(p44) and cyclic AMP response element binding protein

Dual blockade of mitogen-activated protein kinases ERK-1 (p42) and ERK-2 (p44) and cyclic AMP response element binding protein (CREB) by neomycin inhibits glioma cell proliferation Pedro Cuevas*{, Diana Diaz-Gonza ´ lez*, Fernando Carceller* and Manuel Dujovny{ *Departamento de Investigacio ´ n, Hospital Universitario Ramo ´ n y Cajal, Universidad de Alcala ´ de Henares, Madrid, Spain { Department of Neurosurgery, Wayne State University, Detroit, MI, USA

Several growth factors and their receptors are expressed in inappropriately high abundance in gliomas and are further upregulated during the transition from low- to high-grade malignancy. In glioma cells growth factors induce expression of mitogen-activated protein kinase (MAPK) pathways. Here we report that neomycin restrained glioma cell proliferation in vitro by inhibition of p42/44 MAPK and the cyclic AMP element binding protein (CREB)-directed transcription pathways. Since alteration of gene transcription by inhibition of specic transcriptional regulatory proteins has important therapeutic potential, neomycin offers great promise for treating cancer and other diseases associated with a sustained MAPK activity [Neurol Res 2003; 25: 13–16] Keywords: Glioma; neomycin; mitogen-activated protein kinase (MAPK); cAMP response element binding protein (CREB); growth factors

INTRODUCTION The importance of altered gene expression in disease pathophysiology is now well established. Thus, modication of gene expression has emerged as an important therapeutic strategy. Agents that inhibit the transcription of disease-mediating genes or transactivate the expression of genes whose products disrupt pathophysiological processes are being developed1 ,2 . Regulation of gene expression is a complex biological process involving transcriptor factor–DNA interaction. Transcriptor factors are proteins residing in the cell nucleus or cytoplasm that, upon activation by phosphorylation, bind with specic DNA motifs in the promoter region of the target gene. A variety of growth factors and cytokines activate the mitogen-activated protein kinases (MAPK) which are a group of highly conserved and ubiquitously expressed proteins that have been shown to become activated by phosphorylation and therefore may serve to integrate input from a variety of different receptor types3 . There are at least ve subfamilies of mammalian MAPK. Of these, the p42 and p44 (also known as ERK1 and ERK2) are the most extensively studied. Phosphorylation of the ERK proteins is followed by translocation to the nucleus where they activate a number of important transcriptor factors as cyclic AMP (cAMP) response element binding protein (CREB)4 , which in turn activates genes that synthetize proteins with a key role in cell proliferation. CREB is a transcription factor that binds and activates

genes containing cAMP response element (CRE-consensus sequence: TAGCGTCA). CREB-activating activity can be activated by a number of different stimuli including mitogens, stress, and agonist that elevate cAMP. Since ERK1 and ERK2 have been implicated in mediating tumor cell proliferation, their down-regulation would appear to be a potentially useful means of reducing tumor growth5 . This hypothesis was tested in this study by attempting to down-regulate MAPK expression in glioma cell cultures. For accomplishing this hypothesis we used neomycin because this aminoglycoside antibiotic has been proved an effective antiproliferative agent in experimental glioma6 .

Correspondence and reprint requests to: Departamento de Investigacioń, Hospital Universitario Ramo ´ n y Cajal, Universidad de Alcala ´ de Henares, Madrid, Spain. [[email protected]] Accepted for publication July 2002.

Immunocytochemistry For immunocytochemical studies rat C6 glioma cells were grown in coverslides and processed according to

# 2003 Forefront Publishing Group 0161–6412/03/010013–04

MATERIALS AND METHODS Glioma cell cultures The rat cells line C6 were grown in Dulbecco’s modied Eagle’s medium (DMEM, Paisley, Scotland, UK), supplemented with 7.5% (v/v) fetal calf serum. Antibiotics [penicillin (10 units ml¡1 ) and streptomycin (10 mg ml¡1 )] were added to the medium. Cultures were incubated at 37° C in a humidied chamber with 5% CO2 . Cells were set in cover slides, and then treated with neomycin sulfate (10 mM) (Sigma Chemicals, St. Louis, MO, USA) for 1 h, an effective dose for inhibition of glioma proliferation6 .

