- Email: [email protected]
Dual-specificity phosphatases in the hypo-osmotic stress response of keratin-defective epithelial cell lines
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m
w w w. e l s e v i e r. c o m / l o c a t e / y e x c r
Research Article
Dual-specificity phosphatases in the hypo-osmotic stress response of keratin-defective epithelial cell lines Mirjana Liovica,b,⁎, Brian Leec , Marjana Tomic-Canicd , Mariella D'Alessandrob , Viacheslav N. Bolshakove , E. Birgitte Laneb,f a
National Institute of Chemistry, Ljubljana, Slovenia CRUK Cell Structure Research Group, University of Dundee College of Life Sciences, MSI/WTB Complex, Dundee, UK c Genentech, San Francisco, USA d Hospital for Special Surgery at Weill Medical College of the Cornell University, New York, USA e University of Reading BioCentre, Reading, UK f A⁎STAR Institute of Medical Biology, Immunos, 8A Biomedical Grove, Singapore b
ARTICLE INFORMATION
ABS T R AC T
Article Chronology:
Although mutations in intermediate filament proteins cause many human disorders, the
Received 2 November 2007
detailed pathogenic mechanisms and the way these mutations affect cell metabolism are
Revised version received
unclear. In this study, selected keratin mutations were analysed for their effect on the
24 February 2008
epidermal stress response. Expression profiles of two keratin-mutant cell lines from
Accepted 26 February 2008
epidermolysis bullosa simplex patients (one severe and one mild) were compared to a
Available online 8 March 2008
control keratinocyte line before and after challenge with hypo-osmotic shock, a common physiological stress that transiently distorts cell shape. Fewer changes in gene expression
Keywords:
were found in cells with the severely disruptive mutation (55 genes altered) than with the
Keratin
mild mutation (174 genes) or the wild type cells (261 genes) possibly due to stress response
Intermediate filaments
pre-activation in these cells. We identified 16 immediate-early genes contributing to a
Cytoskeleton
general cell response to hypo-osmotic shock, and 20 genes with an altered expression
Stress response
pattern in the mutant keratin lines only. A number of dual-specificity phosphatases (MKP-1,
ERK
MKP-2, MKP-3, MKP-5 and hVH3) are differentially regulated in these cells, and their
p38
downstream targets p-ERK and p-p38 are significantly up-regulated in the mutant keratin
MKP
lines. Our findings strengthen the case for the expression of mutant keratin proteins
EBS
inducing physiological stress, and this intrinsic stress may affect the cell responses to
Epithelium
secondary stresses in patients' skin.
Microarray
Introduction Intermediate filaments are a class of cytoskeleton filament proteins that are especially important for the structure and function of cells within tissues. This large family of 70 genes [1,2], encoding at least 73 proteins in humans, constitutes a
© 2008 Elsevier Inc. All rights reserved.
class of important biomarkers of differentiation, which are widely used in tumour diagnosis. In recent years these proteins have become associated with a growing number of human genetic disorders. Mutations are now linked to at least 70 clinically distinct pathological conditions [3,4]. Amongst the two keratin subclasses of intermediate filaments alone there
⁎ Corresponding author. National Institute of Chemistry, L-11 Room 133, Hajdrihova 19, SI-1000 Ljubljana, Slovenia. Fax: +386 1 4760 300. E-mail address: [email protected] (M. Liovic). 0014-4827/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2008.02.020
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
are over 680 reported mutations in 18 genes identified as causative in a range of human genetic diseases that affect the epidermis, its appendages and other epithelial tissues [4]. Intermediate filament disorders are phenotypically diverse, ranging from skin fragility (as in epidermolysis bullosa simplex (EBS), a skin blistering disorder caused by mutations in keratins K5 or K14) to premature ageing syndromes (such as Hutchinson-Gilford progeria, caused by mutations in lamin A/C). Many intermediate filament disorders are clearly associated with cell and tissue fragility, suggesting that a major function of intermediate filaments is to provide physical resilience to cells in tissues. The disease phenotype is the end point of a sequence of events, often requiring years, sometimes even decades to fully develop. A disruptive dominant mutation in a keratin gene will not only structurally impair the cytoskeleton, but it will also activate a cascade of biochemical processes that will result in the disease phenotype. Cells with mutant keratins have been shown to respond abnormally to a range of stresses [5–8]. The morphological effect of the mutations on intermediate filament structure varies: morphologically normal filaments can usually still be synthesized in the cell, but protein aggregates can also accumulate in the cytoplasm and these have been suggested to contribute to the severity of the phenotype [9–12]. Identification and understanding of the intracellular pathological process will be essential for developing better therapeutic approaches for this group of hereditary disorders. Amongst the intermediate filament disorders, EBS is still the best studied. The causative link between keratin mutations (in K5 or K14) and EBS has been known of the longest, and because of the superficial location of the affected tissue (epidermis), early lesions are accessible and visible in vivo and mutant cells can be isolated, cultured and manipulated experimentally. Keratinocyte cell lines derived from EBS patients are thus promising model systems in which to analyse EBS disease mechanisms [7]. In earlier studies we have used patient-derived cell lines to demonstrate abnormalities in stress responses of EBS mutant keratinocytes, such as their response to heat, osmotic and mechanical stress [5,7,8]. The emerging complexity of the disease process in EBS lends itself to analysis by gene expression profiling techniques. In the study presented here, we have used microarray analysis to compare gene expression in “mild” or “severe” EBSderived keratinocyte cell lines with that of keratinocytes with wild type keratins, and then to analyse the responses of these different cell lines to a common physiological stress, hypoosmotic stress. Based on time-course observations from our previous study [5], we assessed gene expression changes at 1 h after exposure to hypo-osmotic stress as being the time-point at which differences between effects of different mutations were most distinct. Results obtained from this study highlight significant differences in the way the cells handle the stress, in comparisons between “mild” EBS cells, “severe” EBS cells and control cells. Altered expression was observed in genes involved in regulating cell survival and in controlling the major MAPK pathways, specifically for certain dual-specificity phosphatases. Our results support the hypothesis put forward earlier [5,7] that the expression of mutant keratin per se induces a state of stress. The results also identify new potential targets
2067
for modifying the stress response in keratin-mutant cells in the quest for new therapeutic approaches to EBS. This study provides new insights into intermediate filament function in cells, which may have implications for the many other disease situations in which intermediate filament expression is modified, from tissue fragility disorders or premature ageing to tumour progression in cancer.
Results Expression profiling of the hypo-osmotic stress response To identify the transcriptome of the hypo-osmotic shock response in keratinocytes expressing mutant keratins, three immortalized keratinocyte cell lines were used: NEB-1 (a keratin-normal control line), KEB-4 (mild EBS phenotype, K14 V270M mutation) and KEB-7 (severe EBS phenotype, K14 R125P mutation) [7]. Cells were exposed to 150 mM urea for 5 min, which causes transient cell swelling, with severe but transient disruption of the cytoskeleton and activation of the JNK/SAPK pathway as previously described [5]. Cells were then allowed to recover for 1 h, and untreated and treated cells harvested. RNA was purified, labeled and hybridized to Affymetrix HU95A chips, and the data extracted from the differential hybridization were analysed. All experiments were performed in triplicate. To test the reproducibility of the microarray data, hierarchical clusters of gene expression were generated using GeneSpring™ software for both conditions, before and after hypo-osmotic shock (Fig. 1A). Highly similar patterns were obtained in all three experimental replicates, indicating good reproducibility of methodology. Clusters of genes that were differently regulated in the EBS lines vs control cells could be readily identified (see arrows in Fig. 1A), as could other clusters of genes showing altered expression unique to an individual cell line (see asterisks in Fig. 1A). Scatter correlation graphs were also generated for each triplicate experiment (data not shown), which further confirmed the hierarchical clustering data and reproducibility of the methods used. Median expression level values for the two states, before and after hypo-osmotic stress, were compared for each cell line individually. As previously reported for SAPK (stress activated protein kinase) activation [5], the degree of response to hypoosmotic stress was found to be correlated with the severity of the keratin mutation expressed, and thus with the severity of the EBS phenotype. By 1 h of recovery from hypo-osmotic stress, the wild type keratin cell line (NEB-1) showed significantly altered expression of 261 genes as compared to unstressed NEB1 cells; of these changes, 203 were down-regulated and 58 upregulated (Supplemental Table 5). The mild EBS mutant cell line KEB-4 showed fewer changes from its baseline, with altered expression in only 174 genes after comparison to before hypoosmotic shock, of which only 39 were down-regulated and 135 were up-regulated (Supplemental Table 6). The most severe EBS mutant (KEB-7 cells, with the severe K14 R125P mutation) showed the least response to osmotic stress, with significant alterations in only 55 genes, i.e. a third as many as the mild mutant and only a quarter of the number altered in the wild type cell line. As in the mild mutant, more genes were up-regulated (43 genes) than down-regulated (12 genes) (Supplemental
2068
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
Table 7). Comparison between these three cell lines is summarized in Figs. 1B and C. It was notable that while the control NEB-1 cell line responded to hypo-osmotic shock predominantly by down-regulation of differentially expressed genes (77% — only 23% were up-regulated), both the KEB-4 (mild) mutant cells and the KEB-7 (severe) mutant cells predominantly up-regulated genes (78% up, 22% down) (see Fig. 1B and Supplemental Tables 5, 6, 7). This suggests that the cell lines are initiating their response to the osmotic stress from a very different starting point.
Hypo-osmotic stress affects a shared group of genes independent of keratin mutation status The 16 genes whose expression was affected by hypoosmotic stress in all three cell lines are shown in Table 1. These genes, which respond in all three cell lines, are likely to have a core role in the response to hypo-osmotic stress in epithelial cells. In the two EBS lines, thirteen of these genes were clearly induced by hypo-osmotic stress, and three (a transcription factor, RQCD1 [13], and two integral
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
2069
Table 1 – The 16 genes responding to hypo-osmotic stress in all three cell lines
The results are shown fold changes compared unstressed cells. Green indicates reduced RNA levels while shows elevated signal The results are shown asas fold changes asas compared to to unstressed cells. Green indicates reduced RNA levels while redred shows elevated signal (over the + 2/− 2-fold threshold).
membrane proteins, AFURS1 (with a role in cation transport) [14] and B4GALT1 (an enzyme with a role also in cell– cell recognition and cell adhesion) [15]) were suppressed. Amongst the thirteen induced genes are the dual-specificity phosphatase hVH3 (also known as DUSP5) and several transcription factors that are important regulators of cell differentiation, such as ID2, FOS, JUN, SOX4, MAFF. In this broadly responsive set of genes, the dual-specificity phosphatases are particularly interesting because of their known role in MAPK inactivation, and therefore in stress response modulation. Inhibitor of DNA binding, ID2, belongs to the family of proteins that inhibit basic helix– loop–helix transcription factors in a dominant-negative way, and is known to play a role in promoting cell proliferation. The oncogenes FOS and JUN are components of the AP-1 transcription factor complex, which like MAFF, have a leucine zipper domain and function as regulators of cell death and proliferation. All of these genes function as promoters of cell survival and positive regulators for cell cycle progression.
