Mannose-Binding Lectin-Deficient Mice Display Defective Apoptotic Cell Clearance but No Autoimmune Phenotype
http://www.100md.com
免疫学杂志 2005年第6期
Abstract
Mannose-binding lectin (MBL) is a circulating serum protein that is sequestered to sites of inflammation and infection. MBL is a member of the collectin family with structural similarities to the lung collectins and functional similarities to C1q. Both MBL and C1q activate complement; C1q activates the classical pathway and MBL the lectin pathway. Here we demonstrate that MBL binds apoptotic cells in vitro and confirm a role for MBL in clearance of apoptotic cells in vivo. Despite MBL null mice demonstrating defective apoptotic cell clearance they did not develop spontaneous autoimmunity, lymphoproliferation, or germinal center expansion although increased numbers of peritoneal B1 cells were detected. These data demonstrate an important in vivo role for MBL in clearance of dying cells and adds the MBL null animals to the few animals with demonstrable in vivo apoptotic cell clearance defects. Moreover, it demonstrates that failure of apoptotic cell clearance can be dissociated from autoimmunity.
Introduction
Mannose-binding lectin (MBL)4 is a collectin as it displays an amino-terminal collagen stalk followed by an -helical coiled domain that orients the spatial organization of the carboxyl-terminal C-type lectin domains (1). Like the lung collectins, MBL contributes to first-line host defense and innate immune responses (2, 3). However, unlike surfactant protein A (SPA) and surfactant protein D (SPD), which are found predominantly in the lung, MBL circulates in serum as a multimer of trimers. The production of MBL in the liver is increased 2- to 3-fold in response to infection, and it is rapidly sequestered to sites of inflammation. One consequence of MBL-ligand interaction is the activation of complement via the lectin pathway. MBL forms a complex with three serine proteases termed mannose-binding associated proteases one of which (MASP2) co-opts the classical complement pathway convertase that in turn activates the complement cascade (4). In addition, the lectin pathway can also be activated by a family of related molecules, the ficolins.
MBL bears many functional and structural similarities to C1q; both molecules form a trimeric structure resembling a bunch of tulips, which further oligomerize to form higher order multimers and are able to activate complement. Importantly, C1q unlike MBL requires recognition of Ab-opsonized particles to activate complement. Defects in the first component of complement are strongly associated with autoimmunity such that >95% of patients deficient in C1q develop a systemic autoimmune syndrome (5). In addition, MBL haplotypes that specify low levels of an aberrant form of MBL unable to activate complement are also reported to be associated with a low but increased risk of autoimmunity (6, 7). Apoptotic cells have been shown to be a preferential source of many autoantigens (8) and failure to promptly remove dying cells has been implicated in the pathogenesis of systemic lupus erythematosus (5, 9, 10), a prototypic autoimmune disease. C1q has been shown to bind and opsonize dying cells (11, 12), and consequently, it has been suggested that this failure of C1q-mediated clearance of apoptotic cell renders dying cells immunogenic with subsequent development of autoimmunity. In further support of this "waste disposal" model, C1q-deficient mice, upon certain backgrounds, develop a lupus-like syndrome and nephritis characterized by multiple uncleared apoptotic cells (13). However, the mechanism by which dying cells come to incite an inflammatory or autoimmune response is poorly understood.
Recent data demonstrates that all of the collectins (MBL, SPA, and SPD) bind cells undergoing apoptosis in vitro (12). Furthermore, SPD null animals have defective in vivo clearance of apoptotic cells in the lung (14). These observations suggest that MBL could also have a role in apoptotic cell clearance in vivo and, along with the reports of increased risk of autoimmunity in MBL low individuals (6, 7), have raised the question of whether lack of MBL might result in autoimmunity as is observed with C1q deficiency (15). To study the role of MBL in vivo, we have generated mice deficient in both genes encoding mouse MBL (MBLA and MBLC) (3), and here we report the role of MBL in autoimmunity. We show that lack of MBL does indeed result in delayed removal of dying cells in vivo but, despite this defect, MBL null mice do not develop autoimmunity even when they are aged for 18 mo on a lupus-prone genetic background (129xC57BL/6). Taken together, these observations suggest that in MBL-deficient mice, failure of apoptotic cell clearance can be dissociated from autoimmunity.
Materials and Methods
Mice
MBL null mice were generated as described previously (3). Briefly, embryonic stem cells from 129 mice were targeted and injected into a C57BL/6 blastocyst and then fostered into a pseudo-pregnant female. Transmission from the chimeric progeny was determined by crossing with a C57BL/6 to generate MBLA or MBLC null heterozygous animals. MBLA/C–/– mice were generated and maintained thereafter by intercrossing on a mixed 129xC57BL/6 background in a specific pathogen-free facility. At 18 mo, mice were sacrificed, and the histology of all tissues was examined by the Charles River pathology service. Mice used for other experiments were between 8 and 12 wk of age. Age-, sex-, and strain-matched controls were maintained in the same facility. Thymocytes were isolated from 6-wk-old mice (The Jackson Laboratory).
Generation of apoptotic cells
Apoptotic cells were generated by allowing single cell suspensions of thymocytes to undergo constitutive apoptosis by ageing overnight in RPMI 1640 supplemented with antibiotics and 0.5% BSA. For most experiments, cells were stained using a green cell tracker dye (CMFDA) (Molecular Probes) before overnight culture.
Analysis of apoptotic cell—MBL binding and fluorescent microscopy
Apoptotic cells were stained for 15 min on ice with annexin V Alexa 488 (Molecular Probes) or Cy3 rhMBL (NatImmune) in high calcium buffer (2.5 mM CaCl2 in HEPES). Specificity of MBL binding was determined by comparing with an irrelevant Cy3-conjugated protein (BSA). For certain experiments, MBL binding was inhibited by 10 mM EDTA or 200 mM mannan. For microscopy, cells were washed once with high calcium buffer and examined immediately by fluorescent microscopy. Images were captured and analyzed using Openlab software (Improvision).
In vivo phagocytosis assay
Apoptotic cells were resuspended in PBS, and mice were injected i.p. with 10 x 106 cells/per mouse. Mice were sacrificed at 30 min and the peritoneum was lavaged. Phagocytosis was analyzed by FACS gating on F4/80-positive cells. Percentage of phagocytosing cells was determined. In addition, the "phagocytic index" was used to estimate the total amount of apoptotic cells removed by calculating (the mean fluorescence intensity of the phagocytosing cells) x (the percentage of cells that had phagocytosed).
Phenotyping of macrophages, dendritic cells (DCs), and B cells
Macrophages and B1 cells were isolated from the peritoneum and germinal center B cells, and DCs were isolated from spleen. For DCs, a single cell spleen preparation was generated and purified using CD11c MACS beads (Miltenyi Biotec). For FACS analysis, all Abs used were from BD Pharmingen except F4/80-allophycocyanin (Caltag Laboratories) and peanut agglutinin (Vector Laboratories).
Assessment of autoimmunity
Aged mice were assessed for proteinuria using urine dipsticks (Multistix). At 18 mo, mice were sacrificed and organs examined for histological evidence of inflammation or autoimmunity. Terminal bleed serum was examined for autoantibodies as follows: anti-dsDNA Abs were assayed in a 1/100 dilution using an anti-mouse dsDNA-specific ELISA kit (Alpha Diagnostics International); 2-fold serial dilutions from 1/80 to 1/2560 of serum anti-nuclear Abs (ANAs) were determined by binding to prepared Hep2 cell slides (Scimedx Corporation) and detected with FITC-labeled anti-mouse Ig. Slides were then scored by fluorescent microscopy. Images were captured at 5 and 10 exposures and scored positive if nuclear binding was detected at 5 exposure and intermediate if only seen at 10. Intermediate binding was repeated twice and deemed negative if only detected with 10 exposures on both occasions. Final ANA titer was the highest dilution that was positive.
Results
MBL clusters on apoptotic cells
Previous studies demonstrating MBL binding to dying cells often used MBL purified from serum, and it is possible that such preparations contain IgM, known to associate with MBL in circulation. IgM (particularly natural Abs) are also able to recognize apoptotic cells, raising the possibility that it is IgM and not MBL that interacts with dying cells in these experiments. We therefore set out to determine whether pure rMBL could bind only to apoptotic cells or whether rMBL also bound to live cells before exposure of phosphatidyl serine (PS). To do this, we directly conjugated Cy3 to rMBL and used this to define the interaction of MBL with dying cells after 0, 16, or 24 h in culture. As a negative control binding was compared with an irrelevant Cy3-conjugated protein (BSA). Surprisingly, even immediately after isolation, MBL binding could be detected, and cells could be separated into three categories: annexin–/rMBL–, rMBL+/annexin–, and annexin+/rMBL+ (Fig. 1) indicating that MBL can bind to live cells before the externalization of PS. This binding was specific and could be inhibited by 10 mM EDTA (Fig. 1) and 200 mM mannan (data not shown) suggesting that MBL was binding through the carbohydrate recognition domain. Binding was detected not only with thymocytes but also to a viable murine fibroblast cell line, L929 (data not shown).
