Complement Protein C1q Enhances Macrophage Foam Cell Survival and Efferocytosis

Macrophage foam cells are key contributors to the progression of atherosclerosis, a chronic inflammatory disease and leading cause of death in the United States and around the world (1). In atherosclerosis, the accumulation of low-density lipoproteins (LDLs) in the subendothelium of arteries leads to monocyte recruitment via chemotaxis, and M-CSF promotes their differentiation into monocyte-derived macrophages. Localized inflammation leads to oxidative damage and modifications to the LDL (2). These modified forms of LDL, such as oxidized LDL (oxLDL) and acetylated LDL (acLDL), are recognized by macrophage scavenger receptors, so-called pattern recognition receptors capable of recognizing damage-associated molecular patterns. Ingested modified forms of LDL are hydrolyzed in the late endosome to free cholesterol (FC) and fatty acids (3). Under normal conditions, FC is removed from the cell via reverse cholesterol transport. Cholesterol efflux is mediated by transporter proteins ABCA1 and ABCG1, which load the FC onto apolipoprotein A1 and high-density lipoprotein, respectively. However, in hyperlipidemic conditions, this process can become overwhelmed, and FC may build up in the cell or be re-esterified and stored as lipid droplets. These lipid droplets give rise to the name macrophage foam cell because of their foamy appearance under a microscope.

Lipid-laden foam cells exhibit increased proinflammatory responses in in vitro and in vivo studies (4–6), and markers of inflammation are associated with human plaques (7). Cholesterol crystal formation in macrophage foam cells was shown to activate the NLRP3 inflammasome, leading to IL-1β proinflammatory cytokine signaling (8, 9). Proinflammatory M1-polarized macrophages are detrimental in late-stage atherosclerosis and may have a role in driving plaque rupture (10, 11). Macrophage foam cell production of matrix metalloproteinases and other proteases also was implicated in the rupture of plaque through degradation of lesions (reviewed in Ref. 12). In addition to inflammation, FC accumulation and exposure to oxysterols were shown to contribute to endoplasmic reticulum (ER) stress in macrophages (13). Excess cholesterol activation of the unfolded protein response during ER stress results in foam cell apoptosis (14). An accumulation of apoptotic cells is associated with advanced stages of plaque development in human plaque tissue (15). In vitro studies in murine macrophage foam cells formed via ingestion of oxLDL demonstrated that ingestion of modified lipoproteins led to activation of apoptosis effector molecule caspase-3, as well as TUNEL⁺ DNA fragmentation (16). Interestingly, the role of apoptosis in atherosclerosis is influenced by the stage of disease. In murine studies, early lesional apoptosis was associated with smaller lesion development, whereas reductions in apoptosis were associated with larger lesions (17–19). However, protection from ER stress and the resulting apoptosis in advanced lesional macrophages were shown to be protective against plaque necrosis (20–23). These contrasting results are likely due to changes in efferocytosis, the removal of apoptotic cells. In early atherosclerotic lesions, efferocytosis efficiently clears apoptotic cells and cellular debris. Larger numbers of apoptotic macrophages are associated with advanced lesions, suggesting inefficient or inadequate removal of apoptotic macrophages that can lead to plaque necrosis and rupture (reviewed in Ref. 24). Thus, mechanisms of cell death and timely apoptotic cell removal strongly influence the progression of atherosclerosis.

