Cytoplasmic vacuolation, adaptation and cell death: A view on new perspectives and features

485

Biology oj the Cell 91 (1999) 485498 o 1999 Editions scientifiques et medicates Elsevier SAS. All rights reserved

Review

Cytoplasmic vacuolation, adaptation and cell death: A view on new perspectives and features Tam& Henicsa* and Denys N Wheatleyb aDepartment of Medical Microbiology and Immunology, University Medical School of f&s, H-7643 P&s, Hungary; bDepartment of Cell Pathology, MacRobert Building, University of Aberdeen, 581 King Street, Aberdeen AB24 5UA, UK

This review focuses on a widely-observed morphological phenomenon, a unique class of cytoplasmic vacuolation, found in cultured (mammalian) cells. This vacuolation is quite distinct from autophagosomal and heterophagosomal, ie excessive lysosomal vacuolation, and occurs in most cell types spontaneously or via a wide range of inductive stimuli. Apart from vacuolation arising artefactually (usually due to poor fixation), spontaneous vacuolation occurs in individual or small clusters of cultured cells without apparent change in their local environment, while neighbouring cells remain completely unaffected. Since spontaneous vacuolation is unpredictable, the process of vacuolation - or ‘vacuolisation’ - (‘Vacuolation’ is the state of being with vacuoles; ‘vacuolisation’ therefore implies the process of becoming vacuolated. However, only the quicker term vacuolation will be used throughout this review to refer to the process of vacuole development.) induced experimentally, and hence relatively reproducibly by a range of substances and disturbances, offers an experimental approach which should give further insight into its physiology and pathophysiology. Unfortunately, our knowledge here remains woefully inadequate compared with the purely morphological aspects of the phenomenon. Vacuolation following disturbances could have an underlying common mechanism; however, a review of the literature suggests that this is not the case, and that it occurs via several different pathways, involving many different cell organelles and structures. All cells appear to retain the capacity to vacuolate for some physiological purpose, and it can be a permanent feature in many cell types, particularly ‘lower’ organisms and plants. Vacuolation in cells is generally seen as an adaptive physiological response, presumably for ‘damage limitation’, but very little is known about the intracellular homeostatic mechanisms which operate to restore the status quo. Where damage limitation fails, cells usually die quickly, but no clear evidence has been found that this is in any way ‘programmed’. It is argued that the demise which occurs via the vacuolation route may, in fact, be a distinct form of cell death which is difficult to fit into the conventional lytic and apoptotic modes. 0 Editions scientifiques et medicales Elsevier SAS

cytoplasmic vacuoles /vesicles /toxins / injury / reversibility / cell death

“Our study of isolated cells has made us familiar with reversible cell disturbances. Varieties of degeneration, *Correspondence and reprints: present address: Institute for Microbiology and Genetics, Vienna Biocenter, Dr Bohr Gasse 9, 1030 Vienna, Austria

Vacuolation of cells

too, are agreedupon, the distinction being basedin most caseson somesfrucfuraf change in the tissuesaficted. The ideal classification would be built upon functional disturbances, but our knowledge of celi physiology is imperfect and perhapswe are still under the tyranny of morphology.” (Cameron, 2952)

Henics and Wheatley

486

INTRODUCTION Modern cell and tissue techniques allow an increasing variety of mammalian cells to be maintained and examined in vitro under rigorously controlled conditions. Cultured cells respond to changes in their environment with a number of easily recognisable morphological signs, shrinkage or swelling, detachment, appearance of cytoplasmic processes, amongst which progressive ‘vacuolation’ of the cytoplasm is one of the more dramatic phenomena. It has been described in extenso, and description has through use led to many assumptions about exactly what happens in this biological response. By far the most detailed coverage can be found in chapters 20 and 21 of ‘P&zoIogy ofthe Cell’ (Cameron, 1952). It should be mentioned from the outset that vacuolation can all too easily be found in circumstances where cells have been badly prepared by standard procedures used in their visualisation, as in the preparation of a histological section or a cell smear. Since vacuolation can be seen in the living cell by phase-contrast microscopy in situations where no disturbance other perhaps than an inducing agent being added, this is clearly where the process is not artefactual. While vacuolation can be induced by a wide variety of agents and conditions, it also occurs ‘spontaneously’. The degree of vacuolation depends to a considerable degree on the cell type, some becoming vacuolated very easily, whereas others might be quite resistant. In some cell populations, almost all the cells can become vacuolated, whereas in others only a sporadic cell may be affected. Vacuolation is not only a response to injury, but a natural process associated with the sequestration of materials and fluids taken up by cells, and also with secretion and digestion of cellular products, for which elaborate mechanisms and infrastructure are already present. The Golgi and the endoplasmic reticulum both play their parts, and a relationship also exists between the GERL and the lysosomal system of the cell. Small vacuoles can be called ‘vesicles’, and these may have very specific roles to play in the Golgi and elsewhere. ‘The fusion of many vesicles together is one way of creating a larger vacuole, or small vesicles might alternatively/additionally expand by secretion of material from the cytoplasm into them to create larger vacuoles. Vacuolation is therefore a manifestation of a number of complex cellular activities, and unless we identify the types by their associated membrane markers, contents or site of origin, the tendency has been to lump them together as ‘cellular degenerations’. There are situations in which cell organelles are affected by the environment, causing swelling, Vacuolation of cells

Biology of the Cell 91 (1999) 485-498

which often gives cells a vacuolated appearance. The classic example of this is mitochondrial distension following hypotonic shock, referred to as ‘cloudy swelling’ by histopathologists. Finally, we must keep firmly in mind that vacuoles are often permanent features of cells. Many lower organisms and most mature plant cells have vacuoles which play a daily role in life. In plants, they are structurally important because they give the turgor pressure required to maintain some physical position, ie they are part of the hydraulic system, eg in keeping stems of grasses upright. The vacuole of a yeast cell is undoubtedly a centre of regulation of the many enzymatic activities under a controlled pH (Calahorra et al, 1998). It also plays a wider role in the physiology of the cell, as seen from the presence of a ‘cytoplasmic to vacuole’ targeting system, associated with autophagic activity in yeast (Baba et al, 1997). The molecular genetics of the different control processes are currently being explored in considerable depth (Bonangelino et al, 1997; Paidhungat and Garrett, 1998; Wang et al, 1998), clearly demonstrating the biological importance of vacuolation as a regular cell function, involving a plethora of genes. Protozoan cells living in fresh water have a permanent contractile vacuole, which shows rhythmical activity as it fills and empties. In evolutionary terms, the ability to form vacuoles is an early cellular attribute. To return to mammalian cells, intense vacuolation leads to cell death, and is expected therefore to fall into one of the two umbrella categories, either the lytic or apoptotic type. Perhaps that ‘cardinal’ feature of apoptosis (shrinkage necrosis, the cell cytoplasm becoming intensely crowded and dense; see Kerr, 1971) is not compatible with, or is indeed the opposite of, vacuolation. Although we may know very little about how vacuoles form, we remain equally ignorant of how the water content of the cytoplasm is reduced in apoptosis. As cytoplasmic tonicity increases, the medium effectively becomes hypotonic. The cell might therefore be expected to compensate by swelling and vacuolating. Since this does not occur in the vast majority of cases (see below), the cell membrane must become more permeable during apoptosis. Vacuoles reported in apoptotic cells tend to be either lipidfilled or autophagic vacuoles, quite unlike those induced by, eg procaine. We have examined DNA fragmentation in cells exposed to procaine hydrochloride for 36 h (fig l), and failed to detect the characteristic laddering of DNA in apoptosis (unpublished observations), arguing against this type of mechanism (Henics and Wheatley, 1997). Equally, vacuolation can occur which has little in common with that seen in autophagosomallydriven cell death, the feature of the lytic mode. Henics and Wheatley