Genes responding to hypo-osmotic stress in the keratinmutant lines The two K14-mutant cell lines had 20 genes in common that responded to hypo-osmotic stress but remained unchanged in the control cell line. These genes, which were all up-regulated in both K14 mutant cell lines, are listed in Table 2. This group is of interest as it includes several genes with a role in growth regulation and growth inhibition. Transforming growth factor β-inducible early growth response protein 1 (TIEG1) is a transcriptional repressor involved in the regulation of cell growth [16]. Cyclic AMP-dependent transcription factor 3 (ATF-3) may repress transcription by stabilizing the binding of inhibitory cofactors at the promoter, whereas GADD45 can recognize an altered chromatin state and modulate DNA accessibility to cellular proteins. Cyclin-dependent kinase inhibitor 1C (CDKN1C) is a negative regulator of cell proliferation [17] that may play a role in maintenance of the nonproliferative state throughout life. Kruppel-like factor 4 (KLF4) plays a role in the development of the epidermal permeability
Fig. 1 – Microarray analysis of one control and two EBS patient-derived cell lines. (A) Hierarchical clustering of the microarray data obtained from triplicate experiments of hypo-osmotic stress response in cell lines with wild type vs mutant keratins. The red color represents expression level above (2× or more) mean expression of a gene across all samples, the yellow color represents less than 2× deviation from the mean expression and the green color represents expression lower (2× or more) than the mean. Keratinocyte cell lines NEB-1 (wild type keratins), KEB-4 (carrying a mildly disruptive EBS K14 mutation, V270 M) and KEB-7 (with a clinically severe EBS K14 mutation, R125P), were used. Relative amounts of mRNA for each of over 12,000 genes were assayed before and 1 h after exposure to hypo-osmotic stress. Clustering of (more than 2-fold) up- and down-regulated genes in each of the triplicate assays for each cell line was analysed using GeneSpring™ software and shows highly similar patterns in all three experimental replicates, indicating good reproducibility. Specific clusters of regulated genes that are unique for each of the cell lines are easily identified (see asterisks), as well as those common for both mutant cell lines (arrows). (B) Different numbers of genes respond to hypo-osmotic stress in each cell line. Total mRNA expression profiles of the three cell lines were cross-compared before and after exposure to hypo-osmotic stress and the results are summarized here (for a full list see Supplemental Tables 5, 6 and 7). The biggest change is seen in the control cells and the least change in the cells expressing a clinically severe mutation, suggesting that the severe cells are most “pre-stressed” cells and require fewer changes in gene expression to convert to a full stress response. (C) Venn diagram summarizing the microarray data obtained on the investigated keratinocyte cell lines.
2070
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
Table 2 – Genes, which showed significantly altered transcription in response to hypo-osmotic stress in the two K14 mutants (KEB-4, KEB-7), but were not significantly altered in the wild type keratin (NEB-1) cells
The results are shown as fold changes compared to unstressed cells.
barrier and accelerates terminal differentiation in the epidermis [18]. Immediate early response 3, IER3, may be important in cell survival by regulating anti-apoptotic genes [19]. IL-6 mediates activation of STAT3 in squamous cell carcinoma of the head and neck, while JUNB is a member of AP1 family and takes part in cell cycle regulation. Thus the two K14 mutant cells appear to be predominantly responding to hypo-osmotic stress by inhibiting transcription and proliferation.
Dual-specificity phosphatases are differentially regulated in keratin-mutant cell lines In addition to hVH3/DUSP5, several other members of the dualspecificity phosphatase family (DUSP, or MKP (MAP kinase phosphatase)) were found to respond to hypo-osmotic stress (Tables 3 and 4). In the resting state (before stress), the expression levels of MKP genes in the two K14 mutant cell lines are constitutively lower than in the control cell line (NEB-1). The microarray results (MA, Table 3) were confirmed by real-time PCR (RT-PCR) as summarized in Table 3. MKP-1, MKP-3 and hVH3 are significantly down-regulated in both mutants when
Table 3 – Dual-specificity phosphatases show lower mRNA levels during resting state in the K14 mutant lines (KEB-4, KEB-7) than in the wild type cells (NEB-1)
MA = microarray data; RT-PCR = real-time PCR data; – = expression MA = microarray data; RT-PCR = real-time PCR data; – =(outside expression levels not significantly different from the wild type the + 2/ −2-fold threshold) indifferent this cell from line. the wild type (outside the +2/−2levels not significantly
compared to the wild type. MKP-5 was also found downregulated in KEB-7 cells (severe mutant), but was similar to the wild type control level in KEB-4 cells (mild mutant) (Table 3). When we compared the transcription of MKP genes in the unstressed state (before hypo-osmotic stress) with that at 1 h of recovery for each cell line, we found that all these phosphatases were differentially regulated in response to stress (Table 4). In addition, dual-specificity phosphatase expression in the control NEB-1 cells was down-regulated from the starting point, whereas in the two K14 mutant lines (KEB-4 and KEB-7) these genes were up-regulated (Table 4). Changes in these phosphatases were also evaluated by RT-PCR (Table 4). While most of the data was confirmed (apart MKP-2 down-regulation in the control, NEB-1 cells), RT-PCR revealed an additional up-regulation of MKP-1 and MKP-3 in the severe mutant (KEB-7 cells), which was not detected by microarray analysis (less than 2-fold change). In view of their role in attenuating MAP kinase activation and the observed down-regulation of MKPs in both mutants during resting state, we would expect this to affect the regulation of the stress response and MAPK activation. We therefore investigated the time-course of the stress response to hypo-osmotic shock in these cell lines [5], and analysed protein levels of total and phosphorylated ERK and p38, and phosphatases MKP-1 and hVH3 in response to stress (Fig. 2A). Total p38 and ERK levels were unchanged in response to stress in all three cell lines. However, activation of p38 and ERK (as deduced from their phosphorylation) was stronger and persisted for longer in the mutant (KEB-7 and KEB-4) than in the wild type cells (NEB-1). In the latter, the stress response was already shutting down after the 1 h time-point whereas it was still strong and persisting in the two mutants (Fig. 2A). Both hVH3 and MKP-1 proteins showed an obvious induction in response to stress in all three cell lines (Fig. 2A). In addition, hVH3 was not detected in any of the cell lines until the 30 min recovery time-point, indicating that in keratinocytes hVH3 is clearly a hypo-osmotic stress
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
2071
Table 4 – Dual-specificity phosphatases showing significantly altered transcription at 1 h of recovery after hypo-osmotic stress
MA = microarray data; RT-PCR = real-time PCR data; – = expression levels not significantly different from the wild type (outside the + 2/ MA = microarray data;inRT-PCR = real-time PCR data; – = expression levels not significantly different from the wild type (outside the + 2/−2-fold − 2-fold threshold) this cell line.
inducible phosphatase. Although all cell lines responded to hypo-osmotic stress with hVH3 synthesis and increased phosphorylation of both ERK and p38, this was more prominent in the two K14 mutants (KEB-4 and KEB-7 cells). We also examined the protein levels of total and phosphorylated keratin in the samples (Fig. 2B), which due to the method of preparation (see Materials and methods) would be depleted in filamentous (pelleted) keratin. Like other
intermediate filaments, phosphorylation of keratins is associated with proteins in a non-filamentous state. Consistent with the presence of unpolymerized (misfolded) keratin protein, in the mild (KEB-4) and severe K14 mutant (KEB-7), phosphorylated keratin (p-K5 and possibly p-K6, due to known cross-reactivity of the antibody used (LJ4) and the type of cell (keratinocytes in culture)) was clearly detectable in the absence of hypo-osmotic stress. In response to stress, the
Fig. 2 – Time-course experiment of the response of wild type and K14 mutant keratinocytes to hypo-osmotic stress. Protein levels of target proteins were tested in wild type (NEB-1) and two K14 mutant cells, KEB-4 (mild mutant) and KEB-7 (severe mutant), before (“0”) and after hypo-osmotic stress at 5, 10, 30 min and 1 h and 2 h of recovery time-points. (A) Total ERK and p38 protein levels are unchanged in response to stress in all cell lines. Hypo-osmotic stress activates ERK and p38, with phosphorylated ERK (p-ERK) and phosphorylated p38 (p-p38) levels higher and persisting for longer in the two K14 mutants, in particular in the severe K14 mutant (KEB-7). hVH3 (DUSP5) phosphatase is a hypo-osmotic shock inducible protein, as it is absent before stress and not detectable until 30 min of recovery. Higher hVH3 levels are detected in the K14 mutants than in the wild type. MKP-1 (DUSP1) is also seen as induced in response to hypo-osmotic stress. (B) Phosphorylated keratin (p-K5) is higher in the K14 mutants than in the wild type cells prior to stress. Hypo-osmotic stress induces keratin phosphorylation in the wild type cells (NEB-1) up to 1 h of recovery, but results in significant reduction of keratin phosphorylation in the two K14 mutants (KEB-4, KEB-7) after 10–30 min. Total K5 is reduced at late time-points (“1 h” and “2 h”) in the wild type, but continues to increase in the severe mutant (KEB-7). GAPDH was used as the loading control; 10 μg of protein from total cell lysate samples was loaded in each track on the gel.
2072
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
levels of phosphorylated keratin increased in the wild type cells (NEB-1), but decreased at later time-points of recovery in the two mutants. Total K5 also decreased at later time-points in the wild type and the mild mutant (KEB-4), while it increased in the severe mutant (KEB-7). This suggests that (hypoosmotic) stress may increase the level of unpolymerized keratin protein in the cytoplasm.
