| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
Regular Articles |
Department of Pathology, Fox Chase Cancer Center, Philadelphia, Pennsylvania 19111 [H. M., A. M. G-G., R. L. D. C., D. E. B., T. G., A. J. P. K-S.], and Neuroscience Research Center of the Charité, Humboldt University, Faculty of Medicine, 10117 Berlin, Germany [K-H. B.]
 |
ABSTRACT |
|---|
 Top ABSTRACT INTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
|
INTRODUCTION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Chemical carcinogenesis protocols of the mouse skin give rise to benign and malignant tumors, including SCC of different aggressive behavior. These experimental tumors have been extensively studied because they represent an excellent model of human SCC, one of the most common malignant tumors, that has similar histopathological and biological features in several organs, such as lung, oral cavity, larynx, esophagus, cervix, skin, etc.
In a previous study, using a pair of low-grade/high-grade SCC cell lines derived from the same primary tumor (CC4B/CC4A cells), we identified by DD a gene product that was overexpressed in the invasive tumor cells. This protein, PACE4, was established to be a member of the pro-protein convertase family and later characterized, together with furin, as having a paramount role in tumor progression (3 , 4) .
In the present study, we describe another gene product, VILIP-1, identified by DD. This gene product was highly expressed in the relatively indolent SCC line CC4B but was drastically reduced in its aggressive counterpart CC4A. VILIP-1 is a member of the visinin–recoverin or neuronal calcium sensor protein family (5 , 6) . The VILIP protein family contains four EF-hand calcium-binding motifs and a consensus myristoylation site at its NH2 terminus. VILIP-1 modulates the levels of cyclic nucleotides by indirect or direct interactions with adenylyl and guanylyl cyclases (6, 7, 8) . The modulation of cAMP by VILIP-1 has been studied previously in nerve cells and advanced as the link between the effect of VILIP-1 on cell proliferation, differentiation, and possibly migration (6, 7, 8) . Of special interest is the strong evidence that RhoA, a small GTPase involved in cell adhesion, migration, and invasiveness, is inhibited by cAMP-elevating agents (9 , 10) . We now report that VILIP-1 is differentially expressed in murine skin tumors and cell lines of different degrees of aggressiveness and that transfection of two high-grade mouse SCC lines with the VILIP-1 cDNA increased cAMP levels, leading to diminished MMP-9 and RhoA activity, together with a significant reduction in the invasive properties of the carcinoma cells.
|
MATERIALS AND METHODS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
and p117 (from Dr. S. Fischer, Smithville, TX) and the SCC-derived CarB (a gift from Dr. A. Balmain, San Francisco, CA), CC4A, CC4B, CH72, and CH72T3 (11
, 12)
. Cells were grown in S-MEM (Sigma, St. Louis, MO) medium containing FCS (10%), L-glutamine (2 mM), and penicillin/streptomycin (100 µg/ml). Primary epidermal keratinocyte cultures were prepared from normal SENCAR mice (13)
. CC4A and CC4B were derived from the same tumor. When injected s.c. into nude mice, CC4A gave rise to a high-grade SCC or spindle cell carcinoma (or SCC IV), whereas CC4B gave rise to a well-differentiated, less aggressive, and low-grade SCC (SCCII; 3 ). CH72 also gave rise to a low-grade SCC after s.c. inoculation, and CH72T3 is a subcloned cell line obtained by in vivo passaging of CH72 into nude mice that resulted in a high-grade SCC (11) .
DD.
DD was performed as described previously (3)
. Total RNA was extracted from subconfluent cell lines using RNAzol (Cinna/Biotecx, Houston, TX). DD was performed according to the manufacturer’s instructions (GenHunter, Nashville, TN) in conjunction with synthesized in-house arbitrary primers. Briefly, CC4A or CC4B RNA was reverse transcribed using T11MC, T11MA, or T11MG oligonucleotide primers, deoxynucleotide triphosphate, and Moloney murine leukemia virus reverse transcriptase and incubated for 60 min at 37°C. The cDNAs were amplified by the PCR by adding the reverse transcriptase reaction to a PCR mix containing T11MC, an arbitrary 10-mer primer (5'-GATCTAAGGC-3'), deoxynucleotide triphosphates, l 33P-dATP, and TaqDNA Polymerase (AmpliTaq, Perkin-Elmer, Boston, MA; Ref. 3
). Each sample was amplified and subjected to electrophoresis in duplicate. The cDNAs were denatured, diluted with a running buffer, and processed through a 6% polyacrylamide denaturing gel. The gels were then dried and exposed to autoradiography overnight at -80°C.
cDNAs of interest were cut out of the gel, boiled in double-distilled water, ethanol precipitated, and resuspended in H2O. This was reamplified by PCR as described above but without radioactive nucleotide. Fragments were subcloned into the T/A cloning vector (Invitrogen, Carlsbad, CA), used as probes in Northern analysis, and sequenced by an automated ABI sequencer (Perkin-Elmer).
