Research Article | Open Access
Selection of DNA Aptamers for Ovarian Cancer Biomarker CA125 Using One-Pot SELEX and High-Throughput Sequencing
CA125 is a mucin glycoprotein whose concentration in serum correlates with a woman’s risk of developing ovarian cancer and also indicates response to therapy in diagnosed patients. Accurate detection of this large, complex protein in patient samples is of great clinical relevance. We suggest that powerful new diagnostic tools may be enabled by the development of nucleic acid aptamers with affinity for CA125. Here, we report on our use of One-Pot SELEX to isolate single-stranded DNA aptamers with affinity for CA125, followed by high-throughput sequencing of the selected oligonucleotides. This data-rich approach, combined with bioinformatics tools, enabled the entire selection process to be characterized. Using fluorescence anisotropy and affinity probe capillary electrophoresis, the binding affinities of four aptamer candidates were evaluated. Two aptamers, CA125_1 and CA125_12, both without primers, were found to bind to clinically relevant concentrations of the protein target. Binding was differently influenced by the presence of Mg2+ ions, being required for binding of CA125_1 and abrogating binding of CA125_12. In conclusion, One-Pot SELEX was found to be a promising selection method that yielded DNA aptamers to a clinically important protein target.
Ovarian cancer is diagnosed at local stage in only 15% of cases; five-year survival following this early diagnosis averages 92%. The majority of ovarian cancer patients are diagnosed with more advanced disease, for which the five-year survival rate is 46% . Given the dramatic improvement in long-term survival rates enabled by early detection of ovarian cancer, the development of analytical tools to aid in screening and early diagnosis is of great importance .
Cancer Antigen 125 (CA125) is the clinical gold standard biomarker for ovarian cancer . Serum levels of CA125 in ovarian cancer patients are regularly monitored, and a resurgence in CA125 correlates strongly with cancer recurrence [4–6]. Along with levels of the complementary biomarker HE4, serum levels of CA125 are considered in the Risk of Ovarian Malignancy Algorithm (ROMA), which is FDA-approved for the classification of women presenting with a pelvic mass into high- or low-risk categories [7, 8]. Another clinical use of CA125 is the Risk of Ovarian Cancer Algorithm (ROCA). ROCA predicts the probability of ovarian cancer based on serial measurements of serum CA125, exploiting the frequently observed phenomenon that ovarian cancer’s initial development causes a change-point in these levels over a woman’s individual baseline [9, 10]. The use of ROCA in ovarian cancer screening was assessed in the United Kingdom Collaborative Trial of Ovarian Cancer Screening (UKCTOCS), a randomized controlled trial that enrolled >200,000 postmenopausal women and followed them for 10 years . The investigators of UKCTOCS report a significant reduction in mortality in the multimodal screening group—who underwent ROCA and sonography—relative to the no-screening control group, when women who entered the study with cancer are excluded from analysis . These findings underscore the importance of excellent analytical assays for CA125 as part of the ongoing effort to refine existing ovarian cancer screening strategies.
CA125 is currently detected via double determinant immunoassay. Reliance on antibody-based detection is problematic for this biomarker, however, because of its heterogeneity. CA125 contains a variable number of repeat domains that are not identical and are recognized to differing extents by the most widely used classes of antibody . The inability of available affinity reagents to recognize all repeat domains of CA125 suggests that existing clinical assays may significantly underestimate levels of this biomarker . The desired clinical outcome of early detection may be enabled by the development of novel affinity probes that recognize CA125 in a manner different from existing antibodies, enabling more informative new detection methods.
As a first step towards the development of assays for ovarian cancer that do not rely on antibodies, we have selected single-stranded DNA aptamers with affinity for CA125. Aptamers are functional oligonucleotides selected through an in vitro evolution process to possess a binding or catalytic property of interest; the aptamer selection process is referred to as SELEX, an acronym for Systematic Evolution of Ligands by EXponential enrichment [14, 15]. SELEX typically occurs over multiple cycles, or rounds, each involving the incubation of candidate aptamer molecules with a target of interest; the partitioning of good binding oligonucleotides from those displaying weaker binding; and the amplification of good binders via polymerase chain reaction. At the conclusion of SELEX, the enriched oligonucleotide pool is subject to some form of sequence determination, typically via cloning into a bacterial expression system followed by Sanger sequencing.