Discussion Tissue injury and stress play important roles in the phenotype of all diseases, but the actual molecular events leading to such changes are not well documented. Intermediate filaments of the cytoskeleton are increasingly becoming recognized as contributing to the defence against stress of cells in tissues, and experiments on chemical stress, heat stress, osmotic stress and mechanical stress in cells with mutant intermediate filament proteins continue to accumulate data that support this view [20]. In this study we have used hypo-osmotic stress to examine the effect of mutations in keratin intermediate filament proteins on a stress response. Osmotic stress, especially hypo-osmotic stress, is one of the most common types of stress affecting cells in tissues. In the body it results rapidly from loss of energy supply to tissue cells, as caused by anything from indirect ischemia to direct wounding: the energy-dependent activity of membrane ion pumps cannot be maintained and osmotic balance is lost, leading to influx of water and swelling of cells. Experimentally induced hypo-osmotic shock causes transient cell swelling and extensive disruption of all three cytoskeleton filament systems (actin, tubulin and intermediate filaments [5]), which cells counteract by raising the activity of membrane pumps. We previously showed that the persistence, intensity, and the timing of induction of the JNK hypo-osmotic stress response varied between keratinocytes expressing different keratin mutations [5]. Here we present further analysis of the hypo-osmotic stress response, using microarray expression profiling of three keratinocyte cell lines expressing normal keratins or keratin mutations of differing clinical severity. The cell lines selected are morphologically similar, immortalized in the same way and showing similar differentiation and growth properties under standard culture conditions, but they show a quantitatively different response to stress that correlates with the clinical severity of the keratin mutations they express [7]. Here we demonstrate that already within the first hour after hypoosmotic stress, the K14 defective cell lines had altered the expression of many genes. At least 16 genes appear to be related to the hypo-osmotic stress itself, rather than to the stress of expressing and handling mutant keratin proteins, as they are altered in both mutant and wild type cells, albeit sometimes in different ways. In the present experiments the number of genes seen to “respond” to hypo-osmotic stress is inversely correlated with clinical severity associated with the keratin mutation in the cell — more severe disease phenotype, fewer genes altered. This is consistent with earlier observations [5,7] that suggested that due to the presence of mutant keratins, the EBSderived cells show an active stress response even before osmotic stress is applied. Both K14 mutant lines respond to
hypo-osmotic stress in a similar way, which is strikingly different from the wild type. In the control NEB-1 cells, 78% of the altered genes were down-regulated, while in the mutant cells 77% of genes are up-regulated. If the presence of mutant keratin protein induces an intrinsic physiological stress by itself, then these cells may have fewer additional adjustments to make when challenged by an additional extrinsic stress. The gene expression profile also suggests that upon hypoosmotic stress, these keratin-mutant cells are being pushed to another level of stress response. 20 of the genes are only altered (up-regulated) in the mutant lines (Table 2), and these genes include several transcriptional factors with a role in growth arrest. Of this group, ATF3, TIEG, JunB, IER3, GADD45A and CDKN1C have been recently found strongly up-regulated upon treatment with oncostatin M in several breast tumour cell lines (T47D, MCF-7, MDA-MB-231) [21]. Oncostatin M is a potent cytostatic cytokine belonging to the IL-6 family that causes cell cycle arrest in late G1 phase. Similar to these breast tumour cell lines, the K14 mutant keratinocytes respond to hypo-osmotic stress by up-regulating the transcription of IL-6 and also the same group of genes: ATF3, TIEG, JunB, IER3, GADD45A and CDKN1C. Transcriptional profiles of these genes remained unchanged in the control keratinocyte cell line at equivalent time-points after stress. This suggests a significant parallel between the sets of genes activated by hypo-osmotic stress in K14-mutant keratinocytes and the ones activated by a cytostatic cytokine in several breast cancer cell lines. In both situations the increase in transcription of an IL-6 family cytokine results in the increase of transcription of specific downstream targets involved in regulating growth arrest and apoptosis [22,23]. It would appear that the EBS keratinocytes, with their constitutive stress already elevated by the presence of mutant keratin protein, are propelled into growth arrest by the additional stress of hypo-osmotic shock. This may be a protective mechanism to permit damage assessment and prevent proliferation of fatally damaged cells. We have previously shown (see “metabolic recovery after hypo-osmotic stress” assay in [5]) that growth arrest may occur in keratin-mutant cells after hypo-osmotic shock. Here we have now identified a number of genes that may be involved in this process. Amongst the differentially expressed genes, the dualspecificity phosphatases (MKPs or DUSPs) were considered especially significant, as these are important regulatory proteins for phosphorylation-driven MAP kinase pathways [24]. MKPs are involved in the regulation of the cell stress response by establishing a negative regulatory feedback loop that targets and dephosphorylates activated MAP kinases, the ERK, JNK and p38 pathways [24–26]. The importance of MKPs has been further emphasized by recent findings that implicate them in the development of some forms of cancer [27–34]. Our microarray data has shown that hVH3, an ERK MAP kinase specific MKP [35], is part of the set of genes that may contribute to the generic way epithelial cells respond to hypo-osmotic shock or even other stresses. Other MKPs were also found differentially regulated in response to hypo-osmotic shock, but not in all of cell lines investigated (Table 4). Furthermore several MKPs were already significantly down-regulated in the two K14 mutant lines before exposure to stress (Table 3, compared to the control cells). This is mirrored by the increased phosphorylation of their downstream targets, p38 and ERK (Fig. 2A).
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
There is substantial evidence in the literature that MKPs have important roles in regulating cell metabolism and p38, JNK and ERK-mediated responses in a variety of cells and tissues [24– 26]. The data presented here indicates that MKPs may also be important in a different pathological situation, i.e. in epidermal cell lines expressing mutant K14 associated with a tissue fragility disorder (EBS). At present, it is not clear which specific mechanism is involved in triggering the observed increased phosphorylated MAPK levels prior to stress, or the differential regulation of MKPs in keratin-mutant cells. Nevertheless, phosphorylated p38 has been recently found to co-localize with keratin granules in mutant keratin expressing cells [36]. Fig. 2 shows that phosphorylated keratin is detectable in the Triton-soluble phase of total cell extracts from the two K14 mutants prior to stress, which would be expected in the presence of unpolymerised keratin, although this appears to decrease after stress. Hypo-osmotic stress induces cytoskeleton remodelling [5], and this is marked by an increase in phosphorylated soluble keratin in the wild type cells (Fig. 2B). Total K5 decreased at later time-points of recovery in the NEB-1 and KEB-4 cells, while it increased in the KEB-7 cells (severe mutant). Seen as a whole, hypo-osmotic stress may augment not only the levels of unpolymerized (soluble) keratin protein in the cytoplasm (part of the general keratinocyte response), but when combined with a severe keratin mutation this may also increase the amount of misfolded protein, interfering with all aspects of keratin dynamics and regulation. Intermediate filament protein aggregates are associated with many pathologies and could arise, or be exacerbated, by many different mechanisms, from altered chaperone behaviour to defective protein degradation [37,38]. It is also known that the phosphorylation state of intermediate filament proteins drives their assembly and disassembly during mitosis or exposure to stress [39,40]. It has been suggested that keratins may act as a “phosphate sponge” for activated MAPKs, whereby depending on the protein variant (wild type or mutant) they may either have a protective role (by buffering MAPK activity) or predispose to disease development (if misfolded mutant keratin restricts phospho-site availability), by deflecting and generating a more intense MAPK signal transduction into the nucleus instead [41]. Our data may support this hypothesis since in response to stress there is a decrease in phosphorylated keratin in the two K14 mutants, while it increases in the wild type cells (Fig. 2B). In response to stress, the levels of p-p38 and p-ERK were also higher and persisted for longer in the two K14 mutants. Thus depending on the overall detrimental effect that a mutation has on the structure and function of keratin intermediate filament proteins, several signalling pathways may be triggered. Therefore, a better understanding of the way in which epithelial cells respond to stress is needed, as this will undoubtedly have a bearing on the development of new therapeutic strategies.
Materials and methods Cell lines, culture conditions and osmotic shock Immortalized keratinocyte cell lines NEB-1 (control keratinocyte cell line), KEB-4 (mild, Weber-Cockayne EBS phenotype caused by the K14 V270M mutation) and KEB-7 cells (severe,
2073
Dowling-Meara EBS phenotype caused by the K14 R125P mutation) [7] were cultured under standard conditions (RM medium, 37 °C, 5% CO2, without feeder cells) on 10 cm Petri dishes to 80% confluence. Each cell line was grown in triplicate (3 plates). For the two experimental situations, i.e. before and 1 h after osmotic shock, we grew 3 × 2 plates per cell line. At 80% confluence, one triplicate set was hypo-osmotically challenged by adding 150 mM urea to the culture medium for 5 min [5]; cells were then incubated in fresh tissue culture medium for a further 1 h at 37 °C to allow the response to develop in a recovery state. At this point all cells were taken out of the incubator, put on ice, washed with two changes of phosphatebuffered saline (PBS) and processed as below.
Total RNA isolation and labeling Total cellular RNA was isolated using RNeasy Mini Kit (Qiagen Ltd, UK) according to the manufacturer's instructions [42]. The optical density was measured and 8 μg of total RNA extract was used per reaction for double-stranded cDNA synthesis, using SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen Ltd., UK). Products were then precipitated and resuspended in RNase-free water and the entire sample was used for a subsequent in vitro transcription and cRNA labeling reaction (BioArray HighYield RNA Transcript Labeling kit, Enzo Life Sciences Inc., New York). Labeled samples were fragmented and hybridized to a test chip to check the efficiency and quality of samples.
Transcriptional profiling and microarray data analysis The Affymetrix Human Genome U95A GeneChip with 12627 probes for the human genome was used (Affymetrix Inc., California). Microarray Suite 5.0 (Affymetrix) was used for data extraction and for further analysis, data mining tool 3.0 (Affymetrix) and GeneSpring™ software 5.1 (Agilent Technologies, California) were used for normalization, extent of change calculations and clustering [42]. Differential expression levels of transcripts were determined as the increase or decrease in expression in relation to the paired sample (as 2-fold, 3-fold change etc.). To compare data from multiple arrays, the signal of each probe array was scaled to the same target intensity value. Genes were considered to show altered expression if the expression levels differed more than 2-fold relative to untreated control at any time-point. Using GeneSpring™, clustering was performed based on experiments or individual genes expression profiles. To aid analysis we assembled an extensive gene annotation table listing the molecular function and biological category of the genes represented on the chip, based primarily on data from J.M. Ruillard and [43].
Real-time PCR Extracted total RNA was used as template for cDNA synthesis using SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen Ltd, UK) following the manufacturer's instructions. Subsequently, 1 μl of cDNA product was used per each real-time PCR reaction using Platinum Taq DNA polymerase (Invitrogen Ltd, UK) and SYBR Green I nucleic acid stain (Molecular Probes, USA). The MKP gene specific primers were obtained from Qiagen (QuantiTect Primer Assays), and used along Qiagen's
2074
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
standard protocol on an iQ5 Biorad real-time PCR detection system. GAPDH and β-actin were used as internal controls.
[2]
Total protein extraction [3]
Cells were cultured as described above. After hypo-osmotic shock cells were washed with PBS and total protein was extracted using an extraction buffer containing the following: 50 mM Tris/HCl pH 7.5, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 5 mM sodium pyrophosphate, 10 mM β-phosphoglycerate and 1% Triton X-100, plus the following which were added fresh, immediately prior to use: 1 mM Na-orthovanadate, 0.1% (β-mercaptoethanol, 1 mM PMSF, 1 mM benzamidin, 10 μg/ml leupeptin and 10 μg/ml pepstatin A). Protein concentration was determined by Bradford assay (BioRad protein assay, Bio-Rad Gmbh., Germany) and 10 μg of total protein per sample was loaded on denaturing protein gels.
Antibodies To detect MAP kinases, rabbit anti-phosphorylated-ERK, rabbit anti-phosphorylated-p38, rabbit anti-ERK and rabbit anti-p38 (Cell Signaling Technology Inc., Massachusetts) antisera were used. Mouse monoclonal LJ4 was used to detect phosphorylated keratin [39], and rabbit polyclonal antibody BL-18 was used to detect K5 [44]. Rabbit anti-MKP-1 (Santa Cruz Biotechnology, USA) and rabbit anti-hVH3 (a gift from J.E. Dixon, University of San Diego) antibodies were used to detect DUSP1 and DUSP5 respectively. To detect GAPDH, a mouse monoclonal antibody was used (Abcam plc., UK). As secondary antibodies, anti-rabbit immunoglobulin (Cell Signaling) and antimouse immunoglobulin (DakoCytomation, Denmark) antisera conjugated to HRP were used.
[4]
[5]
[6]
[7]
[8]
[9]
[10]
Acknowledgments [11]
We would like to thank V. Makrantoni, D. Martin, S. Keyse and D. Russell for their useful discussions, and J.K. Heath and N. Underhill-Day for their helpful comments and access to unpublished information. We would also like to thank J.E. Dixon for supplying the hVH3 antibody. This work was funded by Cancer Research UK (grant C26/A1461 to E.B. Lane), the Wellcome Trust (grant 5090/A/98/Z to E.B. Lane), the Slovenian Research Agency (grant J3-6132-0381-04 to M. Liovic), and the National Institutes of Health (grants AR45974 and NR08029 to M. Tomic-Canic).
