Rapamycin was purchased from Sigma Chemical (St. Louis, MO, USA). CN-A was purified from a stock ethyl acetate extract obtained from the culture filtrate of Cladosporium fungus sp. 501-7W by flash chromatography on silica gel with more than 99% purity 1011. A stock solution of CN-A was prepared in absolute ethanol at 20 mg/ml.

Human breast carcinoma cell lines (MCF-7, T-47D and MDA-MB-231) and human promyeloblastic leukaemia cell line NB-4 were cultured in RPMI 1640 supplemented with 10% foetal bovine serum at 37°C in a humidified atmosphere of 5% carbon dioxide in air.

The cells were seeded at 1–3 × 10⁴/ml in a 24-well multidish. After culture with or without the test compounds for the indicated times, viable cells were examined using either the trypan blue dye exclusion test or a modified MTT (3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide) assay. In the MTT assay, 100 μl MTT solution (1 mg/ml in PBS) was added to each well and cells were incubated for 4 hours. The cells were then centrifuged at 1000 g for 10 min and the precipitates were dissolved in 1 ml dimethyl sulphoxide; their absorption at 560 nm was determined. Assay of the cumulative cell number was determined as described elsewhere 17.

Isobologram analysis was used to determine the effects of combinations of drugs on MCF-7 cells. Dose-dependent effects were determined for each compound and for one compound with fixed concentrations of another. The interaction of two compounds was quantified by determining the combination index (CI), in accordance with the following classic isobologram 18:

CI = (D)₁/(Dx)₁ + (D)₂/(Dx)₂

Where Dx is the dose of one drug alone required to produce an effect, and (D)₁ and (D)₂ are the doses of compounds 1 and 2, respectively, in combination that produce the same effect. From this analysis, the combined effects of the two drugs can be summarized as follows: CI = 1 indicates summation (additive and zero interaction); CI < 1 indicates synergism; and CI > 1 indicates antagonism.

The cell cycle was analyzed using propidium iodide-stained nuclei. Samples of 2 × 10⁶ cells were harvested at the time points indicated, washed in ice-cold PBS, fixed by the addition of 100% ethanol and left for 30 min on ice. The cell pellet was washed and suspended in 200 μl 1.12% sodium citrate containing RNase A (250 μg/ml) for 30 min at room temperature. Thereafter, the cells were stained with 50 μg/ml propidium iodide in the presence of 1.12% sodium citrate and analyzed in a fluorescence-activated cell sorter.

The expression of E-cadherin was detected by immunocytochemistry. Cells were fixed in 4% paraformaldehyde in PBS and permeabilized in acetone at 4°C. Slides were pretreated in PBS containing 0.2% Tween-20, blocked with 5% normal goat serum and 0.2% Tween-20 in PBS, and incubated with rabbit polyclonal antibody to E-cadherin (Santa Cruz, CA, USA). Antibody–antigen complexes were visualized using DAKO ENVISION/AP kit (DakoCytomation Japan, Kyoto, Japan). Senescence-associated β-galactosidase (SA-βGal) activity was determined as described by Dimri and coworkers 19. SA-βGal activity was monitored visually by scoring blue precipitate in the cytoplasm.

Total RNA was isolated from MCF-7 cells treated with or without compound for 12 hours using Isogen (Nippon Gene, Toyama, Japan) 20. Poly(A)⁺RNA was reverse transcribed with the concomitant incorporation of Cy3- and Cy5-labelled nucleotides. The labelled probes were hybridized with a cDNA microarray, representing about 16,000 different human genes (TaKaRa Bio Inc., Tokyo, Japan), and their fluorescent intensities were scanned according to the protocol standardized by TaKaRa Bio Inc. The genes were screened by analyzing the difference in expression profiles between two genes.

Total RNA was extracted using Isogen (Nippon Gene), in accordance with the manufacturer's instructions. Total RNA (1 μg) from tumour cells was converted to first-strand cDNA primed with random nonamer in a reaction volume of 20 μl using an RNA PCR kit (TaKaRa Bio), and 4 μl of this reaction was used as a template in the PCR. The oligonucleotides used in PCR amplification are summarized in Table 1. PCR consisted of 25 cycles for transforming growth factor-β-induced 68 kDa protein (TGFBI), BCL2-interacting killer (BIK) and early growth response (EGR)3; 22 cycles for cyclin G₂; 27 cycles for growth factor receptor-bound (GRB)7; and 17 cycles for glyceraldehyde-3-phosphate dehydrogenase (GAPDH); with denaturing at 94°C for 30 s, annealing at 60°C for 30 s and extension at 72°C for 30 s. The linearity of the quantitation of RT-PCR products was determined using [α-³²P]dCTP and various amounts of total RNA, as described in the literature 20. Under these conditions, the amounts of PCR products increased linearly up to 0.4 μg total RNA.

