Keratin mutations of epidermolysis bullosa simplex alter the kinetics of ...

Research Article

4341

Keratin mutations of epidermolysis bullosa simplex alter the kinetics of stress response to osmotic shock Mariella D’Alessandro, David Russell, Susan M. Morley, Anthony M. Davies* and E. Birgitte Lane‡ Cancer Research UK Cell Structure Research Group, University of Dundee School of Life Sciences, MSI/WTB Complex, Dow Street, Dundee DD1 5EH, UK *Present address: Tayside Institute of Child Health, Ninewells Hospital and Medical School, Dundee DD1 9SY, UK ‡Author for correspondence (e-mail: [email protected])

Accepted 23 August 2002 Journal of Cell Science 115, 4341-4351 © 2002 The Company of Biologists Ltd doi:10.1242/jcs.00120

Summary The intermediate filament cytoskeleton is thought to confer physical resilience on tissue cells, on the basis of extrapolations from the phenotype of cell fragility that results from mutations in skin keratins. There is a need for functional cell assays in which the impact of stress on intermediate filaments can be induced and analyzed. Using osmotic shock, we have induced cytoskeleton changes that suggest protective functions for actin and intermediate filament systems. Induction of the resulting stress response has been monitored in keratinocyte cells lines carrying K5 or K14 mutations, which are associated with varying severity of epidermolysis bullosa simplex. Cells with severe mutations were more sensitive to osmotic stress and took longer to recover from it. Their stress-activated response

Introduction Intermediate filaments remain the least-understood component of the cytoskeleton of eukaryotic cells. The 65 or more (Hesse et al., 2001) members of this large multigene family are expressed in a tissue-specific manner that suggests they have an important role in tissue cell differentiation and function. This has been difficult to test experimentally, owing to the robust nature of these filament proteins, with their long halflife and their relatively slow turnover when compared to actin and tubulin. The microscopic morphology of intermediate filaments suggests a role for these networks in physically reinforcing cells, a hypothesis that was greatly strengthened by finding mutations in several classes of intermediate filament genes that lead to cell fragility and pathological consequences (for reviews, see Corden and McLean, 1996; Quinlan et al., 1999). The first identified of these disorders, and the best studied, is epidermolysis bullosa simplex (EBS). In this disorder, mutations in keratins K5 or K14, expressed in the basal cells of stratified epithelia, lead to fragile epidermal basal cells that lyse upon mild physical trauma to produce intraepithelial blisters (Bonifas et al., 1991; Coulombe et al., 1991; Lane et al., 1992). The clinical severity of the EBS phenotype varies from case to case and can be largely correlated with the position of the mutation in the keratin gene (Letai et al., 1993). Mutation ‘hotspots’ at either end of the central α-helical rod domains are associated with the most severe type of EBS (the Dowling-Meara form), whereas several cluster sites elsewhere

pathways were induced faster, as seen by early activation of JNK, ATF-2 and c-Jun. We demonstrate that the speed of a cell’s response to hypotonic stress, by activation of the SAPK/JNK pathway, is correlated with the clinical severity of the mutation carried. The response to hypo-osmotic shock constitutes a discriminating stress assay to distinguish between the effects of different keratin mutations and is a potentially valuable tool in developing therapeutic strategies for keratin-based skin fragility disorders.

Key words: Epidermolysis bullosa simplex, Keratins, Keratinocyte cell lines, Osmotic shock, Stress response

in the keratin protein are associated with milder (WeberCockayne) forms of the disorder affecting predominantly hands and feet (for a review, see Lane, 1994). EBS is however only one of many epithelial fragility disorders now known to be caused by mutations in keratin genes (Irvine and McLean, 1999). We recently characterized a set of keratinocyte cell lines, derived from EBS patients, as potential model systems in which to analyze and try to reverse the impact of defective keratin intermediate filaments on cell function (S.M.M., I. M. Leigh, E.B.L. et al., manuscript in preparation). These cells are robust in culture and amenable to various types of stress assays and have already been tested for cell spreading and response to heat shock (Morley et al., 1995). In such assays the mutant keratin lines were slower to spread out after re-plating than the control cells with wild-type keratins. The mutant keratin filament networks showed a reduced stability, with a tendency to break up into small aggregates upon thermal stress. This assay was, however, unable to distinguish between cells carrying mutations with different degrees of clinical severity. In this paper we describe the effects of a different type of stress, hypo-osmotic shock, on keratinocytes expressing EBS-associated keratin mutations. Hypo-osmotic shock is a common physiological stress that body tissues encounter, and it leads to transient cell swelling. The extreme cell-shape changes induced by this shock led us to investigate the possibility that the status of the cytoskeleton, and specifically the keratin filaments, is important for cell survival. Volume-

4342

Journal of Cell Science 115 (22)

regulatory responses to osmotic shock have been reported to be sensitive to pharmacological disruption of the cytoskeleton in a number of cell models (Cornet et al., 1994; Foskett and Spring, 1985). We observed that exposure to hypo-osmotic stress, which causes a rapid cell swelling and loss of membrane surface elaborations, is characterized by rapid changes in the configuration of actin and tubulin, and in the case of severely mutant cells, in intermediate filaments as well. Monitoring the early stress response to hypo-osmotic shock by tracking its effect on JNK activation revealed a faster response in EBS keratinocyte cell lines, with a direct correlation between the severity of the keratin mutation carried in these cell lines and the speed of JNK activation. These observations have a bearing on our understanding of the function of integrated cytoskeleton systems, as well as practical applications in gene therapy. Materials and Methods Cell culture Five keratinocyte cell lines, expressing EBS-associated keratin mutations of differing severity, were compared (Table 1). KEB-1, KEB-2 [both K5 E475G (Lane et al., 1992)] and KEB-7 [K14 R125P (S.M.M., unpublished)] represent severe (Dowling-Meara) disease phenotypes, whereas KEB-3 and KEB-4 [both K14 V270M (Rugg et al., 1993)] derive from a milder Weber-Cockayne form of EBS. KEB1, KEB-2 and KEB-3 cells were immortalized with a temperaturesensitive SV-40 T antigen (tsA 58), HPV16 (E6^E7) was used to immortalize KEB-7 and KEB-4 and also the control (wild-type) keratinocyte line, NEB-1 (S.M.M., unpublished). All the experiments described in this paper were carried out on mutant and control cell lines between 10-15 passages after immortalization. The cell lines were cultured in DMEM with 25% Ham’s F12 medium, 10% FCS, plus additional growth factors hydrocortisone (0.4 µg/ml), cholera toxin (10–10 M), transferrin (5 µg/ml), lyothyronine (2×10–11 M), adenine (1.9×10–4 M) and insulin (5 µg/ml). All cell lines were fibroblast feeder cell independent and were cultured at 36.5°C in 5% CO2. They were grown in plastic tissue culture grade dishes and in 96 microwell plates for osmotic shock and on 13 mm glass coverslips for light microscopy. Osmotic shock Cells were cultured for 48 hours to reach 80% confluence and then subjected to hypo-osmotic shock by immersion in 150 mM urea for 5 minutes at 37°C. Cells were then returned to normal tissue culture medium for varying periods of recovery, before further analysis. Keratinocytes are quite resistant to hypo-osmotic effects of water, diluted DMEM or PBS, so urea was added (Kucerova and Strbak,

