Chromogranin a Expression in Neoplastic Cells Affects Tumor Growth and Morphogenesis in Mouse Models1

Home , Chromogranin A

[CANCER RESEARCH 62, 941–946, February 1, 2002] Chromogranin A Expression in Neoplastic Cells Affects Tumor Growth and Morphogenesis in Mouse Models1

Barbara Colombo, Flavio Curnis, Chiara Foglieni, Antonella Monno, Gianluigi Arrigoni, and Angelo Corti2 Departments of Biological and Technological Research [B. C., A. M., C. F., A. C.], and Histopathology [G. A.], San Raffaele H Scientific Institute, 20132 Milan, Italy

ABSTRACT many others (14, 20–22). CgA antigen, released in high amounts in the blood of patients, has proven to be a sensitive and specific serum Chromogranin A (CgA), a secretory protein expressed by many neu- marker for diagnosis of various types of neuroendocrine tumors. roendocrine cells, has been recognized as a useful tissue and serum Moreover, serum CgA is an independent marker of prognosis in marker of neuroendocrine tumors. To investigate the effect of CgA secretion on neoplastic morphogenesis and progression, we have transfected patients with carcinoid tumors (23, 24). mouse RMA lymphoma and TS/A adenocarcinoma cells with the cDNA The effect of this protein on tumor growth is unknown. In this work encoding human CgA and selected several CgA-positive (secreting) and we have studied the effect of CgA secretion on tumor growth and CgA-negative (nonsecreting) clones. In both models, the growth rate morphogenesis in two experimental animal models. To this aim, we of CgA-positive clones implanted s.c. in nude mice was slower than that of have transfected mouse RMA lymphoma and TS/A adenocarcinoma CgA-negative clones. Histological analysis of each RMA tumor showed cells (CgA-negative) with the cDNA encoding CgA and studied the that CgA-expression was associated with multinodular growth patterns, proliferation and tumorigenicity of several CgA-producing and non- whereas CgA-negative tumors appeared more compact and similar to producing clones in vitro and in vivo. We show that CgA secretion wild-type RMA tumors. Moreover, CgA production was associated with affects the tumor growth and tissue architecture. increased tumor necrosis. The number of nodules in each RMA tumor .(0.0004 ؍ P ,0.537 ؍ r ,40 ؍ correlated with the serum levels of CgA (n The reduced growth rate of CgA-positive RMA and TS/A tumors was not MATERIALS AND METHODS related to reduced in vitro proliferation or to changes in cell adhesion and shape, suggesting that the mechanism is indirect and host-mediated. These Natural CgA and Anti-CgA Antibodies. Mouse antihuman CgA mAbs results suggest that abnormal secretion of CgA by neuroendocrine neo- (mAb 5A8, B4E11, and A11) and rabbit anti-CgA polyclonal antibodies were plastic cells could affect neoplastic growth and morphogenesis. described previously (25–27). HSFs of human pheochromocytoma tissues and mouse adrenal glands were prepared as described previously (27). Human CgA was purified from pheochromocytoma tissues by immunoaffinity chromatog- INTRODUCTION raphy on the mAb A11-agarose column as described (28). CgA3 is a glycoprotein expressed by many neuroendocrine cells CgA Assays. CgA quantification in cell supernatants, cell extracts, and animal sera was carried out by sandwich ELISA using mAb B4E11 and and neurons (1–3). Under physiological conditions CgA is concen- polyclonal rabbit anti-CgA IgGs, as described (28). Samples were diluted trated and stored within secretory granules, and is released in the 2-fold with 0.15 M sodium chloride, 0.05 M sodium phosphate buffer (pH 7.3) extracellular environment together with coresident hormones. After containing 0.5% BSA, 2.5% normal goat serum, and 0.05% (v/v) Tween 20. secretion, CgA can reach the blood stream via the capillaries or the Rabbit IgGs were detected using a goat antirabbit IgG-horseradish peroxidase lymphatic vessels (2). Biochemical studies have shown that CgA is a conjugate and o-phenylenediamine, as a chromogenic substrate. A dose- polypeptide of 439 amino acids (4–7), characterized by several post- response curve, covering the range between 0.15 and 10 nM human CgA, was ␮ translational modifications including glycosylation, sulfation, and obtained. Spiking serum samples with 5 g/ml of CgA68–91 peptide, contain- phosphorylation (2, 3). Although the extracellular function of CgA is ing the B4E11 epitope (25), efficiently inhibited the ELISA signal, confirming not yet clearly understood, it is believed that this protein is a multi- the assay specificity for CgA. The assay does not detect chromogranin B (data valent precursor of several polypeptides that may exert autocrine, not shown). Moreover, the assay, based on mAb B4E11 as a “capturing” reagent, efficiently detected human CgA but not murine CgA (Fig. 1). Detec- paracrine, and endocrine effects (8). For instance, NH -terminal frag- 2 tion of murine CgA in mouse adrenal gland HSF was carried out using a ments released from the adrenal medulla and from sympathetic nerve similar assay based on an antibody that cross-react with both species (mAb terminals (9, 10) can suppress vasoconstriction in isolated blood 5A8; Fig. 1). vessels (8, 11, 12). Other fragments inhibit the secretion of hormones, Western blot analysis was carried out using mAb B4E11 and the ECL such as insulin, parathormone, and catecholamines, from neuroendo- Western Blotting kit (Amersham), as described (28). crine cells (13–16). Recent works have also shown that CgA and its Transfection of RMA Cells. RMA lymphoma cells of C57BL/6 origin

