Lefranc et al. Neurosurgery 2003.pdf

EXPERIMENTAL STUDIES

CHARACTERIZATION OF GASTRIN-INDUCED CYTOSTATIC EFFECT ON CELL PROLIFERATION IN EXPERIMENTAL MALIGNANT GLIOMAS Florence Lefranc, M.D. Department of Neurosurgery, Erasmus University Hospital, Brussels, Belgium

Niloufar Sadeghi, M.D. Department of Radiology, Erasmus University Hospital, Brussels, Belgium

Thierry Metens, Ph.D. Department of Radiology, Erasmus University Hospital, Brussels, Belgium

Jacques Brotchi, M.D., Ph.D. Department of Neurosurgery, Erasmus University Hospital, Brussels, Belgium

Isabelle Salmon, M.D., Ph.D. Department of Pathology, Erasmus University Hospital, Brussels, Belgium

Robert Kiss, Ph.D. Laboratory of Histopathology, Faculty of Medicine, Free University of Brussels, Brussels, Belgium Reprint requests: Robert Kiss, Ph.D., Laboratoire d’Histopathologie, Faculté de Médecine, Université Libre de Bruxelles, 808 route de Lennik, 1070 Brussels, Belgium. Email: [email protected] Received, May 9, 2002. Accepted, December 4, 2002.

NEUROSURGERY

OBJECTIVE: Growth patterns of astrocytic tumors can be modulated in vitro by gastrin. In this study, the influence of gastrin on the in vitro cell cycle kinetics and the in vivo growth features of three experimental malignant gliomas was investigated. METHODS: Gastrin-induced influence on overall growth was assayed in vitro by means of the 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium colorimetric assay for human U373 and rat C6 gliomas and for rat 9L gliosarcoma. Although cell cycle analyses were performed by means of computer-assisted microscope analyses of Feulgen-stained nuclei, the gastrin-induced effects of the levels of expression of cyclins D3 and E, CDK2, CDK4, CDK5, CDK7, p15, p16, E2F1, and E2F2 were assayed by means of quantitative Western blot test. The presence of ribonucleic acids for the CCKB and CCKC gastrin receptors in the U373, C6, and 9L models was assayed by means of quantitative reverse transcriptase-polymerase chain reaction, and the presence or absence of ribonucleic acids for CCKA receptor was checked by means of conventional polymerase chain reaction. The influence of gastrin was also characterized in vivo in terms of the survival periods of conventional rats orthotopically grafted with the C6 and 9L models and nude rats with the U373 model. RESULTS: Gastrin significantly decreased the overall growth rate in the rat C6 and the human U373 high-grade astrocytic tumor models with either CCKB or CCKC gastrin receptor but not in the 9L rat gliosarcoma, which had no CCKB gastrin receptor (but had CCKA receptor) and only weak amounts of CCKC receptor. This effect seems to occur via a cytostatic effect; that is, an accumulation of tumor astrocytes occurs in the G1 cell cycle phase. The cytostatic effect could relate to a gastrin-induced decrease in the amounts of the cyclin D3-CDK4 complex in both C6 and U373 glioma cells. In vivo, gastrin significantly increased the survival periods of C6 and U373 gliomabearing rats, but not of 9L gliosarcoma-bearing rats. CONCLUSION: Gastrin is able to significantly modify the growth levels of a number of experimental gliomas. KEY WORDS: cdk proteins, Cell cycle kinetics, Cyclins, Gastrin, Gliomas, Proliferation Neurosurgery 52:881-891, 2003

W

DOI: 10.1227/01.NEU.0000053366.00088.80

e were the first to ascertain that gastrin is able to significantly modify the level of proliferation of human gliomas—that is, both gliomas obtained from surgery and maintained ex vivo under typical culture conditions for organic material and commercially available glioma cell lines (7–9). Whether gastrin significantly modifies growth levels in gliomas depends on whether gastrin receptors are present or absent. Two gastrin receptors have been cloned: the CCKA recep-

www.neurosurgery-online.com

tor, which is predominant outside the central nervous system and which binds cholecystokinin (CCK) with a much greater degree of affinity than gastrin, and the CCKB receptor, which is largely dominant in the brain and which binds gastrin and CCK with similar levels of affinity (11, 27, 28, 30). Some gastrinbinding proteins have also been identified (11, 30). One of these is the so-called CCKC receptor (2). Normal astrocytes possess significant amounts of CCKB receptor (13) and could well

VOLUME 52 | NUMBER 4 | APRIL 2003 | 881

LEFRANC

ET AL.

be a major primary target for CCK-gastrin peptides in the central nervous system (23). We recently identified in human tumor astrocytic cells a new gastrin-binding protein that is different from the CCKB and CCKC receptors (25). We observed in gliomas that when gastrin receptors or gastrin-binding proteins were present, gastrin was able either to stimulate or to inhibit tumor astrocyte cell proliferation (7–9). These apparently discrepant results can be at least partly explained in terms of what is already known regarding human colon cancers: one intracellular signaling pathway targeted by gastrin concerns protein kinase A (PKA) (11, 30), which contains several isoforms of regulatory and catalytic units. The specific type of PKA regulatory unit isoforms present in human colon cancer cells directly influence the fact that gastrin either stimulates or inhibits human colon cancer cell proliferation (4, 36). Although at present we do not know the exact status of PKA activation in human tumor astrocytes, we did find approximately 70 genes overexpressed by gastrin in human glioblastoma U373 cells, among which we identified several directly involved in cell proliferation features such as prohibitin, the HR1f␤ subunit (p102), and HSJ2 associated with the pituitary tumortransforming oncogene (PTTG) (20). In the present study, we tried to describe, at least partly, how gastrin exerts its effects on cell proliferation in two experimental glioma cell lines in which we had previously observed a gastrin-induced decrease in influence on global growth. Kleihues and Cavenee (18), referring to the new World Health Organization classification, recently reviewed the genetics of all types of brain tumors, and with respect to the diffuse astrocytic group, they emphasized the point regarding mutations occurring in the genes controlling two major pathways. The first of these is TP53/MDM2/p21/p27, and the second relates to p16/p15/CDK4/CDK6/Rb. Mutations occurring in one or the other of these pathways are responsible, directly or indirectly, for the occurrence of these tumors and the progression of malignancy (18, 24). The p21WAF1/CIP1 protein both binds to and inhibits a range of cyclin-dependent kinase (CDK)-cyclin complexes, including the G1 cyclins coupled to CDK2; these complexes are highly relevant to the G1/S-phase arrest mediated by TP53 (18). The p27kip1 protein also affects cell cycle regulation in the G1/S transition by the inhibition of cyclin-CDK complexes, including cyclin D-CDK4, cyclin E-CDK2, and cyclin A-CDK2 (18). We therefore decided to investigate by means of quantitative Western blot analyses whether the levels of expression of certain cyclin-CDK complexes could be modulated by gastrin, and whether gastrin is able to induce cytostatic effects (i.e., the accumulation of tumor astrocytes into the G1 phase of their cell cycle) in the specific context of a gastrin-induced decrease in global glioma growth. Gastrin-mediated effects on global growth were investigated by means of the 3-(4,5dimethylthiazol-2-yl)-2,5-diphenyltetrazolium (MTT) colorimetric assay (7, 8); the gastrin-induced influences at cell cycle kinetics level were monitored by means of computer-assisted microscope analyses of Feulgen-stained nuclei (17, 22).

