Enrichment of human Vγ9Vδ2 T lymphocytes by magnetic poly(divinylbenzene-co-glycidyl methacrylate) colloidal particles conjugated with specific antibody

γδ T cells play a significant role in protection against cancer. Purification of γδ T cells is needed for insight when studying their anti-cancer functionality and their utilization in adoptive cell therapy. To improve the purification of γδ T cells, in this work, a composite material based on magnetic nanoparticles was developed for purification of Vγ9Vδ2 T cells, the predominant subset of γδ T lymphocytes in human peripheral blood. The epoxy-functionalized magnetic poly(divinylbenzene-co-glycidyl methacrylate) particles (mPDGs) were bio-conjugated with anti-human Vδ2 antibody to provide specific recognition sites for T cell receptors of Vγ9Vδ2 T cells. Using fluorescence-activated cell sorting (FACS) analysis, separation of Vγ9Vδ2 T cells from peripheral blood mononuclear cells of healthy donors was confirmed with high purity [89.77% (range 87.00–91.80, n = 3)]. More interestingly, the immobilized particles did not affect the viability of purified cells as high cell viability was indicated (>90%). By combining the properties of magnetic nanoparticles with specific antibodies, these immobilized particles were shown to be used as a cell-friendly purification tool of Vγ9Vδ2 T lymphocytes without any limits for the further use of cells. The purified Vγ9Vδ2 T cells using the antibody-immobilized epoxy-functionalized mPDGs could be used directly without a detachment step for further cultivation and expansion. This highlights the advantages of this method in allowing the study of cell function and further investigation of such rare T cell populations in immunotherapy.


Introduction
T lymphocytes are classied into two subsets, ab T cells and gd T cells, based on their T cell receptors (TCRs). These receptors have been characterized using monoclonal anti-TCR antibodies (mAb) and methods to identify gene rearrangement of cytotoxic T lymphocytes. [1][2][3][4] ab T cells play an important role in adaptive immune responses, mostly in regard to their peptide antigen recognition in the context of major histocompatibility complex (MHC) class II or class I molecules and in the production of effector molecules. 5 gd T cells are unconventional T lymphocytes that express TCRs consisting of heterodimeric g and d chain complexes for antigen recognition. 6 In normal human peripheral blood, most gd T cells are dened by their expression of the variable domains Vg9 and Vd2 of the TCR and are referred to as Vg9Vd2 T cells, and represent only 1-10% of total T cells. 7 These cells show unique features with antigen recognition, which is unrestricted to MHC molecules, and TCR gene usage. 8 As a result, Vg9Vd2 T cells act as immune surveillance cells which respond rapidly with respect to pathogens, stressed-or infected-cells and cytokines. 9 Because of their capacity to inltrate tumors and then express cytotoxic activity, these cells are also involved in anti-cancer immune responses. 10,11 Moreover, Vg9Vd2 T cells also have been shown to display the principal characteristics of professional antigenpresenting cells. 12 Based on the attractive characteristics of Vg9Vd2 T cells, several clinical trials were initiated to purify these cells from peripheral blood for further study. In addition, the culture of Vg9Vd2 T cells was developed in order to obtain high numbers of cells for adoptive cell transfer in cancer treatment. 13 Utilizing their nonpeptide, phosphoantigen recognition ability, 14 this cell population can be expanded in vitro using phosphoantigens such as bromohydrin pyrophosphate (BrHPP) or nitrogen-containing bisphosphonates (N-BPs), and can be puried for use in clinical treatments. 15,16 Recent clinical trials have included approaches where magnetic labeling systems and column isolation were used for selection of puried TCR gdexpressing T cells. By using separation kits, the purity of these cells is more than 90% aer separation from peripheral blood mononuclear cells (PBMCs). [17][18][19] However, the process is complex and time-consuming. Moreover, long incubation time may lead to non-specic cell labeling. To address this problem various functionalized magnetic polymer nanoparticles, such as magnetic nanoparticles modied with hydrazine functionalized polymer, uorescent chitosan functionalized magnetic polymeric nanoparticles and epoxy-functionalized magnetic poly(divinylbenzene-co-glycidyl methacrylate) particles (mPDGs), were utilized. [20][21][22] By reason of their properties, along with highly specic surface areas and versatile surface functionality, these particles can be used widely as carriers for biomolecules including proteins, antibodies and antigens. [23][24][25] Moreover, a great advantage of using magnetic nanoparticles is the rapidity of separation upon applying an external magnetic eld, a one-step process which can avoid many of the timeconsuming steps of other separation processes. 26 With the usefulness of mPDGs, there is current report with respect to using the immobilized mPDGs with monoclonal anti-human IL-10 antibody to obtain specic and selective recognition sites for the recombinant human IL-10 protein in an immunoassay. 25 In addition, these mPDGs were successfully applied for purication of CD4 + T lymphocytes suggesting that mPDGs could be a good candidate for Vg9Vd2 T cell separation. 27 In this study, in order to improve the efficiency of Vg9Vd2 T cells isolation, we aimed to investigate the properties of epoxyfunctionalized mPDGs intended for use in the in vitro separation and purication of Vg9Vd2 T cells. The immobilized magnetic particles were rst bioconjugated with monoclonal anti-human Vd2 antibody to provide specic recognition of the TCR Vd2 presented on Vg9Vd2 T cells. To determine percentage of purity and viability of Vg9Vd2 T cells, conjugation with uorescence antibodies which recognized specic surface markers of Vg9Vd2 T cells was performed followed by uorescence-activated cell sorting (FACS) analysis.

