MICROSCOPY RESEARCH AND TECHNIQUE 57:530 540 (2002)

Phagocytosis of Apoptotic Cells by Liver: A Morphological Study

LUCIANA DINI,* PATRIZIA PAGLIARA,
AND

` EMANUELA C. CARLA

Department of Biological and Environmental Science and Technologies, University of Lecce, Lecce, Italy

KEY WORDS

apoptosis; phagocytosis of apoptotic cells; hepatocytes; Kupffer cells; endothelial cells; lymphocytes

ABSTRACT The present review deals with the morphological features of the removal of apoptotic cells by liver. The engulfment of cells undergoing apoptosis can be considered a specialized form of phagocytosis, playing a major role in the general tissue homeostasis in physiological and pathological conditions. In fact, defects of phagocytosis of apoptotic cells might have deleterious consequences for neighboring healthy cells, i.e., pathogenesis of inammatory disease or dysregulation of the immune system. Phagocytosis of apoptotic cells by liver is a complex phenomenon, involving multiple molecular mechanisms of recognition (i.e., lectin-like receptors and receptors for externalized phosphatydilserine) of both parenchymal (hepatocytes) and nonparenchymal (Kupffer and endothelial cells) liver cells, often operating in cooperation. The data discussed in the present review are drawn from studies of phagocytosis of apoptotic cells in the liver, carried out with in vivo and in situ adhesion experiments as well as in vitro assays. Our results indicate that the three main liver cell types (hepatocytes, Kupffer, and endothelial cells) are able to recognize and internalize apoptotic cells by means of specic receptors (galactose and mannose-specic receptor; receptor for phosphatydilserine) and by cytoskeletal reorganization that favors the engulfment of the apoptotic cells. The ags for the identication of apoptotic cells by the liver are modications of the surface of dead cells, i.e., sugar residues and phosphatydilserine exposition. Vitronectin receptor is not involved in such a recognition. The adhesions between modied cell surfaces of apoptotic cells and phagocytes generate cytoplasmatic signaling pathways that drive apoptotic cells to their nal fate within the phagocytes (i.e., lysosomal digestion). Microsc. Res. Tech. 57:530 540, 2002. 2002 Wiley-Liss, Inc. LIVER CELLS PHAGOCYTOSIS OF APOPTOTIC CELLS Although apoptosis occurs at a negligible rate in normal liver, a variety of physiological conditions, diseases, and xenobiotic treatments can cause this form of cell death (Columbano et al., 1985; Bursch et al., 1986, 1992; Tessitore et al., 1989; Grasl Kraupp et al., 1994; Ledda-Columbano et al., 1996). In liver, as in other organs, the execution of apoptosis may be initiated by many different signals, either from within or outside the cell, involving ligand-receptor or TGF-beta/TGFreceptor, or potentially by more unspecic signals such as ceramide or DNA damage. During the modulation/ induction phase of liver apoptosis many different genes, such as p53, c-myc, or Bcl-2/Bax have been shown to be able to shift the balance either to cell survival or cell death (Kanzel and Galle, 2000; Valente and Calabrese, 1999). Therefore, liver apoptotic process can be divided into four phases: an induction phase, the nature of which depends on the specic death-inducing signals; an effector phase, during which the central executioner is activated and the cell becomes committed to die; a degradation phase, during which the cell acquires the biochemical and morphological features of endstage apoptosis. In this cascade of events, the point of no return would be the step at which the cell becomes irreversibly committed to the loss of essential cellular functions. The fourth, and last phase, is the engulfment of the dead corps by macro

phages and other occasional phagocytes. The fact that free or nonphagocytosed dying cells are rarely observed or identied as a frequent physiological event in the liver is the result of the swift in vivo removal into adjacent phagocytic cells (Fig. 1). The rapid ingestion of apoptotic cells is beautifully performed by liver sinusoidal cells (Dini, 2000), thus preventing secondary necrosis and subsequent leakage of potentially harmful materials; this, in turn, limits the potential for inammatory reactions and autoimmune responses. In fact, in spite of the nature of the receptor involved, the molecular mechanism by which apoptotic cells are removed is important in impeding the subsequent proinammatory response (Meagher et al., 1992; Savill et al., 1997; Fadok et al., 1998b; Savill and Fadok, 2000). Specic receptors mediate the particular phagocytic activities of the sinusoidal cells. Among the several alternative mechanisms reported for removal of apoptotic cells, which are mainly related to the cell type and system used (i.e., lectins, thrombospondin (TPS), CD14, scavenger receptors, v 3, CD36, ABC1 (ATP binding Cassette transporter),

*Correspondence to: Prof.ssa Luciana Dini, Department of Biology, University of Lecce, strada prov.le per Monteroni 73100 Lecce Italy. E-mail ldini@ilenic.unile.it Received 20 March 2001; accepted 13 July 2001 DOI 10.1002/jemt.10107 Published online in Wiley InterScience (www.interscience.wiley.com).