882 | VOLUME 52 | NUMBER 4 | APRIL 2003

We used human U373 and rat C6 (1, 10) glioma models and added rat 9L gliosarcoma (21, 26) as a negative control because it does not express CCKB gastrin receptor. We had previously observed that the U373 model retains many characteristics that, biologically, place it close to highly malignant human astrocytic tumors (3, 5, 6), as others did with respect to the 9L model (21). The presence or absence of ribonucleic acids (RNAs) for the CCKA receptor in the U373, C6, and 9L models was investigated by means of the polymerase chain reaction (PCR) technique, whereas the amounts of RNAs for the CCKB and CCKC receptors were analyzed by means of quantitative reverse transcriptase (RT)-PCR. We also investigated whether gastrin is able to significantly modify the level of growth of the C6, U373, and 9L tumor models implanted into the brains of rats by means of stereotactic procedures. These C6 and U373 glioma- and 9L gliosarcoma-bearing rats underwent surgery, after which gastrin was delivered into the tumor resection cavities by means of micropumps. As explained in the Discussion, protein kinase Cs (PKCs) (12) exert major roles in the intracellular signaling pathways targeted by gastrin. We therefore quantitatively analyzed the levels of expression of PKC-␣, PKC-␦, PKC-␥, PKC-⑀, PKC-␩, and PKC-␨ in 9L, C6, and U373 tumors.

MATERIALS AND METHODS Cells, Compounds, and Media The rat C6 and human U373 gliomas and the rat 9L gliosarcoma were obtained from the American Type Culture Collection (Manassas, VA). They were maintained as monolayers, as detailed elsewhere (7). The cells were cultured at 37°C in sealed (airtight) Falcon plastic dishes (Nunc, Roskilde, Denmark) containing Eagle’s minimal essential medium (Invitrogen, Carlsbad, CA) (for U373 and C6 cells) or RPMI 1640 (for 9L cells) supplemented with 5% fetal calf serum. All of the media were supplemented with a mixture of 0.6 mg/ml glutamine, 200 IU/ml penicillin, 200 IU/ml streptomycin, and 0.1 mg/ml gentamycin (all purchased from Invitrogen). The fetal calf serum was heat inactivated for 1 hour at 56°C. Gastrin (pGlu-Gly-Pro-Trp-Leu-Glu-Glu-Glu-Glu-Glu-AlaTyr-Gly-Trp-Met-Asp-Phe-NH2) was purchased from Sigma Chemical Co. (St. Louis, MO) and was further purified by means of high-pressure liquid chromatography on a Vydac C18 column (300 Å, 5 ␮m) (Vydac, Hesperia, CA). The solvent was 0.1% trifluoroacetic acid in acetonitrile-water (3:1), and the gradient was from 33% to 55% over 15 minutes.

Determination of the Levels of Expression of Six PKC Isoforms The levels of expression of six different PKC isoforms were quantitatively analyzed by means of computer-assisted fluorescence microscopy on the 9L, C6, and U373 tumor cells. The six anti-PKC antibodies were purchased from Santa Cruz Biotechnology (Boechout, Belgium) and included anti-cPKC-␣


GASTRIN MODIFIES GLIOMA CELL CYCLE KINETICS

(dilution 0.4 ␮g/ml), anti-cPKC-␥ (dilution 0.4 ␮g/ml), antinPKC-␦ (dilution 0.4 ␮g/ml), anti-nPKC-⑀ (dilution 0.4 ␮g/ ml), anti-nPKC-␩ (dilution 1 ␮g/ml), and anti-aPKC-␨ (dilution 1 ␮g/ml) antibodies. Tumor cell suspensions were seeded on glass coverslips 48 hours before the immunofluorescence staining, after which the cells were fixed with 4% formaldehyde in phosphate-buffered saline, pH 7.4, for 20 minutes. Three coverslips were available for each cell line and for each PKC isoform. To detect these PKC isoforms, the cells were permeabilized with 0.2% Triton X-100 for 5 minutes, and unspecific bounds were blocked for 20 minutes with 1% horse serum (Vector Laboratories, Burlingame, CA). The cells were then incubated for 1 hour at room temperature with the various primary and secondary antibodies. The specificity of each of the six anti-PKC antibodies was checked by means of Western blot analyses (data not shown). The level of expression of each PKC isoform (relating to the fluorescence staining intensity) was quantitatively analyzed by means of computer-assisted fluorescence microscopy with a Provis Olympus microscope (Omnilabo S.A., Antwerp, Belgium) coupled to a Megaview 2 camera (Omnilabo) feeding digitized information to a computer with an AnalySIS software (Soft Imaging System GmbH, Munster, Germany). One hundred cells were analyzed per cell line for each of the six PKC isoforms.

Determination of Overall in Vitro Growth As detailed elsewhere (7, 8), C6, 9L, and U373 cell growth was assessed by means of the modified colorimetric MTT assay (Sigma). This method of assessing of cell population growth is based on the ability of living cells to reduce the yellow product MTT (Sigma) to a blue product, formazan, by a reduction reaction occurring in the mitochondria. The number of living cells is directly proportional to the intensity of the blue, which is quantitatively measured by spectrophotometry on a DIAS microplate reader (Dynatech Laboratories, Guyancourt, France) at a 570-nm wavelength (with a reference of 630 nm). The C6, 9L, and U373 cells were incubated for 72 hours in the culture medium supplemented with either 10⫺12, 10⫺11, 10⫺10, 10⫺9, 10⫺8, 10⫺7, or 10⫺6 mol/L gastrin or left unsupplemented (control). Each experiment was conducted in sextuplicate.

Determination of in Vitro Cell Cycle Kinetics The C6 and U373 cells were incubated for 72 hours in Eagle’s minimal essential medium with or without 10⫺10, 10⫺8 or 10⫺6 mol/L gastrin. The identification of the C6 and U373 cells involved in the G1, S, G2, and M phases of the cell cycle was performed for each experimental condition by means of computer-assisted, microscopy-related analyses of the C6 Feulgen-stained cells, as described elsewhere (17, 22).

Quantitative Western Blot Analyses Ten antibodies were used, all of which were monoclonal, except the one directed against CDK4 (polyclonal, rabbit). The