[12]
[13]
[14]
Appendix A. Supplementary data [15]
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.yexcr.2008.02.020. [16] REFERENCES [17] [1] M. Hesse, T.M. Magin, K. Weber, 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 (2001) 2569–2575. M. Hesse, A. Zimek, K. Weber, T.M. Magin, Comprehensive analysis of keratin gene clusters in humans and rodents, Eur. J. Cell Biol. 83 (2004) 19–26. M.B. Omary, P.A. Coulombe, W.H. McLean, Intermediate filament proteins and their associated diseases, N. Engl. J. Med. 351 (2004) 2087–2100. I. Szeverenyi, A.J. Cassidy, C.W. Chung, B.T.K. Lee, J.E. Common, S.C. Ogg, H. Chen, S.Y. Sim, W.L.P. Goh, K.W. Ng, J.A. Simpson, L.L. Chee, G.H. Eng, L. Bin, D. Chuon, A. Venkatesh, K.H. Khoo, W.H.I. McLean, Y.P. Lim, E.B. Lane, The human intermediate filament database: comprehensive information on a gene family involved in many human diseases, Human Mutat. 29 (2008) 351–360 Electronic publication ahead of print, Nov 2007. M. D'Alessandro, D. Russell, S.M. Morley, A.M. Davies, E.B. Lane, Keratin mutations of epidermolysis bullosa simplex alter the kinetics of stress response to osmotic shock, J. Cell Sci. 115 (2002) 4341–4351. M. Liovic, M.M. Mogensen, A.R. Prescott, E.B. Lane, Observation of keratin particles showing fast bidirectional movement colocalized with microtubules, J. Cell Sci. 116 (2003) 1417–1427. S.M. Morley, M. D'Alessandro, C. Sexton, E.L. Rugg, H. Navsaria, C.S. Shemanko, M. Huber, D. Hohl, A.I. Heagerty, I.M. Leigh, E.B. Lane, Generation and characterization of epidermolysis bullosa simplex cell lines: scratch assays show faster migration with disruptive keratin mutations, Br. J. Dermatol. 149 (2003) 46–58. D. Russell, P.D. Andrews, J. James, E.B. Lane, Mechanical stress induces profound remodelling of keratin filaments and cell junctions in epidermolysis bullosa simplex keratinocytes, J. Cell Sci. 117 (2004) 5233–5243. H. Bar, N. Mucke, A. Kostareva, G. Sjoberg, U. Aebi, H. Herrmann, Severe muscle disease-causing desmin mutations interfere with in vitro filament assembly at distinct stages, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 15099–15104. A. Letai, P.A. Coulombe, E. Fuchs, Do the ends justify the mean? Proline mutations at the ends of the keratin coiled-coil rod segment are more disruptive than internal mutations, J. Cell Biol. 116 (1992) 1181–1195. J. Liu, M. Tang, R. Mestril, X. Wang, Aberrant protein aggregation is essential for a mutant desmin to impair the proteolytic function of the ubiquitin-proteasome system in cardiomyocytes, J. Mol. Cell Cardiol. 40 (2006) 451–454. T. Sasaki, T. Gotow, M. Shiozaki, F. Sakaue, T. Saito, J.P. Julien, Y. Uchiyama, S. Hisanaga, Aggregate formation and phosphorylation of neurofilament-L Pro22 Charcot-Marie-Tooth disease mutants, Hum. Mol. Genet. 15 (2006) 943–952. N. Hiroi, T. Ito, H. Yamamoto, T. Ochiya, S. Jinno, H. Okayama, Mammalian Rcd1 is a novel transcriptional cofactor that mediates retinoic acid-induced cell differentiation, EMBO J. 21 (2002) 5235–5244. N. Habtemichael, G. Kovacs, Cloning the AFURS1 gene which is up-regulated in senescent human parenchymal kidney cells, Gene 283 (2002) 271–275. K.A. Masri, H.E. Appert, M.N. Fukuda, Identification of the full-length coding sequence for human galactosyltransferase (beta-N-acetylglucosaminide: beta-1,4-galactosyltransferase), Biochem. Biophys. Res. Commun. 157 (1988) 657–663. M. Subramaniam, J.R. Hawse, S.A. Johnsen, T.C. Spelsberg, Role of TIEG1 in biological processes and disease states, J. Cell Biochem. 102 (2007) 539–548. J.M. Scandura, P. Boccuni, J. Massague, S.D. Nimer, Transforming growth factor beta-induced cell cycle arrest of human hematopoietic cells requires p57KIP2 upregulation, Proc. Nat. Acad. Sci. 101 (2004) 15231–15236.
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
[18] J. Jaubert, J. Cheng, J.A. Segre, Ectopic expression of kruppel like factor 4 (Klf4) accelerates formation of the epidermal permeability barrier, Development 130 (2003) 2767–2777. [19] L. Shen, J. Guo, C. Santos-Berrios, M.X. Wu, Distinct domains for anti- and pro-apoptotic activities of IEX-1, J. Biol. Chem. 281 (2006) 15304–15311. [20] E.B. Lane, M. Pekny, Stress models for the study of intermediate filament function, Methods Cell Biol. 78 (2004) 229–264. [21] N. Underhill-Day, J.K. Heath, Oncostatin M (OSM) cytostasis of breast tumor cells: characterization of an OSM receptor beta-specific kernel, Cancer Res. 66 (2006) 10891–10901. [22] S. Bosze, F. Hudecz, P. Igaz, Z. Ortutay, G. Csik, A. Falus, S. Toth, Interleukin-6 N-terminal peptides modulate the expression of junB protooncogene and the production of fibrinogen in HepG2 cells, Biol. Chem. 384 (2003) 409–421. [23] T. Hai, C.D. Wolfgang, D.K. Marsee, A.E. Allen, U. Sivaprasad, ATF3 and stress responses, Gene Expr. 7 (1999) 321–335. [24] D.M. Owens, S.M. Keyse, Differential regulation of MAP kinase signalling by dual-specificity protein phosphatases, Oncogene 26 (2007) 3203–3213. [25] A. Theodosiou, A. Ashworth, MAP kinase phosphatases, Genome Biol. 3 (2002) REVIEWS3009. [26] A. Farooq, M.M. Zhou, Structure and regulation of MAPK phosphatases, Cell Signal. 16 (2004) 769–779. [27] H.Y. Wang, Z. Cheng, C.C. Malbon, Overexpression of mitogen-activated protein kinase phosphatases MKP1, MKP2 in human breast cancer, Cancer Lett. 191 (2003) 229–237. [28] C.D. Swartz, C.A. Afshari, L. Yu, K.E. Hall, D. Dixon, Estrogen-induced changes in IGF-I, Myb family and MAP kinase pathway genes in human uterine leiomyoma and normal uterine smooth muscle cell lines, Mol. Hum. Reprod. 11 (2005) 441–450. [29] R. Steinmetz, H.A. Wagoner, P. Zeng, J.R. Hammond, T.S. Hannon, J.L. Meyers, O.H. Pescowitz, Mechanisms regulating the constitutive activation of the extracellular signal-regulated kinase (ERK) signaling pathway in ovarian cancer and the effect of ribonucleic acid interference for ERK1/2 on cancer cell proliferation, Mol. Endocrinol. 18 (2004) 2570–2582. [30] S. Vicent, M. Garayoa, J.M. Lopez-Picazo, M.D. Lozano, G. Toledo, F.B. Thunnissen, R.G. Manzano, L.M. Montuenga, Mitogen-activated protein kinase phosphatase-1 is overexpressed in non-small cell lung cancer and is an independent predictor of outcome in patients, Clin. Cancer Res. 10 (2004) 3639–3649. [31] J.K. Warmka, L.J. Mauro, E.V. Wattenberg, Mitogen-activated protein kinase phosphatase-3 is a tumor promoter target in initiated cells that express oncogenic Ras, J. Biol. Chem. 279 (2004) 33085–33092. [32] T. Furukawa, R. Fujisaki, Y. Yoshida, N. Kanai, M. Sunamura, T. Abe, K. Takeda, S. Matsuno, A. Horii, Distinct progression
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
2075
pathways involving the dysfunction of DUSP6/MKP-3 in pancreatic intraepithelial neoplasia and intraductal papillary-mucinous neoplasms of the pancreas, Mod. Pathol. 18 (2005) 1034–1042. S. Xu, T. Furukawa, N. Kanai, M. Sunamura, A. Horii, Abrogation of DUSP6 by hypermethylation in human pancreatic cancer, J. Hum. Genet. 50 (2005) 159–167. L. Nonn, L. Peng, D. Feldman, D.M. Peehl, Inhibition of p38 by vitamin D reduces interleukin-6 production in normal prostate cells via mitogen-activated protein kinase phosphatase 5: implications for prostate cancer prevention by vitamin D, Cancer Res. 66 (2006) 4516–4524. M. Mandl, D.N. Slack, S.M. Keyse, Specific inactivation and nuclear anchoring of extracellular signal-regulated kinase 2 by the inducible dual-specificity protein phosphatase DUSP5, Mol. Cell Biol. 25 (2005) 1830–1845. S. Wöll, R. Windoffer, R.E. Leube, p38 MAPK-dependent shaping of the keratin cytoskeleton in cultured cells, J. Cell Biol. 177 (2007) 795–807. E. Janig, C. Stumptner, A. Fuchsbichler, H. Denk, K. Zatloukal, Interaction of stress proteins with misfolded keratins, Eur. J. Cell Biol. 84 (2005) 329–339. M. Der Perng, M. Su, S.F. Wen, R. Li, T. Gibbon, A.R. Prescott, M. Brenner, R.A. Quinlan, The Alexander disease-causing glial fibrillary acidic protein mutant, R416W, accumulates into Rosenthal fibers by a pathway that involves filament aggregation and the association of alpha B-crystallin and HSP27, Am. J. Hum. Genet. 79 (2006) 197–213. D.M. Toivola, Q. Zhou, L.S. English, M.B. Omary, Type II keratins are phosphorylated on a unique motif during stress and mitosis in tissues ad cultured cells, Mol. Biol. Cell. 13 (2002) 1857–1870. T. Yamaguchi, H. Goto, T. Yokoyama, H. Silljé, A. Hanisch, A. Uldschmid, Y. Takai, T. Oguri, E.A. Nigg, M. Inagaki, Phosphorylation by Cdk1 induces Plk1-mediated vimentin phosphorylation during mitosis, J. Cell Biol. 171 (2005) 431–436. N. Ku, M.B. Omary, A disease and phosphorylation-related nonmechanical function for keratin 8, J. Cell Biol. 174 (2006) 115–125. O. Stojadinovic, B. Lee, C. Vouthounis, S. Vukelic, I. Pastar, M. Blumenberg, H. Brem, M. Tomic-Canic, Novel genomic effects of glucocorticoids in epidermal keratinocytes: inhibition of apoptosis, IFN gamma pathway and wound healing along with promotion of terminal differentiation, J. Biol. Chem. 282 (2007) 4021–4034. Gene Ontology Consortium Data [http://cgap.nci.nih.gov/ Genes/GOBrowser] [dot.ped.med.umich.edu:2000/ourimage/ pub/shared/JMR_pub_affyannot.html]. Purkis, J.B. Steel, I.C. Mackenzie, W.B. Nathrath, I.M. Leigh, E.B. Lane, Antibody markers of basal cells in complex epithelia, J. Cell Sci. 97 (1990) 39–50.
