A Novel Regulator of Telomerase

Recently we reported a differentiation-dependent inhibition of telomerase activity in human epidermis. Consistent with this observation we found that in keratinocyte cultures calcium-induced differentiation correlates with a decline in telomerase activity. To get further support for a role of calcium in the regulation of telomerase and to elucidate the underlying molecular mechanisms we investigated the effect of calcium on telomerase in the human epidermal keratinocyte line HaCaT. Treatment with thapsigargin, which increases intracellular calcium concentrations, inhibited telomerase activity without down-regulating the expression of hTERT (human telomerase reverse transcriptase). This observation together with the fact that increasing calcium reduced telomerase activity in cell-free extracts suggests that calcium directly interacts with the telomerase complex. This interaction could be mediated by the calcium-binding protein S100A8 as indicated by its ability to mimic the inhibitory effect of calcium. S100A8-induced reduction in telomerase activity was abrogated by S100A9. The ratio of both proteins remained constant in cells treated with thapsigargin, but their interactions were altered similarly in intact cells after thapsigargin treatment and in cell-free extracts in response to calcium. We hypothesize that calcium binds to S100A8/S100A9 complexes and alters their composition, thus enabling S100A8 to interact with the telomerase complex and inhibit its activity.

The enzyme telomerase stabilizes the length of telomeres, the ends of linear chromosomes, by de novo addition of telomeric repeats, which have been lost due to incomplete DNA replication, nucleolytic degradation, and oxidative stress. The maintenance of telomeres and thus telomerase appear to play a critical role in senescence, aging, and cancer (1).
Human telomerase is composed of an RNA component, the telomerase RNA (hTR) 2 (2), and a catalytic subunit, the telom-erase reverse transcriptase (hTERT) (3). hTR acts as an anchor and template for telomeric DNA; hTERT catalyzes the addition of telomeric repeats to the 3Ј-ends of the chromosomes. Human telomerase also contains multiple accessory proteins, which bind either to hTR or hTERT, thus forming a large ribonucleoprotein complex with a molecular weight of more than 1000 kDa (4). Whereas hTR is widely expressed (2,5), expression of hTERT is more restricted and is closely linked to telomerase activity, thus representing its limiting factor (6,7).
Telomerase activity is observed in about 85% of all human tumors and during the last couple of years its inhibition has become an attractive tool for therapeutic intervention. For a long time it was believed that in the non-pathological situation telomerase is only functional during embryonic development and in the germ line of the adult organism, whereas normal somatic cells are telomerase-negative. Now it is generally accepted that also cells of highly proliferative and periodically or continuously renewing tissues, such as the hematopoietic system and the epidermis show telomerase activity (8,9). In epidermis telomerase activity is restricted to basal actively proliferating keratinocytes and is inhibited upon differentiation (9). We showed that this inhibition of telomerase is not a mere consequence of differentiation but rather a prerequisite for this process. In an organotypic co-culture model constitutive expression of hTERT in HaCaT cells did not abolish differentiation per se as indicated by histochemical and immunofluorescence analyses. However, whereas stratification and early stages of epidermal differentiation were unaffected, terminal differentiation was blocked (10).
The molecular mechanisms underlying this differentiationdependent decline in telomerase activity are still unknown. Even the regulation of epidermal differentiation itself is only poorly understood, but experimental evidence points to calcium as a key regulator of this process (11)(12)(13)(14)(15)(16)(17). Subsequent studies of our laboratory suggested that calcium could also play a functional role in the inhibition of telomerase activity in epidermal keratinocytes (18). We found that the inhibition of telomerase activity by differentiation-inducing factors, such as thapsigargin, correlated with their ability to increase intracellular calcium levels. Furthermore, calcium appeared to inhibit telomerase activity directly in a differentiation-independent manner in a cell-free system. However, the mechanism, through which calcium exhibited its inhibitory effect, remained unclear.
Calcium is an intracellular second messenger, which regulates a variety of biological processes. In most cases calcium does not act directly but uses calcium-binding proteins as mediators of its signals. Calcium binds to these calcium-binding proteins and induces conformational changes, thus exposing initially inaccessible protein binding sites and allowing their interaction with specific target proteins (19). Among the calcium-binding proteins are the S100 proteins S100A8 (calgranulin A; MRP8) and S100A9 (calgranulin B; MRP14).
The genes for S100A8 and S100A9 are co-localized and clustered with genes coding for various epidermal differentiation markers in the so-called epidermal differentiation complex (20 -22). S100A8 and S100A9 appear to be expressed in normal skin only at very low levels (23), but their expression is up-regulated during differentiation (24) and correlates in epithelial cell lines with their ability to terminally differentiate (25). For this reason they present excellent candidates for regulating epidermal differentiation and are potential mediators of the effect of calcium on telomerase activity.
