ARTICLES. Rajendra G. Mehta, Robert M. Moriarty, Rajeshwari R. Mehta, Raju Penmasta, Gianluca Lazzaro, Andreas Constantinou, Liang Guo*

Transcription

1 ARTICLES Prevention of Preneoplastic Mammary Lesion Development by a Novel Vitamin D Analogue, 1 -Hydroxyvitamin D 5 Rajendra G. Mehta, Robert M. Moriarty, Rajeshwari R. Mehta, Raju Penmasta, Gianluca Lazzaro, Andreas Constantinou, Liang Guo* Background: The form of vitamin D (vitamin D 3 ) in fortified milk and the provitamin D produced by the body undergo metabolic activation to a biologically active form, 1,25- dihydroxyvitamin D 3 [1,25(OH) 2 D 3 ]. This compound can induce cell differentiation and can prevent proliferation of cancer cells. However, because 1,25(OH) 2 D 3 is hypercalcemic (effective in increasing serum calcium level), it is not suitable for use in cancer prevention or cancer therapy trials. Purpose: We synthesized a vitamin D 5 series analogue, 1 hydroxy, 24-ethyl-cholecalciferol, or 1 -hydroxyvitamin D 5 [1 (OH)D 5 ], and evaluated its chemopreventive activity in carcinogen-treated mammary glands in organ culture experiments. Methods: The analogue 1 (OH)D 5 was synthesized from sitosterol acetate and was characterized by nuclear magnetic resonance. Its purity was evaluated by high-pressure liquid chromatography. The calcemic activities of vitamin D 3 and D 5 analogues were determined in vitamin D-deficient Sprague-Dawley rats. Mammary glands of BALB/c mice were placed in organ culture and treated with the carcinogen 7,12-dimethylbenz[a]anthracene (DMBA) to induce preneoplastic lesions. Vitamin D analogues were added to the culture medium at four different concentrations, and formation of mammary lesions was evaluated. The effects of 1 (OH)D 5 and 1,25(OH) 2 D 3 on the expression of vitamin D receptors (VDRs) and transforming growth factor- 1 (TGF- 1) were studied by immunohistochemistry. Statistical significance was determined by the chi-squared test. All reported P values were two-sided. Results: 1,25(OH) 2 D 3 was fourfold more calcemic than 1 (OH)D 5 at a dose of µg/kg per day in rats. Both 1,25(OH) 2 D 3 and 1 (OH)D 5 inhibited the development of DMBA-induced preneoplastic lesions in mouse mammary glands compared with untreated glands. The effect of the vitamin D 3 analogue was observed at a much lower concentration (0.01 µm). Treatment with 1 (OH)D 5 resulted in a dose-related ( µm) inhibition without any toxicity, whereas the vitamin D 3 analogue was highly potent but toxic at concentrations of 1.0 µm or higher. Normal mouse mammary glands poorly express VDR and TGF- 1; incubation with 1 (OH)D 5 or 1,25(OH) 2 D 3 dramatically induced their expression. Conclusions: This is the first report showing the possibility of chemoprevention by a vitamin D 5 series compound. We conclude that 1 (OH)D 5 is less calcemic than 1,25(OH) 2 D 3.It is nontoxic at a wide range of concentrations, but it is potent in inhibiting the development of preneoplastic lesions in mammary glands in organ culture. In addition, we show for the first time the induction of TGF- 1 in normal mammary tissues by a chemopreventive agent. Implications: 1 (OH)D 5 is a good candidate for in vivo chemoprevention studies. It may mediate its action by inducing expression of VDR and of TGF- 1, as is seen in other systems. [J Natl Cancer Inst 1997;89:212-8] Vitamin D is a secosteroid and is classified as a hormone within a steroid hormone family (1,2). It has been separated on the basis of its chemical structure into different series, e.g., D 2, D 3,D 4,D 5, and D 6. To date, attention has been focused almost exclusively on the vitamin D 3 series of compounds. In its biologic form, vitamin D 3 is inactive unless it is metabolized to 1,25-dihydroxyvitamin D 3 [1,25(OH) 2 D 3 ]. The inactive 24- hydroxy form of the hormone is excreted from the body (3,4). The active metabolite 1,25(OH) 2 D 3 has been shown to suppress the growth in vitro of many neoplastic cells, including breast cancer cells (5,6). In addition, treatment of colon cancer cells and leukemia cells with 1,25(OH) 2 D 3 results in a reduction in the growth rate of these cells (7,8). One of the limiting factors in the successful use of vitamin D 3 in cancer prevention or cancer therapy is its calcemic activity (9). The concentrations of vitamin D 3 required to suppress growth of neoplastic cells cause hypercalcemia and death. Therefore, in recent years, numerous analogues of vitamin D have been synthesized that can reduce calcemic activity without compromising its antiproliferative activity (9,10). The differences in structures of these new compounds arise mostly from modifications in the C and D rings *Affiliations of authors: R. G. Mehta, R. R. Mehta, G. Lazzaro, A. Constantinou (Department of Surgical Oncology, College of Medicine), R. M. Moriarty, R. Penmasta, L. Guo (Department of Chemistry), University of Illinois, Chicago. Correspondence to: Rajendra G. Mehta, Ph.D., Department of Surgical Oncology, College of Medicine, University of Illinois, 840 S. Wood St. (M/C 820), Chicago, IL See Notes following References. 212

