Doxazosin Induces Activation of GADD153 and Cleavage of Focal

Cardiovascular Research 71 (2006) 118 – 128 www.elsevier.com/locate/cardiores

Doxazosin induces activation of GADD153 and cleavage of focal adhesion kinase in cardiomyocytes en route to apoptosis Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 Sonia Eiras, Patricia Fernańdez, Roberto Pin˜eiro, Marıá Jesu´s Iglesias, Jose´ Ramoń Gonza´lez-Juanatey, Francisca Lago *

Unidad de Investigacioń del Servicio de Cardiologıá, Hospital Clıńico Universitario, Santiago de Compostela, Spain

Received 14 October 2005; received in revised form 3 March 2006; accepted 17 March 2006 Available online 24 March 2006 Time for primary review 31 days

Abstract

Objective: The a1-adrenoreceptor blocker doxazosin, which in the ALLHAT trial was associated with a greater risk of heart failure than the diuretic chlorthalidone, induces the apoptosis of human and murine cardiomyocytes regardless of a1-adrenoreceptor blockade. We aimed to throw light on the mechanism of this process. Methods: Murine cardiomyocytes (HL-1) and primary cultures of human and neonatal rat cardiomyocytes were treated with 25 Amol/L doxazosin for between 0.5 and 48 h. cDNA microarray analysis, real-time RT-PCR, and Western blotting were performed to detect possible changes in gene expression and/or activation of proteins that could be involved in doxazosin-induced apoptosis. Results: Microarray analysis revealed changes in the expression of genes directly involved in the apoptotic end-stage of the cellular response to endoplasmic reticulum (ER) stress. Doxazosin considerably increased transcription and translation of gadd153, C/epbb, and DOC-1 in cardiomyocytes as well as translocation of GADD153 to the nucleus, phosphorylation of p38 MAPK (a GADD153 activator), and the initial phosphorylation and subsequent cleavage of focal adhesion kinase (FAK). Experiments repeated following blockade of a1-adrenoreceptors showed no alteration of the above effects of doxazosin. Conclusion: Doxazosin induces the apoptosis of cardiomyocytes via the ER pathway, with increased production of C/EBPh, GADD153 and DOC-1. Likewise it increases phosphorylation of the GADD153 activator p38 MAPK and induces first the phosphorylation, and then the cleavage, of FAK. These effects are not mediated by a1-adrenoreceptors. D 2006 European Society of Cardiology. Published by Elsevier B.V. All rights reserved.

Keywords: Apoptosis; Antihypertensive agents; Gene array analysis; Cell culture; Protein kinases

1. Introduction regression of left ventricular hypertrophy) [1], and are extensively used to alleviate the symptoms of benign Like other selective a1-adrenoceptor blockers, the well- prostatic hyperplasia (BPH). However, in the Antihyperten- known antihypertensive drugs doxazosin and prazosin lower sive and Lipid-Lowering Treatment to prevent Heart Attack blood pressure by dilating resistance and capacitance Trial (ALLHAT), the risk of congestive heart failure (CHF) arteries. They also have a number of other cardiovascularly was found to be twice as high with doxazosin as with the beneficial effects (improvement of lipid profile, fibrinolysis diuretic chlorthalidone, which forced the discontinuation of and insulin sensitivity, reduction of platelet aggregation, and the doxazosin arm of the study and suggested that doxazosin, and possibly a-blockers in general, should no longer be used as first-line antihypertensive therapy [2].A * Corresponding author. Laboratorio de Investigacioń Nu´mero 1, A´ rea de possible explanation for the ALLHAT results may be Investigacioń y Docencia, Hospital Clıńico Universitario de Santiago de Compostela, Travesıá Choupana s/n, 15706 Santiago de Compostela, Spain. pointed to by our group’s previous demonstration that Tel.: +34 981 950902; fax: +34 981 950757. doxazosin and prazosin induce apoptosis in cultured human E-mail address: [email protected] (F. Lago). and murine cardiomyocytes in a dose- and time-dependent

