Forced Expression of the H11 Heat Shock Protein Can Be Regulated by DNA Methylation and Trigger Apoptosis in Human Cells*

H11, the eukaryotic homologue of a herpes simplex virus protein, has the crystallin motif of heat shock proteins (Hsp), but it differs from canonical family members in that mRNA and protein levels were reduced in various tumor tissues and cell lines (viz. melanoma, prostate cancer and sarcoma) relative to their normal counterparts. In these cells, expression was not restored by heat shock, but rather by the demethylating agent 5-aza-2′-deoxycytidine (Aza-C). Forced H11 expression by Aza-C treatment, transient transfection with H11 expression vectors, or retrovirus-mediated delivery of H11 under the control of a tetracycline-sensitive promoter triggered apoptosis. This is evidenced by a significant (p < 0.001) increase in the percentage of cells positive for terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling (TUNEL) and for activation of caspase-3 and p38MAPK and by the co-localization of TUNEL+ nuclei with increased H11 levels. Apoptosis was partially inhibited by the pancaspase inhibitor benzyloxycarbonyl-Val-Ala-Asp-fluoromethyl ketone or the p38MAPK inhibitor SB203580. It was abrogated by co-treatment with both inhibitors, suggesting that H11-triggered apoptosis is both caspase- and p38MAPK-dependent. A single site mutant (H11-W51C) had cytoprotective activity related to MEK/ERK activation, and it blocked H11-induced apoptosis in co-transfected and Aza-C-treated cells, indicating that it is a dominant negative mutant. This is the first report of a heat shock protein with proapoptotic activity.

Apoptosis is a tightly regulated irreversible process that results in cell death in the absence of inflammation and is considered a promising alternative to cancer chemotherapy (1). Morphologic changes associated with apoptosis include nuclear and cytoplasmic condensation, intranucleosomal DNA cleavage, and blebbing into membrane-bound apoptotic bodies. Apoptosis is primarily mediated by caspases, which are cysteine proteases with aspartate specificity that are activated by the cleavage of inactive zymogens (procaspases) in a sequential cascade or by autocatalysis. Caspase-3, one of the main effectors of apoptosis, is activated by cleavage of the inactive pro-caspase-3 into 17-20-and 11-kDa fragments. It is responsible for the proteolytic cleavage of many key proteins, such as the nuclear poly(ADP-ribose) polymerase, which is involved in DNA repair (2).
Environmental stress stimuli and the oligomerization of death receptors recruit apoptosis-regulating kinase 1 and trigger activation of the proapoptotic c-Jun N-terminal kinase (JNK) 1 and p38MAPK protein cascades (3). Apoptosis induced by p38MAPK is both caspase-dependent (4) and -independent (5). Survival stimuli activate various effector pathways, one of which involves c-Raf-1 or B-Raf kinases and the activation of MAPK kinase (MEK) and extracellular signal-regulated kinase (ERK, also known as MAPK) through phosphorylation. ERK targets include genes required for cell cycle progression and the altered balance between pro-and antiapoptotic members of the Bcl2 family of apoptosis regulatory proteins (6). Activation of the ERK survival pathway can override apoptotic cascades triggered by various stimuli (7)(8)(9).
Heat shock proteins (Hsp) are a new family of apoptosis regulatory proteins that are overexpressed in tumor cells (10). Family members have antiapoptotic activity (11,12), which is mediated by their ability to block the activation of apoptosisregulating kinase 1, JNK, and/or p38MAPK (11,13,14), interfere with c-Raf-1 function (15) or formation of the death receptor (16), or inhibit procaspase autocatalysis (17). However, Hsp with proapoptotic activity have not been previously described. The recently identified protein H11 (also known as Hsp22 and HspB8) is the eukaryotic homologue of the herpes simplex virus type 2 protein ICP10 protein kinase (PK). H11 retains the structural motif of the crystallin family of Hsp, but it differs from canonical family members in that it has an N-terminal membrane domain, is associated with the cell surface, and has auto-and trans-phosphorylating serine/threonine PK activity (18 -21). The studies described in this report were designed to examine the role of H11 in apoptosis regulation.