a v a i l a b l e a t w w w. s c i e n c e d i r e c t . c o m
w w w. e l s e v i e r. c o m / l o c a t e / y e x c r
Research Article
Dual-specificity phosphatases in the hypo-osmotic stress response of keratin-defective epithelial cell lines Mirjana Liovica,b,⁎, Brian Leec , Marjana Tomic-Canicd , Mariella D'Alessandrob , Viacheslav N. Bolshakove , E. Birgitte Laneb,f a
National Institute of Chemistry, Ljubljana, Slovenia CRUK Cell Structure Research Group, University of Dundee College of Life Sciences, MSI/WTB Complex, Dundee, UK c Genentech, San Francisco, USA d Hospital for Special Surgery at Weill Medical College of the Cornell University, New York, USA e University of Reading BioCentre, Reading, UK f A⁎STAR Institute of Medical Biology, Immunos, 8A Biomedical Grove, Singapore b
ARTICLE INFORMATION
ABS T R AC T
Article Chronology:
Although mutations in intermediate filament proteins cause many human disorders, the
Received 2 November 2007
detailed pathogenic mechanisms and the way these mutations affect cell metabolism are
Revised version received
unclear. In this study, selected keratin mutations were analysed for their effect on the
24 February 2008
epidermal stress response. Expression profiles of two keratin-mutant cell lines from
Accepted 26 February 2008
epidermolysis bullosa simplex patients (one severe and one mild) were compared to a
Available online 8 March 2008
control keratinocyte line before and after challenge with hypo-osmotic shock, a common physiological stress that transiently distorts cell shape. Fewer changes in gene expression
Keywords:
were found in cells with the severely disruptive mutation (55 genes altered) than with the
Keratin
mild mutation (174 genes) or the wild type cells (261 genes) possibly due to stress response
Intermediate filaments
pre-activation in these cells. We identified 16 immediate-early genes contributing to a
Cytoskeleton
general cell response to hypo-osmotic shock, and 20 genes with an altered expression
Stress response
pattern in the mutant keratin lines only. A number of dual-specificity phosphatases (MKP-1,
ERK
MKP-2, MKP-3, MKP-5 and hVH3) are differentially regulated in these cells, and their
p38
downstream targets p-ERK and p-p38 are significantly up-regulated in the mutant keratin
MKP
lines. Our findings strengthen the case for the expression of mutant keratin proteins
EBS
inducing physiological stress, and this intrinsic stress may affect the cell responses to
Epithelium
secondary stresses in patients' skin.
Microarray
Introduction Intermediate filaments are a class of cytoskeleton filament proteins that are especially important for the structure and function of cells within tissues. This large family of 70 genes [1,2], encoding at least 73 proteins in humans, constitutes a
© 2008 Elsevier Inc. All rights reserved.
class of important biomarkers of differentiation, which are widely used in tumour diagnosis. In recent years these proteins have become associated with a growing number of human genetic disorders. Mutations are now linked to at least 70 clinically distinct pathological conditions [3,4]. Amongst the two keratin subclasses of intermediate filaments alone there
⁎ Corresponding author. National Institute of Chemistry, L-11 Room 133, Hajdrihova 19, SI-1000 Ljubljana, Slovenia. Fax: +386 1 4760 300. E-mail address: [email protected] (M. Liovic). 0014-4827/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.yexcr.2008.02.020
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
are over 680 reported mutations in 18 genes identified as causative in a range of human genetic diseases that affect the epidermis, its appendages and other epithelial tissues [4]. Intermediate filament disorders are phenotypically diverse, ranging from skin fragility (as in epidermolysis bullosa simplex (EBS), a skin blistering disorder caused by mutations in keratins K5 or K14) to premature ageing syndromes (such as Hutchinson-Gilford progeria, caused by mutations in lamin A/C). Many intermediate filament disorders are clearly associated with cell and tissue fragility, suggesting that a major function of intermediate filaments is to provide physical resilience to cells in tissues. The disease phenotype is the end point of a sequence of events, often requiring years, sometimes even decades to fully develop. A disruptive dominant mutation in a keratin gene will not only structurally impair the cytoskeleton, but it will also activate a cascade of biochemical processes that will result in the disease phenotype. Cells with mutant keratins have been shown to respond abnormally to a range of stresses [5–8]. The morphological effect of the mutations on intermediate filament structure varies: morphologically normal filaments can usually still be synthesized in the cell, but protein aggregates can also accumulate in the cytoplasm and these have been suggested to contribute to the severity of the phenotype [9–12]. Identification and understanding of the intracellular pathological process will be essential for developing better therapeutic approaches for this group of hereditary disorders. Amongst the intermediate filament disorders, EBS is still the best studied. The causative link between keratin mutations (in K5 or K14) and EBS has been known of the longest, and because of the superficial location of the affected tissue (epidermis), early lesions are accessible and visible in vivo and mutant cells can be isolated, cultured and manipulated experimentally. Keratinocyte cell lines derived from EBS patients are thus promising model systems in which to analyse EBS disease mechanisms [7]. In earlier studies we have used patient-derived cell lines to demonstrate abnormalities in stress responses of EBS mutant keratinocytes, such as their response to heat, osmotic and mechanical stress [5,7,8]. The emerging complexity of the disease process in EBS lends itself to analysis by gene expression profiling techniques. In the study presented here, we have used microarray analysis to compare gene expression in “mild” or “severe” EBSderived keratinocyte cell lines with that of keratinocytes with wild type keratins, and then to analyse the responses of these different cell lines to a common physiological stress, hypoosmotic stress. Based on time-course observations from our previous study [5], we assessed gene expression changes at 1 h after exposure to hypo-osmotic stress as being the time-point at which differences between effects of different mutations were most distinct. Results obtained from this study highlight significant differences in the way the cells handle the stress, in comparisons between “mild” EBS cells, “severe” EBS cells and control cells. Altered expression was observed in genes involved in regulating cell survival and in controlling the major MAPK pathways, specifically for certain dual-specificity phosphatases. Our results support the hypothesis put forward earlier [5,7] that the expression of mutant keratin per se induces a state of stress. The results also identify new potential targets
2067
for modifying the stress response in keratin-mutant cells in the quest for new therapeutic approaches to EBS. This study provides new insights into intermediate filament function in cells, which may have implications for the many other disease situations in which intermediate filament expression is modified, from tissue fragility disorders or premature ageing to tumour progression in cancer.
Results Expression profiling of the hypo-osmotic stress response To identify the transcriptome of the hypo-osmotic shock response in keratinocytes expressing mutant keratins, three immortalized keratinocyte cell lines were used: NEB-1 (a keratin-normal control line), KEB-4 (mild EBS phenotype, K14 V270M mutation) and KEB-7 (severe EBS phenotype, K14 R125P mutation) [7]. Cells were exposed to 150 mM urea for 5 min, which causes transient cell swelling, with severe but transient disruption of the cytoskeleton and activation of the JNK/SAPK pathway as previously described [5]. Cells were then allowed to recover for 1 h, and untreated and treated cells harvested. RNA was purified, labeled and hybridized to Affymetrix HU95A chips, and the data extracted from the differential hybridization were analysed. All experiments were performed in triplicate. To test the reproducibility of the microarray data, hierarchical clusters of gene expression were generated using GeneSpring™ software for both conditions, before and after hypo-osmotic shock (Fig. 1A). Highly similar patterns were obtained in all three experimental replicates, indicating good reproducibility of methodology. Clusters of genes that were differently regulated in the EBS lines vs control cells could be readily identified (see arrows in Fig. 1A), as could other clusters of genes showing altered expression unique to an individual cell line (see asterisks in Fig. 1A). Scatter correlation graphs were also generated for each triplicate experiment (data not shown), which further confirmed the hierarchical clustering data and reproducibility of the methods used. Median expression level values for the two states, before and after hypo-osmotic stress, were compared for each cell line individually. As previously reported for SAPK (stress activated protein kinase) activation [5], the degree of response to hypoosmotic stress was found to be correlated with the severity of the keratin mutation expressed, and thus with the severity of the EBS phenotype. By 1 h of recovery from hypo-osmotic stress, the wild type keratin cell line (NEB-1) showed significantly altered expression of 261 genes as compared to unstressed NEB1 cells; of these changes, 203 were down-regulated and 58 upregulated (Supplemental Table 5). The mild EBS mutant cell line KEB-4 showed fewer changes from its baseline, with altered expression in only 174 genes after comparison to before hypoosmotic shock, of which only 39 were down-regulated and 135 were up-regulated (Supplemental Table 6). The most severe EBS mutant (KEB-7 cells, with the severe K14 R125P mutation) showed the least response to osmotic stress, with significant alterations in only 55 genes, i.e. a third as many as the mild mutant and only a quarter of the number altered in the wild type cell line. As in the mild mutant, more genes were up-regulated (43 genes) than down-regulated (12 genes) (Supplemental
2068
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
Table 7). Comparison between these three cell lines is summarized in Figs. 1B and C. It was notable that while the control NEB-1 cell line responded to hypo-osmotic shock predominantly by down-regulation of differentially expressed genes (77% — only 23% were up-regulated), both the KEB-4 (mild) mutant cells and the KEB-7 (severe) mutant cells predominantly up-regulated genes (78% up, 22% down) (see Fig. 1B and Supplemental Tables 5, 6, 7). This suggests that the cell lines are initiating their response to the osmotic stress from a very different starting point.
Hypo-osmotic stress affects a shared group of genes independent of keratin mutation status The 16 genes whose expression was affected by hypoosmotic stress in all three cell lines are shown in Table 1. These genes, which respond in all three cell lines, are likely to have a core role in the response to hypo-osmotic stress in epithelial cells. In the two EBS lines, thirteen of these genes were clearly induced by hypo-osmotic stress, and three (a transcription factor, RQCD1 [13], and two integral
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
2069
Table 1 – The 16 genes responding to hypo-osmotic stress in all three cell lines
The results are shown fold changes compared unstressed cells. Green indicates reduced RNA levels while shows elevated signal The results are shown asas fold changes asas compared to to unstressed cells. Green indicates reduced RNA levels while redred shows elevated signal (over the + 2/− 2-fold threshold).
membrane proteins, AFURS1 (with a role in cation transport) [14] and B4GALT1 (an enzyme with a role also in cell– cell recognition and cell adhesion) [15]) were suppressed. Amongst the thirteen induced genes are the dual-specificity phosphatase hVH3 (also known as DUSP5) and several transcription factors that are important regulators of cell differentiation, such as ID2, FOS, JUN, SOX4, MAFF. In this broadly responsive set of genes, the dual-specificity phosphatases are particularly interesting because of their known role in MAPK inactivation, and therefore in stress response modulation. Inhibitor of DNA binding, ID2, belongs to the family of proteins that inhibit basic helix– loop–helix transcription factors in a dominant-negative way, and is known to play a role in promoting cell proliferation. The oncogenes FOS and JUN are components of the AP-1 transcription factor complex, which like MAFF, have a leucine zipper domain and function as regulators of cell death and proliferation. All of these genes function as promoters of cell survival and positive regulators for cell cycle progression.
