The Activation of the c-Jun N-terminal Kinase and p38 Mitogen-activated Protein Kinase Signaling Pathways Protects HeLa Cells from Apoptosis Following Photodynamic Therapy with Hypericin*

In this study, we elucidate signaling pathways induced by photodynamic therapy (PDT) with hypericin. We show that PDT rapidly activates JNK1 while irreversibly inhibiting ERK2 in several cancer cell lines. In HeLa cells, sustained PDT-induced JNK1 and p38 mitogen-activated protein kinase (MAPK) activations overlap the activation of a DEVD-directed caspase activity, poly(ADP-ribose) polymerase (PARP) cleavage, and the onset of apoptosis. The caspase inhibitors benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone (zVAD-fmk) and benzyloxycarbonyl-Asp-Glu-Val-Asp-fluoromethylketone (zDEVD-fmk) protect cells against apoptosis and inhibit DEVD-specific caspase activity and PARP cleavage without affecting JNK1 and p38 MAPK activations. Conversely, stable overexpression of CrmA, the serpin-like inhibitor of caspase-1 and caspase-8, has no effect on PDT-induced PARP cleavage, apoptosis, or JNK1/p38 activations. Cell transfection with the dominant negative inhibitors of the c-Jun N-terminal kinase (JNK) pathway, SEK-AL and TAM-67, or pretreatment with the p38 MAPK inhibitor PD169316 enhances PDT-induced apoptosis. A similar increase in PDT-induced apoptosis was observed by expression of the dual specificity phosphatase MKP-1. The simultaneous inhibition of both stress kinases by pretreating cells with PD169316 after transfection with either TAM-67 or SEK-AL produces a more pronounced sensitizing effect. Cell pretreatment with the p38 inhibitor PD169316 causes faster kinetics of DEVD-caspase activation and PARP cleavage and strongly oversensitizes the cells to apoptosis following PDT. These observations indicate that the JNK1 and p38 MAPK pathways play an important role in cellular resistance against PDT-induced apoptosis with hypericin.


MAPK pathways play an important role in cellular resistance against PDT-induced apoptosis with hypericin.
Mitogen-activated protein kinases (MAPKs) 1 are proline-directed Ser/Thr protein kinases activated by dual phosphorylation on both a tyrosine and a threonine residue (1). These enzymes are critical components of a complex intracellular signaling network that ultimately regulates gene expression in response to a variety of extracellular stimuli. The three well known mammalian MAPK families are: the extracellular signal-regulated kinases (ERKs), the c-Jun N-terminal kinases/ stress-activated protein kinases (JNKs/SAPKs), and the p38 MAPK. Each of these enzymes is a target for discrete but closely related phosphorylation cascades in which the sequential activation of three kinases constitutes a common signaling module (2). The best characterized MAPK pathway is the Ras/ Raf/MEK cascade leading to the activation of ERK1/2 in response to growth factors (3). JNK and p38 MAPK are key mediators of stress signals and inflammatory responses evoked by a variety of agents such as UV-and ␥-irradiation, heat shock, osmotic stress, and inflammatory cytokines (4,5). JNKs are activated by the dual specificity kinases, MKK4/SEK1 (6) and MKK7 (7), while p38 MAPKs are activated by the MKK3/6 homologues (8). Once activated, JNKs mediate the phosphorylation and activation of the transcription factors c-Jun, ATF2, and Elk1 (9,10). The p38 MAPK cascade is likewise involved in the transcriptional regulation of ATF2 and Elk1 (11) as well as in the activation of MAPKAP2/3 kinases, which in turn phosphorylate small heat shock proteins (12,13).
