We have identified two synthetic oligonucleotides (aptamers) that bind to prostate cancer cells,with low nanomolar affinity, via the extracellular portion of the prostate-specificmembrane antigen (PSMA). These two specific aptamers were selected from an initial 40mer library of ∼6 × 1014 random-sequence RNA molecules for their ability to bind to a recombinant protein representing the extracellular 706 amino acids of PSMA, termed xPSM. Six rounds of in vitro selection were performed, enriching for xPSM binding as monitored by aptamer inhibition of xPSM N-acetyl-α-linked acid dipeptidase (NAALADase) enzymatic activity. By round six, 95% of the aptamer pool consisted of just two sequences. These two aptamers, termed xPSM-A9 and xPSM-A10, were cloned and found to be unique, sharing no consensus sequences. The affinity of each aptamer for PSMA was quantitated by its ability to inhibit NAALADase activity. Aptamer xPSM-A9 inhibits PSMA noncompetitively with an average Ki of 2.1 nm, whereas aptamer xPSM-A10 inhibits competitively with an average Ki of 11.9 nm. Distinct modes of inhibition suggest that each aptamer identifies a unique extracellular epitope of xPSM. One aptamer was truncated from 23.4 kDa to 18.5 kDa and specifically binds LNCaP human prostate cancer cells expressing PSMA but not PSMA-devoid PC-3 human prostate cancer cells. These are the first reported RNA aptamers selected to bind a tumor-associated membrane antigen and the first application of RNA aptamers to a prostate specific cell marker. These aptamers may be used clinically as NAALADase inhibitors or be modified to carry imaging agents and therapeutic agents directed to prostate cancer cells.
Molecules able to bind tightly and specifically to the surface of malignant cells would greatly benefit cancer diagnosis and treatment. Whereas antibodies have the ability to specifically recognize tumor cell markers, large size and immunogenicity often limit their pharmacological value. Humanized antibodies, antibody fragments, and short peptides show great promise but are still limited by peptidase susceptibility and the immune response. The recent development of the SELEX 4 process has provided a new alternative, yielding nuclease-stabilized oligonucleotides that can be selected to bind tightly and specifically to a given ligand (1, 2, 3) . The diversity of structures exhibited by an aptamer library allows selection of tight binding aptamers for simple targets, such as a single amino acid (4) , or to complex targets such as tumor cell lines (5) . Such oligonucleotides, termed “aptamers,” have been made to >100 different ligands and are emerging as a new class of molecules that contest antibodies in therapeutics, imaging, and diagnostics (6 , 7) .
Here we describe the in vitro selection of two novel nuclease-stabilized RNA aptamers evolved to bind tightly to the extracellular portion of a well-characterized surface antigen overexpressed on human prostate cancer cells PSMA (8) .
MATERIALS AND METHODS
The cell lines Sf-9, LNCaP, and PC-3 were obtained from the American Type Culture Collection (Rockville, MD). Ready Plaque Sf-9 cells, obtained from Novagen, Inc. (Madison, WI), were applied to serum-free protein purification. A long-term passage LNCaP derivative, termed LNCaP-Parent, was generously provided by Dr. Joel Nelson (Department of Urology, University of Pittsburgh School of Medicine, Pittsburgh, PA). LNCaP-Parent, known to be PSMA positive, was used in fluorescent microscopy because the cells were found to adhere better during formalin fixation than earlier passage LNCaP cells. All of the cells were grown according to the manufacturer’s specifications.
Cloning PSMA cDNA from LNCaP Cells.
First-strand cDNA was synthesized from 2 μg total RNA using 1 μg oligo(dT)12–18 primer and Superscript II RNase H reverse transcriptase (Life Technologies, Inc.) according to the manufacturer’s instructions. The PSMA coding region was cloned by primers TGCAGGGCTGATAAGCGAG and TCTTTCTGAGTGACATAC using high-fidelity Pfu DNA Polymerase (Stratagene, La Jolla, CA). The isolated product was TA-cloned into the pCR-2.1 vector (Invitrogen Corporation, Carlsbad, CA) and verified by sequencing.