2002 WILEY-LISS, INC.

PHAGOCYTOSIS OF APOPTOTIC CELLS

531

Fig. 1. Light (a d) and electron (e,f) micrographs of rat liver 5 days after a single injection of lead nitrate (10 mmoles/100 g b.w.). The intoxication with the heavy metal generates apoptotic cells that are visible as single scattered apoptotic hepatocytes (a,d,e) inside parenchymal liver cells (arrows) and (b) inside the sinusoids (arrows). e: A dying cell embedded in the organ that has lost contact with its neighbors, its microvilli, and any other rufings of the cell surface (arrows). Cellular organelles are well preserved, as is the nuclear envelope. The chromatin is condensed along the nuclear envelope. The entire dying cell, at the later stages of apoptosis, collapses (i.e., decreased volume) and is quickly engulfed by macrophages or by living hepatocytes. c,f: Apoptotic cells phagocytosed by Kupffer cells (arrows). The apoptotic cell is entirely surrounded by tiny protrusions of the liver macrophages, occupying most of the sinusoid (f). Highly condensed chromatin is observed with light as well as electron microscopy. Magnications: (a) 1,000 ; (b,c) 1,200 ; (d) 800 ; (e) 13,000 ; (f) 5,500 .

Ced-6, Ced-7, Ced-5, Ced2, Ced10, DOCK180), in the liver the recognition and phagocytosis of apoptotic cells are mostly performed by means of hepatic lectin-like receptors (Morris et al., 1984; Savill et al., 1990, 1992; Dini et al., 1992; Flora and Gregory, 1994; Ren et al., 1995; Luciani and Chimini, 1996; Devitt et al., 1998; Fadok et al., 1998a; Liu and Hengartner, 1998; Savill, 1998; Wu and Horvitz, 1998; Schlegel et al., 1999; Dini, 2000; Savill and Fadok, 2000). The rst demonstration that liver carbohydrate receptors are involved in the phagocytosis of apoptotic liver cells by healthy ones was performed on newborn hepatocyte cultures induced to undergo apopto-

sis by hormonal treatments (Dini et al., 1992) and further conrmed on normal adult cells (Fig. 2). Inhibition experiments, in which competing saccharides or antibodies to the asialoglycoprotein receptors were added to the culture medium, demonstrated that the hepatocyte recognition and internalization of apoptotic cells is due to the exposition of several glycans (in particular, galactose/Nacetyl-galactosamine) on the surface of apoptotic cells, rendering them available for interaction with lectin-like receptors on hepatocytes (Dini et al., 1992). The vitronectin receptor was not involved in this recognition, since the tetrapeptides Arg-Gly-

532

L. DINI ET AL.

Fig. 2. Normansky light micrographs of cultured isolated hepatocytes treated for 48 hours with 30 g/ml of lead nitrate. The incubation with the heavy metal generates apoptosis in the hepatocytes cultures. In both a and b, micrographs at different stages of the apoptotic process can be recognized: (a) cells are condensing cytoplasm and losing their connections with the Petri dish (arrows).

Round hepatocytes showing condensed nuclei and detaching from the culture dish are observed (b, arrows); (b, double arrow) a oating apoptotic body recognized by a hepatocyte is visible. The nal stage of apoptotic process is the engulfment of apoptotic hepatocyte by the living ones (b arrowheads). Magnications: (a,b) 1,200 .

Fig. 3. In situ adhesion experiments of lymphocytes to sinusoidal liver cells. a: An apoptotic lymphocyte showing an extensive bleb (asterisks) is establishing contact with an endothelial cell (arrow). Note the absolute lack of cellular organelles inside the bleb. c: An endothelial cell is surrounding an apoptotic lymphocyte (asterisks), which occupies the entire sinusoidal lumen. b,d: Micrographs showing phagosomes containing apoptotic lymphocytes (asterisks) at different degrees of digestion. In b, bodies derived from fragmentation of apoptotic cells are shown (arrows). In d, a large phagosome is shown in which chromatin is still recognizable, but not any other organelles. Magnications: (a) 4,500 ; (b) 9,500 ; (c) 8,000 ; (d) 10,500 .

