a diverse set of molecules, including cell surface receptors and surface components of microorganisms, and is believed to be necessary for optimal activity of the protein (4, 5).

The interaction of SP-A with macrophages has multiple effects on the macrophage functional state, including chemotaxis, increased phagocytosis, and modulation of cytokine and oxidant production (1, 2, 6, 7). We have determined that the increase in macrophage phagocytosis is related in part to SP-A-enhanced activity of the mannose receptor (MR) (8), a pattern recognition receptor (9). SP-A effects on macrophages are believed to be mediated by interactions of this protein with its receptor(s) on macrophages (1). Chroneos et al. (10) have reported a 210-kDa SP-AR on U937 cells and rat macrophages called SP-R210. Other SP-A-binding proteins have also been reported (reviewed in Ref. 11). Little is known about the signaling pathways activated by SP-A binding to its receptor(s). Ohmer-Schröck et al. (12) reported that human SP-A activates a phosphoinositide/calcium-signaling pathway involved in phagocytosis by rat alveolar macrophages. However, the signaling pathway(s) induced by SP-A in primary human macrophages is unknown.

Several mechanisms induce the release of small bursts of Ca²⁺ into the cytosol, which acts as a second messenger for signal transduction (13). In macrophages, the generation of D-myo-inositol 1,4,5-triphosphate (InsP₃) is important in regulating Ca²⁺ release from the endoplasmic reticulum (ER) (reviewed in Ref. 14). Signals initiated by either G protein-linked receptors or receptors linked directly or indirectly to tyrosine kinases (14) activate phospholipase C (PLC), which hydrolyzes phosphatidylinositol (4,5)-bisphosphate to form InsP₃ and 1,2-diacylglycerol. In some cases, PI3Ks may act upstream of PLC activation because PLCγ can be activated by lipid products of PI3Ks (15, 16). Although a variety of agonists are capable of activating both PLC and PI3Ks in human macrophages (17), it is not known whether SP-A is one of them.

We hypothesized that SP-A binding to human macrophages activates a signal transduction pathway that uses Ca²⁺ as a second messenger. In this study, we examined Ca²⁺ mobilization in human monocyte-derived macrophages (MDM) in response to SP-A. Using wild-type and variant SP-A proteins with site-directed mutations, we investigated whether the attached carbohydrates and/or collagen-like domain of SP-A are necessary for activation. Using inhibitors, we studied the possible link between components of the Ca²⁺/InsP₃ signal transduction pathway and SP-A up-regulation of MR expression. Our data indicate that SP-A induces Ca²⁺ mobilization by activating PLC, which hydrolyzes membrane phospholipids and yields InsP₃, and that activation of PI3K(s), potentially through a PLC-signaling pathway, plays a role in SP-A up-regulation of MR.

RPMI 1640 medium with l-glutamine was purchased from Invitrogen Life Technologies. RPMI 1640 medium was used alone or with 20 mM HEPES (Sigma-Aldrich) and 1 mg/ml human serum albumin (HSA) (Calbiochem) at pH 7.2. Fura 2-AM was purchased from Molecular Probes. BAPTA-AM, wortmannin, LY294002, U73122, and U73343 were purchased from Calbiochem. Thapsigargin, polymyxin B sulfate (PMB), and LPS from Escherichia coli 055:B5 strain were purchased from Sigma-Aldrich. Platelet-activating factor (PAF) was a kind gift from Dr. L. Stoll (University of Iowa, Iowa City, IA). PE-conjugated mouse anti-human MR mAb and PE-conjugated anti-mouse IgG1κ Ab were purchased from BD Pharmingen. Anti-SP-R210 Ab was a generous gift from Dr. Z. A. Chroneos (University of Texas, Tyler, TX).

The SP-A proteins used in this study were produced and purified as previously described (18) and shown schematically in Beharka et al. (8). In brief, bronchoalveolar lavage was used to obtain SP-A from patients with alveolar proteinosis protein ((APP) SP-A) and from healthy volunteers (native human SP-A) (19). Native rat SP-A was purified from silica-pretreated Sprague-Dawley rat lungs. Recombinant rat SP-A (SP-A^hyp), which is deficient in hydroxyproline content and hence designated SP-A^hyp, was produced from SF-9 insect cells following infection with a recombinant baculovirus containing a 1.6-kb cDNA for rat SP-A (18). SP-A^hyp proteins devoid of oligosaccharides at one or both of the consensus sequences for N-linked glycosylation were generated by amino acid substitutions at the Asn¹ (SP-A^thr1) or Asn¹⁸⁷ site (SP-A^ser187) or at both the Asn¹ and Asn¹⁸⁷ sites (SP-A^thr1ser187) (18). Carbohydrate-deficient SP-A proteins retained structural and biologic functions, including partial oligomerization, aggregation of phospholipid liposomes, binding to immobilized carbohydrate, inhibition of lipid secretion from type II cells, and competition for receptor occupancy on type II cells (18). The synthesis of the mutant SP-A^hyp proteins containing a nested deletion of the proximal collagen-like domain (Gly⁸-Gly⁴⁴) (TM2), a truncation of the protein at the neck region resulting in a protein lacking the collagen-like and the N-terminal regions (TM1-2-3), and an amino acid substitution at the Asn¹⁸⁷ site (TM1-2-3^ser187) were made as described previously (20, 21). SP-A proteins were suspended in buffer containing 10 mM HEPES + 20 mM NaCl and kept at −20°C until use. Purity of the SP-A proteins was assessed by SDS-PAGE and silver staining. Endotoxin levels in SP-A preparations were determined using an Limulus amoebocyte lysate kit (BioWhittaker) and ranged from undetectable to 300 pg/μg protein, with an average of 15 pg/μg protein.