Preparation of Recombinant xPSM-expressing Baculovirus.
Primers were designed to overlap the extracellular portion of PSMA, with a single NH2-terminal amino acid change from lysine to glycine, for direct fusion to the enterokinase cut site of the pBacgus-10 vector (Novagen, Inc.) using primers PSMA (394)63 GCCCTCTAGAGGTGGAGGTGATGACGACGACAAGGGTTCCTCCAATGAAGCTACTAACATTAC and 2PSMA(2491)31 CCCAAGCTTAGGCTACTTACACTCAAAGTCTC. The integrity of the resulting plasmid, pBACgus-PSM was confirmed by sequencing. Sf-9 cells were cotransfected with pBACgus-PSM and linearized high efficiency BacVector-3000 Triple Cut Virus DNA (Novagen, Inc.). Individual recombinant virus plaques were assayed for recombinant protein expression using S-tag assays and S-tag Westerns (Novagen, Inc.) according to manufacturer’s instructions. Positive clones were amplified to high titer (≥107 pfu/ml) in 200 ml suspension cultures.
Large Scale xPSM Expression and Purification.
Ready Plaque Sf-9 cells were infected with recombinant virus at a multiplicity of infection of 5 pfu/cell. Infected cell medium was harvested 72–80 h after infection, and recombinant protein levels were quantitated by S-tag assay. Before purification, S-protein agarose beads (Novagen, Inc.) were washed several times in Bind/Wash buffer [20 mm Tris Tris-HCl (pH 7.5) and 150 mm NaCl] to remove all of the EDTA. Fusion protein was bound to S-protein agarose (1 ml S-protein beads/500 μg fusion protein) for 12–18 h at room temperature, washed, and cleaved by 10–20 units of rEK (Novagen, Inc.) overnight at 37°C. The rEK was removed using EKapture agarose beads (Novagen, Inc.) according to manufacturer’s instructions. Protein was concentrated by Ultra-Free 15, MWCO Mr 50,000 concentration spin columns (Millipore Co., Bedford, MA) and stored at −20°C.
Approximately 100–500 ng of purified protein was separated by 7.5% SDS-PAGE and stained using the Silver Stain Plus kit (Bio-Rad Laboratories, Hercules, CA). All of the purified xPSM protein was screened for size and purity using silver-stained gels.
NAAG hydrolysis was performed essentially as described previously (9) . In short, LNCaP cell extracts were prepared by sonication in NAALADase buffer [50 mm Tris (pH 7.4) and 0.5% Triton X-100]. Using this buffer, either cell lysate or purified xPSM was incubated in the presence of the radiolabeled substrate N-acetyl-l-aspartyl-l-(3,4-3H)glutamate (NEN Life Science Products, Boston, MA) at 37°C for 10–15 min. The reaction was stopped by the addition of an equal volume of ice-cold 100-mm sodium phosphate and 2 mm EDTA. Products were partitioned by AG 1-X8 formate resin (Bio-Rad Laboratories) anion exchange chromatography, eluted with 1 m sodium formate, and quantitated by scintillation counting. In general, aptamer IC50 was determined in the presence of 8 nm NAAG with serially diluted RNA. Aptamer Ki was determined using 5–30 nm aptamer in the presence of serially diluted substrate. General NAALADase assays and aptamer inhibition experiments showed no significant differences when done in the presence of 1 mm of cobalt chloride. Trend lines were determined using linear regression for Ki plots and logarithmic trend lines for enzyme kinetics, applying R2 to determine fitness. Points >3 SD away from the mean were considered outliers. Competitive inhibition Ki s were calculated by changes in Km, where noncompetitive inhibition Kis were calculated by changes in Vmax. Aptamer inhibition experiments were repeated at least once to confirm affinity and mode of inhibition. Experiments were designed to allow 20% or less of the total substrate to be cleaved.
xPSM Protein Magnetic Bead Preparation.
