The antitumor effect of LJ-529, a novel agonist to A3 adenosine receptor, in both estrogen receptor–positive and estrogen receptor–negative human breast cancers Free

Synthetic agonists to the Gi-protein-coupled A3 adenosine receptor (A3AR) have recently been suggested as a possible therapeutic agent to various cancers (1, 2). The A3AR agonist N⁶-(3-iodobenzyl)-5′-N-methylcarbamoyladenosine (IB-MECA) and its 2-chloro derivative, 2-chloro-N⁶-(3-iodobenzyl)-5′-N-methylcarbamoyladenosine (Cl-IB-MECA), inhibit the in vitro and in vivo growth of B16-F10 melanoma cells (3, 4). IB-MECA has also been reported to inhibit prostate carcinoma cell growth, breast cancer cell proliferation, and primary colon carcinoma growth and liver metastasis (5–8). Cl-IB-MECA is known to induce apoptosis of HL-60 and MOLT-4 leukemic and Li-7A hepatoma cells (9, 10). Although the action is not directly on cancer cells, Cl-IB-MECA is also known to potentiate natural killer cells, which is strongly implicated to have tumor surveillance activity (11).

Despite many reports suggesting synthetic A3AR agonists as a promising therapeutic agent against various cancers, the underlying molecular mechanisms seem to vary considerably depending on the tumor system. In melanoma cells, the A3AR internalization/externalization and the deregulation of the Wnt signaling pathway by IB-MECA has been reported. IB-MECA in melanoma cells rapidly internalizes, sorts, resynthesizes, and externalizes the receptor to the cell surface (12). In addition, it decreases cyclic AMP, prevents the activation of both protein kinase A and Akt, diminishes the phosphorylation of glycogen synthase kinase 3β (GSK-3β), and by maintaining GSK-3β in its active form, inhibits cell proliferation via deregulation of the Wnt pathway (13). Furthermore, Cl-IB-MECA has been reported to inhibit human melanoma cell proliferation via phosphatidylinositol 3-kinase/Akt–dependent inhibition of the extracellular signal-regulated kinase (ERK) 1/2 phosphorylation (14). The suppression of the Wnt signaling pathway by IB-MECA has also been reported in prostate carcinoma cells (5). Cl-IB-MECA induces Fas and apoptosis in a p53-independent manner in human leukemia cell lines (9), and apoptosis most likely by activating of caspase-3 in human hepatoma cells (10). In breast cancer cells, IB-MECA suppresses cell proliferation by down-regulating estrogen receptor (ER) α (6) and inhibits anchorage-dependent cell growth in both ER-positive and ER-negative cell lines (7).

We have recently reported that the 4′-thio analogues of the Cl-IB-MECA are associated with higher potency and selectivity than IB-MECA and Cl-IB-MECA (15). In this study, we tested the effect of one of the 4′-thio analogues of the Cl-IB-MECA generically known as LJ-529 [2-chloro-N⁶-(3-iodobenzyl)-5′-N-methylcarbamoyl-4′-thioadenosine; previously reported as thio-Cl-IB-MECA in ref. 15] in breast cancer cells. LJ-529 is reported to have a higher binding affinity to the human A3AR than Cl-IB-MECA (K_i = 0.38 versus 1.00 nmol/L; ref. 15). We report that LJ-529 was effective on both ER-positive and ER-negative breast cancer cell lines, both in vitro and in vivo, by affecting various signaling pathways. It is of significance that LJ-529 was effective in ER-negative, c-ErbB2-overexpressing breast cancer cells because this type of breast cancer is resistant to hormonal therapies and tends to have a more aggressive phenotype (16, 17). However, such effect was independent of the A3AR because none of the cells tested expressed endogenous A3AR.

MCF7, T47D, SK-BR-3, and MDA-MB-231 were from American Type Culture Collection (Manassas, VA). Cells were maintained in DMEM (Welgene, Daegu, Republic of Korea) supplemented with 10% fetal bovine serum (Welgene) in a humidified atmosphere with 5% CO₂ at 37°C. For drug treatment on cells, a stock of 100 mmol/L was prepared in DMSO. When cells reached 60% to 70% confluency, cells were treated with LJ-529 diluted in DMEM supplemented with 5% fetal bovine serum at indicated concentrations for indicated times.

