Magnetic Separation of Melanoma-Specific Cytotoxic T Lymphocytes

Jan 14, 2009 - Target antigen-specific cytotoxic T lymphocytes (CTLs) play a key role in anticancer and antivirus immunity in the body, and purificati...
0 downloads 0 Views 542KB Size
Download PDF
304

Bioconjugate Chem. 2009, 20, 304–309

Magnetic Separation of Melanoma-Specific Cytotoxic T Lymphocytes from a Vaccinated Melanoma Patient’s Blood Using MHC/Peptide Complex-Conjugated Bacterial Magnetic Particles Masayuki Takahashi, Yasuto Akiyama, Junpei Ikezumi, Takeshi Nagata, Tomoko Yoshino, Akira Iizuka, Ken Yamaguchi, and Tadashi Matsunaga* Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16, Naka-cho, Koganei, Tokyo 184-8588, Japan, and Immunotherapy Division, Shizuoka Cancer Center Research Institute, 1007, Shimo-nagakubo, Nagaizumi-cho, Sunto-gun, Shizuoka, 411-8777, Japan. Received September 17, 2008; Revised Manuscript Received November 26, 2008

Target antigen-specific cytotoxic T lymphocytes (CTLs) play a key role in anticancer and antivirus immunity in the body, and purification of CTLs from heterogeneous immune cells is desired for an efficient cancer immunotherapy and fundamental research. Herein, a novel magnetic nanoparticle conjugated with major histocompatibility complex (MHC)/peptide complexes was developed for the magnetic separation of melanomaspecific CTLs. To conjugate biotinylated MHC/peptide complexes on nanosized bacterial magnetic particles (BacMPs), which were synthesized intracellularly by magnetotactic bacteria, phased modification of biotin and streptavidin onto BacMPs was investigated. When biotin was modified on BacMPs, a polyethylene oxide (PEO)linker contributed to the maintenance of the high dispersion properties of BacMPs. Furthermore, nonspecific binding of BacMPs to cell surface was prevented by controlling the level of streptavidin bound on BacMPs and through PEO blocking of the empty streptavidin sites on BacMPs. Finally, single-step magnetic separation of melanoma-specific CTLs was demonstrated using developed MHC/MAGE-1 A24 peptide complex-conjugated BacMPs. Melanoma-reactive cells and melanoma-specific CTLs were successfully separated from stimulated peripheral blood mononuclear cells derived from a vaccinated melanoma patient with 93.1% and 87.7% purity, respectively, and specificity of antigen recognition and cytokine secretion from separated CTLs were confirmed. The potential of MHC/peptide complex-conjugated BacMPs was indicated for efficient separation of antigenspecific CTLs in cancer immunotherapy and fundamental research.

INTRODUCTION Cytotoxic T lymphocytes (CTLs) become activated when, through T cell receptors (TCRs), they recognize major histocompatibility complex (MHC)/peptide complexes that are present on the surface of cancer cells and other infected cells. Once activated, they clonally expand and trigger a cytotoxic response against these cells. CTLs play a key role in anticancer and antivirus immunity. Single CTLs, which express individual TCRs, recognize a pair of antigenic peptides and polymorphic helices of MHC proteins with one-to-one correspondence (1, 2). For effective immunotherapy using CTLs, target antigen-specific CTLs must be purified from various cell mixtures, such as those in blood samples, and significant demand persists for the development of efficient CTL separation technology. Separation of target cells from a heterogeneous cell mixture is used widely in biological and medical applications. Among current cell separation techniques, magnetic cell sorting is a popular tool for isolating cells such as monocytes, B-lymphocytes, and hematopoietic stem cells (CD14, CD19, and CD34 positive cells, respectively) from peripheral blood (3-6). The use of antibody-conjugated magnetic particles (immunomagnetic particles) is a proven, rapid, and simple technique relative to fluorescence-activated cell sorting (FACS) (7). Antibodies conjugated on magnetic particles usually recognize antigens on the surface of target cells (e.g., CD markers), and cells labeled with magnetic particles are successfully separated from unlabeled cells. However, for the separation of antigen-specific * Direct fax: +81-42-385-7713; phone: +81-42-388-7020; e-mail: [email protected].

