Anti-cancer Effect of Broccoli:

Selective Inhibition of Transformed Mammalian Cell Growth

by Brassica oleracea Extract

Adam A. Friedman

Saint Andrew's Episcopal School, Ridgeland, MS 39157

[Adam A. Friedman received the Mississippi Junior Academy of Sciences Clyde Sheely Award for 1996 and Overall Award for 1997. Special thanks to First Chemical Corporation in Pascagoula, Mississippi, for underwriting the publication of this research paper.]

Broccoli extract has been shown to prevent neoplastic transformation by carcinogens in mammalian cells by inducing detoxification enzymes. To test whether broccoli could have other anticancer properties, the ability of a broccoli extract to inhibit the growth of three transformed mammalian cell lines was examined. The proliferation of MCF-7 human breast cancer cells, H-ras-transformed MCF-7 cells, and v-ras-transformed RAT1 fibroblast cells was measured in the presence of the broccoli extract and also compared to the proliferation of non-transformed cultured human lymphocytes. Broccoli extract was found to inhibit significantly the growth of all three transformed cell lines in a dose-dependent form, whereas non-transformed human lymphocytes were found to be initially induced, and later only moderately inhibited. Comparison of the extract's effect on the three transformed cell lines suggested that ras-transformed cells were more sensitive to inhibition by broccoli extract. Since induction of adenylate cyclase was recently reported to inhibit mammalian cells by down-regulating the ras growth pathway, RAT1 cells treated with broccoli extract were assayed for adenylate cyclase activity using a cAMP enzyme immunoassay. Broccoli was found to induce adenylate cyclase during the exponential growth of the RAT1 cells. In summary, these investigations identify broccoli as a specific inhibitor of transformed cell growth, possibly through induction of adenylate cyclase, and suggests potential therapeutic properties of the extract for inhibition of certain cancers.

An anticarcinogenic effect of a diet rich in broccoli and other cruciferous vegetables has been suggested (Zhang et al., 1992). The recent discovery that a component of a broccoli extract prevents transformation of cultured mammalian cells has generated hope for a potential cancer prophylaxis (Zhang et al., 1992, 1994). Sulforaphane [CH₃SO(CH₂)₄NCS] was reported to be found in high concentration in broccoli and to induce Phase II detoxification enzymes, which promote rapid excretion of potential carcinogenic substances (Talalay, 1989). It was postulated that broccoli could help prevent cancers by inducing the elimination of carcinogens.

It is also possible that broccoli could have direct anticancer effects on already established neoplasms through as yet unidentified mechanisms. Preliminary, unpublished research by the author in Saccharomyces cerevisiae suggested that broccoli extract modulates growth possibly through the induction of adenylate cyclase activity. The induction of adenylate cyclase, which increases intracellular cAMP concentrations, has previously been reported to inhibit growth of certain mammalian cell lines (Marx, 1993; Smets, 1972; Kurth and Bauer, 1973) possibly via inhibition of the Ras growth pathway (Cook and McCormick, 1993). Because ras-transformed cells may therefore be particularly sensitive to induced adenylate cyclase activity (Pastan et al., 1975), the ability of a broccoli extract to inhibit the growth of cancer cells was tested against ras-transformed cell lines (v-ras and H-ras transformed) and another not transformed by ras. The specificity of inhibition was tested by comparing the extract's effect on cultured normal cells. A possible mechanism of action of broccoli extract was explored by comparing the sensitivity of various cell lines and assaying treated ras-transformed cells for adenylate cyclase activity.

MATERIALS AND METHODS

Preparation of broccoli extracts--Fresh broccoli extract was prepared immediately before use, following the procedure for extraction of sulforaphane (Zhang et al., 1992). Whole fresh broccoli (Brassica oleracea) was washed, the florets homogenized with two volumes of distilled water, and the homogenate lyophilized to a fine powder. 800 g lyophilized broccoli was added to 28 ml acetonitrile, covered, and stirred overnight at 4ºC. The liquid was filtered and rotary evaporated until a solid residue remained. The residue was resuspended in 200 µl dimethyl sulfoxide and stored at -20ºC.

Transformed mammalian cell lines--MCF-7 breast cancer cell line (human adenocarcinoma, pleural effusion) was obtained from the American Type Culture Collection (Rockville, MD). H-ras-transformed MCF-7 cells were kindly supplied by Dr. Edward P. Gelmann, Georgetown University Medical Center (Washington, DC; Gelmann et al., 1992). V-ras-transformed RAT1 rat fibroblasts were kindly supplied by Dr. Nancy E. Kohl, Merck Research Laboratories (West Point, PA; Kohl et al., 1993).