Because of the strong activation of JNK and p38 MAPK observed in cells treated with several stress signals that ultimately lead to apoptosis, considerable attention has recently been focused on the potential role of these kinases in apoptotic signaling. A causative link between the JNK/p38 MAPK signaling pathways and apoptosis has been suggested by several 1 The abbreviations used are: MAPK, mitogen-activated protein kinase; ERK, extracellular signal-regulated kinase; JNK/SAPK, c-Jun N-terminal kinase/stress-activated protein kinase; MKK, MAPK kinase; SEK1, SAPK/ ERK kinase; MKP-1, MAPK phosphatase-1; MAPKAP, MAPK-activated protein; MEKK, MAPK/ERK kinase kinase; zVAD-fmk, benzyloxycarbonyl-Val-Ala-Asp-fluoromethylketone; zDEVD-fmk, benzyloxycarbonyl-Asp-Glu-Val-Asp-fluoromethylketone; amc, amino-4-methyl-coumarin; CrmA, cytokine response modifier A; PARP, poly(ADP-ribose) polymerase; TNF, tumor necrosis factor-␣; MOPS, 4-morpholinepropanesulfonic acid; PDT, photodynamic therapy; GST, glutathione S-transferase; X-gal, 5-bromo-4-chloro-3indolyl-␤-D-galactopyranoside; FCS, fetal calf serum; EGF, epidermal growth factor. studies. In some reports, it was shown that overexpression of constitutively active forms of MEKKs (14), the upstream regulators of MKK4/SEK1, or MKK6 (15) results in apoptosis. Xia et al. (16) reported that in PC12 cells, JNK and p38 MAPK play a critical role in mediating apoptosis caused by nerve growth factor withdrawal. The JNK pathway also seems to be required for the induction of apoptosis by ceramide, ␥and UVC-irradiation, some chemotherapeutic drugs (17)(18)(19), as well as for the Daxx-mediated, Fas-induced cell death (20). Inversely, other reports have suggested that the activation of JNK is not mechanistically implicated in the apoptotic process. Inhibition of the JNK pathway by the expression of the dominant negative forms of MEKK1, SEK1, or c-Jun mutant did not prevent Fasor TNF-mediated cell death (21,22). Similarly, a recent study has shown that the activation of JNK/p38 MAPK does not correlate with apoptosis induced by the detachment of epithelial cells (23). Furthermore, thymocytes from sek1 Ϫ/Ϫ mice were found to be more susceptible to apoptosis induced by Fas and CD3 than their wild type counterparts (24), suggesting that the JNK pathway may also have a protective function. In conclusion, the exact role of JNK and/or p38 MAPK in programmed cell death remains highly ambiguous. Moreover, these reports stress the dependence of specific cellular responses (death or survival) on the cellular background as well as on the causative agent.
Photodynamic therapy (PDT) is a new cancer treatment based on the topic or systemic application of a photosensitizing agent that accumulates in hyperproliferating benign and malignant tissues (reviewed in Ref. 25). Upon irradiation with tissue-penetrating light, the photosensitizer is activated and generates reactive oxygen species, which then cause cell death. PDT is a promising alternative to conventional cancer treatments involving cytotoxic drugs or ionizing radiation, as it combines a minimal systemic toxicity with a highly selective photodynamic destruction of tumor cells. However, the molecular mechanism underlying its antitumor activity is poorly understood. PDT with porphyrin and a porphyrin derivative has recently been reported to induce both c-Jun and c-Fos expression (26) and activation of the JNK/p38 MAPK signaling pathways (27,28), but a causative link between these observations and the PDT-induced cytotoxicity is still missing.
In the present study, we have used the photodynamic agent hypericin as a stress stimulus and investigated its potential signaling to the different MAPK cascades. Hypericin is a photoactive natural pigment extracted from plants of the genus Hypericum with a phenanthroperylenequinone structure. Interest in this compound was renewed in recent years because of its potential as a photosensitizing anticancer drug. We (29 -31) and others (32,33) have shown recently that photo-activated hypericin has a powerful in vivo antitumor activity. Although hypericin has also been reported to cause apoptosis in various cancer cell lines (34,35), the signal transduction pathways involved have never been investigated.
Here we show that PDT with hypericin leads to a strong and sustained activation of both JNK1 and p38 MAPKs, while causing a drastic and irreversible inhibition of ERK2 in all cancer cell lines tested. Specifically in HeLa cells, we show that hypericin-induced apoptosis requires a DEVD-directed caspase activity, while the activation of the JNK/p38 MAPK pathways occurs via a caspase-independent mechanism and is not related to the process of apoptosis. Furthermore, by expression of dominant negative mutants of components of these stress-induced kinase signaling pathways, or by the use of the p38 MAPK inhibitor PD169316, we show that concomitant inhibition of the JNK and p38 MAPK significantly enhances cell death. These observations indicate that the JNK/p38 MAPK path-ways contribute to a survival response that counteracts PDTinduced apoptosis in HeLa cells.