CTLs, immunomagnetic particles are not effective, and alternative magnetic particles are required. For the separation of antigen-specific CTLs from cell mixtures, tetrameric MHC/ peptide complexes, which combine purified and biotinylated single MHC/peptide complexes with PE-labeled streptavidin (MHC/peptide tetramer), have been used to label target cells that are then separated by flow cytometric sorting (8-10). However, improvements in throughput and absolute sterility of the flow cytometer are required, and thus, a magnetic separation system that uses magnetic particles to recognize antigen-specific CTLs was considered suitable for the separation of CTLs. The magnetotactic bacterium synthesizes intracellular nanosized bacterial magnetic particles (BacMPs) of 50 to 100 nm in diameter that are covered with a stable lipid bilayer. BacMPs, which have a single magnetic domain of magnetite and exhibit strong ferrimagnetisms, can be collected easily with commercially available permanent magnets. The lipid membrane provides BacMPs with a high surface area and good dispersion properties, creating various applications for BacMPs. Magnetic cellseparationusingBacMPshasbeenreportedpreviously(11-14), and many cell types were separated with high purity from peripheral blood mononuclear cells (PBMCs) or whole blood directly using BacMPs displaying protein A or protein G, which are IgG binding proteins, conjugated to antibodies. Furthermore, this system is not only simple but also noninvasive to the separated cells, because BacMPs are collectable using available permanent magnets rather than the special magnetic columns needed for the collection of commercialized magnetic nanoparticles. Therefore, we hypothesized that the magnetic cell separation system using BacMPs will be suitable for efficient

10.1021/bc800398d CCC: $40.75  2009 American Chemical Society Published on Web 01/14/2009

MHC/Peptide Complex-Conjugated BacMPs

Bioconjugate Chem., Vol. 20, No. 2, 2009 305

Figure 1. Schematic diagram for preparation of MHC/MAGE-1 peptide complex-conjugated BacMPs.

separation of antigen-specific CTLs, furthering the application of CTLs in immunotherapy. In this report, novel magnetic particles were developed for the separation of melanoma-specific CTLs. First, MHC/ MAGE-1 A24 peptide complexes that are presented on the surface of melanoma cells were conjugated on BacMPs by a chemical procedure. The prevention of nonspecific binding of BacMPs to the cell surface was considered critical in this process. Furthermore, antibody-independent magnetic separation of melanoma-specific CTLs from in vitro stimulated PBMCs derived from a vaccinated melanoma patient was performed using the MHC/MAGE-1 A24 peptide complex-conjugated BacMPs, and the function of separated CTLs was confirmed.

EXPERIMENTAL PROCEDURES Cultivation of Magnetic Bacteria. M. magneticum AMB-1 was microaerobically cultured in magnetic spirillum growth medium at 28 °C as previously described (15). Microaerobic conditions were established by sparging the cultures with argon gas. Batch cultures were prepared in 10 L media bottles. AMB-1 transformants were cultured under these conditions in 5 µg/mL ampicillin. Preparation of BacMPs. BacMPs were purified from AMB1, according to the methods of Tanaka and Matsunaga (16). Cultured cells were collected by centrifugation, suspended in 40 mL of HEPES (10 mM) buffer (pH 7.4), and disrupted by three passes through a French press cell at 1500 kg/cm2 (Ohtake Works Co. Ltd., Tokyo, Japan). BacMPs were collected from cell debris using a neodymium-iron-boron (Nd-Fe-B) magnet. Collected BacMPs were washed with HEPES buffer at least 10 times. During each wash, BacMPs were dispersed by ultrasonication and collected using the Nd-Fe-B magnet. The concentration of BacMPs in suspension was estimated by optical density at 660 nm (OD660) using a spectrophotometer (UV-2200, Shimadzu, Kyoto, Japan). Purified BacMPs were stored at 4 °C in phosphate buffered saline (PBS) solution. Immobilization of Streptavidin on BacMPs. BacMPs (5 mg) were incubated with 1 mM Sulfo-NHS-LC2-Biotin (Pierce, Rockford, IL, USA) or 1 mM Sulfo-NHS-PEO12-Biotin (Pierce) in 5 mL of borate buffer (pH 8.0) for 1 h at room temperature with pulsed sonication every 3 min (Figure 1-I). After incubation, biotin-modified BacMPs were magnetically separated from the reaction mixture using the Nd-Fe-B magnet and washed with PBS buffer containing 0.1% (v/v) Tween 20 (one time) and then PBS buffer (two times). BacMPs (3 mg) were then incubated with 0.125-1 µM Alexa488-labeled streptavidin (Invitrogen, Carlsbad, CA, USA) in 3 mL of PBS buffer for 1 h at room temperature with pulsed sonication (Figure 1-II). Streptavidin-immobilized BacMPs (SA-BacMPs) were washed three times with PBS buffer using the Nd-Fe-B magnet to remove excess non-immobilized streptavidin and suspended in