Blood was obtained from healthy adult volunteers using an approved Institutional Review Board protocol for human subjects at the University of Iowa and the Veterans Affairs Medical Center. PBMC from single donors were isolated from heparinized blood on Ficoll-sodium diatrizoate (Pharmacia) and cultured in Teflon wells (Savillix) in the presence of 20% autologous serum at 37°C for 5 days (22). On the day of each experiment, PBMC containing MDM were removed from Teflon wells, washed extensively, and counted. For Ca²⁺ experiments, 3 × 10⁵ MDM were adhered to 25-mm acid-cleaned glass coverslips in single wells for 2 h in the presence of medium containing RPMI 1640 medium + 20 mM HEPES + 10% autologous serum. For measurement of InsP₃, 1 × 10⁵ MDM were adhered to 10-mm acid-cleaned glass coverslips for 2 h in single wells.

U937 cells, obtained from the American Type Culture Collection, were maintained at 37°C in a humidified 5% CO₂ incubator in RPMI 1640 medium with heat-inactivated 10% FBS (HyClone), 1 mM glutamine, 20 mM HEPES, and penicillin/streptomycin (100 U/ml and 100 μg/ml, respectively). Cells were maintained in 175-cm² flasks and passaged every 3–4 days by scraping the flask with a cell scraper. Cells were not used after passage 30.

Measurement of [Ca²⁺]_cyt was performed as described previously (23, 24). Briefly, MDM on 25-mm round glass coverslips were washed and loaded with fura 2 by addition of the cell-permeate form, fura 2-AM, in RPMI 1640 medium for 30 min at 37°C in 5% CO₂. MDM were then washed with HEPES-buffered saline containing glucose and stimulated by the addition of SP-A proteins, LPS or PAF, as indicated (<5% v/v). Measurements of the apparent [Ca²⁺]_cyt were done using two different spectrofluorometers. In the first set of experiments, [Ca²⁺]_cyt was measured using the Photoscan II spectrofluorometer (Photon Technology International) with a Nikon microscope (Nikon) at the Cell Fluorescence Core Facility (Veterans Affairs Medical Center). The second set of experiments used an ImageMaster Ratio Fluorescence imaging system (Photon Technology International) at the Central Microscopy Research Facility (University of Iowa). In both cases, the final Ca²⁺ concentrations were determined from the ratio of emission intensities (emission wavelength (λ_em) 510 nm) to excitation wavelengths of (λ_ex) 340 and 380 nm using a Photon Technology International software package. Background fluorescence intensities for each λ_ex were obtained using non-fura 2-loaded cells and were subtracted from the raw data. The ratios of the corrected fluorescence intensities were then converted to [Ca²⁺] values using the formula [Ca²⁺] = K_d × (R − R_min)/(R_max − R), where the maximum and minimum ratios (R_max and R_min, respectively), as well as the apparent dissociation constant (K_d), were empirically derived from [Ca²⁺] curves generated using the instrument. Positive (PAF), vehicle (Tris buffer; DMSO), and HSA controls were done for each experiment. Experiments were repeated using MDM isolated from a minimum of three different blood donors.

In the second method, MDM were adhered to 24-well tissue culture plates in RPMI 1640 medium + 10% autologous serum, then washed and repleted in Dulbecco’s PBS plus 10 mM HEPES plus 1 mg/ml human serum albumin plus 0.1% glucose (DPBS-HHG) containing 10 μM fluo-4 AM (Molecular Probes) for 1 h at 37°C. Cells were then washed three times to remove unhydrolyzed dye and repleted with DPBS-HHG. Calcium measurements were made using a BMG FluoStar microplate fluorometer (BMG Lab Technologies) with excitation and emission wavelengths of 485 and 520 nm, respectively. Readings were taken every 3 s for 2 min to establish baseline, then 10 μg/ml SP-A were added and readings taken every 3 s for 10 min. A second bolus of SP-A (10 μg/ml)- or PAF (100 nM)-positive control was added at 3–60 min after the first addition of SP-A and fluorescence changes monitored. The maximal increase in fluorescence over baseline was determined for each SP-A or PAF addition. Responses of SP-A-treated cells to a second treatment with SP-A were expressed as percentage of initial response.

SP-A was fluorescently labeled using an Alexa Fluor 647 protein labeling kit (Molecular Probes).⁴ Cells were adhered to 4-well plates, washed, and rested overnight in RPMI 1640 medium + 10% serum. Macrophages were lifted using rubber policemen and resuspended in DPBS-HHG. Macrophages in suspension at 1E5/tube were incubated in triplicate tubes with 10 μg/ml unlabeled SP-A, 10–100 nM wortmannin, 3–9 μM Ly294002, 5 μM U73122, 5 μM U73343, or medium controls for 30 min at 37°C, then incubated with 1 μg of fluorescent-labeled SP-A at 4°C for 2 h. For conditions of SP-A preincubation, cells were pelleted once to remove medium containing unbound SP-A before the addition of labeled SP-A. Cells were washed, fixed in 2% paraformaldehyde, and fluorescence indicative of SP-A binding was measured in triplicate tubes using a Beckman FACSCalibur flow cytometer. Background fluorescence (macrophages with no SP-A) was subtracted from all test conditions and values set relative to the positive control (macrophages + SP-A, no inhibitors).