NH2-terminal fragments can regulate the adhesion of fibroblasts and (obtained from Dr. Paolo Dellabona, San Raffaele H. Scientific Institute, smooth muscle cells in vitro (17, 18) and may increase deposition of Milan, Italy) were cultured in RPMI 1640, 5% fetal bovine serum, 100 units/ml basement membrane components by mammary ductal epithelial cells penicillin, 100 ␮g/ml streptomycin, 0.25 ␮g/ml amphotericin B, 2 mM gluta- ␮ in vitro (19). mine, and 50 M 2-mercaptoethanol. The cDNA coding for the full-length human CgA, including the leader sequence, was prepared by reverse transcrip- CgA is abnormally expressed by various tumors, including pheo- tion-PCR on SK-N-BE cells total RNA as described (26). The amplified DNA chromocytoma, carcinoid tumors, medullary thyroid carcinoma, pan- (1389 bp) was subcloned into a SmaI-digested pUC19 vector and sequenced creatic islet cell tumors, small cell lung cancer, prostate cancer, and using an automatic DNA sequencer (Perkin-Elmer 373A). This plasmid was named pUC19/CgA. The CgA coding region was then cloned into the mam- Received 8/1/01; accepted 11/30/01. malian expression vector pRS1-neo using EcoRI and HindIII (pRS1Neo-CgA). The costs of publication of this article were defrayed in part by the payment of page pRS1Neo-CgA (15 ␮g) was electroporated into 107 RMA lymphoma cells charges. This article must therefore be hereby marked advertisement in accordance with using a Bio-Rad Gene Pulser apparatus (250 V; 960 ␮F). Transfected RMA 18 U.S.C. Section 1734 solely to indicate this fact. 1 Supported by Associazione Italiana Ricerca sul Cancro and Ministero della Sanita`of cells, surviving selection with 1.8 mg/ml geneticin, were selected and sub- Italy (Ricerca Finalizzata RF99.54). cloned. The supernatant of each clone was tested by CgA-ELISA. Five clones 2 To whom requests for reprints should be addressed, at San Raffaele H Scientific secreting CgA and five clones nonsecreting CgA were selected, amplified, and Institute, via Olettina 58, 20132 Milan, Italy. Phone: 39-02-26-43-48-02; Fax 39-02-26- stored for subsequent studies. 43-47-86; E-mail: [email protected]. 3 The abbreviations used are: CgA, chromogranin A; mAb, monoclonal antibody; HSF, The presence of human CgA cDNA in transfected cells was analyzed by heat stable fraction; TBS, Tris-buffered saline; WT, wild-type. PCR on genomic DNA using the following primers: 5Ј-TGCATGCGCTC- 941

