Inhibition of constitutively activated nuclear factor-κB radiosensitizes human melanoma cells

Melanoma cells are well known to be resistant to radiation and commonly used chemotherapy agents, and this resistance is associated with a poor prognosis for this metastatic disease. As such, radiotherapy is not commonly used as a primary treatment for melanoma. Instead, melanoma is typically treated with surgery, but adjuvant therapies such as IFN, combination chemotherapy, and the use of other biological response modifiers have begun to have major impact on the success of treatment (1). The incidence of melanoma is on the rise worldwide, with 43,000 new patients diagnosed in the United States. Therefore, improving the treatment of melanoma could benefit many patients. For example, the use of small molecules or gene therapy strategies that overcome the relative radioresistance of melanoma tumors could prove to be useful. To improve radiotherapy, however, it is important to understand the biological response these cells have to radiation, and any new strategies for sensitizing cells to radiation should target those pathways in the cell that contribute to radiation resistance. One such pathway may involve nuclear factor-κB (NF-κB).

Constitutive activation of NF-κB has been showed in a variety of human malignancies including pancreatic cancer, colon cancer, breast cancer, T-cell leukemias, and lymphomas. Several reports have also showed that NF-κB is constitutively activated in human melanoma cells (2, 3), and it has recently been shown that this constitutive activity is a result of elevated IκB kinase (IKK) activity arising from aberrant NF-κB-inducing kinase activation (4). Recent studies have shown NF-κB to be a critical regulator of apoptosis by controlling the transcription of genes with products that block cell death. It has been implied that constitutive activation of NF-κB induces overexpression of its downstream targets such as bcl-x_L, bcl-2, vascular endothelial growth factor, and interleukin-8, which may in turn mediate resistance to apoptosis induced by chemotherapy and radiation. Several genes involved in tumor initiation, promotion, and metastasis are also regulated by NF-κB, suggesting that NF-κB could mediate tumorigenesis and can thus be used as target for chemoprevention and treatment of cancer. Agents that suppress NF-κB activation have been shown to suppress the expression of genes involved in carcinogenesis and tumorigenesis in vivo (5).

The role of NF-κB as a modifier of radiosensitivity has only been analyzed in limited studies. Regulation of NF-κB is mainly controlled by the inhibitory IκB proteins, which include IκBα. On stimulation, IκB is rapidly phosphorylated by IKK and degraded via the ubiquitin-proteasome pathway, permitting nuclear import of NF-κB. There is sufficient evidence in the literature demonstrating the ability of proteasome inhibitors such as Z-Leu-Leu-Leu-H (MG132), lactacystin, and PS-341 in blocking NF-κB activation in cultured cells by preventing the degradation of IκB without interfering with its phosphorylation (6). Russo et al. (7) showed that inhibition of proteasome function dramatically enhanced the apoptotic response of colorectal cells to radiation both in vitro and in an in vivo xenograft model. In their study, treatment of colorectal cancer cells with the proteasome inhibitor PS-341 decreased radiation-induced NF-κB activation and resulted in an enhanced apoptotic response to radiation. Other studies have also showed an enhanced antitumor effect when PS-341 was combined with cytotoxic agents (8) or ionizing radiation (9). Dominant-negative mutant IκBα (IκBαM) has been engineered, which cannot be phosphorylated and degraded; thus, the nuclear importation and DNA binding of NF-κB complexes are repressed constitutively if cells are stably transfected with IκBαM. The stable expression of IκBαM has been shown to inhibit the activation of NF-κB in several cell types (10).

There are several other kinases known to act upstream of IKK. These include members of the mitogen-activated protein kinase (MAPK) kinase (MEK) kinase (MEKK) family such as MEKK-1, MEKK-2, and MEKK-3 (11–15). MEKK-1 activity has been associated with pancreatic cancer (16); MEKK-3 activity has been associated with hepatocellular carcinoma (17); and Tpl-2 activity has been associated with breast, colon, and gastric cancers (18, 19). Thus, cancers that have aberrant kinase activity upstream of IKK would also be expected to have constitutively active nuclear NF-κB, and this activity could in turn contribute to the expression of genes important to the survival of those cancers.

