Antisense oligonucleotides targeted to the human α folate receptor inhibit breast cancer cell growth and sensitize the cells to doxorubicin treatment

The human α class folate receptor (αhFR) is one member of a multigene family of surface glycoproteins with a very high affinity for folic acid and reduced folates (K_d = 0.1–20 nmol/L; refs. 1–4). Folates are essential for cell survival. They are required for numerous biochemical processes, including DNA and RNA synthesis and transmethylation reactions (1–4). Internalization of folates by αhFR involves receptor-mediated endocytosis (5). Consequently, αhFR is considered an essential component in the cellular accumulation of folates and folate analogues used in chemotherapy.

αhFR is expressed in some (6) but cannot be detected in most normal tissues. In contrast, high levels of αhFR expression have been reported in cancer cells. These include ovarian, breast, brain, lung, and colorectal cancers (7, 8). The relevance of αhFR overexpression in malignant tissues remains unclear. It is possible that elevated levels of αhFR induce cell proliferation not only by mediating folate uptake but also by generating other regulatory signals. αhFRs represent a class of cell surface proteins that are inserted into the membrane through a tail called the glycosylphosphatidylinositol anchor (9–11). Glycosylphosphatidylinositol-linked proteins are enriched in membrane clusters termed lipid rafts, in which various components of the cell signaling machinery are concentrated (12, 13). In the ovarian carcinoma cell line IGROV1, Miotti et al. (14) showed that αhFR distributes in low-density membrane microdomains and that the receptor coprecipitates with the Src family tyrosine kinase lyn and the Gα_i-3 heterotrimeric G-protein subunit. An invitro kinase assay revealed that both of these signaling molecules become phosphorylated in αhFR immunoprecipitates (14). Furthermore, transfection and expression of αhFR in murine NIH/3T3 fibroblast cells, which do not endogenously express αhFR, increased cell growth invitroand in vivo (15). The data suggest that αhFR participates in a macromolecular complex that generates intracellular signals potentially involved in modulating cell survival/proliferation processes.

The data presented above suggested the feasibility of targeting the αhFR for the treatment of cancers with elevated αhFR. To explore this idea, we designed αhFR antisense oligodeoxyribonucleotides (ODN). We have showed previously that ligands such as folate and transferrin can beeffective to target liposomal delivery systems to cancer cells in vitro and in vivo (16, 17). To avoid the complication of activating the αhFR, we used the transferrin-liposome–mediated gene delivery system to examine the effect of αhFR antisense ODNs in cultured human breast cancer cell lines. We found that αhFR antisense ODNs reduced αhFR levels, and this reduction correlated with decreased cell survival. Furthermore, we show that treatment of MDA-MD-435 breast cancer cells with αhFR antisense ODNs sensitized them to doxorubicin. The data support the potential use of αhFR antisense ODNs as an anticancer agent in combination with doxorubicin.

Human nasopharyngeal epidermoid carcinoma (KB cells) as well as MCF7, ZR-75-1, SK-BR-3, MDA-MB-231, MDA-MB-453, T47D, and MDA-MB-468 breast cancer cell lines were obtained from American Type Culture Collection (Rockville, MD). Human breast carcinoma cell line MDA-MB-435, normal mammary immortalized cell lines MCF10 and HS578T-BST, and normal mammary primary cells HMEC were provided by the Georgetown University Medical Center Tissue Culture Core Facility. Breast cancer cell lines were maintained in DMEM with folic acid and 10% fetal bovine serum. KB cells were maintained in DMEM without folic acid and 10% fetal bovine serum. The final folic acid concentration was ∼1 to 10 nmol/L (18). Cell culture medium and medium ingredients were purchased from Biofluids, Inc. (Rockville, MD).

Unmodified ODNs (Table 1) were synthesized by using an automated synthesizer (Applied Biosystems, FosterCity, CA). Purity of 21-mer sense and antisense ODNs was assessed by measurement of 260/280-nm absorbance ratios and by electrophoretic analysis on a 10% polyacrylamide gel. End-modified phosphorothioated ODNs were purchased from Ampligene Biotechnologies (Rockville, MD).