PHAGOCYTOSIS OF APOPTOTIC CELLS

533

a net negative charge by means of scavenger receptors and receptors specic for galactose and mannose residues (Steer and Clarenburg, 1979; Kolb-Bachofen et al., 1982; Steffan et al., 1986; Praaning-van Dalen et al., 1987; Van Berker et al., 1992). Due to their location in the sinusoids and combined with the fact that they represent the majority of the bodys xed macrophages, Kupffer cells are the rst cells of the mononuclear phagocyte system to come into contact with particulate and immunoreactive materials coming from the blood. Therefore, together these properties enable them to exert an efcient and fast recognition and engulfment of apoptosing cells. In fact, they are primary actors in clearing potentially noxious materials like apoptotic cells. MORPHOLOGICAL ASPECTS OF APOTOTIC CELL PHAGOCYTOSIS Here the morphological aspects of recognition and engulfment of apoptotic cells by hepatocytes, endothelial, and Kupffer cells are highlighted. The data reported below have been obtained from observations with light and electron microscopy of in vivo, in situ, and in vitro experiments. Briey, the experiments were carried out as follows. In Vivo Induction of Apoptosis Inbred 5-week-old male Swiss mice (20 30 g) or male Wistar rats (150 200 g) purchased from Morini (Reggio Emilia, Italy) were used. All the animals, maintained on a 12-hour daynight rhythm, with free access to water and food (standard diet), received humane care and the study protocols complied with national laws. Before surgery, animals were anesthetized by an i.p. injection with Farmotal (Farmitalia, Italy) 10 mg/ 100 g body weight. Lead nitrate (Carlo Erba, Milano, Italy), dissolved in distilled water, was injected i.v. at a dose of 10 mmoles/100g body weight. The animals were killed at various time intervals after treatment. Control rats received an equivalent amount of NaCl 0.9%. In Situ Adhesion Experiments Livers were perfused in a nonrecirculating system at a ow rate of 1 ml/min. Apoptotic lymphocytes 1 106 labeled with Hoechst 33342 were injected through the portal vein into liver circulation. Peripheral blood mononuclear cells were obtained after Ficoll gradient separation of buffy coats from blood donations of nonsmoking healthy males, age 25 45 years. Lymphocytes, separated from monocytes by double adherence to plastic and maintained at a cell density of 1 106 cells/ml in complete culture medium, were used on the rst day of explant. Apoptosis, induced with cycloexhimide (CHX) 10-2 M for 18 hours, followed by 1-hour recovery in fresh medium, 10 g/ml puromycin (PMC), or by keeping lymphocytes in water baths equilibrated to 43C for 1 hour followed by 8 hours of recovery at 37C was evaluated by ow cytometry, by light microscopy on cytospin, and by electron microscopy. Modications of the cell surface were evaluated by using a broad panel of uorescent lectin conjugates with different specicities: Concanavalin-A (Con-A) and succinyl Concanavalin-A (sCon-A) (a-D-mannosyl); Phaseoulus limensis (PHAE) (N-acetyl-D-galactosamine); Ricinus communis (RCA) (D-Galactosyl); Ulex europaeus (UEA)

Fig. 4. Phagosomes containing apoptotic cells can present different morphologies that are related both to digestion as well as to the type of particle ingested (i.e., apoptotic cell, apoptotic body with or without condensed chromatin). a: Two phagosomes are inside an endothelial cell: one still shows chromatin (arrows), while the other (asterisks) is at the very late stage of digestion. b: A large phagosome inside a Kupffer cell containing remnants of one unfragmented apoptotic cell. Condensed chromatin is still present associated with a nucleolus (arrows); cytoplasmatic organelles are not more morphologically recognizable. Magnications: (a) 12,500 ; (b) 21,000 .

Asp-Ser (RGDS) and Arg-Gly-Glu-Ser (RGES) fail to exert any inhibitory action (Dini et al., 1992). Sinusoidal liver cells are able to clear from the blood galactose and mannose-terminated particles (even those of large size) and a wide range of molecules with

534

L. DINI ET AL.

(a-L-Fucosyl); Triticum vulgare (WGA) (N-acetyl-Dglucosamine); Dolichos biorus (DBA) (N-acetylD-galactosamine); Pisum sativum (PSA) (fucosyl residues); Arachis hypogaea (PNA) (N-acetyl-D-galactosamine); Limulus polyphemus (LPA) (N-acetyl-D-

galactosamine, N-acetyl-D-glucosamine, N-acetylneuramicic acid). Samples were processed for light and electron microscopy to assess the adhesion of lymphocytes to the sinusoidal wall. Adhesion specicity was tested in parallel inhibition experiments by adding 80 mM (nal concentration) of N-acetyl-D glucosamine (GlcNAc) and N-acetyl-D-galactosamine (GalNAc) into the perfusion tube before lymphocyte administration. In Vitro Adhesion Experiments Mouse liver sinusoidal cells (endothelial and Kupffer cells) were isolated by enzymatic (0.05% D-Collagenase and 0.1% Pronase), perfusion of livers and separated from parenchymal and blood cells through centrifugation in a 30% metrizamide gradient. Cells were plated onto type I collagen-coated 24-well plates at a concentration of 1 106 cells/ml/well in DMEM medium. Normal or apoptotic 5 105 lymphocytes labeled with Hoechst 33342 were added to 24-well-plate cultured sinusoidal liver cells and incubated for 20, 60, and 120 minutes. Sinusoidal liver cells were incubated in the presence of 1 ng/ml lypopolysaccharide (LPS) for 6 hours or 300 ng/ml recombinant human interleukin 1 (rhIL-1 ) for 4 hours before addition of lymphocytes. Inhibition experiments were performed, incubating the cells with a solution of galactose, N-acetyl-galactosamine, and mannose (Gal/GlcNAc/Man) (nal concentration of sugar cocktail 80 mM) for 20 minutes at 37C before incubation with apoptotic lymphocytes. The number of adhering cells was determined by a uorescence measurement system. In Vivo Studies In spite of the complexity of the phenomenon, the investigation of liver apoptosis in vivo or as a whole in situ led to study of different aspects of the process at the same time. In particular, taking into account the advantage of the simultaneous presence of dying and healthy cells in the same sample, the distribution, the morphology of dying liver cells, and the recognition activities of the healthy ones were investigated. By using an in vivo model of liver apoptosis (Columbano et al., 1985), the rapid removal from the tissue of apoptotic liver cells was clearly shown. Dead cells were also frequently phagocytosed by nonprofessional phagocytes. In fact, dying hepatocytes, hampered in reaching the circulation, were engulfed by the neighboring ones (Fig. 1). These data are in line with others in the literature, describing (even in invertebrates, i.e., Caenorhabditis elegans) examples of nonprofessional phagocytes phagocyting dying cells (Ellis et al., 1991).

