Apoptosis is an essential process for normal physiology and plays a key role in the resolution of inflammation. Clearance of apoptotic cells (ACs) involves complex signalling between phagocytic cells, ACs, and the extracellular vesicles (EVs) they produce. Here, we discuss apoptotic cell-derived extracellular vesicles (ACdEVs) and how their structure relates to their function in AC clearance and the control of inflammation, focussing on the ACdEV proteome. We review the current knowledge, ongoing work and future directions for research in this field.
Apoptosis plays a key role in physiology; it is the process by which unwanted cells are destroyed, which is central to development and immune regulation, in addition to normal cell turnover and homeostasis [1,2]. The process is genetically programmed and mediated by caspases which are involved in a proteolytic cascade . Defects in apoptosis are implicated in many diseases including cancer [3,4], inflammatory conditions and autoimmune diseases [5–7]. Extracellular vesicles (EVs) are membrane-enclosed structures secreted by cells, which are important mediators of intercellular communication in physiological and pathophysiological processes . EVs are recognised as important mediators of inflammation and their immunomodulatory roles are significant in inflammatory diseases . Apoptosis involves the disassembly of apoptotic cells (ACs) into smaller fragments, including apoptotic cell-derived extracellular vesicles (ACdEVs). ACdEVs are emerging as mediators of inflammation and promoters of AC clearance [10,11], although few of their specific functions have been identified [12–14]. Here, we discuss the significance of ACdEVs in AC clearance and immune modulation, and the identification of key ACdEV molecules that mediate their functions. Many questions regarding ACdEVs remain to be answered and will be discussed here.
EVs: complex mediators of intercellular communication
The biogenesis of EVs incorporates components of the cell and, as such, EVs contain and signal via lipid mediators, nucleic acids and proteins [15–18]. Cells produce exosomes (∼30–150 nm) and microvesicles (MVs, ∼100–1000 nm) which have distinct mechanisms of biogenesis . Exosomes are generated through an endosomal pathway where multivesicular bodies, formed by the maturation of early endosomes into late endosomes, fuse with the plasma membrane and release their intraluminal vesicles . MVs are produced by direct budding of the plasma membrane . It is likely that different stimuli change the rate of EV release and the composition of the EVs, as the phenotype of the stimulated cell changes. ACs produce ACdEVs of ‘greatly varying size’ , including apoptotic bodies (1–5 µm) and smaller EVs which may or may not have the same biogenesis pathways as EVs produced by viable cells. Apoptotic bodies are produced by the process of membrane blebbing and other mechanisms of ACdEV biogenesis include the formation of microtubule spikes, apoptopodia and beaded apoptopodia [21,22]. The release of exosomes by ACs remains to be confirmed in the literature. One key function of EVs may be to prevent degradation of their cargoes in the extracellular environment, and EVs can travel to cells and tissues at distant sites. EV-associated signals can be delivered to recipient cells via EV–cell surface interactions involving specific receptors, which facilitates targeting of specific cell types [8,18]. Otherwise, the content of EVs must be transferred into the recipient cell; internalisation of EVs can occur by membrane fusion, endocytosis, phagocytosis or macropinocytosis [23–27] and the release of EV contents into the cytosol requires fusion of the EV membrane with the cell membrane or the membrane of endocytic compartments . EVs and their contents may simply be destroyed in lysosomes after internalisation.