Immunoprecipitations were done as described previously (19). KB cells (10⁶ cells per 35-mm well) were plated in 3mL DMEM without folic acid and with 10% fetal bovine serum. Cells were washed once with 2 mL DMEM (Select Amine, Life Technologies, Inc., Rockville, MD) containing no methionine or cysteine and then incubated for 1 hour in 2 mL of the same medium. Newly synthesized proteins were radiolabeled by the addition of 50 μCi [³⁵S]methionine and 50μCi[³⁵S]cysteine in 1 mL of growth medium containing no methionine or cysteine. Cells were incubated in medium containing radiolabeled amino acids for 16 hours in the presence of unmodified αhFR sense and antisense ODNs at the indicated concentrations. After labeling, cells were scraped and lysed by repeated freezing and thawing. Membrane proteins were solubilized in 1 mL PBS containing 1% Triton X-100 (v/v). The protein concentrations were determined by a protein assay reagent (Pierce Chemical Co., Rockford, IL). Protein samples (10 μg) were immunoprecipitated using 50 μL of polyclonal rabbit anti-αhFR antiserum and protein A-Sepharose in 750 μL of immunoprecipitation buffer [10 mmol/L Tris, 150 mmol/L NaCl, 2mmol/L EDTA (pH 7.4), 1% Triton X-100, and 0.1% SDS] for 3 hours at 4°C (20). Immunoprecipitates were boiled for5 minutes in the same buffered solution and pelleted at2,500 × g for 5 minutes. Supernatants were fractionated by12% SDS-PAGE and visualized and quantitated by PhosphorImager (Molecular Dynamics, Sunnyvale, CA).

The relative abundance of αhFR transcripts in normal human tissue was determined with RNase protection assay as described previously (21). RNA was purchased from Clontech (Palo Alto, CA) and the integrity and quantity of RNA was verified by Northern analysis. The 5′ EcoRI-HincII restriction fragment from the KB4 cDNA clone was radiolabeled using the Promega (Madison, WI) in vitro transcription kit and used as the αhFR riboprobe. Total RNA (25 μg), internal control RNA (2 μg), and αhFR riboprobe (100,000 counts/min) in 30 μL of hybridization buffer [40 mmol/L PIPES (pH 6.4), 80% (v/v) formamide, 0.4 mol/L NaCl, and 1 mmol/L EDTA] were denatured at 85°C for 8 minutes and then reannealed at 45°C for 18hours. The ssRNA was digested by the addition of 350μL of buffer [10 mmol/L Tris-HCl (pH 7.5), 200 mmol/L NaCl, 100 mmol/L LiCl, and 2 mmol/L EDTA] containing RNase A (14 μg) and RNase T₁ (0.7 μg) followed by incubation at 25°C for 15 minutes. The digestion was terminated by the addition of 10% SDS (w/v) and proteinase K (100 μg) and incubation at 37°C for 30 minutes. After extraction in an equal volume of phenol/chloroform/isoamyl alcohol (25:24:1, v/v/v), RNA was precipitated with ethanol and fractionated on a 6% sequencing wedge gel. The dried bands were visualized using a PhosphorImager (Molecular Dynamics).

Cellular lysates for Western analyses were prepared as described previously (22). Breast cancer and normal mammary cells were washed twice in room temperature PBS and harvested by scraping in cold lysis buffer containing PBS, 1% NP40, 0.5% sodium deoxycholate, 0.1% SDS, 0.1mg/mL phenylmethylsulfonylfluoride, 30 μg/mL aprotinin, and 1 mmol/L sodium orthovanadate. After 20minutes in an ice bath, lysates were passed through a 21-gauge needle, incubated on ice an additional 30 minutes, and then centrifuged at 15,000 × g for 20 minutes at 4°C. Protein concentrations were determined using the Pierce Micro bicinchoninic acid protein assay reagent and the proteins were stored at −80°C.

Primary anti-human αhFR polyclonal rabbit antibody was generated as described previously (20). To visualize the transferrin receptor in breast cancer cells, we used mouse anti-human transferrin receptor (Zymed Laboratories, Inc., San Francisco, CA). The anti-human glyceraldehyde-3-phosphate dehydrogenase rabbit polyclonal antibody was purchased from Trevigen (Gaithersburg, MD). The anti-actin (C-11) goat polyclonal antibody and all secondary horseradish peroxidase–conjugated antibodies were obtained from Santa Cruz Biotechnologies (Santa Cruz, CA).