0008-6363/$ - see front matter D 2006 European Society of Cardiology. Published by Elsevier B.V. All rights reserved. doi:10.1016/j.cardiores.2006.03.014 S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128 119 fashion [3], since heart failure is known to involve the to phosphorylate GADD153 [14]. Finally, the observation apoptosis of cardiomyocytes [4]. Doxazosin also induces the that doxazosin-treated cardiomyocytes underwent marked apoptosis of other cell types, including prostate cancer cells detachment from the culture plate during apoptosis drew [5]. our attention to FAK, since it is known that depression of Proteins that have previously been described as possibly FAK results in loss of cell adhesion and apoptosis [20]. being actively or passively involved in mediating the Moreover, FAK is depressed by doxazosin in other cell apoptotic effects of doxazosin, include transforming growth types [7,10]. factor h (TGF-h) [6], caspase 3 and focal adhesion kinase Our results suggest that doxazosin induces the apoptosis (FAK) [7], tumor necrosis factor a and reactive oxygen of cardiomyocytes via the ER pathway, with increased species [8], ether-a-go-go-gene-related potassium channels production of C/EBPh, GADD153 and DOC-1; likewise [9], and vascular endothelial growth factor (VEGF) and increases phosphorylation of the GADD153 activator p38 Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 fibroblast growth factor 2 (FGF-2) [10]. MAPK; and induces first the phosphorylation, and then the Apoptosis can occur through either of two types of cleavage, of FAK. pathway: ‘‘extrinsic’’ pathways initiated by specific extracellular ‘‘death ligands’’, and ‘‘intrinsic’’ pathways initiated by intracellular or nonspecific extracellular stressors (e.g., in 2. Materials and methods the case of cardiomyocyte apoptosis, infectious pathogens, toxins, or morphological or functional cardiac abnormalities All products (doxazosin included) were from Sigma such as ischaemia, congenital malformations or contractile Chemical Co. (St. Louis, MO, USA) unless otherwise protein defects) [11,12]. Both types have been found to be stated. involved in myocardial pathologies [11]. The early, stress- signalling stages of the intrinsic pathways converge on the 2.1. Cells and doxazosin treatment induction of apoptogenic events in mitochondria and/or the endoplasmic reticulum (ER). Experimental and clinical HL-1 adult murine cardiomyocytes (a gift of Dr. W.C. evidence that cardiomyocytes of hypertrophic and failing Claycomb, Louisiana State University Medical Center, hearts exhibit ER stress [13], suggests that an intrinsic USA) were cultured as previously described [3]. After 24 h pathway via the ER may play a significant role in the (by which time confluence was approximately 80%), the elevated rate of cardiomyocyte apoptosis known to take medium was replaced with supplement-free ExCell 320, place in hearts with these conditions. and 12 h later the cells were used in experiments in which Our initial objective was to determine which type of non-control cells were treated with doxazosin at a pathway may be involved in doxazosin-induced cardio- concentration of 25 AM. This concentration is within the myocyte apoptosis. To this end we used a cDNA micro- range in which the drug induces cardiomyocyte apoptosis array to screen for genes with altered expression in in vitro (0.1–50 AM [3]), and is the least of the doxazosin-treated cardiomyocytes. This procedure detected concentrations that have been used in other studies of increased expression of ten genes, five of which are known the molecular mediators of its apoptotic effects [6–8].A to be directly related to ER stress [14–17] and none to concentration of 1 AM in vitro is considered [5,21] to mitochondrion-mediated apoptosis. In the work described ensure intracellular concentrations similar to those here we focused on two of the five genes related to ER achieved in vivo by therapeutic doses (in patients, serum stress: C/ebpb, which encodes for the eponymous protein doxazosin concentration reaches 0.122 AM with a single C/EBPh (CCAAT/enhancer-binding protein h), and 8 mg dose, and 0.244 AM with a 16 mg dose [22]). It gadd153, which encodes for growth-arrest- and DNA- should be noted, moreover, that in our experiments damage-inducible protein 153 [GADD153, also known as doxazosin, that is usually employed for long-term treat- CHOP (C/EBP-homologous protein), DDIT3 (DNA-dam- ment of hypertensive patients, caused apoptosis in vitro age-inducible transcript 3) or C/EBP~]. We performed within just an hour. experiments to confirm the increased transcription and Primary cultures of neonatal rat cardiomyocytes and translation of these two genes, and we also explored the human cardiomyocytes (the latter obtained from fragments effects of doxazosin on events downstream of GADD153 of right atrial appendage following the informed consent of and C/EBPh. Specifically, we investigated its effects on the the patients) were cultured as described previously [3] and expression of downstream-of-CHOP gene 1 (DOC-1) [18]. serodeprived for 12 h before being used in experiments as DOC-1 encodes for a stress-inducible form of carbonic above. All procedures with animals were carried out in anhydrase VI that, by catalysing the hydration of CO2 to accordance with NIH guidelines (Guide for the Care and H2CO3, may increase stress and favour apoptosis by Use of Laboratory Animals, NIH Publication No. 85-23, acidifying the intracellular medium [19]. We also investi- revised 1996) and with the Declaration of Helsinki. gated the possible involvement of the stress-signalling To note, in all the experiments, after the treatment with protein mitogen-activated protein kinase p38 (hereinafter doxazosin, only cells adhering to the plate were collected ‘‘p38 MAPK’’ or simply ‘‘p38’’), which has been reported and analysed. 120 S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128

2.2. Microarray analysis (experiment versus baseline). It is computed using a one- step Tukey’s Biweight method by taking a mean of the log Genes undergoing altered expression in doxazosin- ratios of probe pair intensities across the two arrays. The treated murine cardiomyocytes were screened for using Tukey’s Biweight method gives an estimate of the amount HL-1 cells. Doxazosin treatment lasted 3 h. Total RNA was of variation in the data, exactly as standard deviation extracted from the cells using an Rneasy Total RNA measures the amount of variation for an average. A Signal Extraction Kit (Quiagen K.K, London, UK), and its integrity Log Ratio of 1 indicates an increase of the transcript level was assessed with an AGILENT 2100 Bioanalyzer (Agilent by 2 fold, and À1 indicates a decrease by 2 fold. In our Technologies, Waldbronn, Germany) using an RNA Nano microarray analysis, only fold changes of 2 or higher have LabChip (Agilent Technologies). been considered. Double-stranded complementary DNA (cDNA) was Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 synthesized from 8 Ag of total RNA using a Super- 2.3. Real-time RT-PCR Scripti double-stranded cDNA synthesis kit (Invitrogen, San Diego, CA) with oligo(dT)24 primer linked to the T7 Following two 2 h cycles of treatment with 10 IU/AL RNA polymerase promoter. Complementary RNA RNAase-free DNAase at 37 -C, reverse transcription of 1 (cRNA) containing biotinylated cytidine triphosphate Ag of RNA was performed and the abundance of gadd153, and uridine triphosphate was prepared therefrom using a Cebpb or DOC-1 in the resulting cDNA was quantified Bioarray High Yield RNA Transcript Labeling Kit (Enzo relative to gapdh by real-time polymerase chain reaction Diagnostics, Farmingdale, NY), and was then purified using a SybrGreen kit (Roche Diagnostics, Barcelona, on an RNeasy column and fragmented by incubation at Spain) and the specific primers listed in Table 1. MgCl2 94 -C for 35 min in 40 mmol/L Tris acetate (pH 8.1) concentration was 4 mmol/L for gadd153 and Cebpb, and containing 100 mmol/L potassium acetate and 30 mmol/L 2 mmol/L for DOC-1. In each amplification cycle, magnesium acetate. The integrity of cDNA, cRNA and denaturalization was performed for 10 min at 95 -C, fragmented cRNA was assessed by electrophoresis of annealing for 40 s at the temperatures listed in Table 1, samples on 1% agarose gels. and extension for 30 s at 72 -C; numbers of cycles are The expression of some 14,000 mouse genes was listed in Table 1. Baseline and threshold cycle (Ct) evaluated using Genechip MOE 430A mouse genome calculations were performed with Chromo 4 software microarrays (Affymetrix Inc., Santa Clara, CA, USA). The (MJ Research, Inc, Reno, NV, USA); Ct was defined as scans were analyzed with the Affymetrix Microarray Suite the cycle in which the fluorescence reached 10 times the 5.0 software. Intensities were scaled so that the average baseline signal, and this latter was defined as the intensity in each array was 100, and each array was fluorescence measured at the first minimum of the analyzed independently. Only genes with consistent mean fluorescence curve (identified as the first maximum of its intensities in three independent experiments were consid- second derivative). The specificity of fluorescence was ered; those with low signal intensity, high background, and shown in each case by the presence of just a single peak in high variability among experiments were ignored. The the melting curve. Fold change in gene expression relative -DDCt change values were determined using the algorithm named to controls was calculated as 2 ,whereDCt is ‘‘Signal Log Ratio’’ (Signal Log Ratio=Log2 Ratio). The [Ct(target gene)ÀCt(gapdh)] and DDCt is the DCt value Signal Log Ratio estimates the magnitude and direction of for doxazosin-treated samples minus the DCt value for change of a transcript when two arrays are compared controls.