PBS buffer containing 0.1% sodium azide before use. In the three washing steps, all supernatants were collected, and the fluorescence intensity of each solution was measured using a microplate reader at an excitation of 485 nm and an emission of 520 nm. The amount of Alexa488-labeled streptavidin that was not immobilized on biotin-modified BacMPs was calculated from the fluorescence intensity attained from a calibration curve, and streptavidin immobilized on biotin-modified BacMPs was then calculated by deducing the amount of non-immobilized streptavidin from initially added streptavidin. The number of streptavidin immobilized on a single BacMP was calculated from the amount of streptavidin immobilized on biotin-modified BacMPs (BacMP diameter ) 75 nm and BacMP density ) 5.2 g/cm3). Analysis of Size Distribution of SA-BacMPs. The size distributions of SA-BacMPs biotinylated with Sulfo-NHS-LC2Biotin or Sulfo-NHS-PEO12-Biotin were analyzed using an electrophoretic light scattering spectrophotometer (ELS-8000, Otsuka Electronics, Osaka, Japan). SA-BacMPs were added to 3 mL of deionized distilled water to achieve a final concentration of 10 µg/mL. Solutions were sonicated prior to measurement. Conjugation of MHC/Peptide Complexes to SA-BacMPs. SA-BacMPs (300 µg) were incubated with 0.75 µM biotinylated MHC/MAGE-1 A24 peptide (NYKHCFPEI) complexes (MBL, Nagano, Japan) or biotinylated MHC/HIV A24 peptide (RYLRDQQLL) complexes (MBL) in 600 µL of TNE buffer (10 mM Tris, 150 mM NaCl, 0.5 mM EDTA, and 0.1% sodium azide; pH 8.0) for 30 min at room temperature with pulsed sonication every 10 min (Figure 1-III). After three washes in PBS buffer, Amine-PEO3-Biotin (10 mM, Pierce) in PBS buffer was added to MHC/peptide complex-conjugated BacMPs (500 µg/mL) to block empty streptavidin sites on BacMPs (Figure 1-IV). After incubation for 30 min at room temperature, MHC/peptide complex-conjugated BacMPs with PEO blocking were washed magnetically and preserved in PBS buffer containing 0.1% sodium azide until use. The average number of MHC/peptide complexes conjugated on a single SA-BacMP was evaluated using alkaline phosphatase (ALP)-labeled antibodies. MHC/peptide complexes conjugated on SA-BacMPs (20 µg) were incubated with mouse IgG2a antiβ2 microglobulin monoclonal antibodies (40 µL sample, 30 µg/ mL, Immunotech, Marseille, France) for 60 min at room temperature. After three washes, BacMPs were incubated with ALP-labeled antimouse IgG (20 µL sample, 10 µg/mL, Santa Cruz Biotechnology Inc., Santa Cruz, CA, USA) for 30 min at room temperature. After three washes, the resuspended complexes (20 µL sample, 100 µg of BacMPs/mL) were mixed with 80 mL of Lumi-Phos 530 including Lumigen PPD {4-methoxy4[3-phosphatephenyl]spirio[1,2-dioxetane-3,2′-adamantane] disodium salt} (3.3 × 10-4 M) as a luminescence substrate for alkaline phosphatase, and the resulting luminescence intensity

306 Bioconjugate Chem., Vol. 20, No. 2, 2009

was measured by a Lucy-2 luminescence reader (Aloka, Co. Ltd., Tokyo, Japan). The numbers of ALP-labeled IgG (molecular weight of F(ab)2 IgG ) 100 kDa) were calculated from the luminescence intensity attained from a calibration curve. The values for the number of ALP-labeled IgGs were used to attain the number of MHC/peptide complexes conjugated on a single SA-BacMP. Evaluation of Nonspecific Binding of BacMPs to Cell Surfaces. Raji cells (2 × 105 cells) in 50 µL of separation buffer (PBS buffer containing 0.5% bovine serum albumin (BSA) and 2 mM EDTA) were mixed with 10 µg of SA-BacMPs, MHC/ MAGE-1 A24 peptide complex-conjugated BacMPs, or MHC/ MAGE-1 A24 peptide complex-conjugated BacMPs with PEO blocking for 10 min at 4 °C. After washing by centrifugation, the cell suspension (1 mL) was transferred to a test tube (10 mm × 75 mm) and magnetically collected using the Nd-Fe-B magnet for 5 min. The dimensions of magnets used for cell separation were 20 × 12 × 5 mm3. The surface magnetic flux density of the magnets was 400 mT. The number of cells that were nonspecifically separated during magnetic separation was determined by direct cell counting under a microscope. Preparation of CTLs. All PBMCs were derived from a metastatic melanoma patient given dendritic cell (DC) vaccines. Clinical research using PBMCs from melanoma patients was approved by the Institutional Review Board of the Shizuoka Cancer Center, Shizuoka, Japan. All patients gave written informed consent. PBMCs were stimulated twice in vitro with MAGE-1 A24 peptide-pulsed autologous DCs at a ratio of 1:10 to 1:100 in GT-T507 medium (Kohjin Bio Co., Saitama, Japan) supplemented with 2 mM L-glutamine, 100 U/mL penicillin, 100 µg/mL streptomycin, 50 µg/mL gentamycin, 50 µM β-mercaptoethanol, 1 mM sodium pyruvate, and 5% AB human serum (Cambrex, Walkerville, MD, USA) (referred to as CTL medium). Additionally, cultured CTLs were boosted weekly with two rounds of stimulation at a ratio of 1:10 with MAGE-1 A24 peptide-pulsed T2-A24 cells, which are a human hybrid (B and T lymphoblast) cell line transfected with the HLA-A24 gene (17). Magnetic Cell Separation using MHC/MAGE-1 A24 Peptide Complex-Conjugated BacMPs. Stimulated PBMCs (1.5 × 107 cells) in 120 µL of staining buffer (PBS containing 2% fetal bovine serum (FBS) and 0.1% sodium azide) were reacted with 120 µL of allophycocyanin (APC)-labeled antiCD8 mAb (BD Biosciences, San Jose, CA, USA) for 30 min on ice. After removing excess antibody by centrifugation, cell aliquots (2 × 106 cells) in 80 µL of separation buffer were incubated with 20 µL of MHC/MAGE-1 or HIV A24 peptide complex-conjugated BacMPs for 10 min at 4 °C. After washing by centrifugation, cells suspended in 1 mL of separation buffer were transferred to a test tube, and cells labeled by BacMPs were separated magnetically as described above. The magnetic separation was performed three times at intervals of 5 min. Magnetically separated cells were resuspended in staining buffer as a positive fraction. To analyze the ratio of melanoma-specific CTLs to CTLs suspension, CTLs (2 × 106 cells) stained with APC-labeled anti-CD8 mAb were reacted further with PElabeled MHC/MAGE-1 A24 peptide tetramer (MBL) for 30 min on ice and washed with excess tetramer by centrifugation. The purity of targeted cells in separated cells and CTL suspensions was analyzed using a BD FACS Canto flow cytometer (Becton Dickinson, Franklin Lakes, NJ, USA). A total of 50 000 events were analyzed for each sample. Measurement of IFN-γ Secreted by Separated Melanoma-Specific CTLs. MHC/MAGE-1 A24 peptide complex-conjugated BacMPs were fabricated by immobilizing 1 µM streptavidin and conjugating 0.5 µM MHC/MAGE-1 A24 peptide complexes on the biotin-modified BacMPs, and mela-

