|Home | About | Journals | Submit | Contact Us | Français|
Apoptosis occurs in many tissues, during both normal and pathogenic processes. Normally, apoptotic cells are rapidly cleared, either by neighboring or recruited phagocytes. The prompt clearance of apoptotic cells requires that the apoptotic cells announce their presence through the release of chemotactic factors, known as ‘find-me’ signals, to recruit phagocytes to the site of death, and through the exposure of so-called ‘eat-me’ signals, which are ligands for phagocytic uptake. The importance of prompt apoptotic cell clearance is revealed by findings that decreasing the efficiency of engulfment results in the persistence of apoptotic cells, which is often associated with chronic inflammation and autoimmunity. Additionally, the proper clearance of apoptotic cells is actively anti-inflammatory, which is thought to play a crucial role in immunologic tolerance. Therefore, defects associated with clearance of apoptotic cells may contribute to the pathogenesis of several inflammatory diseases, including autoimmunity and atherosclerosis. Here, we review the role of nucleotides in the apoptotic cell clearance process, and discuss their implications for disease pathogenesis.
Apoptosis occurs in a variety of tissues during the development of an organism, tissue homeostasis, and pathogenic processes [1, 2]. During development, apoptosis plays a role in sculpting structures and getting rid of excess cells. Cell turnover is also a vital part of tissue homeostasis, as senescent cells are disposed of before they become less functional or even dysplastic. In addition to the normal processes, apoptotic cells are also observed in tumors , atherosclerotic plaques , and neurodegenerative diseases . In healthy tissues, the efficient and rapid clearance of apoptotic cells occurs through a concerted effort by the apoptotic cells themselves, and the neighboring or recruited phagocytes. The quick clearance of these dying cells helps to prevent an inflammatory insult resulting from the uncontrolled release of intracellular contents. In this review, we provide an overview of the components of apoptotic cell clearance and discuss the general link between apoptotic cell clearance and disease pathogenesis. We also specifically evaluate the role of nucleotides in apoptotic cell clearance.
Apoptotic cell clearance can be broken down into four general steps: migrating towards the dying cell; binding/recognition of the apoptotic cell; phagocytosis/internalization of the target; and processing the ingested apoptotic cell (see Fig. 1) . Here we will briefly discuss various components of how this process occurs.
The clearance of apoptotic cells by so-called professional phagocytes requires that the phagocyte can properly locate the dying cell. Locating cells that are undergoing apoptosis involves the first two steps of the clearance process: migrating towards the dying cell (if there are no phagocytes already in the vicinity of the cell) and recognizing the apoptotic cell amongst its living neighbors. Cells undergoing apoptosis release factors, such as the nucleotides ATP and UTP [6-8], the chemokine fractalkine (CX3CL1) , and the lipids lysophosphatidylcholine (LPC)  and spingosine-1-phosphate (S1P) , to promote the recruitment of motile phagocytes. In the context of apoptotic cell clearance, these chemotactic factors are called ‘find-me’ signals for the role they play in the recruitment of phagocytes to the dying cells (a molecular ‘apoptotic cell beacon’). These factors, released by different mechanisms, set up a concentration gradient that allows phagocytes expressing their cognate receptors to migrate toward the site of death. Receptors for some of these find-me signals have been identified, including the purinergic receptor P2Y2 for ATP and UTP  and the chemokine receptor CX3CR1 for fractalkine (CX3CL1) . The receptor G2A has been shown to mediate chemotaxis to LPC . However, the original study that identified LPC as a G2A ligand has been retracted , which raises some questions regarding the ligand that mediates chemotaxis in these systems. We will focus on the role of nucleotides as a find-me signal in the section titled ‘Nucleotides as a find-me signal’, and in doing so will discuss some aspects of find-me signals in general (for a review of all potential find-me signals see Muñoz et al. Autoimmunity Reviews 2010 ).