FIGURE 1. MBL binds to live cells before the induction of apoptosis. Murine thymocytes were isolated, and PS exposure was determined by annexin V Alexa-488 and MBL binding by using purified rhMBL-Cy3 in high calcium buffer. A, Specificity of MBL binding was determined by comparison to control Cy3-BSA and in the presence of 10 mM EDTA. B, MBL binding was assessed both to the live cells (annexin V negative) and to the small population of contaminating apoptotic (annexin V positive) cells. Control Cy3-BSA binding is shown (left). C, Microscopic examination of MBL binding to cells before overnight culture. Arrow, A MBL and annexin V-negative cell; arrowhead, a MBL-Cy3 and annexin V-double- positive cell. Original magnification, x40.
To determine whether MBL binding altered during the process of apoptosis, binding was assessed at intermediate time points (Fig. 2). Using two-color FACS, MBL binding between live cells (determined either by forward and side scatter to sit in (R1) or as annexin– cells) and apoptotic cells (R2 or annexin+ cells) was comparable, and no difference in MBL binding was detected between the two populations. Furthermore, although the number of annexin+ cells increased with time indicating that the cells were entering into apoptosis, the majority of the cells remained intact as determined by the forward/side scatter, and no increase in MBL binding was seen. After 24 h, the majority of the cells were annexin+ (varying from culture to culture from 60 to 90%), and MBL remained bound to all cells. Importantly, when examined by fluorescent microscopy, the pattern of binding of MBL to apoptotic cells was patchy (Fig. 2) and differed from the diffuse pattern seen on the live cells (Fig. 1C). Taken together, these data demonstrate conclusively that MBL binds not only to dying cells but also to certain live cells and that it does so directly and independently of contaminating Ab. However, as death progresses, MBL adopts a distinct clustered pattern on the cell surface suggesting alteration in ligand distribution on the dying cell surface. This suggests that a clustered configuration of ligands on the surface of apoptotic cells may be required for MBL-mediated phagocytosis.
FIGURE 2. MBL clusters on cells during the process of apoptosis. Murine thymocytes were aged in serum-free conditions for 24 h. A, Representative two-color FACS analysis of MBL and annexin V binding (a 16-h time point is shown). B, FACS analysis of MBL binding to "live cells" (R1) and apoptotic cells (R2) at different time points after isolation. C, Microscopy demonstrating clustering of MBL on the surface of an apoptotic cell at 24 h.
MBL mediates clearance of apoptotic cells in vivo
SPD and C1q have been shown to increase phagocytosis of apoptotic cells by macrophages in both in vitro and in vivo systems. However, although MBL can mediate apoptotic cell engulfment in vitro, the contribution of MBL in vivo had not been defined. To determine whether MBL was able to mediate the removal of dying cells in vivo, we injected fluorescently labeled apoptotic thymocytes i.p. into either wild-type (WT) or MBL null mice. After 30 min the peritoneum was lavaged and macrophages were stained for F4/80. In a typical experiment the percentage of macrophages from MBL null mice that phagocytosed was 16% whereas WT macrophage phagocytosis was higher and typically greater than 25% of macrophage-ingested apoptotic cells (Fig. 3). This reduction equated to a 38% decrease in percentage of phagocytic cells in the absence of MBL. In addition, a more profound defect in the total number of apoptotic cells cleared (as shown by 49% reduction in the phagocytic index) was also demonstrated (Fig. 2). Taken together, these observations indicate an important role for MBL in clearance of apoptotic cells in vivo. Furthermore, because these mice are ficolin and C1q sufficient, our results indicate a non-redundant role for MBL in apoptotic cell clearance in vivo.
FIGURE 3. The role of MBL phagocytosis of apoptotic cells in vivo. Fluorescently labeled apoptotic fluorescent thymocytes were instilled into the peritoneum of WT and MBL null animals. Peritoneal lavage was collected at 30 min and macrophages were stained using F4/80-APC. A, Representative FACS density plot after 30 min of phagocytosis. Percentage of macrophages phagocytosing dying cells (B) and phagocytic index (C) were determined by FACS (mean ± SEM; n = 4 WT and n = 6 MBL null). Significance was determined by Student’s t test. *, p < 0.05; **, p < 0.01.
Failure of apoptotic cell clearance in MBL null mice is not associated with spontaneous activation of APCs
MBL might be predicted to alter macrophage phenotype for a number of reasons. First, collectins can directly modulate macrophage function (16), and second, delayed apoptotic cells clearance itself might result in accumulation of proinflammatory secondarily necrotic cells able to stimulate APCs. To exclude the possibility that impaired removal of apoptotic cell by macrophages in MBL null animals was a consequence of altered macrophage phenotype, we choose to determine whether lack of MBL might be associated with spontaneous activation of APCs in unchallenged mice. Our results revealed that peritoneal macrophages expressed similar amounts of costimulatory molecules and MHC in both WT and MBL null animals (Fig. 4). In addition, splenic DCs isolated from MBL null mice did not display enhanced levels of MHC II, CD40, CD80, or CD86 compared with cells isolated from WT mice (Fig. 4) indicating that lack of MBL does not result in spontaneous activation of APCs.
FIGURE 4. Phenotype of APC in WT vs MBL null mice. Splenic DCs and peritoneal macrophages isolated from resting MBL null and WT animals were phenotyped for activation markers CD40, CD80, CD86, and MHC II (light gray is unstained control).
MBL null mice have an increased number of peritoneal B1 cells
B1 cells are a subpopulation of B cells that populate the pleural and peritoneal cavities. B1 cells express a restricted number of germline-encoded B cell receptors and are the major source of natural Abs (17). These Abs demonstrate low affinity for certain bacteria and self-Ags. It has been suggested that B1 cells have a possible role in the development of autoimmunity because abnormalities in B1 cells have been reported in a number of autoimmune models (18, 19). The best defined B1 clone produces a T15 anti-phosphoryl choline Ab that recognizes not only pathogens but modified lipids and apoptotic cells (20, 21), and because of its receptor specificity, it has been suggested that this BCR may be selected on apoptotic cells encountered during B cell ontogeny. Thus, we hypothesized that the failure of apoptotic cell clearance seen in the peritoneum might be associated with a change in the number of B1 cells. To evaluate this question we analyzed peritoneal B cells by FACS, which revealed an expansion of peritoneal B1 cells as defined by B220int, CD11b+, and CD43+ (Fig. 5). These B cells expressed low or no CD5 and hence could be defined as B1-b cells.
FIGURE 5. Phenotype of B cells in WT vs MBL null mice. A, Peritoneal B cells were phenotyped, and percentages of B220int/high, CD11b+ B1 cells (R1) vs B220high, CD11b– B2 cells (R2) are shown. CD5 expression on B1 cells is shown for representative WT (light gray histogram) vs MBL null animal (dark gray histogram). B, Spleen CD95+, peanut agglutinin+ germinal center B cells (indicated in gated region) in WT vs MBL null animals (percentage of B220+ cells shown) (FACS density plots are shown for clarity).
MBL null mice have reduced germinal center B cells and no spontaneous lymphoproliferation
B1 cells normally produce low affinity IgMs or IgG3s but are able to differentiate to produce pathogenic autoreactive Abs if they exit the peritoneum and are activated in secondary lymphoid organs where they are educated in germinal centers on follicular DCs that have captured autoantigens (9). To determine whether the expanded B1 cell numbers would correlate with generalized lymphoproliferation, we phenotyped splenic B cells from unchallenged animals. Interestingly, germinal center B cell numbers were decreased by >60% in MBL null animals when compared with WT animals (Fig. 5). Taken together these results suggest that, although increased numbers of possibly autoreactive B1 cells might be associated with failed clearance of apoptotic cells, in the absence of MBL neither germinal center expansion nor lymphoproliferation occurred spontaneously.