MATERIALS AND METHODS
Cells-Human melanoma cells SK-MEL-2 and SK-MEL-28 and human embryonic kidney cells HEK293 were grown in Eagle's modified minimal essential medium with 10% fetal bovine serum (FBS), 1 mM sodium pyruvate, and 0.1 mM nonessential amino acids. For SK-MEL-28, the medium was additionally supplemented with 4.5 g/liter glucose, 1500 mg/ml sodium bicarbonate, and 4 mM glutamine. Human melanoma cells A375 were grown in similarly supplemented Dulbecco's * These studies were supported by NCI, National Institutes of Health (NIH), Public Health Service Grant CA75453 and in part by NIAMD, NIH, Grant AR42647. The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
ʈ To whom correspondence should be addressed: Dept. of Pharmacology and Experimental Therapeutics, University of Maryland, School of Medicine, 655 W. Baltimore St., Baltimore, MD 21201-1559. Tel.: 410-706-3895; Fax: 410-706-2513; E-mail: laurelia@umaryland.edu. modified minimal essential medium. Human breast cancer cells MCF-7 and MDA468 were grown, respectively, in Dulbecco's modified minimal essential medium and Eagle's modified minimal essential medium with 10% FBS, 0.1 mM nonessential amino acids, 1 mM sodium pyruvate, and 2 mM glutamine. Human breast cancer cells MDA231 and prostate cancer cells DU145 were grown in Eagle's modified minimal essential medium with 10% FBS and 2 mM glutamine. Human prostate cancer cells PC-3 and TSU were grown in similarly supplemented RPMI 1640. Human Ewing's sarcoma family tumor cell line TC32 was grown in Dulbecco's modified minimal essential medium with 10% FBS. HEK293 cells stably transfected with H11-W51C (TAG51) were established as described (18) and cultured in Eagle's modified minimal essential medium plus 10% FBS with 400 g/ml G418. Normal human melanocytes were purchased from Clonetics (San Diego, CA).
Expression Vectors and Transfection-Wild type (WT) H11 was cloned from human heart cDNA by PCR amplification with primers that flank the open reading frame (5Ј-GGATTCTATGGCTGACGGTCA-GAT-3Ј (sense) and 5Ј-GTCGACAATCTCAGGTACAGGTGA-3Ј (antisense)) and ligated into the pCI vector (Promega) at the EcoRI/XbaI sites (pCI-H11). Its sequence was identical to that in the human genome data base. H11 similarly cloned from SK-MEL-2, A375, SK-MEL-28, PC-3, and HEK293 cells had the WT sequence. To construct the WT H11 expression vector pFLAG-H11, the EcoRI/XbaI fragment from pCI-H11 was subcloned into the expression vector p3ϫFLAG-CMV-10 (pFLAG; Sigma) downstream of the CMV promoter and the N-terminal 3ϫFLAG peptide sequence. The original H11 clone (18) has a G to C transition in codon 51 (exon 1), which leads to the exchange of tryptophan for cysteine. This mutation (henceforth designated H11-W51C) results in seven additional ␤-turns in the secondary protein structure predicted by the Garnier-Osguthorpe-Robson model (Genetics Computer Group Program (Madison, WI), PeptideStructure (26)). The construction of the expression vector for H11-W51C (pYXH11) was previously described (18). The c-Raf-1 dominant negative mutant (c-Raf-DN) was obtained from Dr. Bernard Weinstein (Columbia University, New York). Its ability to inhibit c-Raf-1 was previously described (9,27). Transient transfection was with 2 or 4 g of DNA and FuGene 6 Transfection Reagent (Roche Applied Science) according to the manufacturer's instructions. DNA concentrations in single vector controls for co-transfection experiments were equilibrated with pCI DNA.
