Phenylacetate in Chemoprevention: In Vitro and in Vivo Suppression of 5-Aza-2’ .Deoxycytidine-induced Carcinogenesis1

Differentiation inducers selected for their low cytotoxic and genotoxic potential could be of major value in chemoprevention and maintenance therapy. We focus here on phenybacetate, a naturally occurring plasma component recently shown to affect the growth and differentiation of established neoplasms in experimental models. The ability of phenybacetate to prevent carcinogenesis by the chemotherapeutic hypomethylating drug 5-aza-2’-deoxycytidine (5AzadC) was tested in vitro and in mice. Transient exposure of immortalized, but poorly tumorigenic ras-transformed 4C8 fibroblasts to SAzadC resulted in neoplastic transformation manifested by loss of contact inhibition of growth, acquired invasiveness, and increased tumorigenicity in athymic mice. The latter was associated with elevation in ras expression and a decline in collagen biosynthesis. These profound phenotypic and molecular changes were prevented by a simultaneous treatment with phenylacetate. Protection from SAzadC carcinogenesis by phenylacetate was: (a) highly efficient despite DNA hypomethylation by both drugs, (b) free of cytotoxic and genotoxic effects, (c) stable after treatment was discontinued, and (d) reproducible in vivo. Whereas athymic mice bearing 4C8 cells developed fibrosarcomas following a single i.p. injection with 5AzadC, tumor development was significantly inhibited by systemic treatment with nontoxic doses of phenylacetate. Phenylacetate and its precursor suitable for oral administration, phenylbutyrate, may thus represent a new class of chemopreventive agents, the efficacy and safety of which should be further evaluated. INTRODUCTION The multistep nature of neoplastic transformation makes this disease process amendable to chemopreventive intervention. Several agents have been shown to inhibit carcinogenesis and thereby prevent the development of primary or secondary cancers (1-4). Of major interest are natural products and their analogues, including vitamins (A, B12, C, D3, and E), retinoids, and terpenes. These agents can suppress neoplastic transformation subsequent to a carcinogenic insult by regulating cell growth and differentiation. We focus here on the efficacy of the growth regulator phenylacetate. Phenylacetate is a common metabolite of phenylalanine implicated in growth control and differentiation in diverse organisms throughout phylogeny (5-7). Phenylacetate was mecently shown to induce tumor cytostasis and reversal of cancer in tissue culture and in animal models (8-1 1). Although high drug concentrations were required to produce antitumon effects in the experimental systems (1-5 msi, a 1000-fold above physiological human plasma levels), no cytotoxicity or genotoxicity was observed (8). Consistent with the laboratory findings, dinical experience obtained during phenylacetate treatment of patients with urea cycle disorders indicates that millimolar levels can be achieved in humans without significant adverse effects (12, 13). The demonstratable antitumor activities, lack of loxicity, and convenient p.o. administration prompted us to explore the role of phenylacetate in chemoprevention. The efficacy of phenylacetate as a chernopreventive agent was tested using in vitro and in vivo models of SAzadC3induced cancinogenesis. Despite the promise of SAzadC in the treatment of cancer and of 3-chain hemogbobinopathies, its clinical applications have been hindered by concerns regarding carcinogenic potential (14-16). Our preliminary studies with embryonic mouse C3H 1OT1/2 mesenchyrnal cells showed that high concentrations of phenylacetate (5-10 mM) prevent SAzadC-induced neoplastic transformation;4 these studies were however limited by the low frequency of transformation (7 X i0 ; see Ref. 8). The model used in the present studies involved pnemalignant munine fibnoblasts (cell lines 4C8 and PR4) which express a transcniptionally activated c-Ha-ras proto-oncogene. These poorly turnonigenic cells are highly susceptible to