MBL null mice dissociate failure of apoptotic cell clearance from autoimmunity and failed to develop high titer ANAs
Despite the lack of overt lymphoproliferation, the expansion of possibly autoreactive B1 cells in the peritoneum of MBL null mice suggested that these mice might also be predisposed to autoimmunity. Thus to determine whether MBL deficiency per se predisposes these mice to autoimmunity, we maintained MBL mice on autoimmune mixed 129/C57BL/6 background for 18 mo. Histological examination failed to demonstrate any significant autoimmune pathology in the organs of MBL null mice, although we did detect an acidophilic macrophage infiltrate in the lungs of older animals (L. M. Stuart, manuscript in preparation) (Table I). In particular, we failed to detect proteinuria or the appearance of free apoptotic cells in the glomeruli (Fig. 6) when compared with age-matched WT controls. dsDNA titers as assessed by ELISA were comparable between WT and MBL null animals at 18 mo and were significantly lower than those seen in the positive control MRL serum (Fig. 6). Similarly, although ANAs were detectable (particularly in the female animals; Fig. 6), no increase in experimental animals over WT was seen. In contrast, MBL deficiency appeared to provide protection from developing high titer ANAs (median dilution for WT 1/640-1280 vs MBL null 1/80–160). These observations indicate that failure to clear apoptotic cells is not sufficient to generate an autoimmune phenotype.
Table I. Histology of aged MBL null vs WT mice
FIGURE 6. Assessment of autoimmunity in MBL null animals. A, Representative renal histology of 18-mo-old WT and MBL null animals (n = 5 for both groups). B, ANA titers in serial diluted 18-mo serum and dsDNA titers (1/100 dilution) in male and female MBL null and WT animals (n = 5 or more for each group). Median is represented as horizontal red bar. Significance was determined by a Mann-Whitney U test. *, p < 0.025.
Discussion
Collectins are a family of soluble opsonins able to recognize a broad range of pathogens and have important and diverse roles in innate immune responses. In addition, MBL, through its ability to co-opt the classical convertase, is also able to activate complement and hence bears functional similarities to the first component of the classical pathway, C1q. Apoptotic cells are a preferential source of many of the auto-Ags found in systemic lupus erythematosus, and the importance of the prompt removal of dying cells in preventing this disease has been emphasized by a number of recent reports indicating that delayed apoptotic cell removal can precipitate autoimmunity (11, 22, 23). However, exactly how apoptotic cells are removed in vivo and what determines whether the uncleared cells come to incite a proinflammatory or autoimmune response is poorly understood. Both the collectin SPD and complement component C1q are known to mediate recognition and removal of apoptotic cells in vivo but the contribution of MBL in vivo has not been explored. Here, using mice deficient in both MBL genes (MBLA and MBLC), we demonstrate that MBL contributes to clearance of apoptotic cells in vivo. Furthermore, analysis of MBL null animals revealed abnormal B cell homeostasis suggesting an important immunoregulatory role for this molecule in vivo. However, despite the delay in removal of dying cells in this context, MBL-deficient mice do not develop autoimmunity despite being on an autoimmune prone (129xC57BL/6) background indicating surprisingly that failed apoptotic cell clearance can be dissociated from autoimmunity.
Although the absence of C1q results in delayed clearance of apoptotic cells that are introduced into the peritoneum, the dying cells are ultimately removed, indicating that alternative mechanisms are able to partially compensate for the C1q defect (11). All of these partially redundant recognition molecules have not been identified although recent work has suggested that both the Mer receptor tyrosine kinase (22) and MFGE8 (23, 24) might contribute in vivo (reviewed by Savill et al. (25)). Here we demonstrate that MBL, a molecule with both structural and functional similarities to C1q, also contributes to apoptotic cell removal in vivo. Clustering of collectins on the surface of apoptotic cells and the distinct pattern of binding they adopt is thought to be important for these opsonins to trigger phagocytosis (12, 14). Interestingly, not only do both C1q and MBL cluster on apoptotic cells as death progresses, but they also both contribute to clearance in vivo. MBL and C1q have been shown to recognize distinct and only partially overlapping surface ligands on dying cells (26), and it is likely, therefore, that the similarities in the in vivo phenotype of the C1qA–/– and MBL null animals is a consequence of their shared ability to trigger a common pathway or mechanism of uptake. This hypothesis is in keeping with the observed similarity in the extent of the defect in apoptotic cell clearance in the peritoneum of both C1q and MBL null animals. Furthermore, these data raise the intriguing possibility that these molecules work together, either as a macromolecular complex or to trigger a common engulfment pathway, rather than mediate distinct and independent pathways of apoptotic cell removal.
Delayed clearance apoptotic cells as seen in the MBL null animals might be predicted to modulate macrophage activation for two reasons. First, apoptotic cells that are not rapidly cleared in vivo are thought to undergo secondary necrosis and spill their potentially proinflammatory intracellular contents into surrounding tissues. Second, previous in vitro studies have demonstrated that phagocytosis of apoptotic cells actively inhibits APCs (27, 28, 29, 30). In addition, recent work has emphasized a role for the other collectins, SPA and SPD, as negative regulators of macrophages, a function mediated through their ability to ligate the inhibitory receptor SIRP- (16). However, surprisingly our data suggest that neither the absence of MBL itself nor the delayed removal of apoptotic cells is sufficient to result in spontaneous activation of APCs. Importantly it indicates that, in these circumstances, late apoptotic/necrotic cells are not proinflammatory and are ultimately cleared in a nonphlogistic manner. In support of this, our previous in vitro data suggests that removal of late apoptotic neutrophils can be mediated by a 2 integrin (complement receptor) independent mechanism, involving thrombospondin-1 and v3, which allows macrophages to ingest late apoptotic cells without eliciting inflammatory cytokine secretion (31).
Recent data have highlighted the importance of apoptotic cell clearance and regulation of germinal centers reaction in autoimmune disease; MFGE8 null animals demonstrate impaired apoptotic B cell clearance by tingible body macrophages, expanded germinal centers, and autoimmunity (23). Here, we have demonstrated that B cell homeostasis and germinal centers are affected by the absence of MBL, indicating an important immunoregulatory role for this molecule. Although we do not know the exact mechanism for the expansion of B1 cells in MBL null animals we would suggest that it is either in response to delayed apoptotic cell clearance or that MBL has a negative homeostatic role on B1 cell development. Intriguingly, in support of the former hypothesis, mice deficient in soluble IgM also demonstrate impaired clearance of dying cells, expanded B1 numbers, and accelerated autoimmunity (32). Our observation that the numbers of germinal center B cells, both resting and after immunization (L. M. Stuart and K. Takahshi, unpublished data), are lower in MBL null mice and that the autoantibodies made in the MBL null mice are neither associated with pathology nor are they made at high titer is intriguing. One speculation is that the absence of MBL does not support FDC-B interactions and that germinal center formation may be impaired in these animals. It is possible, therefore, that although MBL deficiency is associated with abnormal disposal of apoptotic cell-derived self-Ags, the absence of MBL is not permissive for expansion of truly pathogenic self-reactive B cells, protecting these mice from developing systemic lupus erythematosus. In this regard, it will be important to determine what role MBL plays in germinal center reactions (33) and whether this might underlie the different outcomes of failed apoptotic cell clearance in C1q and sIgM null animals.
The data we have presented indicates that both MBL and C1q null mice display a similar phenotype in the defect in apoptotic cell clearance, but they diverge in susceptibility to autoimmunity. An important difference between C1q and MBL is the ability of C1q to trigger complement deposition on the surface of the apoptotic cell (26) and to recognize Ab-opsonized apoptotic cells and immune complexes. It is possible that it is specifically failed clearance of Ab opsonized apoptotic cells, and associated proinflammatory responses, that renders apoptotic cells more immunogenic in the context of C1q deficiency. Of note, unlike C1q-deficient patients, MBL-deficient humans show only a weak increased risk for development of autoimmunity suggesting that these mice mimic human diseases (34). Furthermore, in support of this animal model, a recent report has suggested a protective role of MBL deficiency in female lupus patients in the Canary Islands who, in its absence, produce lower autoantibody levels and have a later onset of disease (35). C1qA–/– animals develop spontaneous nephritis and autoimmunity only on the mixed 129xC57BL/6 or MRL/Mp background (36), and in this regard, it will be of interest to determine whether lack of MBL can modulate other autoimmune strains such as the MRL/Mp or MRL/lpr where different genetic factors contribute.