Herein, we report on our use of One-Pot SELEX, in which selection and amplification of candidate oligonucleotides occur in a single container , to identify aptamers with affinity for CA125. The One-Pot approach minimizes transfer steps, mitigating both contamination and the loss of good binders. We conclude the aptamer selection process by sequencing DNA from each round on an Illumina platform. This data-rich approach enables the entire selection process, and not merely its endpoint, to be characterized. We have recently reported on a bioinformatics pipeline that enabled the identification of DNA aptamers with affinity for the ovarian cancer biomarker HE4 . This pipeline has now been used to mine another large dataset of sequence information to identify aptamers for another important clinical target.
2. Experimental Section
Oligonucleotides—including unselected DNA library, PCR, and sequencing primers, and labeled aptamers for in vitro testing—were purchased from Integrated DNA Technologies (Coralville, IA). The sequences, previously reported by Bowser and coworkers  were forward primer 5′-FAM-AGC AGC ACA GAG GTC AGA TG-3′, reverse primer 5′-biotin-TTC ACG GTA GCA CGC ATA GG-3′, ssDNA library 5′-FAM-AGC AGC ACA GAG GTC AGA TG (N)25 CCT ATG CGT GCT ACC GTG AA-3′. Hand-mixing was used in the synthesis of the DNA library’s random region to mitigate differential coupling efficiency of nucleobases that can bias syntheses using machine mixing. All oligonucleotides were reconstituted in TE buffer (10 mM Tris, 0.1 mM EDTA, pH 8.0) to 100 μM. Nuclease-free water, 25 mM MgCl2, 5.0 U/L Taq polymerase (for PCR), and Blue/Orange 6x loading dye (for gel loading) were purchased from Promega (Madison, WI). Deoxyribonucleotide triphosphates (dNTPs, 10 mM stock) were obtained from QIAGEN, Inc. (Valencia, CA). NuSieve GTG agarose was purchased from Cambrex BioScience (Rockland, ME). Selection and PCR amplification were performed in 0.5 mL thin-walled polypropylene tubes (Eppendorf North America, Hauppauge, NY). Sodium fluorescein was purchased from Sigma Aldrich (Saint Louis, MO). Tween-20 was purchased from Tokyo Chemical Industry (Portland, OR). CA125 from human ascites fluid (>95% purity) was purchased from Fitzgerald Industries International (Acton, MA) and supplied as a liquid in PBS buffer, pH 7.2, containing 3% sucrose and 0.05% sodium azide. Cancer antigen CA125 enzyme immunoassay test kit (ELISA kit) was purchased from BioCheck (Foster City, CA).
All buffers were prepared using 18.2 MΩ-cm water from a Milli-Q water purification system (Millipore Corp., Bedford, MA). Preweighed salts for preparing Tris-Glycine and Phosphate Buffered Saline were from Thermo Scientific (Rockford, IL). Protein incubating buffer contained 0.01 M NaPO3 and 0.15 M NaCl at pH 7.2. Protein washing buffer consisted of 0.1 M NaPO3, 0.15 M NaCl, and 0.05% v/v Tween-20 at pH 7.2 (PBT buffer). DNA binding buffer consisted of 25 mM Tris, 192 mM glycine, 5 mM KH2PO4, and 1 mM MgCl2 (TGKM buffer) adjusted to pH 8.3 with 1 M NaOH solution. CE separation buffer contained 25 mM Tris, 192 mM glycine, and 5 mM KH2PO4, pH 8.3 (TGK) prepared from Thermo Scientific Tris-Glycine powder and KH2PO4 from Mallinckrodt Chemical Works (St. Louis, MO). 10x single-stranding buffer consisted of 100 mM Tris, 20 M NaCl, and 10 mM EDTA (10x TNE buffer) at pH 7.5. Streptavidin agarose was purchased from Thermo Scientific Pierce Biotechnology Inc. (Rockford, IL). BioRad columns were purchased from BioRad Technologies (Hercules, CA). Binding and Washing (B&W) buffer (10 mM Tris, 2 mM NaCl, and 1 mM EDTA at pH 7.6, 2x concentration) was made from NaCl purchased from VWR (Bridgeport, NJ). All buffers were filtered through a 0.45 μm nylon filter prior to use.
2.2. In-Tube ELISA
A modified ELISA was performed to identify optimal conditions for immobilizing CA125 in PCR tubes. One tube (“Overnight”) was incubated with 100 μL of 8000 U/mL CA125 overnight at 4°C. A second tube (“SpeedVac”) contained 100 μL of 8000 U/mL CA125; solvent was evaporated on a Savant DNA 120 OP SpeedVac Concentrator (Thermo Scientific, Asheville, NC) for four hours with no heating. Tubes were washed 4x with 200 μL PBT and dried, after which the recommended protocol from the CA125 ELISA kit was followed; however, rather than using the CA125-coated microplate provided with the kit, the PCR tubes were used. Along with ELISA kit reagents, 30 μL of 8000 U/mL CA125 was added to the positive control tube. The negative control tube was incubated with PBS overnight.