Biology of the Cell 91 (1999) 485-498

487

Fig 1. Phase contrast view of subconfluent human foreskin fibroblasts incubated in normal medium (a) or in the presence of 3.7 mM procaine-hydrochloride for 6 (b) and 12 h (4. Prominent dose- and time-dependent cytoplasmic vacuolization is induced which is fully reversible upon removal of the inducer. Light microscopic and ultrastructural characteristics of such vacuolar structures are remarkably similar (if not identical) to those induced by a variety of other treatments (see text for details). Phase-contrast, x 800.

Thus, it was considered worthwhile looking deeper into the ways in which cell vacuolation occurs, its reversibility and how it leads to cell demise.

GENERAL FEATURES OF INDUCED VACUOLATION Extensively vacuolated cells were probably first seen by early pathologists who examined tissue sections of v&ious origins. In old pathology textbooks, a form of cellular damage, called ‘hydropic or vacuolar degeneration’ is mentioned and listed as a basic form of cell degeneration (eg references in Cameron, 1952; ChevilleJ983). Indeed, pathologists often see parenchymal cells containing large ‘empty’ vacuoles along with other attendant signs of cell damage. Highly vacuolated cells often have undiminished metabolic activity, as judged by overall protein synthetic rates and mitosis frequency (Belkin et al, 1962; Yang et al, 1965; Henics et al, 1993, 1997). While this is the case, vacuolation probably remains fully reversible, and normal morphology can be regained within a comparable period to that required for vacuole formation. Reversibility on the same time-scale has been observed by others, eg the swelling of mitochondria Vacuolation of cells

to form a vacuolated cytoplasm in cultured liver cells of rats treated with 1,2-dimethylhydrazine (Pollanen et aI, 1990), from which we infer an adaptive significance, although on many occasions the process occurs without any clear evidence of the perturbations which brought it about (‘spontaneous’ vacuolation). But it can also become irreversible and lead to cell death. Exactly what happens at this juncture is not at all clear; functionally defining death as the ‘irreversible breakdown of the energydependent function of the cell’ (Woolf, 1986) in such circumstances is unhelpful, and merely begs the question. A regular sequence of events seems to follow where a vacuolating stimulus induces a detectable response; vacuoles start to grow in number and size, occupying increasingly larger areas of the cytoplasm. Up to a certain threshold, cells can recover and continue to function normally. Beyond this threshold, structural damage caused by vacuolation, or some more direct toxic effect of the inducer, leads to cell death. Various cell types, indeed individual cells themselves, may have different thresholds (discussed in Henics and Wheatley, 1997). With some drugs, their continued presence leads to either some means of excluding Henics and Wheatley

488

the drug and/or metabolising it to an innocuous product, so that reversal occurs in its presence, whereas in other cases reversal is dependent on drug withdrawal. Additionally, environmental factors can have profound effects on the response, as demonstrated by the remarkable example of Wibo et al (1974) that, despite the high vacuolating tendency of their ascites cells in culture, the same tumour cells failed to vacuolate with comparable concentrations of inducing agents in vivo. There can also be times at which cells are less vulnerable to vacuolation. Procaine-induced vacuolation in Hep2 or human foreskin fibroblasts (fig 2) is sensitive to hyperosmotic conditions imposed by a variety of agents, but only at late phases (> 24 h) of the phenomenon (Finnin et al, 1969).

GENERAL FEATURES OF ‘SPONTANEOUS VACUOLATION Many cells show small numbers of vacuoles in their cytoplasm, associated with a range of ‘normal’ functions. This is emphasised in some cells, eg the human ovarian cancer cell line, PEOl, grown in culture, which can show anything from cells with little of no discernible vacuoles to cells completely filled with them (fig 2c, d). Similar appearances can be seen in vivo, not only in normal tissues of the female genital tract, but also in endometrioid adenocarcinoma cells (Kurihara ef al, 1997). In contrast, the PtKl cell line of epithelial cells from the kidney of a kangaroo rat grows into a very flat monolayer sheet, with generally no vacuoles to be seen during the proliferative stage. After several days at confluence, individual cell loss becomes increasingly obvious. The cells tending to undergo vacuolation are the large ones, small tightly clustered cells rarely having vacuoles. Characteristically, the cells which are going to round up and drift off into the medium (being moribund or completely dead, since they cannot be replated to form colonies) first produce this obvious vacuolation in the perinuclear cytoplasm, which then becomes increasingly

Biology of the Cell 91 (1999) 485498

intense with time, but in this case it does not regularly progress towards the cell periphery. Coalescence occurs around the nucleus, progressively isolating it from the cytoplasm, with only a few strands left connecting the two, like a moat being formed around a castle (fig 2a, b). The cells are shed within ~24 h of this occurrence. The nuclei become darker, but the cytoplasm peripheral to the vacuoles shows no condensation or difference from non-moribund, non-vacuolated cells. This typical cell death in a PtKl culture differs from traditional descriptions of both lytic and apoptotic modes of death (see Farber, 1996); there is no question that these changes are substantively different from those in type 3 or ‘non-lysosomal vesiculate degradation’ under Clarke’s (1990) classification, which are not localised, and can have vacuoles filled with all different sorts of materials (unlike the clear vacuoles described above).