Genome Walk.
Upstream sequences were identified using a Mouse GenomeWalker Kit (Clontech, Palo Alto, CA) according to the manufacturer’s instructions. Two rounds of PCR were performed using gene-specific primary and secondary (nested) primers, together with adapter-specific primary and secondary primers (Clontech). Products of the secondary PCR were separated on a 1.2% agarose gel; bands of interest were extracted, purified, and cloned into PCR2.1-Topo TA-cloning vector (Invitrogen). The ligation mixture was used to transform DH5
bacteria cells (Life Technologies, Inc., Rockville, MD) by heat shock. Single colonies were selected, and the DNA was purified and sequenced.
PCR Primers.
Gene-specific primers used in the first round of Genome Walk were GATGATAAGTTGGACAGTGACTTTCCC (primary) and CAGCTCTTTCCAGATTG AGAGGTTACA (secondary). The gene-specific primer used in the second round of Genome Walk was CACATATCACATTGTGGCAGGTTGTCA (used both as primary and secondary primer).
Measurement of Intracellular cAMP.
Cells were grown overnight at a density of 5 x 105 cells/well in a 96-well plate. Cells were either lysed directly or after being exposed to forskolin (either 25 or 50 µM) for 20 min. Levels of intracellular cAMP were measured with an enzyme immunoassay kit essentially as described by the manufacturer (Amersham Pharmacia, Piscataway, NJ).
Northern Analysis.
Cells were grown in culture media to 80% confluency. Total RNA was extracted using TRIzol reagent (Life Technologies, Inc.) according to the manufacturer. Ten µg of RNA were loaded and run on 1% formaldehyde agarose gels and transferred to nylon membranes (Amersham Pharmacia). Gene-specific 32P-radiolabeled DNA fragments were used to detect VILIP transcripts.
Western Analysis.
Cells were grown to 80% confluency, trypsinized, and washed with 1 x PBS buffer. Cells were then treated with radioimmunoprecipitation assay buffer (1 x PBS, 1% NP40, 0.5% sodium deoxycholate, and 0.1% SDS) containing protease inhibitors (phenylmethylsulfonyl fluoride, Aprotinin, and sodium orthovanadate) for 30 min at 4°C. Supernatants were measured for total protein, extracted, and used in Western blots. After extracting proteins, 40 µg of total protein were run on 4–20% gradient precast gels (Novex/Invitrogen) under denaturing conditions. Proteins were transferred to nitrocellulose membranes (Amersham Pharmacia), and VILIP-1 was detected using a specific rabbit polyclonal antibody (14)
as the primary antibody and horseradish peroxidase-conjugated antirabbit immunoglobulin (Amersham) as the secondary antibody. MMP-9 was detected in culture supernatants using antibody sc-6841 from Santa Cruz Biotechnology (Santa Cruz, CA). The antibody was reacted with a peroxidase-conjugated antigoat antibody (sc-2020; Santa Cruz Biotechnology). Western analysis of primary mouse skin tumors (nine papillomas, three low-grade SCCs, and two high-grade SCCs) was performed following similar procedures.
RhoA Activity.
