Aptamer-Based Chromatographic Methods for Efficient and Economical Separation of Leukocyte Populations

Abstract

The manufacturing process of chimeric antigen receptor T cell therapies include isolation systems that provide pure T cells. Current magnetic-activated cell sorting and immunoaffinity chromatography methods produce desired cells with high purity and yield but require expensive equipment and reagents and involve time-consuming incubation steps. Here, we demonstrate that aptamers can be employed in a continuous-flow resin platform for both depletion of monocytes and selection of CD8⁺ T cells from peripheral blood mononuclear cells at low cost with high purity and throughput. Aptamer-mediated cell selection could potentially enable fully synthetic, traceless isolations of leukocyte subsets from a single isolation system.

Graphical Abstract

INTRODUCTION

Chimeric antigen receptor (CAR) T cell therapy is a revolutionary cancer treatment in which autologous or allogeneic T cells are genetically modified with tumor-targeting receptors that directs them to identify and destroy tumor cells.¹ As of writing, the clinical success of CAR T cell therapy has led to six FDA-approved products for treating relapsed/refractory B cell acute lymphoblastic leukemia, various B cell lymphomas, and multiple myeloma, along with over 500 active global clinical trials.^1–9

Production of autologous CAR T cells consists of T cell extraction, activation, genetic modification for CAR expression, expansion, and re-infusion into the patient. In the first step, T cells are enriched from a patient’s own peripheral blood mononuclear cells (PBMCs) that are collected through leukapheresis. Removal of undesired cells, such as monocytes or tumor cells, enhances the safety, efficacy, and production yield of CAR T cells.^10–14 Cellular composition and phenotype of extracted T cells, such as CD4⁺:CD8⁺ T cell ratio and amount of early-differentiated memory T cells, also contribute to the treatment success of CAR T cell therapy.^10,15,16

T cells are currently isolated by magnetic-activated cell sorting (MACS) or by immunoaffinity chromatography (IAC). MACS involves cell incubation with antibody-coated magnetic beads followed by separation via magnetic columns, resulting in high purity and yield of target cells.^17,18 However, the required materials for MACS are expensive, owing to the high cost of biologically produced antibodies and there being a sole provider of goods. Demonstrating this, the cost for the CD4⁺ and CD8⁺ Microbeads and appropriate tubing alone for a single patient on the CliniMACS instrument exceeds $10,000.^19,20 Additionally, MACS-purified target cells retain bound Microbeads if positively selected, which can alter the phenotype of isolated cells, raises immunogenicity concerns for adoptive cell therapy applications, and prohibits further isolation of cell subsets defined by two or more markers. IAC uses reversible affinity reagents, such as the Twin-Strep-tag:Strep-Tactin platform.²¹ In this system, Strep-Tactin is immobilized on the selection matrix that is subsequently bound with Strep-tagged Fabs specific for target molecules. Following cell loading and washing to remove unbound cells, bound target cells are eluted with biotin, which outcompetes the Strep-tagged Fabs due to higher affinity for the Strep-Tactin multimer.^22,23 While IAC reduces the magnetic equipment cost and the cell-preprocessing steps associated with MACS, it still relies on expensive, biologically produced affinity reagents.

Aptamers are short, single-stranded nucleic acid molecules that bind to targets with affinity and specificity comparable to antibodies.^24,25 Discovered by a library selection process called Systematic Evolution of Ligands by Exponential Enrichment (SELEX), aptamers offer many advantages over antibodies. Aptamers are produced synthetically with low cost, high uniformity, and if desired, well-defined chemical modifications. Additionally, aptamers exhibit significantly improved stability and shelf-life compared to antibodies, especially when stored lyophilized, and have low immunogenicity and toxicity.^26,27 Given these favorable qualities, aptamers have become popular affinity reagents for the separation of small molecules,²⁸ proteins,²⁹ and cells.^2,30 Importantly, complementary oligonucleotides can be designed to selectively disrupt aptamer binding for traceless and multiplexed isolation of different cell populations in a single apparatus.³¹