Table 1. Cell lines used in this study Cell line

NEB-1

KEB-1

KEB-2

KEB-3

KEB-4

KEB-7

Phenotype

Normal

DM severe

DM severe

WC mild

WC mild

DM severe

Mutation

– –

K5 E475G

K5 E475G

K14 V270M

K14 V270M

K14 R125P

Stress response*

1

3

3

2

2

4

DM, Dowling-Meara form; WC, Weber-Cockayne form of EBS (see text). *Ranked order of strength and speed of stress response activation, 1 is the lowest.

2001). This small permeant molecule diffuses rapidly into the cells, causing water to enter and increase the rate of cell swelling (Stein, 1990). Cell survival was not affected by this procedure. To assess the effect of cell confluence on sensitivity to osmotic shock, different concentrations of NEB-1 cells were grown in three 96-well plates and tested at 40%, 80% and 100% confluence. The recovery of the metabolic activity of the cell was assessed at 1 hour, 3 hours and 24 hours after osmotic shock, using the CellTiter 96® AQueous One Solution Cell Proliferation Assay (Promega). The same assay was used to measure the recovery of metabolic activity of the cell lines KEB-4, KEB-7 and NEB-1 after osmotic shock. The cells were grown in 96-well plates, subjected to hypoosmotic shock and then measured before and immediately after osmotic shock and at 30 minutes, 3 hours, 4 hours, 24 hours, 48 hours, 5 days, 6 days, 7 days and 8 days after osmotic shock. Absorbance was measured at λ=490 nm and compared with readings from untreated control cells. Microscopy For scanning electron microscopy, cells were grown on 6 mm coverslips and subjected to hypo-osmotic shock as described above. Both test and control cells were fixed with 4% paraformaldehyde for 10 minutes, washed in PBS and dehydrated with absolute alcohol and then acetone. They were critical point dried in a Polaron E3000 series II unit, sputter-coated (Polaron ES100 series II coating unit) with gold-palladium and viewed using a Joel JSM 35 scanning electron microscope operated at 15kV. For light microscopy, cells were grown on 13 mm glass coverslips and processed for immunostaining after 2 minutes and 4 minutes of osmotic shock, and at 10 minutes and 4 hours recovery in normal medium, after 5 minutes osmotic shock. Cells were fixed in 3% paraformaldehyde for 10 minutes at 4°C, permeabilized in 100% methanol-acetone (1:1) for 10 minutes at –20°C and treated with blocking buffer (5.5% normal goat serum in PBS) for 1 hour at room temperature. The cells were stained with mouse monoclonal antibodies LL001 [to K14 (Purkis et al., 1990)], α-tubulin (Amersham, dilution 1:200) and phospho-SAPK/JNK (dilution 1:200; New England Biolabs), and with rabbit polyclonal antibodies BL18 [to K5, dilution 1:200 (Purkis et al., 1990)] and SAPK/JNK (dilution 1:200; New England Biolabs). SAPK/JNK antibody detects the total SAPK/JNK levels and is phosphorylation independent, whereas the phospho-SAPK/JNK antibody (Thr183/Tyr185) detects, after stress, the doubly phosphorylated threonine 183 and tyrosine 185 of JNK1 (p46) and JNK2 (p54) (Kyriakis et al., 1994; Rochat-Steiner et al., 2000; Weiss et al., 2000; Shukla et al., 2001). Rhodamine-conjugated phalloidin (Molecular Probes, Leiden, Netherlands; dilution 1:150) was used to detect actin. Mouse monoclonal antibodies were detected with FITC-conjugated sheep anti-mouse serum (Sigma Aldrich, UK, F3008) at 1:50 dilution; rabbit antibodies were detected using an Alexa-fluorescein-conjugated goat anti-rabbit serum (Molecular Probes, Leiden, Netherlands; dilution 1:1000). Coverslips were mounted onto glass slides using CitiFluor (Sigma Aldrich, UK), air-dried and visualized with a Zeiss Axioplan epifluorescence microscope. Confocal images were collected using a BioRad MRC 600 confocal laser-scanning microscope. Immunoblotting Cell lysates were extracted after 0 minutes, 2.5 minutes, 5 minutes, 10 minutes, 30 minutes, 1 hour and 2 hours of recovery from osmotic shock. Cells were kept on ice and washed twice with ice-cold PBS. Lysis buffer, containing 100 mM Tris (pH 7.4), 50 mM βglycerolphosphate, 0.5 mM Na3VO4, 1 mM ethyleneglycol-bis (β-aminoethyl ether)-N,N,N′,N′-tetra-acetic acid, 1% Triton, 1 µM microcystin and one complete (EDTA-free) protease inhibitor cocktail tablet (Roche) per 50 ml lysis buffer, was added and left on ice for 5