3-(4,5-Dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide assays were done with CellTiter 96 Assay (Promega, Madison, WI) according to the instructions of the manufacturer. The absorbance at 570 nm was recorded using a 96-well plate reader (Bio-Rad, Hercules, CA).

After drug treatment, cells were detached by trypsinization, collected by centrifugation, and washed once with PBS. Staining of cells with propidium iodide and analyzing them by flow cytometry (FACSCalibur and CellQuest software, BD Biosciences, San Jose, CA) were done as described (18).

After drug treatment, cells were lifted by trypsinization and subjected to nuclear staining with 4′,6-diamidino-2-phenylindole as previously described (19), except that the cells were mounted with 50 μL VectaShield (Vector Laboratories, Burlingame, CA).

Cells were lysed in lysis buffer [50 mmol/L Tris (pH 7.4), 150 mmol/L NaCl, 1 mmol/L EDTA, 1% NP40, 0.25% sodium deoxycholate] supplemented with phosphatase inhibitors (Phosphatase Inhibitor Cocktail I and II, Sigma, St. Louis, MO) and protease inhibitors (Complete, Mini, EDTA-free, Roche Applied Science; Indianapolis, IN) and the cleared lysates by centrifugation were used for immunoblotting. Fifty to 100 μg total protein were loaded per lane, separated by SDS-PAGE, transferred to nitrocellulose filter, and subjected to immunoblotting with appropriate antibodies. Antibodies against caspase-3, A3AR, GSK-3β, cyclin D1, p27^kip, and ERα were from Santa Cruz Biotechnologies (Santa Cruz, CA); Akt, phospho-Akt, and phospho-GSK-3β were from Cell Signaling Technology (Beverly, MA); c–poly(ADP)ribose polymerase was from BD Biosciences; and β-actin was from Sigma. Bands were detected with ECL Plus Western Blotting Detection Reagents from GE Healthcare (Chalfont St. Giles, United Kingdom).

Total RNA was extracted with TRIzol (Invitrogen, Carlsbad, CA) and reverse transcription-PCR (RT-PCR) was done with Access RT-PCR System (Promega) according to the instructions of the manufacturer. PCR primers and thermal profiles for A3AR, glyceraldehyde-3-phosphate dehydrogenase, and β-actin were as previously described (6, 20). Specific primers used for PCR were as follows: A1AR (forward, 5′-GATGCCACCTTCTGCTTCAT-3′; reverse, 5′-AGACCATGTACTCCATGCTG-3′), A2aAR (forward, 5′-CTCCGGTACAATGGCTTGGT-3′; reverse, 5′-CTCCATCTGCTTCAGCTGTC-3′), and A2bAR (forward, 5′-TCCTCTGGGTCCTTGCCTTT-3′; reverse, 5′-GGGATTGACAACTGAATTGGC-3′).

Experiments were done in accordance with the guidelines established by the USPHS. Female nude BALB/cAnNCrjBgi-nu mice (Orient, Seoul, Republic of Korea) were used at 7 to 8 weeks of age and maintained in the specific pathogen-free laboratory animal facility at Hanyang University (Seoul, Republic of Korea). A xenograft was established by s.c. injection of in vitro cultured T47D (3 × 10⁶ cells/200 μL) or SK-BR-3 (4 × 10⁶ cells/150 μL) into the flank of mice. For xenograft with T47D cells, mice were supplemented with estradiol pellets (0.72 mg, released over 60 days; Innovative Research of America, Sarasota, FL). Tumors were measured in two diameters with calibers to permit calculation of tumor volume, V = {(D + d) / 2}³, where D and d were the larger and smaller diameters, respectively. For drug treatment in mice, a stock of 0.5 mg/mL was prepared by sonication in water. Five to seven days after injection, when the tumor width reached 1 to 3 mm, 20 animals were randomly selected, grouped, and treated with the indicated doses of LJ-529 daily and p.o. for 28 days. Each group contained five mice. The first day of drug treatment was set as day 0 and the tumor size and body weight were measured twice a week up to day 28.