malignant conversion by pharmacological doses of SAzadC (17, 18). We show here that phenylacetate can protect such vulnerable cells from 5AzadC-induced carcinogenesis both in culture and in mice. MATERIALS AND METHODS Cell Cultures and Reagents. The subclones of mouse NIH 3T3 fibroblasts, PR4N and 4C8-A10 (designated here PR4 and 4C8), have been described previously (19, 20). Both cell lines are phenotypic nevertants isolated from bong terminal Received 8/24/94; revised 3/10/95; accepted 4/13/95. I This work was supported by funds from Elan Pharmaceutical Research Corporation through a Cooperative Research and Development Agreement (CACR-0139). 2 To whom requests for reprints should be addressed, at the Clinical Pharmacology Branch, National Cancer Institute, Building 10, Room 12C103, 9000 Rockvilbe Pike, Bethesda, MD 20892. 3 The abbreviations used are: 5AzadC, 5-aza-2’deoxycytidine; NaPA, sodium phenylacetate; NaPB, sodium phenylbutyrate; 5AzaC, 5-azacytidine; 5mC, 5 methylcytosine. 4 5. Shack and D. Samid, unpublished data. Research. on June 15, 2017. © 1995 American Association for Cancer clincancerres.aacrjournals.org Downloaded from 866 Phenylacetate in Chemoprevention nepeat/c-Ha-rasl-transfonmed 3T3 cells after long-term treatment with munine IFN-a/f3. Cultures were maintained in DMEM supplemented with 10% heat-inactivated FCS (GIBCO) and antibiotics. The sodium salts of phenylacetic and phenylbutynic acids (Elan Pharmaceutical Corporation) were dissolved in distilled water. 5AzadC (Sigma, St. Louis, MO) was dissolved in PBS and stoned in aliquots at -20#{176}C until use. Exposure of 5AzadC to direct light was avoided at all times to prevent drug hydrolysis. Treatments with SAzadC. For treatment in culture, cells were plated at 1-2 X iO cells in 100-mm dishes, and 5AzadC was added to the growth medium at 20 and 48 h later. The cells were subsequently subcultuned in the absence of the nucleoside analogues and observed for phenotypic alterations. For our in vitro studies, we used 0.1 p.M 5AzadC, a concentration within the pharmacological range that has previously been shown to efficiently transform the ras-expressing 3T3 cells (18). For in vivo treatment with 5AzadC, 6-9-week-old female athymic nude mice (Division of Cancer Treatment, National Cancer Institute Animal Program, Frederick Cancer Research Facility) were inoculated s.c. with 0.5 X 106 cells. Twenty-four h later 400 p.g of freshly prepared 5AzadC in 200 p.1 PBS were administered i.p. into each animal (approximately 20 mg/kg). Systemic treatment with NaPA is described in the text. Growth and Invasion through Matrigel. The ability of cells to degrade and cross tissue barriers was assessed by a qualitative in vitro invasion assay that utilized matnigel, a meconstituted basement membrane (Collaborative Research). Cells were exposed for 48 h in T.C. plastic dishes with 5AzadC alone on in combination with NaPA. NaPA treatment continued for an additional 1-2 weeks. Cells were then replated (at S x 10”! point) onto 16-mm dishes (Costar), which were previously coated with 250 p.1 matrigel (10 mg/mI). NaPA was either added to the dishes on omitted in order to determine the reversibility of effect. Net-like formation characteristic of invasive cells occunred within 12 h; invasion into the matnigel was evident after 6-9 days. Quantitative analysis of invasion was performed using a Biocoat Matnigel invasion chamber (Becton Dickinson Labware, Bedford, MA) according to the manufacturer’s instructions. Briefly, cells pretreated with the tested drugs in culture were plated onto the upper chamber at a density of 3 X io cells/well and incubated at 37#{176}C for 16-20 h. NIH 3T3conditioned medium was used as a chemoattractant. Filters were then removed, fixed with methanol, and stained with Giernsa. Cells attached to the upper side of the matrigel membrane were removed, and the number of invading cells found on the inner side was determined by microscopy. Tumor Formation in Athymic Mice. Cells were injected s.c. (5 x i05 cells/site) into 4-6-week-old female athymic