Table 1 Primer sequences and real-time PCR conditions used to evaluate the expression of gadd153, C/ebpb and DOC-1 relative to gapdh (m = mouse, r = rat, h = human) Genes Primers T a -C Cycles Size bp gadd153 (m,r) Forward: 5VAGAGGAAGAATCAAAAACCTTCACT3V 64 40 187 Reverse: 5VACTCTGTTTCCGTTTCCTAGTTCTT3V gadd153 (h) Forward: 5VGGCAGCTGAGTCATTGCC3V 63 40 496 Reverse: 5VGCAGATTCACCATTCGGTCA3V C/ebpb (m,r) Forward: 5VGTTTCGGGACTTGATGCAAT3V 60 40 268 Reverse: 5VCGAAACGGAAAAGGTTCTCA3V DOC-1 (m) Forward: 5VGCCCTCCATGTACCTTGAAA3V 63 35 115 Reverse: 5VCCCATCAATGGTGTGTTCAG3V DOC-1 (r) Forward: 5VCAGCCAGAAGAGAAGGGATG3V 60 40 221 Reverse: 5VGGAGGTTCGGGGTGATTTAT3V gapdh (m,r,h) Forward: 5VTCCATGACAACTTTGGCATCGTGG3V 376 Reverse: 5VGTTGCTGTTGAAGTCACAGGAGAC3V a Temperature. S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128 121

Table 2 HL-1 cardiomyocyte genes with significant microarray-evaluated change in expression (>2-fold) in response to 3 h exposure to 25 Amol/L doxazosin, grouped by gene product type or function (means of three independent experiments). Positive and negative values indicate increased and decreased expression, respectively Signal Log Ratio Fold change Signal Log Ratio Fold change Stress and/or apoptotic related genes (14) Fxyd3 À1.1 À2.2 GADD153/Ddit3 1.9 3.8 Scp2 À1.0 À2.0 CEBPb 1.3 2.6 Slc6a14 À1.3 À2.6 Herpud1 1.4 2.8 Scn9a À1.7 À3.4 Cyp2c29 À1.8 À3.6 Slc5a8 À2.6 À5.2 Psmc6 À4.0 À8.0 Slco1b2 À1.8 À3.6 Edaradd À1.5 À3.0 Atp2b2 À2.5 À5.0

Dub2 À2.0 À4.0 Scn3b À1.2 À2.4 Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 Slc5a8 À2.6 À5.2 Atp À1.4 À2.8 Ros 1 À2.0 À4.0 Scamp5 À1.2 À2.4 Elavl4 À4.8 À9.6 Clca2 À1.0 À2.0 Casp12 À1.1 À2.2 Growth and differentiation factors(7) Tnfsf6 À1.2 À2.4 Gdf15 1.2 2.4 Ubce8 À3.1 À6.2 Nppb 1.0 2.0 Dnajb3 À2.8 À5.6 Lgals2 À2.7 À5.4 Enzymes (28) Eltd1 À1.3 À2.6 Plcb4 1.1 2.2 Fgf10 À1.8 À3.6 Snk 1.7 3.4 Respondin-pending À3.7 À7.4 Gcat À1.2 À2.4 Fgf22 À2.1 À4.2 Ndufs4 À1.0 À2.0 Inflammatory processes (12) Has2 À1.1 À2.2 C1r À2.2 À4.4 Siat8d À3.6 À7.2 Icos À1.6 À3.2 Gcs1 À2.7 À5.4 Magea5 À1.3 À2.6 Indo À1.9 À3.8 Ian3 À1.0 À2.0 Gsta3 À2.1 À4.2 Igk-V28 À2.6 À5.2 G6pc-rs À1.2 À2.4 Ifi205 À4.7 À9.4 Idi1 À1.1 À2.2 Tcrb-V8.2 À2.6 À5.2 Ace2 À1.0 À2.0 C8a À2.0 À4.0 Hsd3b2 À3.4 À6.8 Cd80 À3.1 À6.2 Asph À1.1 À2.2 Msr2 À1.1 À2.2 Ppap2a À3.6 À7.2 Cd38 À1.6 À3.2 PAN1B-like À2.4 À4.8 Ptger3 À3.8 À7.6 Gnb1 À3.3 À6.6 Cytoskeleton genes (34) Bdh À1.3 À2.6 Panx1 À2.1 À4.2 Dusp14 À2.2 À4.4 Calb1 À1.2 À2.4 Ywhae À1.0 À2.0 Cdh 17 À2.9 À5.8 Oxct À1.3 À2.6 Adamdec1 À1.7 À3.4 Mor2 À1.1 À2.2 Adam 18 À2.7 À5.4 Dusp19 À1.0 À2.0 Col9a3 À1.1 À2.2 Chst1 À1.4 À2.8 Klk4 À2.8 À5.6 Ces1 À2.8 À5.6 Nrln1-pending À1.4 À2.8 Sultx1-pending À1.0 À2.0 Cntnap1 À3.0 À6.0 Gsto1 À1.0 À2.0 Alx4 À2.5 À5.0 Bhmt À2.3 À4.6 Catnd2 À3.8 À7.6 Receptors and other Pcdhb9 À2.3 À4.6 ligand molecules (22) Zp3r À2.9 À5.8 Krt2-6a À2.2 À4.4 Abcc5 À1.0 À2.0 Pcdh18 À2.1 À4.2 Tnfrsf1b À1.3 À2.6 Cldn18 À1.2 À2.4 Igfbp1 À3.5 À7.0 Krtap6-1 À2.9 À5.8 Il18rap À1.4 À2.8 Krt2-10 À3.4 À6.8 F2rl2 À1.2 À2.4 Pcdha10 À2.8 À5.6 Creb1 À2.7 À5.4 Defb13 À3.8 À7.6 LOC56544 À1.6 À3.2 Clasp2 À4.8 À9.6 Bmx À2.3 À4.6 Kif2c À1.9 À3.8 Tacr2 À1.0 À2.0 Ndn À1.1 À2.2 Sstr2 À1.2 À2.4 Timp4 À1.2 À2.4 Chrm3 À2.2 À4.4 Coro1c À1.6 À3.2 Gtpbp4 À1.4 À2.8 Itgb1 À2.6 À5.2 Egfr À2.5 À5.0 Plek À3.4 À6.8 Emr4 À2.3 À4.6 Catna2 À1.2 À2.4 Kcnip2 À1.8 À3.6 Adam24 À3.2 À6.4 122 S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128