Genes responding to hypo-osmotic stress in the keratinmutant lines The two K14-mutant cell lines had 20 genes in common that responded to hypo-osmotic stress but remained unchanged in the control cell line. These genes, which were all up-regulated in both K14 mutant cell lines, are listed in Table 2. This group is of interest as it includes several genes with a role in growth regulation and growth inhibition. Transforming growth factor β-inducible early growth response protein 1 (TIEG1) is a transcriptional repressor involved in the regulation of cell growth [16]. Cyclic AMP-dependent transcription factor 3 (ATF-3) may repress transcription by stabilizing the binding of inhibitory cofactors at the promoter, whereas GADD45 can recognize an altered chromatin state and modulate DNA accessibility to cellular proteins. Cyclin-dependent kinase inhibitor 1C (CDKN1C) is a negative regulator of cell proliferation [17] that may play a role in maintenance of the nonproliferative state throughout life. Kruppel-like factor 4 (KLF4) plays a role in the development of the epidermal permeability
Fig. 1 – Microarray analysis of one control and two EBS patient-derived cell lines. (A) Hierarchical clustering of the microarray data obtained from triplicate experiments of hypo-osmotic stress response in cell lines with wild type vs mutant keratins. The red color represents expression level above (2× or more) mean expression of a gene across all samples, the yellow color represents less than 2× deviation from the mean expression and the green color represents expression lower (2× or more) than the mean. Keratinocyte cell lines NEB-1 (wild type keratins), KEB-4 (carrying a mildly disruptive EBS K14 mutation, V270 M) and KEB-7 (with a clinically severe EBS K14 mutation, R125P), were used. Relative amounts of mRNA for each of over 12,000 genes were assayed before and 1 h after exposure to hypo-osmotic stress. Clustering of (more than 2-fold) up- and down-regulated genes in each of the triplicate assays for each cell line was analysed using GeneSpring™ software and shows highly similar patterns in all three experimental replicates, indicating good reproducibility. Specific clusters of regulated genes that are unique for each of the cell lines are easily identified (see asterisks), as well as those common for both mutant cell lines (arrows). (B) Different numbers of genes respond to hypo-osmotic stress in each cell line. Total mRNA expression profiles of the three cell lines were cross-compared before and after exposure to hypo-osmotic stress and the results are summarized here (for a full list see Supplemental Tables 5, 6 and 7). The biggest change is seen in the control cells and the least change in the cells expressing a clinically severe mutation, suggesting that the severe cells are most “pre-stressed” cells and require fewer changes in gene expression to convert to a full stress response. (C) Venn diagram summarizing the microarray data obtained on the investigated keratinocyte cell lines.
2070
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
Table 2 – Genes, which showed significantly altered transcription in response to hypo-osmotic stress in the two K14 mutants (KEB-4, KEB-7), but were not significantly altered in the wild type keratin (NEB-1) cells
The results are shown as fold changes compared to unstressed cells.
barrier and accelerates terminal differentiation in the epidermis [18]. Immediate early response 3, IER3, may be important in cell survival by regulating anti-apoptotic genes [19]. IL-6 mediates activation of STAT3 in squamous cell carcinoma of the head and neck, while JUNB is a member of AP1 family and takes part in cell cycle regulation. Thus the two K14 mutant cells appear to be predominantly responding to hypo-osmotic stress by inhibiting transcription and proliferation.
Dual-specificity phosphatases are differentially regulated in keratin-mutant cell lines In addition to hVH3/DUSP5, several other members of the dualspecificity phosphatase family (DUSP, or MKP (MAP kinase phosphatase)) were found to respond to hypo-osmotic stress (Tables 3 and 4). In the resting state (before stress), the expression levels of MKP genes in the two K14 mutant cell lines are constitutively lower than in the control cell line (NEB-1). The microarray results (MA, Table 3) were confirmed by real-time PCR (RT-PCR) as summarized in Table 3. MKP-1, MKP-3 and hVH3 are significantly down-regulated in both mutants when
Table 3 – Dual-specificity phosphatases show lower mRNA levels during resting state in the K14 mutant lines (KEB-4, KEB-7) than in the wild type cells (NEB-1)
MA = microarray data; RT-PCR = real-time PCR data; – = expression MA = microarray data; RT-PCR = real-time PCR data; – =(outside expression levels not significantly different from the wild type the + 2/ −2-fold threshold) indifferent this cell from line. the wild type (outside the +2/−2levels not significantly
compared to the wild type. MKP-5 was also found downregulated in KEB-7 cells (severe mutant), but was similar to the wild type control level in KEB-4 cells (mild mutant) (Table 3). When we compared the transcription of MKP genes in the unstressed state (before hypo-osmotic stress) with that at 1 h of recovery for each cell line, we found that all these phosphatases were differentially regulated in response to stress (Table 4). In addition, dual-specificity phosphatase expression in the control NEB-1 cells was down-regulated from the starting point, whereas in the two K14 mutant lines (KEB-4 and KEB-7) these genes were up-regulated (Table 4). Changes in these phosphatases were also evaluated by RT-PCR (Table 4). While most of the data was confirmed (apart MKP-2 down-regulation in the control, NEB-1 cells), RT-PCR revealed an additional up-regulation of MKP-1 and MKP-3 in the severe mutant (KEB-7 cells), which was not detected by microarray analysis (less than 2-fold change). In view of their role in attenuating MAP kinase activation and the observed down-regulation of MKPs in both mutants during resting state, we would expect this to affect the regulation of the stress response and MAPK activation. We therefore investigated the time-course of the stress response to hypo-osmotic shock in these cell lines [5], and analysed protein levels of total and phosphorylated ERK and p38, and phosphatases MKP-1 and hVH3 in response to stress (Fig. 2A). Total p38 and ERK levels were unchanged in response to stress in all three cell lines. However, activation of p38 and ERK (as deduced from their phosphorylation) was stronger and persisted for longer in the mutant (KEB-7 and KEB-4) than in the wild type cells (NEB-1). In the latter, the stress response was already shutting down after the 1 h time-point whereas it was still strong and persisting in the two mutants (Fig. 2A). Both hVH3 and MKP-1 proteins showed an obvious induction in response to stress in all three cell lines (Fig. 2A). In addition, hVH3 was not detected in any of the cell lines until the 30 min recovery time-point, indicating that in keratinocytes hVH3 is clearly a hypo-osmotic stress
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
2071
Table 4 – Dual-specificity phosphatases showing significantly altered transcription at 1 h of recovery after hypo-osmotic stress
MA = microarray data; RT-PCR = real-time PCR data; – = expression levels not significantly different from the wild type (outside the + 2/ MA = microarray data;inRT-PCR = real-time PCR data; – = expression levels not significantly different from the wild type (outside the + 2/−2-fold − 2-fold threshold) this cell line.
inducible phosphatase. Although all cell lines responded to hypo-osmotic stress with hVH3 synthesis and increased phosphorylation of both ERK and p38, this was more prominent in the two K14 mutants (KEB-4 and KEB-7 cells). We also examined the protein levels of total and phosphorylated keratin in the samples (Fig. 2B), which due to the method of preparation (see Materials and methods) would be depleted in filamentous (pelleted) keratin. Like other
intermediate filaments, phosphorylation of keratins is associated with proteins in a non-filamentous state. Consistent with the presence of unpolymerized (misfolded) keratin protein, in the mild (KEB-4) and severe K14 mutant (KEB-7), phosphorylated keratin (p-K5 and possibly p-K6, due to known cross-reactivity of the antibody used (LJ4) and the type of cell (keratinocytes in culture)) was clearly detectable in the absence of hypo-osmotic stress. In response to stress, the
Fig. 2 – Time-course experiment of the response of wild type and K14 mutant keratinocytes to hypo-osmotic stress. Protein levels of target proteins were tested in wild type (NEB-1) and two K14 mutant cells, KEB-4 (mild mutant) and KEB-7 (severe mutant), before (“0”) and after hypo-osmotic stress at 5, 10, 30 min and 1 h and 2 h of recovery time-points. (A) Total ERK and p38 protein levels are unchanged in response to stress in all cell lines. Hypo-osmotic stress activates ERK and p38, with phosphorylated ERK (p-ERK) and phosphorylated p38 (p-p38) levels higher and persisting for longer in the two K14 mutants, in particular in the severe K14 mutant (KEB-7). hVH3 (DUSP5) phosphatase is a hypo-osmotic shock inducible protein, as it is absent before stress and not detectable until 30 min of recovery. Higher hVH3 levels are detected in the K14 mutants than in the wild type. MKP-1 (DUSP1) is also seen as induced in response to hypo-osmotic stress. (B) Phosphorylated keratin (p-K5) is higher in the K14 mutants than in the wild type cells prior to stress. Hypo-osmotic stress induces keratin phosphorylation in the wild type cells (NEB-1) up to 1 h of recovery, but results in significant reduction of keratin phosphorylation in the two K14 mutants (KEB-4, KEB-7) after 10–30 min. Total K5 is reduced at late time-points (“1 h” and “2 h”) in the wild type, but continues to increase in the severe mutant (KEB-7). GAPDH was used as the loading control; 10 μg of protein from total cell lysate samples was loaded in each track on the gel.
2072
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
levels of phosphorylated keratin increased in the wild type cells (NEB-1), but decreased at later time-points of recovery in the two mutants. Total K5 also decreased at later time-points in the wild type and the mild mutant (KEB-4), while it increased in the severe mutant (KEB-7). This suggests that (hypoosmotic) stress may increase the level of unpolymerized keratin protein in the cytoplasm.