Frozen cells of both MCF-7 cell types were seeded in supplemented phenol-red-free minimum essential medium eagle (MEM, 100 Units/ml penicillin, 100 µg/ml streptomycin, 2 µg/ml insulin, 29.3 ml/L 7.5% sodium bicarbonate, 0.292 g/L L-glutamine, and 10% fetal bovine serum). RAT1 cells were seeded in supplemented Dulbecco's modified eagle's medium (DMEM, 100 U/ml penicillin, 100 µg/ml streptomycin, 49.3 ml/L 7.5% sodium bicarbonate, and 10% fetal bovine serum). All cell cultures were maintained at 37C, 5% CO₂. Cell culture medium was changed three times weekly with supplemented MEM (MCF-7 cell lines) or DMEM (RAT1). After reaching confluency, the cells were washed with Dulbecco's phosphate-buffered saline, treated with a trypsin-EDTA solution, and subcultured.

For experimentation, some confluent cells were trypsinized and used as a stock cell solution. A 96-well microtitre plate was seeded with 5,000 cells/well. Rows of cells were maintained in 200 µl of the appropriate medium containing one of the following: no extract or solvent (control), 20 µl DMSO per 10 ml medium (low solvent control), 40 µl DMSO per 10 ml medium (high solvent control), 20 µl broccoli extract per 10 ml medium (low concentration), or 40 µl broccoli extract per 10 ml medium (high concentration). The cell culture medium was changed every other day for two weeks. At each feeding, the entire well volume of cells from two wells from each row were washed, trypsinized, and counted with a Coulter Counter.

One of the 96-well microtitre plates growing MCF-7 cells (non-ras-transformed) was designated for "release." After one week of experimentation, all the remaining cells were "released" from the extract, i.e., washed twice and maintained for the remainder of the experiment with control MEM. The cells were counted as previously described.

Normal mammalian cell line--Ten milliliters of blood from a healthy male volunteer was carefully layered on top of 10 ml Ficoll-Hypaque solution and centrifuged. The resulting middle layer (lymphocytes) was removed and washed twice with supplemented RPMI 1640 medium. The cells were seeded in RPMI medium at a density of 100,000 cells/well in a 96-well plate. Phytohemagglutinin (PHA) lectin was added to stimulate lymphocyte growth. Cells in rows were fed once with RPMI containing one of the following: no PHA, solvent, or broccoli (control); PHA (5 mg/ml), 20 µl broccoli extract/10 ml medium; 40 µl broccoli extract/10 ml medium; PHA+20 µl broccoli extract/10 ml medium; PHA+40 µl broccoli extract; or PHA+40 µl DMSO. Each day, cells from three wells from each row were removed and counted as described.

Adenylate cyclase activity--V-ras-RAT1 fibroblast cells were prepared in microtitre plates and fed DMEM supplemented with or without broccoli extract as previously described. Three times weekly, cells from two wells were trypsinized, centrifuged, and sonicated (Ultrasonics W-375) for 10 seconds in buffer [2 mM Tris-HCl, 1 mM DTT (Chew, 1990)]. The sonicated cells were assayed for total protein concentration using biuret reagent and assayed for adenylate cyclase activity, which was then normalized for protein concentration.

Sonicated cells were combined with creatine phosphokinase and buffer [50 mM Tris-HCl, 10 mM MgCl₂, 5 mM KCl, 1 mM DTT, 1 mM 3-isobutyl-1-methylxanthine, 1 mM ATP, 10 mM creatine phosphate, 0.1 mM GTP, 0.1% BSA] for a total reaction volume of 500 µl. After 60 min. at 37C, the reaction was halted by acidification to pH 4.5 with 5 µl 1 N HCl and heating to 100C for one minute. After centrifugation, the supernatant was harvested and cAMP concentration was assayed by enzyme immunoassay (EIA: Amersham Biotrak cAMP EIA system, RPN 225). Activity was expressed as femtomolar (fM) cAMP/min/mg protein.

RESULTS

Cells treated with DMSO at any concentration showed no statistically significant difference from control cells in growth or adenylate cyclase activity.

