Novel Approach to Overcome Defects of Cell-SELEX in Developing Aptamers against Aspartate β-Hydroxylase.

Cell-based aptamer selection (Cell-SELEX) against predefined protein targets that benefits using the native form of the targets is the most promising approach to achieve aptamer probes capable of recognizing targets under both in vitro and in vivo conditions. The major disadvantages in Cell-SELEX are the imperfectness of the negative selection step and the lengthy procedure of selection. Here, we introduced the Counter-SELEX as part of our modified Cell-SELEX and implemented deep sequencing to overcome these shortcomings in developing aptamers against aspartate β-hydroxylase (ASPH) as a known tumor marker. In parallel with the conventional Cell-SELEX, five consecutive cycles of counter selection were accomplished using sequences bound to negative cells (the Counter-SELEX) to detect oligos that are not specific for ASPH. After high-throughput sequencing, the representative of each promising achieved family was subjected to further confirmatory analysis via flow cytometry, followed by the fluorescence immunostaining of histopathological sections. Implementing our innovative complementary method, annoying mis-selected sequences in Cell-SELEX enriched pools were effectively identified and removed. According to the affinity assay on the cells displaying ASPH, three aptamers, AP-Cell 1, AP-Cell 2, and AP-Cell 3, with K d values of 47.51, 39.38, and 65.23 nM, respectively, were obtained, while AP-Cell 1 and 3 could then successfully spot ASPH displayed on the tissues. Our study showed that the Counter-SELEX could be considered as a complementary method for Cell-SELEX to overcome the imperfectness of the negative selection step. Moreover, high-throughput nucleotide sequencing could help to shorten the overall process.

Introduction

Molecular recognition of tumor-specific markers like proteins is a critical issue in both the fields of cancer detection and therapy.[1−3] An emerging class of biological probes, named aptamer, is overthrowing antibodies that are classically applied in this context.[4,5] Aptamers are typically developed via the iterative cycles of selection referred to as systematic evolution of ligands by exponential enrichment (SELEX) through enriching a random library of short single-stranded nucleic acids, which can specifically bind to individual targets, ranging from small molecules to large proteins, as affinity probes.[6−8] Several innovative SELEX procedures have been proposed in different studies using which aptamers are systematically developed under different binding conditions; each one has its own specific advantages and disadvantages. Therefore, choosing a proper procedure for clinical and histopathological applications is not simply feasible.[9,10] The clinical value of using aptamers is faced with serious problems as many sequences obtained from the conventional methods of aptamer selection show nonspecific or poor detection performance in clinicopathological studies. Among various approaches, cell-based aptamer selection (Cell-SELEX), which benefits using the native form of the protein targets, is the most promising approach to achieve appropriate probes capable of recognizing under both in vitro and in vivo conditions.[11−13]

Cell-based aptamer selection (Cell-SELEX) is a systematic aptamer evolution process that proceeds with selection on a wide range of targets including intact complexes or a predefined target protein overexpressed and displayed on the surface of a specific living cell.[14−17] Utilizing Cell-SELEX to generate aptamers against predefined target proteins on the cellular membrane of cells is obviously superlative to targeting purified proteins because of two reasons. First, there is no need to go through the laborious process of protein purification. Second, the proteins displayed on the surface of living eukaryotic systems remain in their native conformation with the proper post-translational modifications.[1,18] Developing aptamers against native targets would enhance the chances of success in detecting histopathological targets. However, there are some technical difficulties with the procedure of Cell-SELEX such as the interference of dead cells during selection, lengthy experimental procedure, and the imperfectness of the counter selection.[14]

In this study, our aptamer target is human aspartate β-hydroxylase (ASPH), a cell surface α-ketoglutarate-dependent dioxygenase, which has recently been considered as a tumor biomarker for detecting different types of cancer, predicting prognosis, and observing the response to treatment.[19,20] Upregulated ASPH expression is associated with a broad range of human malignancies tested to date, such as breast, colon, ovary, prostate, lung, liver, and bile duct tumors.[20−23] Today, there are some chemiluminescence immunoassay (CLIA)-compliant diagnostic tests based on ASPH detection utilizing enzyme-linked immunosorbent assay (ELISA) to diagnose or monitor a series of cancers.[24] ASPH is also known as the target of immunohistochemical staining in some tumors including hepatocellular carcinoma, cholangiocarcinomas, and lung carcinoma.[21,25−28] Moreover, ASPH has recently attracted more attention as a potent target for tumor therapeutic applications.[29,30]