2 and side chain of the vitamin. No attempt, however, has been directed toward evaluating vitamin D compounds in other series. More work needs to be directed toward evaluating such compounds in other series. The effectiveness of a variety of chemopreventive agents has been evaluated by organ culture of the mouse mammary gland (11-13). In organ culture, mammary glands from mice respond to a short stimulation with a carcinogen in the presence of appropriate hormones by developing preneoplastic lesions (14). When implanted in syngeneic hosts, mammary cells prepared from these lesions give rise to adenocarcinomas (15). Effective chemopreventive agents (e.g., certain retinoids, selenium, oltipraz, and limonene) inhibit the formation of these lesions. The relative activity of chemopreventive agents in vitro correlates well with their activity in in vivo carcinogenesis experiments (13,16). Using this model system, we have evaluated the efficacy of 1 hydroxyvitamin D 5 [1 (OH)D 5 ] in preventing 7,12-dimethylbenz[a]anthracene (DMBA)-induced mammary lesion formation in a mouse mammary gland organ culture model. The mechanism of the vitamin D action is not completely understood. Nuclear vitamin D receptor (VDR) protein binding to 1,25(OH) 2 D 3 has been identified and is shown to be present in a variety of tissues, including normal mammary glands and mammary tumors, as well as in breast cancer cells (17,18). In the cytosol of target organs or cells, [ 3 H]1,25(OH) 2 D 3 binds specifically to receptors with a dissociation constant (K d ) ranging from M to M. An increased nuclear VDR concentration has been found to be associated with an enhanced expression of messenger RNA for vitamin D 3 receptors (19,20). The VDR gene has been cloned, and the molecular structure of the receptor protein has been determined. The results have demonstrated that the VDR belongs to the steroid-, thyroid-, and retinoid-receptor superfamily (21). All of these receptors act as ligand-dependent transcription factors that bind to specific DNA sequences. Two classes of response elements have been identified that are activated either by VDR alone or by heterodimers of VDRs and retinoid X receptor (RXR) alpha (22,23). In recent years, considerable attention has been given to the regulation of cell growth by autocrine antiproliferative factors [e.g., (24)]. Inhibition of cancer cell growth is often related to enhanced production of transforming growth factor- (TGF- ) (25,26). TGF- is further subclassified into the following three isoforms of polypeptides: TGF- 1, TGF- 2, and TGF- 3 (27). These isoforms are present in mammalian cells, including breast cancer cells. The isoforms of TGF- are regulated differentially by steroid and protein hormones. In one report (28), a hexafluoro analogue of vitamin D 3,1,25-dihydroxy-16-ene-23yne-26,27- hexafluorocholecalciferol (Ro ), induced expression in HL-60 human leukemia cells of TGF- 1 and its type 2 receptors. These results suggest a possible interaction between the function of VDR and TGF- regulation. Induction of TGF-, however, is often reported only in transformed cells. Although the growthinhibitory role of TGF- has been reported in the normal mammary gland (29), induction of TGF- in response to chemopreventive agents in this tissue has not been reported previously. In this study, we compare the chemopreventive action of vitamin D 5 with that of the active metabolite of vitamin D 3 and attempt to determine the possible mechanism of such action by studying the expression of VDRs and TGF- 1 in normal mammary epithelial cells. Materials and Methods Synthesis of 1 (OH)D 5. As a first step, -sitosterol acetate was converted to 7-dehydro- -sitosterol acetate by allelic bromination (using dibromantin and sodium bicarbonate in hexane and refluxing for 30 minutes) and dehydrobromination (using tetrabutylammonium fluoride, s-collidine, and tetrahydrofurane and refluxing for 6 hours). 7-Dehydro- -sitosterol was reduced to 7-dehydro-3 sitosterol by refluxing for 6 hours with lithium aluminum hydride and tetrahydrofurane. The reaction mixture was subjected to photolysis and then to thermolysis (by ethanol reflux) to yield vitamin D 5. We converted vitamin D 5 to 1 (OH)D 5 by following the Paaren DeLuca hydroxylation sequence (30). 