2.3. One-Pot SELEX
100 μL of CA125 (8000 U/mL in 1x PBS) were added to five 0.5 mL PCR tubes. Three tubes were used for aptamer selection, one tube was used for PCR cycle determination, and one tube served as a negative control. Solvent was evaporated on a SpeedVac Concentrator for 4 hours with no heat. Tubes were then washed 4x with 200 μL PBT. 25 N DNA library in DNA binding buffer was heat cycled at 95°C for 3 minutes and slowly cooled on ice prior to selection. In the initial SELEX round, 50 μL of 1 μM DNA library (diluted using DNA binding buffer) was added to the tubes and incubated for 1 hour at room temperature. The volume of DNA solution was less than that of the CA125 solution to minimize nonspecific interaction of DNA and the tube surface; for the same reason care was taken to pipet DNA solution directly into the center of the tube. 50 μL of DNA binding buffer was added to negative control tube. DNA was removed from the tubes, and the tubes were washed with 200 μL PBT. The number of washes at this step provided a means of increasing the selective pressure; over the course of four selection rounds, 6, 9, 12, and 15 wash steps were used. Vacuum was used to completely dry the tubes. In subsequent selection rounds, the highest available ssDNA concentration was used (typically less than 1 M). DNA concentration was determined by absorbance at 260 nm on a NanoDrop 2000 UV-vis spectrophotometer.
2.4. PCR Amplification
A PCR mastermix (600 μL total volume) was prepared from 375 μL nuclease-free water, 12 μL 10 mM dNTPs, 9 μL 10 μM FAM-labeled primer, 9 μL 10 M biotinylated primer, 72 μL 25 mM MgCl2, 120 μL 5x colorless PCR buffer, and 3 μL 5 U/μL Taq polymerase. 100 μL of this mixture was aliquoted to the five 0.5 mL PCR tubes that underwent selection; these tubes were stored in the freezer while the optimal PCR cycle number was determined. 49 μL of the mixture was added to separate PCR tube with 1 μL of 100 nM library DNA as a PCR positive control. 49 μL of the mixture was added to another PCR tube with 1 μL of nuclease-free water as a PCR negative control. The cycle determination tube, One-Pot negative control tube, PCR positive control tube, and PCR negative control tube were placed in Mastercycler Personal (Eppendorf, Hamburg, Germany). All tubes were heated to 95°C for 10 minutes. 16 to 24 cycles of PCR were performed, each involving denaturation at 95°C for 30 seconds, primer annealing at 55°C for 15 seconds, and extension at 72°C for 15 seconds. Sample (10 μL) was removed from the thermocycler after each round, mixed with gel loading dye, and stored on ice until analyzed. Electrophoresis was performed in a 4% agarose gel to determine the optimal cycle for amplification and to check for contamination in the two negative controls (the FAM primer enabled visualizing of bands without the use of ethidium bromide staining). Three selection tubes were stored at −20°C during cycle determination. The contents of these tubes were amplified using the optimal cycle number, with the addition of a final extension at 72°C for 5 minutes. The contents of the three selection tubes were pooled (total volume = 300 μL).
2.5. Single-Stranding and Ethanol Precipitation
The dsDNA produced via PCR was converted to ssDNA using a streptavidin column to capture dsDNA through the biotinylated (undesired) strand, followed by incubation in alkaline solution to denature the two DNA strands. 300 μL streptavidin agarose was loaded into a BioRad column and washed 4x with 500 μL 2x single-stranding buffer. 300 μL of PCR product and 300 μL of 2x single-stranding buffer were added to the column and incubated for 30 minutes at room temperature with occasional vortexing. The column was allowed to drain completely and was washed 4x with 500 μL 1x single-stranding buffer, followed by a final wash with 500 μL nuclease-free water. 200 μL of ~0.1 M NaOH was added to the column, and the column was incubated for 10 minutes at 37°C. Column contents, containing the fluorescein-labeled ssDNA of interest, were eluted into a 0.5 mL Eppendorf tube containing acetate buffer made from 116 μL 0.27 M acetic acid and 35 μL 3 M sodium acetate. The elution process was repeated with a second portion of NaOH solution; the products were collected into a second Eppendorf tube. 1000 μL of 100% ethanol was added to the tubes, which were incubated in an ice bath for 1 hour or stored in a −20°C freezer overnight to precipitate ssDNA. Tubes were centrifuged at 14,000 rpm, 4°C, for 1 hour to pellet ssDNA. All but 50 μL of solvent was aspirated from above the pellet, which was not always visible. The pellet was resuspended in 70% ethanol and tubes were again centrifuged for 15 minutes at 14,000 rpm and 4°C. This step was repeated, after which the DNA was dried on a SpeedVac for 15 minutes under high heat. 30 μL water or DNA binding buffer was added to the dried tubes to reconstitute the DNA for the next round of selection or for sequencing. Samples were stored at −20°C until further use.