In this paper, we report on our use of modified Cell-SELEX (the combination of conventional Cell-SELEX with a novel Counter-SELEX) in developing single-stranded DNA aptamers against human ASPH as a diagnostic target. This allows us to overcome some drawbacks of the Cell-SELEX strategy while benefitting from its practical advantage of using native ASPH. The enriched pool was eventually subjected to next-generation sequencing (NGS) to gain a more comprehensive view and to better identify the most frequent sequences in the pool.

Results

Cell-SELEX

Stable ASPH expressing HeLa cell lines (HeLaASPH) was established as demonstrated via a significant increment of mRNA level (quantitative reverse transcription polymerase chain reaction (RT-PCR)) and the obvious detection of surface-displayed ASPH (flow cytometry) as depicted in Figures S1 and1. In the process of cell-based aptamer selection, the obtained oligomers of each round of selection were amplified using PCR with an appropriate number of cycles. The optimal number of cycles in the PCR procedure related to each round of selection was determined using agarose electrophoresis (Figure S2). Excessive cycles of PCR may lead to byproduct formation, while inadequate numbers result in insufficient production of the desired dsDNA. Cell-based aptamer selection was then successfully applied, and the progression of enrichment was monitored using the flow cytometry assay as depicted in Figure ; however, due to the implementation of NGS, there was no need to reach the highest fluorescence intensity (the endpoint of selection). The 15 top oligomers obtained from deep sequencing of the last round of selection are presented in Table .

Figure 1

Verification of ASPH expression on the surface of HeLaASPH. Flow cytometric analysis of the ASPH expression on the surface of cells using the anti-ASPH antibody showed that the transfected HeLa cells (HeLaASPH) display ASPH on their surface, while the untransfected ones lack ASPH on their surface; mean fluorescence intensity (MFI).

Figure 2

Flow cytometry assay of the enriched pools of various rounds of Cell-SELEX. Evaluation of selection progress in consecutive rounds of Cell-SELEX is depicted using the flow cytometry assay. Gradually enhanced fluorescence intensity from rounds 3–9 indicated the proper trend of aptamer evolution during the process: (a) transfected HeLa cells (HeLaASPH) and (b) untransfected HeLa cells; mean fluorescence intensity (MFI).

Table 1

First 15 Sequences from Deep Sequencing of Cell-SELEX Last Round of Selection

library: ATACCAGCTTATTCAATT-N52-AGATAGTAAGTGCAATCT
cell-SELEX oligomer name	sequence	count per million	approval statusa
#AP-Cell 1	-CGGGACAAGACAACGAACGAACAGGAAGAGAACCGGAATGCAGACGTCAGGG-	25 931	approved
#AP-Cell 2	-CCGGAGAAAAAAAACTTGATGAGCCAGGAAAACGAGGGTTGGAACGCGACTG-	23 541	approved
#AP-Cell 3	-CCGTATCGCCCAGGCAACTGGGCTAAACTTCCCAGAGGGAACGAAACCTGGG-	20 304	approved
#AP-Cell 4	-CGAGACCACCAACCGAACCAAACCACACGAGTCAAAACCCACGGAACGAACG-	16 601	rejected
#AP-Cell 5	-AGGAACGCGAAAAAAAGGACTGAATACTCGGGGAAGAAAGATGGCCGGAACG-	11 445	approved
#AP-Cell 6	-GGGACAAGACAACGAACGAACAGGAAGAGAACCGGAATGCAGACGTCAGGGG-	8416	rejected
#AP-Cell 7	-CCGTATCGCCCAGGCAACTGGGCTAAACTTCCCAGAGGGAACGAAGGTTGGG-	8371	approved
#AP-Cell 8	-CGAACAAGGAACGAAAACCGAACGGGAAAGGACCGCGACAAGAGATTACGGG-	7491	rejected
#AP-Cell 9	-CGACAAAAAAACGAACAACCAGAACCGGAAGATAGAGGAAAGAGCGCGACGG-	6525	rejected
#AP-Cell 10	-CCGTATCGCCCAGGCAACTGGGCTAAACTTCCCAGAGGGAACGAAACTTGGG-	5994	approved
#AP-Cell 11	-CGGGACAAGACAACGAACGAACAGGAAGAGAACCGGAATGCATGCAGCAGGG-	5485	approved
#AP-Cell 12	-CCGGGAAAACAGGAACGACAAGAGAAAGACCAACACAGACGACAAAGTGGGA-	5020	rejected
#AP-Cell 13	-CGGGACAAGACAACGAACGAACAGGAAGAGAACCGGAATGCAGAGACTGCGG-	4728	approved
#AP-Cell 14	-CGGACGGAACAACAAACGGGACTGGAACAGGAAGAAAAAGACGGAGGACGAC-	4283	rejected
#AP-Cell 15	-CGGAGAAAAAAAACTTGATGAGCCAGGAAAAACGAGGGTTGGAACGCGACTG-	4274	approved