The observed similarities and differences between the C1q- and MBL-deficient animals leave many unanswered questions about the role of these opsonins in control of autoimmunity and inflammation. It is curious that MBL and C1q have many broad similarities in that they both co-opt the classical pathway convertase to initiate the complement cascade and, pertinent to this study, appear to have a similar role in apoptotic cell clearance in vivo. However, this study demonstrates that the two molecules have non-redundant functions. In the absence of C1q, apoptotic cell clearance is not blocked completely but delayed, occurring with time by a compensatory mechanism. Here, we have shown that one such mechanism is MBL, and it is possible that this alone might be sufficient to alter the balance in favor of a proinflammatory outcome in the absence of C1q. Perhaps it is increased MBL-mediated clearance in the C1q-deficient animals that is deleterious and that its absence actively protects against end organ damage. This hypothesis is particularly pertinent in considering the kidney, where MBL is highly expressed and is likely therefore to be a significant contributor regulating response to, and removal of, uncleared apoptotic cells in the glomeruli of C1q null animals. We suggest that further definition and careful comparison of the fate of dying cells and autoreactive B cell clones in the C1qA–/– vs MBL null animals will be vital in determining why one predisposes to lupus whereas the other does not. We believe that, in the future, such comparative studies will lead to new insights into the genesis of the autoimmune syndromes associated with failed apoptotic removal.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Drs. Marina Botto and Adam Lacy-Hulbert for critical reading of the manuscript and helpful comments.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by a Wellcome Trust Clinician Scientist Grant R36731 (to L.M.S.), Wellcome Trust Programme Grant R068089 (to J.S.), and National Institutes of Health Grants R01 AI42788 and P01 AI52343–03 (to R.A.B.E.).
2 L.M.S. and K.T. contributed equally to this study.
3 Address correspondence and reprint requests to Dr. Lynda M. Stuart, Laboratory of Developmental Immunology, Massachusetts General Hospital, Jackson 14, 55 Fruit Street, Boston, MA 02114. E-mail address: lstuart{at}partners.org
4 Abbreviations used in this paper: MBL, mannose-binding lectin; PS, phosphatidyl serine; ANA, anti-nuclear Ab; DC, dendritic cell; WT, wild type; SPA, surfactant protein A; SPD, surfactant protein D.
Received for publication October 8, 2004. Accepted for publication December 27, 2004.
References
Sheriff, S., C. Y. Chang, R. A. Ezekowitz. 1994. Human mannose-binding protein carbohydrate recognition domain trimerizes through a triple -helical coiled-coil. Nat. Struct. Biol. 1:789.
Epstein, J., Q. Eichbaum, S. Sheriff, R. A. B. Ezekowitz. 1996. The collectins in innate immunity. Curr. Opin. Immunol. 8:29.
Shi, L., K. Takahashi, J. Dundee, S. Shahroor-Karni, S. Thiel, J. C. Jensenius, F. Gad, M. R. Hamblin, K. N. Sastry, R. A. Ezekowitz. 2004. Mannose-binding lectin-deficient mice are susceptible to infection with Staphylococcus aureus. J. Exp. Med. 199:1379.
Matsushita, M., Y. Endo, M. Nonaka, and T. Fujita. 1998. Complement-related serine proteases in tunicates and vertebrates. Curr. Opin. Immunol.
Pickering, M. C., M. Botto, P. R. Taylor, P. J. Lachmann, M. J. Walport. 2000. Systemic lupus erythematosus, complement deficiency, and apoptosis. Adv. Immunol. 76:227.
Ip, W. K., S. Y. Chan, C. S. Lau, Y. L. Lau. 1998. Association of systemic lupus erythematosus with promoter polymorphisms of the mannose-binding lectin gene. Arthritis Rheum. 41:1663.
Davies, E. J., N. Snowden, M. C. Hillarby, D. Carthy, D. M. Grennan, W. Thomson, W. E. Ollier. 1995. Mannose-binding protein gene polymorphism in systemic lupus erythematosus. Arthritis Rheum. 38:110.
Casciola-Rosen, L. A., G. Anhalt, A. Rosen. 1994. Autoantigens targeted in systemic lupus erythematosus are clustered in two populations of surface
Herrmann, M., O. M. Zoller, M. Hagenhofer, R. Voll, J. R. Kalden. 1996. What triggers anti-dsDNA antibodies?. Mol. Biol. Rep. 23:265.
Mevorach, D., J. L. Zhou, X. Song, K. B. Elkon. 1998. Systemic exposure to irradiated apoptotic cells induces autoantibody production. J. Exp. Med. 188:387.
Taylor, P. R., A. Carugati, V. A. Fadok, H. T. Cook, M. Andrews, M. C. Carroll, J. S. Savill, P. M. Henson, M. Botto, M. J. Walport. 2000. A hierarchical role for classical pathway complement proteins in the clearance of apoptotic cells in vivo. J. Exp. Med. 192:359.
Ogden, C. A., A. deCathelineau, P. R. Hoffmann, D. Bratton, B. Ghebrehiwet, V. A. Fadok, P. M. Henson. 2001. C1q and mannose binding lectin engagement of cell surface calreticulin and CD91 initiates macropinocytosis and uptake of apoptotic cells. J. Exp. Med. 194:781.
Botto, M., C. Dell’Agnola, A. E. Bygrave, E. M. Thompson, H. T. Cook, F. Petry, M. Loos, P. P. Pandolfi, M. J. Walport. 1998. Homozygous C1q deficiency causes glomerulonephritis associated with multiple apoptotic bodies. Nat. Genet. 19:56.
Vandivier, R. W., C. A. Ogden, V. A. Fadok, P. R. Hoffmann, K. K. Brown, M. Botto, M. J. Walport, J. H. Fisher, P. M. Henson, K. E. Greene. 2002. Role of surfactant proteins A, D, and C1q in the clearance of apoptotic cells in vivo and in vitro: calreticulin and CD91 as a common collectin receptor complex. J. Immunol. 169:3978.
Roos, A., W. Xu, G. Castellano, A. J. Nauta, P. Garred, M. R. Daha, C. van Kooten. 2004. Mini-review: a pivotal role for innate immunity in the clearance of apoptotic cells. Eur. J. Immunol. 34:921.
Gardai, S. J., Y. Q. Xiao, M. Dickinson, J. A. Nick, D. R. Voelker, K. E. Greene, P. M. Henson. 2003. By binding SIRP or calreticulin/CD91, lung collectins act as dual function surveillance molecules to suppress or enhance inflammation. Cell 115:13.
Martin, F., J. F. Kearney. 2001. B1 cells: similarities and differences with other B cell subsets. Curr. Opin. Immunol. 13:195.
Mohan, C., L. Morel, P. Yang, E. K. Wakeland. 1997. Genetic dissection of systemic lupus erythematosus pathogenesis: Sle2 on murine chromosome 4 leads to B cell hyperactivity. J. Immunol. 159:454.
Ito, T., S. Ishikawa, T. Sato, K. Akadegawa, H. Yurino, M. Kitabatake, S. Hontsu, T. Ezaki, H. Kimura, K. Matsushima. 2004. Defective B1 cell homing to the peritoneal cavity and preferential recruitment of B1 cells in the target organs in a murine model for systemic lupus erythematosus. J. Immunol. 172:3628.
Kim, S. J., D. Gershov, X. Ma, N. Brot, K. B. Elkon. 2002. I-PLA2 activation during apoptosis promotes the exposure of membrane lysophosphatidylcholine leading to binding by natural immunoglobulin M antibodies and complement activation. J. Exp. Med. 196:655.
Shaw, P. X., S. Horkko, M. K. Chang, L. K. Curtiss, W. Palinski, G. J. Silverman, J. L. Witztum. 2000. Natural antibodies with the T15 idiotype may act in atherosclerosis, apoptotic clearance, and protective immunity. J. Clin. Invest. 105:1731.
Lu, Q., G. Lemke. 2001. Homeostatic regulation of the immune system by receptor tyrosine kinases of the Tyro 3 family. Science 293:306.
Hanayama, R., M. Tanaka, K. Miyasaka, K. Aozasa, M. Koike, Y. Uchiyama, S. Nagata. 2004. Autoimmune disease and impaired uptake of apoptotic cells in MFG-E8-deficient mice. Science 304:1147.
Hanayama, R., M. Tanaka, K. Miwa, A. Shinohara, A. Iwamatsu, S. Nagata. 2002. Identification of a factor that links apoptotic cells to phagocytes. Nature 417:182.
Savill, J., I. Dransfield, C. Gregory, C. Haslett. 2002. A blast from the past: clearance of apoptotic cells regulates immune responses. Nat. Rev. Immunol. 2:965.
Garcia-Laorden, M. I., I. Rua-Figueroa, P. Perez-Aciego, J. C. Rodriguez-Perez, M. J. Citores, F. Alamo, C. Erausquin, C. Rodriguez-Gallego. 2003. Mannose binding lectin polymorphisms as a disease-modulating factor in women with systemic lupus erythematosus from Canary Islands, Spain. J. Rheumatol. 30:740.