Construction of the H11 Rev-Tet Vector and Cell Infection-The tetracycline-regulated RevTet-On expression system (Clontech, Palo Alto, CA) was used for inducible H11 expression according to the manufacturer's instructions. Briefly, the PT67 packaging cell line was transfected with the pTet-On vector (contains a neomycin resistance cassette) by the calcium phosphate method, and a stable cell line (Tet-On) was selected with G418 (500 g/ml). A 0.6-kb DNA fragment representing the H11 open reading frame was isolated by EcoRI/SalI digestion of the H11 plasmid (pAS2-1/H11), blunt-ended with Klenow fragment, and ligated into the SalI blunt-ended pRev-TRE retrovirus vector in which the murine mammary tumor virus promoter drives a hygromycin resistance gene and the Tet-sensitive promoter (Tet response element upstream of the minimal CMV IE promoter) regulates the inserted H11. The vector was transfected into Tet-On cells, and clones resistant to hygromycin (300 g/ml) were selected and expanded to generate stable virus-producing cell lines. Virus titers in the culture supernatants were assayed as described by the manufacturer and used to infect TC32 cells at a multiplicity of infection of 50. H11 expression was induced with 0.5 g/ml Dox, as per the manufacturer's instructions.
Terminal Deoxynucleotidyltransferase-mediated dUTP Nick End Labeling (TUNEL)-The in situ cell death detection kit (Roche Applied Science) was used as previously described (9,28). Briefly, cells were fixed with 4% paraformaldehyde and permeabilized with 0.1% Triton X-100 (in 0.1% sodium citrate), and DNA breaks were labeled by incubation (60 min; 37°C) with terminal deoxynucleotidyltransferase and nucleotide mixture containing FITC-conjugated dUTP. After examination by fluorescence microscopy, the cells were incubated (30 min; 37°C) with an anti-FITC antibody conjugated with AP, and the chromogenic reaction was carried out with the AP substrate 0.4 mg/ml nitro blue tetrazolium chloride and 0.2 mg/ml 5-bromo-4-chloro-3-indolyl phosphate, toluidine salt (Roche Applied Science). Apoptotic cells (characterized by a dark nuclear precipitate) and nonapoptotic cells (unstained or displaying a diffuse, light, and uneven cytoplasmic staining) were counted in five randomly chosen microscopic fields (at least 250 cells), and results are expressed as percentage of apoptotic cells Ϯ S.E.
DNA Fragmentation-The assay was as previously described and used 1.5% agarose gels with 0.5 mg/ml ethidium bromide (28).
Hybridization of Multiple Tissue Arrays-The multiple tissue expression and matched tumor/normal expression arrays were obtained from Clontech. They consist of a positively charged nylon membrane to which poly(A) ϩ mRNAs from different human tissues are immobilized in separate dots. They are normalized to the mRNA levels of eight housekeeping genes. Hybridization was with H11 cDNA (29) and the control ubiquitin cDNA probe provided by the manufacturer, and the results were visualized by autoradiography (exposure to the x-ray film was at Ϫ80°C for 21 h).
Reverse Transcriptase-Polymerase Chain Reaction (RT-PCR)-Total cellular RNA prepared as previously described (30) was used to generate cDNA with the Advantage TM RT-PCR kit (Clontech) and 10 pmol (0.5 M) of the antisense H11 or ␤-actin primers according to the manufacturer's instructions. Normal and tumor prostate, colon, and breast cDNAs and skeletal muscle cDNA (human MTC panels) were purchased from Clontech. PCR was done with cDNA and 10 pmol of H11 primers (5Ј-CCATGGCTGACGGTCAGATGCCCTTCTCCT-3Ј (sense) and 5Ј-TCCATGCCAAAGCCATCATCCAGCAG-3Ј (antisense)) and ␤-actin primers (5Ј-GTGGGGCGCCCCAGGCACCA-3Ј (sense) and 5Ј-CTCCTTAATGTCACGCACGATTTC-3Ј (antisense)) using the Advan-Taq Plus kit (Clontech). The PCR program was 1 min at 94°C for 1 cycle followed by 94°C for 30 s and 68°C for 2 min for 26 cycles and a final extension of 5 min at 68°C on a PerkinElmer Life Sciences thermocycler (model 9700). Products were analyzed by electrophoresis on 8% acrylamide gels and stained with 0.5 g/ml ethidium bromide.
