Abstract
Removal of cells dying by apoptosis is essential to normal development, maintenance of tissue homeostasis, and resolution of inflammation. Surfactant protein A (SP-A) and surfactant protein D (SP-D) are high abundance pulmonary collectins recently implicated in apoptotic cell clearance in vitro. Other collectins, such as mannose-binding lectin and the collectin-like C1q, have been shown to bind to apoptotic cells and drive ingestion through interaction with calreticulin and CD91 on the phagocyte in vitro. However, only C1q has been shown to enhance apoptotic cell uptake in vivo. We sought to determine the relative importance of SP-A, SP-D, and C1q in pulmonary clearance of apoptotic cells using knockout and overexpressing mice, and to determine the role of calreticulin and CD91 in this process. SP-A, SP-D, and C1q all enhanced apoptotic cell ingestion by resident murine and human alveolar macrophages in vitro. However, only SP-D altered apoptotic cell clearance from the naive murine lung, suggesting that SP-D plays a particularly important role in vivo. Similar to C1q and mannose-binding lectin, SP-A and SP-D bound to apoptotic cells in a localized, patchy pattern and drove apoptotic cell ingestion by phagocytes through a mechanism dependent on calreticulin and CD91. These results suggest that the entire collectin family of innate immune proteins (including C1q) works through a common receptor complex to enhance removal of apoptotic cells, and that collectins are integral, organ-specific components of the clearance machinery.
Apoptotic cell death and the noninflammatory removal of cell corpses are hallmarks of embryonic development, tissue remodeling, maintenance of the immune system, normal cell turnover, and resolution of inflammation (1, 2). Efficient clearance of dying cells may be particularly important in the lung in which immune challenges occur regularly, sometimes eliciting the influx of massive numbers of inflammatory cells. Recognition and clearance of apoptotic cells are accomplished by a redundant system of phagocyte receptors (3, 4, 5), soluble bridging molecules (3), and apoptotic cell ligands (3, 6, 7). Despite activity in vitro, relatively few of these clearance mechanisms, alone, have been shown to impair apoptotic cell clearance in vivo when tested in genetically deficient mice (8, 9, 10, 11), and only CD44 deficiency has been shown to impair apoptotic cell removal in the lung (12). The redundancy of apoptotic cell removal mechanisms and the inability of individual deletions to consistently impair apoptotic cell clearance in vivo suggest that apoptotic cell clearance is a critical process to preserve homeostasis. Recent evidence suggests that two members of the collectin family, surfactant protein A (SP-A)4 and surfactant protein D (SP-D), may participate in apoptotic cell removal by alveolar macrophages (AMs) in vitro (13). However, the importance of SP-A or SP-D in pulmonary apoptotic cell removal, and the mechanism by which they act, has not been addressed.
The collectins, including SP-A, SP-D, mannose-binding lectin (MBL), and bovine conglutinin, are pattern-recognition proteins of the innate immune system known for their ability to facilitate the removal of microorganisms through opsonin-mediated phagocytosis. Collectins are composed of structurally similar oligomers that contain an N-terminal collagenous tail, and a C-terminal globular head group with lectin (carbohydrate-binding) activity (14). C1q, the first member of the classical complement cascade, is related to the collectins by structure and function, except that its globular head lacks lectin activity (15). The finding that C1q and MBL participate in apoptotic cell clearance, in vitro (5) and in vivo (9, 10), followed initial observations that C1q binds to blebs on the surface of apoptotic cells (16). Most recently, SP-A and SP-D have been shown to bind rat apoptotic neutrophils (PMNs) and drive their ingestion by rat AMs in vitro (13). These findings, and the abundance of SP-A and SP-D in the lung (17), suggest that collectins may play a significant role in pulmonary clearance of apoptotic cells, maintenance of homeostasis, and resolution of inflammation.
C1q and MBL have been suggested to drive apoptotic cell uptake through engagement of cell surface calreticulin and CD91 (5). Calreticulin is a multifunctional protein located in the endoplasmic reticulum (18) and on the surface of many cell types, including macrophages (5); it has been shown to specifically bind to the collagenous tail of C1q and MBL (14, 19), and has been identified as the collectin receptor, cC1qR (20). However, calreticulin lacks a transmembrane domain. Therefore, for signal transduction, it is likely that calreticulin must associate with a transmembrane receptor. CD91, also known as the α2 macroglobulin receptor, may be this receptor, because CD91 complexes with calreticulin (21) and participates in apoptotic cell phagocytosis mediated by C1q and MBL (5). SP-A (22) and SP-D (P. Eggleton, personal communication) also bind calreticulin, suggesting that they may drive clearance of apoptotic cells through a similar CD91-mediated mechanism (15).
Studies were designed to determine the relative in vivo and in vitro importance of C1q, SP-A, and SP-D in pulmonary clearance of apoptotic cells. Our findings indicate that SP-D appears to be particularly important in apoptotic cell removal from the naive murine lung, because genetic deletion or overexpression of SP-D, alone, alters apoptotic cell removal in vivo. In contrast, genetic deletion of SP-A or C1q has no effect on pulmonary apoptotic cell clearance. In vitro, SP-A, SP-D, and C1q bind to apoptotic cells and drive their ingestion by AMs through interaction with calreticulin and CD91. These results suggest that the entire extended family of collectins, including C1q, may contribute to apoptotic cell removal through a common receptor complex, but that in the naive murine lung SP-D appears to play a dominant role.