Aliquots of M450 magnetic beads (24 mg/ml; Dynal Biotech Inc., Lake Success, NY) were washed in potassium phosphate buffer, HBSMC buffer [HEPES buffered saline (pH 7.4), 1 mm MgCl2, and 1 mm CaCl2], and resuspended in 100 μl of HBSMC buffer. Approximately 10 μg of purified xPSM was incubated with 100 μl magnetic beads overnight at 4°C in HBSMC, washed three times in HBSMC, three washes in HBSMCIT [HBSMC, 0.01% I-block (Tropix Inc., Bedford, MA), and 0.05% Tween 20], and storage at 4°C. Predetermined volumes of xPSM magnetic beads were incubated for 30 min in HBSMCIT at 37°C before incubation with aptamers.
In Vitro Selection.
Iterative rounds of aptamer selection and amplification were performed as described previously (10) . In short, aptamer library template DNA (5′-TCGGGCGAGTCGTCTG-40N-CCGCATCGTCCTCCC), designated 40N7, was prepared on an automated solid-phase synthesizer (Applied Biosystems Inc., Foster City, CA) according to manufacturer’s protocol. The T7 RNA polymerase promoter was attached to the template by Klenow elongation with the primer (TAATACGACTCACTATAGGGAGGACGATGCGG), where the T7 promoter is underlined. The 2′FY-RNA library was prepared by T7 in vitro transcription in 1× T7 RNAP buffer [4% (w/v) polyethylene glycol 8000, 40 mm Tris-HCL (pH 8.0), 12 mm MgCl2, 5 mm DTT, 1 mm spermidine chloride, and 0.002% Triton X-100], 0.000125 units/μl inorganic pyrophosphatase, 2.5 mm guanosine, 3 mm 2′F-CTP, 3 mm 2′F-UTP, 1 mm ATP, 1 mm GTP, and 356 μm T7 RNA polymerase at 37°C for 4–16 h. Products were then treated with 10 units of DNase I (Roche) for 10 min at 37°C and stopped with EDTA to 20 mm. Products were separated by 6 m urea/8% PAGE, excised, purified, and quantitated. The purified 2′FY-RNA library was diluted in HBSMCIT buffer and incubated with the prewarmed xPSM beads for 30 min at 37°C, followed by five washes in 500 μl of 37°C HBSMCIT using a magnetic separator. Remaining xPSM-bound 2′FY-RNAs were bound to primer by first heating the beads to 95°C in 20 μl of 5 μm 3′-primer (TCGGGCGAGTCGTCTG) for 5 min, followed by slow cooling to room temperature and reverse transcription in 1 × reverse transcriptase buffer (Roche), 5 mm deoxynucleotide triphosphates, and 5 units/μl avian myeloblastosis virus reverse transcriptase (Roche) at 48°C for 30 min. The resulting cDNA was PCR amplified and quantitated using the ABI 7700 Sequence Detector (Applied Biosystems Inc.) in 1× SQ buffer [50 mm KCL, 10 mm Tris-HCL (pH 8.3), 7.5 mm MgCl2, 1 mm deoxynucleotide triphosphates, and 2 μm 5,6 ROX], 0.3 μm 5′-FD2 primer (60FAM-GAGCGAAGC-dabcyl-CTAATACGACTCACTATAGGGAGGACGATGCGG; Operon), 0.3 μm 3′-primer (TCGGGCGAGTCGTCTG), and 0.05 units/μl Taq polymerase for 35 cycles. A serial dilution of 105 to 1011 molecules of randomized 40N7 2′FY-RNA was amplified as a quantitative reference, and linear regression analysis was applied to quantitate the number of aptamers bound to xPSM magnetic beads. Signal to noise was calculated by comparison of aptamer binding to protein-free magnetic bead controls. PCR products (50 μl) from each round were used as templates for the next round.
Cloning and Sequencing.
Aptamer pools were amplified by 5′-primer (GTGCTGCAAGGCGATTAAGTTGG) and 3′-primer (ACTTTATGCTTCCGGCTCG), and cloned into pUC9 plasmid. Sixty individual plasmids were sequenced using DYEnamic ET-terminator cycle sequencing premix kit (Amersham Pharmacia Biotech, Inc., Piscataway, NJ) and ABI Prism 377 sequencer.