Table 2 (continued) Signal Log Ratio Fold change Signal Log Ratio Fold change Abcg4 À1.5 À3.0 Cntn1 À1.2 À2.4 Itgb1 À2.6 À5.2 Pcdhb16 À3.1 À6.2 Gene for odorant receptor MOR10 À2.0 À4.0 Cryba1 À3.1 À6.2 Odorant r S1 À2.1 À4.2 Acvrl1 À1.1 À2.2 Gnai1 À2.6 À5.2 Pcdhga11 À1.6 À3.2 Cellular transporters (14) Lasp1 À1.1 À2.2 Slc1a4 5.2 10.4 Pcdhb16 À3.1 À6.2 Cyp2c29 À1.8 3.6 Transcription activator factors (2) Ndufs4 À1.0 2.0 Atf3 1.5 3.0

Atf4 1.1 2.2 Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 RIKEN cDNA (34), Others (96)

2.4. Western blots of GADD153, C/EBPb, p38 MAPK and FAK anti-pFAK and anti-FAK (all at dilutions of 1:1000; all from Santa Cruz Biotechnology, Delaware, CA, USA), and then Briefly, whole cell extracts of HL-1 cells treated with exposed to peroxidase-conjugated goat anti-rabbit IgG or doxazosin for 0.5, 3, 6, 24 or 48 h were electrophoresed on rabbit anti-goat IgG. Immunoreactive proteins were visualized 10–12% SDS-polyacrylamide gels, electrotransferred to using an enhanced chemiluminescence detection system (ECL, PVDF membranes, reacted with rabbit polyclonal anti- Amersham Pharmacia Ltd, London, UK) and quantified by GADD153, anti-C/EBPh, anti-pp38-alpha, anti-p38-alpha, densitometry using Typhoon (Amersham Pharmacia Ltd).

Fig. 1. Panel A: Fold changes in expression of gadd153 and C/ebpb in HL-1 cells and primary cultures of neonatal rat cardiomyocytes (NRPC) treated with 25 Amol/L doxazosin for 3 h, as determined by real-time RT-PCR (means of at least 4 independent experiments; error bars indicate SEMs). Statistical significance with respect to controls: **, p <0.01; ***, p <0.001. Panel B: Expression of gadd153 in primary cultures of human cardiomyocytes treated with 25 Amol/L doxazosin for 3 h, as determined by RT-PCR. Panel C: Upper panel: Western blots showing increased levels of GADD153 and C/EBPh in a typical experiment in which HL-1 cardiomyocytes were treated with 25 Amol/L doxazosin for 3 h. Lower panel: mean optical density (in arbitrary units) in three independent experiments of this kind; error bars indicate SEMs. *, p <0.05. S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128 123

endoplasmic reticulum) might be most involved in doxazosin-induced cardiomyocyte apoptosis, we used microarray analysis. Table 2 shows the functional classification and fold change in expression of the mouse genes shown by microarray analysis to have altered expression in HL-1 cardiomyocytes after doxazosin treatment. Of the ten genes with increased expression, five are related to ER stress: gadd153, C/ebpb, herpud1, atf3 and atf4. Those chosen for confirmatory RT-PCR experiments and further investigation were gadd153 and C/ebpb. We could not detect with the microarray analysis changes in the gene expression of Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 proteins involved in the mitochondrial death pathway after 3 h of doxazosin treatment of HL-1 cardiomyocytes.

3.2. Induction of C/ebpb and gadd153

To corroborate the doxazosin-induced increase in gadd153 and C/ebpb expression indicated by the microarray analysis, real-time RT-PCR experiments were performed using HL-1 cells and primary cultures of cardiomyocytes. Real-time PCR confirmed that 3 h treatment of HL-1 cardiomyocytes with 25 Amol/L doxazosin increased the expression of both C/ebpb [(1.7T0.01)-fold; n =4, p <0.01] Fig. 2. A: Western blots showing the effect of 0.5–24 h 25 Amol/L and gadd153 [(7.5T1.7)-fold; n =7, p <0.001] (Fig. 1A). doxazosin treatment on the levels of unphosphorylated p38a MAPK and phosphorylated p38a MAPK (pp38a) in HL-1 cardiomyocytes. B: Mean pp38a levels, in arbitrary optical density units (means of four independent experiments; error bars indicate SEMs; **, p <0.01).

2.5. GADD153 translocation

HL-1 cardiomyocytes were treated with doxazosin for 0.5, 3 or 24 h and homogenized in a lysis buffer containing 0.5% NP-40, 10 Ag/mL leupeptin, 2 Ag/mL aprotinin and (mmol/ L): HEPES buffer (pH 7.9), 10; EDTA, 1; EGTA, 1; KCl, 10; dithiothreitol, 1; and phenylmethylsulphonylfluoride, 1. Cytoplasmic and nuclear fractions were separated by centrifuging at 8000 Âg for 15 min, and 40 Ag of each fraction was subjected to Western blot with antibodies against GADD153 (1:1000) or histone H1 (as nuclear marker and gel loading control; 1:500) (Santa Cruz Biotechnology).

2.6. Statistical analysis

Data are presented as meansTSEMs of at least three independent experiments. Differences between means were evaluated by Student’s t test, a value of p <0.05 being considered statistically significant.

3. Results Fig. 3. A: Western blots of GADD153 in nuclear (n) and cytoplasmic (c) fractions of HL-1 cardiomyocytes treated with 25 Amol/L doxazosin for 3.1. Microarray analysis 0.5, 3 or 24 h, and in the corresponding controls (histone H1 was used as a nuclear marker and gel loading control). B: Dependence of nuclear GADD153 levels (in arbitrary optical density units) on time of exposure As a first approximation to determine which among the to doxazosin (means of three independent experiments; error bars indicate apoptotic death pathways (death receptors, mitochondria, SEMs; *, p <0.05). 124 S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128

Increased production of GADD153 and C/EBPh was verified by quantitation of Western blots in arbitrary optical density units (Fig. 1C): GADD153 levels increased from 0.70T0.06 to 1.40T0.10 units, and C/EBPh levels from 0.6T0.1 to 1.2T0.1 units (n =3, p <0.05). In real-time RT-PCR experiments, neonatal rat cells showed an (8.4T3.3)-fold increase in C/ebpb expression and a (5.3T1.9)-fold increase in gadd153 expression (n =6, p <0.01 in both cases; Fig. 1A); while RT-PCR showed that doxazosin also induced a marked increase in gadd153 expression in human cardiomyocytes (Fig. 1B). Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021