Cell Culture and PDT Treatment-All cell lines used in these study were maintained in Dulbecco's modified Eagle's medium supplemented with 2 mM L-glutamine, 1% penicillin/streptomycin solution, and 10% FCS (Life Technologies Inc., Paisley, Scotland). Cells were incubated at 37°C in 5% CO 2 . Preparation and storage of hypericin and cell photosensitization with hypericin were performed as described elsewhere (29). Briefly, cells were preincubated for 16 h with different concentrations of hypericin in culture medium containing 10% FCS in strictly subdued light conditions (Ͻ1 microwatt/cm 2 ). Then cells were irradiated in hypericin-free medium for 15 min by placing the plates on a plastic diffuser sheet 5 cm above a set of seven L18W/30 fluorescent lamps (Osram, Berlin, Germany). At the surface of the diffuser, the uniform fluence rate was 4.5 milliwatts/cm 2 (corresponding to a light dose of 4 J/cm 2 ). In the experiments where EGF was used, the cells were first starved for 48 h in serum-free medium. Preparation of cell extracts at the end of incubation periods was carried out exactly as described before (36). Protein concentrations were estimated with BCA (Pierce).
Protein Kinase Assays-JNK1 and ERK2 activities were measured by immunocomplex kinase assays as described in Ref. 36, with the exception that the kinase buffer used here was 20 mM MOPS, pH 7.4, 15 mM MgCl 2 , 2 mM EGTA, 1 mM 1 dithiothreitol, 0.1% Triton X-100, 1 mM Na 3 VO 4 . The p38 MAPK activation was determined by Western blot analysis using antiphospho-p38 MAPK antibody.
Evaluation of Apoptosis and Caspase Activity-Following treatment as described above, cells were incubated in the dark for 24 h, and apoptosis was evaluated by fluorescent microscopic analysis of fragmented nuclei stained with Hoechst 33342. Caspase activity assays were performed as described (37).
Western Blot Analysis-Samples (20 -40 g of protein) from cell lysates were separated by SDS-polyacrylamide gel electrophoresis and electrophoretically transferred to nitrocellulose or polyvinylidene difluoride membranes (Bio-Rad). The membranes were blocked in Blotto (5% non-fat dry milk in TBST (50 mM Tris, pH 7.4, 150 mM NaCl, 0.1% Tween 20)) for 1 h at room temperature and incubated with primary antibody for 2 h at room temperature (anti-PARP) or overnight at 4°C (antiphospho-p38). The membranes were then washed in TBST and incubated for 1 h with horseradish peroxidase-conjugated secondary antibodies. The specific signals were detected using an enhanced chemiluminescence detection system (Amersham Pharmacia Biotech).
Transient Transfection Assay and X-Gal Staining-The construction of mammalian expression vector containing the c-Jun transcriptional activation mutant TAM-67 was done as described in Ref. 38. The pcDNA3 expression vector containing the cDNA insert encoding an enzymatically inactive form of SEK1 (SEK-AL) was a kind gift of Dr. J. R. Woodgett (Ontario Cancer Institute, Canada). MKP-1 plasmid was kindly provided by Dr. N. Tonks (Cold Spring Harbor, NY).
Cells were seeded at 1.5 ϫ 10 5 cells/well in six-well plates. After an overnight culture, 4 g of SEK-AL, TAM-67, or MKP-1 expression vector and 1 g of a ␤-galactosidase expression vector (pUT651) were co-transfected using the FuGene6 transfection reagent (Boehringer Mannheim) according to the manufacturer's protocol. After 20 h of transient transfection, cells were rinsed twice with phosphate-buffered saline and then incubated in complete medium with or without 125 nM hypericin for 16 h at 37°C in the dark. Following this period of recovery, the cells were irradiated as described above. In order to assess the role of p38 MAPK, transfected cells were pretreated with PD169316 1 h before irradiation. Cells were harvested 10 h post-irradiation, washed, and fixed in 1% (v/v) formaldehyde solution in phosphate-buffered sa-line for 10 min at room temperature. To identify ␤-galactosidase enzyme activity, the fixed cells were washed twice with phosphate-buffered saline and stained in buffer containing X-gal (1 mg/ml X-gal in 10 mM Na 3 PO 4 , pH 7, 5 mM K 3 Fe(CN) 6 , 5 mM K 4 Fe(CN) 6 ⅐3H 2 O, 2 mM MgCl 2 , 0.02% Nonidet P-40, and 0.01% SDS) for 2-6 h at 37°C. The blue-colored ␤-galactosidase-expressing cells were examined with a light microscope. Cell survival was determined by calculating the percentage of morphologically apoptotic blue cells in the total number of blue cells in at least ten randomly selected fields.

