Oncogene (2011) 30, 3672–3681 & 2011 Macmillan Publishers Limited All rights reserved 0950-9232/11 www.nature.com/onc ORIGINAL ARTICLE Inhibition of heat shock protein 27 (HspB1) tumorigenic functions by peptide aptamers

B Gibert1, E Hadchity2, A Czekalla1, M-T Aloy2, P Colas3, C Rodriguez-Lafrasse2, A-P Arrigo1 and C Diaz-Latoud1

1Centre de Ge´ne´tique Mole´culaire et Cellulaire, CNRS UMR5534, Universite´ Lyon 1, Universite´ de Lyon, Lyon, France; 2Laboratoire de Radiobiologie Cellulaire et Mole´culaire, EA-3738, Faculte´ de Me´decine Lyon-Sud, Universite´ Lyon 1, Universite´ de Lyon, Oullins, France and 3CNRS USR 3151, Station Biologique, Place Georges Teissier, Roscoff, France

Human heat shock protein 27 (Hsp27, HspB1) is an anti- shock, oxidative stress or death receptor agonist apoptotic protein characterized for its tumorigenic and (Mehlen et al., 1996; Arrigo et al., 2005). Hsp27 directly metastatic properties, and now referenced as a major interferes with upstream events or key components of therapeutic target in many types of cancer. Hsp27 the apoptotic cascade like cytochrome-c release from biochemical properties rely on a structural oligomeric mitochondria, procaspase 3 or DAXX, a mediator and dynamic organization. Downregulation by small protein on Fas/Apo1-induced apoptosis (Charette and interfering RNA or inhibition with dominant-negative Landry, 2000; Pandey et al., 2000; Bruey et al., 2000a). mutant have proven their efficiency to counteract the anti- Hsp27 can modulate cell proliferation by interacting apoptotic and protective properties of Hsp27. In this with the Akt pathway, which has an important role for study, we report the isolation and characterization of the survival of many types of cancer cells (Rane et al., Hsp27-targeted molecules interfering with its structural 2003). Expression of Hsp27 is upregulated in numerous organization. Using the peptide aptamer (PA) strategy, types of tumors (particularly breast, colon, ovarian and we isolated PAs that specifically interact with Hsp27 and head and neck tumors) and promotes unfavorable not with the other members of the small heat shock protein outcome (Ciocca and Calderwood, 2005). Overexpres- family. In mammalian cell cultures, PAs expression sion of this protein is frequently associated with perturbed the dimerization and oligomerization of increased resistance to radiotherapy and to anti-cancer Hsp27, and acted as negative regulators of the anti- drugs, such as cisplatin, doxorubicin and etoposide apoptotic and cytoprotective activities of this protein. (Hansen et al., 1999; Kim et al., 2007; Zhang and Shen, Further studies analyzing SQ20B cell xenografts in 2007). Others and we have shown that targeting Hsp27 immunocompromised mice showed that PAs strongly by antisense strategy increases cancer cell death in vitro reduced tumor development through cell cycle arrest. and in vivo (Aloy et al., 2008; Matsui et al., 2009). Our data suggest that PAs could provide a potential tool Hsp27 pleiotropic functions are dependent on its to develop strategies for the discovery of Hsp27 chemical three-dimensional structure. Hsp27, like the other inhibitors. members of the small Hsp family, presents an Oncogene (2011) 30, 3672–3681; doi:10.1038/onc.2011.73; a-crystallin domain and can form different oligomeric published online 21 March 2011 structures of variable size, from 27 to 800 kDa. This phenomenon depends on the phosphorylation status of Keywords: Hsp27; HspB1; peptide aptamers; prolifera- three serine residues, mostly controlled by the MAP- tion; apoptosis; therapeutic target KAPK2, 3 (Landry et al., 1992). Every oligomeric form maydisplayadifferentrole,present various biochemical functions and interact with multiple partners. Disturbing these oligomeric structures allows to interfere with Hsp27 Introduction biochemical properties and may inhibit tumor growth. As antisense technologies are not easy to handle Heat shock protein 27 (Hsp27) belongs to the family of in vivo, the use of specific peptides that inhibit the ‘survival proteins’, which also includes other members anti-apoptotic activity of Hsp27 by interfering with of Hsp family, anti-apoptotic BH3, survivin and IAPs its biochemical properties without altering its level of family. This stress protein can interfere with a pleio- expression could lead to a new approach for anti-cancer tropic number of cell death pathways induced by heat therapies. We have previously shown that dominant- negative mutants of the protein disrupt several functions of Hsp27 (Diaz-Latoud et al., 2005). Here, we used a Correspondence: Dr C Diaz-Latoud, Centre de Ge´ne´tique Mole´culaire new approach to specifically target and inhibit Hsp27. et Cellulaire, CNRS UMR5534, Universite´Lyon 1, Universite´de Peptide aptamers (PAs) are made of short peptides of Lyon, 43, bd du 11 novembre 1918, Lyon, France E-mail: [email protected] random sequence inserted into a scaffold protein (here, Received 21 September 2010; revised 17 January 2011; accepted the thioredoxin A ( ¼ TrxA) of Escherichia coli). The 7 February 2011; published online 21 March 2011 role of the scaffold is to display a conformationally Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3673 constrained sequence of amino acids and to stabilize point mutant aB-crystallin-R120G, responsible for them in the cellular environment. PAs can specifically human myopathies and cataracts, and deletion mutant bind to and modulate the activities of a wide range of Hsp27D141-175 did not give a detectable interaction intracellular proteins, including oncogenes, transcrip- phenotype with LexA–Hsp27 (Figure 1b) (Vicart et al., tion factors, cell cycle regulators and others (Buerger 1998). However, the weak interaction phenotypes et al., 2003; Chattopadhyay et al., 2006; Nouvion et al., observed with LexA-Hsp27D141-175 were not due to a 2007). lack of expression of this construct as shown by the In this study, we characterized two aptamers that interaction of this mutant with aptamer 54 (PA54) functionally inhibit Hsp27 and sensitize cancer cells to (Figure 1d). Altogether, these results establish that the apoptosis in vitro and in vivo by disrupting the LexA–Hsp27 fusion protein is a well-behaved bait (that biochemical functions of the protein. This study is, well expressed, properly folded, localized in the confirms that Hsp27 is a therapeutic target in cancer nucleus, and without a detectable transcriptional activ- and provides promising guides for the discovery of ity) that can thus be used for a yeast two-hybrid PA chemical inhibitors of Hsp27. selection. They also show that the R120 residue and the region comprised between amino acids 141 and 175 are essential. To identify Hsp27 interacting aptamers, a yeast two- Results hybrid screening was performed using wild-type Hsp27 as bait and a mix of randomized 8- and 13-mer PA Selection of PAs interacting with Hsp27 libraries as preys. Initially, 160 PAs that showed a yeast In order to validate the use of Hsp27 in the yeast two- two-hybrid interaction phenotype with the LexA–Hsp27 hybrid method and to map crucial domains of this construct were isolated from a screen of 11 Â 109 yeast chaperone, interactions between Hsp27 and mutant colonies. The aptamer plasmids were recovered from forms of the protein were tested by a yeast two-hybrid yeast, and the interaction phenotypes were verified after approach (Figures 1a and b). Strong Hsp27 homo- retransformation and mating assays. Sequences of 50 dimerization was observed, as well as interaction of aptamers showing strong interaction phenotypes with wild-type Hsp27, with several mutant forms. In contrast, LexA–Hsp27 were determined, out of which 25 unique