Discussion Tissue injury and stress play important roles in the phenotype of all diseases, but the actual molecular events leading to such changes are not well documented. Intermediate filaments of the cytoskeleton are increasingly becoming recognized as contributing to the defence against stress of cells in tissues, and experiments on chemical stress, heat stress, osmotic stress and mechanical stress in cells with mutant intermediate filament proteins continue to accumulate data that support this view [20]. In this study we have used hypo-osmotic stress to examine the effect of mutations in keratin intermediate filament proteins on a stress response. Osmotic stress, especially hypo-osmotic stress, is one of the most common types of stress affecting cells in tissues. In the body it results rapidly from loss of energy supply to tissue cells, as caused by anything from indirect ischemia to direct wounding: the energy-dependent activity of membrane ion pumps cannot be maintained and osmotic balance is lost, leading to influx of water and swelling of cells. Experimentally induced hypo-osmotic shock causes transient cell swelling and extensive disruption of all three cytoskeleton filament systems (actin, tubulin and intermediate filaments [5]), which cells counteract by raising the activity of membrane pumps. We previously showed that the persistence, intensity, and the timing of induction of the JNK hypo-osmotic stress response varied between keratinocytes expressing different keratin mutations [5]. Here we present further analysis of the hypo-osmotic stress response, using microarray expression profiling of three keratinocyte cell lines expressing normal keratins or keratin mutations of differing clinical severity. The cell lines selected are morphologically similar, immortalized in the same way and showing similar differentiation and growth properties under standard culture conditions, but they show a quantitatively different response to stress that correlates with the clinical severity of the keratin mutations they express [7]. Here we demonstrate that already within the first hour after hypoosmotic stress, the K14 defective cell lines had altered the expression of many genes. At least 16 genes appear to be related to the hypo-osmotic stress itself, rather than to the stress of expressing and handling mutant keratin proteins, as they are altered in both mutant and wild type cells, albeit sometimes in different ways. In the present experiments the number of genes seen to “respond” to hypo-osmotic stress is inversely correlated with clinical severity associated with the keratin mutation in the cell — more severe disease phenotype, fewer genes altered. This is consistent with earlier observations [5,7] that suggested that due to the presence of mutant keratins, the EBSderived cells show an active stress response even before osmotic stress is applied. Both K14 mutant lines respond to
hypo-osmotic stress in a similar way, which is strikingly different from the wild type. In the control NEB-1 cells, 78% of the altered genes were down-regulated, while in the mutant cells 77% of genes are up-regulated. If the presence of mutant keratin protein induces an intrinsic physiological stress by itself, then these cells may have fewer additional adjustments to make when challenged by an additional extrinsic stress. The gene expression profile also suggests that upon hypoosmotic stress, these keratin-mutant cells are being pushed to another level of stress response. 20 of the genes are only altered (up-regulated) in the mutant lines (Table 2), and these genes include several transcriptional factors with a role in growth arrest. Of this group, ATF3, TIEG, JunB, IER3, GADD45A and CDKN1C have been recently found strongly up-regulated upon treatment with oncostatin M in several breast tumour cell lines (T47D, MCF-7, MDA-MB-231) [21]. Oncostatin M is a potent cytostatic cytokine belonging to the IL-6 family that causes cell cycle arrest in late G1 phase. Similar to these breast tumour cell lines, the K14 mutant keratinocytes respond to hypo-osmotic stress by up-regulating the transcription of IL-6 and also the same group of genes: ATF3, TIEG, JunB, IER3, GADD45A and CDKN1C. Transcriptional profiles of these genes remained unchanged in the control keratinocyte cell line at equivalent time-points after stress. This suggests a significant parallel between the sets of genes activated by hypo-osmotic stress in K14-mutant keratinocytes and the ones activated by a cytostatic cytokine in several breast cancer cell lines. In both situations the increase in transcription of an IL-6 family cytokine results in the increase of transcription of specific downstream targets involved in regulating growth arrest and apoptosis [22,23]. It would appear that the EBS keratinocytes, with their constitutive stress already elevated by the presence of mutant keratin protein, are propelled into growth arrest by the additional stress of hypo-osmotic shock. This may be a protective mechanism to permit damage assessment and prevent proliferation of fatally damaged cells. We have previously shown (see “metabolic recovery after hypo-osmotic stress” assay in [5]) that growth arrest may occur in keratin-mutant cells after hypo-osmotic shock. Here we have now identified a number of genes that may be involved in this process. Amongst the differentially expressed genes, the dualspecificity phosphatases (MKPs or DUSPs) were considered especially significant, as these are important regulatory proteins for phosphorylation-driven MAP kinase pathways [24]. MKPs are involved in the regulation of the cell stress response by establishing a negative regulatory feedback loop that targets and dephosphorylates activated MAP kinases, the ERK, JNK and p38 pathways [24–26]. The importance of MKPs has been further emphasized by recent findings that implicate them in the development of some forms of cancer [27–34]. Our microarray data has shown that hVH3, an ERK MAP kinase specific MKP [35], is part of the set of genes that may contribute to the generic way epithelial cells respond to hypo-osmotic shock or even other stresses. Other MKPs were also found differentially regulated in response to hypo-osmotic shock, but not in all of cell lines investigated (Table 4). Furthermore several MKPs were already significantly down-regulated in the two K14 mutant lines before exposure to stress (Table 3, compared to the control cells). This is mirrored by the increased phosphorylation of their downstream targets, p38 and ERK (Fig. 2A).
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
There is substantial evidence in the literature that MKPs have important roles in regulating cell metabolism and p38, JNK and ERK-mediated responses in a variety of cells and tissues [24– 26]. The data presented here indicates that MKPs may also be important in a different pathological situation, i.e. in epidermal cell lines expressing mutant K14 associated with a tissue fragility disorder (EBS). At present, it is not clear which specific mechanism is involved in triggering the observed increased phosphorylated MAPK levels prior to stress, or the differential regulation of MKPs in keratin-mutant cells. Nevertheless, phosphorylated p38 has been recently found to co-localize with keratin granules in mutant keratin expressing cells [36]. Fig. 2 shows that phosphorylated keratin is detectable in the Triton-soluble phase of total cell extracts from the two K14 mutants prior to stress, which would be expected in the presence of unpolymerised keratin, although this appears to decrease after stress. Hypo-osmotic stress induces cytoskeleton remodelling [5], and this is marked by an increase in phosphorylated soluble keratin in the wild type cells (Fig. 2B). Total K5 decreased at later time-points of recovery in the NEB-1 and KEB-4 cells, while it increased in the KEB-7 cells (severe mutant). Seen as a whole, hypo-osmotic stress may augment not only the levels of unpolymerized (soluble) keratin protein in the cytoplasm (part of the general keratinocyte response), but when combined with a severe keratin mutation this may also increase the amount of misfolded protein, interfering with all aspects of keratin dynamics and regulation. Intermediate filament protein aggregates are associated with many pathologies and could arise, or be exacerbated, by many different mechanisms, from altered chaperone behaviour to defective protein degradation [37,38]. It is also known that the phosphorylation state of intermediate filament proteins drives their assembly and disassembly during mitosis or exposure to stress [39,40]. It has been suggested that keratins may act as a “phosphate sponge” for activated MAPKs, whereby depending on the protein variant (wild type or mutant) they may either have a protective role (by buffering MAPK activity) or predispose to disease development (if misfolded mutant keratin restricts phospho-site availability), by deflecting and generating a more intense MAPK signal transduction into the nucleus instead [41]. Our data may support this hypothesis since in response to stress there is a decrease in phosphorylated keratin in the two K14 mutants, while it increases in the wild type cells (Fig. 2B). In response to stress, the levels of p-p38 and p-ERK were also higher and persisted for longer in the two K14 mutants. Thus depending on the overall detrimental effect that a mutation has on the structure and function of keratin intermediate filament proteins, several signalling pathways may be triggered. Therefore, a better understanding of the way in which epithelial cells respond to stress is needed, as this will undoubtedly have a bearing on the development of new therapeutic strategies.
Materials and methods Cell lines, culture conditions and osmotic shock Immortalized keratinocyte cell lines NEB-1 (control keratinocyte cell line), KEB-4 (mild, Weber-Cockayne EBS phenotype caused by the K14 V270M mutation) and KEB-7 cells (severe,
2073
Dowling-Meara EBS phenotype caused by the K14 R125P mutation) [7] were cultured under standard conditions (RM medium, 37 °C, 5% CO2, without feeder cells) on 10 cm Petri dishes to 80% confluence. Each cell line was grown in triplicate (3 plates). For the two experimental situations, i.e. before and 1 h after osmotic shock, we grew 3 × 2 plates per cell line. At 80% confluence, one triplicate set was hypo-osmotically challenged by adding 150 mM urea to the culture medium for 5 min [5]; cells were then incubated in fresh tissue culture medium for a further 1 h at 37 °C to allow the response to develop in a recovery state. At this point all cells were taken out of the incubator, put on ice, washed with two changes of phosphatebuffered saline (PBS) and processed as below.
Total RNA isolation and labeling Total cellular RNA was isolated using RNeasy Mini Kit (Qiagen Ltd, UK) according to the manufacturer's instructions [42]. The optical density was measured and 8 μg of total RNA extract was used per reaction for double-stranded cDNA synthesis, using SuperScript Double-Stranded cDNA Synthesis Kit (Invitrogen Ltd., UK). Products were then precipitated and resuspended in RNase-free water and the entire sample was used for a subsequent in vitro transcription and cRNA labeling reaction (BioArray HighYield RNA Transcript Labeling kit, Enzo Life Sciences Inc., New York). Labeled samples were fragmented and hybridized to a test chip to check the efficiency and quality of samples.
Transcriptional profiling and microarray data analysis The Affymetrix Human Genome U95A GeneChip with 12627 probes for the human genome was used (Affymetrix Inc., California). Microarray Suite 5.0 (Affymetrix) was used for data extraction and for further analysis, data mining tool 3.0 (Affymetrix) and GeneSpring™ software 5.1 (Agilent Technologies, California) were used for normalization, extent of change calculations and clustering [42]. Differential expression levels of transcripts were determined as the increase or decrease in expression in relation to the paired sample (as 2-fold, 3-fold change etc.). To compare data from multiple arrays, the signal of each probe array was scaled to the same target intensity value. Genes were considered to show altered expression if the expression levels differed more than 2-fold relative to untreated control at any time-point. Using GeneSpring™, clustering was performed based on experiments or individual genes expression profiles. To aid analysis we assembled an extensive gene annotation table listing the molecular function and biological category of the genes represented on the chip, based primarily on data from J.M. Ruillard and [43].
Real-time PCR Extracted total RNA was used as template for cDNA synthesis using SuperScript III First-Strand Synthesis System for RT-PCR (Invitrogen Ltd, UK) following the manufacturer's instructions. Subsequently, 1 μl of cDNA product was used per each real-time PCR reaction using Platinum Taq DNA polymerase (Invitrogen Ltd, UK) and SYBR Green I nucleic acid stain (Molecular Probes, USA). The MKP gene specific primers were obtained from Qiagen (QuantiTect Primer Assays), and used along Qiagen's
2074
E XP E RI ME N TA L CE LL RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
standard protocol on an iQ5 Biorad real-time PCR detection system. GAPDH and β-actin were used as internal controls.
[2]
Total protein extraction [3]
Cells were cultured as described above. After hypo-osmotic shock cells were washed with PBS and total protein was extracted using an extraction buffer containing the following: 50 mM Tris/HCl pH 7.5, 1 mM EDTA, 1 mM EGTA, 50 mM NaF, 5 mM sodium pyrophosphate, 10 mM β-phosphoglycerate and 1% Triton X-100, plus the following which were added fresh, immediately prior to use: 1 mM Na-orthovanadate, 0.1% (β-mercaptoethanol, 1 mM PMSF, 1 mM benzamidin, 10 μg/ml leupeptin and 10 μg/ml pepstatin A). Protein concentration was determined by Bradford assay (BioRad protein assay, Bio-Rad Gmbh., Germany) and 10 μg of total protein per sample was loaded on denaturing protein gels.
Antibodies To detect MAP kinases, rabbit anti-phosphorylated-ERK, rabbit anti-phosphorylated-p38, rabbit anti-ERK and rabbit anti-p38 (Cell Signaling Technology Inc., Massachusetts) antisera were used. Mouse monoclonal LJ4 was used to detect phosphorylated keratin [39], and rabbit polyclonal antibody BL-18 was used to detect K5 [44]. Rabbit anti-MKP-1 (Santa Cruz Biotechnology, USA) and rabbit anti-hVH3 (a gift from J.E. Dixon, University of San Diego) antibodies were used to detect DUSP1 and DUSP5 respectively. To detect GAPDH, a mouse monoclonal antibody was used (Abcam plc., UK). As secondary antibodies, anti-rabbit immunoglobulin (Cell Signaling) and antimouse immunoglobulin (DakoCytomation, Denmark) antisera conjugated to HRP were used.
[4]
[5]
[6]
[7]
[8]
[9]
[10]
Acknowledgments [11]
We would like to thank V. Makrantoni, D. Martin, S. Keyse and D. Russell for their useful discussions, and J.K. Heath and N. Underhill-Day for their helpful comments and access to unpublished information. We would also like to thank J.E. Dixon for supplying the hVH3 antibody. This work was funded by Cancer Research UK (grant C26/A1461 to E.B. Lane), the Wellcome Trust (grant 5090/A/98/Z to E.B. Lane), the Slovenian Research Agency (grant J3-6132-0381-04 to M. Liovic), and the National Institutes of Health (grants AR45974 and NR08029 to M. Tomic-Canic).