NEUROSURGERY

anti-cyclin D3 (dilution 1/100), p15 (dilution 1/1000), p16 (dilution 1/1000), and E2F2 (dilution 1/500) antibodies were purchased from Neo Markers (Freemont, CA), whereas the anti-cyclin E (dilution 1/250), CDK2 (dilution 1/100), CDK4 (dilution 1/100), CDK5 (dilution 1/200), CDK7 (dilution 1/500), and E2F1 (dilution 1/1000) antibodies were purchased from Santa Cruz Biotechnology. Western blot analyses were performed as recently detailed (3). Briefly, C6 and U373 glial cell extracts were prepared by the lysis of subconfluent cells in radioimmunoprecipitation assay lysis buffer (Tris-HCl 100 mmol/L, NaCl 150 mmol/L, ethylenediamine tetra-acetic acid 1 mmol/L, Triton X-100 1%, sodium deoxycholate 0.5%, sodium dodecyl sulfate 0.1%) containing 1 mmol/L phenylmethylsulfonyl fluoride and 10 ␮g/ml aprotinin. After electrophoresis, the proteins were transferred onto a Polyscreen polyvinylidene fluoride membrane (NEN Life Science Products, Boston, MA) by tank blotting. Ten molecules were immunodetected by purified specificaffinity primary antibodies (in Tris-buffered saline containing 5% milk powder) in conjunction with secondary antibodies involving goat anti-mouse or goat anti-rabbit immunoglobulin G conjugated with horseradish peroxidase (0.2 ␮g/ml; NEN Life Science Products). The secondary antibody used for the primary antibodies specific to both the human and rat proteins (cyclins D3 and E and CDK2, CDK4, CDK5, and CDK7 proteins) was goat anti-mouse or goat anti-rabbit immunoglobulin G. For the primary antibodies specific to the rat proteins only (E2F1 and E2F2) and the primary antibody specific to the human proteins only (p15 and p16), the secondary antibody was goat anti-mouse immunoglobulin G. The control experiments included the omission of the incubation step with the primary antibodies (negative control). The levels of expression of each of the 10 proteins under study were quantitatively analyzed by means of a gel scanner and Bioprofil Image Analysis Software (Vilber Lourmat, Marne La Vallée, France).

Analysis of CCKA, CCKB, and CCKC RNA Expression PCR analyses were performed as detailed elsewhere (3). Briefly, the U373, C6, and 9L tumor cells were grown to subconfluence in plastic Falcon dishes (Nunc) and lysed by adding tripure isolation reagent (Roche, Mannheim, Germany). Total RNA was prepared according to the manufacturer’s recommendations. Before complementary DNA (cDNA) synthesis, the RNA was incubated with DNase I (1.71 IU/␮l; Roche) for 15 minutes at 37°C and purified by phenolchloroform extraction. One microgram of total RNA was used as a template for the cDNA synthesis; RT was performed for 50 minutes at 42°C in RT buffer (250 mmol/L Tris-HCl pH 8.3, 375 mmol/L KCl, 15 mmol/L MgCl2, 10 mmol/L dithiothreitol; Invitrogen), oligo(dT) (3, 4, 8, 11, 30, 32, 38) primers (25 ng/␮l; Invitrogen), 500 nmol/L deoxynucleoside triphosphate, 8 IU RNAs in ribonuclease inhibitor (Promega, Leyden, The Netherlands), and 200 IU Superscript RNase H RT (Invitrogen). The reaction was terminated by incubation for 15 minutes at 70°C. The integrity of the cDNA was confirmed by


LEFRANC

ET AL.

␤-actin-specific PCR analysis. The PCR products were analyzed by electrophoresis on a 1% agarose gel. With respect to quantitative RT-PCR, a standard curve was set up by generating a PCR product by means of external primers for each messenger RNA tested. The primers used for this are listed below. The amplified products were revealed by gel electrophoresis in 2% Tris-acetate-ethylenediamine tetraacetic acid gels and visualized by ethidium bromide staining under ultraviolet light. The purification of the PCR products was realized by means of the High Pure PCR Product Purification kit (Roche) in accordance with the manufacturer’s instructions, and by the measurement of the copy number obtained by spectrophotometry analysis at 260 nm (Beckman, Analis, Ghent, Belgium). The setting up of the standard curve for each messenger RNA under study was effected by serial dilutions (107 to 10 copies/␮l). The quantitative PCR reactions were conducted in capillaries with 20 ␮l of reaction media (LC-Fastart DNA Master SYBR Green 1; Roche), 2 ␮l of cDNA, and 0.5 ␮mol/L of both sense and antisense internal primers. The reactions were performed in a Lightcycler thermocycler (Roche). After amplification, data analysis was performed by the “fit points” algorithm of the Lightcycler quantification software; the standard curve obtained enabled the quantification of the samples to take place. Primers for CCKA, CCKB, and CCKC were designed by HYBsimulator software (Adavanced Gene Computing Technologies, Irvine, CA) (14) and were purchased from Biosource Europe S.A. (Nivelles, Belgium). The primers used for conventional PCR were as follows: for the rat (C6 and 9L tumor cells), CCKA: 5'-CTTCATGAACAAACGCTTTCG-3' sense and 5'-CCATAATTCTACAGGAGCAGAA-3' antisense; for the human (U373 cells), CCKA: 5'CTCGAAAATGAGACGCTTTT-3' sense and 5'-GTAGAGTTCCAAAGAGATAATCC-3' antisense. The primers used for quantitative RT-PCR were as follows: for the rat (C6 and 9L tumor cells), internal primers, CCKB: 5'-CTACCTGAAACAGATAGGAGT-3' sense and 5'-GAGTTCCCCTTTATAGGTAAAG-3' antisense; CCKC: 5'-CCCATATTAACTATGGAGTCAAAG-3' sense and 5'-TCTACGAACTCTGATTGTACTT-3' antisense; external primers, CCKB : 5'-GGCTAAGCTATACCACCATC-3' sense and 5'-GGATATGGGGTAGATTAGTCATG-3' antisense; CCKC: 5'-ACCAGAACCCATATTAACTATGG-3' sense and 5'-TCTACGAACTCTGATTGTACTT-3' antisense; for the human (U373 cells), internal primers: CCKB: 5'-TCATTCACTTGCTGAGCTAC-3' sense and 5'-CAGTGTCATGTTTCTATGGG-3' antisense; CCKC: 5'-GAAGTAGAAGCGGTGATTCC-3' sense and 5'-CAGACTCGCTAAAATACTATCCA-3' antisense; external primers, CCKB: 5'-TCATTCACTTG CTGAGCTAC-3' sense and 5'-CAGTGTCATGTTTCTATGGG-3' antisense; CCKC: 5'-GAAGTAGAAGCGGTGATTCC-3' sense and 5'-CAGACTCGCTAAAATACTATCCA-3' antisense.

In Vivo Survival Analyses Sixty male 7- to 8-week-old, 200- to 220-g rats (OFA Sprague-Dawley Ico:OFA-SD [IOPS Caw], IFFA Credo, Brussels, Belgium) had 100,000 C6 cells stereotactically grafted into