Two resulting aptamers, xPSM-A9 and xPSM-A10, were serially 3′-truncated in five nucleotide permutations by PCR primer design, amplification, and in vitro transcription, and were screened for xPSM NAALADase inhibition. The 3′-truncates still able to inhibit NAALADase activity were additionally reduced by 5′-truncation in five nucleotide permutations by solid-phase synthesis and screened for xPSM NAALADase inhibition.
5′-Hexyl-amine-phosphoramidite (Glen Research, Sterling, VA) 2′FY-aptamers A10-3 (GGGAGGACGAUGCGGAUCAGCCAUGUUUACGUCACUCCUUGUCAAUCCUCAUCGGCdT) and A10-3-rndm (CAGGCAUGCCUAGCUAAGCAGCCCAUGGCU-UAUGCGCGGAAUAUUGGCUUCCGUUCdT) were chemically synthesized with a deoxy-T 3′-3′cap. These were 5′ end labeled with rhodamine-red-x succinimidyl ester (Molecular Probes, Inc., Eugene, OR) according to manufacturer’s instructions, gel purified, and quantitated by spectrophotometry. LNCaP-Parent and PC-3 cells were plated at a density of 5 × 104 cells/well on four chamber glass slides (Becton Dickinson, Franklin Lakes, NJ). Twenty-four h after plating, slides were fixed in 10% buffered formalin for 8–16 h at room temperature and stored at 4°C in 1× PBS (pH 7.4), without magnesium or calcium. Slides were stained with 50 nm rhodamine-red-x-labeled aptamer in PBS for 10–15 min at room temperature, washed in PBS, coverslipped, sealed, and imaged with a Zeiss axioskop epifluorescence microscope equipped with a short arc mercury lamp illumination (Carl Zeiss Inc., Thornwood, NY) and cooled CCD camera (Micro MAX Digital Camera; Princeton Instruments, Trenton, NJ). Images were equally processed in Adobe PhotoShop (Adobe Systems Inc., Seattle, WA) to increase brightness and contrast.
Purification of Target PSMA Fusion Protein.
Because the ultimate application of these aptamers is to bind prostate cancer cells in vivo, only the extracellular portion of PSMA was considered a target. Using a baculovirus expression system, we purified a fusion protein target containing a modified extracellular form of PSMA that we term xPSM (Fig. 1A) ⇓ . The recombinant xPSM contains the entire extracellular sequence of PSMA, with a single NH2-terminal alteration of lysine to glycine. This system was designed to secrete fusion protein, which can then be purified from the growth medium by the use of two separate affinity tags, followed by release via peptidase.
Medium from baculoviral-infected cells was harvested at 72–80 h after infection, and protein purified by S-protein agarose and rEK. After digestion, the enterokinase was captured with a second affinity resin, leaving only pure xPSM. NH2-terminal sequencing confirmed that enterokinase correctly removed the entire affinity tag. The protein is pure by evidence of >95% of the silver stained material consisting of a single band of correct molecular weight (Fig. 1B) ⇓ . The size of purified xPSM has been calculated as 88.8 kDa, suggesting glycosylation of the expected 79.5 kDa product.
Demonstration of xPSM NAALADase Activity.
The xPSM fusion protein was tested for enzymatic activity to ensure native protein conformation. This is important to avoid selecting aptamers that recognize an improperly folded fusion protein but not the native enzyme. The purified xPSM protein possessed NAALADase activity with a Km of 16.1 nm for NAAG and a Vmax of 12.1 mmol/mg·min (Fig. 1C) ⇓ . This purified protein was then immobilized on magnetic beads as a means to partition bound RNA aptamers during selection.
The in vitro selection strategy was designed to identify aptamers that would be applicable under some expected physiological conditions. To increase nuclease stability of the aptamer in serum, 2′-fluro-pyrimidines were incorporated during aptamer transcription. Additionally, aptamers were selected to bind xPSM at 37°C (pH 7.4) to approach in vivo conditions.