1 85 171 205 Wild type Hsp27 1-8 Occu- 141-175 Aptamer PEPTIDE SEQUENCE Δ5 Δ rence Hsp27 D100G 11 Q L S G W V G R C L N I N B-crystalline B-crystalline R120G Preys Hsp27 PAc Hsp27 Hsp27 D100G Hsp27 R140G Hsp27 C137A Hsp27 α α 16 I I F Q L P M Y Hsp27 C137A Baits 17 E I L R R W V D 6 Hsp27 R140G Control vector 21 H L L R R V L A 23 Y I L R R A S R 3 51 88 Wild type Hsp27 Hsp27 Δ51-88 33 E I L R R L V C 2 141 175 36 EI LRRALH Hsp27 Δ141-175 Hsp27 D100G 37 S I L R R M A A Hsp27 R140G 39 V T G E L W F I 2 1 66 149 175 αB-crystalline 42 H L L R R V L A Hsp27 C137A 43 A L L R R L L S 2 α B-crystalline R120G 47 A L L R R L A N Hsp27 Δ141-175 50 Y L L R R L C C αB-crystalline 53 L H N E V W V V 3 54 A E C L M W V G 5 57 YI LRRLLD Aptamers 58 Y L L R K L V S 111617 21 23 33 36 37 39 42 4347 50 53 54 5758 82 84 87 89 140 166 172 82 S L L R R L L T Baits: 84 ALLRRVLN Control vector 87 HI LRRMLR 89 R I Y L L N D T Wild type Hsp27 91 GLLRRLLNQEQKH 140 Q I F Q L T D S Hsp27 D100G 166 R F Y I T R V L W A A I F 172 Y L M R R L I A Hsp27 R140G

Hsp27 C137A

Hsp27 Δ141-175

αB-crystalline αB-crystalline R120G

Figure 1 Identification of Hsp27 interacting PAs. (a) Hsp27 and mutants cloned into the two-hybrid interaction system. aB-crystallin and aB-crystallin-R120G were used as a control of aptamers specificity. (b) Interaction mating assay with the different baits and preys. PAc was used as a negative control of interaction with Hsp27. (c) Table of Hsp27 interacting aptamers, two consensus sequences were identified into several aptamers (in yellow and blue). (d) Interactions of Hsp27 and mutants with identified PAs were tested in an interaction mating assay. All PAs presented a positive interaction with wild-type Hsp27.

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3674 variable regions were identified (Figure 1c). A consensus Transfected HeLa cells were treated for 12 h with sequence I/LLRRL can be identified in 50% of the 0.2 mmol/l staurosporine, caspases activation was deter- sequenced peptides. Although the screening was per- mined and 10 PAs that are able to increase apoptosis formed on an equal mix of 8- and 13-mer PA libraries, efficiency were identified (data not shown). Among only 7 aptamers contain a 13-mer variable region and them, PA11 and PA50, displaying 13 and 8 amino acids only one of them presents the consensus sequence. These in their variable region, respectively, were selected upon 25 PAs were tested for their interaction with Hsp27 and their high specificity of interaction with wild-type Hsp27 various mutants, as shown in Figure 1d. The aptamers in yeast. Control experiments were performed to test presented different profiles and strengths of interaction whether PA expression in HeLa cells could induce a with Hsp27 wild-type and mutant forms, suggesting that stress response. This was performed by analyzing the they bind to different regions of the molecular surface of level of expression of Hsp60, 70 and 90 in cells this protein. Aptamers presenting an interaction pheno- transiently expressing PAs 11 and 50. As shown in type with aB-crystallin and/or aB-crystallin-R120G Figure 2a, no significant increase in either major Hsp were removed from further studies because of their lack was observed. of specificity toward Hsp27 (Figure 1d). A WST-1 test was performed to determine cell proliferation. In this transient expression experiment, PAs 11 and 50 induced a decrease of HeLa cell Inhibition of Hsp27 anti-apoptotic activity by PAs proliferation, whereas control aptamer and shRNA27 in HeLa cells did not (Figure 2b). To identify aptamers that are able to counteract Hsp27 Hsp27 has been reported to induce high resistance to protective activity against apoptotic treatments, we some of the drugs used in chemotherapy (Hansen et al., tested the activity of the selected aptamers in HeLa 1999; Kim et al., 2007; Zhang and Shen, 2007). To assess cells. Twenty-two aptamer-coding sequences were sub- effects of PAs on drug response, HeLa cells transfected cloned into the pCINeo vector bearing a neomycin with PA11 or PA50 were treated for 18 h with cisplatin resistance gene. HeLa cells were transfected with either or doxorubicin. As shown in Figure 2c, a 20% increase the pCINeo vectors encoding specific Hsp27 aptamers, in cell death was observed after cisplatin treatment. The three different non-specific control aptamers, or with increase was up to 50% after doxorubicin treatment of a pCINeo vector encoding an short hairpin RNA PA50 or shRNA-transfected cells. In contrast, control (shRNA) directed against Hsp27 messenger RNA. aptamer (PAc) or MsRNA27 did not induce a cell death