Two different methods were used to determine whether SP-A induced InsP₃ formation. In the first method, adherent MDM were incubated with 20 μg/ml APP SP-A for 0, 15, 60 s, and 10 min (time dependence) or with 0, 2.5, 5, 10, 20, or 40 μg/ml APP SP-A for 15 s (concentration dependence). The reaction was stopped, and the cells were lysed by the addition of 0.2× volume of ice-cold 20% perchloric acid on ice for 20 min. After centrifugation for 15 min at 2000 × g at 4°C, supernatants were removed and neutralized with ice-cold 1.5 M KOH containing 60 mM HEPES. KClO₄ was sedimented by centrifugation at 2000 × g for 15 min. Determination of InsP₃ in the supernatant was done using a competitive RIA (Amersham Biosciences), according to the manufacturer’s instructions.

In the second method, adherent MDM were incubated for 48 h in RPMI 1640 medium containing 1 μCi/ml myo-[³H]inositol (Amersham Biosciences), and turnover of inositol phosphates was measured as described previously (23, 25). Briefly, cells were washed with HEPES-buffered saline supplemented with 10 mM glucose (HBS-G) and incubated in HBS-G for 20 min at 37°C, followed by incubation for 20 min with HBS-G containing 10 mM lithium chloride. Finally, cells were treated with SP-A as in the first assay. Inositol phosphates were extracted overnight at 4°C with 0.5 M perchloric acid. The acid extract was neutralized with 2.5 M KOH and 0.5 M HEPES (pH 7.4) and centrifuged to remove the precipitate. The inositol phosphate species were then collected using anion-exchange column chromatography (Dowex AG, 1–8×, 100–200 mesh, formate form; Bio-Rad) as described previously (23, 25).

PBMC were incubated with inhibitors (thapsigargin, wortmannin, U73343, Ly294002, or BAPTA), vehicle control (DMSO), or medium for 30 min followed by SP-A proteins (10 μg/ml) or HSA protein control (10 μg/ml) in Teflon wells for 2 h. As a positive control for up-regulation of MR, MDM were incubated with IL-4 (Genzyme) for 20 h before staining (26). PBMC (5 × 10⁶) were incubated with PE-conjugated mouse anti-human MR or PE-conjugated subtypic control mAb, mouse IgG1κ. Cells were washed and fixed in paraformaldehyde and analyzed for mean fluorescence intensity (MFI) and percentage of positive cells (95/5% cutoff) using a FACScan Flow Cytometry System (BD Biosciences). Macrophages were identified using side scatter vs forward scatter. MFI due to nonspecific binding as determined using subtypic control Ab was subtracted from experimental data to provide a specific MFI. Each experiment was done in duplicate a minimum of three times. Cell viability was unaffected by the presence of inhibitors during the time course examined and was >90% by trypan blue exclusion.

Data were subjected to an analysis of normality. Normally distributed data were analyzed by ANOVA followed by Student’s t test using SPSS statistical program software. Statistical significance was defined as p < 0.05.

To determine the influence of SP-A on Ca²⁺ mobilization, we measured the level of [Ca²⁺]_cyt in adherent fura-2 loaded MDM using a Photoscan II spectrofluorometer. In these studies, the basal [Ca²⁺]_cyt values for MDM were ∼75–150 nM and were similar between experiments. A trace of a typical experiment using an optimal concentration of APP SP-A (10 μg/ml) is shown in Fig. 1,A. Exposure of MDM to SP-A resulted in an significant increase in [Ca²⁺]_cyt concentration over baseline (p < 0.05). The effect was rapid and transient. Cytosolic Ca²⁺ levels began to rise within 10 s of SP-A addition, and the maximal concentration was reached in ∼100 s with an amplitude of 125 ± 21 nM over baseline. The maximal level of [Ca²⁺]_cyt was sustained for <100 s, after which the level of [Ca²⁺]_cyt declined over a period of 5 ± 1 min. However, even after 20 min, [Ca²⁺]_cyt levels never returned completely to baseline (data not shown). The kinetics of the rise in [Ca²⁺]_cyt observed after addition of SP-A were similar to those observed with an optimal concentration (50 nM) of PAF (Fig. 1 B), except the magnitude of the response to SP-A was less. PAF, a known activator of the InsP₃/Ca²⁺ pathway in macrophages (27), was used as the positive control in all experiments.

The preparations of SP-A used contained variable levels of LPS. However, our data provide evidence that the increase in [Ca²⁺]_cyt is due solely to SP-A and not to contaminating LPS because neutralization of LPS through the addition of PMB (5 μg/ml) before stimulation with APP SP-A did not significantly diminish the APP SP-A-induced [Ca²⁺]_cyt response (Fig. 1,C). Furthermore, addition of LPS alone, either at levels found in the SP-A preparations (<300 pg/ml; data not shown) or higher levels used as controls (0.5 μg/ml; Fig. 1,B), did not mobilize [Ca²⁺] _cyt. Failure of LPS to mobilize Ca²⁺ was not due to an inability of the MDM to respond because the addition of PAF after LPS treatment resulted in a significant [Ca²⁺]_cyt response (Fig. 1 B).