RhoA activity was assessed using the Rho-binding domain of Rhotekin as described (15)
. In brief, cells (at 70% confluency) were maintained for 24 h in a serum-free medium and then induced by culturing for different time periods with medium containing 5% FBS. This was followed by extraction with radioimmunoprecipitation assay buffer [50 mM Tris (pH 7.2), 500 mM NaCl, 1% Triton X-100, 0.5% sodium deoxycholate, 1% SDS, 10 mM MgCl2, 0.5 µg/ml leupeptin, 0.7 µg/ml pepstatin, 4 µg/ml aprotinin, and 2 mM phenylmethylsulfonyl fluoride]. After centrifugation at 14,000 x g for 3 min, the extracts were incubated for 45 min at 4°C with glutathione beads (Pharmacia Biotech, Stockholm, Sweden), coupled with bacterially expressed GST–RBD (Rho-binding domain of Rhotekin) fusion protein (Upstate Biotechnology, Lake Placid, NY), and then washed three times with Tris buffer (pH 7.2), containing 1% Triton X-100, 150 mM NaCl, and 10 mM MgCl2. The RhoA content in these samples was determined by immunoblotting samples using the rabbit anti-RhoA antibody sc-179 from Santa Cruz Biotechnology.
Immunohistochemistry.
VILIP-1 immunohistochemistry was performed using paraffin-embedded normal skin and cutaneous tumors from SENCAR mice produced by a complete chemical carcinogenesis protocol using benzo(a)pyrene (16)
. All paraffin sections were subjected to an immunostaining protocol published previously (14)
. The same antibody used in Western analysis was used as primary antibody at 1 of 500 and 1 of 1000 dilutions. An avidin-biotin-peroxidase kit (Vectastain Elite, Burlingame, CA) was then used, followed by the chromagen 3', 3'-diaminobenzidine to develop the immunostain. Negative controls, not incubated with VILIP-1 antibody, were incubated with preimmune serum at the same concentrations as the primary antibody. Specificity of the immunostain was determined by preincubating the VILIP-1 antibody with the blocking peptide at a x100 excess concentration and testing this mixture on normal and tumor tissues. All sections were counterstained with hematoxylin and mounted. Grading of immunostain was based on a semiquantitative evaluation of stain intensity from 0 to II. No or marginal staining of <5% of cells stained was called negative (0), mild to moderate stain of 5–50% of tumor cells was graded as I, and intense stain as seen in most samples of normal epidermis and brain and comprising >50% of the cell population was classified as II. A total of 23 normal skin samples, 14 papillomas, 44 low-grade SCCs (SCC I and II), and 14 high-grade SCCs (SCC II and IV) was used. The SCCs were classified according to a modified Broders classification published previously (17
, 18)
. Murine brain cortex was used as a positive control (5)
.
Cell Transfections.
Human VILIP-1 (VSNL1) cDNA (19)
was digested with DraI and cloned into the SmaI site of a pCi.Neo expression vector (Promega, Madison, WI). Either 4 µg of the resulting construct or 4 µg of the vector was used to transfect CH72T3 cells by Lipofectamine reagents (Life Technologies, Inc.) according to the manufacturer. Two days post-transfection, G418 was added at a concentration of 800 µg/ml. After 1 week of selection, cells were lysed, proteins were extracted, and Western blot was performed to screen for VILIP-1 expression.
In Vitro Invasion Assay.
This assay was performed using BioCoat Matrigel inserts (Becton Dickinson Labware, Bedford, MA) according to the manufacturer’s instructions. Briefly, cells were seeded in the inserts in a serum-free medium without or in the presence of 25 or 50 µM forskolin. Inserts were transferred to wells containing S-MEM medium with 5% FBS as a chemoattractant. Because of the relatively low invasive ability of transfected cells, the invasion chambers were incubated at 37°C, 5% CO2 for 36 h. Invading cells were stained and counted (4)
.
In Vivo Invasion Assay.
Tracheal transplants were prepared as described previously (4)
. Cells (5 x 105) of each cell line were inoculated into each de-epithelialized rat trachea (Zivic, Pittsburgh, PA). Twelve tracheas were used for each cell line. After inoculation of cells into tracheas, the tracheas were sealed and transplanted into the dorsal s.c. tissues of Scid mice. Tracheal transplants were removed surgically at 3 weeks, sectioned into 3-mm-thick rings, and fixed in 10% formalin. After H&E staining, the degree of invasion of the tracheal wall was determined by measuring the length of maximum penetration of the tumor cells into the tracheal wall. Each tracheal transplant was represented by two to six measurements corresponding to the number of cross-sections containing tumor cells. A mean was calculated for each tracheal transplant and group of transfected cells. The results were expressed in micrometers of penetration depth.
Cell Proliferation.