1 (OH)D 5 was crystallized, and the crystalline product was fully characterized by 1 H nuclear magnetic resonance (NMR) at 400 Hz, mass spectroscopy (chemical ionization), and infrared (IR) and ultraviolet spectroscopy. The purity was determined by high-pressure liquid chromatography (HPLC). The following properties were recorded: melting point 150 C-152 C; IR (KBr) cm ( OH stretching) and 2955 ( CH stretching); 1 H NMR 6.35 (doublet [d], C-6H), 6.03 (d, C-7H), 5.33 (singlet [s], C-19H), 4.44 (multiple [m], C-1H), 4.24 (m, C-3H), and 0.55 (s, C-18H 3 ); mass spectrum (CI) 429 (M + 1, 72%), 411 (M+1 H 2 O, 100%), and 393 (M +1 2H 2 O, 14%); and UV max, 265 nm (molar extinction coefficient [ ] ). HPLC analysis of vitamin D analogues. Vitamin D analogues were dissolved in acetonitrile at a final concentration of 0.2 mg/ml. Aliquots (10 L) were injected on a Suplex PKB-100 HPLC column at ambient temperature. The HPLC was carried out with the use of an Hitachi L-6000 pump, an L-4200 UV-VIS detector, and an AS-2000 autosampler (Hitachi Instruments, Inc., Naperville, IL). Vitamin D analogues were eluted with the mobile phase of acetonitrile methanol water (52:30:18, vol/vol) with the flow rate at 1 ml/minute, and the elution profile was monitored at 254 nm. Measurement of calcemic activity. Three-week-old Sprague-Dawley male rats were obtained from the Holtzman Laboratory, Madison, WI. Up to three rats were housed together in a polycarbonate cage. The animal cages were kept under yellow light. The rats (eight to 10 per group per concentration of both vitamin D analogues used) were fed vitamin D-free diet containing 0.47 g/100 g calcium and 0.3 g/100 g phosphorus. After the rats had consumed this diet for 3 weeks, their plasma calcium levels were measured. Rats exhibiting plasma calcium levels of less than 6.0 mg/dl were considered to be vitamin D deficient. Such rats were administered appropriate vitamin D analogues intragastrically for 14 days. At the end of this period, the plasma calcium levels were again measured. Induction of preneoplastic lesions in mammary glands and their prevention by vitamin D 3 and D 5 analogues. Young, virgin BALB/c female mice, 3-4 weeks of age, were obtained from Charles River Laboratories, Wilmington, MA. The entire culture procedure was described in detail previously (12-14). Briefly, the mice were pretreated for 9 days with 17 -estradiol (1 g in 0.1 ml saline per animal) and progesterone (1 mg in 0.1 ml saline per animal). They were then killed by cervical dislocation, and the thoracic pair of mammary glands was dissected out on silk rafts and incubated for 10 days in Waymouth MB752 medium (Life Technologies, Inc. [GIBCO BRL], Gaithersburg, MD) containing the following growth-promoting hormones: insulin (5 g/ml), prolactin (5 g/ ml), aldosterone (1 g/ml), and hydrocortisone (1 g/ml). The carcinogen 7,12-dimethylbenz[a]anthracene (DMBA) at a dose of 2 g/ml was added to the medium on day 3 for 24 hours to induce mammary lesions. The DMBAcontaining medium was then removed, and the mammary glands were incubated for an additional 14 days with medium containing only insulin. This procedure allowed the normal glands to undergo structural regression in which all the normal alveolar structures were disintegrated. However, the alveolar lesions in the carcinogen-treated glands behaved differently. They had acquired altered hormone responsiveness, and these structures did not regress. These structures were termed mammary lesions. The vitamin D analogues (ranging in concentration from 0.01 M to 10.0 M) were included in the medium during the first 10 days of the in vitro culture to determine if they lowered the incidence of mammary lesion formation. Throughout the culture period, the glands were maintained at 37 C in an environment of 95% air and 5% CO 2. The procedure is presented schematically in Fig. 1. At the end of the culture period, the glands were fixed in formalin, stained in alum-carmine solution, and evaluated for the presence or absence of mammary lesions. All hormones and chemicals were purchased from the Sigma Chemical Co., St. Louis, MO. Downloaded Journal from of the National Cancer Institute, Vol. 89, No. 3, February 5, 1997 ARTICLES 213

3 Fig. 1. Protocol for mammary gland organ culture. Thoracic pairs of the mammary glands were removed from female BALB/c mice pretreated for 9 days with estrogen at a dose of 1 g per animal plus progesterone at a dose of 1 mg per animal (E + P). Glands were incubated with IPAH hormones (i.e., insulin [I] at a dose of 5 g/ml, prolactin [P] at 5 g/ml, aldosterone [A] at 1 g/ml, and hydrocortisone [H] at 1 g/ml for 10 days and with 7,12-dimethylbenz[a]- anthracene (DMBA) at a dose of 2 g/ml for 24 hours on the 3rd day in the presence or absence of the vitamin D 3 and the vitamin D 5 analogues as chemopreventive agents (see Materials and Methods section). Glands respond to the presence of hormones by developing alveolar structures. These normal alveolar structures disintegrate when the hormones are withdrawn from the medium. However, if the glands are exposed to DMBA on the 3rd day for 24 hours, as shown in the diagram, they do not disintegrate completely in the absence of hormones. These nonregressed alveolar structures are preneoplastic and are called mammary lesions. The growth of these structures does not require the presence of