2.6. Sequencing and Bioinformatics
The process of sequencing and bioinformatics has been previously reported . Briefly, after aptamer selection was complete, DNA collected from each round was amplified using Illumina sequencing primers. Archived DNA from each round was diluted to 100 nM using ultrapure water. Each sample was assigned a unique reverse primer containing the index used for barcoding. Samples and controls were amplified by PCR using an optimized cycle number. PCR products were imaged on a 3% agarose gel containing 1 μg/mL ethidium bromide to confirm yield and the absence of contaminants or byproducts. Samples were sequenced at the University of Wisconsin Biotechnology Center DNA Sequencing Facility.
Bioinformatic screening of the sequenced DNA used a data pipeline based on freely available software, with the exception of enrichment analysis, which used a locally written Python program. This program (enrichment.py) has been made available on GitHub at https://github.com/rebeccawhelan/PythonEnrichment. After preliminary analysis of FastQC files to ensure the sequencing was successful, data were read into a Biopieces pipeline using read_fastqc. Each line of sequence data, corresponding to one aptamer candidate, was then modified by the removal of the (conserved) primer regions; sequences with length other than were discarded and sequences were counted and ranked based on abundance. Next, the processed data were taken through enrichment analysis, to determine fold enrichment for sequences across rounds of selection. Fold enrichment has been shown to be a more reliable indicator of binding affinity than read counts . A composite score (compScore) was then calculated usingThe compScore is therefore the log10 of the product of each round’s enrichment, with double weight given to the final round, in which the selective pressure was the most stringent. Using CD-HIT-EST , the top 1000 most enriched sequences from each round were clustered by sequence homology to determine possible emergent motifs. Sequences were clustered with their primers attached to a sequence identity threshold of 0.8 and assigned to clusters by the highest identity across all clusters.
2.7. Fluorescence Anisotropy
Fluorescence anisotropy was measured using a SpectraMax M5 multimode plate reader with polarizing optics (Molecular Devices, Sunnyvale, CA). Tested aptamers were ordered from Integrated DNA Technologies, Inc. (Coralville, IA) with a 5′ TEX615 (Texas Red) fluorophore. 100 nM DNA aptamer in buffer (TGKM) was heated to 97°C for 3 minutes, cooled on ice for 5 minutes, and then allowed to warm to room temperature. Heat-cycled DNA solution was combined with CA125 at a range of final concentrations from 0 U/mL to 2000 U/mL in the presence of 0.13 mg/mL bovine serum albumin (BSA) to prevent adsorption to the container walls. PBS buffer was added to achieve the desired final sample volume. After incubating for 90 minutes in the dark at room temperature, samples were loaded in duplicate or quadruplicate (75 μL/well) into a 96-well Fluotrac 200 black immunology plate (USA Scientific, Ocala, FL) and analyzed in the SpectraMax, with temperature held at 22°C. The for fluorescence anisotropy was 585 nm, was 635 nm, and the wavelength cut-off was 610 nm. Raw data (fluorescence emission parallel and perpendicular to the excitation) were blank-corrected before the anisotropy values were calculated. Data were fit with an isotherm function:where is protein concentration (varied), is aptamer concentration (constant), is the anisotropy measured in the presence of protein, and is the anisotropy in the absence of protein, using IgorPro (v. 6.12) graphing software. To test the prediction that magnesium ion affected the affinity of aptamers for CA125, the assay described above was repeated, but with PBS used instead of TGKM at the heat-cycling step; in these assays samples were incubated for 24 hours at 4°C.