Because several A3AR agonists are known to attenuate cancer cell growth, we examined the effect of a novel A3AR agonist, LJ-529, in breast cancer cell lines. MCF7 and T47D cells were tested as ER-positive cells, whereas MDA-MB-231 and SK-BR-3 were examined as ER-negative breast cancer cell lines. As shown in Fig. 1, all of the cells tested yielded attenuated anchorage-dependent cell growth in a dose- and time-dependent manner regardless of its ER status.

Next, we examined whether the attenuated growth upon LJ-529 treatment was due to accumulation of cells at a certain stage during the cell cycle. As seen in Fig. 2, a significant population of cells accumulated at the sub-G₁ phase, suggesting enhanced apoptosis by LJ-529. To confirm that LJ-529 indeed induced apoptosis, nuclear fragmentation was quantified in cells after 4′,6-diamidino-2-phenylindole staining (Fig. 3A). As shown in Fig. 3B, both ER-positive and ER-negative cells showed enhanced apoptosis by LJ-529 in a dose-dependent manner. Biochemical evidence also support that LJ-529 activated the apoptotic pathway as seen by cleavages of caspase-3 and c–poly(ADP)ribose polymerase (Fig. 3C).

Because there are reports that A3AR agonists at high doses (micromolar concentrations) act independently of its receptor, A3AR (6, 9), we examined the expression of A3AR in breast cancer cells by immunoblotting and RT-PCR. As seen in Fig. 4A and B, A3AR expression was not detected in any of the breast cancer cells tested. Similar results were obtained regardless of the growth conditions, such as in culture medium supplemented with 10% or 5% fetal bovine serum with LJ-529 (data not shown). Therefore, we conclude that the effect of LJ-529 occurred independently of the A3AR in breast cancer cells. Intriguingly, when the expression of other adenosine receptors was examined, only the A2 isoforms (A2a and A2b) were detected by RT-PCR (Fig. 4C).

As certain A3AR agonists are reported to deregulate the Wnt signaling pathway in cancer cells (5, 13), we tested the effect of LJ-529 treatment on the Wnt signaling pathway in T47D (ER-positive) and SK-BR-3 (ER-negative) cells. The total amount of Akt was reduced in both T47D and SK-BR-3 cells (Fig. 5A). Furthermore, the phosphorylated form of Akt at serine-473 was reduced in both cells in a dose-dependent manner (Fig. 5A). Whereas the total GSK-3β protein level remained unchanged, the phosphorylated form at serine-9 was reduced in a dose-dependent manner (Fig. 5A). The expression of cyclin D1 was examined because its transcription is regulated as a result of downstream β-catenin and T-cell factor/lymphoid enhancer–binding factor transcription factors (21). The protein level of cyclin D1 was also significantly reduced in a dose-dependent manner (Fig. 5A). Therefore, as evidenced from the changes of Akt, GSK-3β, to cyclin D1, the canonical Wnt signaling pathway was down-regulated by LJ-529 in breast cancer cells regardless of their ER status.

Because Akt activity was reduced upon LJ-529 treatment, and Akt is reported to regulate p27^kip (22, 23), we tested the expression of p27^kip by immunoblotting. As seen in Fig. 5B, the expression level of p27^kip was induced by LJ-529 in a dose-dependent manner. This is intriguing because activated Akt upon phosphorylation at serine-473 is known to phosphorylate p27^kip at threonine-157 (22, 23). As a result, the phosphorylated p27^kip is excluded from the nucleus and remains in the cytoplasm, where protein degradation occurs ultimately. Therefore, it is plausible to conclude that total p27^kip protein accumulated as a consequence of reduced phospho-Akt (at serine-473) by LJ-529 treatment in breast cancer cell lines.