To meet the needs of the expanding cell therapy market and global demand, cell manufacturing technologies that facilitate higher throughput and automation are necessary. A continuous-flow cell isolation system could be incorporated into closed-system manufacturing. Here, we adapted aptamer-mediated cell selection to an affinity chromatography resin platform. Our advance to a fully synthetic, continuous-flow resin system exhibits high-throughput capability and eliminates the need for magnetic supports and laborious cell pre-processing steps. The resin system also enables the shift into a closed system that reduces the potential of contamination, decreases personnel effort, and minimizes production time.²² We evaluated resin coupling techniques, aptamer density, blocking buffer composition, pre-clearing step inclusion, aptamer link length, and resin volume as parameters for optimizing cell isolation and purity. We then successfully applied aptamer-functionalized resins for both negative selection to remove undesirable cell types and positive selection to enrich desired cell types. For the former, we applied our previously-reported monocyte-binding aptamer (Mono.A2)³² to deplete human monocytes from PBMCs. For the latter, we adapted our CD8⁺ T cell-binding aptamer (CD8.A3t) that we previously used in an aptamer-mediated MACS method for isolating CD8⁺ T cells² to a continuous-flow resin platform for CD8⁺ T cell isolation. Importantly, isolated CD8⁺ T cells are recovered label-free by simply flushing the resin with a pipette and can thus undergo successive isolations with the same approach for increased purity. This work demonstrates the versatility of aptamer-based affinity chromatography for cell partitioning, laying the groundwork for next-generation cell separation technologies that will decrease the cost and time of cell therapy manufacturing.

MATERIALS AND METHODS

PBMC Isolation.

Human PBMCs were isolated from Leukocyte Reduction System (LRS) cones (Bloodworks Northwest) with Ficoll-Paque density gradient centrifugation (GE). Briefly, mononuclear cell-enriched blood was sterilely extracted from the LRS cones (8-10 mL) and diluted with 50 mL Dulbecco’s phosphate-buffered saline (DPBS, Gibco) supplemented with 2 mM EDTA (Invitrogen). The diluted blood was then overlaid over two tubes with 15 mL Ficoll-Paque (~1:2 volume ratio of Ficoll-Paque to dilute blood) and centrifuged at 400×g for 40 min at 20°C with zero acceleration and deceleration. After removing the upper plasma layer, PBMC-enriched buffy coats were extracted from above the Ficoll-Paque layer, washed twice in DPBS 2 mM EDTA, and resuspended in DPBS before counting. PBMCs were cryopreserved in RPMI (Corning) supplemented with 20% FBS (Life Tech) and 10% DMSO (Sigma) before use in isolation procedures.

Preparation of Aptamer-Conjugated Resin Supports.

Thiol-labelled aptamers were synthesized by Integrated DNA Technologies, and their sequences and predicted secondary structures are provided in Table S1 and Figure S1, respectively. Epoxidized polystyrene resin (50-150 μm, CytoSorbents, provided by Juno Therapeutics) was rehydrated per gram with 5 mL 20% isopropanol for 15 min at room temperature, where every 0.2 g of dry resin resulted in 0.5 mL of wet bed volume. Resin was then washed and either pre-blocked per gram of resin in 5 mL DPBS with Ca²⁺ and Mg²⁺ (Corning) supplemented with 0.1 mg/mL salmon sperm DNA (SS DNA, Invitrogen) for 30 min at room temperature, passivated in 2.5 mL thiol-C11-sulfobetaine zwitterion (Prochimia) or thiol-PEG12-acid (BroadPharm) for 2 h at room temperature, or incubated with streptavidin (STN-N5116, Acro Biosystems) or neutravidin (31000, Thermo Fisher Scientific) protein for 2 h at room temperature. Meanwhile, various concentrations of thiol-labelled CD8.A3t and Mono.A2 aptamers were reduced with 100-fold excess TCEP for 2 h at room temperature and then folded by heating at 95°C for 5 min in DPBS with Ca²⁺ and Mg²⁺ further supplemented with 5 mM MgCl₂ and titrated to pH 8.0 (Coupling Buffer) followed by snap-cooling on ice. Biotinylated aptamers for streptavidin/neutravidin resin systems were similarly folded without the reduction step. Pre-blocked resin or streptavidin/neutravidin-coated resin was then washed and incubated per gram with either 2.5 mL folded thiol-labelled aptamer, 2.5 mL folded biotinylated aptamer (for streptavidin/neutravidin resin systems), or Coupling Buffer (for uncoupled resin preparations) as a 50% slurry for 2 h at room temperature with gentle rotation. Aptamer-coupled resin was then washed with Coupling Buffer and buffer exchanged into DPBS with Ca²⁺ and Mg²⁺ further supplemented with 5 mM MgCl₂ and 25 mM glucose (Wash Buffer). Unreacted epoxide groups were passivated with Wash Buffer containing 2% bovine serum albumin (BSA, Miltenyi) and 1% Tween-20 (Sigma) for 1 h at room temperature. Pre-blocked, aptamer-coupled, and passivated resin was finally washed and stored in Wash Buffer at 4°C for up to 48 h before cell isolations.

