A new paradigm for aptamer therapeutic AS1411 action: Uptake by macropinocytosis and its stimulation by a nucleolin-dependent mechanism

Materials

Oligodeoxynucleotides were purchased from Invitrogen (Carlsbad, CA). Sequences used for this study were: AS1411, 5’-d(GGTGGTGGTGGTTGTGGTGGTGGTGG); FL-AS1411 (fluorophore-labeled AS1411), 5’-Fluor-d(TTTGGTGGTGGTGGTTGTGGTGGTGGTGG), where Fluor is either 5-carboxyfluorescein (FAM, used for flow cytometry) or Alexa Fluor 488 (used for confocal microscopy); tAS1411, 5’-d(TTTGGTGGTGGTGGTTGTGGTGGTGGTGG); FL-CRO, 5’-Fluor-d(TTTCCTCCTCCTCCTTCTCCTCCTCCTCC); CRO, 5’-d(CCTCCTCCTCCTTCTCCTCCTCCTCC); and tCRO, 5’-d(TTTCCTCCTCCTCCTTCTCCTCCTCCTCC). Unmodified oligonucleotides were purchased in the desalted form, whereas fluorescently labeled sequences were HPLC purified. The 29-mer sequences were used for some experiments because quenching of the fluorophore occurred when it was located adjacent at the 5’-terminal base of AS1411 (Reyes-Reyes and Bates, unpublished observations), so a spacer consisting of 3 thymidines was added. We confirmed that the antiproliferative activities of 29-mer sequences, with and without the fluorophore, were comparable to the synthesized 26-mer AS1411 sequence, and to AS1411 obtained from Antisoma (Supplementary Figure S1A). The dextran, 10,000 MW, anionic fixable (dextran-10K) and transferrin (Tf) conjugated with Alexa Fluor 488 or Alexa Fluor 594 were purchased from Invitrogen. Anti-rabbit and anti-mouse antibodies linked to horseradish peroxidase, anti-histone 3 rabbit polyclonal and anti-pan cadherin (C19) goat polyclonal antibodies were from Santa Cruz Biotech (Santa Cruz, CA). Anti-nucleolin monoclonal antibodies were from Stressgen (4E2) and Santa Cruz (MS-3). The anti-nucleolin mAb (D3) was a generous gift from Dr. Jau-Shyong Deng, University of Pittsburgh School of Medicine. Cytochalasin D, dynasore, and amiloride were from Calbiochem (San Diego, CA). Triton X-100 was from Sigma (Saint Louis, MO), paraformaldehyde was from Electron Microscopy Sciences (Hatfield, PA), and dimethylsulfoxide (DMSO) was from the American Type Culture Collection (ATCC, Manassas, VA).

Cell Culture and Treatment

All cells were obtained from ATCC and routinely grown in a humidified incubator at 37°C with 5% CO₂. Hs27 (non-malignant human foreskin fibroblasts), DU145 (hormone-refractory prostate cancer), HeLa (cervical adenocarcinoma), MCF-7 (hormone-dependent breast cancer) and MDA-MB-231 (hormone-independent breast cancer) cells were grown in DMEM supplemented with 10% fetal bovine serum (FBS; Life Technologies), 62.5 µg/mL penicillin and 100 µg/mL streptomycin (Hyclone Laboratories, Logan, Utah). MCF-10A cells (immortalized human breast epithelial cells) were grown in MEBM supplemented with all the components of MEGM bullet kit (Lonza, Allendale, NJ) except for the GA-1000. We did not carry out additional testing to authenticate cell lines, but their morphology and behavior were consistent with ATCC descriptions. Cells were plated at 50% confluence and incubated 18 h to allow adherence, and then the medium was changed for fresh supplemented medium and treated by addition of oligodeoxynucleotides directly to the culture medium to give the final concentration indicated in the figure legends. Dynasore and cytochalasin D were dissolved in DMSO. Amiloride was dissolved in serum-free medium. Cells were pre-treated with inhibitors in serum-free medium for either 30 min (cytochalasin D, dynasore) or 60 min (amiloride). Cells for biochemical analyses were lysed in lysis buffer (150 mM NaCl, 2 mM EDTA, 50 mM Tris-HCl, 0.25% deoxycholic acid, 1% IGEPAL CA-630, pH 7.5) containing protease and phosphatase inhibitor cocktails (Calbiochem) for 20 min at 4°C and then cleared by centrifugation at 16,000g for 10 min at 4°C. All protein concentrations were determined using the BCA assay (Pierce, Rockford, IL).