The present study was designed to investigate whether melanoma cell lines that are normally resistant to radiation could be made susceptible by inhibiting the activation of NF-κB. The inhibition of NF-κB, either by treatment with MG132 or by stable expression of IκBαM, led to down-regulation of the antiapoptotic proteins bcl-x_L, bcl-2, and survivin and decreased clonogenic survival in response to ionizing radiation. However, exposure of melanoma cells to the MEK inhibitor PD184352 inhibited constitutive extracellular signal-regulated kinase (ERK) activity determined based on phospho-p44/p42 expression and partially suppressed NF-κB activation but was unable to induce radiosensitivity as assessed by clonogenic cell survival assay. Therefore, approaches that inhibit the nuclear translocation or function of NF-κB, including proteasome inhibitors or IκBαM gene therapy, may be beneficial to the treatment of melanoma when combined with standard anticancer therapies such as radiation, but MEK inhibitors may not necessarily act in a synergistic manner to restore sensitivity to radiation in this particular setting.

Two different human melanoma cell lines were used to study the radiosensitizing effect of inhibition of NF-κB: A375 and MeWo. Cells were obtained from American Type Culture Collection (Manassas, VA) and routinely maintained in recommended media supplemented with 10% fetal bovine serum, 10,000 units/mL penicillin-streptomycin, and 2 mmol/L l-glutamine.

The proteasome inhibitor MG132 was obtained from Sigma-Aldrich Co. (St. Louis, MO). For all in vitro assays, a 10 mmol/L stock solution in DMSO was prepared and stored at −20°C until use. The MEK inhibitor PD184352 was prepared as a 10 mmol/L stock in DMSO and stored at −20°C until use.

The IκBαM has been engineered to not be phosphorylated or degraded through the introduction of mutations in Ser³² and Ser³⁶ (20). The plasmid was a generous gift of Dr. Paul Chiao (Department of Surgical Oncology, University of Texas M.D. Anderson Cancer Center, Houston, TX). Semiconfluent cells were transfected with the plasmid construct using FuGENE (Life Technologies, Inc., Rockville, MD). Stable clones were selected in geneticin (Life Technologies)–containing media. Mock-transfected cells were used as controls.

The effectiveness of the combination of proteasome inhibitors/IκBαM or the MEK inhibitor PD184352 and ionizing radiation was assessed by clonogenic assays. Melanoma cells were either mock treated with vehicle alone (DMSO) or pretreated with MG132 (10 μmol/L) for 2 hours. Then, cells were irradiated, drug was washed off, and cells were harvested for clonogenic assay. For experiments comparing the mock-transfected MeWo with those carrying IκBαM, cells were irradiated at various doses and processed for clonogenic cell survival assay. To study the effect of the MEK inhibitor PD184352, cells were exposed to 1 and 2 μmol/L concentrations for 24 hours and were irradiated and further processed for clonogenic cell survival assay. Cells were irradiated with a high-dose rate ¹³⁷Cs unit (4.5 Gy/min) at room temperature. Following treatment, cells were trypsinized and counted. Known numbers were replated in triplicate and returned to the incubator to allow macroscopic colony development. Colonies were stained with 0.5% gentian violet solution and counted after ∼14 days. The percentage plating efficiency and fraction surviving a given treatment was calculated based on the survival of nonirradiated cells treated with the agent in question.

Cells were harvested after treatment, rinsed in ice-cold PBS, and lysed in lysis buffer containing 50 mmol/L HEPES (pH 7.9), 0.4 mol/L NaCl, 1 mmol/L EDTA, 2 μg/mL leupeptin, 2 μg/mL aprotinin, 5 μg/mL benzamidine, 0.5 mmol/L phenylmethylsulfonyl fluoride, and 1% NP40. The lysed cells were centrifuged at 14,000 rpm to remove any cellular debris. Protein concentrations of the lysates were determined by the Dc protein assay system (Bio-Rad, Hercules, CA). Equal amounts of protein were separated by 12% SDS-PAGE, transferred to Immobilon (Millipore, Bedford, MA), and blocked with 5% nonfat dry milk in TBS-Tween 20 (0.05% v/v) for 1 hour at room temperature. The membrane was incubated with the respective primary antibody overnight. Antibody to human bcl-2 was obtained from DAKO (Carpinteria, CA), bcl-x_L and survivin were from Santa Cruz Biotechnology (Santa Cruz, CA), and actin was from Chemicon (Temecula, CA). After washing, the membrane was incubated with the appropriate horseradish peroxidase secondary antibody (1:2,000 dilution; Amersham Pharmacia Biotech, Arlington Heights, IL) for 1 hour. Following several washes, the blots were developed by enhanced chemiluminescence (Amersham Pharmacia Biotech) on a STORM 860 PhosphoImager (Molecular Dynamics, Heyward, CA).