EARLIER STUDIES ON CYTOPLASMIC VACUOLATION PHENOMENA The first systematic description of cells with extensive cytoplasmic vacuolation by induction was by Belkin et al (1961), who tested the effects of a variety of drugs on cultured ascite tumour cells. He found that progressive intracytoplasmic vacuolation occurred within 3 h of incubation and that the phenomenon was fully reversible upon removal of the inducing agent, with vacuoles disappearing in a further 34 h. Yang et al (1965) exposed primary cultures of chick heart vessel fibroblasts, HeLa cells, Fernandes amnion cells, and primary cultures of rat fetal heart cells to weak bases such as procaine hydrochloride. Vacuoles of many different sizes appeared exclusively in the cytoplasm, never in the nucleus. In most instances, prominent vacuolation could be observed within 1 h, but by 12-14 h every cell had extensively altered cytoplasm, usually with one or more large central vacuole. They also noted that vacuolation started in the perinuclear cytoplasm and progressed towards the

Fig 2. a. Late PtKl cell monolayer culture in which some cells have already left the surface denuded. Other cells are in the process of being released from the monolayer in a moribund state. The small black arrow (top-left) points to a large cell in which intense perinuclear vacuolation has occurred. A similar example is seen on the open arrow (bottom-right). The two cells (double arrows) are examples of complete coalescence of vacuoles around the nuclei, isolating them from the rest of the ceil. These cells are beginning to lift off the substratum and nuclei have also become shrunken and dense. Phase-contrast, x 900. b. The larger arrow here points to a cell in which there are discrete vacuoles wrapped around the nucleus to isolate it, the cell itself being still well-spread. A similar example can be seen at the smaller arrow. Both these cells illustrate spontaneous vacuolization amongst a recently established monolayer which is otherwise devoid of vacuolated cells. Phase-contrast, x 900. c. PEO ceils which show intense vacuolation, mostly perinuclear, associated with their active metabolism as ovarian epithelial cells. Large and medium size ceils tend to be vacuolated, but very small cells are relatively vacuole-free (arrow). Phase-contrast, x 900. d. As for c, showing very extensive vacuolation of a large PEO cell (arrow) with different intensities in surrounding cells. Phase contrast, x 900. Vacuolation of cells

Henics and Wheatley

Vat:uolation of cells

Henics and Whe !atley

Biology of the Cell 91 (1999) 485498

490

periphery as exposure continued. Wibo and Poole (1974) reported the formation of similar vacuolar structures in rat embryo fibroblasts exposed to different concentrations of chloroquine. Vacuoles developed within 2-3 h in the perinuclear region and occupied > 25% of the cytoplasmic volume. Similar data are found in Ohkuma and Poole (1981), following treatment of mouse peritoneal macrophages with chloroquine and other weakly basic substances. A common feature of the vacuolation phenomenon observed in our studies is relatively easy reversibility. Others have noted that nearly all cultures treated with weakly basic substances recovered after removal of the inducer (Belkin et al, 1961), vacuoles progressively shrinking, eventually resolving into the background cytoplasm (Yang et RI, 1965), although exactly how this occurs is unknown.

ENVIRONMENTAL CONDITIONS ASSOCIATED WITH CYTOPLASMIC VACUOLATION Induction of vacuolation

bv viruses

Morphological alterations, such as cell rounding or granulation of the cytoplasm, frequently accompany viral infections of cultured monolayer cells. Induction of cytoplasmic vacuoles by viruses or by the expression of viral structural genes in cells has also been described. For example, intensive cytoplasmic vacuolization in primary rat hepatocytes resulted from retrovirus transduction (Weber-Benarous et al, 1993). Pinto et al (1994) showed that the expression of rabies virus structural proteins in insect cells was associated with the appearance of highly vacuolated cytoplasm. Ultrastructurally, these vacuoles were free of electron-dense material, as well as having no obvious membranes surrounding them. Vacuolar generation was specifically related to two viral structural proteins, ‘M2’ matrix protein and ‘G’ glycoprotein, since no such induction accompanied the expression of either the ‘Ml’ or the ‘N’ nucleocapsid proteins. Studies on rabies and rabies-like viral replication also revealed that, besides clear cytopathological changes, large number of electron-lucent cvtoplasmic vacuoles were present in virus-infected neuronal tissues in mammals (Matsamuto et uZ,l974). Enterocyte vacuolation has been described as an indirect consequence of HIV infection in five human patients, all of whom were positive for the viral envelope protein, gp41, in their lamina propria mononuclear cells. This vacuolation was considered from its staining properties to be lipid, and to represent an enterocyte response to injury. Vacuolation of cells

Bacteria and their toxins One of the commonest causes of vacuolation is bacterial infection, which is induced by some of the toxins they release. One particular infection which features prominently is Helicobucfer pyZori, its toxin producing intense vacuolation in HeLa cells (Cover et al, 1992; Smoot et al, 1996). At very low levels, this toxin can greatly emphasise the vacuolation induced by a weak base. Some inhibition was seen in cells treated with monensin, a Na+ ionophore and glycoprotein secretion blocker. More recent studies from the same group indicate that human gastric epithelial cells grown in culture are more sensitive in showing greater vacuolation than HeLa cells, but this nevertheless contrasts with the former cells in situ in patients with Helicobacter pylori infections. The molecular biology and genetics of the gene encoding the toxin and its product are now under investigation, with attempts to map its cytotoxicity-promoting domains (Atherton et al, 1998). This demonstrates that some attempt to understand the vacuolation process in physiological terms is being made. Another organism, Cephu2osporiumcueruhs, produces the toxin, cerulein, which inhibits steroid and lipid biosynthesis. It targets the pancreas and causes intense vacuolation, an effect partially abrogated by allopurinol, an inhibitor of xanthine oxidase (Brunelli and Scutti, 1998). In comparison, the protoxin of Aeromonus hydrophilu, has been shown to be inserted into cell membranes, where its activation causes pore formation through a process of oligomerisation. This alters cell membrane K+ permeability, but the endoplasmic reticulum membranes are also altered, causing considerable ‘vacuolar’ distension (Abrami et uZ,l998). Infectious organisms are frequently introduced into the cell by an endocytotic mechanism. This necessarily involves their containment within a membrane, and the vacuole so created may enlarge in response to a variety of different actions on the cell. This not only occurs in bacterial invasion (eg Legionellapneumophilu), but is seen with other parasitic Protozoa, eg Toxoplusmu go&ii (Sinai and Joiner, 1997).