The purpose of the present study was to further elucidate the functional role of calcium in differentiation-dependent inhibition of telomerase activity and to determine the function of S100A8 and S100A9 in this process. Advancing our understanding of the complex process of telomerase regulation during differentiation may help to develop a physiological strategy to eliminate telomerase in tumors, thus compromising their growth.

EXPERIMENTAL PROCEDURES
Cell Culture-The human keratinocyte line HaCaT was maintained in 4ϫ Eagle's minimal essential medium (MEM) supplemented with 10% fetal calf serum (FCS) and 100 units/ml penicillin/streptomycin (PS) at 37°C as described previously (26). HaCaT cells stably transfected with S100A8, S100A9, both, or vector alone were maintained in 4ϫ MEM/FCS/PS supplemented with neomycin (Geneticin; Invitrogen; 1.2 mg/ml) and/or puromycin (Sigma; 0.8 g/ml). Cells were routinely passaged every 10 days at a dilution of 1:10. Unless stated otherwise, for experiments 7.5 ϫ 10 5 cells per 10-cm plate were plated and grown to 40 -50% confluence. Where indicated, cells were treated with thapsigargin (Molecular Probes) dissolved in dimethyl sulfoxide (Sigma) and added to 4ϫ MEM/ FCS/PS to a final concentration of 1 M; controls were treated with equal amounts of dimethyl sulfoxide. After 30 min at 37°C, the medium was replaced by 4ϫ MEM/FCS/PS without additives. At various time points cells were harvested by incubation with 0.05% EDTA/0.025% trypsine, washed twice with phosphate buffered saline (PBS) and pelleted by centrifugation at 1000 ϫ g for 5 min at 4°C.
Isolation of Nuclear Protein-Crude nuclear protein extracts were prepared by the method of Dignam et al. (27). Cells were lysed in buffer A (10 mM HEPES, pH 7.9, 1.5 mM MgCl 2 , 10 mM KCl, 0.5 mM dithiothreitol) containing protease inhibitors (complete mini EDTA-free protease inhibitor mixture; Roche Diagnostics) and 0.5% Nonidet P-40 for 30 min at 4°C. Nuclei were pelleted by centrifugation at 4000 ϫ g for 30 min at 4°C and resuspended in buffer C (20 mM HEPES, 25% glycerol, 420 mM NaCl, 1.5 mM MgCl 2 , 0.5 mM dithiothreitol) containing protease inhibitors and incubated for 15 min at 4°C. The extracts were centrifuged at 10,000 ϫ g for 15 min at 4°C, and the supernatant was aliquoted and stored at Ϫ80°C. The protein concentration of the extracts was determined with Bio-Rad reagent according to the manufacturer's instructions.
Telomere Repeat Amplification Protocol (TRAP) Assay-Telomerase activity was detected by the PCR-based TRAP using the TRAPeze telomerase detection kit (Qbiogene) according to the manufacturer's instructions. Instead of cell lysates 5 ng of nuclear protein were used. Where indicated, the TRAP assay was performed in the presence of various concentrations of calcium chloride (50, 100, 250, and 500 M) or human recombinant S100A8 and/or S100A9, respectively, bovine serum albumin (Jackson ImmunoResearch Laboratories) (1, 2, 5, and 10 ng/l) in 10 mM Tris, pH 7.5, 150 mM NaCl, 1 mM dithiothreitol. The recombinant S100A8 and S100A9 were generously provided by Drs. Claus Kerkhoff and Clemens Sorg (University of Muenster, Muenster, Germany).
Western Blot Analysis-25 g (anti-hTERT) or 10 g (anti-S100A8 and anti-S100A9) nuclear protein were boiled for 5 min in SDS sample buffer (60 mM Tris, pH 6.8, 2% SDS, 5% 2-mercaptoethanol, 10% glycerol, 0.001% bromphenol blue) and subjected to SDS-PAGE. Samples were electrophoresed through a 3% stacking gel and resolved in a 6% (anti-hTERT) or 15% (anti-S100A8 and anti-S100A9) separating gel in 25 mM Tris, 190 mM glycine, and 0.1% SDS. The proteins were transferred to a nitrocellulose membrane (BA-S85; Schleicher & Schuell) by electroblotting at 16 volts at 4°C for 12-16 h using a transfer buffer containing 25 mM Tris, 190 mM glycine, and 20% methanol. The membrane was blocked in 10% low fat (1.5%) milk in PBST (PBS, 0.5% Tween 20) at room temperature for 1 h. Primary antibody was added at a dilution of 1:500 in milk/PBST. The membrane was incubated at room temperature for 1 h and then washed six times for 5 min with milk/PBST. Horseradish peroxidase-conjugated goat anti-mouse or anti-rabbit IgG was added at a dilution of 1:20,000 in milk/PBST. The membrane was incubated at room temperature for 1 h and then washed six times for 5 min with milk/PBST. Antigen-antibody complexes were detected using an ECL system (Amersham Biosciences) according to the manufacturer's instructions.