Mammalian cell growth analyses

MCF-7 human breast cancer cells-- Control MCF-7 cells grown without broccoli extract followed a standard growth curve reaching a maximal density of 1.8 x 10⁶ cells/ml. In the presence of the low concentration of broccoli extract (20 µl/10 ml MEM), a marked and statistically significant (p0.04) reduced growth was observed after nine days (48% of control growth). The higher concentration (40 µl/10 ml MEM) had an even greater inhibitory effect on the MCF-7 cells (32% of control, p0.01) (Fig. 1). After release on day seven, cells washed free of low concentration broccoli extract resumed normal growth, while unreleased cells did not; unreleased cells grew to 54% of the control value, while released cells grew to 81% of control growth (data not shown).

H-ras-transformed MCF-7 cells-- Control H-ras-MCF-7 human breast cancer cells reached a maximal cell density of 1.75 x 10⁶ cells/ml. After four days, a decrease in growth was seen following exposure to 20 µl broccoli extract/10 ml MEM (59% of control), and an even greater decrease was seen with 40 µl extract/10 ml MEM (34% of control) (Fig. 2).

V-ras-transformed RAT1 fibroblasts--Control RAT1 cells reached a maximal cell density of 3.95 x 10⁶ cells/ml. As with the other two cell types, 20 µl/10 ml DMEM broccoli extract markedly decreased growth beginning on day two (60% of control) (p0.009). With the higher concentration of extract, cell growth was stagnant from day two, falling from 27% of control growth at day four (p0.0009) to 5% of control at day nine (p0.00006) (Fig. 3).

The effect of broccoli extract on growth varied among the transformed cell lines only at the higher concentration. By day nine, H-ras-transformed MCF-7 cells were inhibited by 40 µl extract/10 ml medium significantly more than the MCF-7 cells (12% vs. 32%); the v-ras-transformed cells were, in turn, inhibited by the high concentration of extract more than the ras-MCF-7 cells (5% vs. 12%) (Fig. 4).

Normal human lymphocytes--Both concentrations of broccoli extract produced a biphasic effect on normal lymphocyte growth. Broccoli extract initially induced growth, the higher concentration (p0.002) more so than the lower concentration (143% and 120% of control, respectively). After longer exposure, however, the broccoli-exposed cells grew more slowly than the control cells in a concentration-dependent manner [47% (p0.0001) and 75% (p0.007) of control] (Fig. 5).

Adenylate cyclase activity analysis Adenylate cyclase activity in v-ras-transformed RAT1 cells--In control cells, adenylate cyclase activity was lowest on days four and six and highest on days two and nine. In contrast, at day nine, the broccoli extract inhibited adenylate cyclase activity, the lower more than the higher concentration [18% (p0.008) and 35% (p0.02) of control, respectively]. At day four, this effect was reversed; 40 µl extract induced the activity of adenylate cyclase, as did 20 µl at a lower level [605% (p0.02) and 155% of control, respectively] (Fig. 6).

DISCUSSION

For the three transformed cell lines examined, treatment with broccoli extract resulted in significant growth inhibition. Broccoli extract inhibited the growth of MCF-7 human breast cancer cells, H-ras-transformed MCF-7 cells, and v-ras-transformed RAT1 fibroblasts (Figs. 1, 2, 3). This inhibitory effect was not permanent, as growth inhibition could be partially reversed by removing the broccoli extract from the growth medium. Moreover, this inhibitory effect was more pronounced for transformed cells than for the normal human lymphocytes. While the broccoli extract inhibited the transformed cells throughout the growth period (60% to 5% of control values), the normal lymphocytes were initially induced and only partially inhibited after a number of days (75% to 47% of control values). Thus, although there was later inhibition of the lymphocytes, the inhibition was less marked and discontinuous, suggesting a greater specificity for transformed cells.

A proposed mechanism for this specific inhibition is induction of adenylate cyclase. When a comparison was made between the various transformed cells, at the higher concentration, broccoli extract inhibited RAT1 cells > ras-MCF-7 cells > MCF-7 cells. It has been suggested that Ras interaction with Raf, necessary for stimulation of growth via the Ras-growth pathway, is inhibited by induction of adenylate cyclase, which causes increases in intracellular cAMP concentration and subsequent induction of protein kinase A (Cook and McCormick, 1993). The greater sensitivity of ras-transformed cell lines to broccoli extract suggests that adenylate cyclase may be involved, in part, in the inhibition of cell growth. This is further evidenced by the induction of adenylate cyclase activity in v-ras-transformed RAT1 cells (Fig. 6).