Flow cytometric assays

Cells (2 × 10⁵) in fresh complete medium were plated into 6-well plates for 18 h. After complete adhesion, cells were incubated as indicated in the figures or legends. Cells were washed once with ice-cold PBS, incubated with 1 µg/ml 7-amino-actinomycin D (7-AAD) or 1 µg/ml propidium iodide (PI) for 5 min on ice to allow exclusion of non-viable cells, and washed twice with ice-cold PBS. To harvest, cells were treated with 0.01% trypsin/ 0.5 mM EDTA (300 µl) for 3 min prior addition 3 ml supplemented culture medium. Cells were then centrifuged and resuspended in 0.5 ml of 1% paraformaldehyde for analysis by flow cytometry using a FACScalibur cytometer (BD Biosciences, Mountain View, CA).

Immunofluorescence microscopy

Cells (4 × 10⁴) in fresh complete culture medium were plated on 18 mm diameter glass cover slips for 18 h. The medium was replaced with serum-free medium and cells were treated as described in the figure legends. After incubation, cells were washed 3 times with ice-cold PBS, fixed in 4% paraformaldehyde in PBS for 30 min at room temperature, and washed three times with PBS. After washing, the cover slips were mounted on glass slides with ProLong Antifade (Molecular Probes) according to the manufacturer's directions. Immunofluorescence was documented with an LSM 510 inverted confocal laser-scanning microscope (Carl Zeiss, Oberkochen, Germany) equipped with an Omnichrome argon–krypton laser. Images were obtained with a Zeiss Plan-Apo 63× oil immersion objective (1.4 NA).

Biotinylation and purification of cell-surface proteins

Cells were washed three times with ice-cold PBS and a freshly prepared solution of a cell-impermeable biotinylating agent (sulfo-NHS-biotin, Pierce, Rockford, IL) was added (0.5 mg/ml in PBS). After 30 min incubation at 4°C, cells were washed once with ice-cold TBS (50 mM Tris–HCl, 150 mM NaCl, pH 7.5), incubated with ice-cold complete medium for 10 min at 4 °C, and then washed twice with TBS. Biotinylated proteins were precipitated by incubating with high capacity Neutravidin agarose (Pierce) for 2 h at 4 °C with gentle agitation, and then washed with ice-cold lysis buffer.

RNA interference

Nucleolin siRNA duplexes corresponding to the sequences: 5’-GGUCGUCAUACCUCAGAAGtt (NCL1); 5’-GGCAAAGCAUUGGUAGCAAtt (NCL2); and 5’-CGGUGAAAUUGAUGGAAAUtt (NCL3), were chemically synthesized and annealed by Ambion Inc. (Austin, TX). BLAST analysis showed no homology of the siRNA sequences to any other sequence in the Human Genome Database. The siRNAs were transfected using Lipofectamine 2000 (Invitrogen), according to the manufacturer’s directions. The scrambled siRNA used as a negative control was from Ambion.

Immunoblotting

Samples were resolved by 10% SDS–Tris polyacrylamide gel electrophoresis and then electrotransferred onto polyvinylidine fluoride membranes (Millipore, Bedford, MA) in Tris–glycine buffer containing 20% methanol. Proteins were detected by immunoblotting as described (14). In some cases, membranes were stripped of bound antibodies using 62.5 mM Tris–HCl, pH 6.7, 100 mM 2-mercaptoethanol, 2% SDS for 30 min at 60°C and then reprobed as described in figure legends.

Densitometry and statistical analysis

Densitometry was used to measure band intensities by scanning autoradiographic films and using UN-SCAN-IT gel software (Silk Scientific Corporation). Band intensities were normalized as indicated in the figure legends. The statistical comparisons between AS1411-treated and control groups were performed using Student’s t test.