hormones. Immunohistochemistry of VDRs and TGF- 1. Normal mouse mammary glands were dissected and incubated with growth-promoting hormones either alone or in the presence of vitamin D analogues for only 3 days. In this experiment, the glands were not exposed to DMBA (see protocol described in the previous paragraph). The glands were fixed in buffered formalin, and 5- m-thick sections were prepared for histopathologic evaluations. The sections were mounted on adhesive-coated slides (Superfast; Fisher Scientific Co., Itasca, IL), dried at 60 C for 1 hour, deparaffinized in xylene, dehydrated in a series of alcohol, and finally washed with phosphate-buffered saline (PBS). To block the nonspecific antibody reactions, we treated the tissue sections with 5% dried skim milk for 10 minutes and then incubated them with primary mouse antibody (either against VDR or against TGF- 1; both obtained from BioGenex Laboratories, San Ramon, CA) overnight at 0-4 C. The tissues were rinsed in PBS and incubated with biotinylated rabbit anti-mouse antibody (Dako Corp., Carpenteria, CA) for 10 minutes; the remaining steps were followed according to the manufacturer-specified protocol; i.e., the reaction was stopped by rinsing the sections with PBS, which was followed by 10 minutes incubation with peroxidase-conjugated streptavidin, three 10-minute rinses with PBS, and a 5-minute incubation in a substrate, 3,3 -diaminobenzidine tetrachloride. The tissues were counterstained with hematoxylin eosin, dehydrated through graded series of alcohol and xylene, and finally mounted in Permount (Fisher Scientific Co.). Slides were evaluated for the presence or absence of the VDR or TGF- 1 and for the intensity of staining in the positively stained samples. Statistical analysis. Statistical significance was determined by the chi-squared test. All reported P values were obtained from two-sided tests. Results We synthesized highly purified 1 (OH)D 5. The chemical structures of 1,25(OH) 2 D 3 and 1 (OH)D 5 are shown in Fig. 2. The major differences between these two molecules are the lack of hydroxylation at the C-25 position and the presence of ethyl group at the C-24 position in the newly synthesized vitamin D 5 analogue. The purity of these compounds was evaluated by HPLC analysis. The purity specifications are described in the Materials and Methods section. As shown in Fig. 2, Fig. 2. Chemical structures and high-pressure liquid chromatography (HPLC) profiles of vitamin D analogues. The agents were dissolved in acetonitrile (200 g/ml), and 10- L aliquots were injected on a Suplex PKB-100 HPLC column. The HPLC conditions are described in the Materials and Methods section. The retention time for 1,25-dihydroxyvitamin D 3 [1,25(OH) 2 D 3 ] was about 5 minutes, whereas that for 1 -hydroxyvitamin D 5 [1 (OH)D 5 ] was about 34 minutes. Two plots were created with different chart speeds. Since the retention time was much longer for 1 (OH)D 5, the chart speed was reduced compared with that required for 1,25(OH) 2 D 3. 1,25(OH) 2 D 3 and 1 (OH)D 5 were eluted with retention times of approximately 5 and 34 minutes, respectively. Both of these vitamin D analogues exhibited about 98% purity. Stability studies have suggested that both can be stored in powder form for a year at 20 C, whereas in solution they are stable for 1 month at the same temperature. For all the experiments described in this article, stock solutions of vitamin D compounds were prepared fresh each time. One of the primary reasons to synthesize new vitamin D agents is to prepare analogues that have reduced calcemic activity compared with that of 1,25(OH) 2 D 3, but without compromising the antiproliferative activity. We measured the calcemic activity of both of these vitamin D analogues. As shown in Table 1, the vehicle-treated control rats showed a plasma calcium concentration of 5.4 ± 0.3 mg/dl (mean ± standard deviation). When the rats were treated with vitamin D analogues at a dose of g/kg per day, the following plasma calcium concentrations were observed: 6.0 ± 0.63 mg/dl for 1 (OH)D 5 -treated rats (11% increase over that of the vehicle-treated control group, 214

4 Table 1. Effects of vitamin D analogues on plasma calcium levels in vitamin D-deficient rats Treatment* No. of rats Dose, g/kg per day Plasma calcium, mg/dl P None ± (OH)D ± ± ,25(OH) 2 D ± ± 1.8 <.0001 *1 (OH)D 5 1 -hydroxyvitamin D 5 ; 1,25(OH) 2 D 3 1,25- dihydroxyvitamin D 3. Values means ± standard deviation. Two-sided test. P.121, not statistically significant when compared with that of the control group) and 8.1 ± 1.2 mg/dl for 1,25(OH) 2 D 3 - treated rats (50% increase over that of the control group, P.001, statistically significant difference when compared with that of the control group). At a higher concentration of vitamin D analogues (0.25 g/kg per day), 1 (OH)D 5 treatment resulted in a plasma calcium concentration of 7.9 ± 1.5 mg/dl, compared