Effect of Hypericin on Different MAPK Signaling Pathways-
We initially investigated whether one or more members of the MAPK family were activated following cell photosensitization with hypericin (PDT). In order to assess the general validity of the hypericin-induced effects on MAPK activities, different human cancer cell lines (A431, HaCaT, and HeLa cells) or murine L929 cells were incubated for 16 h with concentrations of hypericin that, after exposure to a light dose of 4 J/cm 2 , result in 20% cytotoxicity as determined by the neutral red assay (29). After PDT, cells were further incubated in the dark and harvested after the specified period of time. Fig. 1 shows that in all cell lines tested, sublethal doses of PDT with hypericin caused a rapid and persistent activation of JNK1 while severely inhibiting the basal levels of ERK2 activity. Hypericin without photo-activation had no effect on the activity of either kinase (Fig. 1, lane 2 in all panels). These observations strongly suggest that the changes in the activities of JNK1 and ERK2 are typical light-dependent, hypericin-mediated responses that are not restricted to a particular cell line. As shown for HeLa cells in Fig. 1, no difference in the JNK1 or ERK2 protein level was observed over the time period examined, indicating that the kinase activity changes reflect modi-fications of pre-existing molecules.
To test whether the inhibition of ERK2 by PDT could be counteracted by the addition of inducing agents of the ERK2 pathway, we evaluated the effect of EGF and FCS on PDTinduced ERK2 inhibition. We observed that in addition to inhibiting the basal cellular levels of ERK2 activity, photo-activated hypericin blocked the EGF-mediated ERK2 activation (Fig. 2). The inhibition of the ERK2 pathway was irreversible, as the addition of EGF either 5 min before (lane 5) or immediately after (lane 6) light exposure could not prevent or reverse the ERK2 inhibition. Similar results were obtained when the signal to ERK2 activation was initiated by the addition of FCS to the cell culture (data not shown). Because a sustained activation of JNKs, with a concomitant inhibition of ERKs, has been reported to occur during the induction of apoptosis (16), we looked next for a correlation between the hypericin-induced effects on MAPK activities and cell death.
PDT with Hypericin Induces Sustained JNK1 and p38 MAPK Activations and Leads to Caspase-mediated PARP Fragmentation and Apoptosis in HeLa Cells-In a separate study undertaken to characterize the type of cell death induced by photo-activated hypericin (39), we have shown that this photosensitizer can induce either apoptosis or necrosis in HeLa cells, depending on the photodynamic conditions used. A 16-h preincubation with 125-250 nM hypericin followed by irradiation with a light dose of 4 J/cm 2 was found to efficiently and specifically trigger the apoptotic program in this cell line.
Here we show that upon Hoechst 33342 staining and fluorescence microscopic analysis, the nuclei of cells treated with photo-activated hypericin appeared highly condensed and fragmented in contrast to the homogeneously stained nuclei of untreated cells (Fig. 3A). Cells treated with hypericin but not irradiated were indistinguishable from untreated control cells (data not shown).
To further characterize the signaling pathways involved in the hypericin-induced cellular responses, we focused on HeLa cells as a model system. As shown in Fig. 3B, treatment of the cells with 125 nM hypericin and a light dose of 4 J/cm 2 caused a sustained activation of both JNK1 and p38 MAPK over a 24-h time period. While the p38 MAPK activity showed a steady increase after 1 h post-irradiation and remained sustained during the time period examined, the activation of JNK1 was also sustained and somehow biphasic. These JNK1 and p38 MAPK activity changes resulted from post-translational modifications, since no difference in protein levels was detected by Western blot analysis over the 24-h time course (data not shown).