Immunofluorescence and Immunohistochemistry-For double immunofluorescence, cells were incubated with terminal deoxynucleotidyltransferase and the nucleotide mixture (containing FITC-dUTP) for 1 h at 37°C and stained (1 h, room temperature) with H11 antibody followed by rhodamine-conjugated anti-rabbit IgG (30 min, room temperature). They were visualized with an epifluorescent confocal microscope fitted with an argon ion laser (Zeiss LSM 410) (9). For immunoperoxidase staining, cells were exposed overnight (4°C) to primary antibodies, and the immunolabeled cells were detected by the streptavidinbiotin method using the DAKO LSAB 2 kit and horseradish peroxidase (DAKO Corp., Carpinteria, CA) according to the manufacturer's instructions. Cells were counted in five randomly chosen microscopic fields (at least 250 cells), and the results were expressed as percentage of positive cells Ϯ S.E.
Immunoblotting and Immunocomplex PK Assays-Cells were treated with radioimmune precipitation assay buffer (20 mM Tris-HCl, pH 8.0, 150 mM NaCl, 1% Nonidet P-40, 1% sodium deoxycholate, 0.1% SDS) supplemented with phosphatase and protease inhibitor mixtures (Sigma) and sonicated for 60 s at 25% output power using the Sonicator/ Ultrasonic Processor (Misonix, Inc., Farmingdale, NY). Total protein was determined by the bicinchoninic assay (Pierce). Proteins were resolved by SDS-PAGE and transferred to nitrocellulose membranes. For immunoblotting with antibodies to P-p38MAPK and P-ERK 1/2, blots were blocked (1 h; room temperature) with blocking buffer (TN-T buffer (0.01 M Tris-HCl, pH 7.4, 0.15 M NaCl, 0.05% Tween 20) containing 1% bovine serum albumin (Sigma)) and incubated with antibody for 12 h at 4°C. Immunoblotting with all other antibodies was with blocking buffer containing 5% nonfat milk (w/v) and a 1-h incubation at room temperature. After three washes, blots were incubated with Protein A-peroxidase diluted in blocking buffer for 1 h at room temperature followed by four additional washes. Detection was with ECL reagents (Amersham Biosciences), followed by exposure to high performance chemiluminescence film (Hyperfilm ECL; Amersham Biosciences). Quantitation was by densitometry using the Bio-Rad GS-700 imaging densitometer (9,18,28). The immunocomplex PK assay was as previously described (18).

H11 Expression Is Altered in Tumor Cells/Tissues-Since
Hsp are overexpressed in tumor cells (10) and H11 has properties distinct from those of canonical Hsp (18,21), we wanted to know whether H11 is also overexpressed in tumor cells. Analysis of multiple tissue expression arrays confirmed previous findings (20,21,29) that H11 mRNA levels (expressed as -fold abundance relative to PBL) were particularly abundant in skeletal muscle, placenta, and heart (Fig. 1A). However, H11 expression was altered in some tumor tissues relative to their matched normal counterparts, with mRNA levels reduced in most colon tumors, increased in many stomach tumors, and individually altered (decreased or increased) in tumors of other sites (Fig. 1B). Focus on some cancer tissues/cell lines and their normal cells counterparts, using the more specific and sensitive RT-PCR assay, indicated that H11 RNA levels were decreased in prostate, colon, and breast cancer cells and a melanoma and a sarcoma cell line (Fig. 1C). We interpret the data to indicate that H11 expression can be altered in human tumors and is significantly decreased in some cancers.
H11 Expression Can Be Increased by Aza-C but Not Heat Shock-Stressful conditions, such as heat shock, increase Hsp expression over basal level (12). To determine whether H11 expression is also regulated by heat shock, we focused on HEK293 (kidney), SK-MEL-2 (melanoma), and PC-3 (prostate cancer) cells, all of which have low levels of H11 expression. Cells were heat-shocked by incubation for 1 h at 42.5°C, and extracts were obtained immediately thereafter (0 h) or at 4, 8, and 12 h after reincubation at 37°C (to allow for protein synthesis) and were immunoblotted with H11 antibody. Heat shock caused a time-dependent increase in the levels of H11 protein in HEK293 cells, with maximum expression seen at 8 ( Fig. 2A, lane 4) and 12 ( Fig. 2A, lane 5) h after treatment. It did not alter the expression of H11 in SK-MEL-2 ( Fig. 2A, lanes 6 -10) or PC-3 (data not shown) cells. By contrast, the levels of H11 protein in SK-MEL-2, A375, SK-MEL28, and PC-3 cells were significantly increased by treatment with the demethylating agent Aza-C (4 days; 2 M) (Fig. 2B, lanes 6, 8, 10, and 12). Aza-C had no effect on H11 expression in HEK293 cells (Fig. 2B, lanes 3 and 4) or normal melanocytes (Fig. 2B, lanes  1 and 2), and actin levels were virtually identical in all samples (Fig. 2B). Preimmune serum was negative (data not shown). The data indicate that (i) heat shock-mediated regulation of H11 expression is cell type-specific, and (ii) loss/diminution of heat shock responsiveness can be due to promoter methylation, a finding not previously reported for human Hsp.