with 10.1 ± 1.8 mg/dl for 1,25(OH) 2 D 3 treatment. Although both analogues at this concentration increased the plasma calcium levels in comparison with those in vehicle-treated control rats, these results showed that 1 (OH)D 5 had overall lower calcemic effects than 1,25(OH) 2 D 3.1,25(OH) 2 D 3 treatment resulted in an 87% increase in the plasma calcium level in rats when compared with the vehicle-treated rats. On the other hand, in animals treated with a higher concentration of 1 (OH)D 5, there was only a 50% increase in the plasma calcium concentration compared with that in the control animals. These results suggest that 1 (OH)D 5 is much less calcemic than 1,25(OH) 2 D 3. To evaluate the efficacy of the newly synthesized vitamin D 5 analogue in preventing mammary lesion formation, we incubated 15 mammary glands per group (135 glands in total) from BALB/c mice with appropriate hormones and exposed the glands to DMBA on day 3 for 24 hours (see Materials and Methods section). The mammary glands were incubated for 10 days with the vitamin D analogues in concentrations ranging from 0.01 M to 10.0 M. The incidence of mammary lesions was calculated for each group and was reported as the ratio of the number of mammary glands showing lesions to the total number of mammary glands at risk. Table 2 shows the incidence of mammary lesions in the various groups treated with vitamin D analogues. There was a dose-related decrease in the number of glands exhibiting lesions in the vitamin D 5 -treated group. In contrast, in the group treated with 0.01 M vitamin D 3, only two of 14 glands developed lesions. At higher concentrations of this analogue, no mammary lesions were observed. We calculated the percent inhibition of formation of lesions in each treatment group by comparing the incidences of lesions between the control group and the treatment group. As shown in Fig. 3, both 1 (OH)D 5 and 1,25(OH) 2 D 3 inhibited the formation of mammary lesions by nearly 100%. At a concentration of 0.01 M, the vitamin D 3 analogue inhibited mammary alveolar lesion formation by 75%; incubation of glands with concentrations of 0.1 M and higher showed 100% inhibition. In contrast, the vitamin D 5 analogue inhibited the lesion formation in a dose-dependent manner, reaching 100% inhibition at a concentration of 10.0 M. 1,25(OH) 2 D 3 was toxic to the glands at concentrations of 1.0 M or higher. Dilation of ducts and disintegration of the morphologic structures were observed. Treatment of mammary glands with 1 (OH)D 5 did not result in any toxicity to the glands. To determine the effects of vitamin D analogues on the structural differentiation as well as their toxic effects on mammary glands, we incubated mammary glands with growth-promoting hormones for 3 days either alone or in the presence of 0.1 M or 1.0 M vitamin D analogues. The histopathologic characteristics of the treated and untreated glands are shown in Fig. 4. The control mammary gland structure was represented by normal alveolar and ductal structures (Fig. 4, A and B, respectively). 1,25(OH) 2 D 3 at a concentration of 0.1 M did not show toxicity (Fig. 4, C). Mammary glands displayed normal ductal and alveolar structures. At a concentration of 1.0 M, vitamin D 3 analogue treatment resulted in disintegration of ducts and structural toxicity to the glands (Fig. 4, D). In contrast, treatment with the vitamin D 5 analogue at a concentration of 1.0 M retained the healthy structural characteristics, as seen in the untreated glands. In fact, some secretion was obvious in the lumen of the ducts (Fig. 4, E). Since the role of chemopreventive agents (including analogues of vitamin D 3 and vitamin D 5 ) on the induction of TGF- in normal mammary epithelial cells has not been studied, the histologic sections of normal mammary glands treated with either only hormones (insulin, progesterone, aldosterone, and hydrocortisone) or hormones plus vitamin D analogues were processed immunohistochemically to investigate the effects of Table 2. Effects of vitamin D analogues on incidence of 7,12-dimethylbenz[a]anthracene-induced lesions in BALB/c mouse mammary glands in organ culture 1 -Hydroxyvitamin D 5 1,25-Dihydroxyvitamin D 3 Concentration, M No. of glands with lesions/total No. of glands treated % incidence* P No. of glands with lesions/total No. of glands treated % incidence* P None (control) 9/ / / / / / / / / / *(Number of glands with lesions/total number of glands treated) 100. Two-sided test. Statistical significance when compared with control group. Downloaded Journal from of the National Cancer Institute, Vol. 89, No. 3, February 5, 1997 ARTICLES 215