Female athymic nude mice with a BALB/c genetic background were supplied by CLEA Japan (Tokyo, Japan). They were housed under specific pathogen-free conditions. The in vivo experiments were performed in accordance with the guidelines of our institute (Guide for Animal Experimentation, Saitama Cancer Center, Saitama, Japan). One week before MCF-7 cell inoculation, mice each received 2 μg of 17β-oestradiol valerate (Sigma), dissolved in 0.2 ml of sesame oil, by subcutaneous injection. Oestrogen injections were repeated every week to sustain tumour growth. Mice were inoculated subcutaneously with 2 × 10⁷ MCF-7 cells. By day 21 after inoculation, all of the tumours were about the same size. The animals were randomly distributed into four groups of 20 mice each. A stock solution of CN-A for administration was prepared in dimethyl sulphoxide at 100 mg/ml, and rapamycin was dissolved in ethanol at 1 mg/ml. Mice were given a daily intraperitoneal injection of 0.1 ml PBS, including 100 ng rapamycin, and/or subcutaneous injections every other day of 0.2 ml of PBS, including 100 μg CN-A, at a site distant from the tumours. Tumour size was measured with vernier calipers every other day. Statistical analysis was performed using Student's t-test.

We examined the synergistic effects of various agents, including differentiation inducers and rapamycin, on growth inhibition of cells of breast cancer cell line MCF-7. Sensitivity to anticancer agents such as 5-fluorouracil, cisplatin and doxorubicin was barely affected by rapamycin. All-trans retinoic acid (a typical differentiation inducer), but not vitamin D₃, affected the growth of MCF-7 cells, and the growth-inhibitory effect of rapamycin was significantly enhanced by all-trans retinoic acid, although a high concentration of retinoic acid was required (data not shown).

Among the differentiation-inducing agents tested, the sensitivity of MCF-7 cells to rapamycin was most affected by CN-A, a novel inducer of the differentiation of myeloid leukaemia 1213141516. Figure 1 shows the synergistic effects of rapamycin and CN-A on the growth of MCF-7 cells. Figure 1a shows the May–Gruenwald–Giemsa staining pattern of MCF-7 cells after 5-days of treatment with various concentrations of both rapamycin and CN-A. Figure 1b shows the time course of the combined effects of rapamycin and CN-A on the growth of MCF-7 cells. The growth of MCF-7 cells was inhibited moderately by rapamycin (0.5 ng/ml) or CN-A (10 μg/ml) alone, but growth was still seen at least until 5 days, whereas the cell number did not change after 1 day of treatment with the combination of rapamycin and CN-A. Figure 2 shows isoboles for the combination of rapamycin with CN-A that were isoeffective (IC₅₀ : the concentration of the drug required for 50 % inhibition of cell growth) for inhibition of proliferation of MCF-7 cells. These isoboles indicate that the combination of these drugs had synergistic effects.

We also examined the long-term effects of combined treatment with rapamycin and CN-A on the proliferation of MCF-7 cells (Fig. 1c). The growth rate of rapamycin-treated cells was almost the same as that of control cells after 5 days. Cell growth was significantly inhibited by CN-A alone after 5 days, but the cell number was still increased even at day 23. On the other hand, cell growth was greatly inhibited by combined treatment with rapamycin and CN-A, and the cell number at day 23 was almost the same as that at day 5 (Fig. 1c).

Similar growth-inhibitory effects of rapamycin and CN-A were observed in two other human breast cancer cell lines (oestrogen receptor-positive T-47D cells and oestrogen receptor-negative MDA-MB-231 cells; Fig. 3a,b). Furthermore, these combined effects of rapamycin and CN-A were also observed in several different tumour cell lines such as human promyelocytic leukaemia NB-4 cells (Fig. 3c) and human non-small-cell lung carcinoma A549 and Lu99 cells (data not shown). These findings indicate that the growth-inhibitory effects of rapamycin plus CN-A were not restricted to MCF-7 cells.