calibration curve. The proliferation index of each clone was calculated by dividing the number of cells in wells cultured for 3 and 0 days. Cell Adhesion Assays. Adhesion of various RMA or TS/A clones to the plastic surface of microtiter plates was carried out as described (17, 18). Cells were stained with crystal violet. In Vivo Studies. In vivo studies on animals were performed according to the prescribed guidelines of the Ethical Committee of the San Raffaele H Scientific Institute. CD-1 nu/nu BR (nude) mice (Charles River Laboratories, Calco, Italy) were challenged with 2 ϫ 106 RMA or TS/A (CgA-negative or CgA-positive) cells, s.c. in the left flank. The tumor growth was monitored by measuring the size with calipers. The tumor volume was estimated by calcu- ϫ ϫ ϫ ␲ lating r1 r2 r3 4/3 , where r1 and r2 are the longitudinal and lateral radii, and r3 is the thickness of tumors protruding from the skin surface. Animals were killed at day 15. The tumor of each animal was then excised for morphological and immunohistochemical examination. Histochemistry and Immunohistochemistry. Each RMA tumor sample was sectioned in two parts. One part was fixed with formalin and embedded in paraffin. The other part was frozen in isopentane-liquid nitrogen and embedded in OCT (BDH Italia, Milan, Italy) for immunohistochemical analysis. Paraffin sections were cut (thickness, 4 ␮m) and stained with H&E or Gomori silver stain for morphological and histochemical evaluation of tumors. The presence of necrotic areas, tumor nodules, and protein casts was evaluated by microscopy analysis of H&E-stained sections and scored from 0 to 4. The amount Fig. 1. Cross-reactivity of mAb B4E11 and 5A8 with human and murine CgA. The (density-intensity) of collagen fibers was evaluated on Gomori-stained sections assay was carried out by ELISA using mAb B4E11 or 5A8 in the capturing step. Samples and graded from 0 to 3. Statistical analysis of the results was performed using include various amounts of protein extracts (HSF) of mouse adrenal glands (A) or human pheochromocytoma (B). mAb B4E11 does not cross-react with murine CgA. the Student unpaired t test (GraphPad software). The microvessels density of each tumor was assessed by immunohistochemical analysis of criocut sections (thickness, 6 ␮m) using the rat-antimouse CD31 mAb 01951 (PharMingen, San Diego, CA) as follows: acetone fixed CGCCGCTGTCCTGGC-3Ј (forward primer) and 5Ј-TCAGGATCCTCAT- sections were incubated with 0.3% hydrogen peroxide for 15 min to quench CAGCCCCGCCGTAGTGCCTGC-3Ј(reverse primer). Sequences were de- endogenous peroxidase and rinsed with 0.05 M Tris-HCl buffer, (pH 7.5; TBS). signed to include the ATG start codon in the forward primer and the 3Ј-end Each section was then incubated (15 min) with TBS containing 5% normal coding sequence of CgA in the reverse primer. Extraction and quantification of goat serum followed by 10 ␮g/ml anti-CD31 mAb in TBS containing 2% BSA genomic DNA were carried out according to standard protocols. The ratio (30 min at room temperature). Bound antibody was detected using a biotin- 260:280 nm of genomic DNA isolated from each clone was 1.8-2. PCR ylated goat-antirat immunoglobulins secondary antibody and avidin-biotin reactions were carried out in 50 ␮l (final volumes) containing 100 ng of complexes (Vectastain ABC Elite Kit; Vector Laboratories, Burlingame, CA) genomic DNA, 1 ␮M primers, 0.25 unit of Taq DNA polymerase, 0.2 mM of followed by a 3,3Ј-diamino-benzidine-tetrahydrocloride chromogenic solution deoxynucleotide triphosphates, 50 mM potassium chloride, 1.5 mM magnesium (Biogenex, San Ramon, CA). The sections were slightly counterstained with chloride, 5% DMSO, and 20 mM Tris-HCl (pH 8.3). After an initial incubation Harris hematoxylin. at 94°C for 3 min, the following temperature cycling was performed: 94°C for Microvessels density was then evaluated by counting the vessels stained by 45 s, 65°C for 45 s, and 72°C for 1 min (30 cycles), followed by 72°C for 10 CD31 in seven different fields/sections at various magnification (ϫ100, ϫ200, min. The PCR products were analyzed by 0.9% agarose gel electrophoresis and ϫ400). In addition, microvessels density was evaluated in areas with a using ethidium bromide staining. high vessel density, identified in each section at low magnification (ϫ40), and Transfection of TS/A Cells. TS/A cells from a BALB/c spontaneous counted at high magnification (ϫ400). mammary adenocarcinoma (29) were cultured in RPMI 1640, 10% fetal bovine ␮ ␮ serum, 100 units/ml penicillin, 100 g/ml streptomycin, 0.25 g/ml ampho- RESULTS tericin B, and 2 mM glutamine. pRS1Neo-CgA (4 ␮g) was mixed with 200 ␮l of 0.15 mg/ml of Lipofectin Reagent (Life Technologies, Inc.) in 0.15 M Transfection of RMA Cells with CgA cDNA Does Not Affect ␮ sodium chloride and incubated for 15 min at room temperature. Then, 100 l Their in Vitro Proliferation. To investigate the effect of CgA secre- 4 ␮ of this mixture was added to 10 TS/A cells and plated 1 day before in 200 l tion on tumor growth and morphogenesis, we have transfected mouse of culture medium. After 48 h of incubation, 1 mg/ml geneticin was added to RMA lymphoma cells with the cDNA coding for residues 1–439 of the culture. One week later, cells surviving selection were subcloned. The supernatant of each clone was tested by CgA-ELISA. Four CgA-secreting human CgA. Among the various clones that acquired geneticin- clones were obtained. resistance, we selected five clones (1E1, 1D3, 4C1, 2G10, and 3A6) TS/A cells transfected with the cDNA coding for the Thy 1.1 antigen were that secrete CgA in the culture supernatants (termed RMA “CgA- also prepared using the pRS1Neo-Thy 1.1 vector (30). positive” clones) and five clones (4H4, 3C3, 1F10, 4D10, and 1D10) In Vitro Proliferation Assay. Various clones of transfected cells were that do not secrete CgA (termed RMA “CgA-negative” clones; ELISA seeded into 96-wells microtiter plates at various densities (5, 10, 20, and detection limit, 0.3 nM; Fig. 2A). No evidence of CgA production by 40 ϫ 103 cells/well) in 200 ␮l of RPMI 1640 with 10% fetal bovine serum, 100 CgA-negative clones was obtained by Western blot analysis and units/ml penicillin, 100 ␮g/ml streptomycin, 2 mM glutamine, and 0.5 mg/ml ELISA of cell extracts (not shown). geneticin, and left to incubate at 37°C, 5% CO2, for 0 or 3 days. Living cells To assess whether the CgA cDNA was present in the CgA-negative were stained with a 5 mg/ml 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazo- clones we analyzed each clone by PCR using CgA-specific primers. A lium bromide solution (Calbiochem, San Diego, CA; 10 ␮l/well,2hat37°C). DNA fragment corresponding to CgA cDNA (1389 bp) was amplified The medium was gently aspirated using a multichannel pipette and replaced with 200 ␮l/well of DMSO. The absorbance was read using a microplate in clone 3A6 (CgA-positive) but not in the CgA-negative clones (Fig. reader at 570 nm (reference wavelength, 670 nm). In parallel, calibration 2C). These results suggest that the geneticin-resistant CgA-negative curves for each clone (cell number versus absorbance) were obtained with clones, despite the fact that they were efficiently transfected, do not known amounts of freshly seeded cells. The number of cells in each well was produce human CgA, because they lost the CgA cDNA. then obtained by interpolating the absorbance of each sample on the relevant The in vitro proliferation and cell morphology of each clone were 942