2.8. Affinity Probe Capillary Electrophoresis
Affinity probe capillary electrophoresis affinity assays were performed using a Beckman P/ACE MDQ (Beckman Coulter, Fullerton, CA) equipped with an argon-ion laser. An unmodified fused silica capillary (Polymicro Technologies, Phoenix, AZ; ID = 50 μm, OD = 360 μm, total length = 49.5 cm, and length from inlet to detector = 10.5 cm using reverse injection and negative polarity) was held at 25°C. Each sample contained 20 nM FAM-labeled aptamer (synthesized as a 25 mer sequence without primer regions), 20 nM fluorescein (internal standard), and 0.2 mg/mL BSA. TGKM was used both as the diluent in sample preparation and as the electrophoresis buffer for experiments with CA125_1_NP; TGK was used as the electrophoresis buffer for CA125_12_NP. To prepare samples, a bulk solution of aptamer in either TE or TGKM was heated to 97°C for 3 min and then cooled on ice. Fluorescein and BSA were then added, and the solution was distributed over an appropriate number of sample tubes. Finally, CA125 was added to a final concentration ranging from 0 U/mL to 2000 U/mL. The volume of protein plus protein buffer (PBS) was constant in all samples (4 μL). Samples were incubated in the dark for 90 min. Pressure injection (0.3 psi, 5 s) was used to introduce sample onto the capillary; 0.5 psi pressure was applied to drive the sample plug past the uncooled region . Separation was achieved by the application (in negative polarity) of 30 kV and 0.3 psi pressure for CA125_1_NP. Separation for CA125_12_NP was conducted at 20 kV with no applied pressure. Run time was 5 min. The fluorescence was excited at 488 nm and detected at 520 nm. Peak heights were determined by the instrument control software (32 Karat). The change in the size of the free DNA aptamer peak, relative to the internal standard, was used to indicate the complex formation between aptamer and protein. Data were fit with an isotherm equation:where is protein concentration (varied) and is aptamer concentration (constant), using IgorPro (v. 6.12) graphing software.
3. Results and Discussion
A single-stranded DNA library with random region was used as the input to the selection process because it provided a good balance of sequence diversity, coverage, and computational tractability. Assuming that each base is equally likely to appear at each position in the random region, there are 425 (~1 × 1015) possible sequences in such a library. In our selection, we used 50 pmol (~3 × 1013 molecules) of DNA as the initial input, giving any individual sequence an expected abundance of 0.03 (a library with would give an expected abundance of 1). Using a longer random region would result in lower coverage of sequence space that could result in the loss of useful motifs, whereas a shorter random region might lack the complexity to form relevant secondary and tertiary structures involved in target binding.
Figure 1 shows the results of an in-tube ELISA used to evaluate different methods for adsorbing CA125 within PCR tubes. Based on a qualitative assessment of color change, both overnight incubation at 4°C and 4 hr of solvent evaporation on a SpeedVac resulted in the adsorption of immunologically active CA125 to the walls of PCR tubes. SpeedVac treatment was observed to cause greater color change for the same input concentration of CA125 and assay reagents and took less time to be completed, so this approach was used in One-Pot SELEX.
One-Pot SELEX is represented schematically in Figure 2. In One-Pot SELEX, selection and amplification occur in the same PCR tube. This approach distinguishes One-Pot from SELEX modes in which other means of separation (nitrocellulose filtration, column chromatography, capillary, or chip electrophoresis) enable the partitioning of high-affinity binding aptamer candidates from those candidates that do not bind target. After incubation of the aptamer candidates with the CA125 adsorbed in the tube, the tube was rinsed with buffer to remove weakly bound and nonspecifically bound DNA. Iteratively increasing the number of wash steps (from 6 to 15 over the course of four selection rounds) applied greater selective pressure and forced the population of DNA to converge on higher affinity binders.
Our group and others have documented that PCR amplification of the random DNA libraries used in aptamer selection is subject to undesirable side-reactions [2, 21]. Specifically, with increasing cycles of PCR, the dsDNA product of interest initially increases in abundance but is consumed in a side reaction yielding byproducts that migrate more slowly under electrophoresis in agarose gels. To minimize the negative impact of byproduct formation, we optimized the number of PCR cycles after each SELEX round. One reaction tube was subject to PCR, and sample was removed from this tube at regular intervals. Samples were electrophoresed together on a 4% agarose gel and imaged on a gel box with ethidium bromide filters. Figure 3 shows a typical gel image. In this experiment, 21 cycles of PCR were found to give the highest yield of the desired product with minimal byproduct formation. The image also shows expected outcomes for the positive and negative PCR controls and for the One-Pot control, in which CA125 was immobilized in the tube but DNA-free buffer was added during the incubation step.