Apoptosis and apoptotic cell clearance
Apoptosis is the process of programmed cell death, first described in 1972 by Kerr et al. , which is essential for normal cell turnover and homeostasis, development and tissue remodelling, as well as control of the immune system . Efficient clearance of ACs prevents secondary necrosis and subsequent inflammatory responses . Apoptosis plays a central role in the resolution of inflammation, as immune cells that are no longer required undergo apoptosis and are cleared by phagocytes . Resolution of inflammation involves the depletion of pro-inflammatory signals, inhibition of leukocyte migration and tissue infiltration, apoptosis and cell clearance. AC clearance is discussed in detail in many reviews [10,30–33]. Firstly, ACs release ‘find me signals’ which stimulate migration and recruitment of phagocytes to the site of cell death. The release of AC ‘find me’ signals is regulated; for example, the release of ATP and UTP is mediated by pannexin 1 channels, which are opened on cleavage by activated caspases [34,35]. As well as the nucleotides ATP and UTP, other ‘find me signals’ released include chemokines such as CX3CL1 (C-X3-C motif chemokine ligand 1) , and lipids such as lysophosphatidylcholine (LPC)  and sphingosine-1-phosphate (S1P) . Some signals released by ACs act to inhibit migration of phagocytes; for example, lactoferrin inhibits migration of granulocytes but not mononuclear phagocytes . Once an AC has been located by a phagocyte, whether engulfment (efferocytosis) occurs is determined by the balance between ‘don't eat me’ and ‘eat me’ signals (apoptotic cell-associated molecular patterns — ACAMPs) presented on the cell surface . Loss of membrane asymmetry and externalisation of phosphatidylserine (PS) is a hallmark of apoptosis, and PS is a key molecule for recognition of ACs which binds to various receptors on phagocytic cells [32,40]. Other ‘eat me’ signals on the surface of ACs recognised by scavenger receptors include the proteins, annexin 1 , ICAM-3 (intercellular adhesion molecule 3)  and calreticulin [43,44]. Recognition and uptake of ACs also involve modification of the structure and function of ‘don't eat me’ signals such as CD31 and CD47 [44,45].
Apoptotic cell-derived EVs
ACs disassemble into smaller fragments in the form of apoptotic bodies, smaller plasma membrane-derived MVs , and ACs potentially also release exosomes. The formation of ACdEVs plays an important role in AC clearance, as they harbour AC ‘find me’ and ‘eat me’ signals and perhaps act as smaller and more easily engulfed cell fragments (Figure 1). ACdEVs are, therefore, also central to control and resolution of inflammation. The surface proteome of EVs likely determines EV–cell interactions. This could be mediated by relatively simple physicochemical means (e.g. charge) and/or more specific and high-affinity ligand–receptor molecular interactions (e.g. through the recognition of proteins, lipids or carbohydrates), which mediate recognition, uptake and signalling . Macrophages preferentially phagocytose ACdEVs with certain surface compositions, such as specific glycosylation patterns; a study by Bilyy et al.  found that macrophages prioritised EVs presenting immature mannosidic glycoepitopes over those with desialylated glycoepitopes. As well as surface molecules, ACdEVs will also signal via the transfer of proteins and nucleic acids into recipient cells. ACdEVs can contain pro- and anti-inflammatory mediators, but apoptosis and AC clearance are generally non-inflammatory processes. The anti-inflammatory functions of ACdEVs are important for prevention of autoimmune disease; for example, ‘apoptotic blebs’ from lymphocytes down-regulate MHC II expression on dendritic cells, but this process is impaired in systemic lupus erythematosus patients . MVs from apoptotic neutrophils promote resolution of inflammation by suppression of the activity of T-helper cells . Production of ACdEVs can create an amplification loop of cell death, promoting apoptosis in recipient cells and further release of ACdEVs . There is a significant gap in our knowledge as the key molecules regulating these effects remain to be identified. The balance between inflammatory mediators changes over the course of apoptosis; therefore, changes in the composition of ACdEVs over time should also be investigated. Tumour cell-derived ACdEVs are often anti-inflammatory, pro-resolution and create an onco-regenerative niche . Additionally, ACdEVs can play roles in promoting inflammation [7,50], antigen presentation  and infection — EVs can be hijacked to suppress immune responses or aid the spread of infection to neighbouring cells [51,52].
Apoptotic cell clearance in the resolution of inflammation.