Mitchell, D. A., M. C. Pickering, J. Warren, L. Fossati-Jimack, J. Cortes-Hernandez, H. T. Cook, M. Botto, M. J. Walport. 2002. C1q deficiency and autoimmunity: the effects of genetic background on disease expression. J. Immunol. 168:2538.(Lynda M. Stuart,, Kazue T)
Mannose-binding lectin (MBL) is a circulating serum protein that is sequestered to sites of inflammation and infection. MBL is a member of the collectin family with structural similarities to the lung collectins and functional similarities to C1q. Both MBL and C1q activate complement; C1q activates the classical pathway and MBL the lectin pathway. Here we demonstrate that MBL binds apoptotic cells in vitro and confirm a role for MBL in clearance of apoptotic cells in vivo. Despite MBL null mice demonstrating defective apoptotic cell clearance they did not develop spontaneous autoimmunity, lymphoproliferation, or germinal center expansion although increased numbers of peritoneal B1 cells were detected. These data demonstrate an important in vivo role for MBL in clearance of dying cells and adds the MBL null animals to the few animals with demonstrable in vivo apoptotic cell clearance defects. Moreover, it demonstrates that failure of apoptotic cell clearance can be dissociated from autoimmunity.
Introduction
Mannose-binding lectin (MBL)4 is a collectin as it displays an amino-terminal collagen stalk followed by an -helical coiled domain that orients the spatial organization of the carboxyl-terminal C-type lectin domains (1). Like the lung collectins, MBL contributes to first-line host defense and innate immune responses (2, 3). However, unlike surfactant protein A (SPA) and surfactant protein D (SPD), which are found predominantly in the lung, MBL circulates in serum as a multimer of trimers. The production of MBL in the liver is increased 2- to 3-fold in response to infection, and it is rapidly sequestered to sites of inflammation. One consequence of MBL-ligand interaction is the activation of complement via the lectin pathway. MBL forms a complex with three serine proteases termed mannose-binding associated proteases one of which (MASP2) co-opts the classical complement pathway convertase that in turn activates the complement cascade (4). In addition, the lectin pathway can also be activated by a family of related molecules, the ficolins.
MBL bears many functional and structural similarities to C1q; both molecules form a trimeric structure resembling a bunch of tulips, which further oligomerize to form higher order multimers and are able to activate complement. Importantly, C1q unlike MBL requires recognition of Ab-opsonized particles to activate complement. Defects in the first component of complement are strongly associated with autoimmunity such that >95% of patients deficient in C1q develop a systemic autoimmune syndrome (5). In addition, MBL haplotypes that specify low levels of an aberrant form of MBL unable to activate complement are also reported to be associated with a low but increased risk of autoimmunity (6, 7). Apoptotic cells have been shown to be a preferential source of many autoantigens (8) and failure to promptly remove dying cells has been implicated in the pathogenesis of systemic lupus erythematosus (5, 9, 10), a prototypic autoimmune disease. C1q has been shown to bind and opsonize dying cells (11, 12), and consequently, it has been suggested that this failure of C1q-mediated clearance of apoptotic cell renders dying cells immunogenic with subsequent development of autoimmunity. In further support of this "waste disposal" model, C1q-deficient mice, upon certain backgrounds, develop a lupus-like syndrome and nephritis characterized by multiple uncleared apoptotic cells (13). However, the mechanism by which dying cells come to incite an inflammatory or autoimmune response is poorly understood.
Recent data demonstrates that all of the collectins (MBL, SPA, and SPD) bind cells undergoing apoptosis in vitro (12). Furthermore, SPD null animals have defective in vivo clearance of apoptotic cells in the lung (14). These observations suggest that MBL could also have a role in apoptotic cell clearance in vivo and, along with the reports of increased risk of autoimmunity in MBL low individuals (6, 7), have raised the question of whether lack of MBL might result in autoimmunity as is observed with C1q deficiency (15). To study the role of MBL in vivo, we have generated mice deficient in both genes encoding mouse MBL (MBLA and MBLC) (3), and here we report the role of MBL in autoimmunity. We show that lack of MBL does indeed result in delayed removal of dying cells in vivo but, despite this defect, MBL null mice do not develop autoimmunity even when they are aged for 18 mo on a lupus-prone genetic background (129xC57BL/6). Taken together, these observations suggest that in MBL-deficient mice, failure of apoptotic cell clearance can be dissociated from autoimmunity.
Materials and Methods
Mice
MBL null mice were generated as described previously (3). Briefly, embryonic stem cells from 129 mice were targeted and injected into a C57BL/6 blastocyst and then fostered into a pseudo-pregnant female. Transmission from the chimeric progeny was determined by crossing with a C57BL/6 to generate MBLA or MBLC null heterozygous animals. MBLA/C–/– mice were generated and maintained thereafter by intercrossing on a mixed 129xC57BL/6 background in a specific pathogen-free facility. At 18 mo, mice were sacrificed, and the histology of all tissues was examined by the Charles River pathology service. Mice used for other experiments were between 8 and 12 wk of age. Age-, sex-, and strain-matched controls were maintained in the same facility. Thymocytes were isolated from 6-wk-old mice (The Jackson Laboratory).
Generation of apoptotic cells
Apoptotic cells were generated by allowing single cell suspensions of thymocytes to undergo constitutive apoptosis by ageing overnight in RPMI 1640 supplemented with antibiotics and 0.5% BSA. For most experiments, cells were stained using a green cell tracker dye (CMFDA) (Molecular Probes) before overnight culture.
Analysis of apoptotic cell—MBL binding and fluorescent microscopy
Apoptotic cells were stained for 15 min on ice with annexin V Alexa 488 (Molecular Probes) or Cy3 rhMBL (NatImmune) in high calcium buffer (2.5 mM CaCl2 in HEPES). Specificity of MBL binding was determined by comparing with an irrelevant Cy3-conjugated protein (BSA). For certain experiments, MBL binding was inhibited by 10 mM EDTA or 200 mM mannan. For microscopy, cells were washed once with high calcium buffer and examined immediately by fluorescent microscopy. Images were captured and analyzed using Openlab software (Improvision).
In vivo phagocytosis assay
Apoptotic cells were resuspended in PBS, and mice were injected i.p. with 10 x 106 cells/per mouse. Mice were sacrificed at 30 min and the peritoneum was lavaged. Phagocytosis was analyzed by FACS gating on F4/80-positive cells. Percentage of phagocytosing cells was determined. In addition, the "phagocytic index" was used to estimate the total amount of apoptotic cells removed by calculating (the mean fluorescence intensity of the phagocytosing cells) x (the percentage of cells that had phagocytosed).
Phenotyping of macrophages, dendritic cells (DCs), and B cells
Macrophages and B1 cells were isolated from the peritoneum and germinal center B cells, and DCs were isolated from spleen. For DCs, a single cell spleen preparation was generated and purified using CD11c MACS beads (Miltenyi Biotec). For FACS analysis, all Abs used were from BD Pharmingen except F4/80-allophycocyanin (Caltag Laboratories) and peanut agglutinin (Vector Laboratories).
Assessment of autoimmunity
Aged mice were assessed for proteinuria using urine dipsticks (Multistix). At 18 mo, mice were sacrificed and organs examined for histological evidence of inflammation or autoimmunity. Terminal bleed serum was examined for autoantibodies as follows: anti-dsDNA Abs were assayed in a 1/100 dilution using an anti-mouse dsDNA-specific ELISA kit (Alpha Diagnostics International); 2-fold serial dilutions from 1/80 to 1/2560 of serum anti-nuclear Abs (ANAs) were determined by binding to prepared Hep2 cell slides (Scimedx Corporation) and detected with FITC-labeled anti-mouse Ig. Slides were then scored by fluorescent microscopy. Images were captured at 5 and 10 exposures and scored positive if nuclear binding was detected at 5 exposure and intermediate if only seen at 10. Intermediate binding was repeated twice and deemed negative if only detected with 10 exposures on both occasions. Final ANA titer was the highest dilution that was positive.
Results
MBL clusters on apoptotic cells
Previous studies demonstrating MBL binding to dying cells often used MBL purified from serum, and it is possible that such preparations contain IgM, known to associate with MBL in circulation. IgM (particularly natural Abs) are also able to recognize apoptotic cells, raising the possibility that it is IgM and not MBL that interacts with dying cells in these experiments. We therefore set out to determine whether pure rMBL could bind only to apoptotic cells or whether rMBL also bound to live cells before exposure of phosphatidyl serine (PS). To do this, we directly conjugated Cy3 to rMBL and used this to define the interaction of MBL with dying cells after 0, 16, or 24 h in culture. As a negative control binding was compared with an irrelevant Cy3-conjugated protein (BSA). Surprisingly, even immediately after isolation, MBL binding could be detected, and cells could be separated into three categories: annexin–/rMBL–, rMBL+/annexin–, and annexin+/rMBL+ (Fig. 1) indicating that MBL can bind to live cells before the externalization of PS. This binding was specific and could be inhibited by 10 mM EDTA (Fig. 1) and 200 mM mannan (data not shown) suggesting that MBL was binding through the carbohydrate recognition domain. Binding was detected not only with thymocytes but also to a viable murine fibroblast cell line, L929 (data not shown).