C1q is a pattern recognition receptor of the innate immune complement system that binds modified lipoproteins, but not unmodified LDL, and directly enhances their uptake by macrophages (25, 26). C1q plays a dual role in atherosclerosis. In later stages, the infiltration of additional complement components, such as C1r and C1s, which are found circulating in blood plasma, contributes to formation and activation of the C1 complex with C1q. The C1 complex (C1qC1r₂C1s₂) activates the classical pathway of the proinflammatory complement cascade and is a major contributor to disease progression (reviewed in Ref. 27). However, recent studies highlighted a beneficial role for C1q in early atherosclerosis, which is likely related to an emerging role for C1q that is not associated with complement. Macrophages secrete detectable levels of C1q when cultured and are likely a major source of C1q production in vivo (28). Therefore, physiologically relevant levels of C1q are likely present in these macrophage-rich lesions in the absence of complement C1 complex proteins C1r and C1s. Ingestion of modified lipoproteins opsonized with C1q increases macrophage cholesterol efflux and decreases foam cell formation. In addition, proinflammatory cytokine signaling is reduced in macrophages that ingest modified lipoproteins in the presence of C1q (25, 29–31). Studies in the LDLR^−/− murine model of atherosclerosis also identified a protective role for C1q in early stages of disease, with C1q-sufficient animals exhibiting a reduced aortic root lesion size compared with C1q-deficient animals, in which an accumulation of apoptotic cells was observed (32). C1q has a well-described and physiologically important role in the clearance of apoptotic cells (efferocytosis) and prevention of autoimmunity (33–35).

Because C1q was shown to have a direct effect on macrophage foam cell formation and inflammatory polarization, we hypothesized that C1q opsonization of modified lipoproteins also modulates macrophage foam cell survival and foam cell function. This study aimed to investigate markers of survival and apoptosis in macrophages during ingestion of modified lipoproteins, as well as to measure efferocytosis in foam cells in the presence or absence of C1q.

C1q is routinely isolated from plasma-derived normal human serum and validated for purity and activity as previously described (29). Preparations of C1q used in this study were determined to be free of endotoxin (<10 pg/ml), as assayed by an LAL Chromogenic Endotoxin Quantitation Kit (Pierce, Waltham, MA). Sterile filtered human lipoproteins (LDL, acLDL, and medium oxLDL) were purchased from Kalen Biomedical (Montgomery Village, MD). Human monocytes were isolated from healthy human blood samples, which were collected according to the guidelines and approval of the California University Long Beach Institutional Review Board, by countercurrent flow elutriation, as described (36, 37). Cell purity was assessed using the Scepter Cell Counter (EMD Millipore, Darmstadt, Germany), and monocyte populations used were >94% pure. Isolated monocytes were stimulated to differentiate into human monocyte–derived macrophages (HMDMs) by culture for 8 d in RPMI 1640, 10% FCS, 2 mM l-glutamine, and 1% penicillin/streptomycin containing 25 ng/ml recombinant human M-CSF (PeproTech, Rocky Hill, NJ). All cell culture reagents, unless otherwise stated, were purchased from Invitrogen (Carlsbad, CA). Pooled HMDMs isolated from multiple donors (5–10) were used to reduce donor variability. Raw264.7 cells (American Type Culture Collection), a murine macrophage cell line, were cultured as described (38). Macrophage foam cells were generated by incubating cells with 10 μg protein per milliliter of oxLDL or acLDL for 24 h at 37°C in 5% CO₂, as described in detail for individual assays and previously (25).

C1q modulation of apoptosis in HMDM foam cells was previously identified by Kyoto Encyclopedia of Genes and Genomes and Gene Ontology (GO) pathway enrichment analysis of RNA-sequencing library data (30, 31). An apoptosis gene network was generated with GeneMANIA (http://www.genemania.org/) (39). Network interactions were limited to pathway and genetic interactions. Cytoscape (40) was used to visualize genes within networks of interest that were significantly modulated by C1q in this library (p < 0.05) (30, 31). Significantly downregulated genes were visualized by color coding, with color intensity corresponding to the extent of modulation.