Samples containing 20 μg of total cellular protein were electrophoresed on a 4% to 20% SDS-PAGE gradient gel and electroblotted onto a 0.2-μm nitrocellulose membrane (Bio-Rad Laboratories, Hercules, CA). To block nonspecific binding, the membrane was incubated at room temperature for 1 hour with 5% nonfat dry milk in 10 mmol/L Tris-HCl buffer (pH 8.0) containing 150 mmol/L NaCl and 0.05% Tween 20 (TBST). The blot was probed for 1 hour with primary antibody and washed thrice for 15 minutes each with TBST. The specific protein was detected using secondary horseradish peroxidase–conjugated immunoglobulin G. The membrane was probed with secondary antibody for 45 minutes and washed thrice for 15 minutes each with TBST. Proteins were visualized using the enhanced chemiluminescence Western blotting detection reagent and Hyperfilm enhanced chemiluminescence (Amersham, Piscataway, NJ).

Transfection of αhFR S6 and AS6 end-modified phosphorothioated ODNs into breast cancer cells was accomplished using a transferrin-mediated liposomal Be delivery system as described previously (16, 17). Briefly, 2 × 10⁴ cells were plated per well in a 96-well plate and washed twice with serum-free DMEM. Transferrin-liposome-ODN complex was prepared by mixing 50 nmol liposome B with 625 μg transferrin and then 5 nmol ODNs in 800 μL of serum-free DMEM at a final ODN concentration of 5 μmol/L and then added to cells at indicated concentrations for 5 hours. Transfections were terminated by theaddition of an equal volume of DMEM containing 10% fetal bovine serum. Cell survival was determined after 48hours by 2,3-bis[2-methoxy-4-nitro-5-sulfophenyl]-2H-tetrazolium-5-carboxanilide inner salt (XTT) assay according to the manufacturer's instructions (Boehringer Mannheim,Indianapolis, IN). In the presence of an electron coupling reagent, XTT is converted into orange formazan by dehydrogenase in the mitochondria of living cells. The formazan absorbance, which correlates to the number of living cells, was measured at 450 nm using a microplate reader (Molecular Devices, Menlo Park, CA). The IC₅₀ was interpolated from the graph of the log of drug concentration versus the fraction of surviving cells.

For chemosensitization studies, MDA-MB-435 cells were plated and transfected with 500 nmol/L of αhFR AS6 or S6 ODN as described above. Twenty-four hours after transfection, the medium was replaced with DMEM containing 10% fetal bovine serum and doxorubicin (Bedford Labs, Bedford, OH) at the indicated concentrations. Cell survival was determined by XTT assay, as described above, after 72hours.

Eight antisense ODNs were designed to target the αhFR open reading frame (Fig. 1A) to determine whether decreased levels of αhFR impact cell survival. For preliminary experiments, we used cultured human nasopharyngeal cells (KB cells), which express high levels of αhFR. KB cells were incubated with αhFR antisense ODNs (AS1, AS2, and AS6) at the indicated concentrations for 6 hours. Figure 1B shows a dose-dependent decrease in the levels of αhFR protein up to 60% 2 days after the treatment as determined by immunoprecipitation analysis. In contrast, αhFR sense ODNs (S1, S2, and S6) had little effect on the levels of αhFR in KB cells (Fig. 1C). The data correspond to survival analysis of KB cells, which decreased after 2 days of αhFR antisense exposure (data not shown).

The levels of αhFR are reportedly higher in cancer compared with normal tissue (7, 8). We determined the levels ofαhFR transcripts in a panel of normal tissue using a RNase protection assay. Figure 2 shows that αhFR transcripts are present in some tissues (i.e., the lung, kidney, breast, salivary gland, and placenta) but are not detected orare present at very low levels in most other normal tissue; in contrast, much higher levels are present inKB cells. The data confirm and extend previous observations (6).

We were interested in evaluating αhFR antisense ODNs as a potential anticancer therapeutic agent for the treatment of breast cancer. Previous studies have reported expression of αhFR in MDA-MB-231 (22) and T47D (3) but not in ZR-75-1 (23) and MCF7 (19) cultured, human breast cancer cell lines. In Fig. 3, we did Western blot analysis of αhFR levels in a panel of eight cultured breast cancer cell lines (MCF7, ZR-75-1, SK-BR-3, MDA-MB-435, MDA-MB-231, MDA-MB-453, T47D, and MDA-MB-468), two normal mammary immortalized cell lines (MCF10A and HS578T-BST), and normal mammary primary cells (HMEC). We have detected αhFR at variable levels in all tested samples. Corrected for protein loading (based on actin levels), it seems that most of the breast cancer cell lines express higher levels of αhFR compared with that in the normal cells.