In vitro cell growth was assessed by 3[H] Thymidine incorporation. VILIP-1-transfected cells were plated at 2 x 105 cells/well in 24-well culture dishes and allowed to adhere overnight. The cultures were synchronized by serum starvation for 24 h, followed by restimulation with serum-containing media. 3[H] Thymidine (1.5 µCi/ml; 20 Ci/mmol) was added to the cultures and incubated for 4 h. The proliferation index was calculated using the mean from three independent experiments, each of which was performed in triplicate. A ratio between the proliferation index of the vector alone-transfected cells and the proliferation index of the respective VILIP-1 transfectant was calculated. In vivo cell proliferation was studied using paraffin sections of tracheal xenografts. A rabbit polyclonal anti-Ki67 serum from Novocastra/Vector (Burlingame, CA) was used to detect the proliferating fraction in mouse tissues. An LI (ratio between positively immunostained nuclei and all nuclei counted) was determined in five high magnification fields (x400) totaling 500-1000 cells/specimen. Two specimens each of tracheal transplants containing either CC4A or CH72T3 cells transfected either with VILIP-1 or with vector alone (Cin) were used.
Zymography.
Cells (1 x 106) were grown overnight in a serum-free S-MEM (Sigma) medium containing L-glutamine (2 mM) and penicillin/streptomycin (100 µg/ml), either in the absence of forskolin or presence of 25 and 50 µM forskolin. The conditioned media were concentrated down to 200 µl using Amicon centripreps (Fisher, Springfield, NJ), and 20 µl of each sample were loaded on a 10% precast zymogram (gelatin) gel. The gel was run, renatured, and developed according to the manufacturer’s instructions (4)
. Gelatinase Zymography Standards were purchased from Chemicon (Temecula, CA).
|
RESULTS |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
. Several differentially expressed DNA fragments (expressed in CC4B low-grade SCC cells but not seen in CC4A high-grade SCC cells) were excised from the gel, purified, and reamplified using the corresponding set of anchoring and arbitrary primers (3)
. The reamplified cDNAs were sequenced and used as probes to screen Northern blots for differential expression. A transcript of 
2 kb was observed in CC4B but not in CC4A (data not shown). A DNA sequence Blast homology search was performed. This resulted in 120 bases of the 550-base-long query exhibiting a 92% homology to a region within the 3'-untranslated region of human VILIP-1 (VSNL1; GenBank accession no. AF 039555). Although CC4A and CC4B cells are of murine derivation, no homology to mouse or rat gene sequences was found. Thus, a genome walk was performed to obtain upstream sequences. After two rounds of genome walk, 99% homology to mouse VILIP-1 sequences was found (GenBank accession no. NM 012038).
|
is inversely proportional to cell aggressiveness as determined previously by histopathological analysis of xenotransplants and invasion assays. CC4A, CH72T3, and CarB produce aggressive spindle cell carcinomas when transplanted s.c. into nude mice and are also invasive in an in vivo invasion assay (3
, 11
, 12)
. Northern analysis showed that these three cells exhibited a remarkable loss of VILIP-1 expression (Fig. 1B)
. CC4B and CH72, which are less aggressive variants of CC4A and CH72T3, respectively, and were derived from the same original skin carcinomas (11
, 16)
, showed high VILIP-1 expression. In addition, two papilloma cell lines exhibited VILIP-1 expression (p117 and MT1/2; Fig. 1B
).
Western blot analysis, using a VILIP-1-specific antibody, showed a similar pattern of VILIP-1 expression as the Northern blot. No VILIP-1 was present in either CC4A or CH72T3; however, the less invasive counterparts of these two cells, CC4B and CH72, exhibited remarkable VILIP-1 expression (Fig. 1C)
. A clear ascending gradient of VILIP-1 expression could be noted from more invasive (CarB, CC4A, and CH72T3), expressing little or no VILIP-1 to the less invasive cell lines (CC4B, CH72, and MT1/2) and normal primary cultures of skin keratinocytes that expressed high levels of VILIP-1 protein. Intermediate expression levels were noted in the moderately invasive cell line p117. The comparison of the cell pairs (CC4A/CC4B and CH72/CH72T3) showed clearly that the less aggressive component (CC4B and CH72 cells) expressed much higher levels of VILIP-1 than their more aggressive counterparts (Fig. 1C)
.
In Vivo Expression of VILIP-1 in Normal and Neoplastic Epidermis.