The role of keratins in response to osmotic shock minutes. Cells were then harvested and centrifuged at 13,000 g for 10 minutes at 4°C. The protein concentration of the supernatant lysate was determined by Bradford assay (BioRad protein assay kit) and standardized using bovine serum albumin. For sodium dodecylsulphate (SDS) polyacrylamide gel electrophoresis, 10 µg of each of the samples was mixed with SDS sample buffer (4% SDS, 20% glycerol, 20 mM Tris-HCl (pH 6.8), 20 mM dithiothreitol and 1% Bromophenol Blue), boiled for 5 minutes and separated on SDS polyacrylamide gels. For immunoblotting, protein transfer was carried out at 300 mA for 60 minutes. Nonspecific binding to the transfer membrane was blocked by incubation in Tris-buffered saline (pH 7.6), containing 0.5% TWEEN-20 and 5% BSA (blocking buffer), for at least 3 hours, with gentle agitation. Blots were then incubated overnight at 4°C with JNK/SAPK antibody or phospho-JNK/SAPK or phospho-ATF-2 antibodies (dilution 1:1000; New England Biolabs) in 10 ml blocking buffer. The membranes were washed with TBS with 0.5% TWEEN-20, and bound primary antibodies were detected using a secondary antibody of swine antirabbit Ig antiserum for JNK/SAPK and phospho-ATF-2, and a secondary antibody of rabbit anti-mouse Ig antiserum for phosphoJNK/SAPK. The secondary antibodies were conjugated to horseradish peroxidase (diluted 1:2000 in 10 ml blocking buffer) and visualized by electrochemiluminescence.

Results The effect of osmotic shock on the cytoskeleton Following immersion in 150 mM urea, cells swell rapidly as water is drawn into them (Fig. 1A,B). We found urea to be a fast and effective way to induce osmotic shock that was no more damaging than distilled water or diluted medium. By scanning electron microscopy, it can be seen that this is correlated with a loss of surface detail of the apical plasma

Fig. 1. Hypo-osmotic shock causes rapid cell swelling (B) and loss of membrane surface elaborations (D). Cells before (A,C) and after (B,D) 5 minutes exposure to hypo-osmotic medium. (A,B) Phase contrast microscopy (Bar, 10 µm) and (C,D) scanning electron microscopy (Bar, 1 µm).

4343

membrane, as microvilli and filopodia are lost and the stretching plasma membrane becomes smooth (Fig. 1C,D). Within the cell, these changes correlate with pronounced reorganization of cytoskeleton proteins (Fig. 2). By immunofluorescence microscopy, it was observed that the microtubules became fragmented within 2 minutes of the initiation of stress (Fig. 2B). This was rapidly reversible, and microtubules were clearly repolymerizing after 4 minutes of urea exposure (Fig. 2C). Ten minutes after return to normal medium, the microtubules were indistinguishable from control cells (Fig. 2D). This fast recovery probably explains why microtubule disturbances were not reported in earlier studies (Cornet et al., 1994; Lang et al., 1998). Actin reorganization was also evident. Concomitant with the loss of surface detail, there was a depletion of the actin associated with the apical cell membrane (Fig. 2G). F-actin appeared to be relocated to sites further into the cell’s interior, and large actin filament bundles were often apparent deep in the cell. After 2 minutes of exposure to urea, residual filopodia were still evident at the free cell margins (Fig. 2G). At this time, the regulatory volume decrease (RVD) would have been initiated, by the activation of ion pumps in the membrane, and the cells were beginning to shrink again. Actin became increasingly reassociated with the cell surface with time (Fig. 2H,I), and cells had all regained their normal morphology by 4 hours after osmotic shock (Fig. 2J). All the above changes were seen equally in wild-type or mutant keratin cells. No signs of keratin fragmentation or disassembly were detected in wild-type cells during the shock and recovery periods (data not shown), although the profound cell-shape changes triggered by osmotic shock must displace filament systems to some extent in all cells. By contrast, the KEB-7 cell line (severe phenotype EBS mutation) showed a degree of keratin filament fragmentation following osmotic shock (Fig. 2K-O) that could be monitored by immunofluorescence. These changes developed more slowly than the changes in actin and microtubules. Tiny keratin aggregates started to form at the cell periphery and were clearly seen 10 minutes after osmotic shock, as shown by staining for keratin 14 (Fig. 2N). After 4 hours, the aggregates appeared bigger (Fig. 2O). The aggregates only ever represented a small proportion of the total keratin in the cell, and the majority of keratin intermediate filaments appeared intact throughout the time course of observation. Keratin aggregate formation depends on mutation severity In EBS patients, the position of the mutation in the keratin protein is generally correlated with the severity of the disease phenotype. When cell lines expressing different keratin mutations were subjected to osmotic stress, differences were seen. Fig. 3 shows this comparison, illustrating the integrity of the keratin filament network 4 hours after stress. Clear aggregates were seen in cells with a severe mutation (Fig. 3F), whereas less severe mutations were associated with some filament fragmentation, but no aggregates were present (Fig. 3B,C). Mild mutations (Fig. 3D,E) showed little deviation from controls (Fig. 3A).

4344


Fig. 2. Changes in the cytoskeleton following exposure to hypotonic medium. Tubulin disruption recovers more rapidly than actin and keratin intermediate filament systems. KEB-7 cells, an EBS-derived cell line expressing the K14 R125P mutation, were fluorescently stained for tubulin (A-E, anti-α tubulin), actin (F-J, phalloidin) and intermediate filaments (K-O, antibody LL001 to K14). Cells were exposed to 150 mM urea for 5 minutes then transferred back to normal tissue culture medium for recovery. Cell were fixed and stained before hypo-osmotic shock (A,F,K), after 2 minutes (B,G,L) and 4 minutes (C,H,M) exposure to urea, and at 10 minutes (D,I,N) and 4 hours (E,J,O) recovery after osmotic shock. Bar, 10 µm.

The role of keratins in response to osmotic shock

4345

Fig. 3. Different effect of osmotic shock on the keratin cytoskeleton in the five EBS-derived cell lines studied. All cell lines were subjected to hypo-osmotic shock and fluorescently stained for K14 (antibody LL001) after 4 hours recovery. (A) NEB-1 (wildtype); (B) KEB-1, (C) KEB-2, (D) KEB-3, (E) KEB-4, (F) KEB-7. Some filament fragmentation was seen in KEB-1 and KEB-2 (arrows in B and C) and clear peripheral aggregates were seen in KEB-7 (arrows in H). Bar, 10 µm.