their brains, and 45 male 200- to 240-g, 8-week-old rats (Fisher 344F; IFFA Credo) received 40,000 9L cells by means of stereotactic procedures. Three groups of 20 C6 glioma- and three groups of 15 9L gliosarcoma-grafted rats were set up 2 days after the stereotactic implantation of the tumor cells into the rats’ brains. In the same way, 21 female 8-week-old 150-g nude rats (Hsd:RH-nu; Harland, Horst, The Netherlands) had 106 human U373 cells stereotactically grafted into their brains. Three groups of seven rats were set up 2 days after the stereotactic implantation of U373 tumor cells. The weekly monitoring of the weight of each rat indicated that 51 of the 60 C6 glioma-implanted rats began to lose weight 19 to 24 days after C6 tumor cell implantation. We did not obtain 100% success with the C6 tumor grafting because of the immunogenic host-graft response (26). In the same way, 39 of the 45 9L gliosarcoma-implanted rats began to lose weight 17 to 21 days after 9L tumor cell implantation, as did 20 of the 21 U373 xenografted rats between 52 and 96 days after U373 tumor cell implantation. Nuclear magnetic resonance (NMR) imaging studies performed on each of the 51 C6 and the 20 U373 glioma-bearing rats and the 39 9L gliosarcoma-bearing rats, all of which had begun to lose weight, evidenced a brain tumor in each case (Fig. 1A). All of the animals were anesthetized (intraperitoneal injection of ketamine and xylazine) for 1 hour during NMR imaging studies (Jyroscan Imager ACS-Power-TRAK6000; Philips, Best, The Netherlands; apparatus for human imaging). Because of the localization of the C6 and U373 gliomas and the 9L gliosarcomas by NMR imaging studies (Fig. 1A), each rat was able to undergo surgery to remove the bulk of its C6 or U373 glioma or its 9L gliosarcoma (Fig. 1B). Surgery was performed on 36 of the 51 C6 and 13 of the 20 U373 gliomabearing rats and 25 of the 39 9L gliosarcoma-bearing rats when each rat had lost 10% of its weight as compared with its weight 3 days before. The rats were anesthetized as for NMR imaging studies. The remaining 15 C6-, 7 U373 glioma-, and 14 9L gliosarcoma-grafted rats acted as control groups (which underwent no treatment). During the operation performed on the 36 C6 and the 13 U373 glioma- and the 25 9L gliosarcoma-bearing rats, a micropump (Alzet micro-osmotic pump, model 1002; Alza Corp., Palo Alto, CA) was installed subcutaneously on the back of each rat just after the removal of the C6 or U373 glioma or the 9L gliosarcoma bulk. The end of the catheter was implanted directly into the brain (i.e., the cavity resulting from the tumor resection and maintained on the skull with cement) (Fig. 1C). These pumps delivered either saline (Group A; 0.15 ␮l/h) or 10⫺8 mol/L gastrin (Group B; 0.15 ␮l/h) for 14 days. The A and B C6 glioma groups each contained 18 rats. Groups A and B also contained 11 and 14 rats, respectively, for the 9L gliosarcoma and 7 and 6 rats, respectively, for the U373 human glioma xenograft. All the in vivo experiments described were performed with the authorization of the Animal Ethics Committee of the Faculty of Medicine of the Université Libre de Bruxelles (agreement 55/LA 1230342).



Statistical Analysis Statistical comparisons of the control and treated groups were made by first carrying out the Kruskal-Wallis test (nonparametric one-way analysis of variance). Where this test revealed significant differences, we investigated whether any of the treated groups differed from controls. To this end, we applied the Dunn procedure (2-sided test) for multiple comparisons adapted to the special case of treatment-versuscontrol comparisons, i.e., where only (k ⫺ 1) comparisons were made among the k groups tested by the Kruskal-Wallis test (in place of k(k ⫺ 1)/2, possible comparisons were considered under the general procedure) (29, 33). All statistical analyses were performed by Statistica software (Statsoft, Tulsa, OK).

RESULTS Morphological and Biological Descriptions of the C6 and U373 Glioma Models

FIGURE 1. A, NMR image of a C6 glioma (white arrow) in the brain of a rat; 100,000 C6 cells had been stereotactically implanted (in 50 ␮l of saline) 15 days earlier. B, intraoperative photograph illustrating surgical removal of the C6 tumor bulk from the brain of an anesthetized rat; surgery was performed on the basis of tumor localization by NMR imaging studies. C, photograph showing a micropump (single white arrow) connected to a catheter (double white arrow) delivering saline (control) or 10⫺8 mol/L gastrin (0.15 ␮l/h) into the tumor resection cavity for 14 days; the catheter is fixed onto the skull of the rat with cement (dotted white arrow). The micropump system is implanted subcutaneously. D, photomicrograph (original magnification, ⫻40) of the development of a human U373 glioma in the brain (NB) of a nude rat. This histological image is consistent with a high-grade glioma exhibiting marked invasion features in the direction of the brain parenchyma. E, photomicrograph (original magnification, ⫻40) of a rat C6 glial tumor (C6) invading the normal brain parenchyma (NB) of the rat into which it has been stereotactically grafted. Note the presence of numerous islets of tumor cells (arrows) invading the brain parenchyma. These features are also typical of biologically aggressive high-grade gliomas. F, photomicrograph of the C6 glioma, which exhibits pseudopalisading processes (PS) surrounding necrotic areas (N) with important areas of hemorrhage (HE) (original magnification, ⫻100). G, photomicrograph showing a 9L gliosarcoma dramatically invading the normal brain parenchyma (NB) (original magnification, ⫻40). H, photomicrograph showing that, like the C6 glioma, the 9L gliosarcoma also exhibits large areas of necrosis (N) surrounded by pseudopalisading processes (PS) (original magnification, ⫻100). All of these features suggest a highly malignant nature for the experimental C6 and 9L tumors.

NEUROSURGERY

Figure 1A illustrates the NMR imaging study that evidenced a C6 glioma developing in the brain of a rat. The site of the glioma pinpointed by the NMR imaging study enabled the bulk of the glioma to be surgically removed (Fig. 1B). A micropump system could thus be implanted subcutaneously to deliver gastrin (or its vehicle, saline, as control) directly into the brain, i.e., into the tumor resection cavity (Fig. 1C). Figure 1D illustrates the morphological appearance of a human U373 glioma developing in the brain of a nude rat; Figure 1, E and F, illustrates the same development for a rat C6 glioma; and Figure 1, G and H, illustrates the development of a rat 9L gliosarcoma. These three models refer to high-grade (malignant) tumors.

Determination of the Levels of PKC Expression Figure 2A illustrates the levels of expression of six PKC isoforms in the C6 rat glioma (open columns), the 9L rat gliosarcoma (gray columns), and the U373 human glioblastoma (black columns). These levels of expression were analyzed by means of computer-assisted fluorescence microscopy. Figure 2, B–E, illustrates the fact that the subcellular distribution of the PKC isoforms dramatically differed from one isoform to another. These features suggest specific roles for each of the PKC isoforms in their intracellular signaling roles, as already proven in the literature. In addition, the distributions for distinct isoforms can be similar for a given cell line, but differ between distinct cell lines. At present, we have no explanations for these features, and additional experiments are warranted.

Identification of CCKA, CCKB, and CCKC RNAs Because the CCKA receptor is not the actual target of gastrin in the brain, we performed conventional PCR analyses only to investigate whether CCKA receptor RNAs were or were not present in the 9L, C6, and U373 experimental tumors. The data


LEFRANC

ET AL.

Although it expressed RNAs for the CCKA receptor (and low levels of RNAs for the CCKC receptor), gastrin was unable to significantly modify the growth level of the 9L tumor cells (black columns in Fig. 3E). In contrast, gastrin significantly decreased the level of overall growth in the C6 rat gliomas in the case of doses ranging between 10⫺10 and 10⫺7 mol/L, with an optimal effect reached at 10⫺8 mol/L (gray columns in Fig. 3E). The same pattern of gastrin-induced modifications at overall growth level was observed with respect to the U373 tumor cells (open columns in Fig. 3E). The gastrin-induced decrease in overall growth was accompanied by a significant gastrin-induced cytostatic effect (i.e., an accumulation of tumor astrocytes in the G1 phase of their cell cycle) in both models, i.e., human U373 (Fig. 3F) and rat C6 (Fig. 3G) gliomas. Maximum effects were observed in both models for the 10⫺8 mol/L concentration (Fig. 3, F and G).