Cells were rinsed twice with ice-cold PBS, harvested by scraping with a cell scraper, and centrifuged at 800 rpm for 10 minutes at 4°C. The cell pellet was resuspended in 400 μL of cold lysis buffer [10 mmol/L HEPES (pH 7.9), 10 mmol/L KCl, 0.1 mmol/L EDTA, 0.1 mmol/L EGTA, 1 mmol/L DTT, 2 μg/mL leupeptin, 2 μg/mL aprotinin, and 0.5 mmol/L phenylmethylsulfonyl fluoride] and incubated on ice for 10 minutes; 10% NP40 (12.5 μL) was added and the mixture was vortexed for 5 seconds. The lysate was centrifuged for 5 minutes at 4°C (14,000 rpm). The supernatant was removed and stored as cytosolic extract. The pellet was resuspended in 30 μL of extraction buffer [20 mmol/L HEPES (pH 7.9), 400 mmol/L NaCl, 1 mmol/L EDTA, 1 mmol/L EGTA, 1 mmol/L DTT, 2 μg/mL leupeptin, 2 μg/mL aprotinin, and 0.5 mmol/L phenylmethylsulfonyl fluoride], mixed thoroughly, and incubated on ice for 30 minutes. The pellet was vortexed every 10 minutes. At the end of 30 minutes, the extract was centrifuged for 10 minutes at maximum speed (14,000 rpm) in a microcentrifuge. The supernatant was designated as nuclear extract, aliquoted, and stored at −70°C until used in the electrophoretic mobility shift assays (EMSA).

To measure suppressed activation of NF-κB in response to treatment with MG132, PD184352, or IκBαM and radiation, EMSA was carried out. Nuclear extracts from these cells were run on EMSA to determine the best dose and timing that leads to maximum inhibition of NF-κB. Basically, nuclear extracts (15 μg) were incubated with poly(deoxyinosinic-deoxycytidylic acid) (1 μg) in binding buffer containing 10 mmol/L Tris, 50 mmol/L NaCl, 20% glycerol, 0.5 mmol/L EDTA, and 1 mmol/L DTT. [³²P]-labeled probe containing the NF-κB site was added and allowed to bind for 15 minutes. The complexes were separated on a 4% polyacrylamide gel and visualized by phosphorimaging.

Data were analyzed using the paired t test (Microsoft Excel). Data are presented as means ± SE. A difference was regarded as significant if P < 0.05.

It has been shown previously that NF-κB is constitutively activated in tumors of different origins including melanomas. We investigated NF-κB activity in two human melanoma cell lines, A375 and MeWo, by EMSA and compared it with the activity in nuclear extracts prepared from a lung cancer cell line, H1299, that does not express NF-κB. NF-κB was constitutively activated in both melanoma cell lines when compared with the H1299 cell line, which lacked any detectable activity (Fig. 1A). Further induction of NF-κB in response to ionizing radiation was observed in MeWo cells 2 hours following exposure to 4 Gy dose (Fig. 1B).

To measure suppressed activation of NF-κB in response to the proteasome inhibitor MG132, EMSA was done. Nuclear extracts from A375 and MeWo cells were analyzed by EMSA to determine the best dose and timing that leads to maximum inhibition of NF-κB. When the cells were pretreated with a relatively low dose (0.5 μmol/L) of the proteasome inhibitor MG132 for 24 hours, neither constitutive nor radiation-induced NF-κB activity was affected (data not shown). In both cell lines, MeWo (Fig. 2A) and A375 (Fig. 2B), we observed an induction of NF-κB activity by ionizing radiation that was blocked by a 2-hour preincubation with 10 μmol/L MG132. Therefore, this treatment with MG132 was picked for all future experiments.