To confirm that the expression of VILIP-1 decreases during skin tumor progression, we analyzed normal murine skin from 14 animals and 72 primary mouse skin tumors of different histopathological grades and aggressive potential. VILIP-1 was localized very specifically in certain normal cells, such as cortical neurons of the brain (Fig. 2A)
, and in the basal and some suprabasal cells of the intrafollicular epidermis, also known as infundibular epithelium (Fig. 2, B and C)
. Both antibody concentrations used gave similar results. Papillomas showed positive immunostain of all basal cells (Fig. 2D)
. Nevertheless, 64% of these benign tumors expressed low to moderate levels of VILIP-1. Approximately 35% of low-grade SCCs had no VILIP-1 expression, whereas another 37% exhibited low to moderate expression (Fig. 2E)
. The remaining low-grade SCCs were scored as negative for VILIP-1 expression. High-grade SCCs were mostly negative (Fig. 2F
; 9 of 14 tumors), whereas 4 of the remaining 5 tumors showed very low levels of VILIP-1 expression (Table 1)
. All specimens incubated with either the preimmune serum or blocking peptide exhibited negative immunostain.
|
|
.
Invasive Properties of SCC Cells.
In vitro invasion assays were performed with the cell pairs CC4A/CC4B and CH72T3/CH72. Confirming previous studies (3
, 11) , CH72T3 and CC4A were 4–5-fold more invasive than the endogenous VILIP-1 expressing CH72 and CC4B cells, respectively. To confirm that the invasion results obtained in parental murine SCC cells were caused by the presence of VILIP-1 in CC4B and CH72 cells and to the marked decrease of this protein in CH72T3 and CC4A cells, the latter cell lines were transfected with VILIP-1 cDNA. Cells were transfected with either vector alone (CH72T3.cin and CC4A.cin) or with vector containing the VILIP-1 (VSNL1) cDNA (CH72T3.VILIP-1 and CC4A.VILIP-1). Western blot of transfected cells showed expression of VILIP-1 in CH72T3.VILIP-1 and CC4A.VILIP-1 cells but no expression in CH72T3.cin or CC4A.cin cells. (Fig. 3A)
.
|
. This effect was less drastic in transfected CC4A cells, where the decrease in invasive ability caused by VILIP-1 was 35% (Fig. 3B)
.
The in vivo invasion assay showed that CH72T3 cells transfected with vector alone or VILIP-1 were invasive 3 weeks after transplantation (Fig. 3C)
. However, the VILIP-1-expressing transfectants did not penetrate as deeply into the peri-tracheal tissues as the vector-alone transfectants (65% reduction in invasiveness).
CC4A transfectants showed a similar pattern of in vivo invasiveness. The penetration of CC4A.VILIP-1 cells was decreased 33% with respect to the control-transfected CC4A cells. (Fig. 3C)
.
Fig. 4
shows that CH72T3.cin cells invaded into the peritracheal tissues (Fig. 4A)
, whereas CH72T3.VILIP-1 remained inside the trachea without invading the pars membranacea nor reaching the peritracheal tissues (Fig. 4B)
. CC4A transfected with vector alone invaded the tracheal wall and surrounding tissues (Fig. 4C)
, whereas CC4A cells transfected with VILIP-1, although growing into the surrounding tissues, did not invade as deeply as its vector alone-transfected counterparts (Fig. 4D)
.
|
50 and 30%, respectively, when compared with vector alone-transfected cells. The mean proliferative ratio between VILIP-1-transfected cells and vector alone-transfected cells was 0.52 ± 0.07 for CH72T3 cells and 0.69 ± 0.13 for CC4A cells (P < 0.05). Similarly, Ki67 LI of these cells grown in vivo as tracheal transplants (in vivo invasion assay) showed that the percentage of cycling cells was much lower in VILIP-transfected cells than in their vector alone-transfected counterparts. The LI of VILIP-transfected CH72T3 cells amounted to 5.5 ± 2.3, whereas it was 16.5 ± 4.2 for the corresponding control (P < 0.01). As for the CC4A cells, the VILIP transfectants exhibited an LI of 29 ± 5, whereas their corresponding vector alone transfectants the LI was 39 ± 6 (P < 0.05).
Effects of VILIP-1 Expression on cAMP.