Takahashi et al.

Figure 2. Particle size distribution of SA-BacMPs biotinylated with Sulfo-NHS-LC2-biotin (empty bars) or Sulfo- NHS-PEO12-biotin (solid bars).

noma-specific CTLs were separated using the MHC/MAGE-1 A24 peptide complex-conjugated BacMPs using the protocol described above. After an overnight incubation in CTL medium, separated CTLs (1 × 105 cells) and MAGE-1 A24 peptidepulsed TISI cells (1 × 105 cells), a human B-lymphoblastoid cell line presenting HLA-A24, were coincubated in a roundbottom 96-well microculture plate for 24 h in RPMI-1640 medium containing 5% FBS. Finally, supernatants were collected, and the amount of IFN-γ in the supernatants was measured using an ELISA kit specific for human IFN-γ (Biosource, Camallilo, CA, USA).

RESULTS Introduction of Biotin on BacMPs. To conjugate biotinylated MHC/peptide complexes on BacMPs, phased modifications of biotin and streptavidin onto BacMPs were studied. In the first step, biotin introduction onto BacMPs was performed by modifying the amine group on the surface of BacMPs. To modify biotin once on BacMPs, two amino-reactive crosslinkers, Sulfo-NHS-LC2-Biotin (linker length: 30.5 Å) and SulfoNHS-PEO12-Biotin (linker length: 56 Å), were individually reacted in the presence of BacMPs for covalent linkage between the amine group of BacMPs and the cross-linker. The dispersion of streptavidin-immobilized BacMPs (SA-BacMPs) that were biotinylated with LC2-linker or PEO12-linker was evaluated as the size distribution of the BacMPs in solution (Figure 2). The average diameter of SA-BacMPs biotinylated with the LC2linker was 159.4 ( 36.0 nm, and the average diameter of SABacMPs biotinylated with the PEO12-linker was 63.3 ( 17.7 nm. SA-BacMPs biotinylated with the PEO12-linker showed better dispersion due to the hydrophilic linker. Furthermore, as the diameter of BacMP is 50 to 100 nm, it was considered that multiple biotin-modified BacMPs did not bind to a streptavidin molecule. Correlation Between the Number of Streptavidin Molecules Bound on BacMPs and Nonspecific Binding of BacMPs to Cell Surfaces. Streptavidin molecules were immobilized onto biotin-modified BacMPs with PEO12-linker. Excessive immobilization of streptavidin, a highly hydrophobic molecule (18), on BacMPs was predicted to cause nonspecific binding of SA-BacMPs to cells. Therefore, the correlation between the number of streptavidin molecules bound on BacMPs and nonspecific binding of SA-BacMPs to cell surfaces was investigated. SA-BacMPs, which were prepared by reacting streptavidin with biotin-modified BacMPs at different concentrations, were incubated with Raji cells, and the number of cells, which were separated magnetically as a result of nonspecific binding of SA-BacMPs to cell surfaces, were counted directly