nude mice (Division of Cancer Treatment, National Cancer Institute animal Program, Frederick Cancer Research Facility). The number, size, and weight of tumors were recorded after 3-4 weeks. For histological examination, tumors were excised, fixed in Bouin’s solution (picric acid:37% formaldehyde:glacial acetic acid, 15:5:1 v/v), and stained with hematoxylin and eosin. Measurement of DNA Methylation. To determine the 5-methylcytosine content, samples of cultures were taken 24 h after the second 5AzadC treatment. The cell pellets were lysed in 0.5% SDS, 0.1 M NaCl, 10 mrvi EDTA (pH 8.0), added with 400 p.g/ml proteinase K (Boehninger Mannheim), and stored at -70#{176}Cuntil DNA isolation and analysis. The content of methylated/unmethylated cytosine residues in the cellular DNA was measured by a 32P-postlabeling technique as described previously (19). Northern Blot Analysis and DNA Probes. Cytoplasmic RNA was extracted from exponentially growing cells and separated by electrophoresis in 1 .2% agarose-formaldehyde gels. RNA preparation, blotting onto nylon membranes (Schleichen and Schuell), hybridization with madiolabeled DNA probes, and autonadiognaphy were performed as described (18). The DNA probes included: 6.2-kb EcoRI fragment of v-Ki-ras, 2.9-kb Sad fragment of the human c-Ha-rasl gene, and a 4.5-kb BamHI fragment of the c-mvc gene. Glyceraldehyde phosphate dehydnogenase cDNA (20) was provided by M. A. Tainsky (University of Texas, Houston, TX), and a mouse transin eDNA was provided by G. T. Bowden (University of Arizona, Tucson, AZ). The eDNA probe for mouse histocompatibility class I antigens was a gift from G. Jay (NIH, Bethesda, MD). Radiolabeled probes were prepared with [32P]dCTP (New England Nuclear) using a random primed DNA labeling kit (Boehningem Mannheirn, Mannheim, Germany). RESULTS In Vitro Carcinogenesis Induced by 5AzadC and Its Prevention by Phenybacetate. Untreated 4C8 and PR4 formed contact-inhibited monolayers composed of epitheliallike cells. In agreement with previous observations (17, 18), transient exposure of these cultures to 0. 1 p.M 5AzadC during logarithmic phase of growth resulted in rapid and massive neoplastic transformation. Within 1 week of 5AzadC treatment, the great majority of the cell population became refractile and spindly in shape and formed multilayered cultures with increased saturation densities (Table 1), indicative of loss of contact inhibition of growth. These phenotypic changes could be prevented by the addition of NaPA. The effect was dose dependent: while 10 mM NaPA completely blocked neoplastic transformation (Table 1), lower doses had an intermediate effect (5 mM) on no significant effects (I mM), as determined by cell morphology and contact inhibition of growth. The profound effect of the aromatic fatty acid observed with the higher dose could not be explained by cytostasis or cytotoxicity per Se, as 10 mM NaPA caused only 45 ± 5% inhibition of 4C8 cell prolifenation, and cell viability remained over 95% (reminisecent of its effect on other nonmalignant cells; Ref. 8). Therefore, all further studies used NaPA at 10 ms . Several different regiments of NaPA treatment were found to be similarly effective. These included: (a) pretreatment with NaPA, starting 1 day prior to the addition of 5AzadC; (b) simultaneous exposure to both drugs; and (c) addition of NaPA 1 day after SAzadC. In all cases, cells were subsequently subjected to continuous treatment with NaPA for at least 1 week. Cells cultured under these conditions, like those treated with NaPA alone, formed contact-inhibited monolayers resembling untreated controls. These cells maintained the benign growth pattern for at least 3 weeks after NaPA treatment was discontinued. That NaPA can prevent neoplastic transformation was furthen indicated by the inability of cells to invade reconstituted Research. on June 15, 2017. © 1995 American Association for Cancer clincancerres.aacrjournals.org Downloaded from Clinical Cancer Research 867 Table I In vitro preven tion by phenylacetate of 5AzadC-induced carcinogenesis