Epoxy-functionalized mPDGs preparation
The epoxy-functionalized mPDGs were synthesized following a previously published method. 25 Briey, seed emulsion copolymerization of divinylbenzene (DVB, 1 mL) (Sigma, USA) and glycidyl methacrylate (GMA, 0.1 mL) (Sigma, USA) monomers was conducted using potassium persulfate (KPS, 0.05 g) (Sigma, USA) as an initiator and in the presence of O/W magnetic emulsion droplets (1.4% w/v) as a seed (kindly provided by Mohamed M. Eissa from University Lyon-1, France). The polymerization reaction was carried out at 70 C while stirring (300 rpm) and under N 2 atmosphere for 24 h. The obtained mPDGs were ltered using glass wool bers to remove any coarse particles before further analyses.
Hydrodynamic size and zeta potential of the prepared particles were measured using Zetasizer (Malvern, Nano ZS2000). In addition, a transmission electron microscope (TEM) (Phillips, CM120) and vibrating sample magnetometer (NETZSCH, TG209) were utilized for studies of morphological and magnetic content.

Antibody immobilization onto epoxy-functionalized mPDGs
The specic anti-human Vd2 antibody (mouse IgG1k, clone B6; BioLegend, USA) was immobilized onto the surface of epoxyfunctionalized mPDGs. In brief, 1 mg of epoxy-functionalized mPDGs (2.2% w/v) were washed twice with PBS buffer (pH 7.4, 500 mL). Aer that, the puried anti-human Vd2 antibody (20 mg mL À1 ) was added to the particles and incubated at room temperature for 20 min using a ThermoMixer (Eppendorf, Germany). Aer centrifugation, the supernatant was collected to determine the residual concentration of antibody at 260 nm using a UV-Vis spectrophotometer (Thermo Scientic, USA). The binding efficiency was calculated using the following equations.
where C i and C f (mg mL À1 ) are the initial and nal concentrations of anti-human Vd2 antibody. The antibody-immobilized particles were washed twice with PBS buffer (pH 7.4) and the supernatant was nally discarded. The non-specic binding of contaminants was blocked with PBS containing bovine serum albumin (0.1% w/v) and sodium azide (0.05% w/v). The immobilized particles were stored at 4 C until used.

Preparation of peripheral blood mononuclear cells (PBMCs)
Whole blood from healthy volunteers was collected at the blood bank unit of Phramongkutklao Hospital. Buffy coats from whole blood (50-70 mL) were collected in blood bags containing anticoagulant for further isolation of PBMCs. Collected buffy coats were overlaid on Lymphoprep™ solution (Axis Shield PoC, Oslo, Norway) at a 1 : 1 ratio by volume and centrifuged (Hettich Zentrifugen, Germany) at 2000 rpm and 20 C for 30 min. The interface layer containing the PBMCs was collected and washed twice with cold incomplete RPMI 1640 medium (Gibco, USA) containing 10 000 U mL À1 penicillin-streptomycin (Gibco, USA). The PBMCs were suspended in cold PBS (100 mL) and kept at 4 C prior to use. Viability of isolated cells was assessed using the trypan blue exclusion method. Obtaining all buffy coats from human donations at the blood bank of Phramongkutklao Hospital followed an informed consent process approved by Mahidol University Ethics Committee, Mahidol University, Bangkok, Thailand (MURA2014/400).