Transferrin-liposome Be–mediated gene delivery has been shown both in vitro and in vivo (24, 25). Transferrin is a plasma glycoprotein used by cells to acquire iron via a process involving receptor endocytosis. There is evidence for overexpression of the receptor for transferrin on tumor cells, including oral (26, 27), prostate (28), and breast (29) cancer cells. We assessed the levels of transferrin receptors by Western blot analysis in cultured breast cancer and normal epithelial mammary cell lines. As shown in Fig. 3, breast cancer cell lines express much higher levels of transferrin receptor compared with immortalized and primary normal cell lines. We have showed previously that both transferrin and folate can be an effective ligand totarget cancer cells in vitro and invivo (16, 17). Thus, in this study, to avoid the complication of activating the αhFR, we have used the transferrin-liposome Be–mediated gene delivery system to examine the effect of αhFR antisense ODNs in cultured human breast cancer cell lines.

To determine the effect of αhFR antisense ODNs on the survival of cultured breast cancer cells, we transfected end-modified phosphorothioated AS6 ODNs and S6 ODNs (Table 1) at various concentrations via the transferrin-liposome Be–mediated gene delivery system. As shown inFig. 4A, increased amounts of transfected AS6 ODNs correlated with decreased cell survival as determined by XTT assay. The IC₅₀ values (μmol/L) for the treated breast cancer cell lines were 0.633 (T47D), 0.600 (MDA-MB-231), and 0.416 (MDA-MB-435). In contrast to these data, cell lines treated with S6 ODNs had significantly less effect onthe survival of model breast cancer cell lines (Fig. 4B).

Antisense ODNs are believed to abrogate protein function by direct hybridization of the ODNs to exposed regions of the targeted mRNA. The resulting RNA-DNA is thought to preclude the translation of the mature, functional protein, thereby diminishing the overall cellular target protein levels (30, 31). To examine if the αhFR ODN effect on cell survival was correlated to decreased levels of αhFR protein, we transfected MDA-MB-435 breast cancer cells with end-modified phosphorothioated AS6 and S6 ODN, as shown in Fig. 5, and did Western analyses. We observed a significant decrease in the levels of αhFR after treatment with 0.25 and 0.50 μmol/L AS6 ODNs for 24 or 48 hours. However, the αhFR levels of the S6 ODN–treated samples remained relatively indistinguishable from the untreated cells or cells treated with liposome alone. These results strongly suggest that the cellular levels of αhFR protein are an important component of cell survival.

To determine whether αhFR antisense ODN sensitizes breast cancer cells to chemotherapy, we transfected, using transferrin-liposome Be, 500 nmol/L of end-modified phosphorothioated AS6 ODN or S6 ODN into MDA-MB-435 cells and treated them with varying concentrations of doxorubicin 24 hours after transfection. Cell survival was determined after 72 hours by XTT assay and the degree of sensitization was calculated by determining the ratio of IC₅₀ values of the combined treatment relative to that of the S6 ODN and untreated controls. In Fig. 6, chemosensitization of MDA-MD-435 cells by AS6 ODN was ∼5-fold (IC₅₀^S6 / IC₅₀^AS6 = 5) relative to the sense control and ∼10-fold (IC₅₀^Untreated / IC₅₀^AS6 = 10) relative to the untreated control.

αhFR is one member of a multigene family of folate binding proteins. αhFR is considered an important component in the maintenance of cellular folate homeostasis. αhFR binds folates with high affinity (K_d = 0.1–20 nmol/L) and internalizes the vitamin by a process involving receptor-mediated endocytosis (1–4).