Because VILIP-1 can affect the levels of intracellular cAMP (6
, 19)
, the concentration of this cyclic nucleotide was measured in the parental cells and in their respective transfected cells. CC4B and CH72 cells with endogenous VILIP-1 showed higher levels of intracellular cAMP than the more aggressive counterparts CC4A and CH72T3. (Fig. 5A)
.
|
). A similar, although smaller, difference was observed between vector alone and VILIP-1-transfected CC4A cells (Fig. 5C)
. The presence of forskolin (25 µM), a potent stimulator of adenylate cyclase and intracellular cAMP levels, further enhanced this difference between VILIP-1 transfectants and vector alone-transfected SCC cells (Fig. 5, B and C)
.
The effect of cAMP on the invasive ability of SCC cells was demonstrated when the in vitro invasiveness of transfectants was evaluated in the absence and presence of forskolin (Fig. 6, A and B)
. This experiment showed that after forskolin treatment, invasiveness of all CH72T3 and CC4A cells is reduced drastically. Although the differential effect between VILIP-1 and vector alone-transfected cells was not that obvious, a significant reduction could be noted with CH72T3 cells treated with 50 µm forskolin. In this specific case, invasiveness without forskolin was reduced 74% in the VILIP-1 transfectants vis a vis its vector alone-transfected counterpart, whereas with forskolin, this difference increased to 87% (Fig. 6A)
.
|
Western analysis of vector alone cells showed that CH72T3 cells were able to process pro-MMP-9 effectively, whereas VILIP-1-transfected cells (CH72T3.VILIP-1) were unable or not very efficient in processing this zymogen. The addition of forskolin to the medium reduced the ability of the cells to process MMP-9 (Fig. 7A)
. Gelatinase zymograms showed that CH72T3.cin cells exhibited higher MMP-9 activity than VILIP-1-transfected cells (CH72T3.VILIP-1 cells; Fig. 7B
).
|
, FBS stimulated a rapid and transient activation of Rho in vector alone-transfected CH72T3 cells. This increase in Rho activity was seen 10 min after stimulation, gradually decreasing thereafter. Interestingly, this activation, although visible (Fig. 8)
, was markedly diminished in VILIP-1-transfected cells.
|
|
DISCUSSION |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
Previous immunohistochemical studies (5) have described the localization of VILIP-1 in defined cell subpopulations of the CNS. In this report, we describe for the first time that VILIP-1 is very precisely localized in the basal cells of the normal intrafollicular epidermis, probably representing reserve or stem cells. In addition, all papillomas showed moderate to high expression in tumor basal cells. Although not conclusive, this observation supports the generally held assumption that many skin papillomas and papilloma-derived SCCs originate from the basal cells of the intrafollicular epidermis (20 , 21) . Moreover, the fact that high-grade SCCs exhibit a decreased or absent expression of VILIP-1 suggests that this molecular alteration occurs late during the process of tumor progression. Although VILIP-1 was originally found to be absent or decreased in high-grade SCC cell lines with a spindle shape morphology, after transfection of this gene into poorly differentiated SCC cell lines with little or no endogenous VILIP-1 expression, no major changes in histopathological grade or the expression of differentiation markers were seen in the resulting tumors of lower aggressiveness (data not shown).
VILIP-1 is an EF-hand calcium-binding protein belonging to the neuronal Ca2+-sensor protein family mainly expressed in the CNS. The founder of this family of proteins was isolated from chick brain and showed 40% homology to the retinal calcium-binding protein visinin and therefore was called VILIP (22) . VILIP orthologs in mouse, rat, and human are NVP-1 and VSNL1 (23 , 24) and have 100 and 96% homology to chick VILIP at the amino acid level.
Although most studies thus far have focused on the role of VILIP-1 in signal transduction in the CNS (5 , 8) , some studies point indirectly to other functions of this protein. This was explored using C6 glioma cells transfected with the wild-type VILIP-1 cDNA. VILIP expression simultaneously increased the expression of glial fibrillary acid protein, a marker of differentiation, together with the intracellular cAMP level (25) . In addition, an enhanced cAMP-dependent morphological differentiation of VILIP-transfected C6 cells was observed. In parallel, treatment with forskolin, an activator of adenylate cyclase, changed the gene expression pattern of C6 glioma cells, similarly resulting in increased glial fibrillary acid protein expression and a decrease in glutamic acid decarboxylase (25) . These studies pointed to the role of VILIP-1 in differentiation and maintenance of the mature cellular phenotype. VILIP-1 expression was found in chick, rat, mouse, and human brain (5 , 19 , 22 , 23) , in the retina of different species (26) , and in rat olfactory neurons (7) . The present study, using another ectodermal derived tissue, i.e., the epidermis and its tumors, suggests that this protein may well have a similar role in maintaining cell differentiation and normal cellular physiology in other cells types.