1 * * HeLa MsRNA27 ShRNA27 PAc PA11 PA50 * * * 80 Hsp90 0.8 Hsp70 60

Hsp60 ion Index HeLa rat * * Hsp27 0.4 Ms27 sh27 40 e of cell death HeLa

Prolife TrxAT PAc Ms27 PA11 actin sh27 PA50 20 0 PAc Percentag PA11 * 0 PA50 * 80 0.05 0.1 * * * Staurosporine concentration (μmol/L) 60 * 60

death

HeLa PAc PA11 PA50 HeLa PAc PA11 PA50 40 procaspase3 of cell * * 40 * HeLa Ms27 p17 20 20 sh27 p12 PAc Percentage actin PA11 ----++++ Staurosporine 0 0 PA50 5 10 2.5 10 Cisplatin concentration (μmol/L) Doxorubicin concentration (μmol/L) Figure 2 PAs inhibit Hsp27 anti-apoptotic functions and decrease cancerous cell survival in response to chemotherapy. PAs 11 and 50 were cloned into pCI-Neo mammalian expression vectors and transiently expressed in HeLa cells that constitutively express a high level of human Hsp27. PAc was used as a negative control and did not present any interaction with Hsp27. ShRNA27 and control MsRNA27 were used as a positive control of Hsp27-targeted induced apoptosis. (a) Samples were collected 48 h after transfection and levels of Hsp27, 60, 70 and 90 were analyzed by western blot. Expression of aptamers was revealed by PA scaffold protein detection with an anti-thioredoxin A antibody ( ¼ TrxA). (b) WST-1 test was performed every 12 h during 3 days, following transfection. Proliferation index was determined and reported for each cell line (*Po0.01). (c) Sensitivity to cell death was determined by crystal violet, as described in Materials and methods. Chemotherapeutic agents (cisplatin and doxorubicin) were used to induce cell death (*Po0.01). (d) Cell viability after 12-h staurosporine treatment was determined by crystal-violet method (*Po0.01). Samples of HeLa treated with 0.15 mmol/l staurosporine for 4 h were analyzed by western blot for caspase 3.

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3675 increase as compared with mock-transfected cells. These control cell line, and transient expression of PA11 or results indicate that PAs interfere with Hsp27 cytopro- PA50 did not modify this extent of apoptosis. This tective properties as much as does an shRNA that indicates that the increase of apoptosis induced by the reduced the expression level of Hsp27 below the aptamers results from their interaction with Hsp27, detection threshold (Figure 2a). clearly demonstrating the specificity and the ability of Studies were also performed with the kinase inhi- aptamers to block Hsp27 anti-apoptotic activities. bitor staurosporine. As observed with doxorubicin, PAs strongly increased, from 50 to 80%, staurosporine- induced cell death (Figure 2d). The effect was similar to Aptamers modify the biochemical properties of Hsp27 that induced by the shRNA targeting Hsp27. The The protection induced by Hsp27 is highly correlated processing of executioner caspase 3 is a hallmark of with its biochemical state (Bruey et al., 2000b). To apoptosis. To assess caspase 3 activation in cells determine whether PAs could modify the properties of expressing or not PAs, protein extracts were analyzed Hsp27, we first analyzed the oxidative stress-induced by immunoblot after a 4-h staurosporine treatment dimerization of the protein in non-reducing polyacryla- (Figure 2d). The high level of processed fragments p17 mide gel electrophoresis (Figure 4a). As a positive and, to a lesser extent, p12 indicated that PAs are also control, we expressed Hsp27-C137A, a dominant- able to counteract Hsp27 anti-apoptotic activity. negative mutant that is able to disrupt Hsp27 dimeriza- tion (Diaz-Latoud et al., 2005). The negative control PAc and PA11 did not induce any modification of the Determination of aptamer specificity dimeric state of the protein. In contrast, PA50 expres- To demonstrate aptamers specificity toward Hsp27, we sion strongly inhibited Hsp27 dimerization, to a much performed further experiments in Hsp27-depleted HeLa higher extent than that of Hsp27-C37A mutant. cells. For this, we used the HSh2.2 cells characterized by To further characterize perturbations of Hsp27 a constitutive 90% decrease of Hsp27, as compared with biochemical properties, we tested the oligomerization control HMs27 cells expressing a mismatch construct of pattern of Hsp27 in HeLa cells expressing PAs using size the shRNA (Figure 3a). HSh2.2 cells were transfected exclusion chromatography. For this experiment, we with PA11, PA50 or PAc, and cell death was determined used myc-tagged PAs, to enhance the detection of PAs after a staurosporine treatment (Figure 3b). Apoptosis in elution fractions. As already shown in exponentially was increased in HSh2.2 cells when compared with growing mock HeLa cells, Hsp27 was recovered in two major structures with heterogeneous native molecular masses, small oligomers (o400 kDa) and large oligo- mers (4400 kDa) (Figure 3b) (Mehlen et al., 1997). The non-interacting control aptamer, PAc, did not induce any modification of Hsp27 oligomerization profile. Hela HMs27 HSh27-2.2 In PA11-transfected cells, we observed strong modifica- Hsp27 tions in Hsp27 oligomerization pattern. Quantification

Actin of western blot pointed out that Hsp27 was redistributed in two distinct populations, with a singular elution profile when compared with the other profiles (Figure 3b, upper right panel). Furthermore, only 31.4% of Hsp27 was detectable in the small oligomers 50 (o400 kDa) (Figure 4b). In contrast, PA50 transfection did not affect Hsp27 oligomerization pattern as shown 40 by the western blot quantification profiles. These data indicate that although both PAs are able to counteract

ell death Hsp27 anti-apoptotic properties, different biochemical mechanisms may underlie this target protein inhibition. To access the above-mentioned point, we next ntage of c analyzed phosphorylation of Hsp27 using specific 20 HeLa

Perce antibodies directed against the three phosphorylated HMs27 serine residues of Hsp27 (Paul et al., 2010). Phosphor- HSh27 2.2 ylation was induced by doxorubicin or cisplatin treat- HSh27 2.2 + PAc ments (Supplementary data 1). In non-treated cells, HSh27 2.2 + PA11 transfection of aptamers did not dramatically modify HSh27 2.2 + PA50 0 phosphorylation of the three serine residues of Hsp27. Staurosporine 0.075 μmol/L It is worth noting that in non-treated cells, PA11 expres- Figure 3 PAs are specific for Hsp27. (a) HeLa cells were sion resulted in a total lack of S78 phosphorylation, constitutively transfected by a specific ShRNA. Hsp27 level was whether this is related to the oligomerization changes determined by immunoblot. (b) Aptamers were transfected in observed in gel filtration assays is not known. In mock Hsp27-depleted HeLa cells. Sensitivity to staurosporine-induced cell death was determined by crystal violet, as described in cells, S15 phosphorylation was induced only by doxor- Materials and methods. ubicin treatment, whereas S78 phosphorylation was