Gastrin-mediated Influences on the Expression of Cyclin and CDK Protein

FIGURE 2. A, quantitative determination (by means of computer-assisted microscopy) of the immunocytochemical expression of six PKC isoforms (PKC-␣, PKC-␦, PKC-␥, PKC-⑀, PKC-␩, and PKC-␨) in the rat 9L gliosarcoma (gray columns), the rat C6 glioma (open columns), and the human U373 glioma (black columns) tumor cells. The levels of expression of each PKC isoforms were quantitatively analyzed by means of computer-assisted fluorescence microscopy, which enables the PKC concentration per cell to be computed (in arbitrary units). Five hundred cells were analyzed for each tumor type and for each PKC isoform. The data are provided as means (columns) ⫾ their standard errors (bars). B–E, photomicrographs showing morphological patterns of PKC-␣ in 9L tumor cells (B), of PKC-␨ in C6 tumor cells (C), of PKC-␥ in C6 tumor cells (D), and of PKC-␥ in U373 tumor cells (E).

indicate that the C6 (Lane 4 in Fig. 3A) and U373 (data not shown) tumor cells did not express RNAs for the CCKA receptor, and the 9L cells did (Lane 5 in Fig. 3A). In contrast, because both CCKB and CCKC receptors are the actual targets for gastrin in the brain, we performed quantitative RT-PCR analyses to characterize their RNA levels in the three tumor models under study. The data in Figure 3D indicate that the C6 tumor cells expressed significant amounts of RNAs for the CCKB receptor, whereas 9L and U373 tumor cells only did so at very low levels. The U373 tumor cells express high levels of RNAs for the CCKC receptor, whereas 9L tumor cells only did so at very low levels (Fig. 3D).


Three independent experiments were performed to analyze the levels of expression of cyclin D3 and CDK4 protein expression in human U373 and rat C6 glioma models either treated with gastrin at different concentrations or left untreated. Figure 4 illustrates the mean values for these three independent experiments, and the resultant data indicate that gastrin decreased the levels of expression of both the cyclin D3 (Fig. 4, A and B) and the CDK4 protein (Fig. 4, C and D) in both the human U373 (Fig. 4, A and C) and the rat C6 (Fig. 4, B and D) glioma models. This feature could at least partly explain the significant gastrin-induced increase in the accumulation of C6 and U373 cells in their G1 cell cycle phase (Fig. 3, F and G), with the result of a significant gastrin-induced decrease in overall growth (Fig. 3E). Quantitative Western blot analyses performed at the levels of cyclin E and the expression of proteins CDK2, CDK5, and CDK7 demonstrated that gastrin induced only weak modifications in the rat C6 and the human U373 glioma cells (data not shown). We were unable to demonstrate the presence of p15 and p16 proteins in human U373 gliomas. The data illustrated in Figure 4, E and F, demonstrate that gastrin increased the levels of expression of E2F1 and E2F2 proteins in the rat C6 glioma model.

In Vivo Gastrin-induced Effects on the Survival of the C6, 9L, and U373 Tumor-bearing Rats Figure 5A shows that the rats bearing the C6 gliomas in their brains (51 of 60 of the stereotactically implanted rats) died approximately 3 weeks after the stereotactic implantation of 100,000 C6 cells into their brains when they were left untreated (control group in Fig. 5A). Surgery significantly extended the survival time of the C6 glioma-bearing rats (surgery group in Fig. 5, A and B). The individual benefit from surgery for each rat is indicated in Figure 5, A, C, and E, by black areas above the gray columns, which represent the survival in days elapsed from the day of tumor grafting till the day of surgery. The delivery of 10 nmol/L gastrin for 14 days (with 0.15 ␮l deliv-



FIGURE 3. A, analysis of the expression of the total CCKA receptor RNA in C6 (Lane 4) and 9L (Lane 5) rat tumor cells. Lane 3 corresponds to the positive control in the shape of normal rat pancreatic tissue. Lane 2 corresponds to the negative control (H2O) and Lane 1 to molecular weight markers (detailed in Fig. 1, D and E). B, analysis of the expression of the total CCKB receptor RNA in C6 (Lane 4) and 9L (Lane 5) rat tumor cells. Lane 3 corresponds to the positive control in the shape of normal rat brain tissue. Lane 2 corresponds to the negative control (H2O), and Lane 1 to molecular weight markers (detailed in Fig. 1, D and E). C, analysis of the expression of total ␤-actin RNA in normal rat pancreatic tissue (Lane 1), rat 9L tumor cells (Lane 2), normal rat brain tissue (Lane 3), and rat C6 tumor cells (Lane 4). The same characterization of total ␤-actin RNA expression was also performed for U373 tumor cells and their corresponding normal positive controls, i.e., Jurkat and LoVo cells (data not shown). D, quantitative RT-PCR analyses of total CCKB (gray columns) and CCKC (black columns) receptor RNAs in rat 9L gliosarcoma, rat C6 glioma, and human U373 glioma cells. The data are presented as means ⫾ standard error of the mean values calculated on three distinct experiments. E, characterization of the gastrin-induced in vitro effects on the growth (analyzed by means of the colorimetric MTT assay) of the U373 (open columns), the C6 (gray columns), and the 9L (black columns) tumor cells. The tumor cells were cultured for 72 hours with (or without; 0 ⫽ control) gastrin. The P values above each column for the 0.001 to 1000 nmol/L gastrin concentration refer to the level of statistical significance of each specific experimental condition as compared with the corresponding control condition, i.e., the “0” nmol/L condition at the extreme left of the figure. F and G, characterization of the gastrinmediated in vitro effects on the U373 and C6 cell kinetics, respectively, i.e., the distribution of the U373 and C6 cells in the G1 (open columns, i.e., the first series of columns), the S (hatched columns, i.e., the second series of columns), the G2 (gray columns, i.e., the third series of columns), and the M (black columns, i.e., the fourth series of columns) phases of their cell cycles. These cell cycle kinetics were evaluated by analyses of the U373 and C6 Feulgen-stained nuclei by means of computerassisted microscopy. The P values above each column for the 0.1, 10, and 1000 nmol/L gastrin concentration refer to the level of statistical significance of each specific experimental condition as compared with the corresponding control condition, i.e., the “0” nmol/L condition at the extreme left of the figure.

ered per h) into the tumor resection cavities significantly increased the survival periods of these C6 tumor-bearing rats as compared with the group of rats that underwent surgery alone (Fig. 5, A and B). Surgery slightly increased the survival periods of the 9L gliosarcoma-bearing rats, whereas surgery plus gastrin delivery added no significant benefit in terms of survival as compared with surgery alone (Fig. 5, C and D). With respect to the U373 human glioma xenografts, surgery brought no significant benefit in terms of survival, whereas surgery plus gastrin delivery did (Fig. 5, E and F).