Forced H11 Expression Triggers Apoptosis-Having seen that H11 expression is reduced/inhibited in some cells, we wanted to know whether its forced expression in these cells triggers apoptosis. We focused on SK-MEL-2, PC-3, and HEK293 because H11 expression is similarly increased in these cells, whether by Aza-C (SK-MEL-2 and PC-3) or heat shock (HEK293). As shown in Fig. 3A for SK-MEL-2, the percentage of TUNEL ϩ cells was significantly (p Ͻ 0.001 by ANOVA) increased in SK-MEL-2 and PC-3 cultures when H11 was expressed (Fig. 3B). HEK293 cells did not exhibit increased apoptosis when given a heat shock to induce H11 expression. To determine whether H11 is involved in apoptosis, duplicate cultures were similarly treated with Aza-C or heat shock but in the presence of increasing concentrations (0 -30 M) of H11-specific aODN (or sODN control). As shown in Fig. 4 for SK-MEL-2 cells, apoptosis was not inhibited by sODN, but there was a dose-dependent decrease in the percentage of TUNEL ϩ cells in cultures treated with Aza-C in the presence of aODN (Fig. 4A). We conclude that the aODN effect is specific, because H11 expression was inhibited in duplicate cultures treated with aODN (Fig. 4B, lane 4) but not sODN (Fig. 4B,  lane 3), and actin levels were similar in all cultures. Indeed, double immunofluorescent staining with FITC-dUTP (TUNEL) and rhodamine-labeled H11 antibody indicated that virtually all of the TUNEL ϩ cells (43-48% of Aza-C-treated cultures) overexpressed H11 (Fig. 5C). Cells with low intensity rhodamine staining (low H11 levels) (Fig. 5B) were TUNEL Ϫ (Fig. 5,  A and C), also in untreated (Fig. 5E) and Aza-C ϩ aODNtreated (Fig. 5D) cultures. In cultures treated with Aza-C ϩ sODN, ϳ39 -45% of the cells were positive for both H11 overexpression and TUNEL (Fig. 5F). Similar results were obtained for PC-3 cells (data not shown).
Because caspase-3 is a major effector of apoptosis (2) and TUNEL has come under scrutiny in terms of its ability to reflect apoptosis (31), duplicates of the Aza-C-treated cultures were examined for caspase-3 activation by (i) immunoblotting with an antibody that recognizes the 32-kDa procaspase-3 and its p20/p11 cleavage products and (ii) immunohistochemistry with an antibody specific for the large (p20) caspase-3 cleavage fragment (caspase3p20). Cleavage products p20/p11 were seen in Aza-C-treated (but not untreated) cultures (Fig. 5G), and the percentage of caspase3p20 ϩ cells was significantly (p Ͻ 0.01 by ANOVA) increased by Aza-C treatment. The percentage of caspase3p20 ϩ cells was similarly increased in cultures treated with Aza-C plus sODN but not in those treated with Aza-C plus aODN (Fig. 5H). Collectively, the data indicate that forced H11 expression triggers apoptosis in some but not other cell types. However, the percentage of caspase3p20 ϩ cells was lower than that of TUNEL ϩ cells (Fig. 4A), suggesting that factors other than caspase activation contribute to apoptosis.