For aptamer conjugation validation experiments, thiol-labelled and unlabeled aptamer were separately coupled to resin at 6.25 nmol/g. 5 uL aliquots of aptamer solution were taken at 0.5 h timepoints and analyzed on both the Qubit 4 Fluorometer (Invitrogen) and NanoDrop One Spectrophotometer (Thermo Fisher Scientific). Aptamer conjugation plots were normalized to the 0 h timepoint.

Cell Isolation Procedure with Aptamer-Conjugated Resin.

Prepared resins were buffer exchanged into Wash Buffer containing 1% BSA and 0.1 mg/mL SS DNA before loading into polypropylene gravity-flow columns with porous 30 µm polyethylene bed supports (Bio-Rad Poly-Prep or Econo-Pac). The resin-loaded columns were washed once with the same buffer before applying thawed PBMCs resuspended at 20×10⁶ cells/mL in the same buffer for gravity-flow isolation. Poly-Prep and Econo-Pac columns were then washed three times with 2 and 5 mL DPBS containing 0.5% BSA and 2 mM EDTA, respectively, and the flowthrough from the initial cell application and washes were collected and pooled. Captured cells were eluted by vigorously pipetting the resin three or five times with 2 mL DPBS containing 0.5% BSA and 2 mM EDTA for the Poly-Prep and Econo-Pac columns, respectively. Cell aliquots were collected at each step of the cell isolation procedure.

For the pre-clearing procedure, cell solutions were first passed through 1 mL of pre-blocked, passivated resin without aptamer coupling in an Econo-Pac column. Cells were then run through 0.5 mL of aptamer-coupled resin in a Poly-Prep column, unless otherwise specified. For the optimized double purification procedure, cells were pre-cleared as described above and then subjected to two rounds of aptamer-coupled resin isolation, first through 1 mL of aptamer-coupled resin in an Econo-Pac column and then through 0.3 mL of aptamer-coupled resin in a Poly-Prep column. For pump-based isolation, an Econo-Pac column with 1 mL of resin was fitted with an Econo-Pac Flow adaptor (7380019, Bio-Rad) and a 3-way stopcock. A syringe pump (Harvard Apparatus) was used to pass cells and buffer through the system at a controlled flow rate (1-4 mL/min) and cells were mechanically eluted from the resin after removal of the column adaptor, as detailed above.

For high-throughput blocking buffer screening experiments, 50 μL of resin (0.02 g dry weight) was loaded in each well of a 25 µm filter microplate (201718-100, Agilent). PBMCs (2.5×10⁶) were loaded at 12.5×10⁶ cells/mL in each well. Blocking buffer formulations of BSA (2 or 5%) and Tween-20 (1, 2 or 4%) were used, with a comparison to 1 M ethanolamine pH 8.5 solution (Sigma-Aldrich). Flowthrough washes were performed with careful addition of DPBS containing 0.5% BSA and 2 mM EDTA and incubation on an orbital shaker (Eppendorf) at 300 rpm for 1 min. Flowthrough washes were then collected by centrifugation at 180×g for 3 minutes. Cells were eluted with vigorous pipetting using a multichannel P200 pipette and subsequent centrifugation spins at 180×g for 3 minutes.

Flow Cytometry Reagents and Staining.