DISCUSSION Gastrins and cholecystokinins derive from a precursor family of genes that generate a variety of products, each with a distinctive spectrum of biological activity that has recently been reviewed by Dockray et al. (11). Gastrin can

NEUROSURGERY

act as a potent cell growth factor that is involved in a variety of normal and abnormal biological processes, including the maintenance of gastric mucosa, the proliferation of enterochromatin-like cells, and neoplastic transformations (30). Gastrin significantly modifies the level of cell proliferation in a large variety of gastrointestinal tumors (27, 30) and also in human astrocytic tumors, as we have previously observed (7–9). Cell proliferation relates to specific cyclin-CDK protein complexes that control the transition of cells from one phase of the cell cycle to the next, and the activities of several cyclin-CDK complexes controlling the transition from the G1 to the S phase of the cell cycle are seriously disturbed in human astrocytic tumors as a result of mutations occurring in the genes whose products are involved in the p16/p15/CDK4/CDK6/Rb pathway (18). The data from the present study indicate that gastrin sig-


LEFRANC

ET AL.

FIGURE 4. Characterization (by means of quantitative Western blot analyses) of the level of expression of cyclin D3 (A, human U373 and B, rat C6 tumor cells), CDK4 (C, human U373 and D, rat C6 tumor cells), E2F1 (E, rat C6 glioma), and E2F2 (F, rat C6 glioma) gene products (from left to right in each group). The data are represented as means ⫾ standard error of the mean values calculated for these three independent experiments.

nificantly decreased the overall growth rate in both the rat C6 and the human U373 glioma cells, with the decrease taking place as a result of an accumulation of tumor astrocytes in the G1 phase of their cell cycles. The data from the present study also suggest that part of this effect could be mediated through a gastrin-induced decrease in the levels of expression of cyclin D3-CDK4. Gastrin triggers the activation of multiple signal transduction pathways that relay the mitogenic signal to the cell nucleus and modify cell proliferation (30). The gastrin-induced activation of various biological processes is mediated by mitogen-activated protein kinase-induced c-fos gene transcription via PKC-


dependent and -independent pathways (34, 35). Among other things, gastrin acts on the Sis-inducible element, the serum response element, and the Ca2⫹-cyclic adenosine monophosphate response element that regulates the elements of the c-fos promoter (34). Although the activation of the serum response element involves the small guanosine triphosphate-binding protein RhoA, gastrin targets the c-fos promoter CarG sequence via RhoA-dependent pathways (30). As mentioned above, gastrin can activate the PKC-dependent pathways. PKC, which is in fact a family of enzymes that play a ubiquitous role in intracellular signal transduction (12), also plays a dramatic role in tumoral astrocyte migration (15, 32, 37). PKC-␣, PKC-⑀, and PKC-␨ have already been demonstrated to play major roles in tumor astrocyte cell proliferation and/or apoptotic features (15, 32, 37). We indicate here that the PKC-␣, -␥, -␦, -⑀, -␩, and -␨ subunits are present in the 9L, C6, and U373 tumor models with dramatic differences in terms of subcellular distribution. These differences in subcellular distribution emphasize distinct subcellular roles for each PKC isoform in terms of signal transduction. In the present study, we observed higher gastrin-induced modifications at tumor astrocyte cell proliferation level at a concentration of 10⫺8 mol/L rather than at one of 10⫺6 mol/L. A similar observation was reported by Kimura et al. (16), who noted that in the case of amylase secretion, micromolar concentrations of gastrin-CCK-related peptides may activate the intracellular mechanism that inhibits PKC activity in acini, whereas physiological concentrations of gastrin-CCK-related peptides activate PKC-dependent signals directly. Pharmacological (micromolar) doses of gastrin can therefore desensitize PKC-mediated signals (16), a feature that could at least partly explain the weak effects observed with 1 ␮mol/L gastrin at C6 cell proliferation level (or even their absence). Thus, numerous data have already been published on the intracellular pathways used by gastrin to effect cell proliferation. In contrast, however, no data have been published, at least to our knowledge, on the effects of gastrin on cyclin-CDK complex expression in relation to cell proliferation. The only report that can be found in scientific databases concerns the work published by Koh et al. (19), which indicates that gastrin is a target for the ␤-catenin/T-cell factor-4 growth-signaling pathway in intestinal polyposis, and that the downstream targets of ␤-catenin include some cyclins. The present data suggest that although gastrin significantly decreases the level of expression of cyclin D3CDK4 complex expression, the levels of expression of other proteins, including cyclin E, CDK2, CDK5, and CDK7, are not modified by gastrin (data not shown), or only marginally so. CDK4 and CDK6 genes encode for proteins with catalytic kinase activities, and both can form complexes with members of the cyclin D family because both are inhibited by p15 and p16 (18). The overexpression of either CDK might be expected to mimic mutations of the p15/p16 inhibitors and to override their function (18). Our data suggest that gastrin lowered the amounts of the cyclin D3-CDK4 complex in both the rat C6 and the human U373 tumor astrocytes, a feature that could explain the gastrin-induced



FIGURE 5. Illustration of the individual (A, C, and E) and overall (B, D, and F) survival periods (in days after tumor cell implantation into the rats’ brains) of the rats bearing orthotopic grafts of either the rat C6 or the human U373 glioma model, or the 9L rat gliosarcoma, and which did not otherwise undergo treatment (control groups), for rats that underwent surgery for tumor debulking (Fig. 1B, with NMR imaging scans used to localize the tumor, as illustrated in Fig. 1A), i.e., the “surgery” (and saline delivery through micropumps) group, and of the rats that underwent surgery and gastrin delivery (Fig. 1C), i.e., the “surgery plus gastrin” group. The individual benefit from surgery for each rat is indicated in A, C, and E by black areas above the gray bars. The gray bars represent the survival in days elapsed from the day of tumor grafting until the day of surgery. Rats underwent treatment when they had lost 10% of their weight as compared with the weight recorded 3 days earlier. B, D, and F, overall effects of surgery (plus saline delivery through micropumps) and surgery plus gastrin on the survival periods of the tumor-bearing rats; the individual survival periods of these rats are illustrated in A, C, and E, respectively. The P levels of statistical significance in B, D, and F are provided for both the surgery (plus saline) and the surgery plus gastrin groups in comparison with the control group.

effect observed at cell cycle kinetics level, i.e., a significant increase in these tumor astrocytes in the G1 phase of their cell cycles. However, functional studies are definitely needed to decide whether these consequences are actually important in the biological behavior characteristic of highly malignant gliomas. We are now involved in such functional studies. The ultimate target of the kinase activities of the cyclin D-CDK4/CDK6 complexes is the Mr 107,000 retinoblastoma (Rb) protein (18). The phosphorylation of Rb allows the release of the E2F1 and E2F2 transcription factors that it complexes with, and this in turn activates the genes necessary for the activation and/or inhibition (depending on the signaling pathway activated) of cell proliferation (18, 31). The antibodies that we used against E2F1 and E2F2 are specific to rat proteins. Therefore, we do not know about the status of these proteins in human U373 cells. The data that we obtained in vivo demonstrate that gastrin significantly increased the survival of the C6 and U373 glioma-bearing rats, both of which had first undergone tumor debulking by means of surgery. This gastrin-induced increase in the survival of the C6 and U373 glioma-bearing rats could relate partly to the in vitro

NEUROSURGERY

observed gastrin-induced cytostatic effects that can delay the regrowth of C6 and U373 tumor cells escaping surgical removal. The C6 and U373 cells exhibited gastrin receptors, whereas the 9L tumor cells did not: the 9L gliosarcoma-bearing rats did not benefit from gastrin treatment. However, the fact remains that the studies into the effects on proliferation and expression of cyclins were undertaken in vitro, whereas the survival studies were in vivo: we cannot extrapolate that the extended survival of the animals is directly attributable to an effect observed in the in vitro studies. Also, cyclin expression studies were performed on cultured cells and not on the treated tumors. The quantitative determination of cyclin-CDK expression should also be performed on in vivo materials before any definitive conclusions can be drawn concerning the potential role of gastrin on cyclin-CDK expression.