Caspases, a group of cysteine proteases that cleave protein substrates after aspartic acids, play a central role in the regulation and execution of the apoptotic program (reviewed in Ref. 40), and the cleavage of poly(ADP-ribose) polymerase (PARP) by activated caspase-3 subfamily members is considered to be one of the hallmarks of apoptosis (41). We therefore examined whether photo-activated hypericin could trigger this event in HeLa cells. As shown in Fig. 3C, lysates of PDT-treated cells contained the characteristic 85 kDa PARP fragment, an indication of the activation of the effector caspase-3 and/or caspase-7 (42) and apoptotic cell death. The apoptotic PARP fragmentation coincided well with the activation of a DEVDdirected caspase activity (Fig. 3D). While the initial PDT-mediated induction of the stress kinase pathways clearly preceded the apoptotic events (Fig. 3B), the sustained activation of p38 MAPK and the late phase of JNK1 activity continued in parallel with the kinetics of PARP cleavage and the DEVD-amc proteolytic activity (Fig. 3, C and D). This observation may suggest a role for these kinases in the hypericin-induced apoptosis.
Impact of Caspase Inhibitors on the Hypericin-induced JNK1 and p38 MAPK Activation and Apoptosis-Because a functional cross-talk between the JNK/p38 MAPK and the caspase signaling pathway has been reported (43)(44)(45)(46), we tested the effect of different caspase inhibitors on hypericin-induced JNK1 and p38 MAPK activations and apoptosis. HeLa cells, incubated for 16 h with 125 nM hypericin, were pretreated for 2 h before irradiation with the caspase-3-directed inhibitor zDEVD-fmk. As shown in Fig. 4A (lower panel), preincubation with zDEVD-fmk did not notably affect the PDT-induced JNK1 or p38 MAPK activations. This caspase inhibitor did not inhibit the JNK1/p38 MAPK activation during the entire time course examined (data not shown). These observations indicate that zDEVD-fmk-inhibitable caspases are not required for the activation of the JNK/p38 MAPK signaling pathways. On the other hand, zDEVD-fmk completely blocked the PDT-induced PARP cleavage (Fig. 4A, upper panel), the DEVD-directed caspase activity and efficiently counteracted the apoptotic cell death (Fig. 4A, lower panel). Similar results were obtained when cells were pretreated with the broad spectrum caspase inhibitor zVAD-fmk (data not shown). These results demonstrate that while the zDEVD-fmk-inhibitable caspases are crucial mediators of hypericin-induced cell death, caspase activation is either unrelated to or lies downstream from the JNK and p38 MAPK pathways. Furthermore, these observations show that JNK1 and p38 MAPK are activated whether or not cells will ultimately undergo apoptosis following PDT.
We also investigated the potential role of caspase-1 and caspase-8 in the process of hypericin-induced JNK1/p38 MAPK activation and apoptosis. For this, we used HeLa cells that stably overexpress the serpin-like protease inhibitor CrmA, which is a selective inhibitor of caspase-1 and caspase-8 (47). HeLa cells stably expressing the empty vector (HeLa-Hyg) were taken as control. Fig. 4B shows that PDT-induced activation of the JNK1 and p38 MAPK was not affected by the CrmA over-expression. For reasons not clear to us, the CrmA expression had a stimulatory effect on JNK1 activation at early time points (Fig. 4B), and a similar effect was also observed in parental HeLa cells treated with the peptide caspase inhibitors (data not shown). Similar observations have also been reported by others (46) in different cell lines. Nevertheless, our results clearly indicate that JNK1 activation in PDT-treated HeLa cells occurs independently from caspase activities. Moreover, hypericin-induced PARP cleavage was not blocked in the CrmA-expressing HeLa cells, while the TNF-mediated PARP fragmentation was completely prevented in the same cell line (Fig. 4C). HeLa-Hyg cells behaved as the parental cells (Fig. 4, B and C). In agreement with the differential PARP cleavage results, PDT treatment with hypericin induced a DEVD-directed caspase activation and apoptosis in the CrmA-expressing HeLa cells, while the TNF-induced apoptosis in the same cell line was blocked (data not shown). Altogether, these results indicate that photo-activated hypericin induces JNK1 and p38 MAPK activation as well as apoptosis in HeLa cells via a caspase-1-and caspase-8-independent mechanism.

Effect of Blocking the JNK and p38 MAPK Signal Transduction Pathways on the Hypericin-induced Programmed Cell
Death-The results presented so far indicate that while the zDEVD-fmk-sensitive executioner caspases play a critical role in the PDT-induced cell death, the early CrmA-inhibitable caspases are not essential in this process. It can also be concluded that the activation of JNK and p38 MAPK signaling pathways do not require these caspase activities.