VILIP-1 is reported to be a calcium-binding protein associated with membranes under physiological calcium concentrations (22 , 27) . A well-known effect of VILIP-1 in brain cells is to modulate the intracellular levels of cAMP, both the basic and extracellularly induced cAMP concentrations. In olfactory membranes, after odor stimulation, VILIP-1 preparations inhibited adenylate cyclase activity in a calciumdependent manner (7) . Conversely, up-regulation of cAMP has been observed after transfection of C6 glioma cells with the full-length VILIP-1 cDNA (5 , 25) . The different regulatory effects of VILIP-1 in different cell systems is most likely explained by expression of different types of adenylyl cyclases. In C6 glioma cells, transfected with VILIP-1 cDNA, the protein was shown to increase the ß-adrenergic receptor- and forskolin-stimulated cAMP levels (25) . A similar effect was observed on basal cAMP levels without forskolin stimulation (19) . To evaluate the effect of VILIP-1 on cAMP accumulation in murine skin cancer cells, we investigated the cAMP levels in transfected and nontransfected SCC cells. In concordance with previous findings (19 , 25) , both endogenous and ectopic expression of VILIP-1 led to increases in the intracellular concentration of cAMP.
Because the parental pair of cell lines (CH72T3 and CH72 and CC4A and CC4B) with different VILIP-1 expression had differences in their invasive properties that correlated with their respective VILIP-1 expression levels, in vitro invasion assays were performed to evaluate the effects of ectopically expressed VILIP-1 on tumor cell invasiveness. As expected, the expression of VILIP-1 in two high-grade SCC cell lines (that had low or no expression of endogenous VILIP-1) decreased the invasive ability of tumor cells in parallel with an increase in cAMP levels.
Increased cAMP levels have been found to be a negative regulator of DNA synthesis and proliferation in normal epidermal keratinocytes (28) . More recently, Howe and Juliano (29) found that cAMP, acting through PKA, induces an anchorage-dependent cell phenotype by inhibiting mitogen-activated protein kinase signaling through p21 kinases (30) . Investigations focusing on cancer cell systems have indicated that up-regulation of cAMP not only inhibits cell proliferation but also induces increased cell differentiation and apoptosis (28, 29, 30, 31, 32, 33) . In this context, in VILIP-1-transfected C6 glioma cells, VILIP expression affects basal cellular cAMP homeostasis, thereby influencing differentiation of C6 cells (19) . In our models of squamous cell cancer cells, VILIP-1 overexpression had a definite effect on increasing cAMP levels and reducing invasiveness and proliferative rates without obvious alterations in morphological patterns.
Several other experiments have demonstrated that not only cAMP and its effector PKA negatively regulate cell proliferation but that it also plays a fundamental role in inhibiting or counteracting RhoA signaling (9 , 10 , 34) , as well as MMP activity (35) . In view of these possible connections, we investigated whether the decrease in cell proliferation seen in our VILIP-1 transfectants was also accompanied by changes in the activity and/or expression of other gene products associated to the invasion/metastatic cascade seemingly influenced by cAMP levels.
The MMPs’ family is instrumental for tumor cell progression (36) . MMP-9 is considered one of the most relevant metalloproteinases in murine skin carcinogenesis models (37) and may represent a key downstream factor triggered by VILIP-1 down-regulation, resulting in tumor cell invasiveness. Interestingly, cAMP elevation can result in decreased expression of several members of the MMP family, including MMP-9 (35 , 38) . McCawley et al. (35) demonstrated in a human SCC cell line that cAMP can interfere with epidermal growth factor-mediated MMP-9 induction and cell locomotion. In squamous tumor cells, VILIP-1-induced cAMP elevation also leads to a decrease in MMP-9 expression and activation. VILIP-1-expressing tumor cells showed decreased levels of MMP-9 activity that paralleled reductions in invasiveness. Furthermore, our findings indicate an important role for VILIP-1 in reducing RhoA activation. Rho and Rho kinase have been implicated in tumor cell invasion (39 , 40) , suggesting that VILIP-1-mediated regulation of Rho pathways could be an important component of carcinoma progression. Our results highlight the importance of cAMP metabolism in the inactivation of RhoA in a model of SCC progression. These results are in accordance with other reports that demonstrated the role of cAMP and PKA in the inhibition of RhoA activity (9 , 34 , 41) .