Macrophages (HMDMs or Raw264.7) were harvested and plated in 24-well tissue culture–treated plates at 5 × 10⁵ cells per well in X-VIVO 15 (Lonza, Walkersville, MD) serum-free defined media. Cells were treated with 10 μg protein per milliliter of oxLDL or acLDL in the absence or presence of 75 μg/ml C1q and incubated for 1–24 h at 37°C in 5% CO₂. RNA was isolated from cells using an RNeasy Mini Kit (QIAGEN, Hilden, Germany). A Moloney murine leukemia virus reverse transcriptase kit (Life Technologies) was used to synthesize cDNA from 100 ng of total RNA, according to the manufacturer’s instructions. Quantitative real-time PCR (StepOne; Applied Biosystems) was performed using TaqMan Gene Expression Assay probes (Life Technologies). Gene expression was normalized to endogenous control GAPDH, and mRNA levels were expressed as fold changes compared with untreated macrophages. The 2^−ΔΔCt method was used to determine the fold change, as described (33).

Levels of active caspase-3 and cleaved (active) PARP-1 were measured in lysates from HMDMs that were fed 10 μg protein per milliliter of oxLDL or acLDL in the absence or presence of 75 μg/ml C1q for 1–24 h at 37°C in 5% CO₂. The MILLIPLEX MAP Cell Signaling Buffer and Detection Kit for Magnetic Beads was used to prepare lysates, and levels of active caspase-3 and PARP-1 were measured using the MILLIPLEX MAP Human Late Apoptosis Magnetic Bead 3-Plex Kit on a Luminex MAGPIX System (all from EMD Millipore). The measured mean fluorescent intensity levels of caspase-3 and PARP-1 were normalized to that of GAPDH within each sample to account for any variations in cell number within wells.

Raw264.7 cells were plated in a 16-well glass Chamber Slide System (Lab-Tek) at 1 × 10⁵ cells per well in complete media (Dulbecco's Modified Eagle Medium supplemented with 10% FCS, 1% penicillin/streptomycin) and incubated overnight at 37°C in 5% CO₂ for cell adhesion. Media were replaced with X-VIVO 15 serum-free defined media, and cells were treated with 10–100 μM etoposide (control apoptosis inducer) or with 10–100 μg protein/ml oxLDL or acLDL in the presence or absence of 75 μg/ml C1q for 24 h at 37°C in 5% CO₂. Media were replaced with 125 μl of a staining solution consisting of two drops each of NucBlue Live ReadyProbes Reagent Stain and CellEvent Caspase-3/7 Green Detection Reagent (Life Technologies) per 1 ml of Opti-Klear Live Cell Imaging Buffer (Marker Gene Technologies, Eugene, OR). Slides were incubated for an additional 30 min and viewed via fluorescent microscopy (Invitrogen EVOS FL Auto Cell Imaging System, Life Technologies) using GFP and DAPI filters. Images were automatically taken in ≥10 random fields within each well. The number of caspase-3/7⁺ cells per field was counted manually and normalized to the number of NucBlue-stained nuclei.

Raw264.7 macrophages were plated in tissue culture–treated 96-well plates at 2.5 × 10⁴ cells per well in X-VIVO 15 serum-free defined media. Cells were incubated for 24 h at 37°C in 5% CO₂ to allow cell attachment. Cells were treated with 10 μg protein per milliliter of oxLDL or acLDL in the absence or presence of 75 μg/ml C1q and incubated for 24 h at 37°C in 5% CO₂. Four hours prior to harvesting (at 20 h posttreatment), 10% v/v alamarBlue reagent was added per well (Thermo Fisher Scientific, Sunnyvale, CA). Fluorescence was measured at an excitation of 570 nm and emission of 630 nm using a Synergy H1 Hybrid plate reader (BioTek, Winooski, VT). Negative control (background fluorescence of cells + alamarBlue reagent without 4 h incubation) and positive control (total 100% reduced sample, prepared by autoclaving alamarBlue reagent in X-VIVO 15 media) values were used to calculate the percentage of alamarBlue reduction in each sample as follows, according to the manufacturer’s instructions: (sample fluorescence – background)/(total reduction – background) × 100.