variable, ranging from 2 to ϳ40 nM, whereas no CgA was detected in the serum of animals bearing CgA-negative tumors. The growth and morphology of each RMA tumor was then evaluated. The volume of CgA-positive tumors was smaller than that of CgA-negative tumors (Fig. 4A). Macroscopically, CgA-positive tumors presented multinodular growth patterns, whereas CgA-negative tumors appeared frequently compact and similar to WT RMA tumors (Figs. 4B and 5, A–D). Interestingly, we observed a significant correlation between serum CgA levels and tumor nodules, in both Exp. 1 and 2 (Table 1). A weak correlation was observed also between tumor volumes and CgA levels in cell supernatants (r ϭ 0.27; P ϭ 0.012) but not with CgA levels in animal sera. This could be explained by the fact that serum levels may directly depend on the tumor mass, which is lower for clones that secrete high amounts of CgA. Histological examination of tissue sections showed the presence of necrotic areas and stroma collagen fibers in all of the samples, though to a different extent. In CgA-positive samples, the necrotic areas resulted more extended and frequently presented a big-branched core accompanied by other smaller areas of necrosis (Fig. 5, C–F). Tumor macro- and micronodules, completely or partially surrounded by extracellular matrix containing septa, were also visible. Tumor necro- Fig. 2. In vitro production of CgA (A) and proliferation (B) of RMA lymphoma clones transfected with human CgA cDNA. A, CgA production was assessed by measuring the CgA levels in the supernatant of each clone by ELISA after 3 days of culture (20 ϫ 103 cells seeded in 200 ␮l of culture medium). B, in vitro proliferation index of clones 1E1, 1D3, 4C1, 2G10, and 3A6 (CgAϩ), and clones 4H4, 3C3, 1F10, 4D10, and 1D105 (CgAϪ; n ϭ 5); bars, ϮSE. OOO, ELISA detection limit (0.3 nM). N.S., not significant by unpaired Student t test. C, PCR analysis of WT RMA cells, CgA-negative clones, and clone 3A6. PCR was carried out using primers specific for the human CgA cDNA (see “Materials and Methods”). The plasmid pRS1neo-CgA was included as positive control; molecular markers (M). then studied. No significant difference in proliferation of CgA-positive and -negative clones was detected in 3-day culture assays (Fig. 2B). Moreover, no difference in morphology or adhesion to culture plates, depending on CgA secretion, was observed (not shown). Because these cells were subjected to identical transfection, cloning, and culture conditions, they may represent a good model for investigating the effect of CgA secretion on tumor progression in vivo. CgA Production by RMA Tumor Cells Correlates with Tumor Growth and Morphogenesis in Vivo. To investigate the effect of CgA expression on the tumorigenicity of lymphoma cells, each RMA CgA-positive and -negative clone was implanted s.c. in nude mice. Immunodeficient animals were used to avoid T-cell-dependent im- mune responses against murine-transfected cells, because these cells express human CgA. Two separate experiments were done with RMA cells (called “Exp. 1” and “Exp. 2”), using 80 mice in total. All of the Fig. 4. Tumorigenicity of RMA CgA-positive and -negative clones. Tumor volumes (A) and number of macroscopically visible tumor nodules (B) of animals bearing CgA- animals bearing RMA CgA-positive tumors had elevated serological negative or -positive tumors (cumulative results of Experiment 1 and 2; bars, ϮSE). .by unpaired t test ,(ء) levels of human CgA by ELISA (Fig. 3, A–C). The CgA levels were CgAϪ versus CgAϩ, P ϭ 0.006

Fig. 3. Circulating levels of CgA in animals bearing RMA CgA-positive and -negative clones. A, CgA serum levels of mice bearing various tumors (Exp. 1). B, cumulative data of Experiment 1 and Experiment 2; bars, ϮSE.