Uptake of FL-AS1411 occurs through an active uptake process

To first identify suitable conditions for our study, we analyzed the timing and serum-dependence of uptake in DU145 prostate cancer cells, which are sensitive to AS1411. Using flow cytometry with gating to exclude non-viable cells, we compared uptake of FL-AS1411, a fluorescently labeled version of the active aptamer, with that of FL-CRO, a fluorescently labeled control oligonucleotide with no antiproliferative activity. It has not been possible to identify a scrambled or point-mutated analog of AS1411 that lacks activity because of the very high proportion of guanines in the sequence. Therefore, a sequence where the guanines are substitututed with cytosines (FL-CRO) is used simply as a control phosphodiester oligdeoxyonucleotide of the same length. We first confirmed that the fluorophore signal was from internalized, rather than surface-bound, material by showing that cell-associated fluorescence was not influenced by washing the cells with dextran sulfate to remove the extracellular oligonucleotide or by adding trypan blue to quench external fluorescent signals prior to flow cytometry, (Supplementary Figure S2A).

FL-AS1411 uptake was detected as early as 5 min, with maximum signal between 2 h and 4 h under these conditions (Figure 1A). FL-CRO uptake was consistently much lower than FL-AS1411 and followed different kinetics. These data are in accord with our recent finding that G-rich phosphodiester sequences, such as AS1411, are taken up more efficiently than other non-G-rich sequences (15). As shown in Figure 1B, uptake of FL-AS1411 was independent of the presence of serum in the medium.

To determine whether AS1411 uptake occurs through an active uptake process, we evaluated the temperature-dependence of AS1411 uptake in various cancer cell types and non-malignant Hs27 skin fibroblasts. In all cell types examined, the uptake of FL-AS1411 and FL-CRO showed strong temperature dependence. However, in contrast to our original hypothesis, Hs27 cells appeared to have a higher uptake of AS1411 than any of the cancer cells analyzed (Figure 1C). The dose-response curve for FL-AS1411 uptake in DU145 cells (Supplementary Figure S2B) suggested at least two components: uptake involving a high affinity, low capacity receptor, which was saturated at high nanomolar concentrations, and a non-saturable uptake that was predominant at micromolar concentrations (the curve presented an almost linear increase between 1.25 µM and 40 µM). Concentrations higher than 40 µM resulted in obvious cytotoxicity, even at this early time point. We believe that the non-saturable uptake component is most relevant for AS1411 activity, because this corresponds to concentrations at which the biological effects are observed in cultured cancer cells and to the serum concentrations in patients treated with AS1411.

FL-AS1411 uptake occurs through different endocytic mechanisms in cancer cells and in non-malignant cells

To confirm that uptake of AS1411 occurs by endocytosis, we tested for involvement of the actin cytoskeleton, which has been implicated in regulating endocytic pathways. To this end, DU145 and Hs27 were pre-treated with an actin polymerization inhibitor, cytochalasin D, and assessed for FL-AS1411 uptake by flow cytometry. Cytocholasin D-treated cells showed a decrease in FL-AS1411 uptake compared with the untreated cells (Figure 2A), consistent with endocytic uptake. Recognized pathways of endocytosis include classical clathrin-mediated endocytosis, caveolae-mediated endocytosis, clathrin- and caveolae-independent endocytosis, and macropinocytosis (16). The GTPase dynamin is required for clathrin- and caveolae-mediated endocytosis and some clathrin and caveolae-independent pathways (16), so we next investigated the effect of dynasore, a potent inhibitor of dynamin function (17), on FL-AS1411 uptake in DU145 cancer cells and nonmalignant Hs27 cells (Figure 2B). Pre-treatment of Hs27 cells with dynasore decreased their uptake of AS1411 (Figure 2B). In contrast, pre-treatment of DU145 cells slightly increased their uptake of FL-AS1411. To rule out the possibility that DU145 cells were unresponsive to dynasore, we demonstrated that uptake of transferrin, a well-established ligand of clathrin-dependent endocytosis, was inhibited in DU145 cells pre-treated with dynasore, (Supplementary Figure S3A). These results indicate that AS1411 may be taken up by a predominantly clathrin or caveolae-dependent route of entry in Hs27 cells, but not in DU145 cells.