Transfection-The full-length coding sequences for human S100A8 and S100A9 including start and stop codons, as well as an additional NheI site at the 5Ј-end and a BamHI site at the 3Ј-end were amplified from the vectors pBUD CE 4.1-hMRP8 and pBUD CE 4.1-hMRP14 (kindly provided by SWITCH Biotech GmbH, Munich, Germany) and cloned into the mammalian expression vectors pIRESneo and pIRESpuro (Clontech). 20 g of plasmid were transfected into HaCaT keratinocytes using Lipofectamine 2000 (Invitrogen) according to the manufacturer's instructions. After selection with neomycin (Geneticin, 1.2 mg/ml) and/or puromycin (Sigma; 0.8 g/ml) clones were picked and analyzed for transgene expression by Northern blot analyses.
siRNA-ON-TARGETplus SMARTpool siRNAs specific for S100A8, S100A9, or non-target control siRNAs (Dharmacon) were introduced into HaCaT keratinocytes by transfection using DharmaFECT according to the manufacturer's instructions.
Cross-linking-Proteins were cross-linked with 1.5 mM bis-(sulfosuccimidmidyl)suberate (Pierce) for 1 h at room temperature. Subsequently, the reactions were quenched with 10 mM ethanolamine, 20 mM N-ethylmaleimide, and 50 mM sodium phosphate, pH 7.4, for 30 min at room temperature.

RESULTS
Recently we reported a differentiation-dependent inhibition of telomerase activity in human epidermis and in primary human skin keratinocytes (9,18). To investigate this phenomenon in more detail and to elucidate the underlying molecular mechanisms we induced differentiation in exponentially growing HaCaT cells by treating them for 30 min with 1 M thapsigargin. This factor raises intracellular calcium levels by inhibiting the Ca 2ϩ -ATPases of the endoplasmic reticulum (28,29). Control cells were treated with equal amounts of the solvent dimethyl sulfoxide. Induction of differentiation was monitored by immunofluorescence analyses for the early differentiation markers keratin 1/keratin 10 and involucrin and for the late differentiation markers filaggrin and transglutaminase, all of which were elevated in response to thapsigargin (data not shown). At various time points crude nuclear extracts were prepared and analyzed for their telomerase activity by TRAP assay. The results are shown in Fig. 1. Dimethyl sulfoxide-treated controls showed a basal telomerase activity, which was significantly reduced after thapsigargin treatment. Thapsigargin-induced inhibition of telomerase activity was already detectable at around day 1. The thapsigargin effect was most prominent at around day 3. Subsequently telomerase activity gradually increased until reaching control levels at around day 9 (data not shown).
To determine whether thapsigargin-mediated inhibition of telomerase activity is due to an inhibition of telomerase expression we first examined steady state protein levels of hTERT, the rate-limiting component of telomerase, in the HaCaT protein extracts, which had been analyzed by TRAP assay by Western blot analysis. Fig. 2A reveals no significant changes in the hTERT protein expression profile after thapsigargin treatment throughout the entire time course up to 6 days. These results indicate that thapsigargin-mediated inhibition of telomerase activity does not result from reduced hTERT expression and thus an insufficient number of telomerase complexes but suggest that calcium directly affects the complex. To confirm calcium as the mediator of the thapsigargin effect, we tested its ability to directly inhibit telomerase activity in a cell-free, gene expression-independent system. For this purpose we analyzed crude nuclear extracts of HaCaT cells for their telomerase activity in the presence of various amounts of calcium (50, 100, 250, and 500 M) by TRAP assay. As shown in Fig. 2B treatment with low concentrations of calcium had only a slight effect on telomerase activity, whereas 250 and 500 M calcium resulted  . At the indicated time points (1, 3, and 6 days) cells were harvested, crude nuclear extracts were prepared, and steady state hTERT protein levels were determined using an hTERT-specific antibody. Horseradish peroxidase-conjugated goat anti-mouse IgG was used as secondary antibody. Protein expression levels of fibrillarin as loading control were determined using an antibody specific for fibrillarin and horseradish peroxidase-conjugated goat anti-rabbit IgG as secondary antibody. Antigen-antibody complexes were visualized by chemiluminescence. B, calcium inhibits telomerase activity in cell-free extracts of HaCaT keratinocytes. TRAP assay of crude nuclear extracts (5 ng of protein) of HaCaT keratinocytes. The extracts were analyzed for their telomerase activity in the presence of 0, 50, 100, 250, and 500 M calcium chloride. S-IC, internal control. The results shown are representative of three independent experiments. in a significant inhibition of telomerase activity. To exclude the possibility that decreased telomerase activity in response to calcium was due to unspecific degradation of the telomerase complex by calcium-activated proteases or RNases present in the cell-free extract we repeated the experiments including the respective inhibitors. Neither treatment with protease inhibitors nor the presence of RNase inhibitors could abrogate the calcium effect (data not shown). Taken together these data indicate a direct role of calcium in inhibiting telomerase activity independent of gene expression.