943

Fig. 5. Morphology of RMA CgA-negative and -positive tumors after H&E staining (A–F) or Gomori- silver staining (H and I). WT RMA (A), RMA CgA- negative (B), and RMA CgA-positive tumors (C and D) at low magnification (ϫ100); microphotographs of RMA CgA-positive tumors at higher magnification (ϫ400, E, F, H, and I). Note the presence of nodules (nod)inC and D separated from the rest of the tumor mass by septa (s). F, intranodular necrotic areas (n). G, results of semiquantitative evaluation of various pa- rameters (Experiment 1, bars, ϮSE). H and I, Gomori- silver stained CgA-positive and CgA-negative sections, respectively (ϫ200) showing collagen fibers (arrows). L and M, representative microphotographs of CgA- positive (L) and CgA-negative (M) tissues immunostained with anti-CD31 mAb 01951 showing tumor vessels .(ء) P ϭ 0.0011 ;(ءء) arrowheads). P ϭ 0.0003)

sis and nodule formation, scored by microscopy analysis, were sig- Transfection of TS/A Cells with CgA cDNA Inhibits Tumor nificantly different between CgA-positive and CgA-negative tumors Growth in Vivo. To investigate the effect of CgA expression on the (Fig. 5G). tumorigenicity of adenocarcinoma cells, TS/A cells were transfected To investigate the effect of CgA secretion on tumor stroma forma- with cDNAs coding for CgA or Thy 1.1 (as a negative control). Four tion, we evaluated the amount of collagen fibers and vessel density in CgA-secreting clones were obtained (A6A, A6B, A5B, and A5C; Fig. each section. The amount of stromal fibers in non-necrotic areas of 6A). Of note, clone A6B secreted Ͼ14 nM of CgA in the supernatant. CgA-positive and -negative tumors was similar, as judged by semi- In contrast, no CgA was detected in the supernatant of a clone quantitative evaluation of Gomori silver-stained sections (Fig. 5G). Of transfected with the Thy 1.1 antigen (Thy 1.1) and of WT TS/A cells. note, thicker fibers were observed in CgA-positive tumors within or The in vivo growth rate of CgA-positive clones was markedly lower close to necrotic areas (Fig. 5H) and, to a lower extent, also in than that of controls, particularly in the case of clone A6B (Fig. 6B). CgA-negative tumors (Fig. 5I). The vessel density, assessed by count- The tumor dimension of CgA-positive tumors at day 17 was very ing the CD31-positive vessels in different areas of each tumor (Fig. 5, small, and histological examination at this stage revealed a single L and M), was not significantly different in the two groups. Protein nodule in most cases (not shown). To assess whether the reduced casts were also observed in both CgA-positive and -negative tumors to growth rate was directly related to changes in cell proliferation and the a similar level (Fig. 5G). adhesion, we analyzed the in vitro proliferation index and adhesion of each clone. The proliferation index of clone A6B but not of A6A, A58, and A5C was significantly lower than that of WT or Thy 1.1 Table 1 Correlation between serum CgA, nodule formation, and tumor volume of animals bearing various clones of RMA cells transfected with the cDNA coding controls (Fig. 6C). The adhesion of A6A and A5B but not of A6B and for CgA A5C to microtiter wells was stronger than that of controls (Fig. 6D). Serum CgAb vs. Thus, also in the case of TS/A cells no significant correlation was observed between tumorigenicity and in vitro proliferation, or be- Nodules Volume No. of No. of tween tumorigenicity and in vitro cell adhesion. clonesa mice rPrP Experiment 1 10 40 0.537 0.0004 0.031 NSc DISCUSSION Experiment 2 10 40 0.467 0.0027 0.170 NS Experiment 1 ϩ 2 10 80 0.318 0.0045 0.050 NS The results of this study suggest that CgA expression by tumor cells a Clone tested: 1E1, 1D3, 4C1, 2G10, 3A6, 4H4, 3C3, 1F10, 4D10, and 1D10 (four can affect tumor development and architecture. The study has been mice/clone in each experiment). b As measured by ELISA in the animal sera, 15 days after tumor implantation. carried out using genetically engineered mouse RMA lymphoma and c NS, not significant. TS/A adenocarcinoma cells expressing variable levels of human CgA 944