The bioinformatics pipeline used after the completion of One-Pot SELEX is represented in Figure 4. This pipeline is an improved version of a similar computational process that we developed to identify DNA aptamers with affinity for ovarian cancer biomarker HE4 in a previous study . DNA archived after each selection round was sequenced on an Illumina platform. This approach yields more reads (max = 10.8 × 106, min = 4.1 × 106, average = 6.6 × 106) than would be achievable with a clone-and-sequence method, which typically yields 10–100 sequences. Table 1 contains information about the resulting sequence data. We note two attributes in this table. First, the—ostensibly random—unselected library contains duplicate sequences. We previously reported such sequence duplication in unselected libraries  and attribute it to errors in synthesis and sequencing. In addition, the percent duplicated sequences increase during selection, from 15.24% in the unselected library to 85.04% in DNA selected in Round 4. Sequence duplication increases for both CA125-binding DNA (designated by + in Table 1) and unbound DNA (designated by –). However, the rate of increase is greater in the population of CA125 binders. We interpret these data to mean that the selective pressures have substantially altered the ssDNA population, and that the effect occurred first in the population of DNA that bound to the protein target. Considering that the input for each selection round was the binding DNA from the previous round, it is reasonable that sequence duplication in the positive and negative selection rounds should proceed as observed.
The data reveal synthesis bias in the unselected DNA library. Ideally, the “random” region of the unselected library will contain an equal proportion of the four nucleobases; in our library the random region contained % G; % A; % T; % C. Although this library—prepared with hand-mixing—does not adhere to the ideal base ratio, it is less biased than other libraries we have examined that were prepared by machine mixing (data not shown). Hand-mixing is therefore advisable in the preparation of libraries for aptamer selection.
Four promising aptamer sequences were chosen for testing in vitro; these sequences are shown in Table 2. Candidates were chosen that were substantially enriched during the selection process, belonged to secondary structure clusters, or both. Sequence CA125_1 was the most enriched sequence by the final round of selection (Round 4). Sequence CA125_2 has a strong composite score and a steady increase over all selection rounds. Unlike the other sequences we tested, CA125_2 is 26 bases long; our bioinformatics analysis was designed to include all sequences of length . This 26 mer sequence is most likely an artifact of PCR or synthesis error, but we chose to characterize its affinity for CA125 because of its high degree of enrichment. Sequence CA125_3 is representative of the sequence cluster that seems dominant in the most highly enriched selection round. Finally, sequence CA125_12 is representative of the highest-scoring cluster.
Two independent analytical methods—fluorescence anisotropy (FA) and affinity probe capillary electrophoresis (APCE)—were used to evaluate the binding affinity of these four aptamer candidates for CA125. When the (or , in the case of CA125_3) random regions of the aptamer candidates were assayed, two of them—CA125_1 and CA125_12—displayed concentration-dependent binding to CA125. Figure 5 shows a binding isotherm for CA125_1, resulting from monitoring FA change as CA125 concentration was increased from 1 to 2000 U/mL. On the -axis is plotted the change in anisotropy relative to the free aptamer in the absence of protein, an indication of complex formation; CA125 concentration is the independent variable. The data are well fit by the binding isotherm equation, yielding an apparent of U/mL. Binding required that CA125_1 be heat cycled in the presence of 5 mM magnesium ion prior to incubation with protein target; CA125_1 heat cycled in buffer depleted of magnesium ion did not display concentration-dependent binding to CA125. Data from FA assays are supported by APCE assays, which yield an apparent of 80. ± 38 U/mL. Aptamer candidate CA125_12 was found to display concentration-dependent binding to CA125 as well, with values of U/mL and U/mL determined by FA and APCE, respectively. In contrast with CA125_1, which required heat cycling in the presence of magnesium ion for binding, CA125_12 bound to CA125 only when heat cycled in PBS or TE without magnesium ion; heat cycling in the presence of magnesium ion substantially reduced the binding affinity of CA125_12. The variation in values reported here is within the range observed by our lab and others for DNA aptamer-protein pairs characterized by FA and APCE [2, 17]. We chose to use a simplified form of the isotherm equation because the generalized form requires the addition of terms with dissimilar units: U/mL for CA125 concentration and nM for aptamer concentration .
The units in which we have chosen to report values—U/mL—merit discussion. Because of the long-documented heterogeneity of CA125 , this protein does not have a well-defined value for molar mass. For clinical and research applications, concentrations of CA125 are always reported in U/mL, a unit that derives from the immunoassay that is the standard for clinical detection. Although this unit is not familiar to chemists and biologists, reporting CA125 in this manner is standard practice for ovarian cancer researchers and clinicians. The current clinical cut-off for CA125 is 35 U/mL, and commercially available research grade immunoassay kits typically have a dynamic range up to 200 U/mL. Our DNA aptamers therefore have an apparent affinity at a clinically relevant concentration.