To determine if inhibition of activated NF-κB can reverse the radioresistance of melanoma cells, MeWo and A375 cells were pretreated with 10 μmol/L MG132 for 2 hours, and the cells were irradiated and plated for clonogenic cell survival. Figure 3 shows that MG132 suppressed the clonogenic survival of both melanoma cell lines: MeWo (representing the mean of four independent experiments; Fig. 3A) and A375 (representing the mean of three independent experiments; Fig. 3B). Survival at 2 and 4 Gy was reduced from 48 ± 0.65% and 11.6 ± 0.33% in the control MeWo cells to 27.7 ± 0.32% and 6.25 ± 0.23% in MG132-treated MeWo cells, respectively (Fig. 3A). The observed sensitization was statistically significant with P < 0.005 and 0.0006 at 2 and 4 Gy, respectively. Similar results were obtained on exposure of A375 cells to MG132 with survival at 2 and 4 Gy being reduced from 48 ± 1.6% and 12.1 ± 0.7% in the control cells to 34.3 ± 0.7% and 7.6 ± 0.22%, respectively (Fig. 3B). The observed sensitization was statistically significant with P < 0.004 and 0.025 at 2 and 4 Gy, respectively. Survival enhancement ratios were calculated at the 50% cell survival by dividing radiation dose of the radiation-only survival curve with that of the corresponding MG132 plus radiation curve. Survival enhancement ratio was 1.47 for the MeWo cells and 1.3 for the A375 cells. Therefore, NF-κB activation may be one of the pathways responsible for radioresistance in melanoma cells. However, because inhibition of the proteasome does not specifically target just NF-κB, it was necessary to test the role of NF-κB by an independent and more specific means.

Because the proteasome selectively eliminates many cellular proteins, the effects observed on treating cells with proteasome inhibitors may not be specifically targeted toward inhibition of NF-κB. Therefore, a specific inhibitor of NF-κB activation, IκBαM, was introduced into MeWo cells to create a stable cell line. Figure 4A shows the expression of the IκBαM in the stably transfected clone by Western blot analysis. To determine if the activation of NF-κB in response to ionizing radiation was suppressed, nuclear extracts from IκBαM-transfected MeWo cells were run on EMSA. We observed that the induction of NF-κB activity by ionizing radiation (4 Gy) and the constitutive activation of NF-κB were completely abolished in the stably transfected cell line (Fig. 4B). These cells were further used to determine the role of NF-κB in mediating radioresistance in melanoma cells. We carried out clonogenic cell survival assay on MeWo IκBαM-transfected cells (representing the mean of four independent experiments) and observed a modest sensitization to ionizing radiation. Surviving fraction at both 2 and 4 Gy was decreased from 47 ± 0.8% to 32.9 ± 0.7% and 10.4 ± 0.4% to 7.4 ± 0.3% in the control versus the IκBαM cells (Fig. 5). The observed sensitization was statistically significant at 6 Gy with P = 0.045. The P values at 2 and 4 Gy were 0.12 and 0.14, respectively. Survival enhancement ratio at the 50% cell survival was found to be 1.28. These findings confirmed that activated NF-κB can be one of the possible mechanisms to explain the radioresistance of melanoma cells.

We determined the effects of NF-κB inhibition on some of its downstream antiapoptotic targets such as bcl-2, bcl-x_L, and survivin, a unique member of the inhibitor of apoptosis protein family of proteins, by Western blot analysis. Treatment with MG132 (10 μmol/L for 2 hours) and transfection of the IκBαM into MeWo cells caused a down-regulation of bcl-2, bcl-x_L, and survivin proteins (Fig. 6A and B). Combination of these treatments with a 4 Gy dose of radiation did not further suppress the levels of bcl-2 or bcl-x_L. Radiation alone slightly suppressed the levels of bcl-2 and survivin but enhanced bcl-x_L. However, combination treatments of MG132 or IκBαM with ionizing radiation suppressed survivin to a greater degree than either treatment alone (Fig. 6A and B).