across the endothelial barrier. In recent studies, we also observed that CgA regulates the adhesion of fibroblasts and smooth muscle cells to solid phases (17, 18). Other studies showed that CgA at a nanomolar concentration may increase deposition of basement membrane components, such as collagen type IV, laminin, and perlecan, by mammary ductal epithelial cells and alter ductal morphogenesis in vitro (19). Thus, it is possible that CgA could affect the tumor architecture also by modulating the physiology of stromal fibroblasts within the tumor, which in turn are important for the production of other extracellular matrix proteins. The presence of thicker stromal fibers in CgA-positive tumors within or close to necrotic areas is apparently in line with this hypothesis. However, because this could also be explained by an increased fibrotic response to damaged tissues, it is difficult to draw conclusions on this point. The question is raised as to which is the cause of increased necrosis in RMA CgA-positive tumors. The finding that in vitro production of CgA does not affect its proliferation index, hence it is not cytotoxic, might suggests that the mechanism of necrosis is indirect. Hypoxia- related mechanisms dependent on changes in vascular density are unlikely, given the similar microvessel density in CgA-positive and -negative tumors. Possibly, necrosis could be a consequence of decreased vessel permeability and reduced supply of macromolecules critical for tumor cell survival. Alternatively, CgA could affect the production of other mediators by cells present in the tumor stroma, Fig. 6. CgA secretion (A), tumorigenicity (B), in vitro proliferation (C), and adhesion (D) of TS/A adenocarcinoma clones transfected with human CgA cDNA or Thy 1.1 which, in turn, are responsible for the increased necrosis. antigen. A, CgA secretion was assessed by measuring the CgA levels in the supernatant The results of this study could have some pathophysiological im- of each clone by ELISA as described for RMA cells in the legend of Fig. 1. B, tumor plications. CgA is present in the blood of normal subjects at 0.4–2nM growth of TS/A CgA-positive clones (A6A, A6B, A5A, and A5B), WT TS/A cells, and TS/A cells transfected with the Thy 1.1 cDNA. C, in vitro proliferation index of each (31, 32). Increased levels of CgA (Յ 200 nM) have been detected in clone. D, adhesion of TS/A clones to the plastic surface of microtiter plates after3hof the blood of a variety of patients with different neuroendocrine tumors incubation in the presence of 0.2% FCS. (33, 34). Given that serum CgA levels in our animal models were 2–40 nM, it is possible that the amount of CgA secreted by neuroendocrine tumors in patients is sufficient to affect tumor morphogenesis implanted in nude mice. The most striking observation is that CgA as we observed in our models. expression was associated with a decreased tumorigenicity in mice. Several studies have been carried out thus far to assess the corre- Moreover, CgA production was associated with increased tumor ne- lation among neuroendocrine differentiation, CgA expression, and crosis and multinodular growth pattern in RMA tumors but not in prognosis in patients with neuroendocrine and nonendocrine tumors. TS/A tumors. Interestingly, the number of tumor nodules significantly For instance, it has been shown that the expression of CgA decreases correlated with the serum levels of CgA. Because the CgA ELISA with increasing malignancy in neuroendocrine tumors, being higher in used in this study does not detect murine CgA, it is very likely that the well-differentiated carcinomas (low grade) and lower in poorly dif- circulating CgA antigen reflects the secretory activity of tumor cells ferentiated (high grade) carcinomas (31). Interestingly, one study and not the production of CgA by the neuroendocrine system of the showed that the number of CgA-immunoreactive neuroendocrine cells animals. Moreover, given that the half life of CgA is 18.4 min (20), remarkably decreases in invasive breast carcinomas compared with the presence of CgA in animal sera 15 days after tumor implantation noninvasive breast carcinomas (32). On the other hand, other studies indicates that this protein was secreted by tumor cells throughout the showed that neuroendocrine differentiation in prostate tumors could duration of the experiments. These notions, together with the observed be associated with poorer prognosis (33) and that large cell carcino- correlation between tumor morphology and circulating CgA, suggest mas of the lung with neuroendocrine features are more clinically that the different growth patterns of CgA-positive and -negative aggressive than classic large cell carcinomas (34). Because neuroen- tumors are related to locally produced CgA. docrine differentiation and CgA production could be associated with Studies on the mechanisms of action showed that CgA expression secretion of many other hormonal peptides, it is difficult to speculate does not affect the in vitro proliferation index of RMA cells, whereas on the role of CgA simply on the basis of these correlations. Our it affects the in vivo growth. This suggests that the effect is indirect results, showing slower progression of mouse mammary adenocarci- and host-mediated. One possibility is that CgA affects the complex noma after transfection with the CgA cDNA, suggest that indeed CgA interplay between neoplastic cells and tumor stroma, which is critical may contribute to regulate the growth of neuroendocrine tumors in a for tumor growth. Components of the tumor stroma include blood negative manner. vessels, inflammatory leukocytes, extracellular matrix molecules, and the cells necessary for their production. Noteworthy, we have found REFERENCES recently that CgA and its NH2-terminal fragments can inhibit vascular permeability.4 Thus, it is possible that CgA affects tumor growth by 1. Rosa, P., and Gerdes, H. H. The granin protein family: markers for neuroendocrine cells and tools for the diagnosis of neuroendocrine tumors. J. Endocrinol. Invest., 17: affecting the vascular compartment of the tumor, e.g., by decreasing 207–225, 1994. the transport of macromolecules critical for tumor cell proliferation 2. Winkler, H., and Fischer-Colbrie, R. The chromogranins A and B: the first 25 years and future perspectives. Neuroscience, 49: 497–528, 1992. 3. Simon, J. P., and Aunis, D. Biochemistry of the chromogranin A protein family. 4 E. Ferrero, E. Magni, C. Foglieni, F. Curnis, A. Villa, M. E. Ferrero, and A. Corti. Biochem. J., 262: 1–13, 1989. Chromogranin A protects vessels against tumor necrosis factor-2-induced vascular 4. Benedum, U. M., Baeuerle, P. A., Konecki, D. S., Frank, R., Powell, J., Mallet, J., and leakage, submitted for publication. Huttner, W. B. The primary structure of bovine chromogranin A: a representative of 945