Neurological Research, 2003, Volume 25, January 13

Neomycin inhibits MAPK pathways: Pedro Cuevas et al.

Lo´pez-Toledano7 . Briey, control and neomycin-treated cultures were xed in 4% paraformaldehyde, washed two times with PBS, permeabilized and then treated with anti-BrdU antibody (1 : 20; Dako), anti-pMAPK antibody (1 : 5,000; Sigma Chemicals, St. Louis, MO, USA), which recognized the same epitope of pERK1/2 isoforms, and anti-pCREB antibody (1 : 1000; Upstate Biotechnology, Lake Placid, NY, USA). For BrdU immunodetection cultures were treated with Cl2N for DNA denaturation, incubated in secondary antimouse antibody conjugated to biotin, and then treated with avidin–biotin–peroxidase complex (Vectastin, Santa Cruz Biotechnology, Inc., CA, USA). DAB (3-3-diaminobenzidine tetrahydrochloride; Dako) was used as chromagen. pMAPK was detected using Fast red (Sigma Chemicals) as the chromagen for alkaline phosphatase conjugated to an anti-rabbit secondary antibody. pCREB was detected using DAB as the chromagen for peroxidase conjugated with anti-mouse secondary antibody. Negative control experiments used cultures that were incubated without primary antibody and subsequently processed with secondary antibody and reacted with chromagen. These cultures demonstrated a complete absence of reaction product. Quantitation of immunocytochemical signals was performed on a Pentium computer with Metamorph software. Briey, integrate pixel intensity was measured in 17 selected elds. A constant pixel area was used for three independent measurements that were averaged. The averaged integrate pixel intensity was statistically analyzed. Image and quantitation were blind conducted. Statistical analysis Data are expressed as mean § SEM and paired means were compared with the Student’s t-test. In all cases, a probability error of < 0.05 was selected as criterion for statistical signicance. RESULTS BrdU immunolocalized to glioma cells was blocked with neomycin To identify proliferating cells in glioma cultures BrdU immunodetection was performed as described in Materials and Methods, and the effect of neomycin in BrdU immunodetection was assessed by immunoreactive intensity quantitation (Figure 1 I ). BrdU labeled nuclear proles were observed in 48.63% of glioma cells (Figure 1A, top), and neomycin treatment signicantly (p < 0.0001) reduced the number of BrdU positive nuclei (8.9%) (Figure 1D, top). Since BrdU immunoreactivity is an indicative mean for detecting proliferating cells, these morphological ndings corroborate the antiproliferative role of neomycin in glioma cells as previously reported6 . Having demonstrated a potent inhibitory effect of neomycin in glioma cells, next we investigated whether neomycin participates in the regulation of MAPK and CREB signaling pathways. For these purposes we analyzed the immunoreactivy of ERK1/2 MAPK isoforms 14 Neurological Research, 2003, Volume 25, January

and CREB immunoreactivity in control and in neomycintreated cultures. Phosphorylated ERK isoforms immunolocalized to glioma cells were blocked with neomycin To identify the localization of active ERK isoforms in glioma cells, cultures were examined for immunoreactivity with an antibody directed to an ERK-derived phosphopeptide that reacts specically with the phosphorylated (activated) form of ERK as described in Materials and Methods. Activated ERK was localized in the cytoplasm of glioma cells in control cultures (Figure 1, middle), and was blocked with neomycin (Figure 1E, middle). As Figure 1 II shows, neomycin decreases fourfold pERK1/2 immunoreactivity when compared with pERK1/2 immunodetection in control cultures. The uniform immunostaining of the phosphorylated ERK observed in control cultures suggested that activation of ERK isoforms in glioma cells was related to their high rate of proliferation. Blocking MAPK activity with neomycin treatment eliminated phosphorylation of ERK as measured by image intensity analysis (Figure 1 II ). Phosphorylated CREB immunolocalized to glioma cells was blocked with neomycin CREB phosphorylation was monitored with an antibody specic for CREB phosphorylated at Ser-133 using an immunocytochemical protocol as described in Materials and Methods. As Figure 1 (bottom) shows, phospho-CREB immunoreactivity was depicted only in the nucleus of glioma cell control cultures. Neomycin produced a signicantly decreased phospho-CREB immunoreactivity in the nucleoplasm of glioma cells (Figure 1F, bottom). Quantitation of immunostaining intensity showed that neomycin treatment decrease three-fold phospho-CREB immunoreactivity in comparison to glioma control cultures (Figure 1 III). Furthermore, neomycin treatment produced changes in glioma cell phenotype from a bipolar spindle-like morphology to a large polygonal appearance (compare left lane to right lane in Figure 1). The mechanistic basis of these morphological changes need further study. DISCUSSION Previous studies have demonstrated that the MAPK system is instrumental in transducing the growth factorinduced proliferative response8 , and also is involved in the intracellular signaling response to several growth factors which have been implicated in glioma cell proliferation9 . Binding of the appropriate ligand to the cell surface receptor leads to activation of the ras/raf system which in turn activates a kinase cascade culminating in the phosphorylation of ERK itself3 . This activated form of ERK translocates to the nucleus where it leads to progression through the cell cycle via the activation of nuclear proteins, as CREB which activates genes that synthetize proteins implicated in cell proliferation and cell survival by confering protection against cell death10 ,1 1 . Experimental evidence supports an important role of MAPK and CREB pathways in