Because an adenosine analogue, IB-MECA, is reported to down-regulate ERα and suppress human breast cancer cell proliferation in MCF7 cells (6), we examined whether LJ-529 acted in a similar manner. As seen in Fig. 5C, down-regulation of ERα occurred by LJ-529 in a time-dependent manner in MCF7 and T47D cell lines. Thus, we conclude that LJ-529 down-regulates ERα, whose signaling is critical in cell proliferation of ER-positive breast cancer cells (6).

We further investigated the additional biochemical pathway that may be responsible for growth-suppressive effect in ER-negative breast cancer cells. As SK-BR-3 cells express high levels of c-ErbB2 (24), we examined the expression level of c-ErbB2 in LJ-529-treated SK-BR-3 cells. As seen in Fig. 5D, the expression level of c-ErbB2 was significantly reduced by LJ-529. Furthermore, the expression of total and phosphorylated forms of both ERK1 (p44) and ERK2 (p42) was also significantly reduced by LJ-529. Therefore, we conclude that LJ-529 down-regulated ErbB2 signaling in c-ErbB2-positive SK-BR-3 cells.

To examine the in vivo effect of LJ-529 in breast cancer cells, T47D and SK-BR-3 cells were engrafted s.c. into nude mice. When tumor reached the width of 1 to 3 mm, the mice were treated daily p.o. with the indicated doses of LJ-529. Tumor growth was clearly suppressed in the LJ-529-treated group compared with the vehicle-treated group as examined by Student's t test (P < 0.05 at indicated time points; Fig. 6). When one-way ANOVA was done to examine the group differences between tumor growth rates of T47D xenografts, statistical significance was seen between the groups [F(3, 10) = 11.536, P = 0.001]. Upon post hoc test using Scheffe, significant differences between vehicle- and all three LJ-529-treated groups were revealed (the mean differences were significant at the 0.05 level), but not within the drug-treated groups. The average tumor sizes are shown in Table 1. The results show that LJ-529 inhibits breast cancer growth in xenograft models.

This study presents data showing that the novel A3AR agonist, LJ-529, attenuated in vitro proliferation and in vivo tumor growth of breast cancer cells by inducing apoptosis and deregulating the Wnt signaling pathway. Such effect was seen in both ER-positive and ER-negative breast cancer cells. In addition, ERα expression was down-regulated by LJ-529 in ER-positive cells, which may partially account for the suppression of cell proliferation as IB-MECA does in MCF7 cells (6). Furthermore, c-ErbB2 expression and its downstream ERK activities were down-regulated by LJ-529 in c-ErbB2-overexpressing, ER-negative SK-BR-3 cells. Finally, such effect of LJ-529 occurred independently of the A3AR, because no A3AR expression was detected by RT-PCR in all four breast cancer cell lines used in this study.

The requirement for the signaling via the A3AR has been examined by using the A3AR-specific antagonists, by quantifying the expression by RT-PCR or immunoblotting or by stably overexpressing A3AR. The anticancer effect of A3AR agonists occurs via A3AR in melanoma, prostate, and hepatoma cells (5, 10, 12–14). In agreement with these results, A3AR is reported to be highly expressed in primary tumors, such as colon carcinoma, breast carcinoma, small cell lung carcinoma, pancreatic carcinoma, and melanoma (25). However, the action of A3AR agonist has been reported to be independent of the A3AR in leukemia cells (9). The requirement for A3AR in breast cancer cells is controversial because two research groups have drawn opposite conclusions with the identical A3AR agonist (IB-MECA) on identical cells (MCF7). One group has detected A3AR by RT-PCR and observed partial attenuation of anchorage-dependent cell growth by addition of the A3AR antagonist (7), whereas the other detected no A3AR by RT-PCR and no change in colony formation by stably overexpressing A3AR (6). The reasons for such discrepancies are not known, but for the RT-PCR we found that the primers and the reaction conditions are quite different. The former group (7) ran PCR for 40 cycles with the primers as previously reported (26), whereas the latter group (6) did PCR for 33 cycles with newly designed primers. When nucleotide-nucleotide Basic Local Alignment Search Tool was ran to search for short, nearly exact matches, we found that the antisense primer used by the former group significantly lacks the specificity to A3AR sequence and that it only matches a stretch of 13 nucleotides (from the 8th to 20th nucleotide) out of a 20-mer. In this study, we adopted the conditions of the latter group because the primers were highly specific to A3AR upon Basic Local Alignment Search Tool analysis and detected no A3AR expression in all four cell lines tested.