CONCLUSION

The data from the present study demonstrate that gastrin decreases the overall growth rate in the malignant rat C6 and human U373 glioma models, with a cytostatic effect evidenced in terms of a significant accumulation of glioma cells in the G1 phase of their cell cycles. This cytostatic effect may have occurred, at least partly, through a gastrin-induced decrease in the amount of the cyclin D3-CDK4 complex in both U373 and C6 glioma cells. Other cell cyclerelated complexes must be targeted by gastrin because we had previously observed that gastrin is also able to stimulate cell proliferation in other glioma models. We are now in the process of trying to identify that intracellular pathways could be activated in gliomas whose cell proliferation is stimulated by gastrin.

REFERENCES 1. Amberger VR, Hensel T, Ogata N, Schwab ME: Spreading and migration of human glioma and rat C6 cells on central nervous system myelin in vitro is correlated with tumor malignancy and involves a metalloproteolytic activity. Cancer Res 58:149–158, 1998. 2. Baldwin GS: Antiproliferative gastrin/cholecystokinin receptor antagonists target the 72-kDa gastrin binding protein. Proc Natl Acad Sci U S A 91: 7593–7597, 1994.


LEFRANC

ET AL.

3. Belot N, Rorive S, Doyen I, Lefranc F, Bruyneel E, DeDecker B, Micik S, Brotchi J, Decaestecker C, Salmon I, Kiss R, Camby I: The molecular characterization of cell-substratum attachments in human glial tumors relates to prognostic features. Glia 36:375–390, 2001. 4. Bold RJ, Alpard S, Ishizuka J, Townsend CM Jr, Thompson JC: Growthregulatory effect of gastrin on human colon cancer cell lines is determined by protein kinase A isoform content. Regul Pept 53:61–70, 1994. 5. Branle F, Lefranc F, Camby I, Jeukens J, Geurts-Moespot A, Sprenger S, Sweep F, Kiss R, Salmon I: Evaluation of the efficiency of chemotherapy in in vivo orthotopic models of human glioma cells with and without 1p/19q deletions and in C6 rat orthotopic allografts serving for the evaluation of surgery combined with chemotherapy. Cancer 95:641–655, 2002. 6. Camby I, Belot N, Rorive S, Lefranc F, Maurage CA, Lahm H, Kaltner H, Hadari Y, Ruchoux MM, Brotchi J, Zick Y, Salmon I, Gabius HJ, Kiss R: Galectins are differentially expressed in supratentorial pilocytic astrocytomas, astrocytomas, anaplastic astrocytomas and glioblastomas, and significantly modulate tumor astrocyte migration. Brain Pathol 11:12–26, 2001. 7. Camby I, Salmon I, Danguy A, Pasteels JL, Brotchi J, Martinez J, Kiss R: Influence of gastrin on human astrocytic tumor cell proliferation. J Natl Cancer Inst 88:594–600, 1996. 8. Camby I, Salmon I, Oiry C, Galleyrand JC, Nagy N, Danguy A, Brotchi J, Pasteels JL, Martinez J, Kiss R: The influence of gastrin and/or cholecystokinin antagonists and agonists on the proliferation of three human astrocytic tumor cell lines. Neuropepides 30:433–437, 1996. 9. Camby I, Salmon I, Rorive S, Gras T, Darro F, Kruczynski A, Danguy A, Pasteels JL, Kiss R: Characterization of the influence of anti-hormone and/or anti-growth factor neutralizing antibodies on cell clone architecture and the growth of human neoplastic astrocytic cell lines. J Neurooncol 20:67–80, 1994. 10. Chicoine MR, Silbergeld DL: Invading C6 glioma cells maintaining tumorigenicity. J Neurosurg 83:665–671, 1995. 11. Dockray GJ, Varro A, Dimaline R, Wang T: The gastrins: Their production and biological activities. Annu Rev Physiol 63:119–139, 2001. 12. Goekjian PG, Jirousek MR: Protein kinase C in the treatment of disease: Signal transduction pathways, inhibitors, and agents in development. Curr Med Chem 6:877–903, 1999. 13. Hosli E, Hosli L: Binding of cholecystokinin, bombesin and muscarine to neurons and astrocytes in explant cultures of rat central nervous system: Autoradiographic and immunohistochemical studies. Neuroscience 61:63– 72, 1994. 14. Hyndman D, Cooper A, Pruzinsky S, Coad D, Mitsuhashi M: Software to determine optimal oligonucleotide sequences based on hybridization simulation data. Biotechniques 20:1090–1097, 1996. 15. Janik P, Szaniawska B, Miloszewska J, Pietruszewska E, Kowalczyk D: The role of protein kinase C in migration of rat glioma cells from spheroid cultures. Cancer Lett 63:167–170, 1992. 16. Kimura T, Honda T, Higashi T, Konishi J: High concentration of cholecystokinin octapeptide suppress protein kinase C activity in guinea pig pancreatic acini. Peptides 17:917–925, 1996. 17. Kiss R, Salmon I, Camby I, Gras S, Pasteels JL: Characterization of factors in routine laboratory protocols that significantly influence the Feulgen reaction. J Histochem Cytochem 41:935–945, 1993. 18. Kleihues P, Cavenee WK (eds): Pathology and Genetics of Tumours of the Nervous System: International Agency for Research on Cancer (IARC) WHO Health Organisation. Oxford, Oxford Press, 2000. 19. Koh TJ, Bulitta CJ, Fleming JV, Dockray GJ, Varro A, Wang TC: Gastrin is a target of the ␤-catenin/TCF-4 growth-signaling pathway in a model of intestinal polyposis. J Clin Invest 106:533–539, 2000. 20. Kucharckzac J, Pannequin J, Camby I, Decaestecker C, Kiss R, Martinez J: Gastrin induces over-expression of genes involved in human U373 glioblastoma cell migration. Oncogene 20:7021–7028, 2001. 21. Liau LM, Jensen ER, Kremen TJ, Odesa SK, Sykes SN, Soung MC, Miller JF, Bronstein JM: Tumor immunity within the central nervous system stimulated by recombinant Listeria monocytogenes vaccination. Cancer Res 62: 2287–2293, 2002. 22. Malonne H, Farinelle S, Decaestecker C, Gordower L, Fontaine J, Chaminade F, Saucier JM, Atassi G, Kiss R: In vitro and in vivo pharmacological characterizations of the antitumor properties of two new olivacine derivatives, S16020-2 and S30972-1. Clin Cancer Res 6:3774–3782, 2000.