The other two aptamer candidates (CA125_2 and CA125_3) did not display binding to CA125 when examined as random region alone. The reason for the absence of positive binding as indicated by these two in vitro assays is not clear. We note that the two successful aptamer candidates (CA125_1 and CA125_12) were, respectively, the most enriched in response to selective pressure and representative of the highest-scoring cluster. The other candidates had less favorable enrichment metrics, as determined by bioinformatics analysis. In addition, we note that whereas the aptamer candidates that displayed target binding contained all of the four nucleobases, the two unsuccessful aptamer candidates tested in vitro contain no G and have a percentage of T higher than the expected percentage (25%). Table 3 shows the base composition of the four aptamer candidates we tested in vitro. Those that did not display binding (CA125_2 and CA125_3) contain 15% and 12%, respectively, of G + C whereas those that displayed binding (CA125_1 and CA125_12) both contained 40% G + C (random distribution would predict that each sequence would contain 50% G + C). Given the stronger intramolecular forces associated with G-C base-pairing over A-T base-pairing, this difference in nucleobase content may be significant in the observed binding to the protein target. Finally, assays involving aptamer candidates bearing the PCR primer-binding regions did not show evidence of concentration-dependent complex formation as the concentration of CA125 was increased (data not shown).
In its original applications, One-Pot SELEX successfully identified aptamers with affinity for antibodies [16, 23]. In this report we have shown its use for selection of aptamers recognizing CA125, a mucin protein that is the gold standard biomarker for ovarian cancer. Given the widespread importance of mucins in many types of cancer [24, 25], this relatively simple and efficient process might enable the selection of clinically useful aptamers, particularly for cancer biomarkers and molecules involved in cancer proliferation, metastasis, and immune suppression. Recently a pair of 2′-fluoro-pyrimidine RNA aptamers with affinity for CA125 were reported . These modified RNA aptamers were selected using a magnetic bead purification strategy on His-tagged CA125, with His-tagged VEGF used for counter-selection. Our future work will involve a comparison of the DNA aptamers reported here with these new RNA aptamers, and the evaluation of their possible synergistic use in the creation of novel assays for the detection of CA125. The availability of multiple aptamers with affinity for this clinically important target opens the possibility of developing a sandwich-format assay, with its inherent advantages of specificity and amplification.
The content is solely the responsibility of the authors and does not necessarily represent the official views of the National Cancer Institute or the National Institutes of Health.
The authors declare that they have no competing interests.
Delia J. Scoville and Tae Kyu Brian Uhm contributed equally.
The project described was supported by the National Cancer Institute (Grant no. R15CA161970). Financial support from a Henry Dreyfus Teacher Scholar Award (to Rebecca J. Whelan) is gratefully acknowledged. The authors wish to thank Rachel Eaton and Kepler Mears for technical support and stimulating discussions.
- American Cancer Society, Cancer Facts & Figures 2016, American Cancer Society, Atlanta, Ga, USA, 2016.
- R. M. Eaton, J. A. Shallcross, L. E. Mael et al., “Selection of DNA aptamers for ovarian cancer biomarker HE4 using CE-SELEX and high-throughput sequencing,” Analytical and Bioanalytical Chemistry, vol. 407, no. 23, pp. 6965–6973, 2015.
- M. Felder, A. Kapur, J. Gonzalez-Bosquet et al., “MUC16 (CA125): tumor biomarker to cancer therapy, a work in progress,” Molecular Cancer, vol. 13, article 129, 2014.
- M. E. L. Van Der Burg, F. B. Lammes, and J. Verweij, “The role of CA 125 in the early diagnosis of progressive disease in ovarian cancer,” Annals of Oncology, vol. 1, no. 4, pp. 301–302, 1990.
- M. E. L. van der Burg, F. B. Lammes, and J. Verweij, “CA 125 in ovarian cancer,” Netherlands Journal of Medicine, vol. 40, no. 1-2, pp. 36–51, 1992.
- M. E. van der Burg, F. B. Lammes, and J. Verweij, “The role of CA 125 and conventional examinations in diagnosing progressive carcinoma of the ovary,” Surgery, Gynecology & Obstetrics, vol. 176, no. 4, pp. 310–314, 1993.
- R. G. Moore, A. K. Brown, M. C. Miller et al., “The use of multiple novel tumor biomarkers for the detection of ovarian carcinoma in patients with a pelvic mass,” Gynecologic Oncology, vol. 108, no. 2, pp. 402–408, 2008.
- R. G. Moore, D. S. McMeekin, A. K. Brown et al., “A novel multiple marker bioassay utilizing HE4 and CA125 for the prediction of ovarian cancer in patients with a pelvic mass,” Gynecologic Oncology, vol. 112, no. 1, pp. 40–46, 2009.