Based on their differentiation-correlated expression pattern and their property to be modulated in their activities by calcium, S100A8 and S100A9 have established themselves as potential components of the intracellular calcium signaling pathway in epidermal differentiation (24). To test whether these calcium-binding proteins could be involved in mediating the inhibitory effect of calcium on telomerase activity we added various amounts of human recombinant S100A8 or S100A9 or bovine serum albumin as control to crude nuclear extracts of HaCaT cells and analyzed the effects on telomerase activity by TRAP assay. The results are presented in Fig. 3A. Addition of bovine serum albumin had no effect on telomerase activity, whereas S100A8 inhibited telomerase activity in a dosedependent manner. The S100A8-induced inhibition of telomerase activity was already detectable at a concentration of 1 ng/l and further increased at 2 and 5 ng/l. In contrast, S100A9 did not alter telomerase activity.
Since S100A8 and S100A9 are able to form heterodimers or higher order complexes with one another (30), we investigated whether the inhibitory effect of S100A8 could be influenced by S100A9 by adding 5 ng/l S100A8 together with various amounts of S100A9 (2, 5, and 10 ng/l). As presented in Fig.  3B, when S100A9 was added together with S100A8 at a con-centration of 2 ng/l the effect of S100A8 was not altered. The addition of equal (5 ng/l) or higher (10 ng/l) amounts of S100A9, however, resulted in an efficient abrogation of S100A8-mediated inhibition of telomerase activity. Similar results were obtained when the extracts were preincubated with S100A8 for 1 h before S100A9 was added (data not shown). These findings demonstrate that both S100A8 and S100A9 are involved in telomerase regulation and that S100A9 controls the effect of S100A8. When S100A8 is present in excess telomerase activity is inhibited. S100A9 counteracts this inhibition and when equal or higher amounts of S100A9 are available telomerase activity is not altered.
To test whether the inhibition of telomerase activity by S100A8, we found under cell-free conditions could also be observed in intact cells we generated HaCaT keratinocytes stably transfected with an expression construct for S100A8. Five S100A8 overexpressing HaCaT clones were analyzed for their telomerase activity (Fig.  4A). Surprisingly S100A8 overexpression did not reduce telomerase activity and telomerase activity did not correlate with S100A8 expression levels. Previous studies have reported that S100A8 and S100A9 are frequently co-expressed (31) and suggested that their expression is coordinately regulated (31,32). Thus, overexpression of S100A8 in HaCaT keratinocytes may lead to a simultaneous up-regulation of S100A9. To test this possibility we compared S100A8 and S100A9 protein levels in S100A8 transfectants. As shown in Fig. 4B S100A8 transfectants also had high levels of S100A9. We additionally transfected HaCaT cells with an expression construct for S100A9 and expression constructs for S100A8 and S100A9. In all clones tested we found a constant ratio of S100A8 and S100A9 (Fig.  4B). To obtain further proof for a coordinate regulation of S100A8 and S100A9 in HaCaT keratinocytes, we independently knocked down S100A8 and S100A9 with siRNAs specific for these proteins. Down-regulation of S100A8 protein correlated with down-regulation of S100A9 protein and vice versa (Fig. 5A) and had no effect on telomerase activity (Fig.  5B). These results clearly demonstrate a coordinately regulated expression of these two calcium-binding proteins. Furthermore, these data confirm that the effect of S100A8 on telomerase activity does not depend on the absolute level of S100A8 but its amount relative to S100A9. Telomerase activity is only reduced when S100A8 protein levels are higher than S100A9 protein levels.