Raw264.7 macrophages were plated in tissue cultured–treated 24-well plates at 0.5 × 10⁶ cells per well in X-VIVO 15 serum-free defined media. Cells were treated with 10 μg protein per milliliter of oxLDL or acLDL in the absence or presence of 75 μg/ml C1q and were incubated for 24 h at 37°C in 5% CO₂. Foam cells were harvested using Cellstripper (Corning, Manassas, VA) and resuspended in 100 μl of PBS/1% BSA. One microliter of 100 μg/ml propidium iodide (PI; Thermo Fisher Scientific) working stock staining solution was added to the cell suspension for 15 min at room temperature. Fluorescence was measured by flow cytometry using the SH800 Cell Sorter (Sony Biotechnology). A total of 10,000 events was acquired for each sample. The percentage of PI⁺ cells in the total population was quantified using FlowJo software (TreeStar, Ashland, OR). Unstained untreated cells were used to determine the PI⁻ gate, with <0.1% of these negative controls falling within the PI⁺ gate. Data are expressed as change in the percentage of PI⁺ cells compared with untreated macrophages.

Raw264.7 macrophages were plated in eight-well glass chambered slides at 2.5 × 10⁵ cells per milliliter in X-VIVO 15 serum-free defined media and were incubated overnight at 37°C in 5% CO₂ to allow cells to attach. Then cells were treated with 10 μg protein per milliliter of oxLDL or acLDL in the absence or presence of 75 μg/ml C1q for 24 h at 37°C in 5% CO₂. Sheep erythrocytes, suboptimally opsonized with IgG, as described previously (41), were used as phagocytosis targets. A total of 10⁷ targets in 100 μl was added to each well and incubated for 30 min at 37°C in 5% CO₂. Uningested targets were lysed with ACK Lysing Buffer (Thermo Fisher Scientific), and cells were fixed in 1% glutaraldehyde in PBS. Cells were stained with Giemsa (Sigma-Aldrich) and counted by light microscopy. The average percentage of cells that had ingested at least one target (percentage phagocytosis) and the average number of ingested targets per 100 cells (phagocytic index) were quantified in ≥200 cells per well.

Raw264.7 macrophages were plated at 1 × 10⁶ cells per milliliter in X-VIVO 15 serum-free defined media in tissue culture–treated 12-well plates and incubated for 24 h at 37°C in 5% CO₂ to allow cells to attach. Then cells were treated with 10 μg protein per milliliter of oxLDL in the absence or presence of 75 μg/ml C1q for an additional 24 h at 37°C in 5% CO₂ to generate foam cells. Additional Raw264.7 cells were labeled with CFSE, as previously described (34), before induction of apoptosis by irradiation at 120,000 μJ/cm² for 30 s. CFSE-labeled apoptotic macrophages were incubated overnight at 37°C in 5% CO₂ prior to use. Foam cells were stained with CellTrace Violet (Molecular Probes) in their wells, according to the manufacturer’s instructions. CFSE-labeled apoptotic macrophages were added to each well at a 3:1 ratio of apoptotic cells/foam cells for 1 h at 37°C. Foam cells were washed twice with PBS to assist in the removal of uningested apoptotic cells and were harvested from wells using Cellstripper, as previously described (29). Cells were resuspended in 300 μl of PBS/1% BSA, and uptake was determined by flow cytometry using a Sony SH800 Cell Sorter. Unstained Raw264.7 cells were used to define the CFSE⁻/CellTrace Violet⁻ quadrant. Single-stained CellTrace Violet–untreated Raw264.7 cells and CFSE-apoptotic Raw264.7 cells were used to compensate for any spillover between fluorophores. The percentage of CellTrace Violet⁺/CFSE⁺ cells was quantified using FlowJo software.

Numerical data represent the mean of at least three independent experiments in which each condition was tested at least in duplicate, unless otherwise stated. ANOVA was performed to account for different sources of variation, treatment, and experiments with a post hoc Tukey or Bonferroni multiple-comparison test, where appropriate. Statistical analyses were performed using GraphPad Prism 5.