Figure 1: Effect of neomycin in BrdU, pERK1/2 and pCREB immunoreactivities in rat glioma cells. Left lane, control cultures; right lane, neomycintreated cultures. Neomycin treatment (10 mM, 60 min) blocks immunoreactivityof BrdU (compare A and D), pERK1/2 (compare B and E) and pCREB (compare C and F). Arrows indicate scattered immunoreactive cells in neomycin-treated cultures. Original magnication £100. Experiments were performed four times and done in triplicate. Quantitation of BrdU (I), pERK1/2 (II) and pCREB (III) immunodetection image analysis in control and neomycin-treatedcultures (10 mM, 60 min) as described in Material and Methods. No positive immunoreactivitywas observed in cultures that lacked the primary antibody but were otherwise processed identically to experimental cultures

Neurological Research, 2003, Volume 25, January 15


tumorigenesis due to blockade of p42/44 MAPK and CREB signals inhibit tumor cell proliferation9,10 . In this report, we focussed on the role of neomycin in MAPK pathways inhibition in glioma cell cultures since MAPK activity has been linked to cell proliferation and neomycin abolished glioma cell proliferation in a dose-dependent manner6,12 . Several lines of evidence suggest that growth factors are required for MAPK activation in tumorigenesis. Binding of FGF or VEGF to their high afnity receptors induces rapid phosphorylation of ERK2 in vivo1 3 , and several endogenous angiogenesis inhibitors like the 16-Ka fragment of prolactin or as angiostatin can block FGF-induced phosphorylation of ERK isoforms p42/p44 in vitro14 ,15 . We therefore investigated whether this important signaling event in glioma cell proliferation was affected by neomycin. Glioma cells depicted a sustained MAPK and CREB activity that was correlated with their high rate of proliferation as evidenced by our immunocytochemical studies. Exposure to neomycin resulted in an immediate relative short-term inactivation of MAPK and CREB pathways in glioma cells that was correlated with a signicant decrease in cell proliferation. Previously it has been reported6 ,12 that neomycin, blocking nuclear translocation of several growth factors, inhibits angiogenesis and glioma cell proliferation. Although mechanisms regulating the translocation of growth factors into cell nuclei is not known, it is reasonable to suppose four possibilities to explain translocation phenomenon: 1. A true nuclear localization signal exists. 2. Simple diffusion. 3. Co-translocation with the receptors for growth factors. 4. Assistance from a carrier protein. Thus, antiproliferative activity of neomycin can be accomplished at least by two mechanisms: by blocking nuclear translocation of growth factor and by inhibition of MAPK signaling pathways. Our study corroborates the importance of the p42 and p44 MAPK proteins in mediating glioma cell proliferation. Thus, inhibitors that block the activity of MAPK proteins and the signaling pathways they activate may provide a useful basis for drug development to treat diseases such as cancer, atherosclerosis, and psoriasis2,8,16,17 . In this context, neomycin is unique in that it is currently in clinical use to treat bacterial infection and, hence, its toxicity to humans is already known. ACKNOWLEDGEMENTS

We thank Chantal Bourdier for editorial assistance, Concha Muela and Argentina Ferna ´ ndez-Ayerdi for technical assistance and Eulalia Baza ń

16 Neurological Research, 2003, Volume 25, January

for gift of antibodies. This work was partially supported by Fundacio ń Futuro grant (Madrid, Spain).

