Reprints
PubMed
CAVEAT LECTOR
Alberto Ortiz, Silvia González Cuadrado, Corina Lorz, Jesús Egido.
Division of Nephrology. Fundación Jiménez Díaz, Universidad Autónoma, Madrid, Spain.
Received 12/01/95; Accepted 01/26/96; On-line 03/01/96
3. APOPTOSIS AND PROGRAMMED CELL DEATH
The functional concept of programmed cell death implies an active participation of the cell in its own death (cell suicide) through the activation of a genetic program (5-7). In general, programmed cell death has the morphologic characteristics of apoptosis, although there are exceptions (7). In fact, there is functional, morphologic, and genetic evidence of heterogeneity in this process (6-9), and there are unanswered questions about the physiologic relevance of this diversity. In any case, the pragmatist may define programmed cell death (and because of the extensive use of the term, also apoptosis) as a process that can be modulated through interference with cell death related genes, independently of the morphology or pattern of DNA degradation. Thus, the main characteristic of apoptosis/programmed cell death would be its susceptibility to therapeutic intervention.
Apoptosis, however, is usually defined by a characteristic morphology and functional changes (10-12). Apoptotic cells display decreased cell and nuclear size with chromatin condensation, detachment from adjacent cells, cell membrane blebbing, and fragmentation with the formation of membrane-bound bodies (10).
Functionally, apoptotic cells express new cell membrane structures that determine a high rate of recognition and phagocytosis by adjacent cells. The integrity of the cell membrane is preserved for some time and cells are cleared before there is any significant leakage of pro-inflammatory molecules. The half-life of the apoptotic cell is a few hours (13, 14). As a consequence, a low percentage of apoptotic cells visible in a tissue section may be associated with a significant loss of cell mass (13). Apoptosis usually affects individual cells and its tissue distribution is patchy and asynchronous. Detection of apoptosis is further impaired in renal injury by the fact that detached apoptotic cells may be flushed away by urine (15).
Apoptosis requires the expression or suppression of certain genes (5). Endonucleases, tissue transglutaminases and proteases are activated (5,11,12). Some kinds of apoptotic cell death can occur in the absence of a nucleus if the required genetic machinery is constitutively expressed (16). In this regard, the inhibition of mRNA or protein synthesis may cause or prevent apoptosis (12), depending on the balance between lethal and protective factors in a given cell.
By contrast, necrosis is a passive mode of cell death that frequently involves fields of contiguous cells, and a prominent inflammatory response. Both apoptosis and necrosis can occur at the same time in the same tissue (17). The occurrence of either one may depend on the intensity of the precipitating events (18).
3.1 EXTRACELLULAR FACTORS IN THE REGULATION OF APOPTOSIS
Apoptosis may be the consequence of withdrawal of survival factors or exposure to lethal factors (1,19). The survival factor requirement may vary with cell type, functional status of the cell, or presence of lethal stimuli. Survival and death factors for extrarenal cells include cytokines, extracellular matrix, lipids, small molecules and microbial products. In addition, physical factors (such as heat, irradiation) and certain drugs may induce apoptosis.
Serum deprivation results in apoptosis of mesangial and tubular epithelial cells (20-22). Little is known about the specific factors that account for survival of renal cells. EGF prevents apoptosis in proximal tubular cells (14), and IGF-1 and bFGF in mesangial cells (23,24). By contrast, a survival activity could not be demonstrated for PDGF-BB and EGF in mesangial cells (24).
Extracellular matrix is a survival factor for epithelial (25) and mesangial cells (26), especially if soluble survival factors are not available. This effect is mediated by integrin receptors (26). The nature of the extracellular matrix is important. While basement membrane supports the survival of mesangial cells, type I collagen (found in sclerotic but not in normal glomeruli) does not (26).