FIGURE 1. MBL binds to live cells before the induction of apoptosis. Murine thymocytes were isolated, and PS exposure was determined by annexin V Alexa-488 and MBL binding by using purified rhMBL-Cy3 in high calcium buffer. A, Specificity of MBL binding was determined by comparison to control Cy3-BSA and in the presence of 10 mM EDTA. B, MBL binding was assessed both to the live cells (annexin V negative) and to the small population of contaminating apoptotic (annexin V positive) cells. Control Cy3-BSA binding is shown (left). C, Microscopic examination of MBL binding to cells before overnight culture. Arrow, A MBL and annexin V-negative cell; arrowhead, a MBL-Cy3 and annexin V-double- positive cell. Original magnification, x40.
To determine whether MBL binding altered during the process of apoptosis, binding was assessed at intermediate time points (Fig. 2). Using two-color FACS, MBL binding between live cells (determined either by forward and side scatter to sit in (R1) or as annexin– cells) and apoptotic cells (R2 or annexin+ cells) was comparable, and no difference in MBL binding was detected between the two populations. Furthermore, although the number of annexin+ cells increased with time indicating that the cells were entering into apoptosis, the majority of the cells remained intact as determined by the forward/side scatter, and no increase in MBL binding was seen. After 24 h, the majority of the cells were annexin+ (varying from culture to culture from 60 to 90%), and MBL remained bound to all cells. Importantly, when examined by fluorescent microscopy, the pattern of binding of MBL to apoptotic cells was patchy (Fig. 2) and differed from the diffuse pattern seen on the live cells (Fig. 1C). Taken together, these data demonstrate conclusively that MBL binds not only to dying cells but also to certain live cells and that it does so directly and independently of contaminating Ab. However, as death progresses, MBL adopts a distinct clustered pattern on the cell surface suggesting alteration in ligand distribution on the dying cell surface. This suggests that a clustered configuration of ligands on the surface of apoptotic cells may be required for MBL-mediated phagocytosis.
FIGURE 2. MBL clusters on cells during the process of apoptosis. Murine thymocytes were aged in serum-free conditions for 24 h. A, Representative two-color FACS analysis of MBL and annexin V binding (a 16-h time point is shown). B, FACS analysis of MBL binding to "live cells" (R1) and apoptotic cells (R2) at different time points after isolation. C, Microscopy demonstrating clustering of MBL on the surface of an apoptotic cell at 24 h.
MBL mediates clearance of apoptotic cells in vivo
SPD and C1q have been shown to increase phagocytosis of apoptotic cells by macrophages in both in vitro and in vivo systems. However, although MBL can mediate apoptotic cell engulfment in vitro, the contribution of MBL in vivo had not been defined. To determine whether MBL was able to mediate the removal of dying cells in vivo, we injected fluorescently labeled apoptotic thymocytes i.p. into either wild-type (WT) or MBL null mice. After 30 min the peritoneum was lavaged and macrophages were stained for F4/80. In a typical experiment the percentage of macrophages from MBL null mice that phagocytosed was 16% whereas WT macrophage phagocytosis was higher and typically greater than 25% of macrophage-ingested apoptotic cells (Fig. 3). This reduction equated to a 38% decrease in percentage of phagocytic cells in the absence of MBL. In addition, a more profound defect in the total number of apoptotic cells cleared (as shown by 49% reduction in the phagocytic index) was also demonstrated (Fig. 2). Taken together, these observations indicate an important role for MBL in clearance of apoptotic cells in vivo. Furthermore, because these mice are ficolin and C1q sufficient, our results indicate a non-redundant role for MBL in apoptotic cell clearance in vivo.
FIGURE 3. The role of MBL phagocytosis of apoptotic cells in vivo. Fluorescently labeled apoptotic fluorescent thymocytes were instilled into the peritoneum of WT and MBL null animals. Peritoneal lavage was collected at 30 min and macrophages were stained using F4/80-APC. A, Representative FACS density plot after 30 min of phagocytosis. Percentage of macrophages phagocytosing dying cells (B) and phagocytic index (C) were determined by FACS (mean ± SEM; n = 4 WT and n = 6 MBL null). Significance was determined by Student’s t test. *, p < 0.05; **, p < 0.01.
Failure of apoptotic cell clearance in MBL null mice is not associated with spontaneous activation of APCs
MBL might be predicted to alter macrophage phenotype for a number of reasons. First, collectins can directly modulate macrophage function (16), and second, delayed apoptotic cells clearance itself might result in accumulation of proinflammatory secondarily necrotic cells able to stimulate APCs. To exclude the possibility that impaired removal of apoptotic cell by macrophages in MBL null animals was a consequence of altered macrophage phenotype, we choose to determine whether lack of MBL might be associated with spontaneous activation of APCs in unchallenged mice. Our results revealed that peritoneal macrophages expressed similar amounts of costimulatory molecules and MHC in both WT and MBL null animals (Fig. 4). In addition, splenic DCs isolated from MBL null mice did not display enhanced levels of MHC II, CD40, CD80, or CD86 compared with cells isolated from WT mice (Fig. 4) indicating that lack of MBL does not result in spontaneous activation of APCs.
FIGURE 4. Phenotype of APC in WT vs MBL null mice. Splenic DCs and peritoneal macrophages isolated from resting MBL null and WT animals were phenotyped for activation markers CD40, CD80, CD86, and MHC II (light gray is unstained control).
MBL null mice have an increased number of peritoneal B1 cells
B1 cells are a subpopulation of B cells that populate the pleural and peritoneal cavities. B1 cells express a restricted number of germline-encoded B cell receptors and are the major source of natural Abs (17). These Abs demonstrate low affinity for certain bacteria and self-Ags. It has been suggested that B1 cells have a possible role in the development of autoimmunity because abnormalities in B1 cells have been reported in a number of autoimmune models (18, 19). The best defined B1 clone produces a T15 anti-phosphoryl choline Ab that recognizes not only pathogens but modified lipids and apoptotic cells (20, 21), and because of its receptor specificity, it has been suggested that this BCR may be selected on apoptotic cells encountered during B cell ontogeny. Thus, we hypothesized that the failure of apoptotic cell clearance seen in the peritoneum might be associated with a change in the number of B1 cells. To evaluate this question we analyzed peritoneal B cells by FACS, which revealed an expansion of peritoneal B1 cells as defined by B220int, CD11b+, and CD43+ (Fig. 5). These B cells expressed low or no CD5 and hence could be defined as B1-b cells.
FIGURE 5. Phenotype of B cells in WT vs MBL null mice. A, Peritoneal B cells were phenotyped, and percentages of B220int/high, CD11b+ B1 cells (R1) vs B220high, CD11b– B2 cells (R2) are shown. CD5 expression on B1 cells is shown for representative WT (light gray histogram) vs MBL null animal (dark gray histogram). B, Spleen CD95+, peanut agglutinin+ germinal center B cells (indicated in gated region) in WT vs MBL null animals (percentage of B220+ cells shown) (FACS density plots are shown for clarity).
MBL null mice have reduced germinal center B cells and no spontaneous lymphoproliferation
B1 cells normally produce low affinity IgMs or IgG3s but are able to differentiate to produce pathogenic autoreactive Abs if they exit the peritoneum and are activated in secondary lymphoid organs where they are educated in germinal centers on follicular DCs that have captured autoantigens (9). To determine whether the expanded B1 cell numbers would correlate with generalized lymphoproliferation, we phenotyped splenic B cells from unchallenged animals. Interestingly, germinal center B cell numbers were decreased by >60% in MBL null animals when compared with WT animals (Fig. 5). Taken together these results suggest that, although increased numbers of possibly autoreactive B1 cells might be associated with failed clearance of apoptotic cells, in the absence of MBL neither germinal center expansion nor lymphoproliferation occurred spontaneously.