Drugs and toxic chemicals Apart from natural toxins, many drugs and other chemicals can induce vacuolation in cells. Weak bases (see above), such as chloroquine, procaine and ammonia, are well known as inducers. The literature contains innumerable reports of substances claimed to cause vacuolation of cells, but remember that this might include the classic example of lipid vacuole formation in liver in response to chloroform, carbon tetrachloride and other haloHenicsand Wheatley

491

Biology of the Cell 91 (1999) 485-498

gen-substituted hydrocarbons. It would be too difficult to gather all such references together, and therefore we will deal here only with the most recent papers which dwell on some aspect of the mechanisms involved rather than on the phenomenon itself. Substances like Ccl,, which damage the liver and cause lipid vacuoles to form, are often metabolised to more injurious species during ‘detoxification’ pathways by the cell, and since the most powerful enzyme are concentrated in the liver, damage is first encountered here. One recent study of note is that of Nayak et af (1996), who looked at small dose treatments of rats with Ccl,, which induced a cytoplasmic vacuolation characterised by accumulation of monoparticulate glycogen rather than the lipid vacuolation found with higher doses. These cells proved more resistant to subsequent high dose Ccl, administration, as if the prior alterations were by way of an adaptation to the toxin, whereas previously it was thought that these same cells were more sensitive. If injury does entail some adaptation which later confers resistance, it should be no surprise to that heat-shock (stress) proteins are involved, yet remarkably few studies of cell vacuolation have included studies on their expression. A case in point which does include them is that of Sharp et al (1992), in which N-methyl-D-aspartate receptor antagonists (ketamine, MK 801 and others) cause vacuolation in neurones of rats, and also induce hsp70 (see Dietrich et af, 1992; Freeman and Goldberg, 1994, for more details on the process of neuronal vacuolation itself). Substances such as adenine can cause pancreatic vacuolation, further illustrating the fact that different tissues handle substances and respond to both natural and xenobiotic compounds in different ways. The effect is exacerbated by dimethylsulphoxide treatment (Okazaki et al, 1992) on the basis that the former substance induced a higher activity of lipid peroxidation in the pancreatic cells. On the other hand, substances which inhibit activating metabolism and free radical production from many xenobiotic compounds (eg ascorbic acid, allopurinol) tend to reduce the severity of vacuolating agents. Oxidative injury in pancreatic cells, leading to vacuolation, was also reported by Niederau et al (1996). An example is the devastating damaging effect of 1-nitronaphthalene on the pulmonary epithelium, which can be offset by O,O,S-trimethylphosphorodithioate, but the associated vacuolation of liver cells in treated rats was not inhibited by this compound. The activating enzymes involved seem to be distinctly different in the two target tissues. A similar study with naphthalene in mice resulted in vacuolation rather than destruction of the bronchial epithelial cells (van Winkle et al, 1995). Vacuolation of cells

Insecticides are regularly among the list of vacuolating agents. A novel investigation involves winter flounder in Boston harbour, in which liver vacuolation was evident (Moore et ~2, 1996). In an earlier study, these fish were caught, checked endoscopitally for liver damage, and then injected intraperitoneally with bromodeoxyuridine. Three hours later, tissues were removed and fluorescent nuclei examined. Hepatocytes with a normal appearance rarely stained, whereas those areas containing vacuolated cells were in proliferative mode as judged by this Sphase specific test (Moore and Stegeman, 1992). Vacuolation has been reported in myelin fibres on nerves, for example, in response to 5-fluorouracil, tegafur and carmofur (Akiba et al, 1996). It is particularly noteworthy that no concomitant vacuolation of adjacent neuronal cells was found, nor in the supporting astrocytes and oligodendrocytes.

Vacuolation response to ultrafiltrates human serum

of

Cytoplasmic vacuoles are induced by exposure of different cell types to ultrafiltrates of human sera (Henics et ~2, 1993; Henics and Wheatley, 1997); we are unaware of any similar reports in the literature. The induced vacuoles arose transiently as ragged edged, seemingly membraneless and generally electron-lucent structures, which progressed from the perinuclear cytoplasm (their primary site of formation) towards the cell periphery. Additionally, the structures often maintained a close contact with filamentous elements, much as in the procaine-treated model (see above). When human foreskin fibroblast and Hep-2 epithelial carcinoma cells were pretreated with cytochalasin D to disrupt actin microfilamen&, interference occurred in both procaine and serum ultrafiltrate induced vacuolation in a doseand time-dependent manner. Proton-NMR relaxation T, and T, times were significantly reduced parameters in these vacuolated cells without concomitant change in their net water content.

Physical conditions eliciting vacuolation A wide variety of physical conditions, other than altered tonicity (osmolarity) of the medium surrounding cells (see Williams et al, 1991; and above), can cause vacuolation. Perhaps the commonest is ischaemia, since cells which have limited oxygen supply soon show signs of developing vacuoles and swelling as pH falls. The ischaemic changes in cells are those which most clearly suggest that, as metabolism becomes disrupted and glycolysis becomes the only form of energy production, illperfused tissues and cells cannot avoid the buildup of acidic products from this process. The most logi.cal explanation for a vacuolation process is that Henics and Wheatley

492 in the earlier stages of lysis, before swelling of organelles such as mitochondria, there is a release of a plethora of hydrolytic enzymes from various storage places to destroy the cell. Milder ischaemia will initially, at least in the body, be felt by the endothelial cells, in which vesicles will soon turn more obviously into vacuoles, and subsequently deeper tissues below the endothelia start to vacuolate (Caceres et al, 1995; Clause11 et al, 1996). Cold induces vacuolation in many cells, but the severity of it is variable. Concern comes when whole tissues and organs are cooled for storage at unnatural temperatures, since cold per se is only relative, and vacuolation of cells certainly does not feature commonly in organisms acclimated to temperatures around zero centigrade. The start of vacuolation in cells of stored organs (such as liver or kidneys of pigs) has been correlated with the onset of damage and loss of viability (Romagnoli et al, 1990). However, this is typically inferred, because no experimental evidence has been presented that truly correlates vacuolation with the loss of viability of those particular cells, especially as the same tissues from different animals did not all show the same changes in response to the cold. If the onset of vacuolation is a true marker of loss of viability, quantification would be desirable and a much stricter correlation drawn up between the dgeree of damage and the loss of viability. Tubular epithelial cells have been shown to become vacuolated in response to shock waves given to isolated kidneys (Kohrmann et al, 1994), as well as by other agents introduced for physical reasons (radio-opaqueness for imaging, eg Tervahartiala et al, 1997). Perhaps of all cells which are at constant risk to a wide variety of vacuolating agents, kidney tubular cells are the ones featured most widely in reports of cytoplasmic changes. Damage can often be met at the proximal part of the tubular system; depending on the severity of the agent, it can become manifest at the same time or later in the more distal epithelial parts of the tubules and collecting ducts (Rinke ef al, 1998). The type of swelling seen in tubular cells are not always well characterised and go from hydropic degeneration, cloudy (mitochondrial) swelling, to clear evidence of large serous vacuoles in the apical part of the cell. This contrasts strongly with the way in which vacuoles develop in spontaneously or procaine-induced vacuolation, which occurs perinuclearly and often around the centrosomal region.