REFERENCES 1 Mann MJ, Dzau VJ. Therapeutic application of transcription factor decoy oligonucleotides. J Clin Invest 2000; 106: 1071–1073 2 Mann M, Whittemore AD, Donaldson MC, Belkin M, Conte MS, Polack JF, Orav EJ, Ehsan A, Dell’Acqua G, Dzau VJ. Ex vivo gene therapy of vascular bypass graft with E2F decoy: The PREVENT single-centre, randomized, controlled trial. Lancet 1999; 354: 1493–1498 3 Davis JR. The mitogen-activated protein kinase signal transduction pathway. J Biol Chem 1993; 268: 14553–14556 4 Kyriakis JM, Avruch J. Mammalian mitogen-activated protein kinase signal transduction pathways activated by stress and inammation. Physiol Res 2001; 81: 807–869 5 Hagedorn M, Zilberberg L, Lozano RM, Cuevas P, Canron X, Redondo-Horcajo M, Gimeńez-Gallego G, Bikfalvi A. A. short peptide domain of platelet factor 4 blocks angiogenic key events induced by FGF-2. FASEB J 2001; 15: 550–552 6 Cuevas P, Diaz-Gonza ´ lez D, Dujovny M. Antiproliferative effect of neomycin in glioma cells. Neurol Res 2002; 24: 389–391 7 Lo´pez-Toledano M. Generacio ´ n de ce´lulas catecolamine ´ rgicas a partir de ce´lulas tronco neurales expandidas en cultivo. Doctoral Thesis (unpublished) 2001. Facultad de Biologia, Departamento de Biologia Celular, Universidad Complutense Madrid, Spain 8 Fisher ML, Liu B, Glennon PE, Southgate KM, Sale EM, Sale GJ, Lewis MJ, Groves PH. Down regulation of the ERK1 and ERK2 mitogen activated protein kinases using antisense oligonucleotides inhibits proliferation of porcine vascular smooth muscle cells. Atherosclerosis 2001; 156: 289–295 9 Chattopadhyay N, Tfeit-Hansen J, Brown EM. PKC, p42/44 MAPK and p38 MAPK regulate hepatocyte growth factor secretion from human astrocytoma cells. Mol Brain Res 2002; 102: 73–82 10 Park YG, Nesterovat M, Agrawal S, Cho-Chung YS. Dual blockade of cyclic AMP response element (CRE) and AP-1-directed transcription by CRE-transcription factor decoy digonucleotide. Gene-specic inhibition of tumor growth. J Biol Chem 1999; 274: 1573–1580 11 Chouinard N, Valerie K, Rouabhia M, Huot J. UVB-mediated activation of p38 mitogen-activated protein kinase enhances resistance of normal human keratinocytes to apoptosis by stabilizing cytoplasmic p53. Biochem J 2002; 365: 133–145 12 Hu G-F. Neomycin inhibits the angiogenicactivity of broblast and epidermal growth factors. Biochem Biophys Res Commun 2001; 287: 870–874 13 Eliceiri BP, Klemke R, Stromblad S, Cheresh DA. Integrin alphavbeta3 requirement for sustained mitogen-activated protein kinase activity during angiogenesis. J Coll Biol 1998; 140: 1255–1263 14 D’Angelo G, Struman I, Martial J, Weiner JI. Activation of mitogenactivated protein kinases by vascular endothelial growth factor and basic broblast growth factor in capillary endothelial cells is inhibited by the antiangiogenic factor 16-kDA N-terminal fragment of prolactin. Proc Natl Acad Sci USA 1995; 92: 6374–6378 15 Redlitz A, Daum G, Sage EH. Angiostatin diminishes activation of the mitogen-activated protein kinases ERK-1 and ERK-2 in human dermal microvascular endothelial cells. J Vasc Res 1999; 36: 28–34 18 Levitzki A, Gazit A. Tyrosine kinase inhibition: An approach to drug development. Science 1995; 267: 1782–1788 17 English JM, Cobb MH. Pharmacological inhibitors of MAPK pathways. Trends Pharmacol Sci 2002; 23: 40–45