To investigate whether thapsigargin-induced inhibition of telomerase activity could result from a shift in the ratio of S100A8 to S100A9 in favor of S100A8, we determined the steady state levels of these proteins in response to thapsigargin. . A, S100A8 but not S100A9 inhibits telomerase activity in cell-free extracts of HaCaT keratinocytes. TRAP assay of crude nuclear extracts (5 ng of protein) of HaCaT keratinocytes. The extracts were analyzed for their telomerase activity in the presence of 0, 1, 2, and 5 ng/l recombinant S100A8 or S100A9. Equal amounts of bovine serum albumin (BSA) were used as control. B, S100A9 abrogates S100A8-mediated inhibition of telomerase activity in cell-free extracts of HaCaT keratinocytes. TRAP assay of crude nuclear extracts (5 ng of protein) of HaCaT keratinocytes is shown. Crude nuclear extracts (5 ng of protein) with 5 ng/l recombinant S100A8 were analyzed for their telomerase activity in the presence of 0, 2, 5, and 10 ng/l recombinant S100A9 or equal amounts of bovine serum albumin. S-IC, internal control. The results shown are representative of three independent experiments.
We treated exponentially growing HaCaT cells for 30 min with 1 M thapsigargin or equal amounts of dimethyl sulfoxide. At various time points crude nuclear extracts were prepared and analyzed by Western blot. As shown in Fig. 6 thapsigargin does not significantly alter the protein levels of S100A8 and S100A9. Thapsigargin does neither increase S100A8 expression nor decrease S100A9 expression to shift the ratio of S100A8 to S100A9 in favor of S100A8.
When present at equal or higher amounts S100A9 prevents S100A8 from reducing telomerase activity. This occurs most likely by sequestering S100A8 in heterodimeric or higher order complexes. Coordinate up-or down-regulation of S100A8 and S100A9 should lead to similar changes in the amount of these complexes. To verify this idea we analyzed S100A8/S100A9 complexes in S100A8/S100A9-overexpressing HaCaT clones and in HaCaT cells, in which S100A8/S100A9 expression had been knocked-down by siRNA. For this purpose we covalently cross-linked non-covalently associated proteins by treatment with bis-(sulfosuccimidmidyl)suberate for 1 h at room temperature. S100A8 and S100A9 complexes were then detected by Western blot analysis. Fig. 7A shows cross-linked versus noncross-linked crude nuclear extracts of exponentially growing HaCaT cells revealing that neither S100A8 nor S100A9 exist as free monomers but form complexes with each other and potentially other proteins. Four major complexes with apparent molecular weights of 50, 45, 32, and 25 kDa were detected. In S100A8/S100A9 overexpressing HaCaT clones the amount of all four complexes was increased and correlated with the expression levels of S100A8 and S100A9 (Fig.  7B). Knock-down of S100A8 or S100A9, on the other hand, resulted in coordinate down-regulation of these complexes (Fig. 7C). Hypothetically, calcium could interact with these complexes and alter their composition. These alterations could lead to the release of S100A8 and allow its interaction with telomerase. In this scenario thapsigargin would keep the overall ratio of S100A8 to S100A9 constant but increase the ratio of free, telomerase-inhibiting S100A8. To test this hypothesis we analyzed the effect of thapsigargin on S100A8/S100A9 complexes. As shown in Fig. 7D thapsigargin did not induce the release of S100A8 but altered its interaction with S100A9. Thapsigargin promoted the formation of the 45-kDa complex and reduced the amount of the two smaller complexes of 32 and 25 kDa. Similar alterations in complex formation were observed in cell-free extracts in response to calcium. These results suggest calcium as mediator of the thapsigargin effect and demonstrate that calcium directly affects S100A8/S100A9 complexes altering their interaction with one another and potentially their target proteins including telomerase.

DISCUSSION
In human epidermis telomerase is only active in basal proliferating keratinocytes. Inhibition of telomerase activity occurs in suprabasal keratinocytes with the onset of differentiation (9). This inhibition is not a mere consequence of differentiation but rather a prerequisite for proper termination of this process (10). Constitutive expression of telomerase allows stratification, although with some morphological alterations, and while early steps of differentiation proceed normally, terminal differentiation is prevented.