Materials and Methods
Materials
Abs obtained from BD PharMingen (San Diego, CA) included mouse monoclonal anti-human FcγRII (CD32) IgG, mouse monoclonal anti-human CD45 IgG, and rat monoclonal anti-mouse FcγRIII/II (CD16/32) IgG. Mouse monoclonal anti-human CD91α IgG was obtained from American Diagnostica (Greenwich, CT). Chicken polyclonal anti-mouse calreticulin IgY was obtained from Affinity Bioreagents (Golden, CO). Human C1q was obtained from Quidel (Santa Clara, CA). C1q tails were prepared, as previously described (5). C1q was confirmed to be endotoxin free (data not shown). Human α2 macroglobulin and bovine calreticulin were obtained from Sigma-Aldrich (St. Louis, MO).
Human and animal experimentation
This study was approved by, and performed in accordance with the ethical standards of, the National Jewish Medical and Research Center Institutional Review Board on Human Experimentation. Written informed consent was obtained from each subject.
Mice were housed and studied under Institutional Animal Care and Use Committee-approved protocols in the animal facility of National Jewish Medical and Research Center. Experiments were performed with 8- to 12-wk-old mice matched for age.
Animals
SP-D-deficient mice (SP-D knockout (KO)) were prepared as previously described (23) and were outbred five generations with National Institutes of Health Swiss Black mice. SP-D KO mice had no detectable SP-D protein in bronchoalveolar lavage fluid. SP-A was decreased by 25%, and saturated phosphatidylcholine was increased 3-fold (23).
SP-D-overexpressing (SP-D OE) mice were prepared as previously described (24) and were outbred five generations with National Institutes of Health Swiss Black mice. SP-D OE mice had ∼5–10 times more SP-D protein in lung homogenates, 10–20 times higher extracellular concentrations of SP-D protein in bronchoalveolar lavage fluid, but no change in SP-A concentrations (24). Rat SP-D expressed in these mice has been shown to be biologically active by virtue of its ability to reverse the SP-D KO phenotype (i.e., emphysema and altered phospholipid homeostasis) in crossbreeding experiments (24).
SP-D control mice (SP-D wild type (WT)) prepared in parallel with SP-D KO and SP-D OE mice were combined in all experiments. These mice had been outbred with National Institutes of Health Swiss Black mice for five generations.
SP-A-deficient mice (SP-A KO) mice were prepared as previously described (25) and were outbred seven generations with National Institutes of Health Swiss Black mice. SP-A KO mice had no detectable SP-A in lung homogenate or in alveolar type II cells, and SP-D was not altered (25). Control mice (SP-A WT) were prepared in the same fashion as the SP-A KO mice (25) and were outbred seven generations into National Institutes of Health Swiss Black mice.
C1q-deficient mice (C1q KO) were prepared as described (10) and were outbred eight generations with C57BL/6 mice. C1q KO mice had no detectable circulating C1q protein and no detectable functional classical pathway hemolytic activity (10). C1q KO mice exhibit an autoimmune phenotype including the development of nephritis, anti-nuclear, and anti-histone Abs (10). There was no alteration in the number or morphology of AMs in the C1q KO mice (data not shown). Control mice (C1q WT) were bred in the same manner and were outbred eight generations with C57BL/6 mice.
In vivo clearance of apoptotic cells
Mice were anesthetized with Avertin (Milwaukee, WI), and apoptotic or IgG-opsonized, human PMNs (10 × 106), suspended in 50 μl PBS, were instilled intratracheally through a modified animal feeding needle (Fisher, Pittsburgh, PA). Whole lung lavage was performed 30 min following cell instillation with a total of 5 ml ice-cold PBS. Total cell counts were made for each mouse. A cytospin slide of lung lavage was stained with modified Wright’s Giemsa (Leukostat; Fisher), and evaluated for macrophage phagocytosis of apoptotic cells (see In vitro phagocytosis assays). An additional cytospin slide was fixed in 10% neutral buffered Formalin (Fisher) for 10 min, washed with PBS, and stained for myeloperoxidase activity with dianisidine (Sigma-Aldrich), as described previously (26). This method stains human PMNs deep brown, but does not stain mouse PMNs.
Isolation and purification of human SP-A and SP-D
SP-A was isolated from whole lung lavage fluid taken from patients with pulmonary alveolar proteinosis, as previously described (27). SP-D was isolated from transfected Chinese hamster ovary cells, as previously described (28). LPS was stripped from these proteins using polymixin.
Cell isolation and culture
Human monocyte-derived macrophages (HMDMs) and human PMNs were isolated using Percoll gradient centrifugation, as previously described (29). Monocytes were plated in 48-well tissue culture plates (BD Biosciences, Franklin Lakes, NJ), and matured into HMDMs by culturing for 6–8 days (30). Approximately 5 × 105 HMDMs were present in each well.