MHC/Peptide Complex-Conjugated BacMPs

Bioconjugate Chem., Vol. 20, No. 2, 2009 307

Figure 3. Correlation between the number of streptavidin molecules bound on a single BacMP and nonspecific binding of BacMPs to cell surface. The number of streptavidin molecules bound on a single biotinmodified BacMP was evaluated by measuring the fluorescence intensity of non-immobilized streptavidin in the supernatant. The numbers of Raji cells separated nonspecifically using each SA-BacMPs were counted, and then, the ratio of nonspecifically separated Raji cells was calculated Table 1. Number of MHC/MAGE-1 Peptide Complexes Conjugated on a Single SA-BacMP on Which Different Number of Streptavidin Molecules Immobilized Number of Streptavidin or MHC/MAGE-1 Peptide Complexes on a Single SA-BacMP (Molecules/a Single BacMP) streptavidin 339 ( 10 132 ( 21

MHC/MAGE-1

peptidecomplex

16.0 ( 1.3 17.9 ( 2.1

under a microscope. Then, the correlation between the number of streptavidin molecules immobilized on a single BacMP and nonspecific binding of SA-BacMPs to cell surfaces was estimated (Figure 3). The number of nonspecifically separated cells increased in proportion to the number of streptavidin molecules immobilized on BacMPs. Immobilization of approximately 100 streptavidin molecules on a single BacMP did not influence the nonspecific binding property of BacMPs. These results show that value of less than 100 streptavidin molecules per BacMP are suitable for efficient cell separation. Fabrication of MHC/Peptide A24 Complex-Conjugated BacMPs with Little Nonspecific Binding to Cells. MHC/ peptide complex-conjugated BacMPs were fabricated by reacting biotinylated MHC/peptide complexes with SA-BacMPs. The number of MHC/peptide complexes conjugated on a single SABacMP was evaluated using mouse anti-β2 microglobulin mAb and ALP- labeled antimouse IgG. Eighteen MHC/MAGE-1 A24 peptide complexes were conjugated per BacMP on which 100 streptavidin molecules were immobilized, and the number of MHC/MAGE-1 A24 peptide complexes per BacMP was almost equal to that of BacMPs on which over 300 streptavidin molecules were immobilized (16 MHC/MAGE-1 A24 peptide complexes, Table 1). From these results, we speculated that only 1 or 0 MHC/peptide complexes was bound to a single streptavidin molecule on BacMPs, and confirmed that many empty streptavidin sites remained immobilized on BacMPs. For blocking of these empty streptavidin sites and, therefore, suppression of nonspecific binding of BacMPs to cells, MHC/ peptide complex-BacMPs were incubated with amine-PEO3biotin. The nonspecific binding of three different MHC/peptide complex-conjugated BacMPs to cells was evaluated (Figure 4), and nonspecific binding of BacMPs to cell surfaces was decreased markedly by PEO blocking on BacMP surfaces. The ratio of cells separated nonspecifically using MHC/MAGE-1 A24 peptide complex-conjugated BacMPs with PEO blocking was approximately one-sixth that observed using BacMPs without PEO blocking. These results suggest that PEO blocking

Figure 4. Nonspecific binding of BacMPs to the cell surface. The numbers of Raji cells separated nonspecifically using wild-type BacMPs (BacMP), MHC/MAGE-1 peptide complex-conjugated BacMPs on which 330 streptavidin molecules per BacMP (MAGE-1-SA(330)BacMP), and 140 streptavidin molecules per BacMP (MAGE-1SA(140)-BacMP) were immobilized, and MHC/MAGE-1 peptide complex-conjugated BacMPs with PEO blocking (PEO-MAGE-1SA(140)-BacMP), on which 140 streptavidin molecules per BacMP were immobilized, were counted. Then, the ratio of nonspecifically separated Raji cells was calculated.