The differentiation-dependent decline in telomerase activity is also detectable in cell culture in primary skin keratinocytes (18). In the present study we show that telomerase activity was similarly reduced in the human epidermal keratinocyte line HaCaT when differentiation was induced with thapsigargin. This finding further supports the hypothesis that inhibition of telomerase activity represents a general phenomenon of epidermal differentiation. In HaCaT cells thapsigargin inhibits telomerase activity in a time-dependent manner, already being detectable after 1 day of treatment, reaching a maximum after ϳ3 days and then gradually increasing until reaching control levels after ϳ9 days. This time course of telomerase activity FIGURE 4. A, S100A8 overexpression in HaCaT cells does not inhibit telomerase activity. Upper panel, TRAP assay of crude nuclear extracts (5 ng of protein) of clones of HaCaT keratinocytes stably transfected with an expression construct for S100A8 or vector alone. Exponentially growing cells were harvested, and crude nuclear extracts were prepared and analyzed for their telomerase activity by TRAP assay. S-IC, internal control. Lower panel, corresponding Western blot analysis (10 g of protein). S100A8 protein expression levels were determined using an antibody specific for S100A8. Horseradish peroxidase-conjugated goat anti-mouse IgG was used as secondary antibody. Protein expression levels of fibrillarin as loading control were determined using an antibody specific for fibrillarin and horseradish peroxidase-conjugated goat anti-rabbit IgG as secondary antibody. Antigen-antibody complexes were visualized by chemiluminescence. B, S100A8 overexpression leads to coordinate overexpression of S100A9 and vice versa. Western blot analysis of S100A8 and/or S100A9 overexpressing HaCaT clones. S100A8, S100A9, and fibrillarin protein expression levels were determined using antibodies specific for S100A8, S100A9, or fibrillarin as in A. The results shown are representative of three independent experiments.
reflects very well the respective stage of differentiation of the cells. Differentiation is induced in a certain portion of the cells, and these cells are subsequently lost and replaced by proliferating cells.
Thapsigargin increases intracellular calcium levels by promoting the release of this factor from internal stores by inhibition of their Ca 2ϩ -ATPases (28,29). An increase in intracellular calcium levels also occurs during differentiation with epidermis displaying a calcium gradient increasing with differentiation (13,14,17,33). This finding together with the fact that most of the differentiation-specific proteins are regulated by calcium (11,12,15,16) point to calcium as a key regulator of differentiation. In cell culture keratinocytes retain their proliferative activity at calcium concentrations below 0.1 mM. A switch to higher calcium levels suppresses proliferation and induces the expression of differentiation-specific keratins and other markers of differentiation, such as involucrin (11,12). Consistent with this idea in vitro induction of differentiation and expression of differentiation-specific proteins is strongly linked with the ability of the inducers to increase intracellular calcium (34,35). On the other hand, down-regulation of intracellular calcium blocks differentiation and prevents the expression of differentiation markers both in vitro and in vivo (36). The finding that in skin diseases, which are characterized by aberrant differentiation, such as psoriasis, the epidermal calcium gradient is perturbed further supports the role of calcium in epidermal differentiation (17).
Confirming our previous hypothesis we now demonstrated that thapsigargin-induced inhibition of telomerase activity in HaCaT cells was not the consequence of reduced telomerase expression. Furthermore we showed that calcium was able to inhibit telomerase activity in cell-free HaCaT extracts indicating that it does not require gene expression but directly interacts with the telomerase complex. We present for the first time experimental evidence that this interaction is mediated by the S100 calcium-binding protein S100A8. In this context it is noteworthy to mention that the concentration of calcium required to inhibit telomerase activity in cell-free extracts exceeds the calcium levels reported for intact cells and epidermis (13,14,17,33). This phenomenon could be explained by the fact that in solution S100 proteins have surprisingly low calcium affinities well below physiological ionic conditions ranging from one to several hundred M (38 -40). There is a consensus that in vivo calcium affinities are increased to physiological levels by the presence of target proteins or by interaction with other cations (41)(42)(43).