Our previous unbiased RNA-sequencing screen of C1q-modulated pathways in macrophages during ingestion of modified atherogenic lipoproteins highlighted a role for C1q in modulating cell death and apoptosis by GO and Kyoto Encyclopedia of Genes and Genomes pathway enrichment analysis (30, 31). Network pathway analysis demonstrated that C1q downregulates a number of genes involved in apoptosis during ingestion of modified lipoproteins (Fig. 1). C1q regulation of CASP7, CASP8, CASP10, BCL2L11, and TNFSF10 in HMDMs during clearance of oxLDL and acLDL was validated by quantitative real-time PCR (Fig. 2A). For most genes tested, ingestion of modified lipoproteins by HMDMs increased apoptosis-related gene expression. However, when lipoproteins were ingested in the presence of C1q, levels were consistently downregulated, including certain conditions in which levels were reduced to below basal levels in untreated HMDMs (CASP8, BCL2L11, TNFSF10). Interestingly, although CASP3 was not significantly modulated in our original RNA-sequencing screen, quantitative real-time PCR was able to detect downregulation by C1q during ingestion of acLDL. To determine whether C1q triggered similar responses in mouse macrophages and to identify whether there were any kinetic differences in regulation, gene expression of apoptosis executioner CASP7 was also measured in Raw264.7 macrophages under similar conditions (Fig. 2B). In Raw264.7 macrophages, oxLDL increased CASP7 gene expression above levels in untreated macrophages at early time points (1, 3 h) that was significantly downregulated by C1q, similar to the kinetics and levels observed in HMDMs. This suggests that Raw264.7 macrophages may provide an acceptable model to investigate the role of C1q in modulation of apoptosis.

To investigate whether the downregulation of apoptosis-related genes by C1q was seen at the protein and functional levels, Luminex multiplex analysis on lysates from HMDM during foam cell formation (ingestion of atherogenic lipoproteins) was performed. Levels of the active form of caspase-3 and the cleaved (active) form of PARP-1 were measured by Milliplex MAPmate assay (Fig. 3). From 3 h, ingestion of oxLDL and acLDL enhanced levels of active caspase-3 and PARP-1 compared with untreated HMDM. In contrast, when C1q was bound to the modified lipoproteins, these levels were significantly suppressed to basal levels.

To directly measure caspase-3/7 activity, a fluorogenic substrate for activated caspase-3/7 was added to Raw264.7 macrophage foam cells that had ingested modified lipoproteins in the presence or absence of C1q for 24 h and then was detected by fluorescence microscopy (Fig. 4A). Raw264.7 macrophages that had ingested modified lipoproteins displayed dose-dependent increases in caspase-3/7 activity (Fig. 4B). However, the presence of C1q during uptake of lipoproteins suppressed caspase-3/7 activity to levels similar to those in untreated macrophages.

To determine whether the C1q modulation of caspase-3/7 activation pathways provided a survival advantage to macrophage foam cells, a cell viability assay was performed (Fig. 5A). alamarBlue redox indicator dye, which yields a fluorescent signal in response to metabolic activity, was provided to Raw264.7 cells during ingestion of modified lipoproteins for 24 h. Foam cells that ingested modified lipoproteins alone had lower metabolic activity (cell viability) than did untreated macrophages. However, the presence of C1q significantly enhanced cell viability (alamarBlue reduction) to levels significantly higher than in untreated macrophages. Conversely, to determine whether the C1q modulation of caspase-3/7–activation pathways altered apoptosis in macrophage foam cells, nuclear stain PI was used (Fig. 5B). Because PI is not permeant to live cells, cells that are positive for PI staining are considered apoptotic. Foam cells that ingested modified lipoproteins alone had a higher percentage of PI⁺ cells than did untreated macrophages, as measured by flow cytometry. However, the presence of C1q significantly reduced the percentage of PI⁺ cells (from 18 ± 1% for untreated to 11 ± 3% for oxLDL+C1q treated) to a level even below that of untreated macrophages (Fig. 5B), which have a low, but measurable, basal level of cell death in culture.