In addition to releasing find-me signals, apoptotic cells also expose ‘eat-me’ signals on their surface, promoting their recognition by the recruited phagocyte. The most well known of these eat-me signals is phosphatidylserine (PtdSer). Normally concentrated on the inner leaflet of the plasma membrane, PtdSer loses its asymmetric distribution during apoptosis and appears on the outer leaflet of the plasma membrane [15, 16]. The exact mechanism by which PtdSer is exposed during apoptosis is still unclear. Normally, phosphatidylserine is concentrated on the inner leaflet of the plasma membrane by aminophospholipid translocase (APLT) activity. Several groups have demonstrated that APLT activity decreases during apoptosis, thereby removing the driving force that maintains PtdSer asymmetry [17, 18]. However, how does PtdSer make it to the outer leaflet? Since the headgroup of PtdSer is polar, spontaneous jumping from the inner to the outer leaflet of a bilayer occurs rather slowly . Several mechanisms, which are not exclusive of each other, have been suggested to mediate the increase in PtdSer on the outer leaflet. One mechanism is that increased ‘scramblase’ activity, which catalyzes the bidirectional transbilayer movement of phospholipids, might allow for PtdSer to ‘diffuse’ down its concentration gradient to the outer leaflet during apoptosis [17, 19]. The exposure of PtdSer on the outer leaflet might also be due to fusion of vesicles with the plasma membrane , perhaps as part of a calcium-induced membrane repair response . While the APLT and scramblase activities are thought to play a role in the exposure of PtdSer during apoptosis, the identity of the proteins mediating these activities are largely unknown and controversial [22-24].
The recognition and subsequent engulfment of apoptotic cells by phagocytes is mediated by receptors that either directly or indirectly (via bridging molecules) bind eat-me signals. Here, we will briefly discuss several receptors and bridging molecules that bind the eat-me signal PtdSer. Bridging molecules (opsonins) are secreted proteins that bind PtdSer on the surface of apoptotic cells and are subsequently recognized by their cognate receptors on the phagocyte. MFG-E8 and Gas6 are two bridging molecules that bind the vitronectin receptor (αVβ3 integrin) and the receptor tyrosine kinase Mer respectively [25, 26]. In addition to the indirect link to PtdSer, several membrane receptors that directly bind PtdSer have been identified. BAI1 , Tim4 and Tim1 [28, 29], and Stabilin-2  have been shown to mediate uptake of apoptotic cells by directly binding PtdSer. Related receptors such as Tim3  and Stabilin-1  have also been shown to play a similar role. For some of these receptors, ligation to PtdSer, either directly or indirectly, results in Rac-dependent cytoskeletal reorganization, which ultimately leads to engulfment of the apoptotic cell . However, Tim-4 does not appear to signal significantly through any of the known intracellular signaling pathways for engulfment, and its cytoplasmic tail appears dispensable .
Activation of Rac during phagocytosis of apoptotic cells occurs through one of two delineated intracellular signaling pathways: through the mammalian intracellular signaling molecules ELMO, Dock180, and CrkII, or the adaptor molecule GULP. ELMO and Dock180 interact together to form a bipartite guanine nucleotide exchange factor (GEF) for Rac, while it is still unknown how GULP leads to Rac activation (for a more detailed overview of intracellular signaling for apoptotic cell phagocytosis see [5, 34]).
Once the target has been internalized, the phagosome is progressively acidified, leading to degradation of the ingested cell  (for a review of phagosome maturation see [35-38]). Processing of the ingested cell within the phagolysosome leads to an increased load of cellular metabolites. How the phagocyte deals with the raw materials and energy derived from this catabolic process is an interesting question that remains largely unanswered . The work that has been done has addressed cholesterol homeostasis in the phagocyte. Engulfment of apoptotic cells results in increased cholesterol efflux by the phagocyte through ABCA1, in addition to incorporation of cholesterol derived from the apoptotic cell into the phagocyte’s membrane . However, it appears that this response was not mediated by metabolites derived from the engulfed cell, rather from binding of PtdSer, as surrogate targets (beads which mimic PtdSer on apoptotic cells) also elicited a similar response. This suggests that simply tickling the engulfment receptors may be sufficient to trigger a homeostatic response in the phagocyte for controlling its own cellular contents (in this case, cholesterol).