Macropinocytosis is the predominant mechanism of uptake for AS1411 in cancer cells

Recent work has showed that internalization of DNA can be mediated through macropinocytosis (18–20), an actin-driven, ligand-independent mechanism in which cells “gulp” the surrounding medium and any macromolecules it contains. This endocytic mechanism has been shown to be sensitive to amiloride, a specific inhibitor of Na+/H exchange (21), so we tested the effect of amiloride on FL-AS1411 uptake. We found that amiloride pre-treatment caused a reduction in FL-AS1411 uptake in DU145 cancer cells, but a slight increase in the non-malignant Hs27 cells (Figure 2C). There was little effect of amiloride treatment on uptake of FL-CRO in either DU145 or Hs27 cells (Supplementary Figure S3B). Amiloride treatment also decreased AS1411 uptake in other cancer cells (MCF7 and MDA-MB-231, data not shown). These data imply that macropinocytosis is responsible for the internalization of AS1411 in cancer cells.

To confirm these results without using chemical inhibitors, uptake was also examined by confocal microscopy. This showed that FL-AS1411 was localized in confined structures in the cytoplasm of both cancer and non-malignant cells (Figure 3A). As expected, uptake of FL-CRO was much lower than FL-AS1411 (Figure 3B). Interestingly, our studies showed that macropinocytosis (indicated by dextran uptake) is much more active in DU145 cancer cells than in the non-malignant Hs27 cells (Figure 3). Moreover, internalized FL-AS1411 was strongly co-localized with the macropinocytic marker, dextran, in DU145 cells (Figure 3A), but not in Hs27 cells (Figure 3A).

Further experiments were performed to confirm the identity of the vesicles containing FL-AS1411 as macropinosomes, which can be distinguished from other endosomes by their comparative inability to concentrate receptors (22). Therefore, we compared localization of transferrin, a ligand for the transferrin receptor, with that of dextran or FL-AS1411. We observed that transferrin and dextran were localized mainly in distinct non-overlapping vesicles in DU145 and Hs27 cells (Figure 3C) and there was very little overlap between transferrin and AS1411 in DU145 cells (Figure 3D). These combined data confirm that the endocytic process regulating the internalization of AS1411 in DU145 cancer cells is macropinocytosis.

It is also notable that no FL-AS1411 was observed in the nuclear region in these studies. This was in contrast to our earlier (unpublished) studies using fluorescence microscopy, where we had observed some cells with intense nuclear fluorescence after 24 h or more treatment with 10 µM FAM-labeled AS1411. This pattern was observed in cancer cell lines, but not in the non-responsive Hs27 cells. We initially interpreted this result as preferential uptake in cancer cells, but later realized that the cells with strong nuclear staining were likely dying in response to AS1411 and were therefore permeable, whereas that the lack of intense nuclear staining in Hs27 cells reflected that they did not die in response to AS1411. The current confocal microscopy and flow cytometry studies (which are gated to exclude dead or dying cells that are propidium iodide positive due to loss of plasma membrane integrity) demonstrate that overall uptake is not cancer-specific and that AS1411 has a non-nuclear localization in viable cells.

AS1411 stimulates macropinocytosis in cancer cells

Several molecules that are internalized by macropinocytosis, including some cell penetrating peptides and viruses, are known to stimulate macropinocytosis and thereby enhance their own uptake (23–25). Moreover, several research groups have recently demonstrated that hyperstimulation of macropinocytosis can cause a novel form of cell death, characterized by vacuolization, irregular nuclei, and swollen cells (26–29), and AS1411 causes changes in cancer cell morphology with exactly these features (4). Therefore, we investigated whether AS1411 could stimulate macropinocytosis in DU145 cells or non-malignant Hs27 cells. Flow cytometry experiments indicated a significant increase in the uptake of the macropinocytic marker, dextran, in DU145 cells treated with tAS1411 (which is FL-AS1411 without the fluorescent label) for 24, 48, or 72 h, whereas there was no increase in the Hs27 cells (Figure 4A). As in all our flow cytometry experiments, cells were gated to exclude permeable cells, discounting the possibility that this increase was due to cell death. No changes in dextran uptake were observed in DU145 cells treated with the control oligonucleotide, tCRO, (Figure 4A) or with AS1411 for shorter times (1h, 2h, and 4h, data not shown). The tAS1411 was also able to induce hyperstimulation of macropinocytosis in other cancer cells lines (MCF-7 and MDA-MB-231) and had a reduced effect in another non-malignant cell type (MCF-10A) (Figure 4B), suggesting that these novel observations may represent a general difference between the response of cancer cells and normal cells, although further studies are needed to verify this idea. Confocal microscopy confirmed that DU145 cells treated with tAS1411 presented a higher dextran uptake compared to untreated or CRO-treated cells (Figure 4C). Additional experiments (Supplementary Figure S4) confirmed that the 26-mer version of AS1411 was able to induce the same response as tAS1411 (which has three additional nucleotides for reasons described in the Methods section).