Fig. 5. In situ adhesion experiments of apoptotic lymphocytes to endothelial (a,c) and Kupffer cells (b,c). The rst step for phagocytosis of apoptotic cells is their recognition and blockade by the sinusoidal wall. Circulating apoptotic lymphocytes injected through the hepatic circulation are retained by both endothelial (a,c, arrows) as well as Kupffer cells (b,c, arrowheads); however, the percentage of Kupffer cells with phagosomes containing apoptotic cells and/or remnants of apoptotic cells is always higher than endothelial cells, thus conrming the well-known phagocytic activity of liver macrophages. In c a large Kupffer cell surrounding an apoptotic cell with a recognizable fragmented nucleus is shown (arrowhead). Magnications: 1,500 .

PHAGOCYTOSIS OF APOPTOTIC CELLS

535

Fig. 6. Lymphocytes stained with Con-A-FITC for sugar residues exposure. a: Control (untreated) cells were unstained. b, c,d: cells treated with PMC 10 g/ml for 2, 4, and 6 hours respectively displayed brightly marked membranes that increase with time of treatment. Magnications: 1,200 .

Inammatory injury in the liver parenchyma was never observed. Hepatocytes surrounding the apoptotic cells showed normal cytoplasm without any signs of organelle swelling and/or degradation (Fig. 1). Therefore, the protective role of the phagocyte recognition of apoptotic self by preventing the leakage of noxious substances and by limiting the development of autoimmune responses was successfully conrmed in the liver. The increased number of apoptotic cells, produced by lead nitrate treatment and mainly poured into the blood, induces sinusoidal liver cells (i.e., Kupffer and endothelial cells) to actively phagocyte both apoptotic hepatocytes and circulating apoptotic cells, and also those derived from other body areas (Figs. 3, 4). The peak of phagocytic activity of Kupffer cells (3-fold the control) was measured at 5 and 15 days from lead nitrate injection (Ruzittu et al., 1999; Dini and Carla, ` 1998; Dini, 2000), thus conrming the capacity (in particular for the interaction with particulate materials) of the hepatic sinusoidal wall to operate as a protective barrier for the systemic circulation (Steffan et al., 1986; Wardle, 1987; Dini and Kolb-Bachofen, 1989; KolbBachofen, 1992; Toth and Thomas, 1992).

In Situ Adhesion Experiments To further explore the liver phagocytosis of apoptotic cells, in situ adhesion experiments were carried out. With this type of experiment the overall recognition capacity of the sinusoidal wall can be assayed. In the experimental protocol, liver blood is replaced by culture medium containing apoptotic cells. In particular, apoptotic lymphocytes were used, due to the fact that the liver is the specialized site where T cells undergoing apoptosis in vivo are eliminated. However, the molecular mechanism(s) that control this accumulation is still unknown (Huang et al., 1994). Once injected into the mouse hepatic circulation, apoptotic lymphocytes, but not normal ones, are efciently removed by sinusoidal cells by means of carbohydrate receptors, as conrmed by inhibition studies (Ruzzittu et al., 1999; Dini, 2000). The amount of retained lymphocytes is strictly dependent on the number of exposed binding sites on the cell surface of sinusoidal liver cells. In agreement with this are data showing that apoptotic lymphocytes retained by sinusoidal cells of the periportal tract are double those retained in the perivenous region. In fact, the number of carbohydrate receptors expressed on cell surface of endothelial cells from the

536

L. DINI ET AL.

periportal tract is double those quantied on endothelial cells of the perivenous tract (Dini and Carla, 1998). ` The immediate fate of apoptotic lymphocytes after injection into liver circulation, independent of the modality of induction of apoptosis (i.e., by oxidative, hyperthermic stress, or drugs) is their absorption onto endothelial and Kupffer cells (Fig. 5), thus indicating that liver recognition of apoptotic cells is a fundamental step for their subsequent sequestration, internalization, and digestion. Even if the ability to block apoptotic cells is similar for both endothelial and Kupffer cells, these latter cells show a higher rate of apoptotic cell engulfment, Kupffer cells being more active and faster than endothelial cells. The very rapid ingestion of apoptotic cells, which occurs immediately after their binding to Kupffer cell surfaces, was repeatedly observed in in situ adhesion experiments, as well as in in vitro adhesion experiments. The recognition process of apoptotic cells, as extensively reported in literature, is triggered by modications of the surface of the dead cells (Platt et al., 1998; Ren and Savill, 1998). Modications of the glycidic residues of glycoproteins of the plasma membrane of apoptotic lymphocyte cell surfaces are major candidates as the eat me signal. Substantial changes of exposed sugar residues were observed on apoptotic lymphocyte surfaces when compared to normal cells (Falasca et al., 1996; Dini, 2000). It is still unknown how these cell surface carbohydrate modications could occur, but they appear to be a common mechanism for recognition of unwanted cells by the liver (i.e., the removal of aged erythrocytes by the liver; Kolb et al., 1981). In addition, progressive glycan modications (Fig. 6) are achieved in parallel with the morphological modications that characterize apoptotic cells. Therefore, taking into account the progressive surface modications, the execution process of apoptosis is divided into three stages, each characterized by qualitative and quantitative modications of the cell surface: early, mature, and late/necrotic. Even if apoptosis is not a synchronized phenomenon, it is nevertheless possible to obtain a cell population enriched for each stage of apoptosis by using time-course experiments. Enriched cell suspension of early, mature, or late/necrotic apoptotic cells were used in in vitro and in situ adhesion experiments. The mature apoptotic cell population was quickly recognized by sinusoidal cells when compared to the recognition time of early or late/ necrotic apoptotic cells (manuscript in preparation). In Vitro Phagocytosis Experiments Liver cells (i.e., hepatocytes, Kupffer cells, endothelial cells, as well as pit and fat storing cells) can be dissociated, puried, and maintained in suspension or in adhesion cultures for some time (depending on the cell type). Therefore, isolated liver cells, and sinusoidal cells in particular, are useful tools for studies of phagocytosis of apoptotic cells. Hepatocytes maintained in adhesion cultures for a short time do not show signicant rates of proliferation and apoptosis unless they are treated with specic proliferative or apoptotic inducers, i.e., retinoic acid and estrogens. Considering that under the above conditions the apoptotic rate is about 30%, it is possible, by means of microscopy analysis, to discriminate in the