The SP-A-induced elevation in [Ca²⁺]_cyt was concentration dependent (Fig. 2). APP SP-A induced a rise in [Ca²⁺]_cyt when added at 0.5 μg/ml (threshold concentration), and the increase over baseline was significant at 1 μg/ml (Fig. 2). The dose response was linear at concentrations of up to 5 μg/ml and then plateaued, with no additional effect at higher concentrations.

SP-R210 has been reported as a receptor for SP-A that is present on rat alveolar and bone marrow-derived macrophages and the human myeloid cell line U937 (10). To determine whether signaling through SP-R210 was responsible for the SP-A-induced rise in [Ca²⁺]_cyt, we blocked binding of SP-A to SP-R210 through the addition of 10 μg/ml rabbit anti-SP-R210 (10) or control rabbit IgG to MDM for 20 min before stimulation with 5 μg/ml APP SP-A. SP-R210 did not affect the subsequent rise in [Ca²⁺]_cyt by APP SP-A (174 ± 35 vs 167 ± 21 nM for MDM treated with anti-SP-R210 Ab or isotype control, respectively, n = 3). Expression of SP-R210 and C1qR, but not the putative SP-AR identified by Strayer et al. (28) and Stevens et al. (29), has been demonstrated on U937 cells (10). Addition of APP SP-A did induce a rise in [Ca²⁺]_cyt in U937; however, a 10-fold greater concentration of SP-A was required for an optimal response in these cells compared with human MDM (50 μg/ml APP SP-A for U937 vs 5 μg/ml for MDM, n = 3).

The SP-A-induced rise in [Ca²⁺]_cyt was not seen after a second stimulation with SP-A at the same concentration as the first (10 μg/ml), even after 3 min (Fig. 3,A). When 20 μg/ml APP SP-A (Fig. 3,B) or 40 μg/ml APP SP-A (Fig. 3,C) were used as the secondary stimulus, [Ca²⁺]_cyt increased within 20 s of APP SP-A addition but never reached levels obtained by the first SP-A stimulus. However, the refractory state was specific to SP-A, as MDM exposed to APP SP-A retained their ability to respond to a second unrelated agonist, PAF (Fig. 3,D), and MDM exposed to PAF were capable of responding to SP-A (Fig. 3 E) to the same extent as MDM treated with SP-A alone.

We examined the length of this refractory period in MDM and found full recovery of the calcium response to a second stimulation with SP-A by 60 min (Fig. 4). The reduced rise in [Ca²⁺]_cyt seen upon a second stimulation with SP-A at time points of <60 min was not due to a reduction in the ability of SP-A to bind to MDM because preincubation of MDM with 10 μg/ml SP-A for 15 or 30 min at 37°C did not inhibit binding of subsequently added SP-A (Fig. 4). These data suggest that SP-A induces a refractory state in macrophages that is specific for SP-A and thus resembles homologous receptor desensitization with regard to mobilization of intracellular Ca²⁺.

We next determined the influence of the source of SP-A protein on Ca²⁺ mobilization. When added to MDM at 10 μg/ml, the increase in [Ca²⁺]_cyt by APP SP-A was significantly higher than from all other sources (Fig. 5). Native human SP-A and native rat SP-A induced similar (p > 0.05) levels of [Ca²⁺]_cyt that were significantly higher (p < 0.005) than that induced by the wild-type SP-A^hyp. Additionally, the amount of SP-A required to achieve a threshold [Ca²⁺]_cyt response varied with SP-A type (0.5, 1, 1, and 2 μg/ml for APP, human native, native rat, and SP-A^hyp, respectively). These results indicate that the source of SP-A (APP, and native, recombinant) affects the degree of Ca²⁺ mobilization.

We next explored which of the domains or structural elements of SP-A mediated the Ca²⁺ response. To study the potential influence of the carbohydrate moieties of SP-A in mediating its effect on Ca²⁺mobilization, we used SP-A^hyp proteins devoid of oligosaccharides at one (Asn¹ site (SP-A^thr1) or Asn¹⁸⁷ site (SP-A^ser187)) or both (Asn¹ and Asn¹⁸⁷ (SP-A^thr1ser187)) of the consensus sequence sites for glycosylation. The absence of carbohydrates at these sites on the variant proteins used in these studies has been demonstrated previously (7, 18). Stimulation of fura 2-labeled MDM with 10 μg/ml SP-A devoid of one (SP-A^ser187 or SP-A^thr1) or both (SP-A^thr1ser187) N-linked carbohydrates did not elicit a significant rise in [Ca²⁺]_cyt over baseline (Fig. 6). When the amount of mutant SP-A protein added to the cells was increased to 40 μg/ml, all SP-A proteins were able to induce small, short duration (<100 s) rises in [Ca²⁺]_cyt; however, only the rise in [Ca²⁺]_cyt induced by SP-A devoid of carbohydrate at Asn¹ (SP-A^thr1) was significantly higher than baseline (Fig. 6). Moreover, the rise induced by SP-A^thr1 was significantly lower than that induced by SP-A in which both carbohydrates were present (SP-A^hyp). These results provide evidence that the presence of carbohydrate at both glycosylation sites is necessary for optimal mobilization of intracellular Ca²⁺.