of MHC/MAGE-1 A24 peptide complex-conjugated BacMPs is necessary for the high-purity magnetic separation required for melanoma-specific CTLs. Therefore, we used PEO-blocked MHC/MAGE-1 A24 peptide complex-conjugated BacMPs for the following CTL separation experiments. Magnetic Separation of Melanoma-Specific CTLs Using MHC/MAGE-1 A24 Peptide Complex-Conjugated BacMPs. Magnetic separation of melanoma-specific CTLs from in vitro stimulated PBMCs prepared from a vaccinated melanoma patient using MHC/MAGE-1 A24 peptide complexconjugated BacMPs was performed and evaluated using a PElabeled MHC/MAGE-1 A24 peptide tetramer and APC-labeled anti-CD8 mAb. The purity of melanoma-reactive cells (PEMAGE-1 tetramer positive in Figure 5A) and melanoma-specific CTLs (PE-MAGE-1 tetramer positive and APC-anti-CD8 mAb positive in Figure 5A) in the stimulated PBMC from a melanoma patient were 31.2% and 30.1%, respectively. Most melanomareactive cells were melanoma-specific CTLs. MHC/MAGE-1 or HIV (negative control) A24 peptide complex-conjugated BacMPs were incubated in stimulated PBMCs, and the number of magnetically separated cells was counted. The number of cells separated nonspecifically using MHC/HIV A24 peptide complex-conjugated BacMPs was approximately one-sixth of the number of cells separated using MHC/MAGE-1 A24 peptide complex-conjugated BacMPs. Therefore, it was concluded that the target cells, which react to melanoma cells, were separated by the interaction with MHC/MAGE-1 A24 peptide complexes conjugated on BacMPs. Then, according to analysis of the separated cells using MHC/MAGE-1 A24 peptide complexconjugated BacMPs, 93.1% of the separated cells were melanomareactive cells (Alexa-488-MAGE-1-BacMPs positive in Figure 5B), and 87.7% of those were melanoma-specific CTLs (Alexa488-MAGE-1-BacMPs positive and APC-anti-CD8 mAb positive in Figure 5B). These results indicate that MHC/MAGE-1 A24 peptide complex-conjugated BacMPs bind to targeted melanoma-specific CTLs with high selectivity and that these cells were separated with high purity using this system. Function of Separated Melanoma-Specific CTLs. The specificity of antigen recognition and cytokine secretion of CTLs following magnetic separation using MHC/MAGE-1 A24 peptide complex-conjugated BacMPs were evaluated. The amount

308 Bioconjugate Chem., Vol. 20, No. 2, 2009

Takahashi et al.

Figure 5. Magnetic separation of melanoma-specific CTLs using MHC/MAGE-1 peptide complex-conjugated BacMPs. Cells before (A) and after (B) magnetic separation were analyzed by flow cytometry and are shown accordingly. Prior to magnetic separation, the cells were stained with APC-labeled anti-CD8 mAb (APC-anti-CD8 mAb). For the analysis of cells before magnetic separation, PE-labeled MHC/MAGE-1 peptide tetramers (PE-MAGE-1 tetramer) were used. Upon magnetic separation, the cells were analyzed with reference to the fluorescence intensity of Alxa-488 labeled to the MHC/MAGE-1 peptide complex-conjugated BacMPs (Alexa-488-MAGE-1-BacMPs). Table 2. Levels of IFN-γ Secreted by Separated Melanoma-Specific CTLs Incubated with TISI Cells or MGAE-1 Peptide-Pulsed TISI Cells Amount of Secreted IFN-γ (pg/mL) MAGE-1 peptide-pulsed TISI cells TISI cells

1049 ( 24.7 34 ( 4.4

of IFN-γ in the supernatants of CTLs cultured with MAGE-1 A24 peptide-pulsed TISI cells was measured (Table 2). The separated CTLs cultured with peptide nonpulsed TISI cells, which did not express MAGE-1 A24 peptide and MHC (HLAA24) complexes on the cell surface, secreted little IFN-γ. It was indicated that the CTLs cultured with peptide nonpulsed TISI cells did not activate. Furthermore, it was suggested that BacMPs on CTLs were not a cytotoxin and did not activate the CTLs. On the other hand, the amount of IFN-γ secreted by the separated CTLs cultured with MAGE-1 A24 peptide-pulsed TISI cells was approximately 30 times more than that of the CTLs cultured with peptide nonpulsed TISI cells. These results indicate that the CTLs separated using MHC/MAGE-1 A24 peptide complex-conjugated BacMPs were stimulated specifically by MAGE-1 A24 peptide and MHC complexes expressed on TISI cells, and the stimulated CTLs were able to secrete IFN-γ, a normal function of melanoma-specific CTLs.

DISCUSSION In preparing a population of target cells for applications such as cell translation or cell immunotherapy, techniques for separating these cells from a heterogeneous cell mixture must produce large numbers of highly pure target cells. For magnetic cell separation, the magnetic strength, size, and surface characteristics of magnetic particles influence the purity of separated cells. It is possible to reduce nonspecific binding to cells, which causes decreased purity, by controlling the particle surface characteristics. Specially, various surface modifications, such as the addition of poly(ethylene glycol) and polysaccharide to add hydrophilicity to the particles, have been demonstrated to reduce nonspecific interactions with the cell surface (19-22). In this report, BacMPs, which are known to elicit little nonspecific binding, were developed for the magnetic separation of melanoma-specific CTLs. First, the surfaces of BacMPs were coated with a PEO linker, and the hydrophilicity and steric barrier of the PEO contributed to the progress of distribution of BacMPs (23). Then, the number of streptavidin molecules immobilized on single BacMPs was controlled. This was performed, as it was speculated that excessive immobilization of streptavidin augments hydrophobicity and therefore acceler-