We utilized a model of Ewing's sarcoma family tumor, the TC32 cell line, to further evaluate the role of H11 in apoptosis. TC32 cells stably transfected with the H11 retrovirus were induced to express H11 by Dox treatment and examined for H11 expression and apoptosis at 3-14 days thereafter. H11 expression was seen in Dox-treated cells at 3 days after treatment (Fig. 6E), and this increase corresponded with increased apoptosis, as measured by TUNEL. The percentage of rounded, refractile, and detaching cells increased with time after Dox treatment, reaching maximal levels at 7-9 days (Fig. 6F, panel  2). All of the cells were dead/detached by day 14 after treatment. The percentage of TUNEL ϩ cells increased from 3 Ϯ 1% before treatment, to 31 Ϯ 2% on day 3 and 56 Ϯ 5% on days 7-9 after Dox treatment (Fig. 6G). Cell death (apoptosis) was not seen in untreated cultures (Fig. 6, F (panel 1) and G) that did not express H11 (Fig. 6E, lane 2). The data support the conclusion that (i) forced H11 expression triggers apoptosis, and (ii) factors other than caspase-3 activation are involved in H11triggered apoptosis.
H11-W51C Antiapoptotic Activity Is MEK/ERK-dependent-Studies from our laboratory (9) and other laboratories (7,8) have previously shown that apoptosis can be blocked by activation of the ERK survival pathway. Therefore, we wanted to know whether MEK/ERK are activated by H11-W51C. Extracts of TAG51 and HEK293 cells, untreated or treated (45 min; 37°C) with 20 M of the MEK-specific inhibitor U0126, were immunoblotted with antibodies specific for ERK1/2 or phosphorylated (activated) ERK1/2 (P-ERK1/2). The levels of P-ERK2 were similar in HEK293 (Fig. 9A, lane 1) and TAG51 cells (Fig. 9A, lane 2), consistent with previous reports that ERK2 is constitutively activated in HEK293 cells (32). However, the levels of P-ERK1 were significantly higher in TAG51 than HEK293 cells (Fig. 9A, lanes 1 and 2). Data analysis as P-ERK/ERK ratios indicated that the levels of P-ERK1 and P-ERK2 were 35-and 2.5-fold higher, respectively, in TAG51 than in HEK293 cells (Fig. 9B). P-ERK1/2 were not seen in U0126 treated TAG51 cells (Fig. 9, A (lane 3) and B), indicating that their activation is MEK-dependent. Significantly, P-ERK1/2 levels were not altered by transfection of TAG51 cells with c-Raf-DN (Fig. 9, A (lane 4) and B), suggesting that c-Raf-1 is not involved in MEK/ERK activation. However, the levels of phosphorylated B-Raf were significantly (34-fold) higher in TAG51 than HEK293 cells (Fig. 9C), as determined by immunoprecipitation/immunoblotting of cells pulse-labeled with [ 32 P]orthophosphate (2.5 mCi, 2 h) with B-Raf antibody (Fig. 9C). The data suggest that B-Raf may be involved in MEK/ERK activation in TAG51 cells.
H11-W51C Overrides H11-induced Apoptosis-To examine whether H11-W51C is a dominant-negative mutant of H11induced apoptosis, SK-MEL-2 cultures were co-transfected with H11 and H11-W51C or H11 and the empty vector (4 g each) and examined by TUNEL at 48 h after transfection. The percentage of TUNEL ϩ cells was significantly lower in cultures co-transfected with H11 and H11-W51C (9.7 Ϯ 1.9%) than H11 and empty vector (40.7 Ϯ 2.3%) or H11 alone (39.4 Ϯ 1.9%). Cultures transfected with H11-W51C or the empty vector alone had base-line levels of apoptosis (8.0 Ϯ 1.2 and 6.0 Ϯ 0.9% TUNEL ϩ cells, respectively) (Fig. 10). This reduction in H11triggered apoptosis (95 Ϯ 2.5%) suggests that virtually all of the cells that received H11 were simultaneously transfected with the antiapoptotic H11-W51C, a conclusion consistent with present understanding of Fugene function. H11-W51C also overrides apoptosis triggered by Aza-C treatment. In these experiments, SK-MEL-2 cultures were transfected with H11-W51C or empty vector (pFLAG) and treated with Aza-C (2 M) for 3 days. Transfection efficiency, confirmed by staining with H11 or FLAG antibody (and expressed as percentage of staining cells in transfected cultures Ϫ percentage of staining cells in untransfected cultures), was 35-40%. Consistent with this transfection efficiency, the percentage of TUNEL ϩ cells in cultures transfected with H11-W51C and treated with Aza-C was 37% lower than in Aza-C-treated cultures transfected with the empty vector. Only base-line levels of apoptosis were seen in transfected cultures that did not receive Aza-C.