Taken together, these results indicate that levels of VILIP-1 expression correlate inversely with SCC cell invasive behavior and histopathological grade in vivo. The increase in VILIP-1 expression was accompanied by a parallel increase in cAMP levels that resulted in a decrease in MMP-9 expression, RhoA activity, and cell proliferation. Similar cascades of events were described previously in human keratinocytes overexpressing cAMP (35) and in other cell systems in which cAMP elevation resulted in RhoA inhibition (9 , 10 , 34) . The present contribution identifies VILIP-1 as a possible trigger of this cascade. The marked decrease in the expression of VILIP-1 found in high-grade murine SCC cells and its expression in benign tumors and normal epidermis points to a possible role for this protein in suppressing the invasive/metastatic phenotype by decreasing cell proliferation and intercellular matrix degradation/tumor cell invasiveness through a cAMP-mediated pathway.
|
ACKNOWLEDGMENTS |
|---|
|
FOOTNOTES |
|---|
1 Supported in part by NIH Grants CA75028, CA06927, and CA71539 and an appropriation from the Commonwealth of Pennsylvania. 
2 H. M. and A. M. G-G. contributed equally to this work.
3 Present address: Merck Research Laboratories, BLA-20, P. O. Box 4, West Point, PA 19486.
4 To whom requests for reprints should be addressed, at Department of Pathology, Fox Chase Cancer Center, 7701 Burholme Avenue, Philadelphia, PA 19111. E-mail: aj_klein-szanto{at}fccc.edu
5 The abbreviations used are: SCC, squamous cell carcinoma; DD, differential display; FBS, fetal bovine serum; LI, labeling index; PKA, protein kinase A; MMP, matrix metalloproteinase; cAMP, cyclic AMP; CNS, central nervous system; VILIP, Visinin-like protein.
Received 2/10/03. Revised 3/21/03. Accepted 5/16/03.
|
REFERENCES |
|---|
|
TopABSTRACTINTRODUCTIONMATERIALS AND METHODSRESULTSDISCUSSIONREFERENCES |
|---|
This article has been cited by other articles:
|
|
 |
|
  Y. Xie, H. Chan, J. Fan, Y. Chen, J. Young, W. Li, X. Miao, Z. Yuan, H. Wang, P. K.H. Tam, et al. Involvement of visinin-like protein-1 (VSNL-1) in regulating proliferative and invasive properties of neuroblastoma Carcinogenesis, October 1, 2007; 28(10): 2122 - 2130. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 C. F. Yen, H. W. Lin, J. C. Hsu, C. Lin, T. F. Shen, and S. T. Ding The Expression of Pituitary Gland Genes in Laying Geese Poult. Sci., December 1, 2006; 85(12): 2265 - 2269. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. F. Dai, Y. Zhang, Y. Kang, Q. Wang, H. Y. Gaisano, K.-H. Braunewell, C. B. Chan, and M. B. Wheeler The Neuronal Ca2+ Sensor Protein Visinin-like Protein-1 Is Expressed in Pancreatic Islets and Regulates Insulin Secretion J. Biol. Chem., August 4, 2006; 281(31): 21942 - 21953. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Fan, Y. Chen, H. M. Chan, P. K. H. Tam, and Y. Ren Removing intensity effects and identifying significant genes for Affymetrix arrays in macrophage migration inhibitory factor-suppressed neuroblastoma cells PNAS, December 6, 2005; 102(49): 17751 - 17756. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Brackmann, S. Schuchmann, R. Anand, and K.-H. Braunewell Neuronal Ca2+ sensor protein VILIP-1 affects cGMP signalling of guanylyl cyclase B by regulating clathrin-dependent receptor recycling in hippocampal neurons J. Cell Sci., June 1, 2005; 118(11): 2495 - 2505. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||
). Pull-down assays were analyzed by 12.5% SDS-PAGE and immunoblotting with antibodies against RhoA (top), alongside corresponding total cell extracts, for normalization of total RhoA protein levels (bottom). B, densitometric quantification of GTP-Rho relative to total RhoA protein.