Mouse AMs were isolated from 8- to 12-wk-old C57BL/6 mice (Harlan Sprague-Dawley, Indianapolis, IN). Unstimulated mouse lungs were lavaged with 5 ml ice-cold PBS. Mouse AMs were pelleted, resuspended in X-vivo medium (BioWhittaker, Walkersville, MD), plated in 96-well tissue culture plates at 1 × 105 AMs/well, and incubated at 37°C in 5% CO2. AMs were allowed to adhere for 30 min, washed with X-vivo medium, and incubated for an additional 2 h before experimentation.
Human AMs were isolated by bronchoalveolar lavage from normal volunteers. The right middle lobe was lavaged with 240 ml 0.9% saline solution at room temperature. Human AMs were plated in 96-well tissue culture plates (BD Biosciences) at 1 × 105 AMs/well, and were treated in the same manner as described above for mouse AMs.
Jurkat cells, a human T cell leukemia line, were obtained from American Type Culture Collection (Manassas, VA) and were cultured as previously described (30).
Induction of apoptosis
Jurkat cells or human PMNs were exposed to UV irradiation at 254 nm for 10 min. Jurkat cells were cultured in RPMI 1640 with 10% FCS (Gemini Bio-Products, Woodland, CA) for 3.5 h at 37°C in 5% CO2. Human PMNs were cultured in RPMI 1640 with 1%, endotoxin-free, fraction V BSA (Sigma-Aldrich) for 2.5 h at 37°C in 5% CO2. Jurkat cells and human PMNs were ∼70% apoptotic by nuclear morphology, and have been shown to maintain intact cell membranes that exclude trypan blue (30)
Opsonization of PMNs
Freshly isolated human PMNs were resuspended in DMEM at 5 × 106 cells/ml. Mouse monoclonal anti-human CD45 IgG was added at 1 μg/1 × 106 PMNs, and PMNs were incubated with Ab for 30 min at 4°C while rotating. Chilled, opsonized PMNs were washed and used immediately.
In vitro phagocytosis assays
Six- to 8-day-old HMDMs, or freshly isolated AMs, were cocultured with apoptotic Jurkat cells or human PMNs in X-vivo medium (BioWhittaker) at 37°C in 5% CO2 in the absence of human serum for 1 h at a 5:1 ratio (apoptotic cell:macrophage). Macrophages were gently washed to remove uningested apoptotic cells, fixed, and stained with a modified Wright’s Giemsa stain (Fisher). Phagocytosis was determined by visual inspection of samples and was expressed as a phagocytic index, as previously described (31).
Single-ligand particles (Ebab): construction and ingestion assay
Human erythrocytes were obtained from normal donors and washed in PBS, pH 8.0. Thirty million erythrocytes were resuspended in 800 μl PBS (pH 8.0) and incubated with EZ-Link sulfo-NHS-LC-biotin (Pierce, Rockford, IL), according to manufacturer’s instructions. These biotinylated erythrocytes were washed twice, resuspended in PBS (pH 8), and incubated with 120 μg streptavidin (Sigma-Aldrich) at room temperature for 30 min. Erythrocytes were washed twice, resuspended in PBS (pH 7.4) and incubated with 5–10 μg biotinylated Ab or protein. Proper construction of Ebab ligand was confirmed by flow cytometry using the appropriate fluorochrome-conjugated Ab.
Target Ebab were added to HMDMs in duplicate wells at a 15:1 ratio and allowed to incubate at 37°C for 30 min. Unbound Ebab were washed away with HBSS. Uningested Ebab were lysed by adding deionized H2O for 10 s in one of the two duplicate wells. Cells were fixed with 0.75% glutaraldehyde, stained with dianisidine and H2O2 (Sigma-Aldrich), and counterstained with eosin. Two hundred cells were scored using light microscopy to quantify binding (before distilled water lysis) and engulfment of Ebab (after lysis).
Surface modulation experiments
Tissue culture plates were precoated with specific proteins to test their ability to modulate HMDM phagocytosis of apoptotic cells. C1q tails were prepared from human C1q (Quidel), as described previously (32). Proteins were suspended in DMEM at 0.1 μg/ml; 100 μl was added to each well and incubated for 2 h at 4°C. HMDMs (5 × 105) were adhered to precoated plates for 30 min at 37°C, following which target Ebab were added to each well at a 15:1 ratio for 1 h. Binding and ingestion of erythrocytes were assessed as described above.
Flow cytometry and fluorescence microscopy
Viable and apoptotic human PMNs were stained with 10 μg/ml FITC-conjugated SP-A or SP-D and analyzed on a FACScan cytometer, using PCLysys software (BD Biosciences), as described previously (5). Viable and apoptotic human PMNs were stained with 10 μg/ml FITC-conjugated SP-A or SP-D and analyzed using fluorescent microscopy, as previously described (5).
Statistics
The means were analyzed using ANOVA for multiple comparisons; when ANOVA indicated significance, the Dunnetts method was used to compare groups with an internal control. All data were analyzed using JMP (version 3) Statistical Software for the Macintosh (SAS Institute, Cary, NC) and are presented ± SEM.