In agreement with Saintigny et al. (25) and Thorey et al. (44) we show that HaCaT keratinocytes similar to several epithelial cell lines, reconstructed epidermis, as well as epidermis in situ (23) express S100A8 and S100A9. Consistent with various reports (30,(45)(46)(47), which show that in human cells the monomers of these proteins are highly unstable and need to be com- FIGURE 5. A, knock-down of S100A8 protein in HaCaT cells also knocks down S100A9 protein and vice versa. Western blot analysis (10 g of protein) of HaCaT keratinocytes transfected with SMARTpool siRNAs specific for S100A8, S100A9, or non-target control siRNAs. At the indicated time points (4 and 5 days after transfection) cells were harvested, crude nuclear extracts were prepared, and steady state protein levels of S100A8 and S100A9 were determined using antibodies specific for S100A8 or S100A9. Horseradish peroxidase-conjugated goat anti-mouse IgG was used as secondary antibody. Protein expression levels of fibrillarin as loading control were determined using an antibody specific for fibrillarin and horseradish peroxidase-conjugated goat anti-rabbit IgG as secondary antibody. Antigen-antibody complexes were visualized by chemiluminescence. B, knock-down of S100A8 or S100A9 protein does not alter telomerase activity. Corresponding TRAP assay of crude nuclear extracts (5 ng of protein). S-IC, internal control. The results shown are representative of three independent experiments. FIGURE 6. Thapsigargin does not alter the ratio of S100A8 to S100A9 in HaCaT keratinocytes. Western blot analysis of crude nuclear extracts (10 g of protein) of HaCaT keratinocytes. Cells were treated for 30 min with 1 M thapsigargin or equal amounts of the solvent dimethyl sulfoxide (DMSO). At the indicated time points (1, 3, and 6 days) cells were harvested, crude nuclear extracts were prepared, and steady state protein levels of S100A8 and S100A9 were determined with antibodies specific for S100A8 or S100A9. Horseradish peroxidase-conjugated goat anti-mouse IgG was used as secondary antibody. Protein expression levels of fibrillarin as loading control were determined using an antibody specific for fibrillarin, and horseradish peroxidaseconjugated goat anti-rabbit IgG as secondary antibody. Antigen-antibody complexes were visualized by chemiluminescence. The results shown are representative of three independent experiments. MARCH 2, 2007 • VOLUME 282 • NUMBER 9

S100A8 Mediates Inhibition of Telomerase Activity
plexed with each other, we found that in HaCaT cells both S100A8 and S100A9 did not exist in their free forms. By Western blot analysis of cross-linked protein extracts using S100A8-and S100A9-specific antibodies we could detect four major complexes with apparent molecular weights of 50, 45, 32, and 25 kDa. Similar complexes have been detected in human mono- FIGURE 7. A, complex formation of S100A8 and S100A9 in HaCaT keratinocytes. Western blot analysis was performed with crude nuclear extracts (10 g of protein) isolated from exponentially growing HaCaT cells. Where indicated, extracts were cross-linked by 1.5 mM bis-(sulfosuccimidmidyl)suberate for 1 h at room temperature. S100A8-and S100A9-containing complexes and monomeric S100A8 and S100A9 were detected with antibodies specific for S100A8 or S100A9. Horseradish peroxidase-conjugated goat anti-mouse IgG was used as secondary antibody. Antigen-antibody complexes were visualized by chemiluminescence. B, increased formation of S100A8/S100A9 complexes in S100A8/S100A9-overexpressing HaCaT clones. Western blot analysis of exponentially growing S100A8/S100A9-overexpressing HaCaT clones as in A. Protein expression levels of fibrillarin as loading control were determined using an antibody specific for fibrillarin and horseradish peroxidase-conjugated goat anti-rabbit IgG as secondary antibody. C, decreased formation of S100A8/S100A9 complexes in HaCaT keratinocytes transfected with siRNAs specific for S100A8 or S100A9. HaCaT cells were transfected with SMARTpool siRNAs specific for S100A8, S100A9, or non-target control siRNAs. At the indicated time points (4 and 5 days after transfection) cells were harvested, and Western blot analysis was performed as in A. D, similar changes in S100A8/S100A9 complexes in response to thapsigargin in HaCaT keratinocytes and calcium in cell-free extracts of HaCaT keratinocytes. HaCaT cells were treated with 1 M thapsigargin or dimethyl sulfoxide (DMSO) for 30 min. After 3 days cells were harvested, and Western blot analysis was performed as described for A. Where indicated, the cross-linking occurred in the presence of 250 M calcium chloride. The results shown are representative of three independent experiments. cytes and granulocytes and have been identified as S100A8 2 / S100A9 2 tetramer (48.5 kDa), S100A8 2 /S100A9 trimer (35 kDa), and S100A8/S100A9 dimer (25 kDa) (31). Based on that information, the 45-kDa complex in HaCaT cells could represent a S100A8/S100A9 2 trimer.
As in monocytes and granulocytes (31) formation of these complexes was calcium-dependent also in HaCaT cells. Increases in calcium concentrations either by thapsigargin in intact cells or by direct addition of calcium to cell-free extracts significantly reduced the association of the two low molecular weight complexes, whereas the amount of the 45-kDa complex was increased. Since the 45-kDa complex contains less S100A8 than the 35-and 25-kDa complex, the amount of S100A8, which is sequestered by S100A9, is decreased. Consequently S100A8 could be released and enabled to interact with telomerase, thus reducing its activity. This idea is supported by our observation that recombinant S100A8 when added to cell-free HaCaT extracts inhibits telomerase activity. When added at equimolar or higher amounts S100A9 abrogates S100A8-induced inhibition of telomerase suggesting a competition for S100A8 binding between telomerase and S100A9 in favor of the latter. Taken together these results suggest that telomerase activity is only inhibited by S100A8 when the amount of this protein is higher than the amount of S100A9. Consequently changes in telomerase activity only occur when the ratio of S100A8 to S100A9 is altered.