- S. J. Skates, F.-J. Xu, Y.-H. Yu et al., “Toward an optimal algorithm for ovarian cancer screening with longitudinal tumor markers,” Cancer, vol. 76, no. 10, pp. 2004–2010, 1995.
- S. J. Skates, D. K. Pauler, and I. J. Jacobs, “Screening based on the risk of cancer calculation from Bayesian hierarchical changepoint and mixture models of longitudinal markers,” Journal of the American Statistical Association, vol. 96, no. 454, pp. 429–439, 2001.
- U. Menon, A. Gentry-Maharaj, A. Ryan et al., “Recruitment to multicentre trials—lessons from UKCTOCS: Descriptive Study,” BMJ, vol. 337, no. 7681, pp. 1283–1286, 2008.
- I. J. Jacobs, U. Menon, A. Ryan et al., “Ovarian cancer screening and mortality in the uk collaborative trial of ovarian cancer screening (UKCTOCS): a randomised controlled trial,” The Lancet, vol. 387, no. 10022, pp. 945–956, 2016.
- A. Bressan, F. Bozzo, C. A. Maggi, and M. Binaschi, “OC125, M11 and OV197 epitopes are not uniformly distributed in the tandem-repeat region of CA125 and require the entire SEA domain,” Disease Markers, vol. 34, no. 4, pp. 257–267, 2013.
- A. D. Ellington and J. W. Szostak, “In vitro selection of RNA molecules that bind specific ligands,” Nature, vol. 346, no. 6287, pp. 818–822, 1990.
- C. Tuerk and L. Gold, “Systematic evolution of ligands by exponential enrichment: RNA ligands to bacteriophage T4 DNA polymerase,” Science, vol. 249, no. 4968, pp. 505–510, 1990.
- S. Missailidis, D. Thomaidou, K. E. Borbas, and M. R. Price, “Selection of aptamers with high affinity and high specificity against C595, an anti-MUC1 IgG3 monoclonal antibody, for antibody targeting,” Journal of Immunological Methods, vol. 296, no. 1-2, pp. 45–62, 2005.
- M. Jing and M. T. Bowser, “Tracking the emergence of high affinity aptamers for rhVEGF165 during capillary electrophoresis-systematic evolution of ligands by exponential enrichment using high throughput sequencing,” Analytical Chemistry, vol. 85, no. 22, pp. 10761–10770, 2013.
- W. H. Thiel, T. Bair, A. S. Peek et al., “Rapid identification of cell-specific, internalizing RNA aptamers with bioinformatics analyses of a cell-based aptamer selection,” PLoS ONE, vol. 7, no. 9, Article ID e43836, 2012.
- W. Li and A. Godzik, “Cd-hit: a fast program for clustering and comparing large sets of protein or nucleotide sequences,” Bioinformatics, vol. 22, no. 13, pp. 1658–1659, 2006.
- M. U. Musheev, Y. Filiptsev, and S. N. Krylov, “Noncooled capillary inlet: a source of systematic errors in capillary-electrophoresis-based affinity analyses,” Analytical Chemistry, vol. 82, no. 20, pp. 8637–8641, 2010.
- M. U. Musheev and S. N. Krylov, “Selection of aptamers by systematic evolution of ligands by exponential enrichment: addressing the polymerase chain reaction issue,” Analytica Chimica Acta, vol. 564, no. 1, pp. 91–96, 2006.
- M. Jing and M. T. Bowser, “Methods for measuring aptamer-protein equilibria: a review,” Analytica Chimica Acta, vol. 686, no. 1-2, pp. 9–18, 2011.
- S. Missailidis, “Targeting of antibodies using aptamers,” in Methods in Molecular Biology, Vol. 248: Antibody Engineering: Methods and Protocols, B. K. C. Lo, Ed., vol. 248, pp. 547–555, Humana Press, Totowa, NJ, USA, 2003.
- M. A. Hollingsworth and B. J. Swanson, “Mucins in cancer: protection and control of the cell surface,” Nature Reviews Cancer, vol. 4, no. 1, pp. 45–60, 2004.
- D. W. Kufe, “Mucins in cancer: function, prognosis and therapy,” Nature Reviews Cancer, vol. 9, no. 12, pp. 874–885, 2009.
- I. Lamberti, S. Scarano, C. L. Esposito et al., “In vitro selection of RNA aptamers against CA125 tumor marker in ovarian cancer and its study by optical biosensing,” Methods, vol. 97, pp. 58–68, 2016.
Copyright © 2017 Delia J. Scoville et al. This is an open access article distributed under the Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.