Cell recovery after hypo-osmotic shock Following osmotic shock, there was a similar cell loss rate in all the cell lines tested, and an 80-85% drop of metabolic activity was detected in all the cultures within 24 hours. A clear difference was observed in the cell lines’ long-term recovery from osmotic shock, depending on the state of their keratins. The cell lines expressing EBS mutations in K5 or K14 took longer for their metabolic activity to build up to pre-osmotic shock levels compared with cells expressing wild-type keratins. In the case of the wild-type control cell line NEB-1, for example, metabolic activity of the culture had reached its pre-shock level by 48 hours, whereas it took 6 days for the cell line carrying a mild EBS mutation (KEB-4) to recuperate. The cell line with the most severe mutation (KEB-7, K14 R125P mutation) had still not recovered by 8 days (Fig. 5).

Different levels of JNK and phospho-JNK in mutant keratin cell lines In view of the variable cell recovery, it was decided to restrict the investigation to immediate or short-term effects of osmotic shock on the cells, to seek a more consistent, cell-number70

60

cell metabolic activity (%)

Sensitivity to hypo-osmotic shock varies with culture confluence Although the short-term effects of osmotic shock were reproducible, initial experiments indicated that longer-term survival of the cells was quite variable and dependent on the time elapsed since last trypsinization and re-plating of the cells and on the degree of confluence of the culture at the time the osmotic shock was applied. Comparison of cell survival after osmotic shock between cells plated at different densities revealed that cells that were 80% confluent upon hypo-osmotic shock were more sensitive to this stress than cells that were either 40% confluent or 100% confluent (Fig. 4).

50

40

30

20

10

0

40% 80%100% 1 hr after OS

40% 80%100%

40% 80%100%

3 hr after OS

24 hr after OS

time after OS

Fig. 4. The effect of cell confluence on sensitivity to osmotic shock. NEB-1 (wild-type keratin) cells were grown to different degrees of confluence (40%, 80% and 100%), before subjecting them to hypoosmotic shock; metabolic activity was measured at 1, 3 and 24 hours of recovery (CellTiter 96, Promega). At all time points, cultures at 80% confluence showed the greatest sensitivity to osmotic shock.


independent readout. We therefore examined the speed of cells’ response to this shock, as reflected by induction of the stress-activated protein kinases (SAPKs) pathway, which is known to be triggered by hypo-osmotic stress (Haussinger, 1996b). On assessing JNK/SAPK activation, by immunoblotting with antibodies to JNK and phosphorylated (activated) JNK, we confirmed the induction of this pathway and noted in addition some differences in the response of cells expressing mutant keratins. Firstly, the total amount of JNK was higher in cell lines with ‘severe’ mutations (KEB1, KEB-2 and KEB-7) than either in the control (NEB-1) or in the ‘mild’ mutant lines (KEB-3 and KEB-4) (Fig. 6). Densitometric measurement of JNK levels showed that this protein was about seven times more concentrated in KEB-7 than in NEB-1 (data not shown). JNK activation also varied between the cell lines, as shown by immunoblot analysis with a phospho-JNK/SAPK antibody (New England Biolabs). A weak signal was detected in the control cell line NEB-1 (wild-type keratin) after 2.5 minutes of recovery from osmotic shock. This signal peaked at around 10 minutes into the recovery period and then gradually reduced over the next three hours (Fig. 6). Two phospho-JNK isoforms, identified by the New England Biolabs antibody as phospho-JNK1 and -JNK2, were detectable before osmotic shock in the severe mutant cell lines (Fig. 6). These bands had both increased greatly in intensity by 2.5 minutes recovery after osmotic shock, and a

Fig. 6. Induction of JNK phosphorylation following hypo-osmotic shock. JNK1 (p46) and JNK2 (p54) were detected by immunoblotting with rabbit antisera to native JNK or with mouse antisera against phospho-JNK pJNK1 and pJNK2 (New England Biolabs), as indicated and visualized by chemiluminescence (see Materials and Methods). Cell lines varied in the time course of JNK1 and JNK2 phosphorylation. Severely mutated cell lines have higher levels of JNK1 and JNK2 than milder mutants or controls, and phosphorylation of JNK is induced faster. B, before OS; 0, start of recovery time after 5 minutes osmotic shock.

120

cell metabolic activity (%)

4346

100

NEB-1 KEB-4 KEB-7

80

60

40

20

0 before OS

0

30 min

1 hr

3 hr

4 hr

1 day

2 days

5 days

6 days

7 days

8 days

time after OS Fig. 5. Cell viability after hypo-osmotic shock (OS). Cells were assayed for metabolic activity (CellTiter 96, Promega) at different time intervals of recovery after 5 minutes exposure to 150 mM urea. Metabolic activity is presented as a percentage of the activity in the untreated control cell population. Wild-type cell cultures returned to starting levels of metabolic activity after 2 days; KEB-4 reached this level in 6 days, and KEB-7 cultures showed no recovery in 8 days.

strong signal was still present 1 hour after osmotic shock. Among all the EBS cell lines tested, KEB-7 (K14 R125P, the one carrying the most severe mutation) consistently showed the

The role of keratins in response to osmotic shock highest levels of phospho-JNK, before, during and after osmotic shock. KEB-1 and KEB-2 (K5 E475G) were the next highest. KEB-3 and KEB-4 (K14 V270M) gave the weakest signal, with the two phospho-isoforms present after 2.5 minutes recovery from osmotic shock, and greatly increased by 5 minutes after shock (Fig. 6). All cells reached a peak of activation after 10 minutes recovery from osmotic shock. This gradation in response to osmotic shock between the cell lines is directly correlated with the gradation in clinical severity of the EBS seen in patients with these mutations (see Table 1). To determine whether JNK phosphorylation was truly a reflection of the JNK/SAPK pathway activation, we looked for effects on known downstream targets of this kinase. The ATF2 transcription factor is a well documented substrate for phosphorylation by activated JNK (Ip and Davis, 1998; Leppa and Bohmann, 1999). Immunoblotting revealed the appearance of phosphorylated (activated) ATF-2 (pATF-2) with the same time course observed for JNK phosphorylation (Fig. 7). In the control cell line, a weak signal was observed at 5 minutes recovery after osmotic shock. This signal increased after 10 minutes and had the strongest peak after 30 minutes (Fig. 7), in keeping with its activation downstream of JNK. In the KEB3 and KEB-4 cell lines, a weak signal was present after 2.5 minutes of recovery from osmotic shock. This then increased progressively, until reaching the highest intensity 30 minutes after osmotic shock (Fig. 7). In the cell lines carrying a more severe mutation (KEB-1, KEB-2, and KEB-7), there was already a signal before osmotic shock, which reached its maximum intensity 10-30 minutes after the shock and remained very strong at 3 hours (Fig. 7). Activation of this pathway was further confirmed by assaying phosphorylated cJun levels, as c-Jun is another downstream target of JNK. Three cell lines (NEB-1, KEB-4 and KEB-7) were tested, and the cJun phosphorylation pattern observed was identical to that observed for ATF-2 (data not shown). It appears therefore that the mutated cell lines have a more rapid activation of JNK in response to osmotic shock than the controls and that there is a direct correlation between the severity of the mutation and the speed of JNK activation.