a class of acidic secretory proteins common to a variety of peptidergic cells. EMBO membrane components by mammary epithelial cells in vitro. Biochem. Biophys. Res. J., 5: 1495–1502, 1986. Commun., 259: 563–568, 1999. 5. Konecki, D. S., Benedum, U. M., Gerdes, H. H., and Huttner, W. B. The primary 20. O’Connor, D. T., and Bernstein, K. N. Radioimmunoassay of chromogranin A in structure of human chromogranin A and pancreastatin. J. Biol. Chem., 262: 17026– plasma as a measure of exocytotic sympathoadrenal activity in normal subjects and 17030, 1987. patients with pheochromocytoma. N. Engl. J. Med., 311: 764–770, 1984. 6. Helman, L. J., Ahn, T. G., Levine, M. A., Allison, A., Cohen, P. S., Cooper, M. J., 21. Sobol, R. E., Memoli, V., and Deftos, L. J. Hormone-negative, chromogranin A- Cohn, D. V., and Israel, M. A. Molecular cloning and primary structure of human positive endocrine tumors. N. Engl. J. Med. 320: 444–447, 1989. chromogranin A (secretory protein I) cDNA. J. Biol. Chem., 263: 11559–11563, 22. Hsiao, R. J., Seeger, R. C., Yu, A. L., and O’Connor, D. T. Chromogranin A in 1988. children with neuroblastoma. Serum concentration parallels disease stage and predicts 7. Wu, H. J., Rozansky, D. J., Parmer, R. J., Gill, B. M., and O’Connor, D. T. Structure survival. J. Clin. Investig., 85: 1555–1559, 1990. and function of the chromogranin A gene. Clues to evolution and tissue-specific 23. Janson, E. T., Holmberg, L., Stridsberg, M., Eriksson, B., Theodorsson, E., Wilander, expression. J. Biol. Chem., 266: 13130–13134, 1991. E., and Oberg, K. Carcinoid tumors: analysis of prognostic factors and survival in 301 8. Helle, K. B., and Angeletti, R. H. Chromogranin A: a multipurpose prohormone? patients from a referral center. Ann. Oncol., 8: 685–690, 1997. Acta Physiol. Scand., 152: 1–10, 1994. 24. Eriksson, B., Oberg, K., and Stridsberg, M. Tumor markers in neuroendocrine tumors. 9. Liang, F., Dillen, L., Zhang, X. Y., Coen, E. P., Hogueangeletti, R., Claeys, M., and Digestion, 62: 33–38, 2000. Depotter, W. P. Vasostatins, N-Terminal products of chromogranin A, are released 25. Corti, A., Longhi, R., Gasparri, A., Chen, F., Pelagi, M., and Siccardi, A. G. Antigenic from the stimulated calf spleen in vitro. Acta Physiol. Scand., 155: 23–30, 1995. regions of human chromogranin A and their topographic relationships with structural/ 10. Helle, K. B., Marley, P. D., Angeletti, R. H., Aunis, D., Galindo, E., Small, D. H., and functional domains. Eur. J. Biochem., 235: 275–280, 1996. Livett, B. G. Chromogranin A: secretion of processed products from the stimulated 26. Corti, A., Sanchez, L. P., Gasparri, A., Flavio, C., Longhi, R., Brandazza, A., retrogradely perfused bovine adrenal gland. J. Neuroendocrinol., 5: 413–420, 1993. Siccardi, A., and Sidoli, A. Production and structure characterization of recombinant chromogranin A N-terminal fragments (vasostatins): evidence of dimer-monomer 11. Aardal, S., and Helle, K. B. The vasoinhibitory activity of bovine chromogranin A equilibria. Eur. J. Biochem., 248: 692–699, 1997. fragment (vasostatin) and its independence from extracellular calcium in isolated 27. Pelagi, M., Bisiani, C., Gini, A., Bonardi, M. A., Rosa, P., Mare, P., Viale, G., Cozzi, segments of human blood vessels. Regul. Pept., 41: 9–18, 1992. M. G., Salvadore, M., Zanini, A., Siccardi, A. G., and Buffa, R. Preparation and 12. Aardal, S., Galindo, E., Aunis, D., and Helle, K. B. Human chromostatin inhibits characterization of anti-human chromogranin A and chromogranin B (secretogranin I) endothelin-1-induced contractures in human blood vessels. Regul. Pept., 47: 25–32, monoclonal antibodies. Mol. Cell. Probes, 3: 87–101, 1989. 1993. 