There is a close relationship between NF-κB and MEK pathways, and MEK activation has been linked to activation of NF-κB. As shown in Fig. 7, MEK is constitutively active in MeWo and A375 cells as detected based on phospho-p44/p42 expression. Therefore, we used the MEK1 inhibitor PD184352 to test whether the MEK/ERK pathway was acting upstream of NF-κB activation. Treatment with 1 and 2 μmol/L doses of PD184352 for 24 hours down-regulated phospho-ERK in both A375 and MeWo melanoma cells (Fig. 7A). MeWo cells were treated with PD184352 (1 and 2 μmol/L) for 24 hours, and nuclear extracts were prepared and run on EMSA to determine its inhibitory effects on NF-κB activation. PD184352 when used at the above-mentioned concentrations suppressed both constitutive and ionizing radiation–induced NF-κB activation (Fig. 7B) but not to the extent seen previously with either MG132 or IκBαM. To determine the ability of PD184352 to produce a radiosensitizing effect, melanoma cells were exposed to 1 and 2 μmol/L doses of PD184352 for 24 hours, and cells were irradiated and harvested for clonogenic assays. In spite of its ability to partially suppress NF-κB activation, PD184352 was unable to mediate any sensitization to ionizing radiation in the MeWo cells within experimental error (Fig. 8).

NF-κB is a transcription factor that is activated by various stress stimuli including oxidative stress and DNA-damaging agents such as ionizing radiation. In general, activation of NF-κB leads to induction of antiapoptotic genes such as bcl-2 and bcl-x_L and members of the inhibitor of apoptosis protein family, thereby promoting cell survival. NF-κB has also been implicated in resistance to radiation-induced cytotoxicity, and its role in radiation response has been addressed in several studies showing that NF-κB activation mediates radiation-induced gene expression in both normal and tumor cells (21–23). Because NF-κB was reported to be constitutively activated in melanoma cells (2), we investigated the effect of modulation of the NF-κB DNA binding activity on the intrinsic radiosensitivity of human melanoma cell lines. Both melanoma cell lines used in this study, MeWo and A375, showed constitutively activated NF-κB (Fig. 1). In addition to melanomas, constitutive activation of NF-κB is a common feature in several other tumor types including glioblastomas and tumors of the pancreas, colon, and breast where it apparently influences oncogenesis in several ways including enhancement of cell survival, metastasis, and angiogenesis (24–27). Blockade of basal NF-κB activity through the use of inhibitors efficiently reduces chemoresistance of human tumor cells and offers the potential for improved therapeutic strategies. For example, it has been shown that selective inhibition of constitutive NF-κB activity with various NF-κB inhibitors (MG132, gliotoxin, sulfasalazine, and transfection with the IκBα superrepressor) renders pancreatic cancer cell lines more sensitive to the chemotherapeutic drugs VP16 and doxorubicin (28). Therefore, we hypothesized that strategies that selectively inhibit NF-κB activity in melanomas could produce a radiosensitizing effect.

The proteasome has recently been proposed to be a molecular target for cancer therapy (29). Proteasome inhibitors have been shown to promote apoptosis (30, 31). Several proteasome inhibitors have been described such as MG132, lactacystin, and PS-341. The effects of proteasome inhibition on NF-κB activation have been well established (32). Typically, the use of proteasome inhibitors has been shown to decrease NF-κB DNA binding activity, thereby leading to induction of a more radioresponsive and chemosensitive phenotype in cancer cells. In the present study, we have shown that the proteasome inhibitor MG132 can inhibit NF-κB activity in melanoma cell lines as determined by EMSA. Suppression of NF-κB DNA binding activity by PS-341 has been shown previously to affect the radioresponse of human prostate cancer cells (9, 10). This earlier report is confirmed by our in vitro clonogenic assays, which revealed that melanoma cells were significantly sensitized to subsequent doses of radiation when pretreated with MG132 to reverse NF-κB activity (Fig. 3).

Activation of NF-κB requires that IκB is phosphorylated at two serine residues (Ser³² and Ser³⁶) by IKKs, polyubiquitinated, and subsequently degraded by the 26S proteasome. This process frees NF-κB for translocation to the nucleus and activation of its target genetic programs (33). IκBαM, which cannot be phosphorylated, has been constructed and used to alter the radiation response of tumor cells. In a previous report by Pajonk et al. (10), PC3 prostate and HD-MyZ Hodgkin's lymphoma cells were transfected with an adenovirus carrying a dominant-negative IκBα (AdCMV-IκBα). An increase in radiation-induced apoptosis was observed, but there was no substantial effect on clonogenic survival. These authors concluded that radiosensitivity of the surviving cells was not significantly altered in PC3 cells transduced with AdCMV-IκBα, although the constitutive DNA binding activity of NF-κB was dramatically decreased and the clonogenicity was decreased to 19.6% of untreated control cells. This report is in contrast to our results (conducted on cells having 80% survival for both untreated controls), which show that inhibition of constitutive NF-κB by proteasome inhibitors or IκBαM restores some radiosensitivity to human melanoma cells. However, due to the toxicity of their vector, Pajonk et al. (10) would have carried out the radiation survival experiments on relatively small subsets of surviving cells, whereas their biochemical assays were done using cellular extracts prepared from the total cell population. In our study, we did EMSA on nuclear extracts rather than on total cellular extracts. Our results revealed a substantial decrease in the amount of radiation-induced nuclear NF-κB activation by MG132 or IκBαM compared with appropriate controls, which correlated with a restoration with radiosensitivity.