In parallel experiments, we investigated the importance of the collagen-like domain of SP-A for induction of Ca²⁺ mobilization (Fig. 7). Truncated SP-A proteins lacking the first half of the collagen-like domain but still possessing both carbohydrate attachment sites (TM2), lacking the collagen-like domain and the N-terminal segment (Asn¹ to Ala⁷) but still possessing the carbohydrate attachment site at Asn¹⁸⁷ (TM1-2-3), or lacking the collagen-like domain, N-terminal segment, and both carbohydrate attachment sites (TM1-2-3^ser187) were added to fura 2-labeled MDM. None of the truncated mutants was capable of inducing a Ca²⁺ response similar in strength or duration to that achieved by APP SP-A or SP-A^hyp. However, similar to the response seen using carbohydrate-deficient SP-A proteins, all of the collagen-deficient mutants induced a very low-level increase of Ca²⁺, which was similar between the mutants. Taken together, these studies demonstrate a critical role for both the carbohydrates and collagen-like domain in the Ca²⁺ response.

The rise in [Ca²⁺]_cyt induced by agonists is accomplished through two mechanisms: the release of Ca²⁺ from intracellular stores, of which the ER is the largest contributor, and by influx across the plasma membrane (30). We next evaluated the relative contribution of these two mechanisms to the SP-A-induced calcium response. Thapsigargin is an irreversible ER Ca²⁺-ATPase inhibitor whose application leads to depletion of intracellular Ca²⁺ stores (31). Fura 2-loaded MDM were pretreated with thapsigargin (0.1, 1, or 10 μM) for 60 min. Preliminary experiments confirmed that this pretreatment did not adversely affect cell viability (data not shown). Pretreatment with thapsigargin resulted in the dose-dependent loss of APP SP-A-induced Ca²⁺ mobilization (Table I). Thapsigargin at 10 μM essentially prevented any increase in Ca²⁺ in response to APP SP-A. These results provide evidence for the involvement of intracellular Ca²⁺ stores in the initial SP-A-induced rise in [Ca²⁺]_cyt. However, the data do not exclude the possibility that calcium influx plays a role later in the Ca²⁺ response.

To further explore the role of the InsP₃/Ca²⁺ pathway in SP-A signaling, we next investigated whether SP-A induced generation of InsP₃ in MDM. The level of InsP₃ in MDM was measured using a commercially available competitive RIA, which determines the InsP₃ content, as well as an assay that measures turnover of inositol phosphates (23, 32). Both assays gave similar results (data not shown); therefore, only the results of the competitive RIA are shown. APP SP-A induced production of InsP₃ in a time- and dose-dependent manner (Fig. 8). A typical time course of InsP₃ formation in response to SP-A is shown in Fig. 8,A. Within 15 s of SP-A addition, InsP₃ levels rose significantly above baseline. Maximal InsP₃ levels were reached at 15 s post-SP-A addition. InsP₃ formation returned to basal values by 10 min after stimulation. These kinetics are similar to those observed for the Ca²⁺ response to SP-A. Additionally, the level of InsP₃ formed was dependent on the concentration of APP SP-A (Fig. 8 B). Although the effective range of SP-A in these assays was similar to that of the Ca²⁺ response, slightly more APP SP-A was needed to induce InsP₃ formation compared with Ca²⁺ mobilization, a result that may be related to differences in assay sensitivity.

InsP₃ is typically formed through PLC hydrolysis of membrane phospholipids (14). Incubation of MDM with an aminosteroid inhibitor of PLC (U73122), but not its analog (U73343), inhibited the SP-A-induced rise in [Ca²⁺]_cyt (Table I), suggesting that SP-A activation of PLC leads to Ca²⁺ mobilization. SP-A binding to MDM was not decreased by treatment with U73122 compared with its analog control U73343 (data not shown), indicating that inhibitors of PLC had no effect on SP-AR expression. Thus, these data provide support for SP-A activation of a PLC/InsP₃ pathway leading to the rise in [Ca²⁺]_cyt.

PI3K, through its lipid products, has been reported to regulate PLCγ-mediated calcium signaling under some conditions (16). Therefore, we next determined whether inhibition of PI3K affected the SP-A-induced rise in [Ca²⁺]_cyt. We pretreated MDM with the PI3K inhibitor wortmannin, a fungal metabolite, before the addition of APP SP-A (Table I). Pretreatment with wortmannin resulted in a significantly decreased Ca²⁺ response (Table I). However, wortmannin did not completely inhibit Ca²⁺ mobilization in response to APP SP-A, suggesting that activation of PI3K plays a role in but may not be essential for Ca²⁺ mobilization.