ates nonspecific binding of BacMPs to nontarget cells (18). It has also been known that animal cells prefer to attach to particles as they increase in hydrophobicity (20). In the present study, nonspecific binding of BacMPs to nontarget cells was reduced markedly by controlling the number of immobilized streptavidin molecules per BacMP and via further modification of PEO on BacMPs. Finally, BacMPs that specifically target CTLs through the conjugation of MHC/peptide complexes on their surfaces allowed for the isolation of target CTLs with high purity. In all previous reports, magnetic separation of antigenspecific CTLs has been demonstrated using both a PE-labeled MHC/peptide tetramer for targeting the cells and magnetic particles displaying anti-PE antibodies for magnet marking of the cells via the tetramer (24-27). This two-step separation lacks precision, is time-consuming, and is difficult to automate, and for these reasons, this technique is limited in its application toward immunotherapies. Therefore, we proposed the novel separation system in which antigenspecific CTLs are separated in a single-step process using BacMPs with directly conjugated MHC/peptide complexes as a more suitable process for use in immunotherapy. Conjugation of MHC/peptide complexes on magnetic nanoparticles for cell separation and single-step separation of antigen-specific CTLs was the first trial. In this report, we demonstrated efficient melanoma-specific CTL separation using MHC/MAGE-1 A24 peptide complexconjugated BacMPs. The BacMPs conjugated with MHC/ MAGE-1 A24 peptide complexes bound to targeted melanomaspecific CTLs with little nonspecific binding to nontarget cells, and a single-step magnetic separation of the melanomaspecific CTLs from stimulated PBMCs derived from a vaccinated metastatic melanoma patient was achieved while maintaining CTL function. CTLs separated efficiently and with high purity will be useful in developing immunotherapies and in fundamental research.

ACKNOWLEDGMENT This work was funded in part by a Grant-in-Aid for Scientific Research on Priority Areas “City Area Program” from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

LITERATURE CITED (1) Anderson, B., Park, B. J., Verdaguer, J., Amrani, A., and Santamaria, P. (1999) Prevalent CD8(+) T cell response against

MHC/Peptide Complex-Conjugated BacMPs one peptide/MHC complex in autoimmune diabetes. Proc. Natl. Acad. Sci. U.S.A. 96, 9311–9316. (2) Pinilla, C., Rubio-Godoy, V., Dutoit, V., Guillaume, P., Simon, R., Zhao, Y., Houghten, R. A., Cerottini, J. C., Romero, P., and Valmori, D. (2001) Combinatorial peptide libraries as an alternative approach to the identification of ligands for tumorreactive cytolytic T lymphocytes. Cancer Res. 61, 5153 5160. (3) Hermentin, P., Doenges, R., Franssen, U., Bieva, C., Vander Brugghen, F. J., Stryckmans, P., Friesen, H. J., Optaczy, B., Schneider, S., Ax, W., et al. (1990) Hinge-thiol coupling of monoclonal antibody to silanized iron oxide particles and evaluation of magnetic cell depletion. Bioconjugate Chem. 1, 411–418. (4) Campbell, J. D., Piechaczek, C., Winkels, G., Schwamborn, E., Micheli, D., Hennemann, S., and Schmitz, J. (2005) Isolation and generation of clinical-grade dendritic cells using the CliniMACS system. Methods Mol. Med. 109, 55–70. (5) Jing, Y., Moore, L. R., Williams, P. S., Chalmers, J. J., Farag, S. S., Bolwell, B., and Zborowski, M. (2007) Blood progenitor cell separation from clinical leukapheresis product by magnetic nanoparticle binding and magnetophoresis. Biotechnol. Bioeng. 96, 1139–1154. (6) Leyendeckers, H., Voth, E., Schicha, H., Hunzelmann, N., Banga, P., and Schmitz, J. (2002) Frequent detection of thyroid peroxidase-specific IgG+ memory B cells in blood of patients with autoimmune thyroid disease. Eur. J. Immunol. 32, 3126– 3132. (7) Miltenyi, S., Muller, W., Weichel, W., and Radbruch, A. (1990) High gradient magnetic cell separation with MACS. Cytometry 11, 231–238. (8) Altman, J. D., Moss, P. A., Goulder, P. J., Barouch, D. H., McHeyzer-Williams, M. G., Bell, J. I., McMichael, A. J., and Davis, M. M. (1996) Phenotypic analysis of antigen-specific T lymphocytes. Science 274, 94–96. (9) Verginis, P., McLaughlin, K. A., Wucherpfennig, K. W., von Boehmer, H., and Apostolou, I. (2008) Induction of antigenspecific regulatory T cells in wild-type mice: visualization and targets of suppression. Proc. Natl. Acad. Sci. U.S.A. 105, 3479– 3484. (10) Vonderheide, R. H., Domchek, S. M., Schultze, J. L., George, D. J., Hoar, K. M., Chen, D. Y., Stephans, K. F., Masutomi, K., Loda, M., Xia, Z., Anderson, K. S., Hahn, W. C., and Nadler, L. M. (2004) Vaccination of cancer patients against telomerase induces functional antitumor CD8+ T lymphocytes. Clin. Cancer. Res. 10, 828–839. (11) Yoshino, T., Hirabe, H., Takahashi, M., Kuhara, M., Takeyama, H., and Matsunaga, T. (2008) Magnetic cell separation using nano-sized bacterial magnetic particles with reconstructed magnetosome membrane. Biotechnol. Bioeng. 101, 470–477. (12) Kuhara, M., Takeyama, H., Tanaka, T., and Matsunaga, T. (2004) Magnetic cell separation using antibody binding with protein a expressed on bacterial magnetic particles. Anal. Chem. 76, 6207–6213. (13) Matsunaga, T., Takahashi, M., Yoshino, T., Kuhara, M., and Takeyama, H. (2006) Magnetic separation of CD14+ cells using antibody binding with protein A expressed on bacterial magnetic particles for generating dendritic cells. Biochem. Biophys. Res. Commun. 350, 1019–1025.