Generation of human Vg9Vd2 T cells from peripheral blood mononuclear cells
To cultivate human Vg9Vd2 T lymphocytes, total PBMCs isolated from buffy coats by density gradient centrifugation using Lymphoprep™ were resuspended in RPMI 1640 supplemented with FBS (10% w/v) and penicillin-streptomycin (10 000 U mL À1 ) in the presence of pamidronate (10 mM) and recombinant human IL-2 (50 IU mL À1 nal concentration). The PBMCs were plated on 24-well plates with addition of IL-2 every 3 days. Aer 12 days of culture, the cells were examined to determine their percentage and viability by ow cytometric analysis (BD FACS-CANTO II, BD Biosciences, USA) and trypan blue exclusion, respectively. The cells in cultures containing high percentages of Vg9Vd2 T cells were harvested and suspended in cold PBS (100 mL) and kept at 4 C before starting isolation experiments.
Separation of human Vg9Vd2 T cells by antibody-immobilized epoxy-functionalized mPDGs from peripheral blood PBMC samples from healthy donors were processed to separate Vg9Vd2 T cells using antibody-immobilized epoxyfunctionalized mPDGs. The separation conditions were optimized by varying the amount of antibody-immobilized particles (20, 100 or 200 mg) and the amount of PBMCs (1 Â 10 6 , 5 Â 10 6 or 10 Â 10 6 cells). Briey, PBMCs at a certain concentration in cold PBS (100 mL) were mixed with 20, 100 or 200 mg of antibodyimmobilized epoxy-functionalized mPDGs. The mixtures were incubated at 4 C for 15, 30 or 60 min, gently stirred, and then complexes were separated by applying a magnet. The percentage and viability of Vg9Vd2 T cells were determined using FACS analysis and trypan blue exclusion, respectively.

Separation of human Vg9Vd2 T cells by antibody-immobilized mPDGs from Vg9Vd2 T cells cultivation system
The optimal conditions determined with PBMCs and antibodyimmobilized mPDGs were used to separate Vg9Vd2 T cells from cell cultures. The percentages of Vg9Vd2 T cells were investigated by FACS analysis and viabilities were determined by trypan blue exclusion. To compare the result, separation of Vg9Vd2 T cells was also done by negative selection using TCRg/ d + T Cell Isolation Kit and immunomagnetic sorting (Miltenyi Biotec, Germany) according to the manufacturer's instructions.

Statistical analyses
Statistical signicance was determined by one-way analysis of variance (ANOVA). One-way ANOVA with Bonferroni's multiple comparison test was used to assess the signicance of differences between varying amounts of antibody-immobilized epoxyfunctionalized mPDGs and incubation times. One-way ANOVA with Dunnett's multiple comparison test was used to compare between groups and controls. In addition, a two-tailed, paired t test was used to assess the signicance of differences between groups. All of the tests utilized GraphPad Prism version 5.01 (GraphPad Soware). In bar graphs, results are expressed as mean AE SEM. P values < 0.05 were considered statistically signicant.

Characterizations of mPDGs
Reactive epoxy-functionalized mPDGs were successfully developed following the published method of Eissa and colleagues. 25 Seed emulsion co-polymerization of DVB and GMA monomers in the presence of Fe 3 O 4 magnetic emulsion droplets as the seed was elaborated. The hydrodynamic size of the prepared functionalized mPDGs was 488 nm, zeta potential at pH 7.3 was À59 mV and the magnetic content was 67% wt. From TEM, the obtained mPDGs exhibited a dark magnetic core and well-dened polymer shell (as shown in Fig. 1).
To date, interest in magnetic nanoparticles or functional polymer hybrid materials has increased greatly within the biomedical eld. Functionalized magnetic polymer nanoparticles can be used not only for in vivo therapeutic applications, 28 but also for in vitro applications including the separation and purication of biomolecules such as proteins and nucleic acids. [29][30][31][32] Their surface functionality make them suitable for wide use as carriers of antibodies which detect specic antigens. 33,34 There are two approaches used for bioconjugated between the surface of the nanoparticle and the antibody, physical adsorption and direct covalent linkage. 35 The covalent binding is currently preferred because the strength of interaction between biomolecule (i.e., chemical properties of the antibodies) and the reactive particle (i.e., surface dispersion) can be manipulated. Moreover, such interaction can prevent an unexpected competitive displacement of the adsorbed antibodies by other blood components. 36 The major advantage of This journal is © The Royal Society of Chemistry 2018 using magnetic polymeric nanoparticles is their easy manipulation by applying a magnet to the material and the resulting rapidity of separation. For these particles, magnetite (Fe 3 O 4 ) is most frequently used as the magnetic core. This iron oxide is superparamagnetic, easily prepared, and biocompatible when utilized in biomedical applications. 37 Immobilization of anti-human Vd2 antibody onto epoxyfunctionalized mPDGs Anti-human Vd2 antibody was covalently bound to the epoxyfunctionalized mPDGs by using appropriate conditions with no need of a coupling agent. The binding efficiency of antibodyimmobilized mPDGs of 70.97% (range 58.35-96.65, n ¼ 5) was calculated by using eqn (1). The results, therefore, indicated that most anti-human Vd2 antibody was immobilized on the particle surface.
As mentioned previously, we chose the GMA monomer as a precursor to produce reactive functional polymers. GMA contains a polymerizable double bond and a reactive epoxy group allowing it to be chemically graed with any biomolecule which contains reactive functional groups without using a coupling agent. 24 This enabled us to covalently attach the selected antibodies via their functional groups to the particle surface. Our result is in line with previous reports. 25,38 The reactive epoxy-functionalized mPDGs were easily bioconjugated with the anti-human Vd2 antibody and the covalent binding process leads to a strong conjugation which is suitable for various types of biologic application. Moreover, the binding efficiency can be further elevated by varying the amount of antibodies or the incubation time of the interaction.