A library of ∼6 × 1014 random-sequence 2′FY-RNA molecules was generated by in vitro transcription of a synthetic template and incubated for 30 min with xPSM bound to magnetic beads (Fig. 2A ⇓ , schematic). The protein-bound RNA was then partitioned by magnetic separation, and amplified and quantitated by reverse transcription and real-time PCR. The products of PCR amplification were then used as templates for in vitro transcription, generating a new pool of 2′FY-RNA enriched for xPSM binding. Six rounds of iterative selection and amplification were performed and quantitated as illustrated in Table 1 ⇓ . The signal to noise peaked at round six and showed no additional improvement for up to nine total rounds of selection.
Aptamer Pool Inhibition of PSMA NAALADase Activity.
Given that enzyme assays provide a sensitive method to identify and quantitate enzyme-ligand interactions, we assayed aptamer pools from several SELEX rounds for their ability to inhibit xPSM NAALADase activity. The original random sequence library had no effect on xPSM NAALADase activity, whereas micromolar aptamer inhibition could be seen as early as round three (Fig. 2B) ⇓ with maximum inhibition at round six.
Sequencing of Individual Aptamers and Calculated Affinities.
Round six RNA was amplified by reverse transcription-PCR and cloned. Sixty randomly picked plasmid clones were sequenced. Ninety-five percent of round six aptamers were represented by only two sequences (Fig. 2C) ⇓ . The identified sequences, termed xPSM-A9 and xPSM-A10, are unique and share no consensus sequences.
The affinity of each aptamer for PSMA was quantitated by its ability to inhibit NAALADase activity. Aptamer xPSM-A9 inhibits PSMA noncompetitively with a calculated Ki of 2.1 nm (Fig. 3B) ⇓ where aptamer xPSM-A10 inhibits competitively with a calculated Ki of 11.9 (Fig. 3A) ⇓ . These two separate modes of inhibition suggest that each aptamer identifies a unique extracellular epitope of PSMA. Both aptamers inhibit native NAALADase activity from LNCaP cells with identical mechanism and similar affinity.
Both aptamers were subjected to 3′-truncation in five-nucleotide increments to identify the minimal required binding elements. Truncation of xPSM-A9 past five nucleotides resulted in loss of PSMA inhibition, whereas xPSM-A10 could be truncated by up to 15 nucleotides while retaining PSMA binding ability. This truncated aptamer, named xPSM-A10-3, is approximately Mr 18,500 and retains the ability to competitively inhibit xPSM NAALADase activity with a calculated Ki of 20.5 nm (Fig. 3C) ⇓ . Additional truncation of xPSM-A10-3 on the 5′-end resulted in loss of the ability to bind xPSM.
Aptamer xPSM-A10-3 was fluorescently end-labeled with rhodamine-red-X (xPSM-A10-3-red) to evaluate if it could bind and label PSMA-expressing prostate cancer cells. Fluorescence microscopy showed that xPSM-A10-3-red specifically bound to the PSMA-expressing LNCaP prostate cancer cell line but not the PSMA-negative PC-3 prostate cancer cell line (Fig. 4) ⇓ . A randomly scrambled version of xPSM-A10-3 was unable to stain either LNCaP or PC-3, demonstrating the sequence specificity of the selected aptamer.
Whereas aptamers have been available for more than a decade, their applications are just emerging. Aptamers have been successfully applied in vitro to enzyme-linked oligonucleotide assays (11) , flow cytometry (12) , protein quantitation (13) , and protein purification techniques (14) . In vivo, aptamers can inhibit growth factor-receptor interactions (15) , block lymphocyte trafficking (16) , and inhibit enzyme activities (17) . More recently, RNA aptamers have been described as having ideal qualities for in vivo imaging (6) , exhibiting superior target:blood ratios when compared with antibodies, mainly because of more efficient blood clearance. Although brief, this history of applications suggests that aptamers may soon be an attractive alternative to antibodies for both in vitro and in vivo applications. Despite the success of SELEX, there are no reported RNA aptamers to membrane bound tumor antigens. Therefore, we explored creating nuclease-stabilized RNA aptamers to bind to and inhibit the enzymatic activity of a well-known prostate tumor cell surface antigen, PSMA.