Aside from dealing with the ingested cellular components, phagocytes also change their ‘behavior’ in several interesting ways. One post-engulfment response that has been described is the LXR nuclear receptor dependent upregulation of Mer (receptor that binds PtdSer indirectly through the bridging molecule Gas6) . This finding suggests that phagocytes may become more efficient after every meal, possibly via signaling by lipid metabolites derived from degradation of the internalized apoptotic cells.
Engulfment-induced modulation of the phagocyte is not limited to metabolic processes or the phagocytic machinery. Apoptotic cell clearance is described as an immunologically ‘silent’ process, due to the fact that it does not elicit an immune response (like necrosis) [41, 42]. Therefore, immunologic consequences of engulfment have been of particular interest in the field. The clearance of apoptotic cells is not only silent, but it is also actively anti-inflammatory; engulfment promotes the secretion of the anti-inflammatory cytokines such as TGFβ by macrophages [43, 44]. This anti-inflammatory effect is even potent enough to suppress LPS-induced inflammatory cytokine release [43, 45]. Interestingly, it appears that internalization of the corpse is not required for the anti-inflammatory effect mediated by the phagocyte [46, 47], and that PtdSer recognition alone can mimic the response [44, 48].
Collectively, it is through the coordinated release of find-me signals and exposure of eat-me signals that apoptotic cells assure their proper disposal. Phagocytes also play a role in promoting the immunologically ‘quiet’ nature of clearance, by releasing anti-inflammatory cytokines and coordinating the proper disposal of the engulfed contents.
It has been a long-standing puzzle in the apoptosis/engulfment field, that very few apoptotic cells are observed, even in tissues that are known to have high turnover of cells (such as the bone marrow or thymus). This led to the notion that apoptotic cells may advertise their presence at the earliest stages of apoptosis, via the release of ‘find-me’ signals; this could attract phagocytes to their proximity and thereby lead to the prompt clearance of the dying cells. Several factors have been suggested to act as find-me signals, including the nucleotides ATP and UTP [6-8], the chemokine fractalkine (CX3CL1) , and the lipids lysophosphatidylcholine (LPC)  and spingosine-1-phosphate (S1P) . Among these, only the nucleotides and fractalkine have been shown to have relevance in clearing apoptotic cells in vivo [6, 9].
How do find-me signals attract the phagocytes? While receptors mediating chemotaxis to these various find-me signals have also been identified in these studies [6, 9, 12], the effective ranges of these find-me signals still need to be defined. For example, it has been shown that ATP added to the basolateral side of an endothelial cell monolayer can promote monocyte transmigration in vitro . Do the find-me signals ATP and UTP permeate into the intravascular space to promote extravasation of circulating monocytes directly? This would require at least two things to occur. The nucleotides would need to cross the endothelial cell barrier somehow (transcytosis or diffusion through the endothelial cell barrier) and then the nucleotides would need to be presented on the luminal surface, in order to provide some spatial information about the location of the dying cells without being swept away in the circulation . Additionally, these would need to occur while avoiding degradation by extracellular nucleotidases. These requirements could be avoided if the find-me signals were to act on endothelial cells, causing the upregulation of some stable molecular signpost on the luminal surface that promotes monocyte extravasation. Along these lines, it has been shown that ATP and UTP are able to induce Vascular Cell Adhesion Molecule-1 (VCAM-1) expression in Human Coronary Artery Endothelial Cells (HCAECs) in vitro . The above ideas can be synthesized into the following model (see Fig. 2). Apoptotic cells release nucleotides, which would promote monocyte extravasation by upregulating adhesion molecule/chemokine expression by vascular endothelial cells. Once the macrophage is in the interstitium, the chemotactic gradient set up by the nucleotides would then mediate the attraction of the phagocyte to the dying cell. In the case where there are motile resident phagocytic cells (such as macrophages or microglia), the nucleotides would not need to promote extravasation of phagocytes. As an example, nucleotides have been shown to mediate microglial (resident phagocytes in the central nervous system) chemotaxis towards injured neurons [52, 53].