The following dyes and antibodies were used to stain cells: Zombie Violet (1:500 in 100μL per 10⁶ cells, BioLegend), APC antihuman CD8a (rhesus cross-reactivity, 1:100, 301014, BioLegend), PE antihuman CD3 (1:200, 300308, BioLegend), FITC antihuman CD14 (1:50, 301804, BioLegend), and Super Bright 702 antihuman CD19 (1:20, 67-0199-42, Invitrogen). Aliquots of 10⁶ cells from pre-sort, flowthrough, and flush fractions were washed with DPBS and first incubated with live/dead Zombie Violet dye for 15 min at room temperature. Cells were then washed with DPBS 1% BSA to neutralize excess Zombie Violet, blocked with 10 μL FcR Blocking Reagent (Miltenyi) for 10 min at 4°C, and stained with 100 μL antibody solutions for 20 min at 4°C. Stained cells were then fixed in 200 μL of DPBS 1% BSA with 0.1% paraformaldehyde (Alfa Aesar) and analyzed on an Attune NxT (Invitrogen) flow cytometer. Unstained and single-stained cells were used for compensation controls. The flow cytometry gating strategies for determining cell yield and purity are shown in Figure S2.

RESULTS AND DISCUSSION

Preparation of Aptamer-Functionalized Resin Supports.

To generate the resin for cell selection and cell depletion, thiolated CD8.A3t and Mono.A2 aptamers were respectively chemically grafted onto epoxidized polystyrene resin supports (Scheme 1). Polystyrene resin supports were selected for their low cost and ease of sterilization compared to other resin supports, such as agarose. The epoxy-thiol chemistry is compatible with the relatively neutral pH conditions required for stable aptamer conformations and utilizes the thiol nucleophile for epoxide ring opening to a stable β-hydroxythioether linkage. Mono.A2 aptamer,³² which binds to human monocytes with a K_d of 45 nM, was used for negative selection because monocytes are a common contaminant in apheresis material that hinders CAR T cell production.³³ CD8.A3t aptamer,² which binds highly specifically to the CD8⁺ cytotoxic subset of T cells with a K_d of 14.7 nM, was selected for positive enrichment of CD8⁺ T cells crucial for subsequent generation of engineered CAR T cells.

Scheme 1.

Conjugation of thiol-labelled aptamers to epoxidized polystyrene resin for chromatographic cell separation. Epoxidized polystyrene resin particles are rehydrated, washed, pre-blocked with salmon sperm DNA, and conjugated with TCEP-reduced, thiol-labelled aptamers in modified DPBS, pH 8 for 2hr. After conjugation, aptamer-coupled resin is washed and blocked with 2% BSA, 1% Tween-20 in modified DPBS solution.

To prevent nonspecific adsorption of aptamers to the resin, we applied salmon sperm DNA as a pre-blocking agent prior to resin conjugation with thiolated aptamer. Aptamer conjugation was confirmed by both UV-Vis Nanodrop and fluorescent Qubit measurements that revealed selective depletion of thiolated aptamer from solution compared to unfunctionalized aptamer control (Figure S3). Following aptamer conjugation, the resin was treated with BSA and Tween-20 for quenching unreacted epoxide groups and passivation.

Optimization of Monocyte Depletion from PBMCs through Mono.A2-conjugated Resin.

Typical monocyte populations in peripheral blood of healthy individuals range from 10% to 30% in collected leukapheresis product.^14,34 Monocytes, circulating white blood cells that can differentiate into antigen-presenting cells, disrupt CAR T cell production when present in high quantities due to their phagocytosis of the CD3/CD28 Dynabeads needed for T cell activation prior to transduction.^13,14,35,36 Noaks et al. demonstrated that the removal of monocytes in healthy leukapheresis product led to improved T cell activation and transduction efficiency for CAR T cell products.¹³ Thus, we applied the continuous-flow aptamer-mediated resin chromatography system for negative selection of monocytes from PBMCs.