Several cytokines and inflammatory mediators induce apoptosis. The cytotoxic effect of TNF-alpha on glomerular epithelial and mesangial cells in culture (27) has recently been characterized as apoptosis (20,28). IL-1-alpha is also lethal for mesangial cells (28). However, neither TNF-alpha nor IL-1-alpha induce glomerular endothelial cell apoptosis (28). Fas activation, oxygen radicals and anti-Thy1 antibodies also induce apoptosis of mesangial cells (29-31). TNF-alpha, Fas activation, inhibition of protein kinase C, nephrotoxins and ceramide induce apoptosis of tubular cells (18,21,32 and unpublished observation). Ceramide is a mediator of TNF-alpha and Fas-induced apoptosis (33). In tubular cells DNA degradation has been dissociated form the morphologic features of apoptosis (34)
3.1.3 Interaction of survival and lethal factors
Cell fate depends on the interaction of survival factors and apoptosis-inducing factors. In oligodendrocytes, ciliary neurotrophic factor (CNTF) prevents both growth factor deprivation-induced apoptosis and TNF-alpha-induced cell death (35). Several interleukins can rescue lymphocytes from glucocorticoid-induced cell death (36). Survival factor deprivation of renal tubular cells increases the susceptibility to apoptosis induced by TNF-alpha (21). While the basis for this interaction is unclear, it may depend on changes in the expression of apoptosis regulatory genes (21,37).
3.2 INTRACELLULAR REGULATORS OF APOPTOSIS
The study of mutations of the genes involved in apoptosis in the development of Caenorhabditis elegans has identified two lethal genes, ced-3 and ced-4, which are required for cell death, and a survival gene, ced-9, which prevents death (38). A third type of gene, reaper, encodes a protein that activates the death program in Drosophila, but appears not to be directly involved itself (39). In C. elegans, additional genes regulate later steps of the apoptotic process, such as cell engulfment and degradation (38).
In mammalians, genes that may be classified as sensors/triggers ("reaper-like") (40), effectors ("ced-3/ced-4-like")(41), and survival factors ("ced-9-like")(42,43) have been identified.
Extracellular factors regulate cell survival and death through activation of sensors. The most intensively studied have been cytokine receptors. This sensors, in turn, could activate intracellular signals that trigger apoptosis.
3.2.1.1. Receptors that mediate cell survival and associated proteins
Integrins and cytokine receptors transduce poorly understood survival signals. The activity of the receptor depends not only on the availability of ligand, but also on intracellular regulators. EGF was one of the first cytokines shown to promote cell survival (14). The phosporylated EGF receptor binds to growth factor receptor-bound protein 2 (Grb2), which acts as a link to another intracellular signaling molecule, a guanine releasing factor of the son of sevenless (Sos) class. A Grb2 isoform, named Grb3-3, cannot bind to the EGF receptor, but does bind to the Sos-related factor (44). In this way, Grb3-3 acts as a dominant negative protein over Grb2. A direct functional consequence is that, when Grb3-3 is abundant, the survival promoting activity of EGF is blocked and apoptotic cell death is triggered (44). Both Grb2 and Grb3-3 are expressed in human kidney, although their possible role in renal physiology remains unexplored (44).
3.2.1.2. Receptors that mediate apoptosis and associated proteins
Several members of the TNF receptor superfamily regulate cell survival. This family is defined by similarities in their extracellular domains, and includes both TNF receptors, Fas, the NGF receptor and others (45). Both Fas (CD95) and 55 Mr TNF receptor (TNFR1)-induced apoptosis requires the integrity of a relatively homologous intracellular domain (46,47). This so-called death domain defines a different family that includes receptors and cytoplasmic proteins that are involved in cell death (40). The 75 Mr TNF receptor lacks the death domain but may also have a limited role in triggering cytotoxicity (48).
Agonistic anti-Fas antibodies and the endogenous ligand for Fas induce apoptosis (49-53). Fas ligand is expressed mainly by T cells (53), but its transcript is also present in macrophages and cultured renal cells (29). Fas plays a role in T cell-mediated cytotoxicity and activation-induced T cell death (54-57). When the microenvironment is appropriate, Fas may activate lymphocytes (58). Thus, it is not surprising that genetic defects in Fas ligand (gld mice) and Fas (lpr/lpr and lprcg mice) result in autoimmunity and a lupus-like syndrome (59-61). Fas defects also cause autoimmunity in humans (62,63). Soluble Fas molecules may be secreted and bind to and antagonize the Fas ligand (64).