[12]
[13]
[14]
Appendix A. Supplementary data [15]
Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.yexcr.2008.02.020. [16] REFERENCES [17] [1] M. Hesse, T.M. Magin, K. Weber, 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 (2001) 2569–2575. M. Hesse, A. Zimek, K. Weber, T.M. Magin, Comprehensive analysis of keratin gene clusters in humans and rodents, Eur. J. Cell Biol. 83 (2004) 19–26. M.B. Omary, P.A. Coulombe, W.H. McLean, Intermediate filament proteins and their associated diseases, N. Engl. J. Med. 351 (2004) 2087–2100. I. Szeverenyi, A.J. Cassidy, C.W. Chung, B.T.K. Lee, J.E. Common, S.C. Ogg, H. Chen, S.Y. Sim, W.L.P. Goh, K.W. Ng, J.A. Simpson, L.L. Chee, G.H. Eng, L. Bin, D. Chuon, A. Venkatesh, K.H. Khoo, W.H.I. McLean, Y.P. Lim, E.B. Lane, The human intermediate filament database: comprehensive information on a gene family involved in many human diseases, Human Mutat. 29 (2008) 351–360 Electronic publication ahead of print, Nov 2007. M. D'Alessandro, D. Russell, S.M. Morley, A.M. Davies, E.B. Lane, Keratin mutations of epidermolysis bullosa simplex alter the kinetics of stress response to osmotic shock, J. Cell Sci. 115 (2002) 4341–4351. M. Liovic, M.M. Mogensen, A.R. Prescott, E.B. Lane, Observation of keratin particles showing fast bidirectional movement colocalized with microtubules, J. Cell Sci. 116 (2003) 1417–1427. S.M. Morley, M. D'Alessandro, C. Sexton, E.L. Rugg, H. Navsaria, C.S. Shemanko, M. Huber, D. Hohl, A.I. Heagerty, I.M. Leigh, E.B. Lane, Generation and characterization of epidermolysis bullosa simplex cell lines: scratch assays show faster migration with disruptive keratin mutations, Br. J. Dermatol. 149 (2003) 46–58. D. Russell, P.D. Andrews, J. James, E.B. Lane, Mechanical stress induces profound remodelling of keratin filaments and cell junctions in epidermolysis bullosa simplex keratinocytes, J. Cell Sci. 117 (2004) 5233–5243. H. Bar, N. Mucke, A. Kostareva, G. Sjoberg, U. Aebi, H. Herrmann, Severe muscle disease-causing desmin mutations interfere with in vitro filament assembly at distinct stages, Proc. Natl. Acad. Sci. U. S. A. 102 (2005) 15099–15104. A. Letai, P.A. Coulombe, E. Fuchs, Do the ends justify the mean? Proline mutations at the ends of the keratin coiled-coil rod segment are more disruptive than internal mutations, J. Cell Biol. 116 (1992) 1181–1195. J. Liu, M. Tang, R. Mestril, X. Wang, Aberrant protein aggregation is essential for a mutant desmin to impair the proteolytic function of the ubiquitin-proteasome system in cardiomyocytes, J. Mol. Cell Cardiol. 40 (2006) 451–454. T. Sasaki, T. Gotow, M. Shiozaki, F. Sakaue, T. Saito, J.P. Julien, Y. Uchiyama, S. Hisanaga, Aggregate formation and phosphorylation of neurofilament-L Pro22 Charcot-Marie-Tooth disease mutants, Hum. Mol. Genet. 15 (2006) 943–952. N. Hiroi, T. Ito, H. Yamamoto, T. Ochiya, S. Jinno, H. Okayama, Mammalian Rcd1 is a novel transcriptional cofactor that mediates retinoic acid-induced cell differentiation, EMBO J. 21 (2002) 5235–5244. N. Habtemichael, G. Kovacs, Cloning the AFURS1 gene which is up-regulated in senescent human parenchymal kidney cells, Gene 283 (2002) 271–275. K.A. Masri, H.E. Appert, M.N. Fukuda, Identification of the full-length coding sequence for human galactosyltransferase (beta-N-acetylglucosaminide: beta-1,4-galactosyltransferase), Biochem. Biophys. Res. Commun. 157 (1988) 657–663. M. Subramaniam, J.R. Hawse, S.A. Johnsen, T.C. Spelsberg, Role of TIEG1 in biological processes and disease states, J. Cell Biochem. 102 (2007) 539–548. J.M. Scandura, P. Boccuni, J. Massague, S.D. Nimer, Transforming growth factor beta-induced cell cycle arrest of human hematopoietic cells requires p57KIP2 upregulation, Proc. Nat. Acad. Sci. 101 (2004) 15231–15236.
E XP E RI ME N TA L CE L L RE S E A RCH 3 1 4 ( 2 00 8 ) 2 0 6 6 –20 7 5
[18] J. Jaubert, J. Cheng, J.A. Segre, Ectopic expression of kruppel like factor 4 (Klf4) accelerates formation of the epidermal permeability barrier, Development 130 (2003) 2767–2777. [19] L. Shen, J. Guo, C. Santos-Berrios, M.X. Wu, Distinct domains for anti- and pro-apoptotic activities of IEX-1, J. Biol. Chem. 281 (2006) 15304–15311. [20] E.B. Lane, M. Pekny, Stress models for the study of intermediate filament function, Methods Cell Biol. 78 (2004) 229–264. [21] N. Underhill-Day, J.K. Heath, Oncostatin M (OSM) cytostasis of breast tumor cells: characterization of an OSM receptor beta-specific kernel, Cancer Res. 66 (2006) 10891–10901. [22] S. Bosze, F. Hudecz, P. Igaz, Z. Ortutay, G. Csik, A. Falus, S. Toth, Interleukin-6 N-terminal peptides modulate the expression of junB protooncogene and the production of fibrinogen in HepG2 cells, Biol. Chem. 384 (2003) 409–421. [23] T. Hai, C.D. Wolfgang, D.K. Marsee, A.E. Allen, U. Sivaprasad, ATF3 and stress responses, Gene Expr. 7 (1999) 321–335. [24] D.M. Owens, S.M. Keyse, Differential regulation of MAP kinase signalling by dual-specificity protein phosphatases, Oncogene 26 (2007) 3203–3213. [25] A. Theodosiou, A. Ashworth, MAP kinase phosphatases, Genome Biol. 3 (2002) REVIEWS3009. [26] A. Farooq, M.M. Zhou, Structure and regulation of MAPK phosphatases, Cell Signal. 16 (2004) 769–779. [27] H.Y. Wang, Z. Cheng, C.C. Malbon, Overexpression of mitogen-activated protein kinase phosphatases MKP1, MKP2 in human breast cancer, Cancer Lett. 191 (2003) 229–237. [28] C.D. Swartz, C.A. Afshari, L. Yu, K.E. Hall, D. Dixon, Estrogen-induced changes in IGF-I, Myb family and MAP kinase pathway genes in human uterine leiomyoma and normal uterine smooth muscle cell lines, Mol. Hum. Reprod. 11 (2005) 441–450. [29] R. Steinmetz, H.A. Wagoner, P. Zeng, J.R. Hammond, T.S. Hannon, J.L. Meyers, O.H. Pescowitz, Mechanisms regulating the constitutive activation of the extracellular signal-regulated kinase (ERK) signaling pathway in ovarian cancer and the effect of ribonucleic acid interference for ERK1/2 on cancer cell proliferation, Mol. Endocrinol. 18 (2004) 2570–2582. [30] S. Vicent, M. Garayoa, J.M. Lopez-Picazo, M.D. Lozano, G. Toledo, F.B. Thunnissen, R.G. Manzano, L.M. Montuenga, Mitogen-activated protein kinase phosphatase-1 is overexpressed in non-small cell lung cancer and is an independent predictor of outcome in patients, Clin. Cancer Res. 10 (2004) 3639–3649. [31] J.K. Warmka, L.J. Mauro, E.V. Wattenberg, Mitogen-activated protein kinase phosphatase-3 is a tumor promoter target in initiated cells that express oncogenic Ras, J. Biol. Chem. 279 (2004) 33085–33092. [32] T. Furukawa, R. Fujisaki, Y. Yoshida, N. Kanai, M. Sunamura, T. Abe, K. Takeda, S. Matsuno, A. Horii, Distinct progression
[33]
[34]
[35]
[36]
[37]
[38]
[39]
[40]
[41]
[42]
[43]
[44]
2075
pathways involving the dysfunction of DUSP6/MKP-3 in pancreatic intraepithelial neoplasia and intraductal papillary-mucinous neoplasms of the pancreas, Mod. Pathol. 18 (2005) 1034–1042. S. Xu, T. Furukawa, N. Kanai, M. Sunamura, A. Horii, Abrogation of DUSP6 by hypermethylation in human pancreatic cancer, J. Hum. Genet. 50 (2005) 159–167. L. Nonn, L. Peng, D. Feldman, D.M. Peehl, Inhibition of p38 by vitamin D reduces interleukin-6 production in normal prostate cells via mitogen-activated protein kinase phosphatase 5: implications for prostate cancer prevention by vitamin D, Cancer Res. 66 (2006) 4516–4524. M. Mandl, D.N. Slack, S.M. Keyse, Specific inactivation and nuclear anchoring of extracellular signal-regulated kinase 2 by the inducible dual-specificity protein phosphatase DUSP5, Mol. Cell Biol. 25 (2005) 1830–1845. S. Wöll, R. Windoffer, R.E. Leube, p38 MAPK-dependent shaping of the keratin cytoskeleton in cultured cells, J. Cell Biol. 177 (2007) 795–807. E. Janig, C. Stumptner, A. Fuchsbichler, H. Denk, K. Zatloukal, Interaction of stress proteins with misfolded keratins, Eur. J. Cell Biol. 84 (2005) 329–339. M. Der Perng, M. Su, S.F. Wen, R. Li, T. Gibbon, A.R. Prescott, M. Brenner, R.A. Quinlan, The Alexander disease-causing glial fibrillary acidic protein mutant, R416W, accumulates into Rosenthal fibers by a pathway that involves filament aggregation and the association of alpha B-crystallin and HSP27, Am. J. Hum. Genet. 79 (2006) 197–213. D.M. Toivola, Q. Zhou, L.S. English, M.B. Omary, Type II keratins are phosphorylated on a unique motif during stress and mitosis in tissues ad cultured cells, Mol. Biol. Cell. 13 (2002) 1857–1870. T. Yamaguchi, H. Goto, T. Yokoyama, H. Silljé, A. Hanisch, A. Uldschmid, Y. Takai, T. Oguri, E.A. Nigg, M. Inagaki, Phosphorylation by Cdk1 induces Plk1-mediated vimentin phosphorylation during mitosis, J. Cell Biol. 171 (2005) 431–436. N. Ku, M.B. Omary, A disease and phosphorylation-related nonmechanical function for keratin 8, J. Cell Biol. 174 (2006) 115–125. O. Stojadinovic, B. Lee, C. Vouthounis, S. Vukelic, I. Pastar, M. Blumenberg, H. Brem, M. Tomic-Canic, Novel genomic effects of glucocorticoids in epidermal keratinocytes: inhibition of apoptosis, IFN gamma pathway and wound healing along with promotion of terminal differentiation, J. Biol. Chem. 282 (2007) 4021–4034. Gene Ontology Consortium Data [http://cgap.nci.nih.gov/ Genes/GOBrowser] [dot.ped.med.umich.edu:2000/ourimage/ pub/shared/JMR_pub_affyannot.html]. Purkis, J.B. Steel, I.C. Mackenzie, W.B. Nathrath, I.M. Leigh, E.B. Lane, Antibody markers of basal cells in complex epithelia, J. Cell Sci. 97 (1990) 39–50.