Approval status indicates which sequences from Cell-SELEX are acceptable according to the related complementary methods, the Counter-SELEX.

Counter-SELEX

To recognize and then eliminate the oligomers undesirably binding to the frequent unwanted surface molecules in Cell-SELEX, we effectively established a novel negative selection procedure named Counter-SELEX. The top 20 sequences obtained from the deep sequencing of the Counter-SELEX are listed in Table , including those presented in Cell-SELEX NGS data. Of the first 15 sequences obtained from Cell-SELEX, 6 sequences were identically present in the Counter-SELEX list (Table ), and therefore, they could be recognized as unwanted oligomers.

Table 2

First 20 Sequences from Deep Sequencing of the Counter-SELEX

library: ATACCAGCTTATTCAATT-N52-AGATAGTAAGTGCAATCT
cell-SELEX oligomer name	sequence	count per million	cell-SELEX ranka
#AP-Cntr 1	-CGAGACCACCAACCGAACCAAACCACACGAGTCAAAACCCACGGAACGAACG-	6856	4
#AP-Cntr 2	-GGGACAAGACAACGAACGAACAGGAAGAGAACCGGAATGCAGACGTCAGGGG-	6768	6
#AP-Cntr 3	-CGACAAAAAAACGAACAACCAGAACCGGAAGATAGAGGAAAGAGCGCGACGG-	4672	9
#AP-Cntr 4	-CGAACAAGGAACGAAAACCGAACGGGAAAGGACCGCGACAAGAGATTACGGG-	3013	8
#AP-Cntr 5	-CCTCCTGAACACATTACATTGTATAAACTGAGCTGAATTGTATCATGAAGGG-	2805	NDb
#AP-Cntr 6	-CGGACGGAACAACAAACGGGACTGGAACAGGAAGAAAAAGACGGAGGACGAC-	2772	14
#AP-Cntr 7	-CCGGGAAAACAGGAACGACAAGAGAAAGACCAACACAGACGACAAAGTGGGA-	2467	12
#AP-Cntr 8	-GGAATTCATCTTGTTCACGGGGAAGTTGATAATGGGAGGTGTGTTAAACAGG-	2175	NDb
#AP-Cntr 9	-CGAACAGGAACAACAGAAAGACACCCAGAGGAACCGGAGGAAGACGCACTGA-	1998	16
#AP-Cntr 10	-CACGCCAAGAAGGGGAACACGGTTCGCACGGAGGAAGACTAAAAGAATATCG-	1827	NDb
#AP-Cntr 11	-CGAAACGCACGGAACAAACAGACAAAGAACCAACAGAAACGGAGTAAACGGG-	1672	17
#AP-Cntr 12	-CGAAACAGCCCGGAACAGATTCAATTCGAGGCACGCCAGCAGGAAACGACAC-	1576	NDb
#AP-Cntr 13	-GGGACAATAAACGGGTACTGGACAAAACGAAAGGGACAGACGAAGGAAGCGG-	1432	24
#AP-Cntr 14	-CCGAATTCAGGAAAGGGTTTAAATCGAAAGCCCGCATCGAAAGATTAACGAC-	1395	NDb
#AP-Cntr 15	-CGAAAACACACCCGGAAAAACGAACGGAAGCGGGAATACAACAAACCGGAAT-	1253	NDb
#AP-Cntr 16	-CACACGACCCAATAAAAGGAAGAACAAGAGAAGAGAAACAGGAGCCGACGAC-	1169	26
#AP-Cntr 17	-CGACAAAACAAGGGGAACACACGGAAAGACCAGAAAACGGGGTAGGACGTAG-	1099	29
#AP-Cntr 18	-CGGACAAAGGGAAGACGAACCACGACAAACAGGGACAAGAATATAGTGATGG-	924	33
#AP-Cntr 19	-CATAGACACGGAACACACAGAACGGAAATAAAGGCTGCGAAGATCGGGATGG-	888	34
#AP-Cntr 20	-CGGGACAAGACAACGAACGAACAGGAAGAGAACCGGAATGCAGACGTCAGGG-	882	64