Mono.A2 aptamer was partially truncated in the stem region to decrease costs and modified with both a 12(oligo)dT (12dT) spacer and PEG spacer for added spacing (Table S1 and Figure S1), which we discuss and characterize further in the upcoming results section for the CD8.A3t aptamer. We first determined the effect of pre-clearing procedures on monocyte depletion. We applied PBMCs to unfunctionalized resin for pre-clearing before loading on Mono.A2-conjugated resin packed into gravity flow columns (Figure 1A). The pre-clearing step removed particularly adherent monocytes and B cells in addition to dead cells and cell debris. This procedure removed 14.8% (±5.8%) of the total PBMC cell number and 21.8% (±10.3%) of total monocytes (Table S2). The pre-clearing step also improved monocyte depletion in the subsequent Mono.A2 resin step by 20% in flowthrough populations (Figure S4) and minimized column clogging for better processing. We then optimized the aptamer density on the resin by reacting various concentrations of thiolated Mono.A2 aptamer (3.125, 6.25, 12.5, and 25 nmol/g resin) with resin. As the concentration of Mono.A2 increased above 6.25 nmol/g resin, the monocyte fold-depletion decreased, likely due to disrupted aptamer folding and restricted steric accessibility of the immobilized aptamer.³⁷ Resins incubated with 6.25 nmol/g resin of Mono.A2 provided the highest monocyte depletion (Figure 1B) and were therefore used in replicate, large-scale studies.

Figure 1.

Monocyte aptamer (Mono.A2) depletion studies. A) Monocyte depletion scheme. Thawed PBMCs are run first through uncoupled resin, yielding pre-cleared PBMCs (PC PBMCS) in the flowthrough that are subsequently loaded into a Mono-A2-conjugated resin. B) Optimization of Mono.A2 aptamer concentration for resin coupling was performed with a single PBMC donor using concentrations ranging from 3 to 25 nmol/g resin, where a concentration of 6.25 nmol/g resin depleted the monocytes the most. C) Representative flow cytometry histograms of CD14 expression in the initial PC PBMCs compared to the monocyte-depleted Mono.A2 flowthrough fractions. D) Final CD14⁺ monocyte purities in pre-sort PBMCs, PC PBMCs, and flowthrough PBMCs. The circles, squares, and triangles represent different donors from separate isolation experiments. Horizontal lines and error bars represent the mean ± s.d.; ns = not significant; ***p<0.001 (one-way ANOVA test)

Overall, we depleted the monocyte population from three donor PBMCs from 23.2-31.9% to 3.66-3.82% (Figure 1C,1D), thereby removing over 94% of total monocytes from PBMCs by a single pass through the Mono.A2-conjugated resin. Notably, CD4⁺ and CD8⁺ T cell concentrations increased in the donor PBMCs by about 11% and 5%, respectively, after monocyte removal (Table S2). For CD4⁺ T cells, CD8⁺ T cells, NK cells, and B cells, we recovered an average cell yield of 49.9%, 47.5%, 40.7%, and 34.5%, respectively, in the final monocyte-depleted flowthrough fractions. The reduced cell yield might be due to monocyte buildup in the resin, which could partially block flowthrough of other cell types.

Optimization of CD8⁺ T Cell Selection from PBMCs through CD8.A3t-conjugated Resin.

To demonstrate resin-based, aptamer-mediated positive cell selection, we used the aptamer-functionalized resin chromatography system to isolate CD8⁺ T cells from PBMCs. We loaded CD8.A3t-functionalized resin into a gravity flow column, followed by addition of PBMCs to the column for CD8⁺ T cell binding. We washed away unbound, undesired cell populations with buffer washes before mechanically flushing columns with buffer to recover captured CD8⁺ T cells. While we have demonstrated in previous work that cells can be recovered by the application of a complementary DNA sequence to the aptamer,² we used mechanical agitation as an alternative approach in this work due to its simplicity. Overall, cell-processing times were low, requiring only about one hour to run a single column, and cell viabilities were high, reaching above 80% for bound cells. However, in our preliminary studies, we observed that both cell purity and yield were sub-optimal, around 40-50% and 45-60%, respectively.

To improve cell purity, we investigated (i) methods for chemical resin passivation, (ii) various blocking buffers, (iii) gravity versus pump-based flow, and (iv) inclusion of a pre-clearing step. Nonspecific adsorption of proteins and cells to polystyrene surfaces, a long-reported phenomenon, can cause a decline in cell purity and yield for cell selection.³⁸ To reduce nonspecific protein and cell adsorption, surface modifications or surface coatings are commonly employed. PEG modifications have successfully reduced nonspecific surface interactions of blood cells, PBMCs, and proteins.^39,40 Likewise, zwitterions have been used to passivate solid surfaces and nanoparticles against protein and cell adsorption.⁴¹ We therefore passivated epoxidized resin beads with varying concentrations of thiol-C11-sulfobetaine zwitterion (zwitterion) or thiol-PEG12-acid (PEG). Washed PBMCs were then loaded onto the resin and bound PBMCs were eluted through mechanical agitation. Nonspecific binding was recorded as the percentage of PBMCs bound and eluted from the resin out of the total number of PBMCs. Three conditions reduced nonspecific binding of PBMCs compared to the non-passivated resin: zwitterion (3.3 mM), PEG (3.3 mM), and PEG (10 mM). However, these modifications only reduced nonspecific binding by less than 1.5% (Figure S5, Table S3). We thus omitted passivation agents from the protocol to minimize processing steps and costs.