Human and murine mesangial and tubular cells express Fas (29,65). Bacterial endotoxin and cytokines thought to play a pathogenic role in kidney damage, such as TNF-alpha, IL-1-ß and IFN-gamma, increase fas mRNA and Fas receptor expression (29,65,65b). Fas receptor expression peaks at 48-72h after stimulation, even though increased mRNA was noted as early as 1 hour (65). This time course is similar to that previously described in thymocytes (66). Agonistic anti-Fas antibodies induce apoptosis in murine and human renal cells (29 and unpublished). Death induced by Fas activation is increased in renal cells activated by cytokines or treated with the inhibitor of mRNA synthesis actinomycin-D (29). While the reason for the latter observation is not clear, several proteins protect against Fas-induced death, such as Fas-associated protein (FAP-1), a protein tyrosine phosphatase that associates to Fas (67).
The fact that Fas underexpression leads to autoimmunity and immune-mediated glomerulonephritis (60,61) does not necessarily exclude a role for Fas in renal damage. The genetic defect of fas in MRL-lpr/lpr lupus mice consists in the insertion of an endotransposon in the fas gene, leading to abnormal splicing of the gene and low levels of expression of normal fas transcript (61). These mice do however express the Fas receptor (68) and fas mRNA is detectable in the kidney (unpublished observation). Thus, the lpr/lpr mouse cannot be considered a knock-out for Fas. By contrast lprcg mice carry an inactivating point mutation in the death domain of the Fas molecule (60). A preliminary report suggests that mice carrying the lprcg gene in the MRL background display immune alterations equal to those of MRL lpr/lpr mice. However, even though glomerular C3 deposition is also similar, glomerular damage is milder in MRL-lprcg mice (69). One possible interpretation is that complete absence of functional Fas partially protects against glomerular injury in mice with autoimmunity.
Overexpression of Fas and of TNFR1 induces apoptosis in the absence of ligand, probably because of self-oligomerization through the death domain (70). Several cytoplasmic proteins share with these receptors the death domain and the ability to induce apoptosis through protease activation. They include the Drosophila protein reaper, mammalian receptor interacting protein (RIP), TNF receptor associated death domain (TRADD), and Fas associated death domain (FADD/MORT-1) (71-74). Beyond the similarity in the death domain, these proteins are unrelated. The death domain allows them to oligomerize and bind to the receptors. RIP and FADD bind to active forms of Fas, but not to mutated, inactive lprcg Fas (71,73,74). TRADD and, with less affinity, RIP, bind to TNFR1 (71,72). FADD appears to play a role in the transduction of the Fas death signal. It should be noted that the death effector domain in FADD is distinct from the so called death domain (74). RIP, TRADD and FADD are expressed in multiple organs, including the kidney (71-74).
3.2.1.3. Transcription factors
Several transcription factors have been implicated in apoptosis.
c-myc appears to activate a common genetic program that may determine both cell division and cell death. The presence or absence of additional signals (such as external survival factors or Bcl-2) may determine cell fate: the uncontrolled expression of c-myc increases cellular proliferation in the presence of growth factors, but in cells deprived of them, it induces apoptosis (75-77). c-myc activates the transcription of p53 (78), and death induced by c-myc requires p53 (79,80). In serum-deprived mesangial and tubular cells c-myc mRNA is induced by stimuli that promote apoptosis such as TNF-alpha (37).
p53 is the most frequently mutated or deleted gene in solid neoplasia (81) and it also plays a role in benign processes: its expression is increased in relation to neuronal cell death during experimental cerebral ischemia (82). In thymocytes p53 is required for apoptosis induced by radiation and DNA damaging drugs, but not for dexamethasone-induced apoptosis (9,83). It also participates in survival factor deprivation induced apoptosis (84,85). The mechanism of cell death by p53 is unclear. p53 decreases the transcription of the antiapoptotic gene bcl-2, and increases that of its antagonist, bax (87,88). However, p53-induced apoptosis may be independent of transcriptional activation of p53-target genes (89). External survival factors or enforced Bcl2 expression protect cells from death induced by p53 (86,90).
Fos and Jun homo or heterodimerize to form the AP-1 transcription factor. The expression of c-fos and c-jun precedes apoptosis and is rapidly and transiently induced upon growth factor deprivation in IL-2 and IL-6-dependent cell lines (91). Antisense inhibition of either of them or the intracellular injection of antibodies protects against death (91,92).