Cell-SELEX rank indicates the frequency rank of each detected sequence in the enriched pool of the Cell-SELEX experiment.

ND refers to each sequence that is not detected in the enriched pool of the Cell-SELEX experiment.

Sequence and Affinity Analysis

Multiple alignments of the obtained sequences using ClustalX2 showed no motifs or consensus sequences (Figure S3). However, the phylogenic analysis revealed at least three promising families containing confirmed members of Cell-SELEX methods (Figure ). The representatives (AP-Cell 1, AP-Cell 2, and AP-Cell 3) were synthesized with a 5′-FAM label and then evaluated for the dissociation constant utilizing flow cytometry (Figure ). In this manner, three proper oligomers, AP-Cell 1, AP-Cell 2, and AP-Cell 3 with reasonable dissociation constants (Kd) of 47.51, 39.38, and 65.23 nM, respectively, were achieved.

Figure 3

Phylogenic analysis of obtained sequences using MEGA7. Phylogenic analysis of enriched pools from Cell-SELEX revealed three major families of sequences that were approved using complementary methods. Checkmark symbol (√) refers to sequences of Cell-SELEX that are approved using the Counter-SELEX.

Figure 4

Affinity assay of selected aptamers using flow cytometry. According to the findings of the flow cytometry assay (mean fluorescence signal) from each FAM-labeled aptamer candidate, a binding saturation curve was obtained to specify the dissociation constant (Kd). FAM-labeled ssDNA library was also considered to determine possible nonspecific binding.

Apta Fluorescence Imaging of Tissue Slides

Given that ASPH is predominantly overexpressed in hepatocellular carcinoma (HCC),[21,25] it was speculated that the obtained aptamers could also be able to spot ASPH on the HCC tissue. To examine the feasibility of clinicopathological diagnosis, the staining of HCC sections with FAM-labeled aptamers along with imaging using fluorescence microscopy was carried out. Obviously, a strong green fluorescence signal was detected in cancerous tissues stained with FAM-labeled AP-Cell 1 and AP-Cell 3, while the adjacent normal tissues displayed negligible signal (Figure ). However, AP-Cell 2 showed no binding signal, indicating the importance of the difference between the microenvironmental condition of cells and tissues. Also, the staining of the same sections with the FAM-labeled random library resulted in no obvious fluorescence signal.

Figure 5

Apta histochemistry using the obtained aptamers. Apta histochemistry binding analysis of promising aptamers against hepatocellular carcinoma (HCC) indicated proper binding of AP-Cell 1 and AP-Cell 3. (a) Anti-ASPH antibody on cancer tissue; (b) Anti-ASPH antibody on adjacent normal tissue; (c) library on cancer tissue; (d) AP-Cell 1 on cancer tissue; (e) AP-Cell 1 on adjacent normal tissue; (f) AP-Cell 3 on cancer tissue; (g) AP-Cell 3 on adjacent normal tissue; (h) AP-Cell 2 on cancer tissue; (i) AP-Cell 2 on adjacent normal tissue. Probe: aptamer or antibody; DAPI: 4′,6-diamidino-2 phenylindole; Merge: merge of Probe and DAPI-related images; Phase: Optic images from the tissue.