Results
SP-D enhances clearance of apoptotic cells from the naive lung
Using a murine model of apoptotic cell removal, experiments were conducted to determine the relative importance of SP-D, SP-A, and C1q in the clearance of apoptotic cells from the naive lung (Fig. 1⇓). In these experiments, apoptotic PMNs were instilled into mice intratracheally. Thirty minutes later, the lungs were lavaged, and clearance of apoptotic cells was assessed by 1) quantitation of recovered apoptotic PMNs, and 2) measurement of AM ingestion of apoptotic cells (phagocytic index). These data suggested that clearance of exogenous, apoptotic PMNs from the naive mouse lung was specifically affected by SP-D, because recovery of apoptotic PMNs was increased in SP-D-deficient mice vs SP-D overexpressors (Fig. 1⇓A), and the AM phagocytic index was decreased in SP-D-deficient mice vs WT or SP-D overexpressors (Fig. 1⇓B). When lung lavage was examined separately, we found that endogenous apoptotic cells did not accumulate in naive SP-D-deficient mice compared with WT controls (data not shown). In contrast, neither SP-A, nor C1q deficiency altered clearance of exogenous, apoptotic PMNs (Fig. 1⇓, C–F). Intratracheal instillation of exogenous apoptotic PMNs did not induce an inflammatory response in any of the mice tested, as assessed by influx of murine PMNs (data not shown).
SP-D, but not SP-A or C1q, enhances clearance of apoptotic PMNs from the naive murine lung. Apoptotic PMNs (10 × 106) were instilled intratracheally. Thirty minutes later, whole lung lavage was performed, followed by analysis of recovered and ingested PMNs for SP-D KO, SP-D WT, and SP-D OE mice (A and B); SP-A KO and SP-A WT mice (C and D); and C1q KO and C1q WT mice (E and F). The mean ± SEM is shown for six to nine animals per group. ∗, Significantly different from SP-D OE (p < 0.05). ∗∗, Significantly different from SP-D KO (p < 0.05). ∗∗∗, Significantly different from all other groups (p < 0.05).
Because SP-D-deficient mice contain increased numbers of AMs, some of which become large and contain excessive amounts of phospholipids (23), we sought to determine whether defective clearance of apoptotic cells in SP-D-deficient mice was due to nonspecific dysfunction of these AMs, or specifically due to SP-D deficiency. To test this, SP-D-deficient, WT, and SP-D OE mice were challenged with IgG-opsonized, viable PMNs, and clearance was assessed. IgG-opsonized PMNs were used because they are recognized and ingested through macrophage FcγR, rather than receptors involved with clearance of apoptotic cells. These experiments did not demonstrate a significant clearance defect for IgG-opsonized PMNs by any of the groups tested (data not shown), suggesting that removal of apoptotic cells from the naive lung is specifically affected by the presence or absence of SP-D. SP-A and C1q appear to play lesser roles, individually, in the clearance of apoptotic cells from the unstimulated lung. Because SP-A has been shown to be involved with AM clearance of apoptotic cells in vitro (13), and because C1q has been shown to promote clearance of apoptotic cells in vitro (5), in the kidney (10), and in the peritoneum (9), experiments were performed to confirm the activity of individual collectins in AM phagocytosis of apoptotic cells in vitro.
Human SP-A, SP-D, and C1q enhance phagocytosis of apoptotic cells by human and murine AMs in vitro, independent of the apoptotic cell type
Maintenance of homeostasis in the naive lung requires that phagocytes recognize and remove a variety of apoptotic cells, including lymphocytes and PMNs. The AM is the professional phagocyte in the lung likely to be most involved in the removal of apoptotic cells (33), although structural cells, such as airway epithelium (34) and fibroblasts (35), almost surely contribute to this process. Accordingly, in vitro experiments were conducted to test whether pulmonary collectins (SP-A and SP-D) and the nonpulmonary, collectin-like C1q enhance AM uptake of apoptotic cells. These data demonstrated that human SP-A and SP-D increased human AM ingestion of apoptotic human PMNs (Fig. 2⇓A) and human Jurkat T lymphocytes (Fig. 2⇓B). C1q also increased human AM ingestion of apoptotic human PMNs and Jurkat T cells, but this increase did not reach statistical significance. Similarly, human SP-A, SP-D, and C1q increased murine AM uptake of apoptotic human Jurkat T lymphocytes (Fig. 2⇓C). In contrast, neither SP-A, SP-D, nor C1q augmented uptake of viable human PMNs by AMs (data not shown), suggesting that the effect of collectins on phagocytosis is specific to apoptotic cells. Therefore, SP-A, SP-D, and C1q enhanced AM ingestion of apoptotic cells independent of the apoptotic cell type.
SP-A, SP-D, and C1q enhance phagocytosis of apoptotic cells by human and murine AMs. Human, resident AMs were cocultured with apoptotic PMNs (A) (n = 5) and apoptotic Jurkat cells (B) (n = 5) in the presence of medium (control), SP-A (25 μg/ml), SP-D (10 μg/ml), or C1q (75 μg/ml). C, Murine, resident AMs were similarly cocultured with apoptotic Jurkat cells (n = 8). The mean phagocytic index as percent control ± SEM is shown for each group. Control mean phagocytic index = A, 6.6 ± 0.9; B, 8.0 ± 1.0; C, 4.1 ± 0.5. ∗, Significantly different from control (p < 0.05).