Fig. 7. Transmission electron micrographs of the interaction between apoptotic lymphocytes and cultured human Kupffer cells at different stage of interaction. Apoptotic lymphocytes when incubated with Kupffer cells at 37C are promptly bound (a) and then phagocytosed (b). An apoptotic lymphocyte (asterisk) whose chromatin begins to aggregate into dense masses adhering closely to the plasma membrane of a human Kupffer cell (Kc) at 5 min of incubation. Within 5 minutes of co-culture almost all the apoptotic lymphocytes are bound to the plasma membrane of Kupffer cells, while after 10 minutes of incubation the majority of apoptotic cells are internalized by the Kupffer cells, thus suggesting a very rapid mechanism of recognition (b). Phagosomes, containing dark material, which represent residual of the partially digested apoptotic lymphocytes, are visible inside Kupffer cells (b, arrow). Magnications: (a) 7,500 ; (b) 4,500 .

PHAGOCYTOSIS OF APOPTOTIC CELLS

537

Fig. 8. Scanning electron micrographs of cocultures of apoptotic lymphocytes and human Kupffer cells. a: Human Kupffer cells are characterized by prominent membrane rufing with microvilli of variable length accompanied by numerous pseudopodia when cultured in normal conditions. Conversely, apoptotic cells are recognized by their round and smooth surface that is a consequence of the disappearance of microvilli during the apoptotic process. a: Apoptotic lymphocytes added to the culture medium adhere to the surface of the Kupffer cells (arrow). b: A Kupffer cell at the beginning of the phagocytic process: the macrophage is tethering the dead corpse (arrows). c: A few minutes, later Kupffer cells, which are very active in phagocytosis, have completely internalized the apoptotic lymphocytes. After 15 minutes of co-culture round protrusions (representing the internalized apoptotic lymphocytes) are often visible inside the cells (arrow). When Kupffer cells were incubated with the carbohydratespecic receptor inhibitors (i.e., sugars or modied glycoproteins), before and during the incubation with apoptotic lymphocytes, their phagocytic activity is dramatically reduced. The addition of healthy lymphocytes to the Kupffer cell cultures does not result in the recognition and internalization of the blood cells. Magnications: (a) 5,500 ; (b) 11,000 ; (c) 10,000 .

same culture dish apoptotic and healthy hepatocytes and to verify the phagocytic ability of the latter for apoptotic cells. The micrographs in Figure 2 show hepatocyte cultures treated with hormones. Many scattered apoptotic cells are observed: some of them are detaching from the culture dish, others are undergoing recognition before engulfment, and others have been internalized by the healthy hepatocytes and are visible as membrane-enclosed phagosomes (Fig. 2). All these data support the idea that hepatocytes are able to internalize apoptotic cells when necessary. In vitro adhesion and uptake experiments were performed by using cultures of isolated and puried endothelial and Kupffer cells. Sinusoidal liver cells were incubated with apoptotic lymphocytes at different times (Figs. 79). As mentioned above, lymphocytes were chosen because in vivo they are a physiological source of apoptotic cells/bodies recognized and phagocytosed by liver cells. In fact, in vivo apoptotic lymphocytes are recognized and phagocytosed well before the nal stages of DNA degradation and cell lysis (Pradhan

et al., 1994; Huang et al., 1994). Kupffer and endothelial cells in culture phagocyte in a very efcient manner lymphocytes undergoing apoptosis induced by different stimuli (heat-shock 43C; cycloheximide), but not normal living ones (Dini and Carla, 1998; Dini, 2000) ` (Figs. 79). Since endocytosis is a multistep process that includes cellular movements, in particular the extension of pseudopodia, cytoskeletal integrity must be important. In fact, a relationship between Kupffer cell shape and phagocytic activity has been recently reported (Dini et al., 1998). To accomplish phagocytosis of apoptotic cells, the recognition process must be followed by internalization. This latter phenomenon requires cytoplasmic movements that generate ne lamentous processes immediately adjacent to the apoptotic lymphocyte (Fig. 8). In the meantime, the internalization of apoptotic cells requires the recruitment of cell-surface receptors on the extending pseudopodia into positions in which they can interact with the appropriate ligands. In fact, we repeatedly found that phagocytosis is inhibited by the presence in the