Fig. 7. Induction of ATF-2 phosphorylation following hypo-osmotic shock. ATF-2 activation was detected by immunoblotting with rabbit antisera to phospho-ATF-2 (pATF2) (New England Biolabs) and visualized by chemiluminescence. Cell lines varied in the time course of ATF-2 phosphorylation. In severely mutated cell lines phosphorylation of ATF-2 was induced faster.

4347

Moreover, the presence of a signal, even before osmotic shock, in KEB-1, KEB-2 and KEB-7, seems to indicate that the cell lines with a severe mutation have constitutively activated the JNK/SAPK pathway. Cellular localization of JNK and phospho-JNK The distribution of JNK, and its relocation after activation, was monitored by immunofluorescence on cultured cells. Before osmotic shock, JNK was localized at the periphery of wild-type cells, often with a distribution suggestive of focal contacts (Fig. 8A). After activation (phosphorylation), JNK was found patchily within the nucleus, both in the control and in the KEB7 mutant cell line (Fig. 8B,D). No phospho-JNK was detected in control lines before stress (Fig. 8E), but it was present in the nucleus of KEB-7 both before and after osmotic stress (Fig. 8G,H), again confirming its constitutive activation in these cells. After osmotic shock, phospho-JNK was also detected in the nucleus of the control cell line (Fig. 8F). These results concur with the immunoblotting data. Discussion The first stage in the development of any new therapy for keratin-based skin fragility disorders, such as EBS, will depend on in vitro culture studies with cell lines expressing EBSassociated mutations. Useful cell model systems must be able to distinguish between the effects of different mutations, since these may require different clinical strategies to treat them. Differences in response to heat shock were reported between keratinocytes expressing wild-type versus EBS mutant keratins (Morley et al., 1995), but this assay was unable to distinguish between EBS mutations of different severity. We therefore sought to develop a more discriminating stress assay, on the basis of hypo-osmotic shock. During initial investigations it became clear that the outcome of the osmotic stress was dependent on the state of confluence of the epithelial cells, to the extent that long-term reproducibility of the assay became hard to establish. Thus, we concentrated on the early effects of

4348


osmotic stress and its consequent changes in c-Jun N-terminal kinase (JNK/SAPK) pathway, as a marker for cellular response to stress in the presence of pathogenic keratin mutations. Hypo-osmotic stress is a common physiological situation with which tissue cells have to cope. Ischaemia and the resulting oxygen deprivation, for example, lead to energy depletion and loss of function of ion pumps in the cell membrane, and so to a rise in the intracellular osmolarity. Cell swelling follows rapidly, with a substantial but transient distortion of the cell architecture. Most mammalian cells have developed compensatory mechanisms to respond to osmotic changes and within minutes will have upregulated the activity of ion channels to restore the osmotic balance. This process, known as regulatory volume decrease (RVD), includes the concerted opening of K+ and Cl– selective ion channels, leading to an efflux of KCl, which causes the loss of cellular water (Haussinger, 1996a; Okada et al., 1994). By contrast, if cells are exposed to hypertonic extracellular fluid, they initially shrink but then return to their normal volume, by a regulatory cell volume increase (RVI), achieved by the activation of ion pumps and carriers. The hypo-osmotic shock was carried out on 80% confluent cells, which were more sensitive to this stress than cells that were either 40% confluent or 100% confluent. As cells become confluent, they establish junctions with all their neighboring cells, and with increasing time these desmosomes ‘mature’ and differentiate to a more stable state, in which for example they cannot be disrupted by removal of calcium from the medium (Wallis et al., 2000). It is possible that the greater sensitivity of cells at subconfluence is due to the transition between cell dependence on actin-mediated attachment junctions to intermediate-filament-mediated attachment junctions, with maturation of desmosomes still incompletely established at the time of stress. Further studies will be needed to understand this observation. Different effects of osmotic shock on actin, tubulin and intermediate filaments Of the three cytoskeleton filament systems, the role of the actin network in response to anisosmotic conditions has received the most attention. There is general agreement that hypo-osmotic shock leads to a transient actin reorganization, with which we concur. This has been variously reported as an increase in Factin (Czekay et al., 1994; Mountain et al., 1998; Tilly et al., 1996) or a decrease in G-actin (Theodoropoulos et al., 1992), with evidence for Rho kinase signaling involvement (Koyama et al., 2001; Tilly et al., 1996). Our time course observations suggest that actin is shifted from filaments subtending membrane structures such as microvilli to stress fiber-like bundles of filaments deeper in the cytoplasm. An intact actin filament cytoskeleton appears to be required for activation of at least some of the volume regulatory mechanisms. Cell swelling requires actin mobilization and is accompanied by a calcium surge, either or both of which may be involved in the activation of membrane channels (Cornet et al., 1993; Foskett and Spring, 1985; Jessen and Hoffmann, 1992; Sudlow and Burgoyne, 1997). The resulting RVD response has been shown to be inhibited by both cytochalasin B (actin depolymerization) (Pedersen et al., 1999) and phalloidin (F-actin stabilization) (Shen et al., 1999).

Fig. 8. Cellular localisation of JNK changes upon phosphorylation. NEB-1 wild-type cells (A,B,E,F) and KEB-7 mutant K14 cells (C,D,G,H) were subjected to hypo-osmotic shock and stained for JNK (A,B,C,D) and phospho-JNK (E,F,G,H). In wild-type cells, native JNK was localized at the cell periphery, in a pattern typical of focal adhesions (A), but, upon phosphorylation, JNK left the focal adhesions and moved to the nucleus (B). In the mutant cells, nuclear JNK was detectable even before osmotic shock (C). Phospho-JNK (pJNK) was absent in wild-type cells before osmotic shock (E) and was only detectable after the shock (F); but it was constitutively present in mutant cells even before osmotic shock (G). Bar, 10 µm.