SP-A and SP-D enhance uptake of apoptotic cells through interaction with calreticulin and CD91 on the surface of AMs
Several collectin receptors have been identified that are associated with phagocytosis of microorganisms (36, 37), but recent evidence suggests that C1q and MBL drive ingestion of apoptotic cells through interaction with a macrophage surface complex composed of calreticulin and CD91 (5). Because SP-A and SP-D are known to bind calreticulin, experiments were performed to test whether SP-A and SP-D enhanced apoptotic cell uptake by AMs through interaction with the calreticulin/CD91 complex. In these experiments, Abs against calreticulin or CD91 inhibited SP-A- and SP-D-amplified phagocytosis of apoptotic cells (Fig. 3⇓). The presence of calreticulin and CD91 on the surface of murine and human AMs was confirmed by flow cytometry (data not shown).
SP-A and SP-D enhance ingestion of apoptotic cells through interaction with macrophage surface calreticulin and CD91. Murine, resident AMs were cocultured with apoptotic Jurkat cells in the presence of medium, SP-A (25 μg/ml), SP-D (10 μg/ml), or both. For each group, AMs received either no pretreatment (control) or pretreatment with Abs against calreticulin (CRT, 2 μg/1 × 105 cells), CD91 (2 μg/1 × 105 cells), or CD16/32 (2 μg/1 × 105 cells). The mean phagocytic index as percent control ± SEM is shown for six replicates per group. Control mean phagocytic index = 7.3 ± 1.5. ∗, Significantly different from control (p < 0.05).
Ingestion of SP-A- or SP-D-coated erythrocytes is dependent on interaction with macrophage surface calreticulin and CD91
Phagocytosis of apoptotic cells occurs through a complex set of interactions between multiple ligands and receptors, making it difficult to test the involvement of a single ligand-receptor pair. Recently, our laboratory has developed an assay system that allows evaluation of particle (erythrocyte) uptake through specific ligand-receptor interactions (35). Biotinylated erythrocytes are coated with biotinylated ligands through an avidin sandwich (Ebab-ligand). Erythrocytes coated in this manner bind to, and are taken up by, HMDMs depending on the ligand-receptor specificity. Because externally bound, but uningested, erythrocytes can be removed with hypotonic lysis, both binding and ingestion can be quantitated. The Ebab ingestion assay was used to confirm the ability of SP-A and SP-D to drive ingestion through interaction with the calreticulin/CD91 complex on HMDMs. In these experiments, Ebab-SP-A (Fig. 4⇓A) and Ebab-SP-D (Fig. 4⇓B) bound to, and were ingested by, HMDMs, and Abs with activity against human calreticulin and CD91 blocked these effects. In both experiments, erythrocytes coated with BSA (Ebab-BSA) were neither bound, nor ingested, by HMDMs (data not shown).
Ingestion of SP-A- or SP-D-coated erythrocytes is dependent on interaction with macrophage surface calreticulin and CD91. Erythrocytes coated with SP-A (Ebab-SP-A) (A) or SP-D (Ebab-SP-D) (B) were cocultured with HMDMs. HMDMs were pretreated with medium or Abs against calreticulin (CRT, 10 μg/1 × 105 cells), CD91 (10 μg/1 × 105 cells), or CD45 (10 μg/1 × 105 cells). The mean number of erythrocytes/200 HMDMs is indicated for three replicates per group. Filled bars indicate engulfed erythrocytes, and open bars indicate adherent erythrocytes. ∗, Significantly different from control (p < 0.05).
Competitive modulation of macrophage surface calreticulin or CD91 decreases ingestion of SP-A- or SP-D-coated erythrocytes
Experiments were conducted to determine whether competitive modulation of calreticulin or CD91 with known ligands would interfere with SP-A- or SP-D-driven ingestion of erythrocytes by HMDMs. To test this, HMDMs were adhered to wells that had been precoated with human serum albumin (HSA; negative control), or to known ligands of calreticulin (C1q collagenous tails) or CD91 (calreticulin and α2 macroglobulin). Importantly, plating HMDMs on surfaces coated with ligands for calreticulin or CD91 has been shown to remove these proteins from the upper surface of the macrophage, where the interaction occurs between the phagocyte and apoptotic cell (5). Modulation of calreticulin with human C1q tails decreased binding and ingestion of Ebab-SP-A (Fig. 5⇓A) and Ebab-SP-D (Fig. 5⇓B) compared with HSA controls. Similarly, competitive modulation of CD91 with either bovine calreticulin or human α2 macroglobulin also decreased binding and ingestion of Ebab-SP-A (Fig. 5⇓A) and Ebab-SP-D (Fig. 5⇓B) compared with HSA controls. In contrast, binding and ingestion of anti-CD32-coated erythrocytes were not impaired by any of the ligands used to precoat wells (data not shown). Therefore, these data suggest that calreticulin and CD91 are involved in phagocytosis mediated by SP-A or SP-D.