538

L. DINI ET AL.

Fig. 9. Scanning electron micrographs of in vitro (a) and in situ (b) adhesion experiments. Co-cultures of apoptotic lymphocytes and endothelial cells is shown in a. After 15 minutes of incubation apoptotic lymphocytes were observed in the process of binding to the surface of the endothelial cells (the fenestrae are clearly visible) (arrows). b: Apoptotic lymphocytes (arrows) engulfed by the sinusoidal wall of mouse liver after intraportal injection. Magnications: (a) 16,000 ; (b) 9,000 .

culture medium of inhibitors of galactose- and mannose-specic receptors (i.e., sugar residues both as single moieties or as cocktail and desialylated glycoproteins, but not by unmodied ones) and to a lower extent by desialylated glycoproteins, but not by unmodied glycoproteins (Dini, 2000). A difference in phagocytic activity is easily observed between isolated endothelial and Kupffer cells, the latter being much more active than endothelial cells. The recognition of the apoptotic lymphocytes once added to human Kupffer cell cultures is almost entirely completed within a few minutes of incubation and the apoptotic cells are detected as dark material inside large phagosomes (Fig. 7). On the other hand, endothelial cells need more time to complete engulfment of apoptotic lymphocytes. One explanation, of course, is related to the different functions in the liver of endothelial and Kupffer cells, that being that macrophages are characterized by high phagocytic activity. However, it is worth noting that apoptotic recognition may be regulated by the state of the phagocyte and by external inuences (Savill et al., 1993). The exposure of phagocytes to cytokines known to be present at inammation sites (i.e., granulocyte-macrophage colony stimulating factor; GM-CSF) or implicated in the repair of injured tissue (i.e., transforming growth factor, TGF platelet-derived growth factor, PDGF) and those involved in the initiation of inammation (i.e., interferon gamma, IFN interleukin-1 , IL-1 , and tumor necrosis factor : TNF ) increased the recognition of apoptotic human neutrophils (Savill et al., 1993). LPS and IL1 upregulate the mannose receptor expression of liver cells and consequently the phagocytic activity of sinusoidal cells (Dini et al., 1995). CONCLUDING REMARKS This brief discussion of the recognition and ingestion of apoptotic cells by hepatocytes, Kupffer, and endothelial cells shows clearly that liver cells are active participants in the removal of apoptotic cells and that this removal is swift and efcient despite its complexity. To achieve recognition of apoptotic cells, signals in the form of molecular modications of the plasma membrane must occur on the dying cell surface (i.e., modication of the membrane lipid asymmetry, external

exposition of phosphatidylserine, and normally hidden sugar residues). The morphological study of apoptotic cells is not sufcient to detect the very early stages of the process, those characterized by plasma membrane modications without visible nuclear modications. Conversely, it is very useful for the detection of apoptotic cells containing phagosomes, especially when dying cells have been labeled with uorescent dyes or electron-dense markers. On the other hand, due to the characteristic chromatin condensation and round shape of the apoptotic cells, the phagocytosis of the mature/late stages of apoptotic cells is easily studied with both light and electron microscopy. The adhesion to the apoptotic cell is the rst step that allows engulfment of dead cells and this in turn allows apoptotic cells to reach their nal fate within the phagocytes. The recognition of dead cells could be a multistep process complicated by the existence of regional specialization and by the display on the apoptotic cells of multiple signals to increase the probability of their removal and consequently the safety of the whole organism. To engulf the apoptotic cells, cytoskeletal reorganization is also necessary, as shown by the dramatic modication of the phagocytic cellular shape. In addition, as reported for the liver, cooperation among different cellular types sharing the same receptor system is shown for the removal of apoptotic cells. In fact, hepatocytes, Kupffer, and endothelial cells operate, at the same time, in the plasma clearance of apoptotic cells generated during the involuting phase of liver hyperplasia induced by a single injection of lead nitrate by means of a sugar recognition mechanism (Dini et al., 1995; Ruzzittu et al., 1999). These data, together with the fact that the phagocytic activity in endothelial cells can be enhanced in macrophage-depleted rats (Bogers et al., 1991) and that IL-1 induces in vitro overexpression of mannose-specic receptors on endothelial cells, further support the idea of cooperation among liver cells during phagocytosis of apoptotic cells (Dini et al., 1995; Dini et al., 1998). However, the process of phagocytosis of apoptotic cells, which is an ancient process present in invertebrates as well as in vertebrates, has developed species-specic mechanisms whose biological signicance is still obscure.