Caution must be exercised in the interpretation of these results because multiple repetitions of the experiments with these and other cells lines must be completed in order to confirm that ras-transformed cells are more sensitive to the broccoli extract. It is also possible that the broccoli extract may demonstrate some trypsin/EDTA-like effects, thus detaching the cells in culture, or that another mechanism caused inhibition of growth and induction of adenylate cyclase independently. Since the extract contains numerous compounds, including sulforaphane, this study cannot differentiate among many possible mechanisms of inhibition of growth in these transformed cells.

In summary, analysis of the growth inhibition of various transformed cell lines and of induction of adenylate cyclase activity in the v-ras-transformed RAT1 line demonstrated that the inhibition of growth of transformed cells by a broccoli extract can be correlated, in part, to induction of adenylate cyclase. This observation is consistent with the suggestion that induction of adenylate cyclase is a mechanism through which broccoli extract exerts its effects. Future investigations should include purification of the extract and identification of the active compounds and their biological role. If these active compounds can be shown to inhibit the growth of transformed mammalian cells with minimal inhibition of non-transformed cells, as demonstrated here, broccoli may indeed have applications to human cancer therapy. Additional knowledge of the mechanism of action of broccoli would also guide future development of the broccoli extract as a pharmaceutical agent against cancer.

ACKNOWLEDGEMENTS

The author wishes to acknowledge the following for their assistance: Demethous Chambliss, Science Department, St. Andrew's Episcopal School (Ridgeland, MS); Dan Rose (Pearl, MS); Dr. Terrence J. Hall, Trinity Medical Center (Moline, IL); Dr. Pratibha Joshi, Dr. James Grogan, and other members of the Department of Surgery, University of Mississippi Medical Center (Jackson, MS); Dr. Steve Case, Dr. Gregory Leno, and other members of the Department of Biochemistry, UMMC; and the UMMC chapter of Sigma Xi for a grant in support of this research.

LITERATURE CITED

Chew, Catherine S. 1990. cAMP Technologies, Functional Correlates in Gastric Parietal Cells. Page 650 in S. Fleischer and B. Fleischer, eds. Vol 191, Methods in Enzymology: Biomembranes, Part V: Cellular and Subcellular Transport: Epithelial Cells. Harcourt Brace Jovanovich, Publishers, New York.

Cook, S.J., and F. McCormick. 1993. Inhibition by cAMP of Ras-Dependent Activation of Raf. Science. 262:1069­1072.

Gelmann, E.P., E.W. Thompson, and C.L. Sommers. 1992. Invasive and Metastatic Properties of MCF-7 Cells and ras^H-Transfected MCF-7 Cell Lines. International Journal of Cancer. 50:665­669.

Kohl, N.E., S.D. Mosser, J. deSolms, E.A. Giuliani, D.L. Pompliano, S.L. Graham, R.L. Smith, E.M. Scolnick, A. Oliff, and J.B. Gibbs. 1993. Selective Inhibition of ras-Dependent Transformation by Farneyltransferase Inhibitor. Science. 260:1934­1937.

Kurth, R., and H. Bauer. 1973. Influence of Dibutyryl Cyclic AMP and Theophylline on Cell Surface Antigens on Oncovirus Transformed Cells. Nature New Biology. 243:243­245.

Marx, J. 1993. Two Major Signal Pathways Linked. Science. 262:988­990.

Pastan, I.H., G.S. Johnson, and W.B. Anderson. 1975. Role of Cyclic Nucleotides in Growth Control. Annual Review of Biochemistry. 44:491­522.

Smets, L.A. 1972. Contact Inhibition of Transformed Cells Incompletely Restored by Dibutyryl Cyclic AMP. Nature New Biology. 239:123­124.

Talalay, P. 1989. Mechanisms of Induction of Enzymes that Protect Against Chemical Carcinogenesis. Advanced Enzyme Regulation. 28:237­250.

Zhang, Y., T.W. Kensler, C.-G. Cho, G.H. Posner, and P. Talalay. 1994. Anticarcinogenic activities of sulforaphane and structurally related synthetic norbornyl isothiocyanates. The Proceedings of the National Academy of Sciences. 91:3147­3150.

Zhang, Y., P. Talalay, C.-G. Cho, and G.H. Posner. 1992. A major inducer of anticarcinogenic protective enzymes from broccoli: Isolation and elucidation of structure. Proceedings of the National Academy of Sciences, USA. 89:2399­2403.