Competitive modulation of HMDM surface calreticulin or CD91 decreases SP-A- or SP-D-mediated ingestion of erythrocytes. Erythrocytes coated with SP-A (Ebab-SP-A) (A) or SP-D (Ebab-SP-D) (B) were cocultured with HMDMs that had been grown on plates precoated with HSA (control), α2 macroglobulin (α2M), calreticulin (CRT), or C1q tails. The mean number of erythrocytes/200 HMDMs is indicated for four to five replicates per group. Filled bars indicate engulfed erythrocytes, and open bars indicate adherent erythrocytes. ∗, Significantly different from control (p < 0.05).
Surface-bound SP-A and SP-D impair ingestion of calreticulin and CD91-coated erythrocytes
Experiments were performed to further explore the interaction among the pulmonary collectins (SP-A and SP-D), calreticulin, and CD91. In these experiments, HMDMs were adhered to wells that had been precoated with HSA (negative control), or human SP-A or SP-D (to competitively modulate calreticulin and CD91), and were then cocultured with Ebab conjugated with anti-human calreticulin Ab (Ebab-anti-calreticulin), or anti-human CD91 Ab (Ebab-anti-CD91). SP-A and SP-D specifically decreased binding and ingestion of Ebab-anti-calreticulin (Fig. 6⇓A) and Ebab-anti-CD91 (Fig. 6⇓B) in comparison with HSA control. In contrast, binding and ingestion of erythrocytes coated with anti-CD32 were not affected by surfaces precoated with HSA, SP-A, or SP-D (data not shown). Therefore, these results support the hypothesis that SP-A and SP-D specifically interact with the calreticulin/CD91 complex to initiate uptake of coated erythrocytes and apoptotic cells.
Surface-bound SP-A and SP-D impair erythrocyte ingestion through calreticulin and CD91. Erythrocytes coated with anti-calreticulin (CRT) (Ebab-anti-CRT) (A) or anti-CD91α (Ebab-anti-CD91α) (B) were cocultured with HMDMs grown on plates precoated with HSA (control), SP-A, or SP-D. The mean number of erythrocytes/200 HMDMs is indicated for three replicates per group. Filled bars indicate engulfed erythrocytes, and open bars indicate adherent erythrocytes. ∗, Significantly different from control (p < 0.05).
Apoptosis alters the binding pattern of SP-A and SP-D to PMNs
Enhanced uptake of apoptotic cells by SP-A and SP-D in vitro suggests that these collectins act as intercellular bridges between the phagocyte and apoptotic cell. The alternative possibility, that SP-A and SP-D enhance phagocytosis through activation of the phagocyte alone (activation-ligation mechanism), is contradicted by receptor modulation experiments, in which SP-A and SP-D down-regulated phagocytosis of coated erythrocytes and apoptotic cells (Fig. 6⇑). In addition, experiments were conducted to test whether SP-A and SP-D specifically bind to apoptotic cells. Binding of SP-A and SP-D to viable and apoptotic PMNs was quantitated using flow cytometry (Fig. 7⇓A). These experiments showed that FITC-labeled SP-A and SP-D bound to viable PMNs, and that the amount of binding was unaffected by apoptosis. However, fluorescence microscopy demonstrated that the pattern by which SP-A and SP-D bound to PMNs was significantly altered by apoptosis; SP-A and SP-D binding to viable PMNs was diffuse, whereas binding to apoptotic PMNs was highly localized to patches (Fig. 7⇓, B–E). Interestingly, this pattern of binding has been seen earlier with C1q (5). These data suggest that collectin aggregation on the surface of apoptotic cells may be required to facilitate ingestion.
Apoptosis alters the binding pattern of SP-A and SP-D to PMNs. A, Representative histograms showing that FITC-labeled SP-A and SP-D bound viable and apoptotic PMNs equally, and that SP-A and SP-D binding was greater than FITC-HSA. B–E, Fluorescence microscopy demonstrating that FITC-labeled SP-A (B) and SP-D (D) bound diffusely to viable PMNs. In contrast, FITC-labeled SP-A (C) and SP-D (E) bound to apoptotic PMNs in discrete patches.
Discussion
Collectins (SP-A, SP-D, and MBL) and C1q are members of the innate immune system that play key roles in the early recognition and removal of microorganisms (15, 17), modulation of the inflammatory response (38), and potentially a vital role in the clearance of cells dying by apoptosis (5, 13). In vivo and in vitro models of apoptotic cell clearance were used to determine the importance of individual pulmonary collectins (SP-A and SP-D) and C1q in the removal of apoptotic cells from the lung. These studies demonstrated that SP-D is a particularly potent modulator of apoptotic cell clearance in vivo and in vitro. In contrast, SP-A and C1q were clearly involved in apoptotic cell clearance in vitro, but genetic deletion failed to delay apoptotic cell removal in vivo, suggesting that SP-A and C1q contribute to the overall clearance machinery, but perhaps to a lesser extent in the naive lung. In vitro, SP-A and SP-D bound to apoptotic cells, and drove apoptotic cell ingestion through engagement of calreticulin and CD91 on the macrophage surface in a manner similar to MBL and C1q (5). These findings support the unifying hypothesis that calreticulin and CD91 function as a common receptor complex for collectin-mediated apoptotic cell removal.