SOME FEATURES OF VACUOLATED CELLS

Biology of the Cell 91 (1999) 485-498

uoles. For instance, Wheatley et al (1984, 1990) described a large number of vacuoles in the cytoplasm of HeLa S-3 cells exposed to hyperosmotic concentrations of sorbitol (fig 3). Ultrastructural examination revealed that the vacuoles had no obvious membranous demarcation commonly seen with other intracytoplasmic bodies such as secondary lysosomes or swollen Golgi vesicles. Moreover, the vacuoles were quite electron-lucent. Interestingly, further ultrastructural studies showed that seemingly identical vacuolar structures were induced by a variety of other agents, including glycerol, urea and amino acids. Clegg et al (1986) reported the formation of cytoplasmic vacuoles with no apparent membrane-like demarcation or obvious electron dense content in L-929 fibroblast cells treated with 0.3 M sorbitol in long-term cultures. The possibility was mooted that a non-membrane-busedphase-separation of intracellular water was occurring in the cytoplasm to explain this phenomenon (Wheatley et al, 1984). This phenomenon has not been studied independently, and has been criticised on the grounds that the received wisdom tells us that all vacuoles should, apparently by definition, be ‘membrane-bound’. (This simply creates a nomenclature problem, ie what we should call serous fluid or any other droplet of phase-separated material in the cytoplasm.) However, until it is corroborated by independent observation that phase separations can occur, which must at least involve interfaces if not membranes, there will be disagreement about this phenomenon. The implication of the latter notion is that not until a ‘true’ membrane is present can pumps and transporters introduce more fluid and solute material into the vacuole; from this argument it follows that all vacuoles are supposedly formed from membranous vesicles generated presumably by the GERL system. But the appearance of non-membranous lacunae within cytoplasm full of ribosomes is not uncommon, of which figure 4 is an obvious demonstration. Nevertheless, while membraneless vacuoles may or may not be a real phenomenon, in many caseswe accept that it can represent a technical artefact, due to tangential sectioning of a thin hollow sphere not revealing a distinct bilipid layer whereas more equatorial cuts do. Also, the surrounding membranes can be shed internally into the vacuoles, leaving them partially or completely denuded (fig 5a). Nevertheless,we do believe that both types can exist, although membrane-bound forms are far more commonly reported than non-membrane-bound ones (fig 5b).

Osmotic changes and water partitioning in cells

Relationship of vacuoles to the cytoskeleton

Osmotic insults of cultured cells frequently result in the formation of various size cytoplasmic vac-

A number of studies have implicated the involvement of cytoskeletal components in the sorting and

Vacuolation of cells

Henicsand Wheatley

Biology of the Cell 91 (1999) 485-498

493

Fig 3. HeLa S-3 cells subjected to 0.3 M sorbitol in basal Eagle’s medium, resulting in a very high degree of cytoplasmic and nuclear shrinkage and condensation with a phase separation of cytoplasmic water in clear vacuoles, some times clustering around the Golgi region, but mostly with a general distribution in the cytoplasm. Vacuoles have never been seen in the nucleus. Electron micrograph, x 4000.

Fig 4. Cytoplasm of a Hep2 cell in which the cytoplasm has developed membrane-less vacuolar regions, and increased ribosomal density in the intervening matrix. A mitochondrion (arrow) indicates no swelling of an hypo-osmotic nature. Electron micrograph, x 37 500. Vacuolation of cells

Henics and Wheatfey

494

Biology of the Cell 91 (1999) 485498

Fig 5. a. Large perinuclear vacuoles in a Hep-2 cell, in which the plane of section sometimesreveals their limiting internal membranes (filled arrows), and in other cases does not (clear arrows). Evidence of demembranation seems to exist in several of them (asterisks). Electron micrograph, x 12 500. b. Periphery of a Hep-2 cell in which large vacuoles are seen coalescing, with demembranation also apparent. Electron micrograph,x 12 500.

maintenance of various vesicular compartments, such as endocytotic or lysosomal vesicles (Durrbath at al, 1996; Murphy et al, 1996; Riezman et al, 1996). The participation of filament elements in the formation and/or maintenance of induced cytoplasmic vacuoles has been pointed out in a number of recent studies. Fenyves et al (1993) studied the effect of interferon-y on cultured microvascular Vacuolationof cells

endothelial cells which had been isolated from bovine corpus luteum and differed mainly in their cytoskeletal organisation. One phenotype of these cells had a unique ‘starburst-like’ pattern of actin filaments, frequently located at the cell centre, a feature which is quite distinct from the only other reported actin cytoskeleton pattern (ie a dense peripheral band) in resting endothelial cells. InterHenicsand Wheatley

Biology of the Cell 91 (1999) 485-498

estingly, certain subtypes of this very same phenotype of cells exhibit lucent intracytoplasmic vacuoles of various size. With interferon-y treatment, the number and size of such cytoplasmic vacuoles grew dramatically, and concomitantly there was rearrangement of their characteristic microfilament pattern. Peripherally located actin intersections became connected by actin cables. More recently, Wilson et al (1996) examined the effect of cannabinoid enantiomers on the cytoarchitecture of pheochromocytoma (PC-12) cells, and found that some synthetic derivatives are capable of inducing cytoplasmic vacuolation in these cells. Using an immunofluorescent approach to visualise cytoskeletal structural changes following cannabinoid treatment, these authors showed that medium and large sized cytoplasmic vacuoles had actin surrounding their periphery. Eisel et aI (1993) noted that expression of tetanus toxin light-chain in Sertoli cells of transgenic mice resulted in the formation of large cytoplasmic vacuoles, parallel with disruption of actin filaments. Involvement of actin in relation to vacuolation of cornea1 cells of rabbit and man can be inferred from the Mitomycin-C treatments given to isolated perfused corneas in the study of Nuyts rt al (1995). Mairesse et al (1996) reported that inhibition of the 27 kDa heat-shock protein by transfecting antisense cDNA into MCF-7 breast carcinoma cells was associated with prominent growth inhibition and changes in microfilament organisation from diffuse to punctate distribution. These studies demonstrated the appearance of cytoplasmic vacuolar structures which were quite distinct in their ultrastructural properties from those of other known vacuolar systems, such as seen in the endocytotic or lysosomal pathways. The question always left by such studies is whether the cytoskeletal rearrangements are integral or incidental to the vacuolation process. In our own studies on cytoplasmic vacuoles, induced by weak bases and ultrafiltrates of human serum, we have shown that there may be a close relationship with F-actin based microfilaments and that cytochalasin mediated disruption of the microfilament system can interfere with the induction of vacuolation (Henics and Wheatley, 1997). Where vacuoles form as a normal physiological process, or have become permanent features through evolutionary necessity, as in the contractile vacuoles of freeliving protozoa, it is noticeable that they can stain for contractile fibre elements in the immediately surrounding cytoplasm. In the case of protozoa, the association is very direct and unquestionably has a role to play in the filling and discharge of vacuolar contents. We have spent even less time exploring the ability of induced vacuoles to discharge their contents than on their filling. Vacuolation of cells