Apta Fluorescence Imaging of HeLa Cells

Using apta fluorescence imaging, the capability of FAM-labeled AP-Cell 1 and AP-Cell 3 to specifically bind ASPH was well confirmed through their strong green fluorescence detectable on HeLaASPH, but not on control untransfected HeLa cells, as shown in Figure .

Figure 6

Apta fluorescence imaging using the obtained aptamers. The Apta cytochemistry binding analysis of the obtained aptamers against negative (HeLa) and positive (HeLaASPH) cell lines indicated proper binding of AP-Cell 1 and AP-Cell 3. (A) AP-Cell 1, HeLaASPH; (B) AP-Cell 1, HeLa; (C): AP-Cell 3, HeLaASPH; (D) AP-Cell 3, HeLa. Aptamer: aptamer candidate; DAPI: 4′,6-diamidino-2 phenylindole; Merge: merge of aptamer and DAPI-related images; Phase: optic images from the cells.

Discussion

Among the various approaches envisaged for developing aptamers for different applications like clinicopathological studies, we used Cell-SELEX to take advantage of the unique superiority of this method in utilizing native targets. Furthermore, we applied some modifications to the Cell-SELEX procedure by establishing the Counter-SELEX to enhance the efficiency of the counter selection in excluding nonspecific interfering sequences and increasing the chances of success in the immunostaining of the tissue slides (Scheme ). In the sequencing step, we harnessed the power of next-generation sequencing (NGS) to deepen our understanding to gain a more inclusive view of the selected sequences in the enriched pool and shorten the selection procedure.

Scheme 1

Schematic Diagram of Modified Cell-SELEX Phases

The conventional phase of the cell-based selection (left) consists of nine iterative cycles of selection including positive selection step (using HeLaASPH), negative selection step (using HeLa, not in round 1), PCR amplification, (d) single stranding, and refolding. The enriched pool of the last round of selection was sequenced via NGS. The Counter-SELEX uses the negative cell-bound oligomers of the second round of the conventional phase as the initiating pool, followed by five iterative cycles of selection using negative cells (HeLa). The enriched pool of the last round of the Counter-SELEX was sequenced via NGS

The traditional Sanger sequencing used frequently in various aptamer selection systems can merely visualize a small portion of all of the different sequences available in the enriched pools.[33,34] Researchers have recently implemented deep sequencing to overcome the defects of traditional sequencing in the process of SELEX.[35−37] In agreement with the previous studies,[38,39] our obtained NGS data showed that the final enriched pools consist of a large number of oligomers, the top sequences of which only make up 0.5–3.6% of the total number of sequences (Table ), indicating that the conventional method of sequencing does not provide an authentic view of the enriched pool. Deep sequencing also enables us to track the evolution of ligands much sooner than the process is completed.[38,39] Powerful owing to its harnessing of natural forms of target proteins, the routine Cell-SELEX procedure is quite time-consuming in practice. Due to the use of NGS, we terminated the process at round 9 before the completion of the overall selection procedure, which typically takes 11–16 cycles of selection.[18,40−42] Also, in the conventional Cell-SELEX, the determination of the final cycle is dependent on additional flow cytometric analyses and will be postponed until the mean fluorescence intensity stops improving.[41] Using deep sequencing, we could bypass the urgent need to determine the final round of SELEX and in doing so could properly shorten the overall process period.

Another main weakness of Cell-SELEX is the imperfectness of the counter selection step, the process aimed at eliminating oligomers bound to the control cells,[14,39] which is started usually from the second round of selection.[41] In fact, other than the intended target, there are numerous undesired molecules on the surface of both positive and control cells and thus could serve as easily available targets. When some undesirable oligomers remain in the first round of selection and then amplified via PCR, there is no guarantee to be eradicated in the counter selection of the next round. The evolution of nonspecific oligomers may easily cause false-positive results in the clinicopathological analysis of paraffin-embedded tissues. Some studies have proposed that adding an excess amount of some natural ligands may be helpful,[14,43] though this could only partially remove any competitive aptamers. Here, we designed a procedure, named Counter-SELEX, and applied it successfully to better recognize and deprive control cell-binding sequences. Some oligos with a notable frequency rank (AP-Cell 4 and 6) were identified invalid, suggesting the imperfectness of conventional Cell-SELEX and the necessity of complementary Counter-SELEX. It should be noted that despite the addition of the Counter-SELEX steps, the overall time of the selection process does not increase. On the one hand, these steps go hand in hand with the original Cell-SELEX procedure, and on the other hand, early detection of nonspecific sequences allows the removal of future confirmatory tests, which in turn reduces cost, time, and extra works.