C1q and SP-A deficiency had no effect on apoptotic cell removal from the naive lung. Given the redundancy of clearance mechanisms, these results were not surprising; however, additional factors may contribute to this lack of an in vivo effect. C1q is mainly produced by cells of the monocyte/macrophage lineage (39), is present at high concentrations in the plasma, but due to lung compartmentalization it is only present at low concentrations in the normal lung (39). Consequently, C1q largely moves into the lung only during times of inflammation with associated plasma leak. Our findings and the requirement for impaired vascular permeability suggest that C1q does not play a major role in homeostatic removal of apoptotic cells from the lung, but may play a larger role in apoptotic cell removal during episodes of severe pulmonary inflammation. In contrast, SP-A is present at high concentrations in the lung (17). High levels of pulmonary SP-A, however, are highly bound to surfactant lipids (38), and the effect of this protein/lipid interaction on the ability of SP-A to direct apoptotic cell clearance is, as yet, unknown. SP-A-deficient mice do not have reciprocal increases in pulmonary SP-D (25) nor increased numbers of AMs, either of which could potentially augment apoptotic cell clearance and mask any effect from SP-A deficiency. Therefore, while C1q and SP-A appear to contribute to the machinery available to drive removal of apoptotic cells from the lung, the redundancy of clearance mechanisms, and perhaps functionally low levels, appear to be sufficient to prevent C1q or SP-A deficiency from interfering with the overall ability of the naive murine lung to remove an apoptotic cell challenge.
In contrast, SP-D deficiency impaired clearance, and SP-D overexpression enhanced clearance of apoptotic cells from the naive lung, implying that a hierarchy exists for collectin potency in vivo. Unlike SP-A, the majority of pulmonary SP-D is not complexed with lipids (38) and may be more available in the lung to drive removal of apoptotic cells. It is important in this study to note that SP-D-deficient mice have 25% less SP-A, and 3-fold more saturated phosphatidylcholine in their lungs compared with WT controls (23). Lower SP-A levels, and perhaps more lipid-bound SP-A, could have contributed to impaired apoptotic cell removal and confounded the interpretation of delayed clearance in SP-D-deficient mice, were it not for data showing that SP-D overexpression enhanced removal of apoptotic cells. Still, it remains possible that lower levels of SP-A could enhance the defect in apoptotic cell clearance observed in SP-D-deficient mice. SP-D-deficient mice contain increased numbers of AMs, some of which are enlarged, filled with phospholipid, and foamy in appearance (23); this suggests the possibility that AM ingestion of apoptotic cells in SP-D-deficient mice is impaired due to a more general defect in phagocytosis. However, this seems not to be the case because these abnormal macrophages ingest IgG-opsonized PMNs through the Fcγ receptor, a receptor not associated with uptake of apoptotic cells. Together, these findings strongly imply that SP-D is directly involved with apoptotic cell removal from the naive murine lung, and that, when tested individually, SP-D is more potent than SP-A or C1q in vivo. However, during active pulmonary inflammation, associated with altered levels of surfactant and movement of C1q into the pulmonary compartment, SP-A and C1q may become more important effectors of apoptotic cell removal in the lung.
Human SP-A, SP-D, and C1q enhanced apoptotic cell ingestion by both human and murine AMs, confirming initial observations made by Schagat et al. (13), showing that SP-A and SP-D accelerate uptake of apoptotic rat PMNs by rat AMs. In contrast to their findings, our data also suggest that C1q enhances AM ingestion of apoptotic cells, albeit to a lesser degree than SP-A or SP-D. Differences in these results may be due to the higher concentration of C1q (75 μg/ml) used in our experiments vs that used by Schagat et al. (25 μg/ml). Alternatively, human C1q may more effectively drive ingestion of apoptotic cells by AMs of human and murine rather than rat origin.
Both SP-A and SP-D enhanced Ebab ingestion by HMDMs, confirming that macrophages derived from blood monocytes contain the surface recognition components necessary to interact with collectins primarily relegated to the lungs. It would seem likely then that, during an inflammatory response, macrophages migrating into the lung from the vascular space might remove dying, apoptotic inflammatory cells by a mechanism that uses the lung collectins. In addition, SP-D may participate in removal of apoptotic cells in sites other than the lung, because SP-D is also present in a variety of tissues (40). Schagat et al. found that SP-A did not increase ingestion of apoptotic cells by thioglycolate-elicited peritoneal macrophages. The reasons for these observed differences in macrophage response to SP-A are not clear, but most likely relate to pre-exposure of the peritoneal macrophages to MBL or C1q present in the inflammatory exudate, or due to up-regulation of other more dominant ingestion mechanisms in the peritoneal macrophage, such as the phosphatidylserine receptor.
Apoptosis altered the binding pattern, but not the overall attachment, of SP-A and SP-D to apoptotic PMNs. Our data demonstrated that SP-A and SP-D bound to both viable and apoptotic PMNs. However, the pattern by which SP-A and SP-D bound to PMNs was markedly influenced by apoptosis; SP-A and SP-D binding to viable PMNs was distributed evenly, while binding to apoptotic PMNs was localized to discrete patches. This binding pattern (i.e., equal intensity with altered distribution) of SP-A and SP-D is similar to that previously observed for C1q (5), and may be important in apoptotic cell interaction with the phagocyte.