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3676 Column elution profiles 29 66 150 200 443 609 2000 kDa Antibody: 30 Hsp27 Mock 20

cells

PA50 PA11 PA11 PA50 PAc Hsp27 C137A Hsp27 HeLa PAc Hsp27 C137A Hsp27 HeLa Myc Hsp27 - HeLa dimers 10 PAc PA11 Hsp27 PA50 Hsp27 - PAc 0 200 400 600 800 kDa monomers cells Myc

Actin

58.3% 55.7% 57.1% 31.4% - - - - - + + + + + H2O2 Hsp27 PA11 cells

Myc

Hsp27 forms <400 kDa <400 forms Hsp27

Input IP Hsp27 IP Input IP Hsp27 IP Mock PAc PA11 PA50 Hsp27 Hsp27 cells PA50 cells Myc Myc

Apt11 Apt50

PA50 PA11 PAc Mock cells Mock Total Hsp27

Total myc

Figure 4 Functional and biochemical inhibition of Hsp27 by PAs. These experiments were performed in HeLa cells transiently expressing myc-tagged PAs. (a) Cells were treated or untreated with 1 mmol/l H2O2 to induce Hsp27 dimerization. Hsp27-C137A mutant was used as a positive control of dimerization inhibition. Western blot was performed with polyacrylamide gel electrophoresis non-reducing conditions. (b) PAs were cloned in pCMV-myc mammals expression vector. Transfected HeLa cells were lysed in the presence of 0.1% Triton X-100 and the 10 000 g soluble cytosol was applied to Sepharose CL-6B gel filtration columns, as described in Materials and methods. The presence of Hsp27 in pooled fractions eluted from the columns was detected by immunoblots. 29, 66, 150, 200, 443, 669 and 2000 indicate the apparent native size of gel filtration markers (kDa). Column elution profiles were realized by quantification of immunoblots. Quantification of Hsp27 oligomers of apparent native size o400 kDa was also determined. Total Hsp27 and Myc-PAs in samples were determined with an immunoblot on 1/100 of the loaded fraction. (c) Intracellular interactions between Hsp27 and aptamers were analyzed by co-immunoprecipitation upon expression of aptamers in HeLa cells. After immunoprecipitation from column fractions with anti-Hsp27 antibody, immunoprecipitated proteins and input cell lysates were analyzed by western blot.

induced by both treatments. In contrast, no phosphor- et al., 2008). To test the radiosensitizing effect of PAs, ylation of S82 was observed after either treatment. cell survival curves were established after irradiation of Aptamers expression did not modify cisplatin-induced SQ20B-PAs and SQ20B-PAc control cells at doses phosphorylation of S78 when compared with control varying between 1 and 8 Gy. PA 11 and 50 increased cells. In contrast, PA11 and PA50 stimulated the the clonogenic cell death after irradiation as revealed doxorubicin-induced phosphorylation of S15 and S78 by an exponential decrease in the percentage of sites, respectively. These results indicate that PA11 and surviving cells (Figure 5b). The survival fraction at PA50 modify Hsp27 phosphorylation, and therefore 2 Gy (SF2), used as an index of radiosensitivity, shifted probably alter the quaternary structure and functions of from 0.75 Gy in SQ20B-PAc cells to 0.40 and 0.51 Gy in this protein (Supplementary data 1) (Paul et al., 2010). SQ20B-PA11 and SQ20B-PA50 cell lines, respectively To confirm the interaction observed in yeast, we (Figure 5b). performed co-immunoprecipitation assays directly using To better characterize PA-induced inhibition, we fractions of sizing chromatography columns. Hsp27 was performed tumor xenograft studies by injecting PA- immunoprecipitated from the 200-kDa column fraction expressing cells (and RNAi27-expressing cells) in female containing both Hsp27 and PAs. The presence of the NUDE mice and measuring tumor volume by external aptamers was detected by immunoblots with an anti- calipers weekly after injection (Figure 5c). Each animal myc antibody (Figure 4c). This experiment confirms the was also injected with the appropriate control (Ms27 for interaction between Hsp27 and PA11 and 50. RNAi27, and PAc for PA11 and PA50) to circumvent inter-animal variability. PA11 and 50 cells showed a strongly reduced tumor growth of 80 and 50%, Inhibition of tumor growth by Hsp27 PAs respectively, as compared with PAc cells (Figure 5d). To assess the ability of PAs to inhibit tumor growth, we shRNA cells, used here as a positive control, displayed a produced constitutively expressing PA clones in the reduced tumor growth but to a lesser extent than PA- head and neck squamous cell carcinoma (HNSCC) expressing cells. These data indicate that targeting SQ20B cell line (Figure 5a). We have previously shown Hsp27 with PAs is more efficient in inhibiting Hsp27 that targeting Hsp27 by antisense approach sensitizes tumorigenic activity than downregulating its expression the radioresistant SQ20B cells to g-irradiation (Aloy with an shRNA.

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3677

1 1400 n=13 1 n=13 1200 SQ20B Ms27 SQ20B RNAi 1000 800 600 400 * SQ20B SQ20B PAc SQ20B PA50 SQ20B PA11 * * 200 GAPDH Tumor volume mm3 0 SQ20B Ms27 SQ20B RNAi27 253 46897 Trx Weeks after xenograft

2 1400 n=13 2 n=13 SQ20B-PAc 1200 SQ20B-PA11 Dose (Gy) 1000 * 800 0 2 4 6 8 * 1.000 600 400 * 200 Tumor volume mm3 0 0.100 SQ20B PAc SQ20B PA11 253 46897 Weeks after xenograft SQ20B-PAc SQ20B-PA11 3 Surviving fraction 1400 SQ20B-PAc n=13 SQ20B-PA50 3 n=13 0.010 1200 SQ20B-PA50 1000 SF2 800 SQ20-PAc 0.75 * 600 * SQ20-PA11 0.40 * 400 SQ20-PA50 0.51 200 Tumor volume mm3 0 SQ20B PAc SQ20B PA50 253 46897 Weeks after xenograft Figure 5 Targeting Hsp27 by PAs inhibits tumor growth. (a) SQ20B squamous oral carcinoma were transfected to constitutively express PAs. Expression of aptamers was detected by immunoblots revealed by an anti-TrxA antibody. (b) Cell survival after exposure of the stably transfected cell lines and their respective controls to radiation at doses varying between 1 and 6 Gy. Colonies with more than 64 cells after six cell divisions were scored. Survival curves were fitted with the linear quadratic equation (Equation (1), see Materials and methods). Table: SF2 ¼ survival fraction at 2 Gy. (c) Photographs of tumor xenografts in immunocompromised BALBc/ nude mice 9 weeks after implantation. Controls and Aptamers or shRNA expressing tumor cells were injected two by two in the inner thigh of mice (n ¼ 13 for each group of mice). (d) Tumor volume was determined by external caliper measurement once weekly after injection. Tumor volume was determined by external caliper measurements and results were expressed as mean±s.d. Statistical significance was determined by unpaired t-test, revealing *Po0.05 for comparison of control with RNAi27, PA11 or PA50 at all times.