In 2019, Pleiko et al. have tried to find a way to overcome defects of Cell-SELEX in the elimination of nonspecific oligos.[44] In this study, they performed a conventional Cell-SELEX including positive and negative selection in 11 rounds. Then, in round 11, they compared the differences between the sequences in the pools obtained after incubation with the control and target cells. They claimed that the differences obtained using high-throughput sequencing could reveal nonspecific sequences. This strategy seems to have some drawbacks. Up to cycle 11, the copy number of each target-specific sequence has increased dramatically. Therefore, in the counter selection step using negative control cells, a large part of these copies may remain in the resulting pool and could be mistaken for sequences that bind to the control cells when NGS is performed. Moreover, during these 11 cycles, the number of sequences that bind to the control cells is constantly reduced using the counter selection step. Therefore, the oligos which potentially bind to the control cells may not be so frequent. A single incubation step in an individual round of selection (like round 11) appears to have limited power to make a significant difference. To better differentiate, we performed five consecutive rounds of selection on the control cells to make the difference more obvious. Moreover, we utilized the pool from the second round of selection as the initiation library for the Counter-SELEX, in which target-specific sequences have not yet reached high frequency.

Meyer et al., have proposed another strategy to deal with this problem.[45] Briefly, first, five rounds of conventional Cell-SELEX including both positive and negative selection steps were performed, then the resulting pool was subjected into two separate directions. One path included six rounds of selection on control cells, while the other path used positive target cells in the selection process (without any counter selection step). In the end, the pools obtained from each path were deeply sequenced and compared. If the sequences with the binding affinity to the negative control cells are to be identified, trying to remove them in the first five rounds of selection using the negative selection step will reduce their frequency, making them difficult to be distinguished and detected as highly specific ligands for controls. At the same time, in these five rounds, the frequency of specific sequences for the target has also increased, making it more difficult to eliminate these sequences during the selection process using control cells. Another drawback of this strategy is removing the counter selection step in the last six rounds of selection on target cells. This could cause a sharp elevation of the sequences binding to nonrelevant molecules that are abundant on the surface of the cells. In our strategy, the counter selection step of the main SELEX procedure has never been omitted to constantly reduce the sequences attached to the interfering molecules or at least to prevent an increase in their frequency. Moreover, as described before, in the Counter-SELEX, the selection of the initial pool (from the second round) along with the five sequential selection cycles on the control cell makes it possible to make the greatest difference in frequency between sequences capable of binding control and target cells.

Even using our modified method including the Counter-SELEX, there may still be sequences that recognize the targets on the surface of cells but are not able to detect them at the tissue level. This may be due to the differences in spatial constraints or the microenvironment around the cell and tissue. Therefore, for clinicopathological applications requiring fluorescence imaging, the binding ability of the candidate sequences needs to be checked. Owing to the power of our novel complementary method and using the phylogenic analysis, the promising candidates were successfully identified and then evaluated using flow cytometry. However, the natural environment of tumor tissue may be completely different from the cell lines, so the binding capacity of aptamers against ASPH was also assessed using a tumor tissue with overexpression of ASPH like hepatocellular carcinoma.[21] Successful binding to HCC indicated the potential of the aptamers for clinical applications, although advanced clinical assessments are needed for definitive confirmation.

Since the counter selection efficiency of Cell-SELEX was enhanced by implementing the Counter-SELEX, we concluded that our novel method could effectively improve aptamer selection for future clinicopathological applications. Also, the overall duration of selection could be apparently reduced using the potential of NGS to track evolving sequences.