We also tested different blocking agents to minimize nonspecific absorption of cells to the resin, such as BSA, Tween-20, and ethanolamine. BSA is an inert plasma protein commonly used as a blocking reagent to reduce nonspecific adsorption of proteins and cells to synthetic surfaces.⁴² Tween-20 is a detergent used to supplement BSA blocking buffer and reduce nonspecific binding. Ethanolamine was chosen as a blocking agent to quench unreacted epoxide groups on the resin through ring-opening hydrolysis. To screen various formulations of blocking buffer, we employed a high-throughput assay format by loading CD8.A3t-coupled resin into 96-well filter microplates in different compositions of BSA (2 or 5%), Tween-20 (1, 2, or 4%), and ethanolamine (0 or 1 M). Following resin loading, we added PBMCs to the resin at 12.5×10⁶ cells/mL and performed flowthrough washes and elution flushes with centrifugation. Analyzing each flowthrough and flush fraction for CD8⁺ purity by flow cytometry, we found that blocking buffer with 2% BSA and 1-2% Tween-20 or 5% BSA and 1% Tween-20 was most effective in reducing nonspecific cell adsorption (Figure S6). To minimize potential cell toxicity from high Tween-20 concentrations and BSA input, we moved forward with 2% BSA and 1% Tween-20 as a blocking buffer for all subsequent studies in resin preparation.

Next, we tested a pump-based chromatography system for comparison to gravity flow. We were able to better control flow rates with the pump-based system and discovered that higher flow rates reduced nonspecific binding, improving purity of CD8⁺ cells from 53% to 66% (Figure S7). However, the higher flow rates reduced the final CD8⁺ cell yield by 29%, and the pump-based system required lengthy column packing procedures. Thus, while useful for eventual automated and closed cell separation applications, we continued to use gravity-flow columns in our laboratory-scale optimization experiments.

An initial pre-clearing step with the non-functionalized support matrix has been shown to minimize nonspecific binding of unwanted proteins to the affinity chromatography column.^43,44 We also showed that pre-clearing improved monocyte depletion in PBMCs with Mono.A2-coupled resin. To further improve T cell purity from resin-based isolation, we compared a pre-clearing series to a single-column platform (Figure 2A). In the pre-clearing approach, we first ran PBMCs through uncoupled resin prepared with the same pre-blocking and blocking steps as the aptamer-coupled resin. We then loaded the collected flowthrough from the uncoupled resin on the CD8.A3t aptamer-coupled resin. In parallel, we compared this method to PBMCs run through a single CD8.A3t aptamer-coupled resin column. The pre-cleared elution fraction exhibited 76.7% CD8⁺ purity, nearly two-fold higher than the 39.7% purity from the single aptamer-coupled elution fraction (Figure 2B). Cell populations determined through flow cytometry demonstrated higher removal of monocytes and B cells through the pre-clearing step, likely due to the large size of monocytes and higher nonspecific binding properties of both monocytes and B cells, which increased the T cell compositions in the PBMC fractions (Table S4). The trade-off for the increase in purity was a decrease in CD8⁺ cell recovery by over 15% (47.0% versus 63.6% yield), which was expected since the PBMCs needed to pass through an additional column.

Figure 2.