PSMA is considered to be an excellent prostate tumor cell marker. PSMA expression is primarily prostate-specific, with very low levels seen in the brain, salivary glands, and small intestine (18) . PSMA expression is elevated in malignant prostate cells, with the highest expression in androgen-resistant cells because of loss of the negative regulation normally exerted by androgens (19) . Furthermore, PSMA is alternatively spliced, where normal prostate cells predominantly express a cytosolic form, PSM′, and malignant cells express the full-length membrane bound form (20) . The first anti-PSMA antibody, known to bind the intracellular portion of PSMA, has been modified into an imaging agent (21) and applied clinically to diagnose metastatic prostate tumors but with limited success. New antibodies to the extracellular portion of PSMA are currently being evaluated as potentially better imaging and therapeutic agents (22) .
To target malignant prostate cells via PSMA, a ligand should be designed to bind to the extracellular portion of the protein. We have designed and purified a fusion protein target representing the extracellular portion of PSMA, termed xPSM (Fig. 1) ⇓ . Two unique xPSM-binding aptamers were identified from an initial modified-RNA aptamer library of 6 × 1014 sequences (Fig. 2) ⇓ . The high affinity of these aptamers is demonstrated by their ability to inhibit PSMA NAALADase activity (Fig. 3) ⇓ . Whereas it is known that NAALADase activity is inhibited by phosphate, and that RNA is a polyphosphate, the demonstrated aptamer inhibition is specific by evidence that random RNA had no effect on xPSM NAALADase activity (Fig. 2B) ⇓ nor could random sequence RNA bind xPSM beads (Table 1) ⇓ or cells (Fig. 4) ⇓ .
One aptamer, truncated from 23.4 kDa to 18.5 kDa, specifically binds to and labels prostate cancer cells expressing PSMA (Fig. 4) ⇓ . Specific labeling of live whole cells with PSMA aptamers has been difficult as dying cells take up aptamer in a sequence independent manner, resulting in strong nuclear and nucleolar staining. Histological tissue sections were also difficult to stain with aptamers, as there is strong, sequence-independent nuclear staining, most likely because of positively charged nuclear proteins. Additional work is required to identify the ideal conditions for these in vitro diagnostic tests.
The work here describes the first application of SELEX to a membrane tumor antigen, the first reported aptamer inhibitors of a metallopeptidase, and the first prostate-cell specific aptamer. Additional modification of these high affinity aptamers could increase their protection against nuclease degradation and allow them to carry imaging agents to metastatic prostate foci. These PSMA aptamers will be additionally studied to evaluate their promise as prostate cancer diagnostic and therapeutic agents.
We thank Don Vindivich, Stuart Criley, and Dr. Miguel Garcia for their support throughout the project. Special thanks to the William Isaacs lab and Dennis Faith for their assistance in sequence analysis. We also thank Dr. Alan Meeker for critical reading of the manuscript. Finally, we thank the faculty and staff of the Brady Urologic Institute for their support of the research laboratories.
The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
↵1 Supported in part by NIH Grants CA15416, CA58236, and DK22000, and the Koch Foundation.
↵2 To whom requests for reprints should be addressed, at Department of Urology, Johns Hopkins University School of Medicine, Marburg 113, 600 North Wolfe Street, Baltimore, MD 21287-2101. Phone: (410) 614-4974; Fax: (410) 502-9336; E-mail:
↵3 Present address: SomaLogic, 1775 38th Street, Boulder, CO 80301.
↵4 The abbreviations used are: SELEX, systematic evolution of ligands by exponential enrichment; 2′FY-RNA, 2′-fluoro-pyrimidine-RNA; PSMA, prostate-specific membrane antigen; NAALADase, N-acetylated-α-linked acid dipeptidase; pfu, plaque-forming unit; rEK, recombinant enterokinase; NAAG, N-acetyl-aspartyl-glutamate.
- Received February 25, 2002.
- Accepted May 10, 2002.
- ©2002 American Association for Cancer Research.