The interaction of SP-A with the macrophage results in modification of macrophage functions, including, but not limited to, up-regulation of MR expression (1, 7, 8). To determine whether SP-A signaling through the InsP₃/Ca²⁺ pathway results in up-regulation of MR, we used inhibitors against the previously identified signaling molecules activated by SP-A and measured MR expression using flow cytometry. As previously reported (8), APP SP-A or IL-4 significantly increased MR expression on MDM (Table II). The addition of the general PI3K inhibitors wortmannin or Ly294002 before SP-A treatment significantly reduced the SP-A-induced up-regulation of MR in a dose-dependent manner. Wortmannin did not reduce basal MR expression levels (data not shown), nor did it significantly inhibit SP-A binding to MDM at concentrations that inhibited MR up-regulation by SP-A (100 nM) compared with concentrations that had no effect (10 nM; data not shown). Additionally, Ly294002 treatment of MDM had no effect on binding of SP-A to MDM (data not shown), indicating that inhibition of PI3K(s) has no effect on SP-AR expression. Therefore, a PI3K(s) plays an important role in the SP-A-induced signaling pathway that leads to MR up-regulation. The role of PLC-linked Ca²⁺ mobilization in SP-A-induced MR up-regulation is less clear. Inhibition of PLC by U73122 did not significantly affect SP-A induction of MR. Furthermore, while preincubation with the Ca²⁺ inhibitor BAPTA reduced SP-A-induced MR expression, inhibition of calcium mobilization by thapsigargin did not significantly affect SP-A induction of MR. However, it should be noted that 5 μM BAPTA tended (p = 0.07) to reduce the basal level of MR surface expression on MDM, suggesting that Ca²⁺ is involved in the normal trafficking of this receptor.

SP-A is increasingly recognized as a key component of innate immunity in the lung. SP-A plays a critical role in host defense against inhaled pathogens, as demonstrated by the increased susceptibility of the SP-A^−/− mouse to infection with a variety of extracellular bacteria (33, 34, 35). SP-A has been shown to have direct antimicrobial activity against both bacterial (36) and fungal (37) microorganisms. However, it is also evident that part of the role of SP-A in pulmonary immunity is due to the ability of SP-A to modify alveolar macrophage immunological function (2). Addition of SP-A to macrophages in vitro increases chemotaxis and phagocytosis, decreases NO production, and alters the respiratory burst (reviewed in Ref. 1). Recently, our laboratory has demonstrated that SP-A increases the surface expression and activity of the macrophage MR (8) and that this up-regulation plays a role in the observed increase in Mycobacterium tuberculosis phagocytosis by SP-A-treated cells (7).

SP-A has been shown to associate with macrophages in a saturable and specific manner, consistent with receptor-mediated binding (38). Although several SP-A binding proteins have been identified on rat macrophages, myeloid cell lines, and in purified protein systems (10), the receptor(s) responsible for the effects of SP-A on MR expression and activity have not yet been determined. Interactions between SP-A and its macrophage receptor(s) activate signaling pathway(s) that ultimately effect the reported alterations in macrophage functions (12, 39, 40, 41, 42, 43). Our laboratory is particularly interested in the identification of signaling pathway(s) leading to up-regulation of the MR. The results presented here indicate that at least one signal transduction pathway activated in human macrophages by SP-A involves Ca²⁺ signaling and that PI3K contributes to the SP-A-induced increase in MR expression.

Ca²⁺ is an important second messenger in phagocytic cells and has been linked to many cellular processes, including vesicle trafficking (reviewed in Ref. 44). [Ca²⁺]_cyt in resting cells is low; upon stimulation, [Ca²⁺]_cyt increases 5- to 10-fold, activating Ca²⁺-sensitive processes. Our data demonstrate that SP-A significantly increases [Ca²⁺]_cyt within seconds of SP-A addition to cultured MDM. The immediate increase in [Ca²⁺]_cyt in response to APP SP-A is similar to that reported by Ohmer-Schröck et al. (12) using rat alveolar macrophages. However, there may be species differences in threshold responsiveness to SP-A: we found that 10 μg/ml APP SP-A induced mobilization of calcium in 100% of MDM. In contrast, Ohmer-Schröck et al. (12) reported 30 μg/ml APP SP-A stimulated Ca²⁺ mobilization in 36% of rat alveolar macrophages, and 120 μg/ml APP SP-A achieved mobilization in 69% of these cells (12).

The effects of SP-A on [Ca²⁺]_cyt were observed in our system at concentrations as low as 1 μg/ml. Although the precise SP-A concentration in pulmonary surfactant is not known, the concentration of SP-A in rat lung hypophase has been estimated at 300 μg/ml to 1.8 mg/ml (reviewed in Ref. 1) based on SP-A levels in lavage fluid. However, this does not account for the possibility that microenvironments could exist where the local concentrations of SP-A are much higher or lower (1). It is important to note that SP-A activity may also be affected by its association with lipid, as SP-A has been shown to be concentrated in tubular myelin in the human lung (45).

In our studies, after the maximum [Ca²⁺]_cyt was reached, the [Ca²⁺]_cyt declined over the following 10 min until reaching a steady-state level slightly above baseline. Interestingly, when these macrophages were stimulated with a second dose of SP-A (lesser or equal concentration) within 10 min of the first, no second rise in [Ca²⁺]_cyt was detected. When 2- or 4-fold greater APP SP-A was used as the secondary stimulus following an initial SP-A treatment, the [Ca²⁺]_cyt was increased but never achieved the level reached in response to the first stimulus. However, SP-A-treated MDM were still able to fully respond to the second stimulus PAF. This suggests that the inhibition is not due to depletion of available Ca²⁺ but that SP-A facilitates an agonist-specific refractory period, which is consistent with the hypothesis that SP-A induces homologous desensitization of its receptor(s) on the macrophage surface.