Bioconjugate Chem., Vol. 20, No. 2, 2009 309 (14) Takahashi, M., Yoshino, T., Takeyama, H., and Matsunaga, T. (2008) Direct magnetic separation of immune cells from whole blood using bacterial magnetic particles displaying protein G, Biotechnol. Prog., in press. (15) Matsunaga, T., Sakaguchi, T., and Tadokoro, F. (1991) Magnetite formation by a magnetic bacterium capable of growing aerobically. Appl. Microbiol. Biotechnol. 35, 651–655. (16) Tanaka, T., and Matsunaga, T. (2000) Fully automated chemiluminescence immunoassay of insulin using antibodyprotein A-bacterial magnetic particle complexes. Anal. Chem. 72, 3518–3522. (17) Akiyama, Y., Kuzushima, K., Tsurumi, T., and Yamaguchi, K. (2004) Analysis of HLA-A24-restricted CMVpp65 peptidespecific CTL with HLA-A*2402-CMVpp65 tetramer. Immunol. Lett. 95, 199–205. (18) Schwidop, W. D., Klossek, P., Muller, R., and Claus, R. (1990) Procedure for the purification of streptavidin by hydrophobic interaction chromatography. J. Chromatogr. 520, 325–331. (19) Horak, D., Babic, M., Jendelova, P., Herynek, V., Trchova, M., Pientka, Z., Pollert, E., Hajek, M., and Sykova, E. (2007) D-Mannose-modified iron oxide nanoparticles for stem cell labeling. Bioconjugate Chem. 18, 635–644. (20) Hu, Y., Xie, J., Tong, Y. W., and Wang, C. H. (2007) Effect of PEG conformation and particle size on the cellular uptake efficiency of nanoparticles with the HepG2 cells. J. Controlled Release 118, 7–17. (21) Lewin, M., Carlesso, N., Tung, C. H., Tang, X. W., Cory, D., Scadden, D. T., and Weissleder, R. (2000) Tat peptidederivatized magnetic nanoparticles allow in vivo tracking and recovery of progenitor cells. Nat. Biotechnol. 18, 410–414. (22) Natarajan, A., Gruettner, C., Ivkov, R., DeNardo, G. L., Mirick, G., Yuan, A., Foreman, A., and DeNardo, S. J. (2008) NanoFerrite particle based radioimmunonanoparticles: binding affinity and in vivo pharmacokinetics. Bioconjugate Chem. 19, 1211–1218. (23) Sadzuka, Y., Nakade, A., Hirama, R., Miyagishima, A., Nozawa, Y., Hirota, S., and Sonobe, T. (2002) Effects of mixed polyethyleneglycol modification on fixed aqueous layer thickness and antitumor activity of doxorubicin containing liposome. Int. J. Pharm. 238, 171–180. (24) Appel, H., Kuon, W., Kuhne, M., Wu, P., Kuhlmann, S., Kollnberger, S., Thiel, A., Bowness, P., and Sieper, J. (2004) Use of HLA-B27 tetramers to identify low-frequency antigenspecific T cells in Chlamydia-triggered reactive arthritis. Arthritis Res. Ther. 6, R521–534. (25) Barnes, E., Ward, S. M., Kasprowicz, V. O., Dusheiko, G., Klenerman, P., and Lucas, M. (2004) Ultra-sensitive class I tetramer analysis reveals previously undetectable populations of antiviral CD8+ T cells. Eur. J. Immunol. 34, 1570–1577. (26) Kim, S. K., Devine, L., Angevine, M., DeMars, R., and Kavathas, P. B. (2000) Direct detection and magnetic isolation of Chlamydia trachomatis major outer membrane protein-specific CD8+ CTLs with HLA class I tetramers. J. Immunol. 165, 7285– 7292. (27) McDermott, A. B., Spiegel, H. M., Irsch, J., Ogg, G. S., and Nixon, D. F. (2001) A simple and rapid magnetic bead separation technique for the isolation of tetramer-positive virus-specific CD8 T cells. Aids 15, 810–812. BC800398D