Aptamers inhibit the proliferation of tumor xenografts tumors. As shown in Figure 6c, levels of p21-waf1, a To define the mechanism of action of PA in inducing a cyclin-dependent kinase inhibitor, were increased in reduced tumor growth inhibition, we performed im- PA11- and 50-expressing tumors. In contrast, the level munohistochemical staining of Ki67 proliferation mar- of p27, another cyclin-dependent kinase-inhibiting ker, Giemsa coloration and terminal deoxynucleotidyl protein, was not affected. To decipher this p21-waf1 transferase dUTP nick end labeling (TUNEL) apoptosis accumulation in xenografts, we performed western blot assay (Figure 6a). Photomicrographs of tumor sections analysis of p21-waf1 level in the constitutively PA- and quantification of Ki67 marker revealed that the lack expressing clones of SQ20B cells. As shown in Supple- of tumor development with PA11 and 50 was due to an mentary Figure 2b, an identical level of p21-waf1 protein alteration of proliferation process (Figures 6a and b). wasdetectedinculturedSQ20BPAccellsorSQ20BPA11 We have previously reported that a Hsp27 RNAi or 50 cells. This indicates that aptamers induce a induces a small delay in tumor progression, without modification of p21-waf1 expression that is restricted to modulation of the Ki67 marker (Hadchity et al., 2009). the in vivo model. Hence, the phenomenon is probably a Moreover, TUNEL assay on tumor section showed that consequence of the tumor environment rather than a direct PA11 and PA50 did not increase apoptosis, probably regulation by the aptamers of p21-waf1 expression. as a result of a lack of pro-apoptotic stimulation (Figure 6a). Although PAs statistically modulated proliferation in cancer cell line, our data suggest that Discussion tumor environment of xenograft may have a crucial role in inhibition of tumor development in mice with SQ20B- Hsp27 is an anti-apoptotic protein over-expressed in a PA11 and 50 cells. variety of tumors (Ciocca and Calderwood, 2005). The above findings show that PA expression induces a Because of its cytoprotective and oncogenic properties, reduced proliferation of SQ20B cells. We therefore Hsp27 has emerged as a major therapeutic target in investigated whether PAs could modulate cell cycle cancer therapy (Arrigo et al., 2007). We have developed components. For this, we analyzed by western blot the a PA approach to functionally target Hsp27 without expression level of cell cycle markers in the recessed modulating its expression.

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3678 SQ20-Ms27 SQ20-RNAi27 SQ20-PAc SQ20-PA11 SQ20-PA50

Ki67

Giemsa

TUNEL assay

60

* * PAc PA11 PA50 40 p21 waf1

p27 20 actin

Ki67 positives cells index 0 PAc PA11 PA50 Figure 6 PAs inhibit tumor cell proliferation in vivo by blocking cell cycle. (a) Detection of tumor cell proliferation using Ki67 immunostaining on tumors excised at the end of week 9 (  100 magnification). Tissue morphology of tumors assessed on paraffin- embedded tissue sections after Giemsa staining (  100). Detection of apoptosis using TUNEL staining on paraffin-embedded tumor sections (  100 magnification). (b) The proliferative index was calculated as the percentage of Ki67-positive cells counted in 50 random fields at  100 magnification for each tumor. (*Po0.05). (c) Immunoblot analysis of p21 waf1, p27 and actin in SQ20B-PAc and SQ20B-PAs xenograft tumors at the end of week 9.

In this work, we have identified PAs that are able to 17-AAG is currently used in clinical trials for its interact with Hsp27. Two of them promoted apoptosis antitumor activity. However, 17-AAG triggers the induced by different radio or chemotherapeutic treat- transcription and elevation of anti-apoptotic Hsp27, ments in HeLa or SQ20B cells to the same extent than a 70 and 90, which lead to chemoresistance of cancer cells shRNA directed against Hsp27. Furthermore, apop- (Cervantes-Gomez et al., 2009). Interestingly, targeting tosis promotion by the aptamers resulted of a specific Hsp27 by PAs did not modulate other Hsp levels, either interaction with Hsp27, as aptamer expression in Hsp27- in cells or in vivo (data not shown), thus validating depleted cells were devoid of effects. In addition, PA11 Hsp27 as a highly promising target in oncology. and 50 counteracted Hsp27 pro-survival functions Two aptamers, PA11 and PA50, (13-mer and 8-mer in vivo and strongly inhibited tumor growth. Interest- variable regions, respectively) were further characterized ingly, PAs were significantly more efficient than the because of their specific features. PA11 was the only shRNA27 to reduce tumor growth. Restricted tumor selected aptamer displaying no interaction with Hsp27- analysis suggested that cell cycle was blocked due to the D100G. PA50 presents the consensus hydrophobic upregulation of the tumor suppressor protein p21-waf1. sequence found in 50% of Hsp27 interacting aptamers This is in agreement with O’Callaghan observation in isolated in this study. PA11 and PA50 interacted HCT116 colon carcinomas cells after Hsp27 depletion specifically with Hsp27 and did not present any (O’Callaghan-Sunol et al., 2007). Interestingly, this interaction with aB-crystallin, an other member of small upregulation was restricted to the in vivo model and heat shock protein (sHsp) family sharing a strong probably due to the tumor environment. In contrast, conformational homology with Hsp27. Interestingly, p27 expression level was not affected in our study, those aptamers that did not interact with aB-crystallin although Hsp27 has been shown to promote the did not present any interaction with Hsp27-R140G ubiquitination and degradation of this cyclin-dependent mutant (except for Apt166). This mutant was designed kinase inhibitor (Parcellier et al., 2006). by homology with aB-crystallin-R120G responsible for One of the major challenges of targeting Hsps in human myopathies and cataracts (Vicart et al., 1998). cancer therapy is the induction of expression of other aB-crystallin-R120G did not present an interaction with Hsp genes, which show functional redundancy and can any of the sHsp tested in our two-hybrid assays. trigger compensatory mechanisms. The Hsp90 inhibitor Arginine 120 residue might be of importance for all