In addition to setting up a chemotactic gradient to aid in the location of the dying cells, find-me signals might also have a role in modulating the phagocytic ability or activity of cells in the direct vicinity of the apoptotic cells (see Fig. 2). Apoptotic cells release UTP , which is degraded by extracellular enzymes through the removal of 5’ phosphates to produce UDP (in addition to other nucleotide metabolites). In the context of neuronal injury, UDP has been shown to promote phagocytosis by microglia . While it was demonstrated that UDP promotes phagocytic activity via the P2Y6 nucleotide receptor, the mechanism of the boosted activity is unknown. This could occur by inducing the upregulation of proteins necessary for the uptake of apoptotic cells, such as secreted opsonins, receptors, or intracellular signaling molecules. In fact, ramping up of proteins involved in the engulfment of apoptotic cells has been seen in the context of increased cell death [55-57], but the mechanism of the induction in these systems is unknown. Even though it was not shown in the context of cell death, it has been demonstrated that the find-me signal fractalkine can induce production of the bridging molecule MFG-E8 by macrophages . This suggests that find-me signals may indeed play a role in modulating the activity of phagocytes.
Even though ATP has been typically thought of as a danger signal , apoptotic cell supernatants appear to preferentially recruit monocytes over neutrophils in vivo . Furthermore, apoptotic cell clearance is usually anti-inflammatory and immunologically silent, as phagocytes release anti-inflammatory mediators (such as TGFβ, IL-10 and Prostaglandin E2) after ingestion of apoptotic cells [43, 45]. Therefore, the recent identification of nucleotides as a find-me signal raises the following question: How can apoptotic cell clearance be immunologically silent if the apoptotic cells release factors that are considered inflammatory molecules, or danger signals?
There are perhaps several differences between the release of nucleotides during cell death via cytolysis and apoptosis (see Fig. 3). First and foremost, we will raise the issue of quantity. The nucleotide release seen during apoptosis appears to represent a rather small quantity of the total cellular ATP content (< 2%) . Therefore, the regulated release of nucleotides during apoptosis is significantly less than that seen during cell lysis, or damage-induced loss of membrane integrity. ATP’s reputation as an inflammatory molecule is based largely on its ability to activate the ionotropic nucleotide receptor P2X7, which in turn results in activation of the inflammasome and release of pro-inflammatory cytokines [60, 61]. Along these lines, ATP derived from necrotic cells has been shown to result in sterile inflammation via inflammasome activation . However, the concentrations necessary for activation of P2X7 (EC50 > 100μM) are much higher than those necessary for activation of the receptors mediating chemotaxis (such as P2Y2; EC50 < 1μM) [59, 60]. Intriguingly, it has been shown that lower concentrations of ATP may actually have an anti-inflammatory effect by suppressing the secretion of inflammatory cytokines, while promoting the release of anti-inflammatory cytokines [61, 63-65]. Therefore, the concept of ATP as a universal danger signal might be too simplistic.
In addition to differences between quantities of ATP released by apoptotic cells versus necrotic cells, there are also other cellular factors that are differentially released . While necrotic cells theoretically release all of their intracellular contents, apoptotic cells become selectively permeable, retaining most of their intracellular contents . HMGB-1 is an example of a molecule that is released by necrotic cells, but not apoptotic cells, that is capable of inciting inflammation . Interestingly, it appears that the lack of HMGB-1 release by apoptotic cells is not simply because of selective membrane permeability. Apoptotic cells actively retain HMGB-1 by deacetylating histones, which increases the affinity of HMGB-1 for the chromatin . Apoptotic cells also release factors that are not released by necrotic cells to modulate the inflammatory signature of ATP. It was recently demonstrated that apoptotic cells release lactoferrin, which acts as a “don’t find-me” signal for neutrophils . Perhaps lactoferrin plays a role in dampening the attraction of neutrophils to ATP released by apoptotic cells. Yet another potential distinguishing factor between nucleotide release during apoptosis versus necrosis is the relative quantity of UTP that is released with ATP. Apoptotic cells release roughly similar quantities of the two nucleotides (even though ATP levels in the cell are several fold higher than UTP), whereas necrotic cells release the nucleotides proportionally to their intracellular levels . Therefore, differential activation of various nucleotide receptors (which have different affinities for the various nucleotides) may also play a role in the ability to distinguish necrosis from apoptosis.