5 Fig. 3. Inhibition of 7,12-dimethylbenz[a]anthracene-induced mammary lesions in BALB/c mouse mammary gland organ culture. Mammary glands were incubated with increasing concentrations of vitamin D 3 and vitamin D 5 analogues for the first 10 days of a 24-day culture. After the 24-day culture, glands were fixed, stained, and scored for alveolar lesions. Percent inhibition was calculated by the formula [1 (% incidence in the treated group/% incidence in the control group) 100]. Closed circles 1 -hydroxyvitamin D 5 ; closed squares 1,25- dihydroxyvitamin D 3. vitamin D analogues on the induction and localization of VDRs and TGF- 1. As shown in Fig. 5, VDRs were localized in the nuclei of mammary epithelial cells. There was no selective localization of VDR in ductal or alveolar cells. Treatment with either 1.0 M 1 (OH)D 5 or 0.1 M 1,25(OH) 2 D 3 induced expression of VDRs detectable in the nuclei of both ductal and alveolar cells (Fig. 5, B and C, respectively). This induction was dependent on the concentration of the analogue; VDR induction was much less at the lower concentration of the vitamin D 5 analogue (data not shown). For the vitamin D 3 analogue, intense staining was evident at a lower concentration (0.1 M) (Fig. 5, C). However, at a concentration of 1.0 M, reduced or absent staining was observed as a result of apparent toxicity (Fig. 5, D). The effects of the vitamin D analogues on the induction of TGF- 1 were also evaluated. We studied tissue sections from the mammary glands treated with the vitamin D analogues or those from untreated control glands for the induction of TGF- 1. We found extensive induction of TGF- 1 in the cytoplasm of mammary epithelial cells (Fig. 6). Again, the pattern of intensity was comparable to that of induction of VDR. The extent of induction of TGF- 1 after treatment with the vitamin D 5 analogue at a concentration of 1.0 M (Fig. 6, B) was similar to that observed with the vitamin D 3 analogue at a concentration of 0.1 M (Fig. 6, C). However, at a concentration of 1.0 M of the vitamin D 3 analogue, TGF- 1 expression was much reduced as a result of toxicity (Fig. 6, D). These results indicate that the vitamin D 5 Fig. 4. Effects of vitamin D analogues on the structural differentiation of mammary glands in vitro. Mammary glands were incubated with insulin, prolactin, aldosterone, and hydrocortisone for 3 days either alone or with vitamin D analogues. The glands were not treated with 7,12-dimethylbenz[a]anthracene and did not undergo regression by hormone withdrawal. The glands were fixed in formalin and processed for histopathologic evaluation. The sections were stained with hematoxylin eosin. Normal alveolar and ductal structures are shown in panels A and B, respectively. Panels C and D represent treatment of the glands with vitamin D 3 analogue at concentrations of 0.1 and 1.0 M, respectively. (Note the dilation and disintegration of the ducts in panel D, whereas no toxicity was observed at 0.1 M 1,25- dihydroxyvitamin D 3, as shown in panel C.) Panel E represents incubation of glands with a concentration of 1.0 M of 1 -hydroxyvitamin D 5. No toxicity was observed (original magnification for panels A-E 40). analogue is considerably less toxic than the vitamin D 3 analogue. Moreover, they indicate that this remarkable induction of TGF- 1 in mammary epithelial cells by the vitamin D 5 analogue may be of importance in cancer chemoprevention. Discussion In recent years, considerable attention has been directed to the process of preventing carcinogenesis. The term chemopreven- 216

6 Fig. 5. Effects of vitamin D on the induction of vitamin D receptors (VDRs) in mammary gland organ culture. Mammary glands were incubated with 1.0 M 1 -hydroxyvitamin D 5 [1 (OH)D 5 ] or 0.1 and 1.0 M 1,25-dihydroxyvitamin D 3 [1,25(OH) 2 D 3 ] for 3 days. Glands were fixed and processed for histopathologic evaluation. Sections (5 m) were processed for VDR immunohistochemistry as described in the Materials and Methods section. Increased VDR expression in the nuclei of normal mammary alveolar epithelial cells was observed in both vitamin D 3 - and vitamin D 5 -treated glands. Control (A); 1 (OH)D 5, 1.0 M (B);1,25(OH) 2 D 3, 0.1 M (C); and 1,25(OH) 2 D 3, 1.0 M (D) (original magnification for panels A-D 40). Fig. 6. Effects of 1 -hydroxyvitamin D 5 [1 (OH)D 5 ] on the induction of transforming growth factor- 1 (TGF- 1) in mammary gland organ cultures. Mammary glands were incubated with vitamin D analogues and processed as described in Fig. 5 legend. The glands were incubated with anti-tgf- 1 antibodies as described in the Materials and Methods section and were evaluated for TGF- 1 expression. Increased cytoplasmic expression of TGF- 1 is seen in the normal alveolar epithelial cells in vitamin D 5 -treated glands. Control (A); 1 (OH)D 5, 1.0 M (B); 1,25-dihydroxyvitamin D 3 [1,25(OH) 2 D 3 ], 0.1 M (C); and 1,25(OH) 2 D 3, 1.0 M (D) (original magnification for panels A-D 40). tion was coined to denote the prevention of cancer by purified and well-characterized chemical agents (31). Numerous chemopreventive agents have been identified that can arrest the process of either initiation or progression of the disease. These