CD8⁺ T cell aptamer (CD8.A3t) selection optimization studies. A) PBMC cell loading protocol for pre-clearing step compared to control for CD8⁺ T cell selection on CD8.A3t-conjugated resin. Thawed PBMCs were run through unfunctionalized resin first in the pre-clearing procedure. The resulting uncoupled flowthrough was run through CD8.A3t-coupled resin. This was compared to the initial procedure of thawed PBMCs run through one round of CD8.A3t-coupled resin alone. B) Flow cytometry histograms of CD8 expression comparing isolation purities from elution fractions with pre-clearing procedure and without pre-clearing procedure. C) Purity and yield analysis comparing PBMCs isolated through CD8.A3t- and 12dT-CD8.A3t-coupled resin, performed on the same donor. D) Optimization of 12dT-CD8.A3t aptamer concentration for resin coupling performed across two PBMC donors using concentrations ranging from 0.625 and 25 nmol/g resin, where the highest purity and yield occurred at 0.625 nmol/g resin. E) Optimization of 12dT-CD8.A3t-coupled resin volumes after pre-clearing using 0.3, 0.4, and 0.5 mL volumes.

We next focused on improving CD8⁺ T cell yield. Many researchers have employed the high affinity interaction between biotin and streptavidin to immobilize biotinylated ligands onto streptavidin-coated resin supports for affinity chromatography.^45–48 This strong, non-covalent grafting technique circumvents complex chemical conjugation conditions and methods. We compared this system of biotinylated aptamer coupled to streptavidin-coated epoxidized resin to our initial system of thiolated aptamer conjugated directly to epoxidized resin. The streptavidin resin system isolated CD8⁺ cells with slightly less purity (66%) compared to those isolated from the thiol-conjugated system (71%). Interestingly, the streptavidin system improved the yield of CD8⁺ cells by over 20% (Figure S8). We hypothesize that the higher CD8⁺ cell capture may be due to the larger spacing of the aptamers from the resin matrix due to the addition of streptavidin. This increased spacing likely reduces steric hindrance of cell-aptamer binding, thus enhancing effective binding interactions. Furthermore, biotinylated CD8.A3t was HPLC-purified while the thiolated CD8.A3t we used was not, increasing grafting rates of the correct aptamer. Despite the increased yield and CD8⁺ capture, the streptavidin system revealed many limitations. Most significantly, the streptavidin system increased nonspecific binding of cells, particularly for CD4⁺ T cells and “Other” cells, which likely included neutrophils and other granulocytes. Streptavidin has been shown to bind cell surface integrins through an RYD (Arg-Tyr-Asp) sequence that mimics the RGD (Arg-Gly-Asp) motif of fibronectin, facilitating nonspecific cell binding.^49–51 Neutravidin, a deglycosylated avidin protein that still binds with high affinity for biotin, lacks the nonspecific binding RYD sequence.⁵² However, we found that biotinylated CD8.A3t coupled to neutravidin-coated resin did not improve CD8⁺ purity in cell isolation experiments (Figure S9). Moreover, a streptavidin- or neutravidin-containing system is more expensive, lower-throughput, and requires milder elution methods than a direct aptamer-functionalized resin system.

We thus increased the linker length on the CD8.A3t aptamer to model the larger spacing between the resin and aptamer from the streptavidin system. The previous aptamer consisted of only a single PEG6 spacer, which provided about 2 nm distance between the aptamer and the resin. We modified the aptamer to include a 12dT spacer in addition to the existing PEG6 spacer (12dT-CD8.A3t), which provided an additional 4 nm between the aptamer and resin, a similar spacing length as the streptavidin protein.⁵³ Moreover, the modified CD8.A3t aptamer was HPLC-purified, guaranteeing at least 85% purity compared to the previous 40-60% aptamer purity that we were using. We first compared the 12dT-CD8.A3t-coupled resin to the previous CD8.A3t-coupled resin and found that while the new linker slightly reduced CD8⁺ purity by 8%, it improved yield by 27% (Figure 2C). Thus, we moved forward and ran a concentration titration experiment with the 12dT-CD8.A3t aptamer (from 0.78 to 25 nmol/g resin). Overall, CD8⁺ cell capture improved with the 12-dT-CD8.A3t aptamer compared to the initial aptamer. Interestingly, lower concentrations of aptamer depleted fewer CD8⁺ cells but resulted in higher CD8⁺ purities and yields in the elution fraction (Figure 2D). This may suggest that higher concentrations of immobilized aptamer bound to CD8⁺ T cells with increased avidity, making the mechanical elution of cells via flushing more difficult. Lönne et al. also showed that higher aptamer densities impaired binding for aptamer-based affinity purification of vascular endothelial growth factor, likely from disrupted aptamer folding or restricted steric accessibility of immobilized aptamers.³⁷ We thus moved forward with the smallest 0.78 nmol/g resin concentration. Furthermore, we optimized resin bed volume to better balance purity and yield. Larger resin volumes could result in more nonspecific binding of PBMCs due to increased contact time during gravity flow. We ran pre-cleared PBMCs through resin volumes of 0.3, 0.4, and 0.5 mL and found the highest CD8⁺ purity with the 0.3 mL resin volume, although yield was slightly reduced compared to the larger resin volumes (Figure 2E).