Release of actin from its membrane tethering proteins (e.g. proteins in the focal adhesion or focal contact zones) could initiate the signal cascade that triggers the stress response, possibly by integrin-mediated detection of membrane stretching (Low et al., 1997; Low and Taylor, 1998). The movement of actin away from the plasma membrane may also

The role of keratins in response to osmotic shock directly help the cell to survive a sudden intracellular increase in pressure, by releasing plasma membrane to accommodate volume expansion and so prevent the cell from rupturing. Raucher and Scheetz have demonstrated the existence of a reservoir of plasma membrane with great capacity for expansion, which is normally tethered by the cytoskeleton (Raucher and Sheetz, 1999). The physical location of such membrane reservoirs may actually be in the numerous small microvillar structures that vanish from cell surfaces upon osmotic shock (Fig. 1). These microvilli and microridges decorate the free apical surface of epithelial tissues in aqueous environments, from internal mammalian epithelia to the epidermal surface of aquatic vertebrates (e.g. Lane and Whitear, 1982). They are readily seen by scanning electron microscopy, but there has been no satisfactory explanation of their existence to date. They may exist to protect epithelial cells against hypotonic damage. We observed that osmotic shock has a wider effect on the cytoskeleton than previously described, as even microtubules are temporarily disrupted. Hypo-osmotic shock causes a dramatic but transient tubulin depolymerization, which is rapidly reversed. Tubulin has been reported to be depolymerized by increased pressure (Engelborghs et al., 1976; Larsen et al., 2000), and the time course of tubulin disruption observed here, that is, disruption within 2 minutes and recovery by 4 minutes, is certainly compatible with a direct linkage to the transient pressure increase inside the cell. Other studies commenting on the effect of osmotic shock on microtubules have suggested that perturbation of microtubule kinetics either has no effect on, or inhibits, the RVD (Downey et al., 1995; Edmonds and Koenig, 1990; Shen et al., 1999; Zhang et al., 1997), and that hypotonicity may stabilize microtubules and stimulate tubulin synthesis (Haussinger et al., 1994). Few studies have been undertaken to date on the response of intermediate filaments to anisosmotic conditions. Astrocytes devoid of any intermediate filaments (GFAP–/–, vimentin–/–) had a less effective response to osmotic stress, as measured by taurine efflux (Ding et al., 1998). Our results also suggest a role for keratin intermediate filaments in protecting cells against osmotic stress. A keratin cytoskeleton network, compromised by dominant-negative mutations known to be pathogenic in situ, will also render cells less able to cope with, and recover from, osmotic stress. The different responses to osmotic stress that we observe in the three cytoskeleton filament systems also point to another functional aspect of intermediate filament evolution. If microtubules are intrinsically unable to withstand pressure, and filamentous actin needs to be remodeled to accommodate cell swelling, this leaves the intermediate filament network as the only one of the three cytoskeleton filament systems that can be maintained through hypo-osmotic stress, to provide a reference framework for the steady state of the cell. Osmotic shock is such a universal stress in multicellular organisms, especially on the epithelial surface, that this may be one of the driving forces behind the evolution of the intermediate filament cytoskeleton.

4349

signal-regulated kinases ERK-1 and ERK-2 (Haussinger, 1996b; Lang et al., 1998). JNK is known to be involved in cellular responses to environmental stresses. These include UV light (Derijard et al., 1994), γ-irradiation (Chen et al., 1996), osmotic shock (Niisato et al., 1999) and mechanical stress (Kippenberger et al., 2000). JNK activation is also induced by mitogenic signals, including growth factors, oncogenic Ras and CD40 ligation. Targets of JNK signaling pathways include the transcription factors ATF-2 and c-Jun, which are members of the basic leucine zipper (bZIP) group that bind to AP-1 and AP-1-like sites (Whitmarsh and Davis, 1996) within the promoters of many genes. JNK phosphorylates two sites within the activation domain of each transcription factor, leading to an increased transcriptional activity. In this paper, we have monitored the osmotic stress response of cell lines carrying a mutation in K5 or K14 by tracking the progress of JNK phosphorylation. Using immunocytochemical analysis, we confirmed that, once activated, JNK moves toward the nucleus, where it can signal to nuclear transcription factors and thus activate transcription. We present evidence for a direct correlation between the severity of the mutation carried in the cell line and the speed of JNK activation. This is the first cell culture assay with a read-out that varies with the graded severity of the mutations in the clinic and should therefore be a useful monitor of the effect of mutations. In addition, we observed that the severely mutated cell lines (only) appear to be in a permanent ‘stress’ condition, as they exhibit a significant level of JNK activation even without osmotic shock. We speculate that this may be due to the increased physiological burden on the cell, incurred by the necessity of handling a defective keratin protein. This aspect of EBS cell physiology may have a direct bearing on the clinical phenotype of epidermolysis bullosa simplex. As we come to understand the molecular basis of keratin disorders, we are faced with the challenge of devising practical and effective therapy for these genetic diseases. Any new therapeutic strategies will need extensive preclinical development work, beginning with in vitro model systems, in which the defects can be induced and then repaired. In conjunction with the appearance of cell culture models, response to hypo-osmotic shock is a practical surrogate measure of the mutation burden on keratinocytes. We expect that this will be useful in monitoring the effect of gene therapy on EBS keratinocyte cell lines. This work was supported by the Dystrophic Epidermolysis Bullosa Research Association, D.E.B.R.A., (LANE3 to E.B.L., supporting M.D.) and by the Cancer Research UK (SP2060/0103 to E.B.L., supporting S.M.M.). D.R. is supported by a Medical Research Council studentship (G78/6810). We are grateful to John James for his help with the scanning electron microscopy and to Martyn Ward and Peter Taylor for their helpful discussions on the osmotic shock procedure. A special thank you to Mirjana Liovic and Mandy Gulbransen for their help and support.

References Stress response activation by osmotic shock Hypo-osmotic stress triggers the MAPK pathway, leading to activation of the Jun-NH2-terminal kinase (JNK/SAPK) pathway, via MAPK kinase (MKK4) or via the extracellular

Bonifas, J. M., Rothman, A. L. and Epstein, E. H., Jr (1991). Epidermolysis bullosa simplex: evidence in two families for keratin gene abnormalities. Science 254, 1202-1205. Chen, Y. R., Meyer, C. F. and Tan, T. H. (1996). Persistent activation of cJun N-terminal kinase 1 (JNK1) in gamma radiation-induced apoptosis. J. Biol. Chem. 271, 631-634.