To determine whether the C1q modulation of survival leads to modulation of macrophage foam cell functions, the ability of C1q-treated foam cells to perform phagocytosis and efferocytosis was measured. Raw264.7 cells were incubated for 24 h with oxLDL or acLDL to induce foam cell formation in the presence or absence of C1q. To investigate immune complex clearance, phagocytosis assays were performed. Sheep erythrocytes suboptimally opsonized with IgG were added to macrophage foam cells (42) (Fig. 6A, 6B). After 30 min of incubation, uningested targets were lysed, prior to fixing and staining macrophages for counting ingested targets by light microscopy. The average basal percentage of macrophages undergoing phagocytosis of at least one target was 46% (±7%, data not shown). Macrophage foam cells formed by ingestion of oxLDL, but not acLDL, showed reduced phagocytic capabilities compared with untreated macrophages (Fig. 6A, 6B). However, the presence of C1q during oxLDL uptake significantly restored the percentage of phagocytic cells back to basal levels. The phagocytic index, or average number of targets ingested by untreated macrophages, was 3.1 (±1.1, data not shown). Macrophage foam cells ingested fewer targets than did untreated macrophages; this was again restored in the presence of C1q to around (acLDL) or above (oxLDL) basal levels. Efferocytosis assays were performed to investigate apoptotic cell clearance (Fig. 6C, 6D). CFSE-labeled apoptotic macrophages were fed to CellTrace Violet–labeled foam cells (formed by ingestion of oxLDL or acLDL in the presence or absence of C1q), and ingestion of apoptotic cells was quantified by flow cytometry. Similar to the phagocytosis assay, macrophage foam cells had a slightly lower ability to clear apoptotic cells compared with untreated macrophages. However, efferocytosis was significantly increased, even above basal levels, in foam cells formed in the presence of C1q.

C1q plays an important dual role in the innate immune response. Activation of the inflammatory complement cascade by C1q is an integral defense mechanism against infectious agents. However, it also has an important physiological role in the direct opsonization of damaged-self targets, such as apoptotic cells, or modified lipoproteins, which helps to prevent autoimmunity by increasing macrophage clearance and anti-inflammatory polarization. In this article, we present data suggesting that C1q plays an additional protective role in maintaining normal tissue homeostasis by enhancing macrophage foam cell survival and function during the clearance of modified lipoproteins.

In an unbiased examination of C1q modulation of gene expression profiles in HMDMs ingesting modified lipoproteins, GO analysis highlighted a role for C1q in modulation of cell death and apoptosis pathways (30, 31). A network pathway analysis in Cytoscape highlighted several apoptosis-related genes that were downregulated by C1q in HMDMs during clearance of modified lipoproteins (Fig. 1); this was validated by quantitative PCR (Fig. 2A). Interestingly, C1q bound to oxLDL was able to downregulate certain apoptosis genes (CASP8, BCL2L11, TNFSF10) to below basal levels in untreated macrophages. This suggests that, by binding to modified lipoproteins, C1q is not simply blocking or interfering with the ability of these lipoproteins to trigger apoptosis but has an active signaling role. In our previous studies, we identified several signaling pathways modulated by C1q during lipoprotein clearance that may impact survival signaling in macrophages. These include suppression of NF-κB and JAK–STAT activation, along with activation of peroxisome proliferator–activated receptor (PPAR) pathways (29, 31). STAT-1 was shown to be critical for ER stress–induced macrophage apoptosis in atherosclerosis (22); therefore, C1q suppression of STAT-1 activation may be directly affecting apoptosis pathways. The mechanism and receptor through which C1q exerts its effects on macrophages are still under investigation. C1q was shown to bind modified LDL via its globular heads domain (43), suggesting that the collagen-like domain of C1q (cC1q) is available and responsible for modulating macrophage functions. A role for calreticulin as a possible receptor for cC1q during uptake of apoptotic cells was identified (44); however, cC1qR is not definitively identified for all targets or C1q-mediated functions (45–47). We showed previously that C1q does not bind the unmodified form of LDL or the isolated fragments of the C1q collagenous domain or C1q globular head domain (25, 31). Importantly, the soluble form of these fragments or intact C1q also does not modulate any macrophage responses tested, including the survival responses detailed in this article (25, 31) (data not shown). Therefore, it is likely that a multivalent presentation of the intact C1q molecule is required to trigger macrophage responses.