Separation of human Vg9Vd2 T cells by antibody-immobilized epoxy-functionalized mPDGs
The antibody-immobilized epoxy-functionalized mPDGs were used for separation of human Vg9Vd2 T cells from peripheral blood of healthy donors. In order to determine optimal conditions, the amount of antibody-immobilized mPDGs (20, 100 or 200 mg), number of PBMCs (1 Â 10 6 , 5 Â 10 6 or 10 Â 10 6 cells), and incubation time (15,30 or 60 min) were varied. It was found that the purity of separation was signicantly higher when there were 10 Â 10 6 PBMCs suspended, irrespective of the amount of antibody-immobilized particles, when compared to the lower numbers of PBMCs (as shown in Fig. 2).
It was noticed that at 10 Â 10 6 PBMCs, the highest purity (mean ¼ 85.60%) of Vg9Vd2 T cells was obtained with the use of 200 mg of antibody-immobilized epoxy-functionalized mPDGs (p < 0.001). Decreases in the amount of the particles (20 and 100 mg) were associated with the purity of the separated Vg9Vd2 T cells, all less than 85.60% (as shown in Fig. 2).
Regarding variation of incubation time, it was shown that at 30 min of co-incubation the separation reaction of Vg9Vd2 T cells was started and that the purity of the separated Vg9Vd2 T cells was signicantly greater (p ¼ 0.0014) than that of controls (mean ¼ 66.97%) (Fig. 3). The results also showed that there was no signicant difference in the purity of separation between incubation times of 30 and 60 min, although the highest purity was seen at 60 min. Of note, the anti-human Vd2 antibody used possessed high specicity for Vg9Vd2 T cells. The optimal conditions for separation of human Vg9Vd2 T cells by immobilized epoxy-functionalized mPDG was the use of 10 Â 10 6 PBMCs with 200 mg antibody-immobilized mPDGs and 30 min of incubation time.
The efficiency of mPDGs for in vitro cell separation was investigated by isolating the rare Vg9Vd2 T cells. This subset of T cells has raised interest from researchers in the last decade because of their distinctive surface molecules and the antigenbinding characteristics leading to their immune-patrolling capacity. Their effector function (initiation of immune responses) upon infectious pathogens is described elsewhere. 7,9 Normally, Vg9Vd2 T cells are a minor population in human peripheral blood representing only 1-10% of the total T cells. 7 Several trials have attempted to achieve pure populations of these cells. The separation techniques commonly applied to this small cell population include both positive and negative selection. Both selection methods are based on magnetic microbead labeling and column separation using different antibody labeling of target cells. For positive selection, the target cells are obtained through the direct binding of anti-TCR g/d hapten-antibody labelling and anti-hapten microbeads along with the magnetic isolation. 19 This method has proved successful for many researchers and oen yields greater than 90% purity of the Vg9Vd2 T cells. However, these separation processes are time-consuming and confounded by contamination with dead cells and non-labeled cells caused by long incubation times.
To overcome this, we generated a simple protocol for Vg9Vd2 T cell purication. This is the rst evidence showing that antibody-immobilized epoxy-functionalized mPDGs can be used for separation of the rare Vg9Vd2 T cell population from PBMCs with high levels of purity. Our protocol includes few steps and allows easy and rapid specic separation of the targeted cells. Furthermore, under the effect of an external magnetic eld, this present method was feasible and rapid.