3.3. Phosphorylation of p38 MAPK

To determine if p38 MAPK activation was involved in doxazosin’s proapoptotic effect on cardiomyocytes, we Fig. 4. Fold change in expression of DOC-1 in HL-1 cells and primary performed Western blot analysis. Doxazosin treatment of cultures of neonatal rat cardiomyocytes (NRPC) exposed to 25 Amol/L HL-1 cells significantly increased the level of phosphory- doxazosin for 6, 24 or 48 h (HL-1) or 24 h (NRPC), as determined by real- lated p38 MAPK with respect to controls (1.3T0.1 vs. time RT-PCR (means of at least 3 independent experiments). Statistical T T T significance with respect to controls: *, p <0.05; ***, p <0.001. 0.5 0.1 after 0.5 h, 1.3 0.03 vs. 0.5 0.07 after 3 h,

Fig. 5. A: Western blots showing the effects of 0.5–48 h doxazosin treatment on phosphorylated FAK (pFAK) and total FAK levels in HL-1 cardiomyocytes. Duplicates confirm the cleavage of FAK after 24 or 48 h exposure. B: pFAK levels, in arbitrary optical density units (means of four independent experiments; error bars indicate SEMs; **, p <0.01). C: Photomicrographs (200Â) of HL-1 cardiomyocytes after 6, 24 or 48 h, with or without 25 Amol/L doxazosin treatment, showing marked detachment from the plate during doxazosin-induced apoptosis. S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128 125

1.2T0.10 vs. 0.7T0.07 after 6 h, 1.4T0.09 vs. 0.6T0.07 and the appearance of a fragment of lower molecular weight after 24 h; n =4, p <0.01 in all cases; Fig. 2). that is not present in control cells or cells treated with doxazosin for 3 h or less (Fig. 5A). 3.4. GADD153 translocation 3.7. Independence of a1-adrenoceptor blockade Western blot experiments were performed to investigate whether doxazosin might increase the ratio between nuclear To confirm that the intracellular signalling changes and cytoplasmic levels of GADD153, which would be in described above were independent of a1-adrenoceptor keeping with its carrying the stress signal into the nucleus. blockade (like the overall apoptotic effect of doxazosin GADD153 was significantly more abundant in the nucleus [3]), we repeated some experiments following 4 h exposure of doxazosin-treated HL-1 cells than in that of controls of the cells to the irreversible a-blocker phenoxybenzamine Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 (values measured in arbitrary optical density units and at a concentration of 1 Amol/L. In HL-1 cells, this exposure corrected using nuclear Histone H1 as a standard or prevented neither the doxazosin-induced cleavage of FAK loading control) after 3 h (1.2T0.09 vs. 0.7T0.04) and (Fig. 6A) nor the doxazosin-induced increase in the after 24 h (1.6T0.4 vs. 0.6T0.2) (n =3, p <0.05 in both expression of DOC-1 [(19.2T4.1)-fold in the presence of cases) (Fig. 3), thus showing that the amount of nuclear both doxazosin and phenoxybenzamine, and (16.9T9.7)- GADD153 increased progressively with the time of ex- fold in the presence of doxazosin alone (n =4, p >0.05, not posure to doxazosin. significant) (Fig. 6B).

3.5. Induction of DOC-1 by doxazosin

We performed real-time RT-PCR to determine the possible GADD153 and CEBPh-dependent induction of the novel carbonic anhydrase VI DOC-1 (recently identified as a novel stress-induced gene downstream of CHOP/ GADD153 [18]) gene expression in HL-1 cardiomyocytes treated with doxazosin. In HL-1 cells, doxazosin increased the expression of DOC-1 (2.9T0.9)-fold after 6 h (n =4; p <0.05), (23.6T4.2)-fold after 24 h (n =4; p <0.001), and (80.5T28.3)-fold after 48 h (n =4; p <0.001) (Fig. 4). In primary cultures of neonatal rat cardiomyocytes, 24 h doxazosin treatment increased DOC-1 expression (8.5T3.2)- fold (n =3, p <0.05) (Fig. 4).

3.6. Cleavage of FAK

To investigate the possible involvement of FAK in doxazosin-induced cardiomyocyte apoptosis, we analyzed, using Western blot, both FAK phosphorylation and FAK total protein expression. Both of them are relevant in order to determine the possible involvement of this kinase in doxazosin’s proapoptotic effects on cardiomyocytes because: a) attenuation of FAK activation and function results in apoptosis [20], b) caspases cleave FAK during apoptosis [23] and c) doxazosin decreases FAK levels in other cell types [7,10]. Treating HL-1 cells with doxazosin for 0.5 or 3 h increased the abundance of phosphorylated FAK (pFAK) T T relative to controls (1.1 0.05 vs. 0.8 0.04 after 30 min, Fig. 6. A: Western blots showing that 4 h prior exposure of HL-1 T T 1.1 0.06 vs. 0.8 0.06 after 3 h; n =4; p <0.01 in both cardiomyocytes to the irreversible a1-adrenoceptor blocker phenoxyben- cases), but after 6 h there was no difference between the zamine did not alter the effects of 24 h treatment with 25 Amol/L doxazosin pFAK levels of controls and doxazosin-treated cells (Fig. on FAK cleavage. B: Fold change in expression of DOC-1 induced in HL-1 5A and B). Total FAK levels at this time were unaltered by cardiomyocytes by 24 h treatment with 25 Amol/L doxazosin, with or without 4 h prior exposure to the irreversible a1-adrenoceptor blocker doxazosin treatment (Fig. 5A). After 24 h, however, the phenoxybenzamine, as determined by real-time RT-PCR (means of 4 cleavage of FAK in doxazosin-treated cells was suggested independent experiments). Statistical significance with respect to controls: by a fall in the levels of both FAK and phosphorylated FAK *, p <0.05; **, p <0.01. 126 S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128