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3679 sHsp-based interactions, strongly suggesting its implica- France). Anti-caspase 3 was from Alexis Biochemical Qbiogen tion in PA binding specificity. (Illkirsch, France). The primary antibodies were detected with To determine the mechanism of action of PAs, we either anti-rabbit or anti-mouse immunoglobulin conjugated studied the structural organization of Hsp27 in PA- to horseradish peroxidase (Amersham Corp., Orsay, France). expressing cells, as anticancerous properties of this sHsp The complete cocktail of protease inhibitors was from Roche rely on its oligomeric status. Although biochemical Diagnostics (Reinach, Switzerland). mechanisms underlying Hsp27 structural dynamic changes are not clearly understood yet, it has been Yeast plasmids shown that dimerization of protein monomers is a LexA-fused bait and B42-fused prey hsp27 or aB-crystallin- crucial event to form larger oligomers. PA50 was able to coding sequences were cloned into EcoRI/XhoI sites in pEG202 or pJG4-5 plasmids, respectively. Wild-type hsp27 inhibit dimerization without disrupting Hsp27 large and D100G, C137A, R140G, D51-88 and D141-175 hsp27 oligomers, thus mimicking the dominant-negative mutant complementary DNA were PCR amplified from the Hsp27-C137A mutant endowed with the ability to pCINeo vectors already described (Gonin et al., 1999; Arrigo disrupt dimer formation and functional properties of et al., 2005), using forward primer: 50-CGGAATTCATGACC wild-type Hsp27 without abolishing large oligomers GAGCGCCGCGTCCCC-30 and reverse primer: 50-CCCTC formation (Diaz-Latoud et al., 2005). Other PAs with GAGCTTTACTTGGCGGCAGTCTCATC-30. pJK103 lex- the conserved amino-acid motif I/LLRRL tested in Aop-lacZ reporter plasmid has been already described (Estojak same conditions did not disrupt dimer formation of et al., 1995). Hsp27 (data not shown). These data suggest that the cysteine residues in the variable region of PA50 are able Yeast strains to interfere with the formation of Hsp27 disulfide Saccharomyces cerevisiae strains MB210 (MATa leu2::6Lex- bridges. PA11 did not interfere with Hsp27 dimeriza- Aop-LEU2 ade2::8LexAop-ADE2 his3 trp1 ura3) and MB226 tion, but inhibited the turnover of oligomeric Hsp27 (MATa leu2 ade2 his3 trp1 ura3) were used for the screening. forms, as the oligomeric profile of Hsp27 was greatly Strains TB50 (MATa) and EGY42 (MATa) were used to confirm interactions (Gyuris et al., 1993). modified in PA11-expressing cells. Interestingly, PAs were only detected in fractions containing small oligomeric forms of Hsp27, suggesting an interference Eukaryotic expression vectors with the dynamic structural organization of Hsp27 to Aptamer DNA sequences from screen-isolated prey plasmids were PCR amplified with specific primers (forward: 50-CC form high oligomeric forms. This was confirmed by the GGAATTCGGATGAGCGACAAGATCA-30 and reverse: 50- observation that PA50 modified doxorubicin-induced CGGGGTACCTTAGGCCAGGTTGGC-30) and cloned into phosphorylation of Hsp27. In addition, PAs were able EcoRI/KpnI sites of the pCMV-myc vector (Clontech, Saint- to inhibit Hsp27 anti-apoptotic activity at different Germain-en-Laye, France). The PAs DNA sequences were levels of conformational state. Hence, these data further EcoRI-NotI digested from pCMV-myc and subcloned confirm that sHsp functions are dependent on the in pCIneo vector (Sigma). psuperNeo vector encoding a structural organization of these proteins, as exogenous shRNA27 or a mismatch control RNA will be described molecules targeting Hsp27 monomers inhibit Hsp27- elsewhere. induced cellular protection by interfering with its conformational state. PA selection In conclusion, our work demonstrates that PAs are PA selection was performed using an improved version of the able to interact with Hsp27 and to interfere with its yeast two-hybrid interaction trap, essentially as previously biochemical properties, thus inducing apoptosis in described (Bickle et al., 2006). Yeast strain MB226a was cancerous cell line and inhibiting tumor proliferation transformed with pJK103 and pEG202-Hsp27. Yeast strain MB210a was transformed with a mix of randomized 8-mer and in vivo. PAs may be exploited to specifically target 13-mer PA libraries in pWP2 vector (Bickle et al., 2006). Hsp27 and discover new insights into Hsp27 structure- Plasmids containing the selected PAs were recovered from the function relationship. The in silico search for mole- yeast using phenol/chloroform procedure and transformed cules interfering with sHsps is a challenge, as the tri- into MH3 E. coli strain onto M9 minimal medium. Interac- dimensional structure of Hsp27 is unknown, because of tions were confirmed by an interaction mating assay after the high difficulty to obtain stable crystals of the retransformation of bait and aptamer plasmids in EGY48a oligomeric form of this protein. Therefore, modeling and TB50a strains, respectively (Finley and Brent, 1994; Bickle of PAs docking sites may represent potential tools to et al., 2006). initiate a screening that may help in the discovery of Hsp27 chemical inhibitors (Baines and Colas, 2006). Cell culture and transfections All cells were grown at 37 1C in a humidified atmosphere containing 5% CO2. HeLa and SQ20B cells were grown in Dulbecco’s modified Eagle’s medium supplemented with 10% Materials and methods heat-inactivated fetal calf serum. For transient expression, exponentially growing HeLa cells were seeded at a density of Reagents 1.5 Â 106 cells/78 cm2 one night before transfection with 7 mgof Nonidet P-40, Triton X-100, desoxycholic acid, dithioerythri- DNA, according to the Lipofectamine reagent procedure tol, staurosporine, doxorubicin, cisplatin and G418 were from (Invitrogen). Forty-eight hours after transfection, cells were Sigma (Saint-Quentin Fallavier, France). Cell culture media submitted to the different treatments. Stably transfected and complements were from Invitrogen (Cergy Pontoise, SQ20B cells were grown under G418 (600 mg/ml) selection.