Despite elucidating several signaling pathways affected by LJ-529 in breast cancer cells, it is not clear how LJ-529 triggers the effect on proliferation and apoptosis in breast cancer cells. One possibility is that LJ-529 at high concentrations binds to a yet unidentified membrane receptor(s) and triggers downstream signaling. Another possibility would be that LJ-529 is transported into the cell via a nucleoside transporter, and the transported compound signals through direct interaction with intracellular targets as reported for adenosine (27) and 2-chloroadenosine (28). Because the growth inhibitory effect of IB-MECA in MCF7 cells was not prevented using the nucleoside transporter inhibitor, Lu et al. (6) have suggested that IB-MECA at high concentrations may compete for the transporter or enter the cell by a nucleoside transporter–independent mechanism. The signaling pathways from the membrane to the cytosol or the transporters of A3AR agonists in cells that do not express A3AR are intriguing and remain to be explored.

The antiproliferative and apoptotic effect of LJ-529 was seen with relatively high doses of the drug at tens-of-micromolar concentrations. Similarly, high doses of IB-MECA are used to attenuate cell growth of MCF7 breast cancer cells (6, 7). Because LJ-529 was effective on all four breast cancer cell lines tested, it is possible that the results were due to toxic effects of the drug at high concentrations. However, we believe this is not the case because we did not observe any toxicity as evidenced by no change in body weight in the animals p.o. treated with LJ-529 for up to 1 month. Therefore, we present the effect of LJ-529 as genuine.

The c-ERBB2 gene is amplified and c-ErbB2 is overexpressed in 25% to 30% of breast cancers, increasing the aggressiveness of the tumor (16). It has also been implicated in the development of resistance to the antiestrogen tamoxifen in both advanced disease and in an adjuvant setting (17). Herceptin, a humanized monoclonal antibody against c-ErbB2, is used clinically in breast cancers that overexpress c-ErbB2 because it seems to block growth signals transmitted by c-ErbB2 to the nucleus and enhances response to chemotherapeutic agents (29, 30). Here, we showed that the expression of c-ErbB2 and it downstream ERK signaling were down-regulated by LJ-529 in c-ErbB2-overexpressing SK-BR-3 cells. Although more study needs to be done to generalize the effect of LJ-529 in c-ErbB2-overexpressing breast cancers, such as testing the effectiveness of the drug in c-ErbB2 transgenic mice, our study strongly suggest LJ-529 as a possible therapeutic agent for this type of aggressive breast cancer.

Recent report of Merighi et al. (14) has shown that Cl-IB-MECA inhibits phosphatidylinositol 3-kinase/Akt–dependent ERK phosphorylation in melanoma cells. Although we have not explored this signaling pathway in breast cancer cells with LJ-529, we reported down-regulation of total and phospho-Akt by LJ-529 in T47D and SK-BR-3 cells. It remains to be seen if ERK phosphorylation is down-regulated in these cells and it is dependent on the inhibition of phosphatidylinositol 3-kinase/Akt by using phosphatidylinositol 3-kinase inhibitors. Apart from the phosphatidylinositol 3-kinase/Akt–dependent regulation of ERKs, we report a novel pathway of c-ErbB2-dependent down-regulation of ERK phosphorylation in SK-BR-3 breast cancer cells by LJ-529.

In summary, we presented LJ-529, an A3AR agonist, as an effective drug that acted independently of the A3AR activity, but with antiproliferative and apoptotic activity in breast cancer cells. We also presented several molecular mechanisms responsible for the effects, and identified a novel pathway of down-regulation of c-ErbB2 by an A3AR agonist in SK-BR-3 breast cancer cells both in vitro and in vivo. Therefore, we suggest LJ-529 as a novel drug that may be effective in the treatment of breast cancer.