Magnetic Cell Separation Using Antibody Binding with Protein A

Oct 7, 2004 - Flow cytometric analysis showed that ∼97.5 ± 1.7% of CD19+ and CD20+ cells were involved in the positive fraction after magnetic sepa...
0 downloads 11 Views 266KB Size
Download PDF
Anal. Chem. 2004, 76, 6207-6213

Magnetic Cell Separation Using Antibody Binding with Protein A Expressed on Bacterial Magnetic Particles Motoki Kuhara,†,‡ Haruko Takeyama,‡ Tsuyoshi Tanaka,‡ and Tadashi Matsunaga*,‡

Technical and Development, Medical & Biological Laboratories Company Ltd., 1063-103, Ohara, Terasawaoka, Ina, Nagano, 396-0002, Japan, and Department of Biotechnology, Tokyo University of Agriculture and Technology, 2-24-16, Naka-cho, Koganei, 184-8588, Tokyo, Japan

Bacterial magnetic particles (BacMPs) are efficient platforms of proteins for surface display systems. In this study, mononuclear cells from peripheral blood were separated using BacMPs expressing protein A on the BacMP membrane surface (protein A-BacMPs), which were complexed with the Fc fragment of anti-mouse IgG antibody. The procedure of positive selection involves incubation of mononuclear cells and mouse monoclonal antibodies against different cell surface antigens (CD8, CD14, CD19, CD20) prior to treatment with protein A-BacMP binding with rabbit anti-mouse IgG secondary antibodies. Flow cytometric analysis showed that ∼97.5 ( 1.7% of CD19+ and CD20+ cells were involved in the positive fraction after magnetic separation. The ratio of the negative cells in the negative fraction was ∼97.6 ( 1.4%. This indicates that CD19+ and CD20+ cells can be efficiently separated from mononuclear cells. Stem cell marker (CD34) positive cells were also separated using protein A-BacMP binding with antibody. May-Grunwald Giemsa stain showed a high nuclear/cytoplasm ratio, which indicates a typical staining pattern of stem cells. The separated cells had the capability of colony formation as hematopoietic stem cells. Furthermore, the inhibitory effect of magnetic cell separation on CD14+ cells was evaluated by measurement of cytokine in the culture supernatant by ELISA when the cells were cultured with or without lipopolysaccharide (LPS). The induction of IL1-β, TNFr, and IL6 was observed in the presence of 1 ng/mL LPS in all fractions. On the other hand, in the absence of LPS, BacMPs had little immunopotentiation to CD14+ cells as well as that of artificial magnetic particles, although TNFr and IL6 were slightly induced in the absence of LPS in the positive fraction. Magnetic particles have been widely used as carriers of antibodies for immunoassay, cell separation, and tissue typing.1,2 Use of magnetic particles is advantageous for full automation, * To whom correspondence should be addressed. E-mail: tmatsuna@ cc.tuat.ac.jp. † Medical & Biological Laboratories Co. Ltd. ‡ Tokyo University of Agriculture and Technology. (1) Lund, V.; Schmid, R.; Rickwood, D.; Hornes, E. Nucleic Acids Res. 1988, 16, 10861-10880. 10.1021/ac0493727 CCC: $27.50 Published on Web 10/07/2004

© 2004 American Chemical Society

resulting in minimizing manual labor and providing more precise results.3 Especially, immunomagnetic particles have been used preferentially in target cell separation from leukocytes4-7 as in vitro diagnosis, because of more rapid and simple methodology compared with cell sorting using a flow cytometer. The kind and role of leukocytes has been well studied. There are many antigens on the cell surface, and antibodies against these antigens are available for analysis and separation of target cells. To separate target cells using antibodies against cell surface antigens, magnetic separation has been frequently used. Commercially available magnetic particles are chemically synthesized compounds of micrometer and nanometer sizes. Use of nanosized magnetic particles has advantages in assay sensitivity, rapidity, and precision. However, it is difficult to synthesize nanosized magnetic particles with uniform size and shape and good dispersity in aqueous solutions. Consequently, advanced technique and high costs are required for the production of such nanosized magnetic particles. Magnetic bacteria have been isolated from fresh and marine sediments and are known to produce nanosized magnetic particles.8-12 Magnetospirillum magneticum AMB-1 synthesizes cubooctahedral bacterial magnetic particles (BacMPs). The BacMP has a single magnetic domain of magnetite. Furthermore, each BacMP is covered with phospholipid membrane.13 Because (2) Fan, Z. H.; Mangru, S.; Granzow, R.; Heaney, P.; Ho, W.; Dong, Q.; Kumar, R. Anal. Chem. 1999, 71, 4851-4859. (3) Sawakami-Kobayashi, K.; Segawa, O.; Obata, K.; Hornes, E.; Yohda, M.; Tajima, H.; Machida, M. Biotechniques 2003, 34, 634-637. (4) Stmpfli, M. R.; Miescher, S.; Aebischer, I.; Zurcher, A. W.; Stadler, B. M. Eur. J. Immunol. 1994, 24, 2161-2167. (5) Schratzberger, P.; Reinisch, N.; Prodinger, W. M.; Kahler, C. M.; Sitte, B. A.; Bellmann, R.; Fischer-Colbrie, R.; Winkler, H.; Wiedermann, C. J. J. Immunol. 1997, 158, 3895-3901. (6) Parra, E.; Wingren, A. G.; Hedlund, G.; Kalland, T.; Dohlsten, M. J. Immunol. 1997, 158, 637-642. (7) Pickl, W. F.; Majdic, O.; Kohl, P.; Stockl, J.; Riedl, E.; Scheinecker, C.; BelloFernandez, C.; Knapp, W. J. Immunol. 1996, 157, 3850-3859. (8) Blakemore, R. P. Science 1975, 190, 377-379. (9) Matsunaga, T.; Kamiya, S. Appl. Microbiol. Biotechnol. 1987, 26, 328-332. (10) Matsunaga, T.; Tadokoro, F.; Nakamura, N. IEEE Trans. Magn. 1990, 26, 1557-1559. (11) Matsunaga, T.; Sakaguchi, T.; Tadokoro, F. Appl. Microbiol. Biotechnol. 1991, 35, 651-655. (12) Sakaguchi, T.; Burgess, J. G.; Matsunaga, T. Nature (London) 1993, 365, 47-49. (13) Balkwill, D. L.; Maratea, D.; Blakemore, R. P. J. Bacteriol. 1980, 141, 13991408.