495

COMMON FEATURES IN SPONTANEOUS AND INDUCED VACUOLATION From the discussion so far, the literature and our own observations, some general features of both spontaneous and induced cytoplasmic vacuolation can be gleaned. i) The prime subcellular region for early vacuole accumulation tends to be perinuclear, probably associated with the Golgi elements close to the centrosome of the cell. Vacuoles move towards the periphery, as they increase in both number and size, the latter often by coalescence, as is also typical of a secretory pathway. This contrasts with some vacuolation processes which occur in the opposite manner, by endocytotic mechanisms from the cell surface. Vacuoles can also arise from lysosomes by inclusion of weak bases, and may also develop from swelling of organelles such as mitochondria. Nuclear vacuolation is rare. ii) In many instances, the process of vacuolation in response to injury is fully reversible, while in others, continued exposure to certain levels of inducers, is irreversible and leads to cell death. The changes which crucially decide whether a cell recovers rather than engages some active or passive process leading to death remain unknown, but have been loosely attributed to the inability of the cell to maintain its energy status, which is tantamount to admitting ignorance. iii) A striking but important ultrastructural feature of some vacuoles is their apparently membraneless appearance (figs 3, 4) and the lack of obvious electron-dense content, other than some flocculent material. Vacuoles of this type are probably quite different in origin from those mentioned under (i), and it is suggested that they arise by phase separation within the cytoplasmic ground substance. iv) Vesicular and vacuolar structures often maintain intimate contact with filamentous elements and stay closely co-ordinated with such filaments as they expand, whether or not they are integral parts of the vacuolation process. There is no single mode by which vacuoles develop; clearly they originate in a number of different ways, and probably from systems which are involved in normal aspects of cellular functioning. Vacuoles of many different types are frequently found, both under physiological and pathological conditions, which can include material of fatty, mutinous, or serous nature. Until some far more consistent classification of vacuolation emerges, it seems almost irresponsible to lump them all together as generally indicative of a degenerative state. This requires a physiological basis, as emphasised by Cameron (1952). Their physiological signiHenicsand Wheatley

496

ficance, particularly in the injured cell, needs to be more fully understood, especially if they are a temporary means of subverting more serious damage through sequestration of toxic compounds and irritants, and/or play a role in maintaining not only osmotic pressure of the cytoplasm itself, but its pH, as seen in the functioning under normal circumstances of vacuoles in organisms such as yeast and protozoa, from which we still have much to learn. In cell culture, where cytoplasmic vacuolation can best be studied under carefully controlled circumstances, many natural and xenobiotic substances can produce this ‘adaptive response’, which we often see as ‘damage’ that has to be repaired/corrected. Left alone, or taken too far, cell functioning is compromised and viability soon lost. But vacuolated cells which are capable of recovery clearly maintain a high metabolic and proliferative activity during the peak period of vacuolation. In the case of spontaneous vacuolation, changes must occur either in the cell itself or in its environment, of which we remain largely ignorant.

VACUOLATION AND CELL DEATH. CONCLUDING REMARKS Terminally, when the process of vacuolation is no longer reversible, the capacity of the cell to ‘compromise’ or ‘sacrifice’ further cytoplasmic substance presumably becomes exhausted. The obvious scenario is of an explosion of the vacuoles tearing the membranes of the vacuoles and the cell apart, a truly lytic cell death. But this is an assumption, for the truth of the matter is that we do not know this to be the case. The vacuolation process probably has to be seen as the visible response a cell makes in an attempt to maintain its functional status under stress. Whether some switch operates to set a program leading to a quick demise from this point onwards remains a moot point, but one worthy of more focused investigation. This argument is raised again here because there are known examples of cells entering a decidedly vacuolated state leading to death in circumstances which suggest some underlying program. During salivary gland development in the metamorphosis of the blowfly larva, cells are eliminated on a particular order around day 9, by a process of intensive vacuolation with: “none of the features associated with classical apoptosis...there is no cell shrinkage or early fragmentation and no margination of chromatin; indeed the cells appear to swell and vacuolate, their contents being dispersed into the lumen” (Bowen et al, 1993; see Bowen et al, 1996, for a broader discussion). These in vine findings agree with our own studies on this process both in Hep-2 cells under an inducer as well as forming spontaneously. It would Vacuolatronof cells

Biology of the Cell 91 (1999) 485498

seem that there is an intrinsically triggered and probably not an entirely randomly initiating process in vacuolation which might be important in the elimination of damage or diseased cells, as in virusinfected cells discussed earlier, eg Pinto et a2 (1994), and monolayer cultures of PtKl cells (fig 2c, d; although here it is difficult to see how it can be other than a randomly initiated event). Finally, Cornillon et al (1994) described a process of this kind in DicfyosteIium discoideum, in which stalk cells undergo extensive vacuolation as they proceeds towards programmed cell death. Careful study of this process reveals a quite different pattern of vacuoles of all different sizes with many different inclusions spread much more evenly throughout the whole cytoplasm. This is not the same type of vacuolation that we have been describing in, for example, spontaneous vacuolation in PtKl cells or in procaine-induced vacuolation. But this is the type of vacuolation to which Carke (1990) had referred in his classification. A similar example is in cell death involving vacuolation in Caenorhabditis elegans(Robertson and Thonson, 1982), which again is little more than the development of large autophagosomes in cells considered to be undergoing apoptosis, but with a more lytic appearance than others (leading to confusion between these two processes in models in which ‘programmed cell death is supposed conventionally to occur). But if free living cells can indulge in death by vacuolation, then one will appreciate more fully why it ought to be seen as a distinct form of demise. We are left with the intriguing possibility that death by vacuolation might indeed have an identity of its own. In agreement with Farber (1996), we see no reason why vacuolation should not be as distinct a form of cell demise as the two to which everything is currently being assigned. While the lytic or apoptotic modes of cell demise have been extensively studied (to the almost complete exclusion of others), and have their established place in a number of well-defined instances (eg as in developing C eleguns), it is questionable whether the processesinvolved in them are indeed as clearly distinguishable as so often indicated. We should perhaps give some credence to at least a third mode of cell degeneration. Although, Clarke’s type 3 or nonlysosomal vesiculate degradation shares some common features with the cytoplasmic vacuolation phenomenon we described above, the cytoplasmic area affected, the lack of involvement of other organelles, the serous nature of the vacuoles, and the relative ease initially of their reversibility together suggest that these types of cell ‘degeneration’ are not identical. We suggest that this often observed, but in not many cases, so clearly identified and well-characterized, process of cytoplasmic vacuolaHenics and Wheatley

Biology of the Cell 91 (1999) 485-498

tion primarily reflects an adaptive, survival response to a plethora of environmental changes, that also has the potential to lead to a particular and distinctive form of cell death.