It has been suggested that the efficient clearance of apoptotic cells is important for the maintenance of self tolerance, and that deficiencies in this process may lead to the development of autoimmunity. Reducing the efficiency of the clearance process, at any of the steps preceding engulfment, results in the persistence of apoptotic cells [6, 31, 40, 57, 69-73]. Early apoptotic cells are relatively intact (i.e., plasma membrane is selectively permeable, allowing regulated release of small molecules ), but if they are not cleared in a timely manner, they eventually become secondarily necrotic. It is thought that secondary necrosis may promote an immune response to intracellular antigens, since the antigens are exposed in an inflammatory environment, caused by the release of various danger signals. Mounting an immune response to normally hidden intracellular proteins does not seem to be simply a product exposure to these antigens due to cytolysis. During apoptosis, some antigens undergo post-translational modifications , and are also concentrated in membrane blebs on the surface of cells during apoptosis [75-77]. Post-translational modifications and altered localization may contribute to bypassing tolerance in various autoimmune diseases [74, 78]. Regardless of the mechanism by which autoantigens are revealed to the immune system, it appears that persistence of apoptotic cells plays a role in the development of autoantibodies, and perhaps autoimmune disease. Interestingly, intravenous administration of apoptotic cells have been shown to result in the development of autoantibodies, presumably due to excessive apoptotic cells that cannot be efficiently cleared, . Hindering clearance in vivo has also been shown to induce the development of autoantibodies, and in some cases autoimmune-disease-like phenotypes [28, 31, 40, 69, 70, 72, 73, 80]. Complicating the interpretation of some of these results, several of the receptors (such as Tim4 and the vitronectin receptor) have other roles in immune tolerance that might be responsible for the phenotype seen . In fact, it is not clear what components are necessary to induce autoimmunity, as reducing clearance does not always result in the development of autoantibodies  and development of autoantibodies does not always result in autoimmunity (for a broader review of the link between apoptotic cell clearance, autoimmunity and other disease processes, please see Elliott and Ravichandran, The Journal of Cell Biology 2010 ).
Systemic Lupus Erythematosus (SLE) is an autoimmune disease characterized by a variety of symptoms, which are thought to be largely due to the presence of antibodies directed against self antigens (autoantibodies) . Interestingly, patients with systemic lupus erythematosus (SLE) have higher amounts of apoptotic cells in their circulation  and in lymph nodes . Macrophages from SLE patients have also been shown to have less phagocytic capacity , suggesting that the accumulation of apoptotic cells is due to a defect in clearance. Whether defects in apoptotic cell clearance cause SLE or are a byproduct of the inflammation is still unclear (for a more in-depth discussion of the role of apoptotic cell clearance in autoimmunity see Muñoz et al. Nature Reviews Rheumatology 2010 ).
Inefficient clearance of apoptotic cells also appears to be a factor in the pathogenesis of atherosclerosis . Defective clearance may contribute to the development of a necrotic core in an atheroma, as uncleared apoptotic cells become secondarily necrotic. Accumulation of secondarily necrotic cells in an atherosclerotic plaque is thought to play a role in the establishment of an inflammatory environment in the plaque and also contributes to its structural instability. This inflammatory plaque, laden with necrotic cells is the late stage plaque that precedes plaque rupture and thrombosis. Knocking out molecules (such as the bridging molecules C1q and MFG-E8 or the receptor tyrosine kinase Mer) that mediate phagocytic removal of apoptotic cells in atherosclerosis-prone mice (LDLR-/- or ApoE-/-), results in accumulation of apoptotic cells and acceleration of atherosclerosis [90-93]. One group demonstrated that functionally knocking out Mer in bone marrow derived cells alone is sufficient to see the increase in apoptotic cells (presumably due to decreased clearance) and acceleration of atherosclerosis . However, it must be noted that these molecules have other roles in the immune system, and therefore the effect cannot be conclusively attributed solely to deficiencies in clearance.