23. Muller W, Heinemann U, Berlin K: Cholecystokinin activates CCKBreceptor-mediated Ca-signaling in hippocampal astrocytes. J Neurophysiol 78:1997–2001, 1997. 24. Nagane M, Lin H, Cavenee WK, Huang HJ: Aberrant receptor signaling in human malignant gliomas: Mechanisms and therapeutic implications. Cancer Lett 162:S17–S21, 2001. 25. Pannequin J, Oiry C, Morel C, Kucharczak J, Camby I, Kiss R, Gagne D, Galleyrand JC, Martinez J: C-terminal heptapeptide of gastrin inhibits astrocytomas mobility by interacting with a new gastrin binding site. J Pharmacol Exp Ther 302:274–282, 2002. 26. Parsa AT, Chakrabarti I, Hurley PT, Chi JH, Hall JS, Kaiser MG, Bruce JN: Limitations of the C6/Wistar rat intracerebral glioma model: Implications for evaluating immunotherapy. Neurosurgery 47:993–1000, 2000. 27. Rehfeld JF, van Solinge WW: The tumor biology of gastrin and cholecystokinin. Adv Cancer Res 63:295–347, 1994. 28. Reubi JC, Schaer JC, Waser B: Cholecystokinin (CCK)-A and CCK-B/gastrin receptors in human tumors. Cancer Res 57:1377–1386, 1997. 29. Rosner B: Fundamentals of Biostatistics. London, International Thomson Publishing, 1995. 30. Rozengurt E, Walsh JH: Gastrin, CCK, signaling, and cancer. Annu Rev Physiol 63:49–76, 2001. 31. Serrano M, Hannon GJ, Beach D: A new regulatory motif in cell-cycle control causing specific inhibition of cyclin/CDK4. Nature 336:704–707, 1993. 32. Sharif TR, Sharif M: Overexpression of protein kinase C epsilon in astroglial brain tumor derived cell lines and primary tumor samples. Int J Oncol 15:237–243, 1999. 33. Siegel S, Castellan NJ Jr: Nonparametric Statistics for the Behavioral Sciences. Singapore, McGraw-Hill, 1988, ed 2. 34. Stepan VM, Tatewaki M, Matsushima M, Dickinson CJ, Del Valle J, Todisco A: Gastrin induces c-fos gene transcription via multiple signaling pathways. Am J Physiol 276:G417–G424, 1999. 35. Todisco A, Takeuchi Y, Urumov A, Yamada J, Stepan VM, Yamada T: Molecular mechanisms for the growth factor action of gastrin. Am J Physiol 273:G891–G898, 1997. 36. Tortora G, Pepe S, Bianco C, Damiano V, Ruggiero A, Baldassare G, Corbo C, Cho-Chung YS, Bianco AR, Ciardello F: Differential effects of protein kinase A sub-units on Chinese-hamster-ovary cell cycle and proliferation. Int J Cancer 59:712–716, 1994. 37. Zhang W, Law RE, Hinton DR, Couldwell WT: Inhibition of human malignant glioma cell motility and invasion in vitro by hypericin, a potent protein kinase C inhibitor. Cancer Lett 120:31–38, 1997.

Acknowledgments FL holds a grant from the Fondation Erasme (Belgium), RK is a director of research with the Fonds National de la Recherche Scientifique (Belgium). The present study was supported by grants awarded by the Fonds de la Recherche Scientifique Médicale (Belgium) and the Fondation Yvonne Boël (Belgium).

COMMENTS

I

n this study, the authors have investigated gastrinmediated growth inhibition on a variety of glial cell lines. The authors confirm their previously published work demonstrating the presence of the three cholecystokinin receptor subtypes in three experimental models (i.e., the U373 human glioma, the C6 rat glioma, the 9L rat gliosarcoma) (3). In addition, the in vitro studies documenting antiproliferative effects are complemented by in vivo studies that apply the local delivery of gastrin. The authors conclude that gastrin is able to significantly decrease the growth levels of gastrin receptor-positive tumor cells in vitro and in vivo. This study is a good first step toward clinical application of gastrin-mediated inhibition of tumor growth. It



complements previous work by this group that has shown that gastrin inhibits the migratory capacity of glioma cells (2). However, the artifactual nature of implantation models makes it difficult to assess the clinical relevance of these findings. The availability of more refined transgenic animal models should allow these investigators to pursue more meaningful preclinical goals (1). Andrew T. Parsa San Francisco, California 1. Holland EC: Brain tumor animal models: Importance and progress. Curr Opin Oncol 13:143–147, 2001. 2. Lefranc F, Camby I, Belot N, Bruyneel E, Chaboteaux C, Brotchi J, Mareel M, Salmon I, Kiss R: Gastrin significantly modifies the migratory abilities of experimental glioma cells. Lab Invest 82:1241–1252, 2002. 3. Lefranc F, Chaboteaux C, Belot N, Brotchi J, Salmon I, Kiss R: Determination of RNA expression for cholecystokinin/gastrin receptors (CCKA, CCKB and CCKC) in human tumors of the central and peripheral nervous system. Int J Oncol 22:213–219, 2003.

G

astrin has been postulated to modulate the growth of a number of tumors, including gliomas. Several gastrin receptors have been identified and cloned, one of which, CCKB, is present in the brain. Normal astrocytes possess significant amounts of this receptor, and therefore the potential exists for regulation of astrocytomas to be in part controlled by gastrin. The authors have done an incredible amount of work investigating the effects of gastrin on glioma cells, in vitro and in vivo. Using a variety of scientific techniques, they examined the effects of gastrin on proliferation, cell cycle kinetics, and the expression of certain cyclins in vitro and then investigated whether gastrin affects the growth of an intracerebral tumor in

vivo. In particular, two of the three experimental gliomas studied (human U373 and rat C6) were shown to express the CCKB receptor, whereas the third (rat 9L gliosarcoma) tumor was used as a negative control since it does not express the CCKB receptor. The results demonstrate that the in vitro gastrin significantly decreased the overall growth rate of the C6 and U373 astrocytoma cells but had no effect on the rat 9L gliosarcoma cells. The effect is thought to occur by maintenance of the tumors in the G1 cell cycle phase as a result of a gastrin-induced decrease in the levels of expression of the cyclin D3-CDK4. In vivo, treatment with gastrin increased the survival of animals with an intracerebral tumor in the C6 and U373 animal models but not the 9L gliosarcoma model. Technically, there is no question about the quality of the scientific work performed. It appears to be well done. However, one problem with this study is in the interpretation of the mechanisms by which gastrin may prolong survival, since the studies on the effects on proliferation and the expression of cyclins were performed in vitro, whereas the survival studies were in vivo. We cannot necessarily extrapolate that the prolongation of survival of the animals in vivo is due to an effect seen in in vitro cell culture studies. Future studies should include in vivo tumor analysis as well. In addition, the results should be interpreted with caution, as it has been reported that gastrin is also able to stimulate proliferation in other tumor models. Nevertheless, this is a thorough analysis with intriguing results that support further study of these cell cycle pathways as a target for future therapy for gliomas. Roberta P. Glick Terry Lichtor Chicago, Illinois

Instructions for Contributors Neurosurgery’s Information for Contributors section of the journal has been updated to include new and modified submission information, including but not limited to general submission requirements, digital formatting guidelines, and Letter to the Editor submission guidelines. The most visible modification of this section involves design elements that should allow readers to retrieve information more easily and quickly than ever before. Unabridged versions of this section appear in the January and July issues of the journal, with abbreviated versions appearing throughout the remaining months’ issues. Contributors may also find these guidelines online at www.neurosurgeryonline.com.

NEUROSURGERY