ACKNOWLEDGMENTS This work was supported by an International Journal of Experimental Pathology Fellowship to TH. DNW wishes to acknowledge the help of Dr Sam Bowser (Wadsworth Laboratories, NY State Public Health Service, Albany, NY, USA) and Dr Simon Langdon (Imperial Cancer Research Fund Unit, University of Edinburgh, UK) for supplying PtKland PEOl cells, respectively.

REFERENCES Abrami L, Fivaz M, Glauser P-E, Parton RG and Van der Goot FG (1998) A pore-forming toxin interacts with a GPIanchored protein and causes vacuolation of the endoplasmic reticulum. J Cell Biol 140,525-540 Akiba T, Okeda II and Tajima T (1996) Metabolites of 5-fluorouracil, alpha-fluoro-beta-alanine and fluoroacetic acid, directly injure myelinated fibers in tissue culture. Actu Nezrropathol92, 8-13 Atherton JC, Peek RM Jr, Tham KT, Cover TL and Blaser MJ (1997) Clinical and pathological importance of heterogeneity in vacA, the vacuolating cytotoxin gene of Helicobacter pylori. Gusfroenferology 112,92-99 Baba M, Osumi M, Scott SV, Klionsky DJ and Ohsumi Y (1997) Two distinct pathways for targeting proteins from the cytoplasm to the vacuole/lysosome. J Cell Bioll39,1687-1695 Belkin M, Hardy WG, Orr HC and La&man AB (1962) Induction in vitro by autonomic drugs of cytoplasmic vacuoles in ascites tumor cells. J Nat1 Cancer lnsf 28,187-201 Bonangelino CJ, Catlett NL and Weisman LS (1997) Vac7p, a novel vacuolar protein, is required for normal vacuole inheritance and morphology. Mel Cell Bioll7,6847-6858 Bowen ID, Morgan S M and Mullarkey K (1993) Cell death in the salivary glands of metamorphosing Calliphora vomitaria. Cell Biol Int 17,13-33 Bowen ID, Mullarkey K and Morgan SM (1996) Programmed cell death during metamorphosis in the blow-fly Culliphoru vomitaria. Microscop RES Tech 34,202-217 Brunelli A and Scutti G (1998) An ultrastructural study to investigate the effect of allopurinol on cerulein-induced damage to pancreatic acinar cells in rat. Int J Puncreutol 23, 25-29 Caceres MJ, Schleien CL, Kuluz JW, Gelman B and Dietrich WD (1995) Early endothelial damage and leukocyte accumulation in piglet brains following cardiac arrest. Acta Neuropathol90,582-591 Calahorra M, Martinez GA, Hernandez-Cruz A and Pena A (1998) Influence of monovalent cations on yeast cytoplasmic and vacuolar oH. Yeast 14.501-515 Cameron GR (1952)‘Putlrology $ the Cell. Oliver and Boyd, Edinburgh Cheville NF (1983) Cell degradation: In: Cell Pathology, 2nd edn. The Iowa State University Press, 75-129 Clarke PGH (1990) Developmental cell death: morphological diversity and multiple mechanisms. Anat Embryol 181, 195-213 Clegg JS, Gallo J and Gordon E (1986) Some structural, biochemical and biophysical characteristics of L-929 cells growing in the presence of hyperosmotic sorbitol concentrations. Exp Cell Res 163, 3546 Cornillon S, Foa C, Davoust J, Buonavista N, Gross JD and Goldstein P (1994) Programmed cell death in Dictyostelium. ICeUSci107,2691-2704 Cover TL, Vaughn SG, Cao I’ and Blaser MJ (1992) Potentiation

Vacuolation of cells

of Helicobucter pylori vacuolating toxin activity by nicotine and other weak bases. J Infect Dis 165,1073-1078 Dietrich WD, Halley M, Alonso 0, Globus MYT and Busto R (1992) Intraventricular infusion of N-methyl-D-aspartate. 2. Acute neuronal consequences. Acta Neuroputhol84,630-637 Durrbach A, Louvard D and Coudrier E (1996) Actin filaments facilitate two steps of endocytosis. J Cell Sci 109,457-465 Eisel U, Reynolds K, Riddick M, Zimmer A, Niemann H and Zimmer A (1993) Tetanus toxin light chain expression in Sertoli cells of transgenic mice causes alterations of the actin cytoskeleton and disrupts spermatogenesis. EMBO J 12,3365-3372 Farber E (1996) Ideas in pathology. Programmed cell death: necrosis Versus apoptosis. Mod PuthoZ7,605-609 Fenyves AM, Behrens J and Spanel-Borowsky K (1993) Cultured microvascular endothelial cells (MVEC) differ in cytoskeleton, expression of cadherins and fibronectin matrix. J Cell Sci 106,879-890 Finnin BC, Reed BL and Ruffin NE (1969) The effects of osmotic pressure on procaine-induced vacuolation in cell culture. J Phurm Phurmacol21,114-117 Freeman JK and Goldberg MP (1994) Confocal microscopic visualisation of MK 801 induced cytoplasmic vacuoles in vitro. Psychophurmucol Bull 30,541-547 Henics T, Koszegi T, Szucs G, Kellermayer MSZ Jr and Kellermayer M (1993) Transient cytoplasmic vacuolization in cultured normal and neoplastic cells treated with low molecular weight human serum ultrafiltrate: Is out inside? Physiol Chem Phys Med NMR 25,227-236 Henics T and Wheatley DN (1997) Vacuolar cytoplasmic phase separation in cultured mammalian cells involves the microfilament network and reduces motional properties of intracellular water. inf J Exp PafhoZ78,343-354 Kerr JFR (1971) Shrinkage necrosis: a distinct form of cell death. J Pathol 105,13-20 Kohrmann KU, Back W, Bensemann J, Florian J, Weber A, Kahmann F, Rassweiler J and Alken P (1994) The isolated perfused kidney of the pig: new model to evaluate shockwave induced lesions. J Endourol8,105-110 Kurihara Y, Kasahara K, Sasano H and Tsutsumi Y (1997) Ovarian endometrioid adenocarcinoma associated with steriodogenic foamy stromal cells. Puthol Int 47,883-888 Mairesse N, Herman S, Mosselmans R and Galand P (1996) Antisense inhibition of the 27-kDa heat shock protein production affects growth rate and cytoskeletal organization in MCF-7 cells. Cell Biol Inf 20,205-212 Matsumoto S, Schneider LG, Kawai A and Yonezawa T (1974) Further studies on the replication of rabies and rabies-like viruses in organized cultures of mammalian neural tissues. J Viral 14,981-996 Moore MJ and Stegeman JJ (1992) Bromodeoxyuridine uptake in hydropic vacuolation and neoplasms in winter flounder liver. Marine Environ Res 34,13-18 Moore MJ, Shea D, Hillman RE and Stegeman JJ (1996) Trends in hepatic tumors and hydropic vacuolation, fin erosion, organic chemicals and stable isotope ratios in winter flounder from Massachusetts. Marine Pollut Bull 32,458470 Murphy C, Saffrich R, Grummt M, Gournier H, Rybin V, Rubino M, Auvinen I’, Lutcke A, Parton RG and Zerial M (1996) Endosome dynamics regulated by a Rho protein. Nature 384,427-432 Nayak NC, Sathar S A, Mughal S, Duttagupta S, Mathur M and Chopra P (1996) The nature and significance of liver cell vacuolation following hepatocellular injury - an anlaysis based on observations on rats rendered tolerant to hepatotoxic damage. Virchow’s Arch 428,353-365 Niederau C, Klonowski H, Schulz H-U, Sarbia M, Luthen R and Haussinger D (1996) Oxidative injury to isolated rat pancreatic acinar cells vs. Isolated zymogen granules. Free Rud Biol Med 20,877-886 Nuyts RMMA, Holley GP and Edelhauser HF (1995) Effects of mitomycin C on the cornea1 endothelium. J Toxicol Cutun Ocul ToxicoZ14,251-264 Ohkuma S and Poole B (1981) Cytoplasmic vacuolation of