The inhibitory effect of S100A8 observed in cell-free extracts could not be reproduced by overexpression of this protein in HaCaT cells. S100A8 overexpressing HaCaT cells did not vary significantly in their telomerase activity and showed, independent of their S100A8 expression level, similar telomerase activity as control cells. Consistent with various reports that S100A8 and S100A9 are co-expressed (31) and that in response to a variety of extracellular signals their expression is altered comparably (32), our expression analyses revealed coordinate changes in S100A8-and S100A9-overexpressing HaCaT variants. S100A8 clones showed a similarly high level of S100A9 and S100A9 overexpression resulted in an equivalently elevated expression of S100A8. This increased expression of S100A8 and S100A9 increased the amount but did not change the pattern of the S100A8/S100A9 complexes. Comparable results were obtained when we knocked down S100A8 and S100A9, respectively. Decreased expression of S100A8 correlated with similarly reduced S100A9 expression, and knock-down of S100A9 also resulted in lower levels of S100A8 protein. The amount of all four S100A8/S100A9 complexes was decreased, but the multimer pattern was preserved. Thus, not only the overall ratio of monomeric S100A8 to S100A9 remained stable but also the amount of S100A8, which was sequestered in S100A8/S100A9 complexes. Consequently no "free" S100A8 was available to interact with telomerase and therefore no effect on telomerase activity was observed. Telomerase activity was only reduced when free S100A8 was provided either by addition of excess S100A8 (Fig. 3) or by calcium-induced release of S100A8 from S100A8/S100A9 complexes (Fig. 7D).
We were not able to detect complexes between S100A8 and hTERT by Western blot analysis. This could be explained by the fact that despite relatively high amounts of hTERT protein the number of telomerase complexes in a cell is very low and most probably under the detection level of Western blot analysis. Most interaction partners of telomerase have been identified by their ability to immunoprecipitate telomerase activity using the highly sensitive TRAP assay. However, this approach requires active telomerase complexes and cannot be applied when the interaction partner significantly inhibits telomerase activity as in the case of S100A8. Accordingly, we failed to precipitate telomerase activity above unspecific background levels using S100A8-specific antibodies.
A substantial amount of experimental evidence implicates S100A8 and S100A9 in epidermal differentiation (24). Their exact function in this process, however, remained unclear and was restricted to the idea that they may be involved in the reorganization of the keratin cytoskeleton (45,48). The observation that S100A8 and S100A9 are also up-regulated during wound healing (44) led to the speculation that these proteins regulate a switch between proliferation and differentiation. Our results support this notion and, moreover present, a candidate through which this switch could be accomplished, namely telomerase. Telomerase is active during proliferation and needs to be inhibited in order for differentiation to occur. The S100A8-mediated inhibition of telomerase based on a direct interaction between these proteins represents an elegant transient mechanism to allow keratinocytes to switch from differentiation back to proliferation as required, e.g. during wound healing. Wounding is FIGURE 8. Model for calcium-mediated inhibition of telomerase activity during differentiation of human epidermal keratinocytes. Increases in intracellular calcium concentrations lead to an interaction between calcium and S100A8/S100A9 complexes. This interaction induces conformational changes of these complexes resulting in the temporary release of S100A8. S100A8 immediately binds to telomerase, thus reducing its activity.
accompanied by a transient decrease in intracellular calcium levels (49). This may lead to the release of S100A8 from telomerase, reactivation of telomerase, and subsequently to cell proliferation. A role for telomerase in wound healing is also suggested by data from transgenic or knock-out mouse models (50,51). Whereas constitutive high levels of telomerase promote reepithelialization, decreased telomerase activity results in delayed wound closure. A similar regulatory mechanism could be operational in psoriasis, where a perturbation of intracellular calcium concentrations (17) and deregulation of S100A8 and S100A9 (52) correlate with changes in telomerase activity (37) and impaired differentiation.
In conclusion, the results presented here suggest S100A8 as the mediator of differentiation-dependent and calciummediated inhibition of telomerase activity in epidermal keratinocytes and thus as a new regulator of telomerase. Based on the experimental evidence we provide in this study, we propose the following working model (Fig. 8): in epidermal keratinocytes all S100A8 is complexed with S100A9 and/or its target proteins. Upon differentiation intracellular calcium levels increase, and calcium interacts with these complexes. This interaction leads to conformational changes of the complexes eventually resulting in the temporary release of S100A8. This enables S100A8 to bind to telomerase and decrease its activity. Interference with S100A8-mediated inhibition of telomerase activity may thus represent an alternative and interesting approach to eliminate telomerase in tumor cells.