4. Discussion cytoskeleton [29,30]. This is interesting, in that alterations of microtubules and other cytoskeletal structures have In a previous paper we reported that doxazosin induces been observed in experimental studies of heart failure apoptosis in human and murine cultured cardiomyocytes by [31], and these gene suppression effects of doxazosin are a mechanism that is independent of a-adrenoceptor block- accordingly being investigated in more detail in our ade [3]. Since apoptosis is known to occur in the early laboratory. In the remainder of the present paper we stages of myocardial dysfunction and to contribute to discuss our results on processes associated with apopto- progressive cardiomyocyte loss in heart failure [4],we sis-related genes that underwent increased expression suggested that this finding might help to explain why in following doxazosin treatment. ALLHAT [2] patients treated with doxazosin had more Five of the ten genes with a 2-fold or greater increase adverse cardiovascular events (particularly heart failure) in expression following 3 h doxazosin treatment are Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 than those treated with the diuretic chlorthalidone (which directly involved in the ER stress response, during which has no effect on the viability of isolated cardiomyocytes; they are upregulated. One, herpud1, encodes for a protein results not shown). Doxazosin also induces apoptosis or cell with an N-terminal ubiquitin-like domain located in the cycle arrest in other cell types: through mechanisms that are ER membrane, and upregulated during endoplasmic independent of its a-adrenergic activity, it halts the cell reticulum stress [17]. The other four, gadd153, C/ebpb, cycle of cells of human smooth muscle [21,24] and bladder atf3 and atf4, encode for bZip transcription factors: [25]; inhibits the adhesion and migration of, and invasion GADD153 dimerizes with C/EBPh to alter the expression by, human vascular endothelial cells [10]; and induces the of numerous genes as part of the final, apoptotic stage of apoptosis of skin fibroblasts [8], normal and cancerous the ER stress response, and its expression is regulated by prostate cells [5,26], and non-prostatic cancer cell lines [8]. ATF3 (activating transcription factor 3) and by ATF4 [14] The unexpected adrenoreceptor-independent activities of (to note, ATF4-C/EBPh heterodimers induce the expres- doxazosin may be relevant to its long-term effects in sion of gadd153 [16]). We focused our attention on the patients treated with this drug. two that lie furthest downstream in the apoptotic process, There is evidence that the molecular mechanisms and for which available information, though incomplete, mediating doxazosin-induced apoptosis in prostate cancer is most abundant: gadd153 and C/ebpb. cells involve TGF-h-1 signaling and InB induction [6], GADD153 is a small nuclear protein that is ubiqui- whereas in normal and BPH-affected prostate stroma cells it tous at very low levels in proliferating cells and is seems that the mechanism involves, not this pathway, but an massively induced by growth arrest signals and toxic increase in reactive oxygen species associated with TNF-a stress, especially ER stress. Specifically, GADD153 signaling [8]. The objective of the work described in this constitutes a major component of the final stage of the paper was to gain insight into the proapoptotic mechanisms response to ER stress, in which cell cycle arrest or triggered by doxazosin in cultured cardiomyocytes. Al- apoptosis is triggered [14]. It heterodimerizes with C/ though most of our experiments were, for convenience, EBPh and other members of the C/EBP family, and also performed on HL-1 cells (a cell line derived from murine with ATF4 and other bZip proteins, and these inter- atrial cardiomyocytes that maintains a heart-specific pheno- actions modify the balance between proapoptotic and type [27] and normally features all the cell-surface and antiapoptotic effects [16]. In particular, the GADD153-C/ intracellular components required for an a1-adrenergic EBP heterodimer does not act as a transactivator to response [28]), we also replicated some of our HL-1 results increase the expression of classical C/EBP-activated using primary cultures of neonatal rat or human cardio- genes, but does activate other target genes [14]. Impor- myocytes. tantly, GADD153 is induced in failing hearts [13], and it As a first step, microarray analysis was used to identify has been found that prolonged ER stress may contribute those cardiomyocyte genes, the expression of which was to cardiomyocyte apoptosis during the progression from significantly affected by doxazosin treatment (Table 2). In cardiac hypertrophy to heart failure [13]. In this work, our particular, we wished to determine whether the affected microarray results showing increased expression of both genes indicated that apoptosis proceeded via the extrinsic gadd153 and C/ebpb following doxazosin treatment were pathway, the intrinsic pathway, or both; and, if the intrinsic confirmed by real-time PCR analyses, and the pathway were affected, whether the mitochondrial branch, corresponding increase in translation was confirmed by the ER branch, or both were involved. We observed no Western blots. alteration of directly apoptosis-related genes that were not GADD153-mediated growth arrest and apoptosis has related to ER stress. However, we cannot exclude the also been reported to be an effect of the antitumoral possibility that non-ER apoptosis paths might be activated quinazoline gefitinib [32]. Since doxazosin and prazosin by exposure to doxazosin for longer than the 3 h used in our are also quinazoline derivatives, the present results, and the microarray experiments. fact that prazosin, like doxazosin, induces apoptosis in Among the genes found to show depressed expression, cardiomyocytes, strongly suggest that the quinazoline many, such as clasp2 and elav14, are related to the structure is central to these activities. S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128 127