Therefore, we looked for a possible functional role of JNK1 in the hypericin-induced apoptosis by transiently transfecting HeLa cells with dominant negative mutants of SEK1 (SEK-AL) and c-Jun (TAM-67) as upstream or downstream inhibitors of the JNK signaling pathway, respectively (2). SEK-AL was generated by mutating the phosphorylation and activation sites S220 and T224 of the wild type SEK1 into alanine and leucine, respectively (6). TAM-67 is a c-Jun mutant in which amino acids 3-122 have been deleted (38). This mutant protein re-  2 and 4 -7) and then either directly irradiated (lane 2) or treated with 100 ng/ml EGF 5 min before (lane 5) or immediately after (lane 6) illumination. Cells treated with 100 ng/ml EGF for 5 min (lanes 3 and 4) or 20 min (lanes 7 and 8) but not irradiated are shown. ERK2 activity was determined as described above.
tains the DNA binding and the leucine zipper domains but lacks most of the transactivation domain, which contains the phospho-acceptor sites for JNK and therefore fails to activate the transcription of AP-1 responsive genes. The role of p38 MAPK was assessed by using its specific pharmacological inhibitor, PD169316 (48). The ectopic expression of the MAPK phosphatase-1 (MKP-1), which can dephosphorylate and inactivate both JNK and p38 MAPKs (49), was also used to analyze the importance of these protein kinases in the apoptotic process. Cells were co-transfected with either TAM-67, SEK-AL, MKP-1 or treated with the p38 MAPK inhibitor (PD169316) in the presence of the ␤-galactosidase reporter plasmid pUT651 and then photosensitized with hypericin. Apoptotic cell death was evaluated 10 h after irradiation by morphological criteria. Fig. 5 shows that PDT treatment with hypericin in the presence of the empty vector alone resulted in about 45% apoptotic cells. Inhibition of the JNK pathway by transfecting the cells with either SEK-AL or TAM-67 increased the number of apoptotic cells by approximately 25%. Co-transfection of SEK-AL and TAM-67 rendered the cells even more susceptible to the cell death and increased hypericin-induced photocytotoxicity by approximately 40% over control. This oversensitization to PDTinduced apoptosis was slightly more pronounced when both kinases were simultaneously inhibited by treating the cells with PD169316 after transfection with either TAM-67 or SEK-AL. Cells treated with the p38 MAPK inhibitor alone or in combination with hypericin without light were indistinguishable from untreated control cells. These results point to a role for both the JNK1 and p38 MAPK signaling cascades in protecting the cells from hypericin-induced apoptosis.
To substantiate the observation that these stress kinase pathways protect cells from hypericin-induced apoptosis or at least delay this process, the kinetics of PARP cleavage and of the DEVD-directed caspase activity following PDT were evaluated in HeLa cells pretreated with the specific p38 MAPK inhibitor. As shown in Fig. 6A, in the PD169316 pretreated cells, PARP cleavage occurred considerably faster than in control cells. Apoptotic PARP fragmentation was nearly complete 6 h after irradiation in cells pretreated with the p38 MAPK inhibitor, while the cleaved 85-kDa product just began to appear in control cells at the same time period. Similarly, the kinetics of the DEVD-directed caspase activation were accelerated by cell pretreatment with the p38 MAPK inhibitor (Fig.  6B). In agreement with these observations, PD169316 pretreatment increased the potency of the PDT cytotoxic effect thereby making the cells oversensitive even to sublethal doses of hypericin (Fig. 6C).
Taken together, these data suggest that the hypericin-induced caspase and JNK/p38 MAPK activations are part of two functionally distinct and opposite pathways: the caspase-mediated pathway being required for apoptosis and the stress kinase cascade being involved in cellular defense against cell death.

DISCUSSION
This study reports that PDT with hypericin induces a strong and persistent activation of the JNK and p38 MAPK signaling pathways while inhibiting ERK2 activity. Most importantly this is the first study that outlines a protective role for the JNK/p38 MAPK pathways during PDT-induced apoptosis which is found to be critically dependent on a DEVD-directed caspase activity.
The effect of photo-activated hypericin on the MAPK activities was observed in several cell lines of human and mouse origin, indicating that it is a general cellular response to this PDT treatment (Fig. 1). In HeLa cells, the activation of JNK1 and p38 MAPK was very rapid and sustained for at least 24 h post-irradiation (Fig. 3). We have shown that photo-activated hypericin, under the conditions used, induces apoptosis in HeLa cells, an observation that has also been reported by others using different cell lines (34,35). Our results further show that following PDT, the JNK1 and p38 MAPK activation patterns overlap with the kinetics of PARP cleavage, a DEVDamc proteolytic activity and the onset of apoptosis (Fig. 3).