Multiple chromatographic separation steps are often needed to achieve sufficient purity of biological molecules.^54–56 As a final step to further improve purity, we developed a double purification platform with the optimized 12dT-CD8.A3t aptamer concentration and resin bed volume for CD8⁺ T cell isolation (Figure 3A). In this method, PBMCs pre-cleared on uncoupled resin were run through 12dT-CD8.A3t aptamer-coupled resin in a larger Econo-Pac column. Then, the CD8⁺-enriched flush fraction was run sequentially on another round of 12dT-CD8.A3t aptamer-coupled resin in a smaller Poly-Prep Column using 0.3 mL of resin, which previously was found to improve purity. The additional purification step boosted the CD8⁺ purity from the first step by 22.5% to a final purity of 85.4% (±5.7%), thus minimizing contamination of undesired cell populations (Figure 3B,C and Table S5). This significant improvement in purity was accompanied by a small reduction in yield of CD8⁺ cells, from 64.2% (±9.1%) in the first purification to 48.3% (±7.1%) after the second purification. The success of this double purification approach highlights the traceless nature of the aptamer-based chromatography system, allowing cells to undergo multiple positive selections in sequence. While we use multiple purifications here with the same aptamer to increase target cell purity, different aptamer-coupled resins could theoretically be applied in sequence to isolate more antigenically complex cell subsets, such as memory T cells that have been shown to be beneficial for CAR T-cell therapy.^54,57

Figure 3.

Final 12dT-CD8.A3t-mediated resin chromatography results for selection of CD8⁺ T cells. A) CD8⁺ selection scheme where thawed PBMCs are first run on unfunctionalized resin, resulting in pre-cleared PBMCs (PC PBMCs) in the flowthrough that are subsequently run in CD8.A3t-coupled resin in the first round (R1). The eluted flush from this round is finally run through a second round (R2) of CD8.A3t-coupled resin. B) Representative flow cytometry histograms of CD8⁺ expression in each round of isolation. C) Final CD8⁺ purity and yield data in samples from cell isolation experiments. The circles, squares, and triangles represent different donors from three separate isolation experiments. Horizontal lines and error bars represent the mean ± s.d.; **p<0.01, ***p<0.001, ****p<0.0001, ns not significant (one-way ANOVA test).

CONCLUSION

Here, we report a proof-of-concept use of aptamers for cell depletion and selection in a continuous-flow affinity chromatography system. Epoxidized polystyrene resin with antigen-specific aptamers composed the affinity matrix to bind target cells for depletion and selection. Light mechanical disruption by pipetting the resin eluted the bound cells. Using the Mono.A2 aptamer for monocyte depletion, we removed >94% of monocytes from PBMCs. Using the CD8.A3t aptamer for CD8⁺ T cell selection, we achieved >85% purity and >48% selection yield of CD8⁺ T cells. A >72% CD8⁺ T cell purity from cell harvest contributed to a 93% remission rate in patients with acute lymphoblastic leukemia.⁵⁸ This resin-based system reduces the costs of expensive affinity reagents, such as antibodies and magnetic supports, with the advantage of simplified procedures, faster processing times, traceless isolation, and high stability. This novel approach can be extended in the future to take advantage of reversible strand displacement for aptamer affinity reagents. Adaptation of these reversible agents will allow for more selective elution of aptamer-bound cells and isolation of multiple cell populations in one column, further improving the overall purity and throughput of this system. Aptamers can thus be successfully applied in manufacturing strategies to prepare engineered T cells through continuous flow methods to decrease cost and increase accessibility of T cell immunotherapies.