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3680 Cell death and/or cell survival determination Sizing chromatography Twenty-four hours after transfection, cells were seeded in Transfected HeLa cells were washed in phosphate-buffered 3 96-well plates (7.5 Â 10 per well). Twelve hours later, cells saline, harvested and lysed in 20 mM Tris, pH 7.4; 5 mM MgCl2; were treated for 18 h with drugs. Viable cells were rinsed 20 mM NaCl; 0.1% Triton X-100; and 0.1 mM EDTA. Cell twice with phosphate-buffered saline and stained for 15 min lysates were spun at 10 000 g for 10 min and supernatant was with 0.5% cresyl-violet in 50% methanol. Afterward, plates loaded on a sepharose CL-6B column (Sigma). Fractions were were rinsed and dried before a medium containing 0.1 M analyzed by SDS–polyacrylamide gel electrophoresis and sodium citrate (pH 5.4), and 20% methanol was added to immunoblotted as described previously (Bruey et al., 2000a). solubilize the stained cells. Absorbance was read at 570 nm with a Wallac1420 Multilabel-Counter (PerkinElmer, Waltham, Tumor growth MA, USA). The percentage of cell survival was based on the A suspension of 2.5 Â 106 SQ20B-transfected cells in 200 mlof ratio of the relative absorbance of the different samples to that phosphate-buffered saline was subcutaneously inoculated in of untreated cells (Javouhey et al., 2008). the inner thigh of a 5-week-old female athymic nude mice Cell proliferation was measured using the WST-1 assay, (Charles River Laboratories, L’Arbresle, France) through a which required pre-incubation of cells in media for 4 h with the 23-gauge needle under ketamine/xylazine anesthesia. All tetrazolium salt (4-[3-(4-Iodophenyl)-2-(4-nitrophenyl)-2H-5- animal procedures were performed according to local guide- tetrazolio]-1,3-benzene disulfonate) followed by absorbance lines on animal care. Tumor volume was measured weekly and measurement at 450 nm. The percentage of cell proliferation calculated according to the formula: 0.5236 (L Â W2), where was calculated based on control absorbance (100%). (L) and (W) are, respectively, the length and width diameters. For each mice group, histological and biochemical analyses Gel electrophoresis and immunoblotting were performed on tumors taken after killing the animal at the After treatment, cells were immediately rinsed twice in ice-cold end of week 9. phosphate-buffered saline and scraped off the dish. At this point, aliquots were withdrawn for determination of protein concentration. Thereafter, cells were lysed in boiling SDS Histological analyses buffer (62.5 mM Tris–HCl, pH 6.8; 1% SDS; 0.1 M dithioery- The morphology of tumors was assessed on formalin-fixed, thritol; 10% glycerol and 0.001% bromophenol blue). Alter- paraffin-embedded tissue sections, after Giemsa staining natively, for detection of Hsp27 dimers, cells were lysed in non- (Hadchity et al., 2009). reducing SDS buffer without dithioerythritol. Cell lysates were then subjected to SDS–polyacrylamide gel electrophoresis (Laemmli, 1970), and immunoblots were performed as Conflict of interest previously described (Mehlen et al., 1993). The detection of immunoblots was performed with the ECL system (Amersham The authors declare no conflict of interest. Corp.). Autoradiographs were recorded onto X-Omat LS films (Eastman Kodak Co.).

Clonogenic assay Acknowledgements Cell survival curves following irradiation were generated by means of a standard colony-formation assay (Hadchity et al., We wish to thank D Guillet for excellent technical assistance 2009). The survival a and b variables were fitted according to and Marc Bickle for helpful discussions. This work was the linear quadratic equation (SF ¼ exp [a  DÀb  D2]) supported by a research grant from the Comite´du Rhoˆne of la (Equation (1)), where SF is survival fraction and D represents Ligue Contre le Cancer and the Re´gion Rhoˆne-Alpes and radiation doses (Hadchity et al., 2009). The SF at 2 Gy (SF2) Retina France for APA. GB was supported by a PhD training was determined as an index of radiosensitivity. grant from the La Re´gion Rhoˆne-Alpes.

References

Aloy MT, Hadchity E, Bionda C, Diaz-Latoud C, Claude L, Rousson Bruey JM, Ducasse C, Bonniaud P, Ravagnan L, Susin SA, Diaz- R et al. (2008). Protective role of Hsp27 protein against gamma Latoud C et al. (2000a). Hsp27 negatively regulates cell death by radiation-induced apoptosis and radiosensitization effects of Hsp27 interacting with cytochrome c. Nat Cell Biol 2: 645–652. gene silencing in different human tumor cells. Int J Radiat Oncol Bruey JM, Paul C, Fromentin A, Hilpert S, Arrigo AP, Solary E et al. Biol Phys 70: 543–553. (2000b). Differential regulation of HSP27 oligomerization in tumor Arrigo AP, Simon S, Gibert B, Kretz-Remy C, Nivon M, Czekalla A cells grown in vitro and in vivo. Oncogene 19: 4855–4863. et al. (2007). Hsp27 (HspB1) and alphaB-crystallin (HspB5) as Buerger C, Nagel-Wolfrum K, Kunz C, Wittig I, Butz K, Hoppe- therapeutic targets. FEBS Lett 581: 3665–3674. Seyler F et al. (2003). Sequence-specific peptide aptamers, interact- Arrigo AP, Virot S, Chaufour S, Firdaus W, Kretz-Remy C, Diaz- ing with the intracellular domain of the epidermal growth factor Latoud C. (2005). Hsp27 consolidates intracellular redox homeostasis receptor, interfere with Stat3 activation and inhibit the growth of by upholding glutathione in its reducedformandbydecreasingiron tumor cells. J Biol Chem 278: 37610–37621. intracellular levels. Antioxid Redox Signal 7: 414–422. Cervantes-Gomez F, Nimmanapalli R, Gandhi V. (2009). Transcrip- Baines IC, Colas P. (2006). Peptide aptamers as guides for small- tion inhibition of heat shock proteins: a strategy for combination of molecule drug discovery. Drug Discov Today 11: 334–341. 17-allylamino-17-demethoxygeldanamycin and actinomycin d. Bickle MB, Dusserre E, Moncorge O, Bottin H, Colas P. (2006). Cancer Res 69: 3947–3954. Selection and characterization of large collections of peptide Charette SJ, Landry J. (2000). The interaction of HSP27 with Daxx aptamers through optimized yeast two-hybrid procedures. Nat identifies a potential regulatory role of HSP27 in Fas-induced Protoc 1: 1066–1091. apoptosis. Ann N Y Acad Sci 926: 126–131.