MBL null mice dissociate failure of apoptotic cell clearance from autoimmunity and failed to develop high titer ANAs
Despite the lack of overt lymphoproliferation, the expansion of possibly autoreactive B1 cells in the peritoneum of MBL null mice suggested that these mice might also be predisposed to autoimmunity. Thus to determine whether MBL deficiency per se predisposes these mice to autoimmunity, we maintained MBL mice on autoimmune mixed 129/C57BL/6 background for 18 mo. Histological examination failed to demonstrate any significant autoimmune pathology in the organs of MBL null mice, although we did detect an acidophilic macrophage infiltrate in the lungs of older animals (L. M. Stuart, manuscript in preparation) (Table I). In particular, we failed to detect proteinuria or the appearance of free apoptotic cells in the glomeruli (Fig. 6) when compared with age-matched WT controls. dsDNA titers as assessed by ELISA were comparable between WT and MBL null animals at 18 mo and were significantly lower than those seen in the positive control MRL serum (Fig. 6). Similarly, although ANAs were detectable (particularly in the female animals; Fig. 6), no increase in experimental animals over WT was seen. In contrast, MBL deficiency appeared to provide protection from developing high titer ANAs (median dilution for WT 1/640-1280 vs MBL null 1/80–160). These observations indicate that failure to clear apoptotic cells is not sufficient to generate an autoimmune phenotype.
Table I. Histology of aged MBL null vs WT mice
FIGURE 6. Assessment of autoimmunity in MBL null animals. A, Representative renal histology of 18-mo-old WT and MBL null animals (n = 5 for both groups). B, ANA titers in serial diluted 18-mo serum and dsDNA titers (1/100 dilution) in male and female MBL null and WT animals (n = 5 or more for each group). Median is represented as horizontal red bar. Significance was determined by a Mann-Whitney U test. *, p < 0.025.
Discussion
Collectins are a family of soluble opsonins able to recognize a broad range of pathogens and have important and diverse roles in innate immune responses. In addition, MBL, through its ability to co-opt the classical convertase, is also able to activate complement and hence bears functional similarities to the first component of the classical pathway, C1q. Apoptotic cells are a preferential source of many of the auto-Ags found in systemic lupus erythematosus, and the importance of the prompt removal of dying cells in preventing this disease has been emphasized by a number of recent reports indicating that delayed apoptotic cell removal can precipitate autoimmunity (11, 22, 23). However, exactly how apoptotic cells are removed in vivo and what determines whether the uncleared cells come to incite a proinflammatory or autoimmune response is poorly understood. Both the collectin SPD and complement component C1q are known to mediate recognition and removal of apoptotic cells in vivo but the contribution of MBL in vivo has not been explored. Here, using mice deficient in both MBL genes (MBLA and MBLC), we demonstrate that MBL contributes to clearance of apoptotic cells in vivo. Furthermore, analysis of MBL null animals revealed abnormal B cell homeostasis suggesting an important immunoregulatory role for this molecule in vivo. However, despite the delay in removal of dying cells in this context, MBL-deficient mice do not develop autoimmunity despite being on an autoimmune prone (129xC57BL/6) background indicating surprisingly that failed apoptotic cell clearance can be dissociated from autoimmunity.
Although the absence of C1q results in delayed clearance of apoptotic cells that are introduced into the peritoneum, the dying cells are ultimately removed, indicating that alternative mechanisms are able to partially compensate for the C1q defect (11). All of these partially redundant recognition molecules have not been identified although recent work has suggested that both the Mer receptor tyrosine kinase (22) and MFGE8 (23, 24) might contribute in vivo (reviewed by Savill et al. (25)). Here we demonstrate that MBL, a molecule with both structural and functional similarities to C1q, also contributes to apoptotic cell removal in vivo. Clustering of collectins on the surface of apoptotic cells and the distinct pattern of binding they adopt is thought to be important for these opsonins to trigger phagocytosis (12, 14). Interestingly, not only do both C1q and MBL cluster on apoptotic cells as death progresses, but they also both contribute to clearance in vivo. MBL and C1q have been shown to recognize distinct and only partially overlapping surface ligands on dying cells (26), and it is likely, therefore, that the similarities in the in vivo phenotype of the C1qA–/– and MBL null animals is a consequence of their shared ability to trigger a common pathway or mechanism of uptake. This hypothesis is in keeping with the observed similarity in the extent of the defect in apoptotic cell clearance in the peritoneum of both C1q and MBL null animals. Furthermore, these data raise the intriguing possibility that these molecules work together, either as a macromolecular complex or to trigger a common engulfment pathway, rather than mediate distinct and independent pathways of apoptotic cell removal.
Delayed clearance apoptotic cells as seen in the MBL null animals might be predicted to modulate macrophage activation for two reasons. First, apoptotic cells that are not rapidly cleared in vivo are thought to undergo secondary necrosis and spill their potentially proinflammatory intracellular contents into surrounding tissues. Second, previous in vitro studies have demonstrated that phagocytosis of apoptotic cells actively inhibits APCs (27, 28, 29, 30). In addition, recent work has emphasized a role for the other collectins, SPA and SPD, as negative regulators of macrophages, a function mediated through their ability to ligate the inhibitory receptor SIRP- (16). However, surprisingly our data suggest that neither the absence of MBL itself nor the delayed removal of apoptotic cells is sufficient to result in spontaneous activation of APCs. Importantly it indicates that, in these circumstances, late apoptotic/necrotic cells are not proinflammatory and are ultimately cleared in a nonphlogistic manner. In support of this, our previous in vitro data suggests that removal of late apoptotic neutrophils can be mediated by a 2 integrin (complement receptor) independent mechanism, involving thrombospondin-1 and v3, which allows macrophages to ingest late apoptotic cells without eliciting inflammatory cytokine secretion (31).
Recent data have highlighted the importance of apoptotic cell clearance and regulation of germinal centers reaction in autoimmune disease; MFGE8 null animals demonstrate impaired apoptotic B cell clearance by tingible body macrophages, expanded germinal centers, and autoimmunity (23). Here, we have demonstrated that B cell homeostasis and germinal centers are affected by the absence of MBL, indicating an important immunoregulatory role for this molecule. Although we do not know the exact mechanism for the expansion of B1 cells in MBL null animals we would suggest that it is either in response to delayed apoptotic cell clearance or that MBL has a negative homeostatic role on B1 cell development. Intriguingly, in support of the former hypothesis, mice deficient in soluble IgM also demonstrate impaired clearance of dying cells, expanded B1 numbers, and accelerated autoimmunity (32). Our observation that the numbers of germinal center B cells, both resting and after immunization (L. M. Stuart and K. Takahshi, unpublished data), are lower in MBL null mice and that the autoantibodies made in the MBL null mice are neither associated with pathology nor are they made at high titer is intriguing. One speculation is that the absence of MBL does not support FDC-B interactions and that germinal center formation may be impaired in these animals. It is possible, therefore, that although MBL deficiency is associated with abnormal disposal of apoptotic cell-derived self-Ags, the absence of MBL is not permissive for expansion of truly pathogenic self-reactive B cells, protecting these mice from developing systemic lupus erythematosus. In this regard, it will be important to determine what role MBL plays in germinal center reactions (33) and whether this might underlie the different outcomes of failed apoptotic cell clearance in C1q and sIgM null animals.
The data we have presented indicates that both MBL and C1q null mice display a similar phenotype in the defect in apoptotic cell clearance, but they diverge in susceptibility to autoimmunity. An important difference between C1q and MBL is the ability of C1q to trigger complement deposition on the surface of the apoptotic cell (26) and to recognize Ab-opsonized apoptotic cells and immune complexes. It is possible that it is specifically failed clearance of Ab opsonized apoptotic cells, and associated proinflammatory responses, that renders apoptotic cells more immunogenic in the context of C1q deficiency. Of note, unlike C1q-deficient patients, MBL-deficient humans show only a weak increased risk for development of autoimmunity suggesting that these mice mimic human diseases (34). Furthermore, in support of this animal model, a recent report has suggested a protective role of MBL deficiency in female lupus patients in the Canary Islands who, in its absence, produce lower autoantibody levels and have a later onset of disease (35). C1qA–/– animals develop spontaneous nephritis and autoimmunity only on the mixed 129xC57BL/6 or MRL/Mp background (36), and in this regard, it will be of interest to determine whether lack of MBL can modulate other autoimmune strains such as the MRL/Mp or MRL/lpr where different genetic factors contribute.
The observed similarities and differences between the C1q- and MBL-deficient animals leave many unanswered questions about the role of these opsonins in control of autoimmunity and inflammation. It is curious that MBL and C1q have many broad similarities in that they both co-opt the classical pathway convertase to initiate the complement cascade and, pertinent to this study, appear to have a similar role in apoptotic cell clearance in vivo. However, this study demonstrates that the two molecules have non-redundant functions. In the absence of C1q, apoptotic cell clearance is not blocked completely but delayed, occurring with time by a compensatory mechanism. Here, we have shown that one such mechanism is MBL, and it is possible that this alone might be sufficient to alter the balance in favor of a proinflammatory outcome in the absence of C1q. Perhaps it is increased MBL-mediated clearance in the C1q-deficient animals that is deleterious and that its absence actively protects against end organ damage. This hypothesis is particularly pertinent in considering the kidney, where MBL is highly expressed and is likely therefore to be a significant contributor regulating response to, and removal of, uncleared apoptotic cells in the glomeruli of C1q null animals. We suggest that further definition and careful comparison of the fate of dying cells and autoreactive B cell clones in the C1qA–/– vs MBL null animals will be vital in determining why one predisposes to lupus whereas the other does not. We believe that, in the future, such comparative studies will lead to new insights into the genesis of the autoimmune syndromes associated with failed apoptotic removal.