H11-W51C Has Higher Kinase Levels than WT H11-Because H11 and H11-W51C trigger distinct signaling pathways, we wanted to know whether this is associated with different levels of autokinase activity (18). Immunocomplex PK/immunoblotting assays were done as previously described (18) with extracts from overexpressing cells (TAG51 (H11-W51C)-and Aza-C-treated SK-MEL-2 (WT H11)) and low expressor cells (HEK293 stably transfected with the empty vector pCI and untreated SK-MEL-2). The levels of phospho-rylated H11-W51C were 5-6-fold higher than those of phosphorylated H11, as determined by densitometric scanning of the bands immunoprecipitated with H11-181 antibody (PK). Protein levels estimated by immunoblotting with antibody H11-10 were similar for both the SK-MEL-2 and TAG51 immunocomplexes, and H11 was barely detectable in HEK293 cells (Fig. 11). The data suggest that H11-W51C has higher autokinase activity than H11, but the role of the PK activity in life-death decisions is still unclear.

DISCUSSION
Hsp are a new family of apoptosis regulatory proteins (11,12) that have antiapoptotic activity and are overexpressed in tumor tissues, where they contribute to cell proliferation, neoplastic transformation, tumor progression, and resistance to chemotherapy (10,33,34). Hsp function upstream and downstream of caspase activation and are involved in the regulation of signaling cascades. Hsp70 prevents cytochrome C release from mitochondria, blocks apoptosis-regulating kinase 1 activation, inhibits the JNK and p38MAPK pathways, and functions downstream of caspase-3 (11,13,35,36). Its cytoprotective effect is both dependent on and independent of JNK activation (11). Hsp90 is required for the maintenance of the c-Raf-1 function (15), Hsp27 sequesters cytochrome C (37), and ␣B-crystallin binds partially processed caspase-3 and inhibits its autocatalytic maturation (17). The salient feature of the studies described in this report is that H11 has proapoptotic activity, to the extent of our knowledge the first Hsp family member with such potential. The following comments seem pertinent with respect to these findings.
Our studies were stimulated by previous observations that H11, the eukaryotic homologue of a viral protein with cell regulatory potential, differs from canonical Hsp in that it is associated with the cell surface, has auto-and transphosphorylating PK activity, is silenced by cell differentiation (18,21,30), and does not translocate to the nucleus upon heat shock. 2 Multiple tissue arrays confirmed previous reports that constitutive H11 expression is tissue-restricted, with a hierarchy and distribution pattern similar to that of other members of the small Hsp subfamily (38). However, tissue arrays comparing tumor to matched normal tissues and RT-PCR with primers that flank the H11 open reading frame, indicated that H11 expression was altered (decreased or increased) in cancers of various tissue origin.
We focused on tumor cell lines SK-MEL-2 (melanoma), PC-3 (prostate cancer), and TC32 (Ewing's sarcoma family tumor) in which H11 expression was reduced relative to that seen in the normal cell counterparts, in order to examine whether H11 is involved in apoptosis regulation. Remarkably, H11 expression in these cells was not increased by heat shock but rather by treatment with the demethylating agent Aza-C, suggesting that expression was inhibited by aberrant DNA methylation. This conclusion is supported by the finding of a CpG island at the 5Ј-untranslated region, 216 bp upstream of the putative H11 transcription start site, as determined by computer-assisted analysis of the human genome data base (available on the World Wide Web at www.bioinformatics.org/sms/). Al-though loss/diminution of heat shock responsiveness due to promoter methylation was previously reported for Hsp70 and ␣B-crystallin in murine cells (39,40), this is the first report of methylation-associated Hsp transcriptional repression in human cells. However, aberrant DNA methylation is cell typespecific, as evidenced by the observation that H11 expression in HEK293 cells was increased by heat shock but not Aza-C treatment. The mechanism and kinetics of the process by which H11 acquires a new methylation pattern in some, but not other cells, and the relationship between altered methylation patterns and loss in accessibility to heat shock transcription factors are still unknown. However, in SK-MEL-2, PC-3, and TC32 cells, forced H11 expression triggered apoptosis, as determined by nuclear morphology, TUNEL, and activation of caspase-3 and p38MAPK.