Recently, PDT with a benzoporphyrin derivative has been shown to induce JNK1 and p38 MAPK by a mechanism involving reactive oxygen species production in murine keratinocytes, but the effects of such activations were not explored (27). Unlike this observation, cell pretreatment with the antioxidants N-acetylcysteine or butylated hydroxyanisole did not affect hypericin-induced stress kinase activations. 2 Moreover, in the aforementioned study, PDT with BDP did not result in an inhibition of the ERKs which could be stimulated by EGF during photosensitization (27). In contrast, here we show that PDT with hypericin caused an inhibition of the ERK pathway which could not be relieved or prevented by EGF or FCS (pre)treatment. Interestingly, we have reported previously that hypericin inhibits the EGF receptor tyrosine kinase activity in vitro in an irreversible and light-dependent manner (50).
The relevance of the hypericin-mediated irreversible inhibition of the ERK signal to the process of PDT-induced cytotoxicity is not yet known. Intriguingly, several studies have suggested that the dynamic balance between the ERK and the JNK/p38 MAPK activities critically determines the cellular fate in response to differentiating, proliferating, or apoptogenic stimuli (16). However, while the role of the ERK signaling cascade in the processes of cell differentiation and proliferation has been clearly recognized, the requirement of JNK1 and p38 MAPK activations in apoptosis remains a matter of controversy. In some instances JNK and/or p38 MAPK activity is strictly required for apoptosis to occur (17)(18)(19), while in other circumstances these kinases are unrelated to the process of programmed cell death (21)(22)(23). Strong and persistent activation of JNK and/or p38 MAPKs with a concurrent inhibition of activities were determined at the indicated time post-irradiation as described above. C, parental, HeLa-CrmA, and HeLa-Hyg cells were either exposed to PDT with 250 nM hypericin or incubated with 10 4 IU/ml TNF in the presence of 1 g/ml cycloheximide. PARP cleavage was analyzed as described 16 h after treatment. ERK has been shown to be critical for apoptosis induced by UVand ␥-irradiation as well as by nerve growth factor withdrawal (16,18). In addition, a cross-talk between the JNK/p38 MAPK and caspase activation pathways has been reported in several systems (43)(44)(45)(46).
These observations prompted us to look into the possible functional relationship between the JNK1/p38 MAPK pathways and caspase activation following PDT. Activation of caspases can be achieved through at least two mechanisms. The best characterized pathway involves the autoactivation of procaspase-8 following its recruitment by the death effector domain of the adapter protein FADD (Fas-associated death domain), which associates to clustered death domains of Fas/ CD95 and TNFR1 (through TRADD, TNF receptor-associated death domain) upon ligand binding. This results in the caspase-8-mediated proteolytic activation of downstream effector caspases, such as caspase-3, caspase-7, and caspase-9, and commitment to cell death (reviewed in Ref. 51). Another pathway for caspase activation, triggered by many apoptogenic agents, involves the release of cytochrome c from the mitochondria into the cytosol (52). Cytosolic cytochrome c induces the ATP-or dATP-dependent formation of the "apoptosome," a complex of proteins composed of cytochrome c itself, Apaf-1, and procaspase-9 (53). The processing and activation of procaspase-9 in the complex ultimately leads to the activation of procaspase-3 and cell death (52). This latter mechanism may also contribute to the Fas-mediated apoptosis by amplifying the effect of caspase-8 on downstream caspases (54). Therefore, we used two approaches to discern the relationship between the JNK/p38 MAPK and the caspase activation pathways. First, we used an inhibitor of the downstream effector caspases, zDEVD-fmk, and showed that while it clearly blocked PARP cleavage and the DEVD-amc proteolytic activity, and significantly counteracted the hypericin-induced apoptosis, it did not affect either JNK1 or p38 MAPK activations (Fig.  4). This indicates that stimulation of the JNK and p38 MAPK signaling cascades, in response to photo-activated hypericin, does not require a zDEVD-fmk-inhibitable caspase activity. This in turn implies that caspase activation is either downstream of the JNK1 and p38 MAPK activations or that it lies on an independent pathway. However, our results demonstrate that a DEVD-directed caspase, most likely caspase-3, is a key mediator of the hypericin-induced cell death.