To establish the most effective treatment procedure using rapamycin and CN-A on MCF-7 cells, we examined the effects of various different treatments on MCF-7 cells. MCF-7 cells were seeded at 1 × 10⁴ cells/ml and the number of viable cells was determined by MTT assay after 7 days. Simultaneous treatment with rapamycin (0.5 ng/ml) and CN-A (2.5 μg/ml) for 7 days was most effective at inhibiting cell proliferation (96% inhibition). Furthermore, we found that the timing of the addition of CN-A was critical to the inhibitory affect on the cell number. When CN-A was added at day 2, the combined growth-inhibitory effect was markedly decreased at 7 days (29% inhibition). On the other hand, when rapamycin was added at day 1 or day 2, the combined growth-inhibitory effects were almost completely conserved at 7 days (95% or 90% inhibition, respectively). We also found that an initial short exposure to CN-A was enough to exert the combined effect. When cells were treated with CN-A for only 1 day from day 0 to day 1 in the presence of rapamycin, and then cultured only in the presence of rapamycin from day 2 to day 7, the combined growth-inhibitory effects were still almost completely preserved at 7 days (90% inhibition). These findings indicate that simultaneous treatment with rapamycin and CN-A is the most effective procedure for inhibiting the cell growth of MCF-7 cells and that initial 1-day treatment with CN-A and continuous treatment with rapamycin has almost the same growth-inhibitory effects as simultaneous treatment with rapamycin and CN-A.

To better understand the combined effects of rapamycin and CN-A on cell growth, we exposed MCF-7 cells to 0.5 ng/ml rapamycin and 10 μg/ml CN-A, and then measured the changes in the cell cycle distribution after 6 days. This concentration of rapamycin or CN-A alone did not affect the cell cycle in our study, whereas rapamycin with CN-A induced growth arrest of cells at the G₁ phase (Fig. 4). The induction of apoptosis (cells in sub-G₁ phase) was not observed in cells with the combined treatment.

Because treatment with rapamycin and CN-A induced G₁ arrest but not apoptosis, these treatments may induce phenotypic changes in MCF-7 cells. Changes in several differentiation-associated phenotypes were examined, but the expression of casein or mucin was not significantly induced by the combined treatment. Intracellular accumulation of lipid droplets was also not observed (data not shown). However, when we cultured MCF-7 cells for 7 days, CN-A greatly enhanced the expression of E-cadherin, and the expression was significantly enhanced by rapamycin (Fig. 5). We also found that CN-A alone or CN-A plus rapamycin could induce activity of SA-βGal (Fig. 5), which is one of the markers of senescence of epithelial cells 21. These findings suggest that this combined treatment with rapamycin and CN-A induced growth arrest of cells at the G₁ phase and may induce phenotypic changes toward cell senescence but not apoptosis.

To elucidate the mechanism underlying the G₁ arrest mediated by rapamycin plus CN-A, we then examined the expression of several genes, using RT-PCR, that are important for cell cycle regulation. We first examined the expression of p53 in MCF-7 cells (Fig. 6a). Unexpectedly, the expression of p53 was not affected when MCF-7 cells were treated with 1 ng/ml rapamycin alone, 2.5–10 μg/ml CN-A alone, or 1 ng/ml rapamycin plus 2.5 μg/ml CN-A for 24 hours. HL60 cells were used as a negative control for p53 expression.

Because the p21^Cip1 and p27^Kip1 Cdk inhibitors play important roles in cell cycle regulation, especially in response to external agents, we then examined the levels of expression of these two genes (Fig. 6a). The expression levels of p21^Cip1, which is a target gene of p53, and p27^Kip1 were also not induced by these treatments. Cyclin D₁ plays an important role in G₁ phase cell cycle progression, and previous work demonstrated that rapamycin decreased expression of cyclin D₁ and c-myc 22. When MCF-7 cells were treated with 0.5 ng/ml rapamycin and 10 μg/ml CN-A for 12 hours, expression of cyclin D₁ or c-myc was not inhibited by the combined treatments (Fig. 6b). On the other hand, expression of cyclin E₁ was slightly inhibited by the combined treatments (Fig. 6b).

Because we could not detect major changes in the expression of cell cycle regulating genes using RT-PCR (Fig. 6), we conducted a cDNA microarray analysis of MCF-7 cells and screened genes that were upregulated or downregulated by rapamycin plus CN-A. MCF-7 cells were cultured with 0.5 ng/ml rapamycin or 10 μg/ml CN-A, or both combined for 12 hours and total RNA was used in the microarray analysis. When MCF-7 cells were cultured with the same treatments for 5-day, the levels of growth inhibition induced by rapamycin, CN-A, and rapamycin plus CN-A were 31%, 40% and 90%, respectively.