Two post-engulfment responses are of particular interest to the field of atherosclerosis research: the upregulation of ABCA1 expression  and the release of anti-inflammatory cytokines  by phagocytes. Lipid-laden macrophages (foam cells) play a crucial role in the atherogenesis . Since ABCA1 expression has been shown to be inversely related to atherogenesis , perhaps apoptotic cell induced ABCA1 upregulation plays a role in protection against atherosclerosis by preventing the accumulation of lipids within macrophages. The release of anti-inflammatory cytokines by phagocytes is also relevant to atherosclerosis for at least two reasons. First, atherosclerosis is an inflammatory disease [95, 96], and therefore quenching inflammation through the release of cytokines could have an effect in promoting disease resolution. Second, TGFβ, which is one of the anti-inflammatory cytokines released after engulfment, has been shown to promote fibrous cap development in atherosclerotic plaques, protecting against thrombotic events due to plaque rupture [97-100]. Furthermore, TGFβ expression correlates with plaque stability in humans . Therefore, it is feasible that defects in phagocytosis of apoptotic cells may exacerbate atherosclerosis, since the post-engulfment responses (cholesterol efflux and anti-inflammatory cytokine release) are generally atheroprotective.
Apoptotic cell clearance is not only relevant to disease from the standpoint of deficiencies leading to inflammatory milieus, but also for its potential role in perpetuating other disease processes. The presence of macrophages plays a critical role in the progression of some disease processes, including atherosclerosis  and cancer . A number of chemokines have been implicated in the recruitment of monocytes to tumors  and atherosclerotic plaques . However, since apoptotic cells are present in both atherosclerotic plaques [3, 105] and tumors , perhaps apoptosis plays a role in the initial recruitment of monocytes in these disease processes. This recruitment of monocytes does not need to be independent of chemokines, since as addressed above, find-me signals may play a role in inducing the production of chemokines by neighboring cells. It has been suggested that lipid deposition in the sub-endothelial space results in macrophage accumulation in atherosclerosis . However, it is not known how lipid deposition results in the recruitment of monocytes. Perhaps lipids promote apoptosis, which leads to monocyte recruitment. Aside from fractalkine release from Burkitt Lymphoma cells , the role of find-me signals in the initial recruitment of macrophages to tumors has not been investigated. However, elevated levels of ATP have been detected in tumors , suggesting that nucleotide find-me signals might be released by apoptotic tumor cells.
Apoptotic cell clearance is an important process that helps to maintain homeostasis in healthy tissue. We have outlined the various steps in the clearance process and discussed how they are related to pathogenesis of disease. We have also discussed the nucleotide find-me signals and how they may play the two seemingly disparate roles as a find-me signal and a ‘danger’ molecule. There have been many contributions to this field recently, including the identification of several receptors for phosphatidylserine, which is exposed on the surface of apoptotic cells, and new find-me signals released by cells undergoing apoptosis. What remains to be determined is the relative contribution of each of these players to clearance in vivo, and how impairment of the molecules or the pathways they regulate could result in disease. A potential future direction that needs to be explored is how to ‘improve’ or restore engulfment in certain disease states and whether such manipulations could be therapeutically beneficial. The recent advances in the field and the pace of discoveries portend a bright future and possibilities for therapies based on targeting the engulfment machinery.
The authors acknowledge funding from the National Institutes of Health. F.B.C. was supported by a Pharmacological Sciences Training Grant (National Institute of General Medical Studies) and a F30 pre-doctoral fellowship (National Heart, Lung and Blood Institute). This work was supported by funding (to K.S.R.) from the National Institutes of Health, American Asthma Foundation and The Goldhirsh Foundation. K.S.R. is a William Benter Senior Fellow of the American Asthma Foundation.
The authors declare no conflict of interests related to this study.