C1q and MBL drive apoptotic cell ingestion through interaction with the calreticulin/CD91 complex on the phagocyte surface (5), suggesting a potential mechanism for SP-A- and SP-D-mediated apoptotic cell uptake. Apoptotic cells that are opsonized with C1q or MBL bind to calreticulin on the phagocyte surface by their collagenous tails. Calreticulin, in turn, has been shown to bind to CD91, which may then transduce an intracellular signal to initiate phagocytosis. Interestingly, the cytoplasmic tail of CD91 contains two NPXY endocytosis signal sequences. This tail region of CD91 also shares sequence homology with other receptors used for endocytosis, including ced-1, a gene from Caenorhabiditis elegans known to play a role in the phagocytosis of apoptotic cells (41). SP-A and SP-D also bind to calreticulin, suggesting that the calreticulin/CD91 complex is a common receptor mechanism for collectin and C1q-initiated apoptotic cell ingestion.
Three independent criteria were used to test whether a common receptor complex composed of calreticulin and CD91 was responsible for SP-A- and SP-D-mediated phagocytosis; these included: 1) Ab inhibition of calreticulin or CD91; 2) competitive modulation of macrophage calreticulin or CD91 using known ligands; and 3) competitive modulation of macrophage calreticulin or CD91 using SP-A or SP-D. Together, these experiments demonstrated that SP-A and SP-D enhanced phagocytosis of coated Ebab and apoptotic cells through a mechanism dependent on macrophage surface calreticulin and CD91. Because a similar relationship has been described for C1q and MBL (5), these data also support the concept that calreticulin and CD91 compose a common receptor complex for apoptotic cell ingestion by all collectins and C1q. However, because blockade of calreticulin or CD91 did not completely inhibit Ebab ingestion mediated through SP-A or SP-D, other receptors such as gp340 or C1qRp may still contribute to this process (36).
The specific form of phagocytosis induced by SP-A or SP-D was not investigated. However, recent evidence suggests that a form of stimulated phagocytosis, macropinocytosis, is the mode of uptake associated with ligation of calreticulin or CD91 (5), stimulation of the phosphatidylserine receptor (35), and more broadly the ingestion of apoptotic cells (5, 35). Therefore, ingestion via macropinocytosis would be the anticipated consequence of SP-A- or SP-D-driven phagocytosis through the calreticulin/CD91 complex.
Is there a relationship between deficiency of surfactant proteins, impaired clearance of apoptotic cells, and the development of disease in the lung? The phenotype of SP-D-deficient mice (i.e., spontaneous emphysema) suggests that a link may indeed exist, especially in view of the finding that apoptotic cells are increased in the alveolar septa of patients with emphysema (42). Chronic airway inflammatory diseases, such as cystic fibrosis, have also been associated with decreased SP-A and SP-D (43), and have been found to have high numbers of airway apoptotic cells (26). Protease/antiprotease imbalance in these diseases may also contribute to impaired collectin-mediated removal of apoptotic cells, either through direct degradation of collectins (44), or through degradation of their receptor. In this vein, we have observed that elastase cleaves calreticulin and CD91 (our unpublished observation), as well as the phosphatidylserine receptor (26). Therefore, impaired clearance of apoptotic cells, either through actual or functional collectin deficiency, may contribute to the development of emphysema or airway inflammation in cystic fibrosis.
The collectins and C1q are proteins of the innate immune system with suggested roles in the clearance of microorganisms and modulation of the inflammatory response. Data reported in this work support the concept that the extended collectin family is also broadly involved in another process for control of inflammation, namely the removal of cells dying by apoptosis. In this context, it would appear that SP-D, and perhaps SP-A, is integral to the maintenance of homeostasis in the naive lung, through the orderly removal of dying cells before they have the opportunity to release internal contents and initiate an inflammatory response. These results may have particularly important implications for human diseases, such as cystic fibrosis, in which clearance of apoptotic cells is defective, control of inflammation is dysregulated, and levels of SP-A and SP-D are decreased.
Acknowledgments
We thank J. Whitsett and T. Korfhagen for kindly providing SP-A-deficient, SP-D-deficient, and SP-D OE mice.
Footnotes
-
↵1 This work was supported by grants from The National Institutes of Health HL03724 (to K.E.G.), GM48211 (to P.M.H.), and HL67671 (to P.M.H.), and the Cystic Fibrosis Foundation (to K.E.G. and R.W.V.).
-
↵2 R.W.V. and C.A.O. contributed equally to the preparation of this manuscript.
-
↵3 Address correspondence and reprint requests to Dr. R. William Vandivier, National Jewish Medical and Research Center, 1400 Jackson Street, Room D505, Denver, CO 80262. E-mail address: vandivierb{at}njc.org
-
↵4 Abbreviations used in this paper: SP-A, surfactant protein A; SP-D, surfactant protein D; AM, alveolar macrophage; Ebab, single-ligand particle; HMDM, human monocyte-derived macrophage; HSA, human serum albumin; KO, knockout; MBL, mannose-binding lectin; SP-D OE, SP-D-overexpressing; PMN, neutrophil; WT, wild type.
- Received April 3, 2002.
- Accepted July 25, 2002.
- Copyright © 2002 by The American Association of Immunologists
References
In this issue