Although the mechanisms by which GADD153 promotes and it appears to be via the integrin-FAK-Src-Ras pathway cell death are far from fully elucidated, Wang et al. [18] found that p38 MAPK is activated in the cardiomyocytes of hearts three target genes that they named DOCs (for downstream of with stress-induced hypertrophy [33]. However, attenuation CHOP). DOC-1 encodes for a stress-inducible form of of FAK activation and function results in loss of cell adhesion carbonic anhydrase VI that reversibly catalyses the hydration and apoptosis [20], and, vice versa, caspases (which are of CO2 to H2CO3, and which may thereby favour apoptosis generated during apoptosis) reduce functional FAK levels by by increasing stress through acidification of the intracellular cleaving FAK, which also generates FAK fragments that medium [19]. DOC-1 is upregulated by GADD153 (in compete with FAK for phosphorylation [23]. keeping with which, a binding site for GADD153-C/EBP The effects of doxazosin on FAK in the present study were heterodimers has been identified in the DOC-1 promoter) in keeping with the above observations: short exposure (30 [18], and its stress-induced expression is in fact dependent on min to 3 h) induced the phosphorylation of FAK (possibly as a Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 both GADD153 and C/EBPh [19]. Our finding in this work FAK-mediated recruitment of adaptive or anti-stress mecha- that doxazosin progressively increased DOC-1 expression in nisms) but prolonged exposure (longer than 6 h) resulted in cardiomyocytes from 3-fold after 6 h exposure to 80-fold FAK cleavage. A similar pattern has been reported in human after 2 days is therefore consistent with the observed glioblastoma cells under oxidative stress [41]. increases in GADD153 and C/EBPh after 3 h exposure. It Finally, just as the induction of apoptosis by doxazosin may be noted that increased expression of DOC-1 has been in human and murine cultured cardiomyocytes is indepen- reported to be followed by the corresponding increase in dent of a-adrenoreceptor blockade [3], we found in the translation [18]. present work that doxazosin-induced expression of DOC-1 GADD153 heterodimerizes when phosphorylated by and cleavage of FAK are also unaffected by prior activated p38 MAPK [14], which could therefore regulate irreversible blockade of a-adrenoreceptors by phenoxyben- apoptosis. In fact, there is evidence that p38 MAPK does zamine. This shows that the mechanism by which do- regulate apoptosis in cardiomyocytes [12]. Activation of xazosin induces apoptosis in cardiomyocytes does not cardiomyocyte p38 MAPK is known to occur in response to involve a-adrenoreceptors. mechanical stress [33] and to be increased in rabbits with In conclusion, we have found that doxazosin treatment of pacing-induced heart failure [34], while inhibition of p38 cardiomyocytes leads to activation of the apoptotic stage of MAPK improves cardiac function and attenuates left the ER stress response, and to initial phosphorylation and ventricular remodelling following myocardial infarction in subsequent cleavage of FAK, and that both these effects are rats [35]. In this work we found that 3 h doxazosin treatment independent of a-adrenoreceptors. Additional research is increased both p38 MAPK phosphorylation and the translo- necessary in order to clarify the possible link between the ER cation of GADD153 to the nucleus, suggesting that dox- stress response and FAK cleavage in doxazosin-induced azosin-induced phosphorylation of p38 MAPK may be apoptosis. Our findings throw light on the cardiac effects of responsible for the activation of GADD153 and its translo- doxazosin, and suggest that the doxazosin-induced apoptosis cation to the cell nucleus. However, it remains to be of cardiomyocytes may have contributed to the incidence of determined how doxazosin may induce the phosphorylation heart failure in the ALLHAT study having been higher in of p38. patients treated with doxazosin than in those treated with As noted in the Introduction, we observed that cardio- chlorthalidone. myocytes treated with doxazosin underwent marked detachment from the plate during apoptosis (Fig. 5C). More specifically, exposure for more than 6 h induced progressive Acknowledgements rounding and detachment from the fibronectin matrix. Such behaviour may be relevant to the apoptotic cardiac effects of This work was supported by the Spanish Ministry of doxazosin, because cardiomyocyte detachment from the Health through Health Research Fund (FIS) contract 03/ extracellular matrix occurs early in left ventricular overload 0115. Francisca Lago’s research is supported by the Spanish and contributes to the progression towards heart failure Ministry of Health and the Galician Health Service [36,37]. We accordingly investigated the effects of doxazosin (SERGAS) under FIS contract 99/3040. on focal adhesion kinase (FAK), since it has been reported that in prostate cancer cells [7] and human vascular References endothelial cells [10] doxazosin induces apoptosis by reducing FAK levels, and that cleavage of FAK accompanies [1] Pool JL. Effects of doxazosin on coronary heart disease risk factors in cardiomyocyte apoptosis [38]. FAK is a non-receptor the hypertensive patient. Br J Clin Pract Suppl 1994;74:8–12. tyrosine kinase that plays a key role in the transmission of [2] ALLHAT Collaborative Research Group. Major cardiovascular events signals from the extracellular matrix to the cytoplasm, and in hypertensive patients randomized to doxazosin vs chlorthalidone: the antihypertensive and lipid-lowering treatment to prevent heart hence in cell proliferation, migration and survival [20]. attack trial (ALLHAT). JAMA 2000;283:1967–75. Phosphorylation of FAK forms part of the initial response of [3] Gonza´lez-Juanatey JR, Iglesias MJ, Alcaide C, Pin˜eiro R, Lago F. cardiomyocytes subjected to acute external stress [39,40], Doxazosin induces apoptosis in cardiomyocytes cultured in vitro by a 128 S. Eiras et al. / Cardiovascular Research 71 (2006) 118–128

mechanism that is independent of a1-adrenergic blockade. Circulation [24] Hu ZW, Shi XY, Hoffman BB. Doxazosin inhibits proliferation and 2003;107:127–31. migration of human vascular smooth-muscle cells independent of [4] Kang PM, Izumo S. Apoptosis and heart failure: a critical review of alpha-1-adrenergic receptor antagonism. J Cardiovasc Pharmacol the literature. Circ Res 2000;86:1107–13. 1998;31:833–9.

[5] Benning CN, Kyprianou N. Quinazoline-derived a1-adrenoceptor [25] Austin PF, Cook BL, Niederhoff RA, Manson RA, Coplen DE, antagonists induce prostate cancer cell apoptosis via an a1-adreno- Weintraub SJ. Inhibition of mitogenic signaling and induction of ceptor-independent action. Cancer Res 2002;62:597–602. apoptosis in human bladder smooth muscle cells treated with [6] Partin JV, Anglin IE, Kyprianou N. Quinazoline-based alpha-1- doxazosin. J Urol 2004;172:1662–5. adrenoceptor antagonists induce prostate cancer cell apoptosis via [26] Chon, JK, Borkowski, A, Partin, AW, Isaacs, JT, Jacobs, SC, TGF-beta signalling and I kappa B alpha induction. Br J Cancer Kyprianou, N. Alpha-1-adrenoceptor antagonists terazosin and dox- 2003;88:1615–21. azosin induce prostate apoptosis without affecting cell proliferation [7] Walden PD, Globina Y, Nieder A. Induction of anoikis by doxazosin in patients with benign prostatic hyperplasia. J Urol 161 (11999)

in prostate cancer cells is associated with activation of caspase-3 and a 2002–2008. Downloaded from https://academic.oup.com/cardiovascres/article/71/1/118/269227 by guest on 01 October 2021 reduction of focal adhesion kinase. Urol Res 2004;32:261–5. [27] White SM, Constantin PE, Claycomb WC. Cardiac physiology at the [8] Zhao H, Lai F, Nonn L, Brooks JD, Peehl DM. Molecular targets of cellular level: use of cultured HL-1 cardiomyocytes for studies of doxazosin in human prostatic stromal cells. Prostate 2005;62:400–10. cardiac muscle cell structure and function. Am J Physiol Heart Circ [9] Thomas D, Wimmer AB, Wu K, Hammerling BC, Ficker EK, Physiol 2004;286:H823–29. Kuryshev YA, et al. Inhibition of human ether-a-go-go-related gene [28] McWhinney CD, Hansen C, Robishaw JD. a1-Adrenergic signaling in