Nur77 is required for TCR-induced apoptosis in T-cell hybridomas (93,94). The ability of cyclosporine A to interfere with this form of apoptosis may be related to its ability to block Nur77 binding to DNA (95). Similar to the other transcription factors already mentioned, Nur77 also plays a role in the cell cycle regulation. nur 77 mRNA is expressed only after appropriate stimulation in cultured murine tubular and mesangial cells (unpublished observation) but it is so far unclear whether it regulates mitosis or death in these cells.
Proteases are key mediators of the effector pathway of apoptosis induced by either survival factor deprivation or receptor activation (41). Several proteins, such as poly(ADP)ribose polymerase (PARP) are cleaved during apoptosis. Protease inhibitors, including the cowpox virus Crm A protein protect cells from death. Several Ced-3-related Cys proteases whose overexpression results in apoptosis have been identified in mammals (41).
The IL-1ß-converting enzyme (ICE) was the first to be identified as the protease involved in apoptosis (96,97). ICE-induced cell death is prevented by CrmA and Bcl-2 (96). However, ICE-deficient mice do not display developmental abnormalities related to a defect in apoptosis, although Fas-induced apoptosis is defective (98). This suggests that either there is a redundancy in apoptosis effector pathways or that ICE is not the ultimate effector. In this regard, ICE cleaves PARP with low efficiency (99). ICE expression in cultured renal cells is low (unpublished observation).
Yama/CPP32ß/apopain is distal in the apoptotic pathway to ICE (100,101). Yama is a zymogen that, when activated, displays a CrmA-sensitive, Asp-specific Cys protease activity and cleaves PARP (100,101). However, in the absence of cytosol, Yama does not provoke apoptotic changes to nuclei, suggesting the involvement of other components of the apoptotic machinery (101). Yama mRNA is readily detected in both cultured renal cells and whole kidney (unpublished observation).
Related enzymes include Ich-1/Nedd2 (102), Ich-2/Tx/ICE-related protein II (103,104), Mch-2 and ICE-related protein III (105). Some are present as zymogens that are activated by proteolytic cleavage. The relationships between them are slowly being unraveled. Thus, Tx/Ich-2 can process pro-TX and pro-ICE (103), ICE is capable of processing both pro-ICE and pro-Yama to their active forms, but Yama cannot exert this function (100,106). This picture is further complicated by the fact that isoforms of these proteins have antagonistic effects on cell death. For example, Ich-1L promotes cell death, while Ich-1S inhibits this effect, presumably because it competes for the same targets (102). The main transcript expressed in kidney and cultured renal cells is Ich-1L (102 and unpublished).
Studies with protease inhibitors have implicated calpain I in apoptosis (107). Calpain cleaves and activates IL-1-alpha. The 17kD C-terminal fragment is referred to as mature IL-1-alpha, and is released to the medium. The 16 kD N-terminal fragment has recently been shown to be targeted to the nucleus and to induce apoptosis (108).
3.2.3 Survival proteins and regulatory factors
Among the proteins related to Ced-9 (Bcl-2, Bcl-xL, Bax, Bad, Bak), Bcl-2 and Bcl-xL protect cells from apoptosis, although their precise mechanism of action is unknown. Another set of proteins interact with the survival proteins, and may enhance or inhibit their survival promoting activity.
Bcl-2 is a membrane-bound protein present in mitochondria and other intracellular membranes (109,110). It has been suggested that Bcl-2 interferes with lipid peroxydation or the production of reactive oxygen species (111,112), but this has not been conclusively proven (113,114). Bcl-2 affords cells partial or complete protection from death induced by survival factor deprivation, Fas, TNF-alpha, c-myc, p53, ICE, Ich-1, Yama, glucocorticoids, oxidants, phospholipase A2, toxins, hypoxia and physical factors such as heat shock and irradiation (100,102,113-115 and reviewed in 116). The survival promoting potential of Bcl-2 has been demonstrated in renal cells (117). Bcl-2 does not protect against all forms of cell death. In some cases this has been explained by the need that this protein has to work in unison with other associated proteins (115).