4350


Corden, L. D. and McLean, W. H. (1996). Human keratin diseases: hereditary fragility of specific epithelial tissues. Exp. Dermatol. 5, 297-307. Cornet, M., Lambert, I. H. and Hoffmann, E. K. (1993). Relation between cytoskeleton, hypo-osmotic treatment and volume regulation in Ehrlich ascites tumor cells. J. Membr. Biol. 131, 55-66. Cornet, M., Isobe, Y. and Lemanski, L. F. (1994). Effects of anisosmotic conditions on the cytoskeletal architecture of cultured PC12 cells. J. Morphol. 222, 269-286. Coulombe, P. A., Hutton, M. E., Letai, A., Hebert, A., Paller, A. S. and Fuchs, E. (1991). Point mutations in human keratin 14 genes of epidermolysis bullosa simplex patients: genetic and functional analyses. Cell 66, 1301-1311. Czekay, R. P., Kinne-Saffran, E. and Kinne, R. K. (1994). Membrane traffic and sorbitol release during osmo- and volume regulation in isolated rat renal inner medullary collecting duct cells. Eur. J. Cell Biol. 63, 20-31. Derijard, B., Hibi, M., Wu, I. H., Barrett, T., Su, B., Deng, T., Karin, M. and Davis, R. J. (1994). JNK1: a protein kinase stimulated by UV light and Ha-Ras that binds and phosphorylates the c-Jun activation domain. Cell 76, 1025-1037. Ding, M., Eliasson, C., Betsholtz, C., Hamberger, A. and Pekny, M. (1998). Altered taurine release following hypotonic stress in astrocytes from mice deficient for GFAP and vimentin. Brain Res. Mol. Brain Res. 62, 77-81. Downey, G. P., Grinstein, S., Sue, A. Q. A., Czaban, B. and Chan, C. K. (1995). Volume regulation in leukocytes: requirement for an intact cytoskeleton. J. Cell Physiol. 163, 96-104. Edmonds, B. T. and Koenig, E. (1990). Volume regulation in response to hypo-osmotic stress in goldfish retinal ganglion cell axons regenerating in vitro. Brain Res. 520, 159-165. Engelborghs, Y., Heremans, K. A., de Maeyer, L. C. and Hoebeke, J. (1976). Effect of temperature and pressure on polymerisation equilibrium of neuronal microtubules. Nature 259, 686-689. Foskett, J. K. and Spring, K. R. (1985). Involvement of calcium and cytoskeleton in gallbladder epithelial cell volume regulation. Am. J. Physiol. 248, C27-C36. Haussinger, D. (1996a). Regulation and functional significance of liver cell volume. Prog. Liver Dis. 14, 29-53. Haussinger, D. (1996b). The role of cellular hydration in the regulation of cell function. Biochem. J. 313, 697-710. Haussinger, D., Stoll, B., vom Dahl, S., Theodoropoulos, P. A., Markogiannakis, E., Gravanis, A., Lang, F. and Stournaras, C. (1994). Effect of hepatocyte swelling on microtubule stability and tubulin mRNA levels. Biochem. Cell. Biol. 72, 12-19. Hesse, M., Magin, T. M. and Weber, K. (2001). Genes for intermediate filament proteins and the draft sequence of the human genome: novel keratin genes and a surprisingly high number of pseudogenes related to keratin genes 8 and 18. J. Cell Sci. 114, 2569-2575. Ip, Y. T. and Davis, R. J. (1998). Signal transduction by the c-Jun N-terminal kinase (JNK)–from inflammation to development. Curr. Opin. Cell Biol. 10, 205-219. Irvine, A. D. and McLean, W. H. (1999). Human keratin diseases: the increasing spectrum of disease and subtlety of the phenotype-genotype correlation. Br. J. Dermatol. 140, 815-828. Jessen, F. and Hoffmann, E. K. (1992). Activation of the Na+/K+/Cl– cotransport system by reorganization of the actin filaments in Ehrlich ascites tumor cells. Biochim. Biophys. Acta 1110, 199-201. Kippenberger, S., Bernd, A., Loitsch, S., Guschel, M., Muller, J., Bereiter-Hahn, J. and Kaufmann, R. (2000). Signaling of mechanical stretch in human keratinocytes via MAP kinases. J. Invest. Dermatol. 114, 408-412. Koyama, T., Oike, M. and Ito, Y. (2001). Involvement of Rho-kinase and tyrosine kinase in hypotonic stress- induced ATP release in bovine aortic endothelial cells. J. Physiol. 532, 759-769. Kucerova, J. and Strbak, V. (2001) The osmotic component of ethanol and urea action is critical for their immediate stimulation of thyrotropinreleasing hormone (TRH) release from rat brain septum. Physiol. Res. 50, 309-314. Kyriakis, J. M., Banerjee, P., Nikolakaki, E., Dai, T., Rubie, E. A., Ahmad, M. F., Avruch, J. and Woodgett, J. R. (1994). The stress-activated protein kinase subfamily of c-Jun kinases. Nature 369, 156-160. Lane, E. B. (1994). Keratin diseases. Curr. Opin. Genet. Dev. 4, 412-418. Lane, E. B. and Whitear, M. (1982). Sensory structures at the surface of fish skin. I. Putative chemoreceptors. Zool. J. Linn. Soc. 75, 141-151. Lane, E. B., Rugg, E. L., Navsaria, H., Leigh, I. M., Heagerty, A. H., Ishida-Yamamoto, A. and Eady, R. A. (1992). A mutation in the conserved