Identifying ACdEV structure–function relationships
The molecular composition of EVs is a major focus for the field and this knowledge has led to an improved understanding of EVs and, in particular, how EV components (protein- and non-protein-based) may be used for the characterisation of different sub-types of EV. This is reviewed within the position statement from the International Society for Extracellular Vesicles . While many studies have identified the roles for EVs in physiology or pathology, there are only a few studies that have characterised the functions of specific ACdEV-associated molecules. Current knowledge of the roles of lipids and lipid mediators has been reviewed previously . In particular, PS is well characterised as an important marker of ACs for recognition by phagocytes ; it is also enriched on the surface of EVs, so it is likely to also play an important role in ACdEV recognition. In addition to lipid composition, the proteome of ACdEVs needs to be explored further: the surface proteome and what ACdEV–cell interactions occur, how specific molecules are recognised by recipient cells, and what molecules on ACdEVs act as ‘find me’ and/or ‘eat me’ signals. As well as those on the surface of ACdEVs, the molecules inside the vesicles, proteins and nucleic acids, which are transferred into recipient cells, will contribute to their functions. Formation of ACdEVs is a highly regulated process , as is packaging of EV cargoes, with differential sorting of cell components depending on the cell state and phenotype [55–58]. Changes in EV phenotype may occur over the course of apoptosis and there may be different waves of EV release. ACdEVs produced in early and late apoptosis have not been well characterised.
In 1999, Segundo et al.  proposed that ‘apoptotic blebs’ function as attractants to recruit macrophages to sites of cell death; this study that found that many CD molecules (including CD11a, CD21, CD22 and CD54) are rapidly lost from the surface of apoptotic germinal centre B cells, partly via secretion of vesicles, and these ‘blebs’ stimulated chemotaxis in human monocytes. Since then, CX3CL1 and ICAM-3 are the only ACdEV-associated proteins that have been studied and characterised in the context of AC clearance [13,14]. CX3CL1, also known as fractalkine, is a chemokine and intercellular adhesion molecule expressed by various cell types; its receptor, CX3CR1 is expressed by mononuclear phagocytes. In a study on macrophage migration towards apoptotic lymphocytes, migration towards apoptotic Mutu Burkitt's lymphoma (BL) cells was found to be dependent on the interaction of CX3CL1 and CX3CR1; CX3CL1 was shown to be rapidly lost from the surface of Mutu-BL cells following induction of apoptosis, in a 60 kDa cleaved form which was found to be mostly ‘microparticle-associated’ . ICAM-3 is an intercellular adhesion molecule expressed by leukocytes and is involved in immune response signalling. ICAM-3 is also shed from ACs within ACdEVs termed ‘microparticles’ . ICAM-3 on the surface of these microparticles was shown to attract macrophages to sites of leukocyte cell death . It is important to note that ICAM-3 is only expressed by leukocytes, so it is, therefore, not essential for clearance of all ACs and ACdEVs. However, it is possible that apoptosis-induced sub-molecular changes on ICAM-3 (e.g. altered glycosylation) may be occurring on other molecules such that there is functional conservation where ICAM-3 is absent.
Many molecules that are on the surface of, or packaged into, ACdEVs have been identified, which are involved in controlling inflammation and AC clearance, although the ACdEV-associated forms have not been functionally characterised. Annexins are commonly found on the surface of EVs , including annexin 1 which is an anti-inflammatory, pro-resolving molecule . Apoptosis induces the presentation of annexin 1 on the cell surface (co-localisation with PS) , where it exhibits immunosuppressive functionality [59,60]. Annexin 1 promotes polymorphonuclear (PMN) cell apoptosis, recruitment of monocytes and clearance of ACs by macrophages. Annexin 1 has been shown to be present in EVs from healthy activated PMN and mediate their anti-inflammatory effects . Annexin 1 on ACdEVs is, therefore, likely to be an important anti-inflammatory mediator. HMGB1 (high mobility group protein 1) is an important nuclear protein, which is also released from immune cells to mediate immune responses such as chemotaxis and cytokine release . HMGB1 can be pro- or anti-inflammatory depending on its posttranslational modifications; it contains three redox-sensitive cysteines, and modification of these residues alters the activity of the protein . The packaging of HMGB1 into EVs from healthy and ACs (including THP-1-derived macrophages, primary T-lymphocytes, Jurkat T-lymphocytes and HL-60 cells) has been demonstrated in many in vitro studies [63–66]; this has also been demonstrated in vivo in humans . However, characterisation of the specific functions of this molecule in EV-mediated communication has not yet been performed. One study interestingly found that when comparing EV subpopulations from activated and apoptotic T-lymphocytes, HMGB1 was present almost exclusively in large ACdEVs  and this highlights the importance of considering the significant vesicle heterogeneity in studies. There are many possible factors that may mediate the activity of ACdEVs. EVs can also contain active enzymes such as enzymes involved in ECM (extracellular matrix) remodelling — proteases and glycosidases are present on the surface of ‘exosomes’ [68–70]. EVs can also contain enzymes which can process other EV components to generate additional mediators . Therefore, individual enzymatically active EVs may be changing over time in structure, function and perhaps potency (e.g. they may become more immune-modulating over time following enzyme activity). How these activities might affect EV function remains to be determined but will require a pathway analysis of EV composition to assess, for example, if the substrates, enzymes and, if required, transporters (components of the EV surface proteome) of products are present and functional.