We and others (6–8) have reported that αhFR levels are higher in cancer cells than in normal tissue (Fig. 2). For example, αhFR is elevated in >90% of human epithelial ovarian malignancies as compared with normal epithelial ovarian cells (32). Other cancers with elevated αhFR relative to their normal counterpart include breast, brain, lung, and colorectal cancers (7, 8). Consequently, several strategies have been developed for the targeted delivery of drugs to αhFR-positive tumor cells (33). Covalent attachment of therapeutic agents to αhFR-targeted monoclonal antibodies has been evaluated for imaging and immunotherapy and have shown significant targeting efficacy in patients with ovarian cancer (34). Alternatively, folate-derivatized anticancer treatments have been successfully applied in vitro for αhFR-specific delivery (35). Of interest, low molecular weight radiopharmaceuticals based on folate conjugates have shown favorable pharmacokinetic properties and tumor selectivity in αhFR-positive animal tumor models (36, 37). Unlike previous reports that have used αhFR as a drug delivery vehicle, the present study isthe first account to target αhFR using gene-specific antisense ODN.

The purpose of elevated αhFR in cancer is unclear. It has been proposed that αhFR has a role in promoting cell proliferation independent of folate internalization. Bottero et al. (15) showed that transfection of αhFR into nonexpressing NIH/3T3 cells increased growth in vitro and invivo. The data presented in this report agree with these observations. We show that reducing the levels of αhFR, by αhFR-targeted antisense ODNs (Fig. 1A), inhibited cell survival in cultured human breast cancer and KB cells after36 hours (Fig. 1B). Decreased cell number is not likely to be attributable to folate starvation, a process that occurs over the order of weeks. The literature suggests that αhFR may be involved in cell processes that are unrelated to folate internalization. For example, αhFR, similar to other glycosylphosphatidylinositol-linked proteins (12, 13), may have a role in cell signaling processes leading to cell growth and proliferation (14). Further studies are required to delineate the mechanism(s) by which αhFR ODN mediates death of cancer cells.

In the present study, we were interested in testing αhFR antisense ODNs as a potential anticancer therapeutic agent for the treatment for breast cancer. We evaluated the levels of αhFR in breast cancer and normal mammary cells. As shown in Fig. 3, αhFR was present in all tested cells; however, unlike other disease models, the levels of αhFR were not significantly elevated in the cultured breast cancer cells compared with normal mammary cells. In an attempt to achieve breast cancer cell–specific targeting, we used a transferrin receptor–mediated delivery system. In Fig. 3, the transferrin receptor is more highly expressed in breast cancer cell lines than in the immortalized and primary normal mammary cells. The data show the feasibility of delivering αhFR antisense ODNs via a transferrin targeting liposomal Be delivery system even where αhFR is not overexpressed.

The mechanism by which αhFR antisense ODNs selectively decreased αhFR protein in MDA-MB-435 breast cancer cells (Fig. 5) is not entirely clear. In general, antisense ODNs are believed to abrogate protein function by direct hybridization of the ODNs to exposed regions of the targeted mRNA. The resulting RNA-DNA is thought to preclude the translation of the mature, functional protein, thereby diminishing the overall cellular target protein levels. The antisense-mediated decrease of protein production is hypothesized to occur by (a) RNase H digestion of the RNA-DNA duplex or (b) physical disruption of the translation machinery (30, 31). Nonantisense mechanisms may also be involved in selective protein targeting by ODN. For example, aptamers are a class of small nucleic acid ligand agonists that exhibit exquisite specificity for proteins (38, 39). Additional studies are required to understand the mechanism of αhFR antisense ODN–mediated selective decrease of αhFR protein in cancer cells.

The current study suggests an application of αhFR antisense ODNs for the treatment of αhFR-positive cancers, particularly ovarian cancer. αhFR has been characterized asa marker for ovarian carcinoma because it is expressed in90% of epithelial ovarian carcinomas and is absent in normal ovarian epithelial cells (40). Furthermore, αhFR appearss very early in the transformation process and even increases with tumor progression (32). The pathology of ovarian carcinogenesis has been well documented. In the early stages, the spread of the disease typically involves theopposite ovary, the uterus, and the fallopian tubes (41). In the more advanced stages, metastases occur at the peritoneal surfaces of the bladder, the rectosigmoid, or thepelvic peritoneum (42). The confinement of ovarian metastatic spread to the i.p. cavity has allowed for treatment strategies involving i.p. therapy (43). Wide expression of αhFR in ovarian carcinoma and its association with malignant transformation and cell survival processes as well as local disease metastasis support i.p. administration of αhFR-targeted antisense molecules for the treatmentof ovarian cancer.