Disclosures
The authors have no financial conflict of interest.
Acknowledgments
We thank Drs. Marina Botto and Adam Lacy-Hulbert for critical reading of the manuscript and helpful comments.
Footnotes
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
1 This work was supported by a Wellcome Trust Clinician Scientist Grant R36731 (to L.M.S.), Wellcome Trust Programme Grant R068089 (to J.S.), and National Institutes of Health Grants R01 AI42788 and P01 AI52343–03 (to R.A.B.E.).
2 L.M.S. and K.T. contributed equally to this study.
3 Address correspondence and reprint requests to Dr. Lynda M. Stuart, Laboratory of Developmental Immunology, Massachusetts General Hospital, Jackson 14, 55 Fruit Street, Boston, MA 02114. E-mail address: lstuart{at}partners.org
4 Abbreviations used in this paper: MBL, mannose-binding lectin; PS, phosphatidyl serine; ANA, anti-nuclear Ab; DC, dendritic cell; WT, wild type; SPA, surfactant protein A; SPD, surfactant protein D.
Received for publication October 8, 2004. Accepted for publication December 27, 2004.
References
Sheriff, S., C. Y. Chang, R. A. Ezekowitz. 1994. Human mannose-binding protein carbohydrate recognition domain trimerizes through a triple -helical coiled-coil. Nat. Struct. Biol. 1:789.
Epstein, J., Q. Eichbaum, S. Sheriff, R. A. B. Ezekowitz. 1996. The collectins in innate immunity. Curr. Opin. Immunol. 8:29.
Shi, L., K. Takahashi, J. Dundee, S. Shahroor-Karni, S. Thiel, J. C. Jensenius, F. Gad, M. R. Hamblin, K. N. Sastry, R. A. Ezekowitz. 2004. Mannose-binding lectin-deficient mice are susceptible to infection with Staphylococcus aureus. J. Exp. Med. 199:1379.
Matsushita, M., Y. Endo, M. Nonaka, and T. Fujita. 1998. Complement-related serine proteases in tunicates and vertebrates. Curr. Opin. Immunol.
Pickering, M. C., M. Botto, P. R. Taylor, P. J. Lachmann, M. J. Walport. 2000. Systemic lupus erythematosus, complement deficiency, and apoptosis. Adv. Immunol. 76:227.
Ip, W. K., S. Y. Chan, C. S. Lau, Y. L. Lau. 1998. Association of systemic lupus erythematosus with promoter polymorphisms of the mannose-binding lectin gene. Arthritis Rheum. 41:1663.
Davies, E. J., N. Snowden, M. C. Hillarby, D. Carthy, D. M. Grennan, W. Thomson, W. E. Ollier. 1995. Mannose-binding protein gene polymorphism in systemic lupus erythematosus. Arthritis Rheum. 38:110.
Casciola-Rosen, L. A., G. Anhalt, A. Rosen. 1994. Autoantigens targeted in systemic lupus erythematosus are clustered in two populations of surface
Herrmann, M., O. M. Zoller, M. Hagenhofer, R. Voll, J. R. Kalden. 1996. What triggers anti-dsDNA antibodies?. Mol. Biol. Rep. 23:265.
Mevorach, D., J. L. Zhou, X. Song, K. B. Elkon. 1998. Systemic exposure to irradiated apoptotic cells induces autoantibody production. J. Exp. Med. 188:387.
Taylor, P. R., A. Carugati, V. A. Fadok, H. T. Cook, M. Andrews, M. C. Carroll, J. S. Savill, P. M. Henson, M. Botto, M. J. Walport. 2000. A hierarchical role for classical pathway complement proteins in the clearance of apoptotic cells in vivo. J. Exp. Med. 192:359.
Ogden, C. A., A. deCathelineau, P. R. Hoffmann, D. Bratton, B. Ghebrehiwet, V. A. Fadok, P. M. Henson. 2001. C1q and mannose binding lectin engagement of cell surface calreticulin and CD91 initiates macropinocytosis and uptake of apoptotic cells. J. Exp. Med. 194:781.
Botto, M., C. Dell’Agnola, A. E. Bygrave, E. M. Thompson, H. T. Cook, F. Petry, M. Loos, P. P. Pandolfi, M. J. Walport. 1998. Homozygous C1q deficiency causes glomerulonephritis associated with multiple apoptotic bodies. Nat. Genet. 19:56.
Vandivier, R. W., C. A. Ogden, V. A. Fadok, P. R. Hoffmann, K. K. Brown, M. Botto, M. J. Walport, J. H. Fisher, P. M. Henson, K. E. Greene. 2002. Role of surfactant proteins A, D, and C1q in the clearance of apoptotic cells in vivo and in vitro: calreticulin and CD91 as a common collectin receptor complex. J. Immunol. 169:3978.
Roos, A., W. Xu, G. Castellano, A. J. Nauta, P. Garred, M. R. Daha, C. van Kooten. 2004. Mini-review: a pivotal role for innate immunity in the clearance of apoptotic cells. Eur. J. Immunol. 34:921.
Gardai, S. J., Y. Q. Xiao, M. Dickinson, J. A. Nick, D. R. Voelker, K. E. Greene, P. M. Henson. 2003. By binding SIRP or calreticulin/CD91, lung collectins act as dual function surveillance molecules to suppress or enhance inflammation. Cell 115:13.
Martin, F., J. F. Kearney. 2001. B1 cells: similarities and differences with other B cell subsets. Curr. Opin. Immunol. 13:195.
Mohan, C., L. Morel, P. Yang, E. K. Wakeland. 1997. Genetic dissection of systemic lupus erythematosus pathogenesis: Sle2 on murine chromosome 4 leads to B cell hyperactivity. J. Immunol. 159:454.
Ito, T., S. Ishikawa, T. Sato, K. Akadegawa, H. Yurino, M. Kitabatake, S. Hontsu, T. Ezaki, H. Kimura, K. Matsushima. 2004. Defective B1 cell homing to the peritoneal cavity and preferential recruitment of B1 cells in the target organs in a murine model for systemic lupus erythematosus. J. Immunol. 172:3628.
Kim, S. J., D. Gershov, X. Ma, N. Brot, K. B. Elkon. 2002. I-PLA2 activation during apoptosis promotes the exposure of membrane lysophosphatidylcholine leading to binding by natural immunoglobulin M antibodies and complement activation. J. Exp. Med. 196:655.
Shaw, P. X., S. Horkko, M. K. Chang, L. K. Curtiss, W. Palinski, G. J. Silverman, J. L. Witztum. 2000. Natural antibodies with the T15 idiotype may act in atherosclerosis, apoptotic clearance, and protective immunity. J. Clin. Invest. 105:1731.
Lu, Q., G. Lemke. 2001. Homeostatic regulation of the immune system by receptor tyrosine kinases of the Tyro 3 family. Science 293:306.
Hanayama, R., M. Tanaka, K. Miyasaka, K. Aozasa, M. Koike, Y. Uchiyama, S. Nagata. 2004. Autoimmune disease and impaired uptake of apoptotic cells in MFG-E8-deficient mice. Science 304:1147.
Hanayama, R., M. Tanaka, K. Miwa, A. Shinohara, A. Iwamatsu, S. Nagata. 2002. Identification of a factor that links apoptotic cells to phagocytes. Nature 417:182.
Savill, J., I. Dransfield, C. Gregory, C. Haslett. 2002. A blast from the past: clearance of apoptotic cells regulates immune responses. Nat. Rev. Immunol. 2:965.
Garcia-Laorden, M. I., I. Rua-Figueroa, P. Perez-Aciego, J. C. Rodriguez-Perez, M. J. Citores, F. Alamo, C. Erausquin, C. Rodriguez-Gallego. 2003. Mannose binding lectin polymorphisms as a disease-modulating factor in women with systemic lupus erythematosus from Canary Islands, Spain. J. Rheumatol. 30:740.
Mitchell, D. A., M. C. Pickering, J. Warren, L. Fossati-Jimack, J. Cortes-Hernandez, H. T. Cook, M. Botto, M. J. Walport. 2002. C1q deficiency and autoimmunity: the effects of genetic background on disease expression. J. Immunol. 168:2538.(Lynda M. Stuart,, Kazue T)