Luminex multiplex analysis of levels of active (cleaved) caspase-3 and the downstream apoptosis effector molecule PARP-1 in HMDMs confirmed that C1q also suppressed protein levels of these molecules over a sustained period of time (3–24 h). Levels were activated above basal levels during macrophage ingestion of oxLDL or acLDL from as early as 1 h (Fig. 3). This suggests that the measured C1q reductions in apoptosis gene expression are functionally relevant. To validate the murine macrophage cell line Raw264.7 as a model system for these studies, CASP7 gene expression was measured in Raw264.7 cells that ingested oxLDL in the presence or absence of C1q (Fig. 2B). Similar to HMDMs, C1q downregulated CASP7 activation at 3 h; thus, additional experiments were performed in the murine macrophage cell line to reduce the donor variability associated with primary human cells.

The activity of caspase-3/7 was measured in Raw264.7 macrophages using a fluorogenic substrate (Fig. 4). Similar to our gene expression and active protein expression data in HMDMs, ingestion of modified lipoproteins led to dose-responsive increases in caspase activation in Raw264.7 macrophages that was suppressed when C1q was present. Although there appears to be consensus that lipid-laden macrophages are prone to apoptosis in atherosclerosis (48), reports of direct cytotoxicity of modified LDL in vitro are variable in the literature. Several studies (49–51) reported that ingestion of oxLDL under certain conditions actually leads to an increase in macrophage survival, although this varies depending on the lipoprotein dosage, time of incubation, and cell type. In general, higher doses of oxLDL (>50 μg protein per milliliter) and longer periods of incubation (>24 h) are associated with increased toxicity/apoptosis in macrophages (52, 53). In our studies, we used physiologically relevant amounts of LDL (from 10 to 100 μg protein per milliliter) for up to 24 h, which we previously showed generates lipid-laden foam cells (25). In our cells, addition of oxLDL or acLDL led to increases in caspase-3/7 activity (Fig. 4), a reduction in viability (Fig. 5A), and an increase in dead cells (Fig. 5B), suggesting that these molecules are measurably cytotoxic in our system, even at 10 μg protein per milliliter. We perform these studies in a serum-free defined media (X-VIVO 15); it is possible that we are depriving our cells of an unknown serum factor that interferes with modified lipoprotein signaling and, thus, our cells are more sensitive to their toxicity. However, again, it is important to note that C1q restores cell viability levels and decreases cell death in macrophages ingesting modified LDL and that levels in the presence of C1q are also significantly modulated compared with basal untreated macrophages. Thus, C1q appears to be actively reprogramming these macrophage foam cells rather than simply interfering with modified LDL signaling. C1q was shown to promote the survival of neurons in response to amyloid beta–induced cytotoxicity (54), in which modulation of cholesterol metabolism was determined to play a role (55). Further studies are needed to identify whether similar mechanisms are involved in C1q-related macrophage survival signaling.