Henics and Wheatley

Biology of the Cell 91 (1999) 485-498

498

mouse peritoneal macrophages and the uptake into lysosomes of weakly basic substances. J Cell BioZ90,656-664 Okazaki Y, Minami T and Natsui K (1992) The effects of adenine and dimethylsulfoxide on the mouse pancreas. J Pharmacobio-Dynamics, 15,605-610 Paidhungat M and Garrett S (1998) Cdcl and the vacuole regulate Mn*+ homeostasis in the yeast Saccharomyces cerevisiae. Genetics 148,1787-1798 Patterson BK, Ehrenpreis ED and Yokoo H (1993) Focal enterocyte vacuolisation: a new microscopic finding in the acquired immune deficiency wasting syndrome. Am J Clin Path01 99,24-27 Pinto RM, Bosch A and Bishop DHL(1994) Structures associated with the expression of rabies virus structural genes in insect cells. Virus Res 31,139-145 Pollanen MS, Quinn BA, Wollenberg GK and Hayes MA (1990) Reversible mitochondrial swelling in cultured rat hepatocytes exposed to 1,2-dimethylhydrazine. Exp Mol Pafhol 52, 170-178 Rinke M, Bomhard EM, Hildebrand H, Leser KH, Loof I and Ruehl-Fehlert C (1998) Serotonin (5-HT(lA)-receptor) agonist-induced collecting duct vacuolation and renal papillary necrosis in the rat. Toxicol PathoZ26,152-159 Robertson AMG and Thomson JN (1982) Morphology of programmed cell death in the ventral nerve cord of C elegans larvae. ] Embryo1 Exp Morph01 67,89-100 Romagnoli P, Rossi S, Filipponi F, Marsili M and Boncinelli S (1990) Alterations in the fine structure of the pig liver upon cold ischemia. Eur J Surg 22,189-196 Sharp FR, Butman M, Wang S, Koistinaho J, Graham SH, Sagar SM, Noble L, Berger P and Longo FM (1992) Haloperidol prevents induction of the hsp70 heat shock gene in neurons injured by phenylcyclidine (PCP), MK801, and ketamine. J Neurosci Res 33,605-616 Sinai AP and Joiner KA (1997) Safe haven: the cell biology of nonfusogenic pathogen vacuoles. Annu Rev Microbial 51, 415-462 Smoot DT, Resau JH, Earlington MH, Simpson M and Cover TL (1996) Effects of Helicobucter pylori vacuolating cytotoxin on primary cultures of human gastric epithelial cells. Gut 39,795-799 Terahartiala P, Kivisaari L, Kivisaari R, Vehmas T and Virtanen I (1997) Structural changes in renal proximal tubular

Vacuolation of cells

cells induced by iodinated contrast media. Studies in dehy drated rats. Nephron 76,96-102 Van Winkle LS, Buckpitt AR, Nisho SJ, Isaac JM and Ploppel CG (1995) Cellular response in naphthalene-induced Clara cell injury and bronchiolar epithelial repair in mice. Am ] Physiol-Lung Cell & Mol PhysioZ296, LSOO-L818 Verschoyle RD, Carthew P and Wolf CR (1993) 1-Naphthalenc toxicity in rat lung and liver. Effects of inhibiting and inducing cytochrome P450 activity. Toxicol Appl Pharmacol 122,208-213 Wang Y-X, Catlett NL and Weisman LS (1998) VacSp, a vacuolar protein with armadillo repeats, function in both vacuole inheritance and protein targeting from the cytoplasm tc vacuole. J Cell Biol 140,1063-1074 Weber-Benarous A, Decaux, JF, Bennoun M, Allemand I, Briand P and Kahn A (1993) Retroviral infection of primary hepatocytes from normal mice and mice transgenic fol SV40 large T antigen. Exp Cell Res 205,91-100 Wheatley DN, Inglis, MS and Clegg JS (1984) Dehydration 01 HeLa S-3 cells by osmosis. 1. Kinetics of cellular responses to hypertonic levels of sorbitol, amino acids and other selected agents. Mol Physiol6,163-182 Wheatley DN, Rimmington JE and Foster MA (1990) Effects 01 osmotic manipulation of intracellular hydration of HeLa S 3 cells on their proton NMR relaxation times. Mag Resow; Imaging 8,285-293 Williams JC Jr, Abrahamson DR and Schafer JA (1991) Structu ral changes induced by osmotic water flow in rabbit proxi. ma1 tubule. Kidney Int 39,672-683 Wibo M and Poole B (1974) Protein degradation in cultured cells. II. The uptake of chloroquine by rat fibroblasts ant the inhibition of cellular protein degradation and cathepsir 81. J Cell Biol63,430-440 Wilson RG Jr, Tahir SK, Mechoulam R, Zimmerman S ant Zimmerman AM (1996) Cannabinoid enantiomer action or the cytoarchitecture. Cell Biol Int 20,147-157 Woolf N (1986) Cell, Tissue and Disease. Bailliere Tindall, Len don Yang WCT, Srasser FF and Pomerat CM (1965) Mechanism 01 drug-induced vacuolization in tissue culture. Exp Cell Re: 38,495-506 Received

23 October

1998; accepted

24 June 1999

Henics and Wheatle)