PHAGOCYTOSIS OF APOPTOTIC CELLS

539

It is worth noting that the study of phagocytosis during the process of apoptosis is not merely a speculative exercise, since defects of phagocytosis of apoptotic cells might have deleterious consequence for neighboring healthy cells. The logical consideration of the importance of phagocytosis leads to thoughts on the contribution of defective clearance as a factor in the pathogenesis of inammatory diseases. The relevance of phagocytosis to the dysregulation of the immune system that underlies specic pathological conditions requires examination: for example, whether compromising the capability to ingest apoptosing cells contributes to autoantibody production (Bellone et al., 1997; Botto et al., 1998; Hermann et al., 1998). The studies of mutations affecting the clearance of dying cells by professional phagocytes in Drosophila will help to unravel the complexity inferred from inhibitor studies in mammalian systems. However, further investigations of the mechanisms of recognition and ingestion of apoptotic cells are urgently required to address regulatory roles in inammation, immune responses, and tissue remodeling. This in turn may allow manipulation of phagocyte responses to apoptotic cell stimuli and the development of novel therapeutic strategies (for example, during tissue repair) as an effective antiinammatory and immunosuppressive strategy (Voll et al., 1997; Fadok et al., 1998; Botto et al., 1998; Herrmann et al., 1998). REFERENCES
Bellone M, Iezzi G, Rovere P, Galati G, Ronchetti A, Protti MP, Davoust J, Rugarli C, Manfredi AA. 1997. Processing of engulfed apoptotic bodies yields T cell epitopes. J Immunol 159:53915399. Bogers WM, Stad RK, Janssen DJ, Prins FA, Van Rooijen N, Van Es LA, Daha MR. 1991. Kupffer cell depletion in vivo as results in clearance of large-sized IgA aggregates in rats by liver endothelial cells. Clin Exp Immunol 85:128 136. Botto M, DellAgnola C, Bygrave AE, Thompson EM, Cook HT, Petry F, Loos M, Pandol PP, Walport MJ. 1998. Homozygous C1q deciency causes glomerulonephritis associated with multiple apoptotic bodies. Nat Genet 19:56 59. Bursch W, Dusterberg B, Schulte-Hermann R. 1986. Growth, regression and cell death in rat liver as related to tissue levels of the hepatomitogen cytoproterone acetate. Arch Toxicol 59:221227. Bursch W, Oberhammer F, Schulte-Hermann R. 1992. Cell death by apoptosis and its protective role against disease. Trends Pharmacol Sci 13:245251. Columbano A, Ledda-Columbano GM, Coni P, Faa G, Liguori C, Santacruz G, Pani G. 1985. Occurrence of cell death (apoptosis) during the involution of liver hyperplasia. Lab Invest 52:670 677. Devitt A, Moffatt OD, Raykundalia C, Capra JD, Simmons DL, Gregory CD. 1998. Human CD 14 mediates recognition and phagocytosis of apoptotic cells. Nature 392:505508. Dini L. 2000. Recognizing death: liver phagocytosis of apoptotic cells. Eur J Histochem 44:217227. Dini L, Carla EC. 1998. Hepatic sinusoidal endothelium heterogene` ity with respect to the recognition of apoptotic cells. Exp Cell Res 240:388 393. Dini L, Kolb-Bachofen V. 1989. Preclustered receptor arrangement is a prerequisite for galactose-specic clearance of large particulate ligands in rat liver. Exp Cell Res 184:235240. Dini L, Autuori F, Lentini A, Oliverio S, Piacentini M. 1992. The clearance of apoptotic cells in the liver is mediated by the asialoglycoprotein receptor. FEBS Lett 296:174 178. Dini L, Lentini A, Diez Diez G, Rocha M, Falasca L, Serano L, Vidal-Vanaclocha F. 1995. Phagocytosis of apoptotic bodies by liver endothelial cells. J Cell Sci 108:967973. Dini L, Ruzittu M, Carla EC, Falasca L. 1998. Relationship between ` cellular shape and receptor-mediated endocytosis: an ultrastructural and morphometric study in rat Kupffer cells. Liver 18:99 109. Ellis RE, Yuan J, Horvitz HR. 1991. Mechanisms and functions of cell death. Annu Rev Cell Biol 7:663 698.