Apoptosis is regulated, in part, by members of the bcl-2 family. The relative abundance of proapoptotic and antiapoptotic members of this protein family determines whether a cell will undergo apoptosis or not. Apoptosis propensity is also regulated to a significant degree by the inhibitor of apoptosis protein family, which suppress apoptosis when expressed. The protein survivin is a member of this family. Thus, enhanced radiation sensitivity may be associated with a suppression of antiapoptotic proteins coupled with a restoration of apoptosis. Therefore, we examined the effect of inhibition of activated NF-κB on its antiapoptotic transcriptional targets such as bcl-2, bcl-x_L, and survivin. These antiapoptotic proteins were found to be down-regulated in the IκBαM transfectants and following treatment with MG132 (Fig. 6). However, in pilot experiments conducted as part of this study, these treatments did not seem to restore apoptosis propensity to these melanoma cells (data not shown), and this end point was not further pursued. Thus, we conclude that the inhibition of NF-κB activation by MG132 and IκBαM leads to a suppressed expression of certain antiapoptotic proteins, but this is not sufficient to restore apoptosis in these cells. Moreover, restoration of radiosensitivity in this context is independent of the apoptotic pathway, suggesting that, in addition to antiapoptotic proteins, NF-κB may target other genes that control radiation response.

Aberrant NF-κB activation in human cancers has been attributed to activation of the signaling kinases acting upstream of IKK. Several MAPK family members have been found to act upstream of IKK. These kinases additionally activate members of the MAPK cascade, leading to the activation of Jun NH₂-terminal kinase, p38 MAPK, and/or ERK (34). In human melanoma cells, the NF-κB activity downstream of NF-κB-inducing kinase seems to be dependent not only on the activation of IKK but also on NF-κB-inducing kinase–regulated activation of the MAPK/ERK cascade (4). Thus, several different MAPK proteins may be capable of activating NF-κB independently of IKK. Apparently, this parallel pathway plays an important role for IKK-induced NF-κB activity. Although previous studies have suggested that MEK inhibitors suppress the growth of various cancers and enhance the effectiveness of certain treatment modalities, some investigators have suggested that MEK inhibitors are not useful for sensitizing carcinoma cells to ionizing radiation (35, 36). Belka et al. (37) found that treatment with PD98059 did not alter radiation responses despite blocking MEK-1 kinase, indicating that MEK-1 and Erk1/2 are not involved in radiation resistance. However, Vrana et al. (38) used PD98059 in combination with ionizing radiation to treat HL-60 cells, which resulted in a large increase in apoptosis and a large decrease in clonogenicity when compared with either treatment alone. This study along with recent findings from Shonai et al. (39) implicate the MEK/ERK pathway in the radioresistance of lymphocytic leukemia cells, suggesting that MEK inhibitors may be useful for radiosensitizing hematopoietic cancers. In the present study, the MEK inhibitor PD184352 was unable to sensitize melanoma cells to ionizing radiation in spite of its ability to partially block radiation-induced NF-κB activation, suggesting that the MEK/ERK pathway may not be a key mediator of NF-κB activation in melanoma cells and that other signaling pathways may mediate NF-κB activation in these cells.

In summary, our results suggest that the NF-κB pathway may play at least a partial role in mediating the radioresistance of human melanoma cells. Identification of upstream signaling pathways associated with its activity will shed additional light on its role in melanoma and could lead to novel strategies that reverse its effects, thereby enhancing response to therapeutics including radiation. Such strategies may also include small molecule proteasome inhibitors and gene therapy vectors that target NF-κB activation.

The MEK inhibitor PD184352 was a kind gift from Dr. Eric Kaldjian (Pfizer, Inc., Ann Arbor, MI).