Desensitization facilitates decreased responsiveness of a cell to successive exposure of the same extracellular stimulus over time. We envision that the functional state of the SP-AR(s) is under continuous regulation in response to changes in SP-A levels in the lung, such as has been reported in bacterial pneumonia (46). Additionally, the degree of SP-AR desensitization is likely influenced by the differentiation and/or activation state of the cell. In the alveoli, alveolar macrophages are in continuous contact with surfactant containing SP-A and ingest abundant amounts of this material (47); therefore, it is possible that resident alveolar macrophages are in a constant state of desensitization to SP-A. In other words, maintenance of the alevolar macrophage phenotype may require constant (tonic) ligation and desensitization of SP-AR(s).

In contrast, monocytes or interstitial macrophages that migrate into the lung would be sensitive to the effects of SP-A upon first encounter. In this model, SP-A would serve as a local tissue factor that contributes to the induction and maintenance of the unique biological properties of alveolar macrophages compared with macrophages in other tissue compartments. These include increased surface expression of receptors of innate immunity such as the MR (9), greater phagocytosis of both nonopsonized and opsonized particles (48), decreased production of inflammatory cytokines (48), a reduced oxidative burst in response to stimuli (6, 49), and decreased expression of costimulatory molecules (50)—an anti-inflammatory phenotype consistent with a state referred to as “alternative activation” (51). Homologous receptor desensitization is dependent on agonist occupancy of the receptor; consequently, under conditions in which SP-A concentration is significantly decreased, the alveolar macrophages would become resensitized to SP-A. This would explain why alveolar macrophages are capable of responding to SP-A in vitro (7, 12).

In the current study, the threshold level of SP-A needed to induce a Ca²⁺ response and the magnitude of this response were both dependent on the source of SP-A used. Although human APP, native human and rat, and SP-A^hyp demonstrate shared functional characteristics, differences in the magnitude of select biological responses related to SP-A type have been observed previously (52, 53). Our laboratory has reported previously that APP SP-A was more effective than either native human or SP-A^hyp in up-regulating MR receptor expression and enhancing the phagocytosis of M. tuberculosis by MDM (7, 8).

One notable difference among the SP-As used is the degree of higher-order molecular organization (54); our results suggest that these differences are related to the ability of SP-A to induce Ca²⁺ mobilization. APP SP-A, which can form complexes larger than the native SP-A octadecamer (52, 55), demonstrated the greatest amount of Ca²⁺ mobilization. In contrast, SP-A^hyp, which oligomerizes to a lesser extent than native SP-A due to a lack of proline hydroxylation, demonstrated the lowest amount of Ca²⁺ mobilization. Additionally, mutants that lacked the collagen-like domain and/or the N-terminal region, both of which are believed to be essential for forming higher order multimers (3), were capable of stimulating only a slight Ca²⁺ response. These data suggest that higher order oligomers are essential for maximally effective receptor engagement, which may involve recognition of a multiunit epitope, and/or receptor clustering by ligation of multiple CRD. It should also be noted that, in addition to differences in multimerization, human APP, native human and rat, and SP-A^hyp also demonstrate differences in glycosylation patterns (3), which may play a role in the current results.

To further investigate the role of individual structural components of SP-A on Ca²⁺ mobilization, we used SP-A^hyp variants lacking part or all of the collagen-like domain or one or both attached carbohydrates. Wild-type SP-A^hyp contains two sugars, one at Asn¹ in the N-terminal segment and one at Asn¹⁸⁷ in the CRD (3). Removal of one or both sugars significantly decreased the SP-A-induced Ca²⁺ burst, which is consistent with studies showing that these carbohydrate moieties are important for some of its biological effects (7, 56). When both sugars remained intact but the collagen-like domain was partially or completely removed, the SP-A-induced calcium response was also diminished. This is supported by work by Chroneos et al. (10), which showed that collagen V inhibits SP-A binding to macrophages, as well as work by Vandivier et al. (57), which showed that C1q collagen tails inhibit macrophage uptake of SP-A-coated erythrocytes. It is not yet clear whether the collagen-like domain participates in direct protein-protein interactions that result in a calcium flux or whether it acts primarily as scaffolding that amplifies the binding activities of other domains of the molecule. None of the mutations completely eliminated the Ca²⁺ response, and inhibition of the Ca²⁺ response by these mutations was not additive because the removal of both sugars and the collagen-like domain or of the N-terminal and collagen-like domains did not further diminish the Ca²⁺ response. It is unclear whether the small remaining Ca²⁺ response is physiologically relevant. Viewed as a whole, our studies demonstrate that both the protein-associated carbohydrates and collagen-like domain of SP-A play a role in SP-A-induced Ca²⁺ mobilization.

The possibility exists that there is more than one receptor for SP-A on macrophages, each one of which induces a different signal pathway and biological effect. Several proteins have been identified as putative receptors for SP-A, not all of which are present on human macrophages (38, 58). One SP-A-binding protein reported to be on U-937 cells and rat alveolar macrophages is the 210-kDa protein SP-R210 (10). Our assays failed to reveal a role for this receptor in the Ca²⁺ response or up-regulation of the MR because Abs against SP-R210 failed to inhibit SP-A-induced increases in [Ca²⁺]_cyt or MR expression (data not shown). U937 cells also required 10 times more APP SP-A for the induction of calcium than did human MDM, suggesting that the SP-AR(s) on U937 cells that signals through Ca²⁺ is expressed at an extremely low level compared with that on primary human macrophages, has a lower affinity for SP-A, or lacks expression of a coreceptor (39, 40, 59).