It would be of further interest to study whether the obtained ASPH-specific aptamers could restrain the ASPH function. Also, their potential in target therapy would be an attractive purpose for future studies.

Material and Methods

Construct Preparation

To achieve recombinant ASPH in mammalian systems, the pcDNA3.1/Hygro(+) vector (Thermo Fisher Scientific, Waltham, MA) containing mammalian cytomegalovirus (CMV) was implemented. This plasmid also involves the Hygromycin resistance gene (Hygromycin-B-phosphotransferase) as the selectable marker of stable transfectants in mammalian cells and the Ampicillin resistance gene (β-lactamase) as the selectable marker of the vector in the properly transformed Escherichia coli strains. A modified cDNA of ASPH (NCBI Accession Number: NM_004318) containing the C-terminal 6-His tag coding sequence before the stop codon and the Kozak consensus sequence (ACC ATGG) just before the start codon was designed before being synthesized and cloned into the vector by General Biosystems, Inc. (Morrisville, NC) to give the pcDNA3.1/Hygro(+)-ASPH expressing the 2310 bp NheI to XhoI fragment just after the CMV and T7 promoter.

Transfection Procedure in HeLa Cells

The pcDNA3.1-ASPH constructs were first transformed into Top 10 F′ strain (Novagen) as a propagation host using the calcium chloride transformation method to make a reserve of the construct. After harvesting GenElute Plasmid Miniprep (Sigma), the plasmids were linearized utilizing FspI (New England Biolabs) according to the manufacturer’s instructions. Then, a stable human cervical carcinoma cell line (HeLa) with overexpression of ASPH on the cell surface (HeLaASPH) was established using linearized pcDNA3.1/Hygro(+)-ASPH plasmid and TurboFect Transfection Reagent (Thermo Fisher Scientific) according to the manufacturer’s instructions, followed by a selection using 200 mg/mL hygromycin B (Solarbio Science & Technology). Transfection was evaluated by measuring mRNA (Forward primer: TTGGCGTGGGATACCTCTTG; Reverse primer: GTCACACTCAGCACCTCTTC) using quantitative RT-PCR and the 2–ΔΔC method.[31] Also, the flow cytometry analysis of cell surface-displayed ASPH was performed using FB-50 biotinylated antibody and PE-streptavidin (BioLegend).

ssDNA Library

A random ssDNA library containing two constant regions as primer binding site flanking a central sequence with 52 randomized bases was purchased from TAG Copenhagen. The sequences of forward primers (with and without 5′-FAM label) and 5′-phosphate reverse primer are as follows:

ssDNA library: 5′-ATACCAGCTTATTCAATT-52N-AGATAGTAAGTGCAATCT-3′

Forward primer: 5′-ATACCAGCTTATTCAATT-3′

Forward primer (FAM-labeled): 5′-FAM-ATACCAGCTTATTCAATT-3′

Reverse primer: 5′-Phosphate-AGATTGCACTTACTATCT-3′

Cell-SELEX and Counter-SELEX

The overall procedure of the modified Cell-SELEX is schematically presented in Scheme . First, 20 nmol of the initial ssDNA library was dissolved in 1000 μL of the binding buffer (5 mM MgCl2, 4.5 g of glucose, 1 g BSA, and 100 mg yeast tRNA in 1 L Dulbecco’s PBS), followed by heating at 95 °C for 5 min and then snap-cooling on ice. The refolded pool was then incubated directly to over 5 million HeLaASPH cells for 1 h. After washing, the bound sequences were recovered and amplified via 10 cycles of PCR (hot start: 2.5 min, 95 °C; denaturation: 0.5 min, 94 °C; annealing: 0.5 min, 46 °C; extension: 0.5 min, 72 °C; final extension: 5 min, 72 °C). An additional PCR was carried out at the cycles of 4, 6, 8, 10, and 12 to determine the optimum number of cycles for a preparative PCR. Single stranding of the products was done using lambda exonuclease III (Thermo Fisher Scientific) according to the manufacturer’s protocols. The counter selection was initiated from the second round when the recovered ssDNA from the positive selection was renatured and then incubated with the control untransfected HeLa. To enhance the affinity of the selected aptamers, the number of washing as well as the duration and volume were gradually increased, while the incubation time was reduced to 30 min. Meanwhile, the cell number was gradually reduced to 1 million in cycle 5. Moreover, up to 20%, FBS was added gradually to the binding buffer (Table S1). After the last round of selection (round 9), the PCR product was sequenced by GenXPro GmbH (Frankfurt, Germany) using Illumina NextSeq. 500 (1 million reads, 1 × 75 bps).