Genes with a treatment:control ratio greater than 2 were considered to be significantly upregulated, and those with a ratio below 0.5 were considered to be significantly downregulated. Many more genes were upregulated and downregulated upon exposure to rapamycin plus CN-A than with rapamycin or CN-A alone. Whereas individually rapamycin and CN-A upregulated 61 and 46 genes and downregulated 33 and 26 genes, respectively, combined treatment with rapamycin and CN-A upregulated 225 genes and downregulated 186 genes. Furthermore, 41 genes were upregulated more than threefold in MCF-7 cells treated with rapamycin plus CN-A, although only one gene and eight genes were upregulated more than threefold in rapamycin-treated and CN-A-treated MCF-7 cells, respectively. These findings suggest that rapamycin and CN-A can also induce synergistic effects on the induction of gene expression. Table 2 shows 25 genes that were highly upregulated (>3.4-fold) by rapamycin plus CN-A. Among these 25 upregulated genes, nine and 17 were upregulated (>2-fold) in rapamycin-treated and CN-A-treated cells, respectively. TGFBI 23, BIK 24, cyclin G₂ 2526 and GRB7 27 were upregulated by more than 5.8-fold in MCF-7 cells treated with rapamycin plus CN-A.

Table 3 shows 25 genes that were highly downregulated (gene expression ratio <0.38) by both rapamycin and CN-A. Thirteen genes were downregulated by less than one-third in cells treated with the combined treatment. In cells treated with CN-A alone or the combined treatment, the expression of EGR3 28 was decreased to about one-fifth of that in control MCF-7 cells. Table 4 shows the effects of rapamycin plus CN-A on the expressions of several cell cycle regulatory genes. As mentioned above, the expression of cyclin G₂ (gene expression ratio 6.08) was markedly induced, whereas the expressions of cyclin E₁ (gene expression ratio 0.44) and E2F transcription factor 3 (gene expression ratio 0.42) were moderately downregulated in the rapamycin plus CN-A treated cells.

To confirm the results of the cDNA microarrray analysis, the expressions of highly upregulated genes were examined by RT-PCR. Figure 7a shows the time course patterns of gene expression of TGFBI, BIK, cyclin G₂ and GRB7. Figure 7b shows graphs of the relative levels of mRNA after normalization to GAPDH mRNA levels. These findings confirm the data from the cDNA microarray analysis. In these genes, the expression of cyclin G₂ was increased fourfold at 3 hours after combined treatment and further increased with time (>8 fold at 24 hours). The expression of GRB7 was also induced by more than threefold at 3 hours after combined treatment, and the expression levels remained almost constant, at least until 24 hours. Expression of the TGFBI and BIK genes were markedly induced by the combined treatment after 12 hours. These genes were quickly and highly upregulated after the addition of rapamycin plus CN-A, and so they may play a crucial role in growth inhibition in human breast carcinoma cells treated with rapamycin plus CN-A.

The in vitro studies described above suggested that combined treatment with rapamycin and CN-A should be more therapeutically effective than treatment with rapamycin or CN-A alone. At day 20 after inoculation of human breast carcinoma MCF-7 cells the tumour volume (mean ± standard deviation) was 119 ± 59 mm³, and treatments were then started. We administered 0.1 μg rapamycin/mouse every day and 100 μg CN-A/mouse every other day. These treatments had no appreciable adverse effects on the mice. The combined treatment significantly inhibited the growth of MCF-7 cells as xenografts (Fig. 8). Injections of solvent alone did not inhibit the growth of MCF-7 cells in vivo (Fig. 8). At day 18 after treatment, the mean tumour volumes in untreated, rapamycin treated, CN-A treated, and rapamycin + CN-A treated nude mice were 1336 ± 332 mm³, 659 ± 181 mm³, 848 ± 237 mm³ and 245 ± 62 mm³, respectively. Although rapamycin and CN-A each significantly retarded tumour growth (P < 0.05), combined treatment induced the arrest of tumour growth. The treatment was continued for 18 days and then stopped, with follow up on day 48. All of the untreated, rapamycin treated and CN-A treated mice had a greater tumour burden at day 48. On the other hand, all mice treated with rapamycin plus CN-A still had a much smaller tumour burden (data not shown), suggesting that the therapeutic effects were still maintained after treatment was terminated. These findings indicate that the combination of rapamycin and CN-A is more therapeutically effective than treatment with rapamycin or CN-A alone, and the combined treatment had a significant antitumour effect (P < 0.001).