Separation of human Vg9Vd2 T cells by antibody-immobilized epoxy-functionalized mPDGs from Vg9Vd2 T-cell cultivation system
In cancer immunotherapy using adoptive T cell transfer, Vg9Vd2 T cells are a critical factor in obtaining favorable results because of their specic binding to small non-peptide antigens derived from various cancer cell lines. 10 A protocol to obtain pure populations and heighten the yield of Vg9Vd2 T cells is important for achieving the best outcomes with gd T-cell immunotherapy. Synthetic phosphoantigens are utilized in cell stimulation to gain high amounts of gd T cells, together with purication steps as shown previously. 39,40 Seeking to improve the outcome of the separation step, the antibodyimmobilized epoxy-functionalized mPDGs were also used for separation of human Vg9Vd2 T cells from in vitro Vg9Vd2 T-cell cultures. The efficiency of the separation of Vg9Vd2 T cells was evaluated by ow cytometry. The results from FACS analyses revealed that the purity of Vg9Vd2 T cells was 67.88% (as shown in Fig. 4).
However, the amount of obtained Vg9Vd2 T cells did not shi signicantly indicating that the antibody-immobilized particles could be used for both cell sources, but that   a greater purication might be reached when used with peripheral blood as shown in our optimal conditions determination, which provide evidence that the obtained amount of human Vg9Vd2 T cells could be reached to 89.77% performed by using PBMC samples (Fig. 3). Hence, to increase purity, the optimization using antibody-immobilized epoxy-functionalized mPDGs for further used in Vg9Vd2 T cell separation from cultivation source may be important.
Regarding cell viability aer separation, results from the trypan blue exclusion tests are shown in Fig. 5. It was found that antibody-immobilized epoxy-functionalized mPDGs did not alter the viability of separated cells, indicating that this protocol was useful for the purication of Vg9Vd2 T cells and would allow further study of cellular function. Furthermore, mPDGs which are the synthetic nanoparticles with the presence of magnetite (Fe 3 O 4 ) as a core and polymer shell are preferred in biomedical applications because they are biocompatible 25 and nontoxic as shown in the result. Therefore, it may be suggested that mPDGs do not have to be removed for downstream application such as cell culture.
Regarding the small size of mPDGs, they could be phagocytosed by antigen presenting cells including macrophages and dendritic cells which are crucial cells for adaptive immune system activation by phagocytosis. 41,42 B cells, a type of lymphocytes, constitute only 1-7% in peripheral blood also function as antigen presenting cells. 43,44 This did not effect the T cells separation as the high purity (mean ¼ 85.60%) of Vg9Vd2 T cells was found with the use of highest amount of mPDGs (Fig. 2).
We showed in this study that in-house, antibodyimmobilized epoxy-functionalized mPDGs could be used to purify Vg9Vd2 T cells from an in vitro cultivation system using pamidronate, a specic activator for expansion of this T cell subtype. 45 However, a few contaminants from other cell populations occurred during the purication, suggesting the importance of further renement of the protocol. But most noteworthy, these antibody-immobilized epoxy-functionalized mPDGs were nontoxic to the cells which was also conrmed by previous study. 27 This developed isolation tool is expected to be employed for providing the ease way in adoptive T cell transfer using Vg9Vd2 T cells which imparts a great benet to the immunological and immunotherapy studies of Vg9Vd2 T cells as important immune effector cells.

Conclusions
A method for Vg9Vd2 T cell separation by using puried human Vd2 antibody immobilized mPDGs (antibody-immobilized mPDGs) was developed as a useful tool in various research elds (as shown in Fig. 6). The specic antibody was immobilized onto epoxy-functionalized mPDGs. When co-incubation with PBMCs or in vitro Vg9Vd2 T cell cultures, the interaction between monoclonal antibody-immobilized particles and Vd2 T cell receptor (TCR) of Vg9Vd2 T cells could be performed, allowing excellent separation. Due to the ease of processing and short incubation time, the use of antibody-immobilized epoxyfunctionalized mPDGs appears attractive for the preparation of Vg9Vd2 T cells. The advantage of using specic antibody was to separate the Vg9Vd2 T cells with good viability, allowing further functional studies which could help to better understand their roles in cancer biology and immunotherapy. This technique could potentially be used in various research studies and could be adapted for the separation of other cells of interest.