28. Corti, A., Gasparri, A., Chen, F. X., Pelagi, M., Brandazza, A., Sidoli, A., and 13. Mahata, S. K., O’Connor, D. T., Mahata, M., Yoo, S. H., Taupenot, L., Wu, H., Gill, Siccardi, A. G. Characterisation of circulating chromogranin A in human cancer B. M., and Parmer, R. J. Novel autocrine feedback control of catecholamine release. patients. Br. J. Cancer, 73: 924–932, 1996. A discrete chromogranin a fragment is a noncompetitive nicotinic cholinergic antag- 29. Nanni, P., de Giovanni, C., Lollini, P. L., Nicoletti, G., and Prodi, G. TS/A: a new onist. J. Clin. Investig., 100: 1623–1633, 1997. metastasizing cell line from a BALB/c spontaneous mammary adenocarcinoma. Clin. 14. Deftos, L. J. Chromogranin A: its role in endocrine function and as an endocrine and Exp. Metastasis, 1: 373–380, 1983. neuroendocrine tumor marker. Endocr. Rev., 12: 181–187, 1991. 30. Moro, M., Pelagi, M., Fulci, G., Paganelli, G., Dellabona, P., Casorati, G., Siccardi, 15. Tatemoto, K., Efendic, S., Mutt, V., Makk, G., Feistner, G. J., and Barchas, J. D. A. G., and Corti, A. Tumor cell targeting with antibody-avidin complexes and Pancreastatin, a novel pancreatic peptide that inhibits insulin secretion. Nature biotinylated tumor necrosis factor ␣. Cancer Res., 57: 1922–1928, 1997. (Lond.), 324: 476–478, 1986. 31. Helpap, B., and Kollermann, J. Immunohistochemical analysis of the proliferative 16. Russell, J., Gee, P., Liu, S. M., and Angeletti, R. H. Inhibition of parathyroid hormone activity of neuroendocrine tumors from various organs. Are there indications for a secretion by amino-terminal chromogranin peptides. Endocrinology, 135: 337–342, neuroendocrine tumor-carcinoma sequence? Virchows Arch., 438: 86–91, 2001. 1994. 32. Kimura, N. Chromogranins in non-endocrine tumours. Adv. Exp. Med. Biol., 482: 17. Gasparri, A., Sidoli, A., Sanchez, L. P., Longhi, R., Siccardi, A. G., Marchisio, P. C., 369–373, 2000. and Corti, A. Chromogranin A fragments modulate cell adhesion. Identification and 33. Young, R. H., Srigley, J. R., Amin, M. B., Ulbright, T. M., and Cubilla, A. L. characterization of a pro-adhesive domain. J. Biol. Chem., 272: 20835–20843, 1997. Carcinoma of the prostate gland (excluding unusual variants and secondary carcino- 18. Ratti, S., Curnis, F., Longhi, R., Colombo, B., Gasparri, A., Magni, F., Manera, E., mas). In: J. Rosai and L. H. Sobin (eds.), Atlas of Tumor Pathology, Ed. 3, Vol. 28, Metz-Boutigue, M. H., and Corti, A. Structure-activity relationships of chromogranin pp. 111–216. Washington DC: Armed Forces Institute of Pathology, 2000. A in cell adhesion. Identification and characterization of an adhesion site for fibro- 34. Iyoda, A., Hiroshima, K., Toyozaki, T., Haga, Y., Fujisawa, T., and Ohwada, H. blasts and smooth muscle cells. J. Biol. Chem., 275: 29257–29263, 2000. Clinical characterization of pulmonary large cell neuroendocrine carcinoma and large 19. Soriano, J. V., Pepper, M. S., Taupenot, L., Bader, M. F., Orci, L., and Montesano, cell carcinoma with neuroendocrine morphology. Cancer (Phila.), 91: 1992–2000, R. Chromogranin A alters ductal morphogenesis and increases deposition of basement 2001.

946

Downloaded from cancerres.aacrjournals.org on October 2, 2021. © 2002 American Association for Cancer Research. Chromogranin A Expression in Neoplastic Cells Affects Tumor Growth and Morphogenesis in Mouse Models

Barbara Colombo, Flavio Curnis, Chiara Foglieni, et al.

Cancer Res 2002;62:941-946.

Updated version Access the most recent version of this article at: http://cancerres.aacrjournals.org/content/62/3/941

Cited articles This article cites 32 articles, 7 of which you can access for free at: http://cancerres.aacrjournals.org/content/62/3/941.full#ref-list-1

Citing articles This article has been cited by 6 HighWire-hosted articles. Access the articles at: http://cancerres.aacrjournals.org/content/62/3/941.full#related-urls

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Subscriptions Department at [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://cancerres.aacrjournals.org/content/62/3/941. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.