We conclude that apoptosis was due to Aza-C-induced H11 overexpression, because (i) the percentage of apoptotic cells was not increased in cultures treated with Aza-C and an aODN that specifically inhibits H11 expression, (ii) the aODN effect was dose-dependent, (iii) sODN, which does not reduce H11 expression, did not alter the ability of Aza-C to increase apoptosis, (iv) apoptosis (TUNEL ϩ nuclei) was present almost exclusively in cells that overexpressed H11, and (v) the percentages of TUNEL ϩ and caspase3p20 ϩ cells were significantly increased in cultures transfected with an H11 expression vector but not the empty vector control that had a similar transfection efficiency. Apoptosis was also observed in Dox-treated TC32 cells stably transfected with a Tet-regulated H11 retrovirus, confirming the involvement of H11 in apoptosis. The percentage of apoptotic cells increased with time after H11 induction and was maximal at 7-9 days post-treatment (4 -6 days after induction of H11 expression). However, heat shockinduced H11 overexpression did not trigger apoptosis in HEK293 cells, and H11 is abundantly expressed in proliferating (basal) keratinocytes (30), suggesting that its apoptotic activity is cell type-specific.
The exact mechanism of H11-induced apoptosis is still unclear, and we do not exclude the possible contribution of distinct signaling cascades in various cell types. In SK-MEL-2 cells, apoptosis triggered by ectopically delivered H11 was partially inhibited by the pancaspase inhibitor z-VAD-fmk or the p38MAPK inhibitor SB203580 (72 Ϯ 6 and 42 Ϯ 8%, respectively). It was completely inhibited (95 Ϯ 2.5%) by treatment with both z-VAD-fmk and SB203580, suggesting that caspase-3 and p38MAPK are independent components of the apoptotic cascade. Presumably, H11 functions upstream of caspase and p38MAPK but downstream of apoptosis-regulating kinase 1, because apoptosis-regulating kinase 1 is involved in the activation of both p38MAPK and JNK (3), and apoptosis was not inhibited by the JNK-specific inhibitor SP600125. However, p38MAPK may also induce caspase activation (41), and/or it may function via a caspase that is not inhibited by z-VAD-fmk (42).
Significantly, the apoptotic activity of H11 was reversed by a single site mutation (H11-W51C). Cells that constitutively express H11-W51C (TAG51) were protected from STS-induced apoptosis involving activation of the MEK/ERK survival pathway. Indeed, the levels of activated ERK1/2 (P-ERK1/2) were higher in TAG51 than parental HEK293 cells, and both ERK activation and cytoprotection were inhibited by the MEK-specific inhibitor, U0126. The preferential activation of ERK1 (relative to ERK2) in TAG51 cells may reflect H11-W51C-initiated alteration of scaffolding proteins that exhibit selectivity for the MEK1/ERK1 isoforms (43) and is consistent with previous suggestions that there may be a physiological distinction between the two ERK isoforms (44). We conclude that B-Raf may be 2 C. C. Smith and L. Aurelian, unpublished results.
Because (i) H11-W51C has 5-6-fold higher levels of PK activity than WT H11, (ii) H11 has transphosphorylating PK activity (22), and (iii) myocardial hypertrophy triggered by WT H11 was accompanied by activation of the Akt but not the ERK survival pathway (21), it is tempting to conclude that the kinase activity is involved in H11-determined survival and death decisions. Consistent with this interpretation, ICP10 PK, the viral homologue of H11, has strong kinase activity that defines its antiapoptotic potential (9,28), and the levels of the tyrosine kinase nerve growth factor receptors and their downstream effectors determine life and death decisions by neurotrophins (46). Moreover, potential clinical relevance is suggested by the finding that H11 can be silenced by aberrant DNA methylation and could function as an apoptosis-based therapeutic with dose-and cell type-dependent specificity. However, since H11-W51C activates the ERK survival pathway and overrides apoptosis triggered by WT H11, mutations that reverse apoptotic activity may be detrimental, even if they occur in only one allele. Ongoing studies are designed to address these questions and identify the signaling pathways involved in H11-triggered apoptosis in various cell types.