agents include retinoids, thiols, inhibitors of polyamine biosynthesis and prostaglandin biosynthesis, and antihormones. An active metabolite of vitamin D 3,1,25(OH) 2 D 3, inhibits the proliferation of many cell types (6-8); moreover, it is a potent inducer of differentiation, as is seen in HL-60 cells (3). However, the use of this analogue to treat cancer patients is precluded because of its calcemic activity. It has been well established that vitamin D treatment results in an increased concentration of calcium in plasma of experimental animals (9). Numerous structural modifications have been reported to generate less calcemic vitamin D analogues with increased antiproliferative or differentiating ability (10). As this article demonstrates, we synthesized a novel vitamin D 5 analogue and determined its calcemic activity. We found that 1 (OH)D 5 exhibited lower toxicity than 1,25(OH) 2 D 3. Nonetheless, 1 (OH)D 5 is not completely devoid of calcemic activity. There are at least two specific vitamin D 3 analogues, Ro , a hexafluoro derivative of vitamin D 3 (28), and 22-oxa-calcitriol (32), that have been shown to be effective against mammary carcinogenesis. Ro is non-calcemic and is capable of inhibiting carcinogen-induced mammary lesions in organ culture; however, it was toxic at a concentration of 0.1 M in organ culture (data not shown), and the toxicity was observed when the analogue was given in vivo at a dose of 10 g/kg diet. Since several hundred analogues of vitamin D 3 have been evaluated without much success (10), the logical step toward finding an ideal compound was to evaluate the vitamin D analogues found within the D 4,D 5, and D 6 series. Various reports (33,34) have shown that TGF- 1 is a negative growth regulator for tumor cells. Increased expression of TGF- 1 is directly related to a reduced rate of growth of cancer cells in vitro. Several chemopreventive agents induce TGF- 1 in vitro in transformed cells, including breast cancer cells (25). In a report (28), a non-calcemic analogue of vitamin D 3, Ro , induced expression of TGF- 1 and its type 2 receptors. Similarly, chemopreventive agents have been shown to induce TGF- 1 in breast cancer cells in culture (34). There is no report, however, yet indicating induction of TGF- 1 in normal mammary epithelial cells. Induction of TGF- 1 by chemopreventive agents in normal cells may be more relevant to chemoprevention Downloaded Journal from of the National Cancer Institute, Vol. 89, No. 3, February 5, 1997 ARTICLES 217

7 than induction of TGF- 1 in transformed cells. In this study, we have shown that TGF- 1 could be induced by 1 (OH)D 5 in mammary ductal cells and mammary alveolar cells. Numerous studies have reported that the action of vitamin D is mediated by nuclear VDRs. Our results are consistent with those published (19-21,23). Our study shows that the analogue of the vitamin D 5 series induces nuclear VDRs in the mammary glands in in vitro culture. Both 1 (OH) 2 D 3 and 1 (OH)D 5 induced VDRs in mammary epithelial cells. The induction of VDR with 1.0 M 1 (OH)D 5 was much greater than with the nontoxic concentration of 1,25(OH) 2 D 3. Since the mammary epithelial cells (e.g., HBL100), which lack VDRs, also fail to respond to 1 (OH)D 5 and do not show induction of TGF- 1 (data not shown), the results presented in this article indicate that the chemopreventive activity of the vitamin D analogue may be mediated by induction of TGF- 1 in mammary tissue and that the vitamin D action may require VDRs. These results also point out that this remarkable induction of TGF- 1 in mammary epithelial cells by vitamin D analogues may be of significant importance in cancer chemoprevention. Experiments presented here constitute the first step toward our long-term goal of investigating the efficacy of chemoprevention by and the mechanism(s) of action of analogues of the vitamin D 5 series of compounds. Reduced calcemic activity and lack of toxicity make 1 (OH)D 5 an attractive candidate for in vivo chemoprevention studies. References (1) Bell NH. Vitamin D endocrine system. J Clin Invest 1985;76:1-6. (2) Haussler MR, Mangelsdorf DJ, Komm BS, Terpening CM, Yamaoka K, Allegretto EA, et al. Molecular biology of the vitamin D hormone. Recent Prog Hormone Res 1988;44: (3) DeLuca HF. New concepts of vitamin D functions. Ann N Y Acad Sci 1992;669: (4) Henry HL, Norman AW. Vitamin D metabolism and biological action. Ann Rev Nutr 1989;4: (5) Abe J, Nakano T, Nishii Y, Matsumoto T, Ogata E, Ikeda K. A novel vitamin D 3 analog, 22-oxa-1,25-dihydroxyvitamin D 3, inhibits the growth of human breast cancer cells in vitro and in vivo without causing hypercalcemia. Endocrinology 1991;129: (6) Skowronski RS, Pechl DM, Feldman D. Vitamin D and prostate cancer: 1,25-dihydroxy vitamin D 3 receptors and action in human prostate cell lines. Endocrinology 1992;132: (7) Abe E, Miyaura C, Sakagami H, Takeda M, Konno