As a second approach, we examined the potential role of the early upstream caspase-1 and caspase-8 in PDT-induced JNK and p38 MAPK activations by stably expressing the serpin-like caspase inhibitor CrmA in HeLa cells. Overexpression of CrmA did not affect either the PDT-induced PARP cleavage, DEVDamc proteolytic activity, or the JNK1/p38 MAPK activations (Fig. 4) and, moreover, did not prevent the PDT-induced apoptosis. On the contrary, TNF-mediated PARP cleavage (Fig. 4) and cell death were completely inhibited in HeLa cells that overexpress CrmA confirming the central role of procaspase-8 in the cytotoxic response to TNF (55). These observations demonstrate that CrmA-inhibitable caspases are not required for PDT-induced apoptosis and clearly dissociate the hypericininduced cytotoxicity from the prototype mechanisms that involve cell surface death domain-containing receptors. As mentioned above, hypericin could lead to the activation of a DEVDdirected caspases and ultimately to apoptosis through mitochondrial damage and release of cytochrome c into the cytosol. Indeed, we have shown that cytochrome c release is one of the earliest events induced by PDT in hypericin-treated HeLa cells (39).
Finally, we present several lines of evidence that the activations of JNK1 and p38 MAPK may be regarded as cellular protective signals against hypericin-induced apoptosis. First, the expression of a dominant negative SEK1 mutant (SEK-AL), a direct upstream activator of JNK, and/or the transfection of the transactivation-deficient mutant of the c-Jun protein (TAM-67), the major downstream effector substrate of JNK, enhanced the rate of hypericin-induced cell death. Co-transfection of both mutant proteins significantly increased the percentage of apoptotic cells. Second, the specific pharmacological inhibitor of p38 MAPK, PD169316, also had an enhancing effect on the hypericin-induced cell death. Combined inhibition of JNK1 and p38 MAPK by PD169316 pretreatment of cells transfected with either SEK-AL or TAM-67 augmented the hypericin-induced apoptosis to nearly 90% (Fig. 5). Third, by expressing the dual specificity phosphatase MKP-1, which dephosphorylates and inactivates both JNK1 and p38 MAPK in vivo, we also observed a potentiation of the hypericin-induced apoptosis in HeLa cells. These observations were further supported by the findings that in the presence of the p38 MAPK inhibitor, PDT with hypericin induced PARP cleavage, DEVDdirected caspase activation, and apoptosis with much faster kinetics. PD169316 pretreatment also highly oversensitized the cells to PDT with sublethal doses of hypericin (Fig. 6). Altogether, the results suggest that JNK and p38 MAPK pathways are required for survival in response to PDT with hypericin. To some extent, our conclusions are in agreement with the results of a recent study on the protective role of the stress kinase pathways in the TNF/Fas-induced apoptosis in the murine cell line L929 (56). In that study, expression of dominant negative mutants of MKK4/SEK1, MKK6 or TRAF2-DN (all of which are mediators of JNK or p38 activation) enhanced the TNF-induced apoptosis. Previous studies have also shown that TNF can induce NF-B but not JNK in cells from TRAF2deficient mice and these cells are more susceptible to apoptosis than those from wild type mice (57,58). These reports suggest that JNK and p38 MAPK pathways may be essential for cell survival in the presence of TNF.
Overall, our results indicate that although JNK1 and p38 MAPK may play a role in protecting HeLa cells from PDTinduced cytotoxicity, they cannot rescue cells from hypericininduced cell death. Most likely they delay the apoptotic process by inducing the transcription of survival-promoting genes until a critical threshold of damage finally commits the cells to apoptosis. An intriguing explanation for the lack of full protection by the JNK/p38 MAPK pathways could reside in the con-current irreversible inhibition of the ERK pathway. It is tempting to speculate that the JNK/p38 MAPK survival pathway cannot cope with the hypericin-induced death signal as it is not complemented with an ERK-mediated proliferation pathway. This could be one of the reasons why PDT with hypericin has been shown to be a powerful in vivo anticancer tool. The observation that JNK and p38 MAPK inhibition enhances the PDTinduced cell death may offer molecular bases for the development of new therapeutic strategies to enhance the effectiveness of hypericin as an anticancer tool in PDT.