In early atherosclerotic lesions, an increase in macrophage foam cell survival is not necessarily beneficial and can be detrimental. For example, in the LDLR^−/− murine model of atherosclerosis, mice deficient in survival signaling protein AIM had smaller lesions and higher amounts of apoptotic macrophages than did AIM-sufficient animals (19). In contrast, larger aortic lesions were observed in different murine models of atherosclerosis (LDLR^−/−, ApoE^−/−) in which macrophages lacked the proapoptotic protein Bax or p53, respectively (17, 18). However, in later stages of disease, increases in macrophage apoptosis are associated with plaque necrosis and rupture (reviewed in Ref. 24). These seemingly contradictory observations may be due, in part, to the process of clearance of apoptotic cells (efferocytosis), which plays an important role in the resolution of inflammation (56, 57). In early stages of atherosclerosis, macrophage apoptosis and efficient efferocytosis may lead to a reduction in M1-polarized, proinflammatory foam cells formed by the ingestion of modified lipoproteins. However, as lesions develop, levels of apoptotic cells are increased, concomitant with defective efferocytosis leading to plaque necrosis (57). The mechanisms behind this are not precisely defined; however, excessive levels of apoptotic cells, as well as a reduction in the efferocytic capabilities of phagocytes in the region, may contribute to the defective efferocytosis. Indeed, our data show that ingestion of oxLDL by macrophages increases cell death and decreases their phagocytic and efferocytic capabilities compared with untreated macrophages (Figs. 5B, 6). C1q has a well-described role in enhancing phagocytosis of foreign targets and immune complexes, as well as in efferocytosis of apoptotic cells during normal tissue homeostasis and the prevention of autoimmunity (reviewed in Refs. 47 and 58). Studies using the LDLR^−/− murine model of atherosclerosis also highlight an important role for C1q in efferocytosis of apoptotic macrophages in atherosclerotic lesions (32). C1q-deficient mice in this study were shown to have increased lesion size, as well as an accumulation of apoptotic cells. Thus, we tested whether C1q could restore defective phagocytosis/efferocytosis in macrophage foam cells. Macrophages that ingested oxLDL opsonized with C1q were significantly more phagocytic than foam cells formed in the absence of C1q and untreated macrophages. Therefore, C1q is able to program even foam cells toward a highly efferocytic phenotype. Interestingly, C1q had an effect on the number of cells capable of performing phagocytosis/efferocytosis (percentage phagocytosis/efferocytosis), as well as on the capability of individual cells to engulf more targets (phagocytic index). One of the molecules involved may be Mer tyrosine kinase, which is known to be involved in efferocytosis (59, 60) and to play a role in the atherosclerotic lesion (61), and we and other investigators identified it as being upregulated by C1q in HMDMs ingesting modified lipoproteins (D. A. Fraser, unpublished observations) or apoptotic cells (59, 60, 62). In a previous study, we also showed that, although C1q increases ingestion of modified lipoproteins at early time points (3 h), after 24 h, C1q-treated HMDMs or human monocytes have enhanced levels of cholesterol efflux, as well as a reduction in total cholesterol accumulation (25). The timing of this reduction in intracellular lipid levels/composition occurs too late to affect the gene expression described in these studies, but it may contribute to the general survival and efferocytic ability of these C1q-treated macrophage foam cells.

Thus, a picture is emerging of C1q as a potent regulator of macrophage function in atherosclerosis (29–32). Macrophages can be a major source of C1q production in vivo (28); therefore, it is likely present in macrophage-rich lesions in the absence of complement proteins C1r and C1s, which are involved in triggering the proinflammatory complement cascade. C1q directly binds modified lipoproteins and enhances their uptake by macrophages. However, opsonization with C1q alters the function of lipid-laden macrophage foam cells toward an anti-inflammatory efferocytic resolving phenotype (25, 29, 31) (Fig. 6). Therefore, therapeutic strategies for atherosclerosis should consider approaches that inhibit complement activation, as well as those that leave intact or enhance the beneficial functions of C1q that are not related to complement.

We thank Christopher Quan for technical assistance with this project.

The authors have no financial conflicts of interest.