K, Yamazaki T, et al. Differentiation of mouse myeloid leukemia cells induced by 1 alpha,25- dihydroxyvitamin D 3. Proc Natl Acad Sci U S A 1981;78: (8) Kragballe K. Vitamin D 3 and skin diseases. Arch Dermatol Res 1992;284 Suppl 1:S30-6. (9) DeLuca HF. The metabolism and functions of vitamin D. Adv Exp Med Biol 1986;196: (10) Bouillon R, Okamura WH, Norman AW. Structure function relationship in the vitamin D endocrine system. Endocrine Rev 1995;16: (11) Mehta RG, Steele V, Kelloff GJ, Moon RC. Influence of thiols and inhibitors of prostaglandin biosynthesis on the carcinogen-induced development of mammary lesions in vitro. Anticancer Res 1991;11: (12) Dickens MS, Custer RP, Sorof S. Retinoid prevents mammary gland transformation by carcinogenic hydrocarbon in whole-organ culture. Proc Natl Acad Sci U S A 1979;76: (13) Chatterjee M, Banerjee MR. Influence of hormones on N-(4-hydroxyphenyl) retinamide inhibition of 7,12-dimethylbenz[ ]anthracene transformation of mammary cells in organ culture. Cancer Lett 1982;16: (14) Lin FK, Banerjee MR, Crump LR. Cell cycle-related hormone carcinogen interaction during chemical carcinogen induction of nodule-like mammary lesions in organ culture. Cancer Res 1976;36: (15) Telang NT, Banerjee MR, Iyer AP, Kundu AB. Neoplastic transformation of epithelial cells in whole mammary gland in vitro. Proc Natl Acad Sci U S A 1979;76: (16) Mehta RG, Liu J, Constantinou A, Thomas CF, Hawthorne M, You M, et al. Cancer chemopreventive activity of brassinin, a phytoalexin from cabbage. Carcinogenesis 1995;16: (17) Carlberg C, Bendik I, Wyss A, Meir E, Sturzenbecker LJ, Grippo JF, et al. Two nuclear signalling pathways for vitamin D. Nature 1993;361: (18) Lamberg-Allardt C, Valtonen E, Polojarvi M, Stewen P. Characterization of a 1,25-dihydroxyvitamin D 3 receptor in FRTL-5 cells. Evidence for an inhibitory effect of 1,25-dihydroxy-vitamin D 3 on thyrotropin-induced iodide uptake. Mol Cell Endocrinol 1991;81: (19) Lawson DE, Muir E. Molecular biology and vitamin D function. Proc Nutr Soc 1991;50: (20) Ozono K, Liao J, Kerner SA, Scott RA, Pike JW. The vitamin D-responsive element in the human osteocalcin gene. Association with a nuclear protooncogene enhancer. J Biol Chem 1990;265: (21) Evans RM. The steroid and thyroid hormone receptor superfamily. Science 1988;240: (22) Elstner E, Linker-Israeli M, Said J, Umiel T, de Vos S, Shintaku IP, et al. 20-epi-vitamin D 3 analogues: a novel class of potent inhibitors of proliferation and inducers of differentiation of human breast cancer cell lines. Cancer Res 1995;55: (23) Pike JW. Vitamin D 3 receptors: structure and function in transcription. Annu Rev Nutr 1991;11: (24) Wakefield L, Kim SJ, Glick A, Winokur T, Colletta A, Sporn M. Regulation of transforming growth factor-beta subtypes by members of the steroid hormone superfamily. J Cell Sci Suppl 1990;13: (25) Knabbe C, Lippman ME, Wakefield LM, Flanders KC, Kasid A, Derynck R, et al. Evidence that transforming growth factor- is a hormonally regulated negative growth factor in human breast cancer cells. Cell 1987;48: (26) Cohen PS, Letterio JJ, Gaetano C, Chan J, Matsumoto K, Sporn MB, et al. Induction of transforming growth factor beta 1 and its receptors during all-trans-retinoic acid (RA) treatment of RA-responsive human neuroblastoma cell lines. Cancer Res 1995;55: (27) Roberts AB, Kim SJ, Kondaiah P, Jakowlew SB, Denhez F, Glick AB, et al. Transcriptional control of expression of the TGF- s. Ann N Y Acad Sci 1990;593: (28) Anzano MA, Smith JM, Uskokovic MR, Peer CW, Mullen LT, Letterio JJ, et al. 1 Alpha,25-dihydroxy-16-ene-23-yne-26,27-hexafluorocholecalciferol (Ro ), a new deltanoid (vitamin D analogue) for prevention of breast cancer in the rat. Cancer Res 1994;54: (29) Silberstein GB, Daniel CW. Reversible inhibition of mammary gland growth by transforming growth factor-. Science 1987;237: (30) Paaren HE, DeLuca HF, Schnoes HK. Direct C (1) hydroxylation of vitamin D 3 and related compounds. J Org Chem 1980;45: (31) Sporn MB, Dunlop NM, Newton DL, Smith JM. Prevention of chemical carcinogenesis by vitamin A and its synthetic analogs (retinoids). Fed Proc 1976;35: (32) Oikawa T, Yoshida Y, Shimamura M, Ashino-Fuse H, Iwaguchi T, Tominaga T. Antitumor effect of 22-oxa-1,25-dihydroxyvitamin D 3, a potent angiogenesis inhibitor, on rat mammary tumors induced by 7,12- dimethylbenz[a]anthracene. Anticancer Drugs 1991;2: (33) Massague J, Cheifetz S, Ignotz RA, Boyd FT. Multiple type-beta transforming growth factors and their receptors. J Cell Physiol 1987;Suppl 5: (34) Wakefield LM, Letterio JJ, Geiser AG, Flanders KC, O Shaughnessy J, Roberts AB, et al. Transforming growth factor-betas in mammary tumorigenesis: promoters or antipromoters? Prog Clin Biol Res 1995;391: Notes Supported by an Illinois Department of Public Health research grant. Manuscript received July 30, 1996; revised October 30, 1996; accepted December 3,