To overcome the imperfectness of the counter selection and to increase the chances of recognizing and eliminating the unduly retained oligomers, we effectively established a novel negative selection procedure named Counter-SELEX. In this method, the oligomers bound to the control cells at the second round of selection were literally recovered, amplified, and then subjected to five iterative rounds of selection, using control (untransfected) HeLa cells as the main target of SELEX (Table S2). Finally, using deep sequencing, the most prevalent sequences bound to the common surface molecules on the control cells were determined.

Sequence Alignments and Phylogenic Analysis

To identify probable consensus sequences, motifs, and families, sequence alignment and phylogenic analysis were accomplished using ClustalX2 and MEGA7 softwares. After confirming each sequence via Counter-SELEX, the most promising oligomers were determined among members of each family according to their frequencies.

Affinity of the Promising Aptamers Using Flow Cytometry

The binding affinity between the selected aptamers and the target protein ASPH displayed on the cell surface was measured and reported by the equilibrium dissociation constant (Kd). Various concentrations of each aptamer or the initial unselected library (600, 300, 150, 75, 37.5, 18.75, and 9.37 nM) were incubated with 5 × 105 HeLaASPH, and the mean fluorescence intensity of each concentration was determined using flow cytometry (FACSCalibur, BD Biosciences). After subtracting the fluorescence background of controls, the equilibrium dissociation constant (Kd) of aptamers was determined using the equation Y = Bmax X/(KD + X) (X: aptamer concentration; Y: MFI of X; Bmax: maximum MFI).

Apta Fluorescence Imaging of Tissue Slides

To examine the capability of the obtained aptamers in clinical applications, tumor/adjacent normal tissue staining using FAM-labeled aptamers was performed. Given that ASPH is predominantly overexpressed in hepatocellular carcinoma 21, the HCC obtained from the pathology department of Alzahra Hospital (Isfahan, Iran) was prepared for staining using a process similar to immunohistochemistry including deparaffinization, dehydration, and antigen retrieval as described before.[32] Briefly, the deparaffinization of the preheated sections (at 60 °C for 2 h) was done using xylene for 10 min twice, followed by dehydration using decreasing concentrations of ethanol (100, 95, 85, and 70%) at 5 min intervals. The sections were then washed (PBS, pH: 8, twice), pretreated in citrate buffer (0.01 mol/L, pH: 6), and heated at 95 °C under pressure for 15 min to retrieve antigens. The slides were then incubated with a precooled binding buffer containing 0.1 mg/mL Herring Sperm DNA and 20% FBS for 1 h at room temperature. Afterward, the blocked sections were treated with 250 nM intended FAM-labeled aptamers in binding buffer 60 min at 4 °C on ice in the dark. Subsequently, the stained tissue sections were washed with PBS three times and imaged via a fluorescence microscope (Labomed LX 400, Labomed). The immunofluorescence imaging of the tissues was also carried out using biotinylated FB50 anti-ASPH antibody and PE-Streptavidin.

To examine the specificity of the obtained aptamers, apta fluorescence imaging of HeLa and HeLaASPH cells was performed. The cells were seeded in a 24-well plate at a density of 105 cells/mL 24 h before the experiment analysis and then fixed using a fixation buffer (PBS with 4% formaldehyde) at the time of experiment for 30 min. After three times of washing with PBS, the wells were blocked using 5% BSA for 60 min. After the next washing step (three times), the cells were stained with FAM-labeled aptamers, enriched ssDNA library, and unselected library (control) for 2 h. After washing again (three times), another staining step was performed using 4′,6-diamidino-2 phenylindole (DAPI; Invitrogen) for 2 min. Finally, the fluorescence images were captured using a Nikon Microphot-5A inverted fluorescent microscope.