The calcium mobilized into the cytoplasm in response to stimulation originates from either the extracellular space or the ER (14). One standard approach to examine whether extracellular Ca²⁺ plays a role in induced rises in [Ca²⁺]_cyt is to chelate extracellular Ca²⁺ using EGTA. However, this approach cannot be used in our system to study the contribution of extracellular Ca²⁺ to signaling by SP-A, as we have recently shown that binding of SP-A to its receptor(s) on primary human macrophages is Ca²⁺ dependent.⁴ We have shown that depletion of ER Ca²⁺ stores by preincubation with thapsigargin completely blocks the SP-A-induced [Ca²⁺]_cyt rise, indicating that the initial calcium response is mobilized from intracellular stores. We also found that SP-A induces a rise in InsP₃, an important molecule involved in release of Ca²⁺ from ER stores (14). Based on these observations, we have concluded that SP-A-induced Ca²⁺ mobilization at least initially originates from intracellular stores. The level of [Ca²⁺]_cyt above baseline seen as late as 20 min after SP-A addition may be due to an influx of extracellular Ca²⁺ through capacitative calcium entry, which is controlled by the efflux of calcium from the ER. Recent evidence supports InsP₃R involvement in the initiation of this entry process (60). Therefore, we propose that the SP-A-induced rise in [Ca²⁺]_cyt is composed of two phases: an initial spike in [Ca²⁺]_cyt through InsP₃-induced release of Ca²⁺ from specific intracellular stores, followed by a plateau phase (essential for replenishment of the intracellular stores) that is sustained by Ca²⁺ influx from the extracellular medium.

Our data further demonstrate that SP-A induction of Ca²⁺ mobilization is mediated through activation of PLC. PLC not only catalyzes the formation of InsP₃ from phosphatidylinositol 4,5-bisphosphate but also controls cellular phosphatidylinositol 4,5-bisphosphate concentrations, thereby intersecting with other signal transduction pathways such as that of PI3K. Pretreatment with wortmannin before stimulation with SP-A resulted in a markedly decreased calcium response, suggesting that PI3K is an upstream signaling molecule involved in Ca²⁺ mobilization by SP-A.

Taken together, these studies demonstrate that SP-A activates a Ca²⁺/PLC/InsP₃ signal transduction pathway. We next investigated whether signaling through this pathway by SP-A results in up-regulation of MR expression. This hypothesis is supported by inhibition of MR up-regulation by the calcium chelator BAPTA. However, BAPTA may also chelate local concentrations of calcium or other cations needed at the membrane for vesicular fusion, thereby inhibiting processes in addition to calcium release that are necessary for MR up-regulation. The role of SP-A signaling through PLC in MR up-regulation is less clear because inhibitors of PLC did not significantly reduce the effect of SP-A on MR expression. However, inhibitors of PLC decreased but did not completely abolish the calcium response to SP-A; the residual low-level calcium response may have been sufficient to induce MR up-regulation in these cells. Alternatively, redundancy in signaling is common in immunological systems, and inhibition of PLC may therefore cause a compensatory increase in the activity of other contributing enzymes or pathways, which may also account for our results.

Using the inhibitors wortmannin and Ly294002, we have shown that PI3K are involved in both SP-A-induced mobilization of Ca²⁺ and in SP-A up-regulation of MR. PI3K, as with calcium, is involved in intracellular trafficking and acts as a regulator of multiple aspects of membrane trafficking (61, 62, 63). A PI3K may be the common molecule involved in two separate signaling cascades—one resulting in MR up-regulation and the other involving the Ca²⁺/PLC/InsP₃ pathway; or, more likely, a different member of the PI3K family is involved in each event. Interestingly, the lipid products of PI3K have been found to directly activate certain protein kinase C (PKC) isoforms. If PKC plays a role in MR up-regulation, it would provide an additional explanation for our data using BAPTA because treatment of cells with BAPTA-AM leads to a translocation and inactivation of PKC (64).

In conclusion, SP-A binding to human macrophages activates a signal transduction pathway(s) that results in increased levels of [Ca²⁺]_cyt. The magnitude of this increase is dependent on 1) the type of SP-A used and 2) the sugars and collagen-like domain of the protein. Our experiments make the novel observation that macrophages become refractory to SP-A after initial exposure, suggesting that the receptor(s) for SP-A undergoes homologous desensitization. This finding may be particularly important for setting the anti-inflammatory “tone” of macrophages within the lung. Our data indicate that SP-A triggers the increase in calcium by inducing PLC activation, leading to the hydrolysis of membrane phospholipids and yielding InsP₃. We have shown that activation of PI3K(s) plays a role in SP-A up-regulation of MR and may act as a component of the PLC signal pathway. These studies provide further insight into the signaling pathways activated in human macrophages by SP-A which play direct roles in SP-A modulation of macrophage biology.

We thank Thomas Kaufman, Jennifer Ufnar, Amanda Dawson, and Thomas Moninger for their expert technical assistance and Deb Nollen-Richter for help with the preparation of this manuscript.

The authors have no financial conflict of interest.