Renal cells, including mesangial cells, tubular epithelium, fibroblasts and metanephric stem cells, express bcl-2 mRNA and protein (21,118). Murine renal cells express bcl-2 mRNA transcripts of several sizes (7.5, 4.1 and 2.4 kb); the 7.5 kb transcript being the most abundant. Gene expression of bcl-2 in renal cells appears to be controlled by environmental factors that regulate cell survival (21). In vivo, Bcl-2 protein is more abundant in embryonic than in adult kidney (119,120).
Derangements in Bcl-2 expression are associated with renal disease. For example, over-expression of Bcl-2 in B-lymphocytes of transgenic mice is associated with autoimmunity and the development of a proliferative glomerulonephritis (121). Mice carrying a targeted mutation in bcl-2 develop neonatal polycystic renal disease that progresses to renal failure (122).
Alternatively spliced isoforms of bcl-x protect from (bcl-xL) or predispose to (bcl-xS) apoptotic cell death in growth factor-deprived cells (123,124). The spectrum of the protection afforded by Bcl-xL against noxious stimuli overlaps with that of Bcl-2 (43,123). Bcl-xL protection may be more effective that Bcl-2, as in the case of cyclosporine A-induced apoptosis in lymphocytes (125). Enforced expression of Bcl-xL decreases the expression of endogenous Bcl-2 and vice versa (43). This observation of reciprocal regulation could explain the finding that antisense oligonucleotides for bcl-2 do not alter mesangial cell susceptibility to apoptosis (118).
Murine kidneys and cultured renal cells express two bcl-x mRNA transcripts, detectable by Northern blot hybridization; the smaller one having a shorter half life (37). Both murine transcripts hybridize to a human bcl-xL specific probe, suggesting that bcl-xL is the main isoform in the kidney (37,126).
3.2.3.2. Regulatory factors that promote apoptosis
Some factors antagonize the protective effect of Bcl-2 and Bcl-xL, and accelerate cell death when the microenvironment is permissive for death. These proteins differ, however, from effectors, because they do not induce cell death when the microenvironment supports survival.
Bax forms homodimers, as well as heterodimers with Bcl-2 and Bcl-xL (127,128). The BH1 and BH2 domains in Bcl-2 and Bcl-xL are required for the cell repressor activity and also for heterodimerization with Bax, but not for homodimerization (129). When optimal relative amounts of Bcl-2/Bax or Bcl-xL/Bax heterodimers are present, cells are protected from death induced by survival factor deprivation (127-129). If the percentage of free Bax is higher, cell death is more likely to occur (127-129).
It was originally reported that the murine kidney abundantly expresses two bax transcripts of 1.5 and 1 kb (127). We have found the 1 kb bax transcript to be preferentially expressed by mesangial cells, proximal tubular cells, metanephric stem cells as well as by the whole kidney (37). Bax immunoreactivity is readily detectable in renal tubular epithelial cells in vivo (130), where it appears to be more abundant than Bcl-2 (120,130).
Bad has homology to the BH1 and BH2 domains of Bcl-2 and has no lethal activity by itself (128). Bad heterodimerizes with Bcl-xL and Bcl-2, displaces Bax and reverses the death repressor activity (128). The Bcl-xS isoform of Bcl-x lacks the BH1 and BH2 domains and inhibits the ability of Bcl-2 to protect from cell death (123). Bak binds to Bcl-xL and antagonizes the protection offered by Bcl-2 when the cell is deprived of survival factors (131-133). Bak is present in the normal kidney (133).
3.2.3.3. Antiapoptotic regulatory factors
Bag-1 binds to Bcl-2 and enhances its survival promoting activity (115). The existence of Bag-1 or related factors may explain the conflicting reports on the ability of Bcl-2 to protect from Fas-induced apoptosis. Bcl-2 protection against this form of cell death is only complete when Bag-1 expression is high (115).
3.2.4 Other apoptosis related factors: clusterin
Clusterin is increased in acute and chronic renal failure associated with the occurrence of apoptosis (reviewed in 134). However, the relationship between clusterin and apoptosis is unclear (135-137). Clusterin inhibits the membrane attack complex of complement and is deposited in immune-mediated glomerular injury (134). In cultured mesangial cells, clusterin expression is regulated by cytokines (20).