potassium channels by a1-adrenoceptor antagonists prazosin, doxazo- a cardiac murine atrial myocyte (HL-1) cell line. Mol Cell Biochem sin, and terazosin. Naunyn Schmiedebergs Arch Pharmacol 2004; 2000;214:111–9. 369:462–72. [29] King PH, Fuller JJ, Nabors LB, Detloff PJ. Analysis of the 5V end of [10] Keledjian K, Garrison JB, Kyprianou N. Doxazosin inhibits human the mouse Elavl1 (mHuA) gene reveals a transcriptional regulatory vascular endothelial cell adhesion, migration, and invasion. J Cell element and evidence for conserved genomic organization. Gene Biochem 2005;94:374–88. 2000;242:125–31. [11] Crow MT, Mani K, Nam YJ, Kitsis RN. The mitochondrial death [30] Akhmanova A, Hoogenraad CC, Drabek K, Stepanova T, Dortland B, pathway and cardiac myocyte apoptosis. Circ Res 2004;95:957–70. Verkerk T, et al. CLASPs are CLIP-115- and -170-associating proteins [12] Baines CP, Molkentin JD. Stress signaling pathways that modulate involved in the regional regulation of microtubule dynamics in motile cardiac myocyte apoptosis. J Mol Cell Cardiol 2005;38:47–62. fibroblasts. Cell 2001;104:923–35. [13] Okada K, Minamino T, Tsukamoto Y, Liao Y, Tsukamoto O, [31] Hein S, Kostin S, Heling A, Maeno Y, Schaper J. The role of Takashima S, et al. Prolonged endoplasmic reticulum stress in cytoskeleton in heart failure. Cardiovasc Res 2000;45:273–8. hypertrophic and failing heart after aortic constriction: possible [32] Ho¨pfner M, Sutter AP, Gerst B, Zeitz M, Scheru¨bl H. A novel contribution of endoplasmic reticulum stress to cardiac myocyte approach in the treatment of neuroendocrine gastrointestinal tumours. apoptosis. Circulation 2004;110:705–12. Targeting the epidermal growth factor receptor by gefitinib (ZD 1839). [14] Oyadomari S, Mori M. Roles of CHOP/GADD153 in endoplasmic Br J Cancer 2003;89:1766–75. reticulum stress. Cell Death Differ 2004;11:381–9. [33] Aikawa R, Nagai T, Kudoh S, Zou Y, Tanaka M, Tamura M, et al. [15] Pan Y, Chen H, Siu F, Kilberg MS. Amino acid deprivation and Integrins play a critical role in mechanical stress-induced p38 MAPK endoplasmic reticulum stress induce expression of multiple Activating activation. Hypertension 2002;39:233–8. Transcription Factor-3 mRNA species that, when overexpressed in [34] Schulz R, Aker S, Belosjorow S, Konietzka I, Rauen U, Heusch G. HepG2 cells, modulate transcription by the human asparagine Stress kinase phosphorylation is increased in pacing-induced synthetase promoter. J Biol Chem 2003;278:38402–12. heart failure in rabbits. Am J Physiol Heart Circ Physiol 2003;285: [16] Kim R, Emi M, Tanabe K, Murakami S. Role of the unfolded protein H2084–90. response in cell death. Apoptosis 2006;11:5–13. [35] See F, Thomas W, Way K, Tzanidis A, Kompa A, Lewis D, et al. p38 [17] Ma Y, Hendershot LM. Herp is dually regulated by both the mitogen-activated protein kinase inhibition improves cardiac function endoplasmic reticulum stress-specific branch of the unfolded protein and attenuates left ventricular remodelling following myocardial response and a branch that is shared with other cellular stress infarction in the rat. J Am Coll Cardiol 2004;44:1679–89. pathways. J Biol Chem 2004;279:13792–9. [36] Michel JB. Anoikis in the cardiovascular system: known and unknown [18] Wang XZ, Kuroda M, Sok J, Batchvarova N, Kimmel R, Chung P, et extracellular mediators. Arterioscler Thromb Vasc Biol 2003;23: al. Identification of novel stress-induced genes downstream of CHOP. 2146–2154. EMBO J 1998;17:3619–30. [37] Ding B, Price RL, Goldsmith EC, Borg TK, Yan X, Douglas PS, et al. [19] Sok J, Wang XZ, Batchvarova N, Kuroda M, Harding H, Ron D. Left ventricular hypertrophy in ascending aortic stenosis mice: anoikis CHOP-dependent stress-inducible expression of a novel form of and the progression to early failure. Circulation 2000;101:2854–62. carbonic anhydrase VI. Mol Cell Biol 1999;19:495–504. [38] Sabri A, Alcott SG, Elourdighi H, Pak E, Derian C, Andrade-Gordon [20] Kurenova E, Xu LH, Yang X, Baldwin AS, Craven RJ, Hanks SK, P, et al. Neutrophil cathepsin G promotes detachment-induced et al. Focal adhesion kinase suppresses apoptosis by binding to the cardiomyocyte apoptosis via a protease-activated receptor-indepen- death domain of receptor-interacting protein. Mol Cell Biol dent mechanism. J Biol Chem 2003;278:23944–54. 2004;24:4361–71. [39] Takahashi N, Seko Y, Noiri E, Tobe K, Kadowaki T, Sabe H, et al. [21] Kintscher U, Wakino S, Kim S, Jackson SM, Fleck E, Hsueh WA, et Vascular endothelial growth factor induces activation and subcellular al. Doxazosin inhibits retinoblastoma protein phosphorylation and G1- translocation of focal adhesion kinase in cultured rat cardiac myocytes. S transition in human coronary smooth muscle cells. Arterioscler Circ Res 1999;84:1194–202. Thromb Vasc Biol 2000;20:1216–24. [40] Seko Y, Takahashi N, Sabe H, Tobe K, Kadowaki T, Nagai R. Hypoxia [22] Frick MH, Halttunen P, Himanen P, Huttunen M, Porsti P, Pitkajarvi T, induces activation and subcellular translocation of focal adhesion et al. A long-term double-blind comparison of doxazosin and atenolol kinase in cultured rat cardiac myocytes. Biochem Biophys Res in patients with mild to moderate essential hypertension. Br J Clin Commun 1999;262:290–6. Pharmacol 1986;21:55S–62S. [41] Sonoda Y, Kasahar T, Yokota-Aizu E, Ueno M, Watanabe S. A [23] Gervais FG, Thornberry NA, Ruffolo SC, Nicholson DW, Roy S. suppressive role of p125FAK protein tyrosine kinase in hydrogen Caspases cleave focal adhesion kinase during apoptosis to generate a peroxide-induced apoptosis of T98G cells. Biochem Biophys Res FRNK-like polypeptide. J Biol Chem 1998;273:17102–8. Commun 1997;241:769–74.