This result also implies that AS1411 might actually promote its own internalization by cancer cells. To test this idea, we pre-treated DU145 cells for 24 h with or without tAS1411, then added FL-AS1411 and evaluated uptake after an additional 2 h using flow cytometry. As predicted, DU145 cells pre-treated with tAS1411, but not those that received control pre-treatment, showed an increase in the uptake of FL-AS1411 in DU145 cells, whereas there was no comparable increase in AS1411-treated Hs27 cells (Figure 4D). All of these results indicate that initial AS1411 uptake leads to the stimulation of macropinocytosis, inducing an increase of its own uptake. This idea is not necessarily inconsistent with the time course data (Figure 1A) because the measured fluorescence signal may decrease over time for a number of reasons, including exocytosis of the ligand or fluorescence quenching due to environmental factors such as protein binding or pH.

Having demonstrated that AS1411-induced macropinocytosis could lead to enhanced uptake of a nucleic acid (AS1411) and a polysaccharide (dextran), we also evaluated uptake of a protein, namely, fluorescently labeled transferrin. We observed that pre-treatment with AS1411 led to increased uptake of transferrin in DU145 cells (Supplementary Figure S5). Although uptake of transferrin in untreated cells occurs by dynamin-dependent receptor-mediated endocytosis (Supplementary Figures S3 and S5), the additional uptake induced by AS1411 was apparently due to macropinocytosis because it was completely inhibited by amiloride (Supplementary Figure S5).

Initial uptake of AS1411 is independent of nucleolin

We have shown that nucleolin is the primary molecular target of AS1411 and had previously hypothesized that surface nucleolin may serve as a receptor for AS1411 (2, 13). However, the new data are not consistent with that hypothesis because they indicate that uptake occurs, not by classical receptor-mediated endocytosis, but by macropinocytosis. Therefore, we were curious to learn whether nucleolin played a role in AS1411 uptake. We first assessed the effect of an anti-nucleolin mAb (D3, which we confirmed could bind to surface nucleolin) on uptake of FL-AS1411 in DU145 cells after 2 h incubation and found no effect (Supplementary Figure S6). Next, we carried out similar experiments using siRNAs to knockdown expression of nucleolin. Immunoblot analyses confirmed that expression of total nucleolin could be reduced by more than 80% in cells transfected with nucleolin siRNAs compared with control-transfected cells (Figure 5A). We also showed that these siRNAs could effectively knockdown the cell surface form of nucleolin (Figure 5B), using techniques described in the Methods section. We next used the transfected DU145 cells to assess the uptake of FL-AS1411 after 2 h by flow cytometry analysis and found that knockdown of nucleolin had no effect on FL-AS1411 uptake under these conditions (Figure 5C).

Nucleolin regulates AS1411-induced stimulation of macropinocytosis

Our results (Figure 4) suggest that induction of macropinocytosis may be an important component of AS1411 activity. Therefore, we also determined whether nucleolin knockdown affects the tAS1411-mediated stimulation of macropinocystosis in DU145 cells. As shown in Figure 6A, inhibition of nucleolin expression by specific siRNAs had only a marginal effect on the baseline macropinocytosis, but caused a significant decrease in AS1411-induced macropinocytosis, almost completely blocking this process. Accordingly, the tAS1411-induced uptake of FL-AS1411 was also completely blocked in DU145 cells transfected with nucleolin siRNAs (Figure 6B). These results indicate that, whereas nucleolin does not appear to play a role in the initial macropinocytic uptake of AS1411, it is essential for the AS1411-induced stimulation of macropinocytosis. Consequently, nucleolin is also essential for the induced uptake of AS1411 that occurs at later time points.