Perspectives and future directions
EVs are recognised as significant mediators of intercellular communication involved in physiology and pathophysiology. While EVs produced by healthy and diseased cells have been studied over the last few decades, EVs derived from ACs have not been the main focus of research efforts. Proteomic analysis shows that the compositions of vesicles derived from ACs are distinct from those released by viable cells [64,72], as does RNA profiling [73,74], however, the functional significance of these differences is poorly understood. ACdEVs have been shown to attract phagocytes to sites of cell death and promote AC clearance. However, exactly how they signal in this context is not well characterised and their immunomodulatory functions need to be studied further. A very limited number of ACdEV-associated molecules have been characterised functionally. There are still some general issues in the field of EVs. Methods for isolation and characterisation of EVs vary across the field, as do methods for defining subpopulations and use of nomenclature . Currently, exosomes, MVs and apoptotic bodies are distinguished by size, which is not entirely accurate, or by the presence of proteins that are exclusively found in exosomes. It may or may not be possible in the future to definitively separate and identify subpopulations of EVs. However, at present, it is most appropriate to discuss ACdEVs as a whole. Previous studies on apoptosis and ACdEVs have focused on apoptotic bodies, but it is becoming apparent that the other populations of ACdEVs also play important roles. Whether ACdEVs include exosomes has not been reported and should be confirmed.
Many important questions regarding the structure and functions of ACdEVs need to be addressed. Investigation of ACdEVs, identifying key molecules and their functions will provide valuable insight into how immune responses are modulated, how inflammation is controlled, how ACs are removed, and the processes behind diseases related to dysfunction in these systems. The impact of EVs on recipient cells is also of interest in the context of tissue environments. For example, cell death is known to modulate the tumour microenvironment to support tumour growth  and ACdEVs will likely play a role in this. Such effects have been noted in other systems where EVs from stressed cells can promote greater resilience to stress in neighbouring cells [76,77]. ACdEVs are often considered to be simply smaller, more readily phagocytosed versions of ACs. However, packaging of lipid, protein and nucleic acid mediators into EVs can be specific as well as nonspecific, and these mediators are lost from the cell; ACdEVs may, therefore, not signal and be processed in exactly the same way as ACs. Current research aims to identify specific molecules present on the surface of ACdEVs, how they are recognised by recipient cells and what their effects are. We must also consider the contribution of the whole proteome of these vesicles — what roles do proteins inside of ACdEVs play in mediating communication between apoptotic and other cells? When investigating the precise responses that ACdEVs induce in recipient cells, it is important to look at what kinds of response EVs are produced by recipient cells and how these amplify signals. Furthermore, it is plausible that different waves of ACdEVs are produced over the course of cell death, with different phenotypes and immunomodulatory functions; the changes in the population of ACdEVs over time should, therefore, be explored. The fact that EVs can carry active enzymes may also contribute to a changing population of ACdEVs. Another key question is what enzymes are present in ACdEVs and how does their enzymatic activity change EV structure and function over time? There is a great deal of complexity in this emerging field.
apoptotic cell-associated molecular pattern
apoptotic cell-derived extracellular vesicle
C-X3-C motif chemokine ligand 1
high mobility group protein 1
intercellular adhesion molecule 3
AD and IM are funded by the Biotechnology and Biological Sciences Research Council (BBSRC) (BB/M006298/1 and BB/S00324X/1).
The Authors declare that there are no competing interests associated with the manuscript.