helix termination peptide of keratin 5 in hereditary skin blistering. Nature 356, 244-246. Lang, F., Busch, G. L., Ritter, M., Volkl, H., Waldegger, S., Gulbins, E. and Haussinger, D. (1998). Functional significance of cell volume regulatory mechanisms. Physiol. Rev. 78, 247-306. Larsen, T. H., Dalen, H., Boyle, R., Souza, M. M. and Lieberman, M. (2000). Cytoskeletal involvement during hypo-osmotic swelling and volume regulation in cultured chick cardiac myocytes. Histochem. Cell Biol. 113, 479-488. Leppa, S. and Bohmann, D. (1999). Diverse functions of JNK signaling and c-Jun in stress response and apoptosis. Oncogene 18, 6158-6162. Letai, A., Coulombe, P. A., McCormick, M. B., Yu, Q. C., Hutton, E. and Fuchs, E. (1993). Disease severity correlates with position of keratin point mutations in patients with epidermolysis bullosa simplex. Proc. Natl. Acad. Sci. USA 90, 3197-3201. Low, S. Y. and Taylor, P. M. (1998). Integrin and cytoskeletal involvement in signalling cell volume changes to glutamine transport in rat skeletal muscle. J. Physiol. 512, 481-485. Low, S. Y., Rennie, M. J. and Taylor, P. M. (1997). Involvement of integrins and the cytoskeleton in modulation of skeletal muscle glycogen synthesis by changes in cell volume. FEBS Lett. 417, 101-103. Morley, S. M., Dundas, S. R., James, J. L., Gupta, T., Brown, R. A., Sexton, C. J., Navsaria, H. A., Leigh, I. M. and Lane, E. B. (1995). Temperature sensitivity of the keratin cytoskeleton and delayed spreading of keratinocyte lines derived from EBS patients. J. Cell Sci. 108, 3463-3471. Mountain, I., Waelkens, E., Missiaen, L. and van Driessche, W. (1998). Changes in actin cytoskeleton during volume regulation in C6 glial cells. Eur. J. Cell Biol. 77, 196-204. Niisato, N., Post, M., van Driessche, W. and Marunaka, Y. (1999). Cell swelling activates stress-activated protein kinases, p38 MAP kinase and JNK, in renal epithelial A6 cells. Biochem. Biophys. Res. Commun. 266, 547-550. Okada, Y., Kubo, M., Oiki, S., Petersen, C. C., Tominaga, M., Hazama, A. and Morishima, S. (1994). Properties of volume-sensitive Cl– channels in a human epithelial cell line. Jpn J. Physiol. 44 Suppl. 2, S31-S35. Pedersen, S. F., Mills, J. W. and Hoffmann, E. K. (1999). Role of the Factin cytoskeleton in the RVD and RVI processes in Ehrlich ascites tumor cells. Exp. Cell Res. 252, 63-74. Purkis, P. E., Steel, J. B., Mackenzie, I. C., Nathrath, W. B., Leigh, I. M. and Lane, E. B. (1990). Antibody markers of basal cells in complex epithelia. J. Cell Sci. 97, 39-50. Quinlan, R. A., Sandilands, A., Procter, J. E., Prescott, A. R., Hutcheson, A. M., Dahm, R., Gribbon, C., Wallace, P. and Carter, J. M. (1999). The eye lens cytoskeleton. Eye 13, 409-416. Raucher, D. and Sheetz, M. P. (1999). Characteristics of a membrane reservoir buffering membrane tension. Biophys. J. 77, 1992-2002. Rochat-Steiner, V., Becker, K., Micheau, O., Schneider, P., Burns, K. and Tschopp, J. (2000). FIST/HIPK3: a Fas/FADD-interacting serine/threonine kinase that induces FADD phosphorylation and inhibits fas-mediated Jun NH(2)-terminal kinase activation. J. Exp. Med. 192, 1165-1174. Rugg, E. L., Morley, S. M., Smith, F. J., Boxer, M., Tidman, M. J., Navsaria, H., Leigh, I. M. and Lane, E. B. (1993). Missing links: WeberCockayne keratin mutations implicate the L12 linker domain in effective cytoskeleton function. Nat. Genet. 5, 294-300. Shen, M. R., Chou, C. Y., Hsu, K. F., Hsu, K. S. and Wu, M. L. (1999). Modulation of volume-sensitive Cl– channels and cell volume by actin filaments and microtubules in human cervical cancer HT-3 cells. Acta Physiol. Scand. 167, 215-225. Shukla, A., Timblin, C. R., Hubbard, A. K., Bravman, J. and Mossman, B. T. (2001). Silica-induced activation of c-Jun-NH2-terminal amino kinases, protracted expression of the activator protein-1 proto-oncogene, fra1, and S-phase alterations are mediated via oxidative stress. Cancer Res. 61, 1791-1795. Stein, W. D. (1990). Channels, Carriers, and Pumps: An Introduction to Membrane Transport. San Diego, California: Academic Press. Sudlow, A. W. and Burgoyne, R. D. (1997). A hypo-osmotically induced increase in intracellular Ca2+ in lactating mouse mammary epithelial cells involving Ca2+ influx. Pflugers Arch. 433, 609-616. Theodoropoulos, P. A., Stournaras, C., Stoll, B., Markogiannakis, E., Lang, F., Gravanis, A. and Haussinger, D. (1992). Hepatocyte swelling leads to rapid decrease of the G-/total actin ratio and increases actin mRNA levels. FEBS Lett. 311, 241-245. Tilly, B. C., Edixhoven, M. J., Tertoolen, L. G., Morii, N., Saitoh, Y., Narumiya, S. and de Jonge, H. R. (1996). Activation of the osmo-

The role of keratins in response to osmotic shock sensitive chloride conductance involves P21rho and is accompanied by a transient reorganization of the F-actin cytoskeleton. Mol. Biol. Cell 7, 14191427. Wallis, S., Lloyd, S., Wise, I., Ireland, G., Fleming, T. P. and Garrod, D. (2000). The alpha isoform of protein kinase C is involved in signaling the response of desmosomes to wounding in cultured epithelial cells. Mol. Biol. Cell 11, 1077-1092. Weiss, L., Whitmarsh, A. J., Yang, D. D., Rincon, M., Davis, R. J. and

4351

Flavell, R. A. (2000). Regulation of c-Jun NH(2) terminal kinase (Jnk) gene expression during T cell activation. J. Exp. Med. 191, 139-146. Whitmarsh, A. J. and Davis, R. J. (1996). Transcription factor AP-1 regulation by mitogen-activated protein kinase signal transduction pathways. J. Mol. Med. 74, 589-607. Zhang, J., Larsen, T. H. and Lieberman, M. (1997). F-actin modulates swelling-activated chloride current in cultured chick cardiac myocytes. Am. J. Physiol. 273, C1215-C1224.