Oncogene Peptide aptamers, inhibitors of Hsp27: new therapeutic agents B Gibert et al 3681 Chattopadhyay A, Tate SA, Beswick RW, Wagner SD, Ko Ferrigno P. oligonucleotides targeting heat shock protein-27 and HTI-286 as a (2006). A peptide aptamer to antagonize BCL-6 function. Oncogene novel strategy for high-grade bladder cancer. Mol Cancer Ther 8: 25: 2223–2233. 2402–2411. Ciocca DR, Calderwood SK. (2005). Heat shock proteins in cancer: Mehlen P, Briolay J, Smith L, Diaz-latoud C, Fabre N, Pauli D et al. diagnostic, prognostic, predictive, and treatment implications. (1993). Analysis of the resistance to heat and hydrogen peroxide Cell Stress Chaperones 10: 86–103. stresses in COS cells transiently expressing wild type or deletion Diaz-Latoud C, Buache E, Javouhey E, Arrigo AP. (2005). Substitu- mutants of the Drosophila 27-kDa heat-shock protein. Eur J tion of the unique cysteine residue of murine Hsp25 interferes with Biochem 215: 277–284. the protective activity of this stress protein through inhibition of Mehlen P, Hickey E, Weber LA, Arrigo AP. (1997). Large dimer formation. Antioxid Redox Signal 7: 436–445. unphosphorylated aggregates as the active form of hsp27 which Estojak J, Brent R, Golemis EA. (1995). Correlation of two-hybrid controls intracellular reactive oxygen species and glutathione levels affinity data with in vitro measurements. MolCellBiol15: 5820–5829. and generates a protection against TNFalpha in NIH-3T3-ras cells. Finley Jr RL, Brent R. (1994). Interaction mating reveals binary and Biochem Biophys Res Commun 241: 187–192. ternary connections between Drosophila cell cycle regulators. Proc Mehlen P, Schulze-Osthoff K, Arrigo AP. (1996). Small stress proteins Natl Acad Sci USA 91: 12980–12984. as novel regulators of apoptosis. Heat shock protein 27 blocks Fas/ Gonin S, Diaz-Latoud C, Richard MJ, Ursini MV, Imbo A, Manero F APO-1- and staurosporine-induced cell death. J Biol Chem 271: et al. (1999). p53/T-antigen complex disruption in T-antigen 16510–16514. transformed NIH3T3 fibroblasts exposed to oxidative stress: Nouvion AL, Thibaut J, Lohez OD, Venet S, Colas P, Gillet G et al. correlation with the appearance of a Fas/APO-1/CD95 dependent, (2007). Modulation of Nr-13 antideath activity by peptide aptamers. caspase independent, necrotic pathway. Oncogene 18: 8011–8023. Oncogene 26: 701–710. Gyuris J, Golemis E, Chertkov H, Brent R. (1993). Cdi1, a human G1 O’Callaghan-Sunol C, Gabai VL, Sherman MY. (2007). Hsp27 and S phase protein phosphatase that associates with Cdk2. Cell 75: modulates p53 signaling and suppresses cellular senescence. Cancer 791–803. Res 67: 11779–11788. Hadchity E, Aloy MT, Paulin C, Armandy E, Watkin E, Rousson R Pandey P, Saleh A, Nakazawa A, Kumar S, Srinivasula SM, Kumar V et al. (2009). Heat shock protein 27 as a new therapeutic target for et al. (2000). Negative regulation of cytochrome c-mediated radiation sensitization of head and neck squamous cell carcinoma. oligomerization of Apaf-1 and activation of procaspase-9 by heat Mol Ther 17: 1387–1394. shock protein 90. EMBO J 19: 4310–4322. Hansen RK, Parra I, Lemieux P, Oesterreich S, Hilsenbeck SG, Fuqua SA. Parcellier A, Brunet M, Schmitt E, Col E, Didelot C, Hammann A (1999). Hsp27 overexpression inhibits doxorubicin-induced apoptosis in et al. (2006). HSP27 favors ubiquitination and proteasomal human breast cancer cells. Breast Cancer Res Treat 56: 187–196. degradation of p27Kip1 and helps S-phase re-entry in stressed cells. Javouhey E, Gibert B, Arrigo AP, Diaz JJ, Diaz-Latoud C. (2008). FASEB J 20: 1179–1181. Protection against heat and staurosporine mediated apoptosis by Paul C, Simon S, Gibert B, Virot S, Manero F, Arrigo AP. (2010). the HSV-1 US11 protein. Virology 376: 31–41. Dynamic processes that reflect anti-apoptotic strategies set up by Kim EH, Lee HJ, Lee DH, Bae S, Soh JW, Jeoung D et al. (2007). HspB1 (Hsp27). Exp Cell Res 316: 1535–1552. Inhibition of heat shock protein 27-mediated resistance to DNA Rane MJ, Pan Y, Singh S, Powell DW, Wu R, Cummins T et al. damaging agents by a novel PKC delta-V5 heptapeptide. Cancer (2003). Heat shock protein 27 controls apoptosis by regulating Akt Res 67: 6333–6341. activation. J Biol Chem 278: 27828–27835. Laemmli UK. (1970). Cleavage of structural proteins during the Vicart P, Caron A, Guicheney P, Li Z, Prevost MC, Faure A assembly of the head of bacteriophage T4. Nature 227: 680–685. et al. (1998). A missense mutation in the alphaB-crystallin Landry J, Lambert H, Zhou M, Lavoie JN, Hickey E, Weber LA et al. chaperone gene causes a desmin-related myopathy. Nat Genet 20: (1992). Human HSP27 is phosphorylated at serines 78 and 82 by 92–95. heat shock and mitogen-activated kinases that recognize the same Zhang Y, Shen X. (2007). Heat shock protein 27 protects L929 cells amino acid motif as S6 kinase II. J Biol Chem 267: 794–803. from cisplatin-induced apoptosis by enhancing Akt activation and Matsui Y, Hadaschik BA, Fazli L, Andersen RJ, Gleave ME, abating suppression of thioredoxin reductase activity. Clin Cancer So AI. (2009). Intravesical combination treatment with antisense Res 13: 2855–2864.

Supplementary Information accompanies the paper on the Oncogene website (http://www.nature.com/onc)

Oncogene