Fadok VA, Bratton DL, Frasch SC, Warner ML, Henson PM. 1998a. The role of phosphatidylserine in recognition of apoptotic cells by phagocytes. Cell Death Differ 5:551562. Fadok VA, Bratton DL, Konoval A, Freed PW, Westcott JY, Henson PM. 1998b. Macrophages that have ingested apoptotic cells in vitro inhibit proinammatory cytokine production through autocrine/ paracrine mechanisms involving TGF- , PGE2, and PAF. J Clin Invest 101:890 898. Flora PK, Gregory CD. 1994. Recognition of apoptotic cells by human macrophages: inhibition by a monocyte/macrophage-specic monoclonal antibody. Eur J Immunol 24:26252632. Grasl Kraupp B, Bursch W, Ruttkay Nedecky B, Wagner A, Lauer B, Schulte-Hermann R. 1994. Food restriction eliminates preneoplastic cells through apoptosis and antagonizes carcinogenesis in rat liver. Proc Natl Acad Sci USA 91:99959999. Herrmann M, Voll RE, Zoller OM, Hagenhofer M, Ponner BB, Kalden JR. 1998. Impaired phagocytosis of apoptotic cell material by monocyte-derived macrophages from patients with systemic lupus erythematosus. Arthritis Rheum 41:12411250. Huang L, Soldevila G, Leeker M, Flavell R, Crispe N. 1994. The liver eliminates T cells undergoing antigen-triggered apoptosis in vivo. Immunity 1:741749. Kanzel S, Galle PR. 2000. Apoptosis and the liver. Semin Cancer Biol 10:173 84. Kolb H, Friedrick E, Suss R. 1981. Lectin mediates homing of neuraminidase-treated erythrocytes to the liver as revealed by scintigraphy. Hoppe-Seylers Z Physiol Chem 362:1609 1614. Kolb-Bachofen V. 1992. A review on the biological properties of C-reactive protein. Immunobiology 183:133145. Kolb-Bachofen V, Schlepper-Schafer J, Vogell W. 1982. Electron microscopic observations of the hepatic microscopic evidence for an asailoglycoprotein receptor on Kupffer cells: localization of lectin mediated endocytosis. Cell 29:859 866. Ledda-Columbano GM, Shinozuka H, Katyal SL, Columbano A. 1996. Cell proliferation, cell death and hepatocarcinogenesis. Cell Death Differ 3:1722. Liu QA, Hengartner MO. 1998. Candidate adaptor protein CED-6 promotes the engulfment of apoptotic cells in C. elegans. Cell 93: 961972. Luciani MF, Chimini G. 1996. The ATP binding cassette transporter ABCD1, is required for the engulfment of corpses generated by apoptotic cell death. EMBO J 15:226 235. Meagher LC, Savill JS, Baker A, Fuller R, Haslett C. 1992. Phagocytosis of apoptotic neutrophils does not induce macrophage release of thromboxane B2. J Leuk Biol 52:269 273. Morris RG, Hargreaves AD, Duvall E, Wyllie AH. 1984. Surface changes in thymocytes undergoing apoptosis. Am J Pathol 115: 426 436. Platt N, Pedro da Silva R, Gordon S. 1998. Recognizing death: the phagocytosis of apoptotic cells. Trends Cell Biol 8:365372. Praaning-van Dalen DP, de Leeuw AM, Brouwer A, Knook DL. 1987. Rat liver endothelial cells have a greater capacity than Kupffer cells to endocytose N-acetylglucosamine- and mannose-terminated glycoproteins. Hepatology 7:672 679. Pradhan D, Williamson P, Schlegel RA. 1994. Phosphatidylserine vesicles inhibit phagocytosis of erythrocytes with a symmetric transbilayer distribution of phospholipids. Mol Membr Biol 11:181 187. Ren V, Savill J. 1998. Apoptosis: the importance of being eaten. Cell Death Differ 5:563568. Ren V, Silverstein RL, Allen J, Savill J. 1995. CD36 gene transfer confers capacity for phagocytosis of cells undergoing apoptosis. J Exp Med 181:18571862. Ruzittu M, Carla EC, Montinari MR, Maietta G, Dini L. 1999. Mod` ulation of cell surface expression of liver carbohydrate receptors during in vivo induction of apoptosis with lead nitrate. Cell Tissue Res 298:105112. Savill JS. 1997. Recognition and phagocytosis of cells undergoing apoptosis. Br Med Bull 53:491508. Savill JS. 1998. Phagocytic docking without shocking. Nature 392: 442 443. Savill JS, Fadok V. 2000. Corpse clearance denes the meaning of cell death. Nature 407:784 788. Savill J, Dranseld L, Hogg N, Haslett C. 1990. Vitronectin receptormediated phagocytosis of cells undergoing apoptosis. Nature 343: 170 173.

540

L. DINI ET AL.
Tessitore L, Valente G, Bonelli G, Costelli P, Baccino FM. 1989. Regulation of cell turnover in the livers of tumor bearing rats: occurrence of apoptosis. Int J Cancer 44:697700. Toth CA, Thomas P. 1992. Liver endocytosis and Kupffer cells. Hepatology 16:255266. Valente M, Calabrese F. 1999. Liver and apoptosis. Ital J Gastrenterol Hepatol 31:7377. Van Berkel TJC, De Rijke JB, Kruijt JK. 1992. Recognition of modied lipoprotein by various scavenger receptors on Kupffer and endothelial liver cells. In: Windler E, Greten H, editors. Hepatic endocytosis of lipids and proteins. Munchen, FRG: Zuckschwerdt Verlag. p 443. Voll RE, Hermann M, Roth EA, Stach C, Kalden JR, Girkontaite I. 1997. Immunosuppressive effects of apoptotic cells. Nature 390: 350 351. Wardle EM. 1987. Kupffer cells and their function. Liver 7:6370. Wu YC, Horvitz HR. 1998. C. elegans phagocytosis and cell-migration protein CED-5 is similar to human DOCK 180. Nature 392:501504.

Savill J, Hogg N, Ren Y, Haslett C. 1992. Thrombospondin cooperates with CD36 and the vitronectin receptor in macrophage recognition of neutrophils undergoing apoptosis. J Clin Invest 90:15131522. Savill J, Fadok V, Henson P, Haslett C. 1993. Phagocyte recognition of cells undergoing apoptosis. Immunol Today 14:131136. Schlegel RA, Krahling S, Callahan MK, Williamson P. 1999. CD14 is a component of multiple recognition systems used by macrophages to phagocytose apoptotic lymphocytes. Cell Death Differ 6:583592. Steer CJ, Clarenburg R. 1979. Unique distribution of glycoprotein receptors on parenchymal and sinusoidal cells of rat liver, J Biol Chem 254:4457 4461. Steffan AM, Gendrault JL, McCuskey RS, McCuskey PA, Kirn A. 1986. Phagocytosis, an unrecognized property of murine endothelial liver cells. Hepatology 6:830 836. Stern M, Meagher L, Savill J, Haslett C. 1992. Apoptosis in human eosinophils. Programmed cell death in the eosinophil leads to phagocytosis by macrophages and is modulated by IL-5. J Immunol 148:35433549.