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004 6207

phospholipids are negatively charged at neutral pH, the surface of BacMPs is also negatively charged enough to have monodispersity in the 40-160-nm size range.14 Therefore, BacMPs do not settle due to gravity, while artificially synthesized magnetite particles of the same size aggregate because of spontaneous magnetization and settle out. Superior dispersion permits various applications of BacMPs, such as highly sensitive, rapid, or precise immunoassays. Fluoroimmunoassay using FITC-labeled antibody conjugated to BacMPs has been developed for a highly sensitive detection of bacteria.15 This assay system was also applied to the removal of microbial cells. Recently, the mechanism of synthesis of magnetic particles in magnetic bacterium M. magneticum strain AMB-1 has been analyzed. An iron transport protein MagA,15 a magnetosome-specific GTPase required for magnetic particles synthesis,16 and a novel protein involved in biological magnetite crystal formation have been also discovered in magnetic bacteria.17 In our previous work, MagA was isolated from transposon mutagenesis in AMB-1.18 The magA gene encodes an integral irontranslocating protein of BacMPs membrane17 and may therefore be used as an anchor for expression of foreign proteins on the surface of the BacMP membrane. The technique of displaying functional protein on the surface of BacMPs using magA gene fusions made it possible to immobilize proteins on magnetic particles without chemical cross-linking. A proteinA-magA fusion gene was cloned in M. magneticum strain AMB-1, and BacMPs bearing protein A were successfully expressed on the BacMP membrane surface.19 The protein A expressing BacMPs have been used for chemiluminescence immunoassay of human insulin in serum, employing a fully automated system.14 The protein A-expressing BacMPs maintain good monodispersity, while chemical conjugation of antibody onto BacMPs yields a decrease in the monodispersity of the BacMPs. Furthermore, the antigen-binding activity of antibody on protein A expressed on BacMPs is 2 times higher than that of chemically conjugated antibody onto BacMPs. These techniques offer an economical advantage in protein A assembling, because in general, protein A is obtained from recombinant Escherichia coli through a time-consuming purification process. In this study, a magnetic separation of mononuclear cells from peripheral blood using BacMPs expressing protein A on the BacMP membrane surface, complexed with the Fc fragment of anti-mouse IgG antibody, was demonstrated. The inhibition effect of magnetic cell separation on CD14+ cells (monocyte separation) and the proliferation and differentiation capacity of CD34+ (stem cells fraction) separated by BacMPs have been also studied, respectively. EXPERIMENTAL SECTION Materials. Rabbit anti-mouse IgG polyclonal antibody was purchased from Cappel. FITC-conjugated goat anti-mouse IgG (14) Tanaka, T.; Matsunaga, T. Anal. Chem. 2000, 72, 3518-3522. (15) Nakamura, N.; Burgess, J. G.; Yagiuda, K.; Kudo, S.; Sakaguchi, T.; Matsunaga, T. Anal. Chem. 1993, 65, 2036-2039. (16) Okamura, Y.; Takeyama, H.; Matsunaga, T. J. Biol. Chem. 2001, 276, 48183-48188. (17) Nakamura, C.; Kikuchi, T.; Burgess, J. G.; Matsunaga, T. J. Biochem. 1995, 118, 23-27. (18) Arakaki, A.; Webb, J.; Matsunaga, T. J. Biol. Chem. 2003, 278, 8745-8750. (19) Matsunaga, T.; Sato, R.; Kamiya, S.; Tanaka, T.; Takeyama, H. J. Magn. Magn. Mater. 1999, 194, 126-134.

6208

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

antibody was obtained from Immunotech (Marseille, France). Mouse monoclonal antibodies against CD8, CD19, CD20, CD14 and CD34 antigens were obtained from Immunotech. Lipopolysaccharide (LPS) from Salmonella minnesota Re 595 was purchased from Sigma (St. Louis, MO.). Magnetite particles (artificial magnetic particles) binding with anti-mouse IgG antibody was obtained from Miltenyi Biotech. Cultivation of Magnetic Bacteria. Expression plasmid, pRZM (12.8 kbp) containing protein A (ezz) and magA fusion gene with the magA promoter were transferred into a wild-type strain of AMB-1 as described previously.19 Transconjugants were microaerobically cultured in magnetic spirillum growth medium20 containing 5 µg/mL tetracycline at 28 °C. Microaerobic conditions were established by sparging with argon gas. Batch culture was performed in a fermentor with a working volume of 8 L. Preparation of BacMPs Expressing Protein A Binding with Antibody. BacMPs expressing protein A (protein A-BacMPs) was purified from strain AMB-1 transconjugant according to the method described by Tanaka and Matsunaga.14 Cultured cells were collected by centrifugation, suspended in 40 mL of phosphate-buffered saline (PBS, pH 7.4), and disrupted by three passes through a French press cell at 1500 kg/cm2 (Ohtake Works Co. Ltd., Tokyo, Japan). Protein A-BacMPs were collected from cell extracts using a columnar neodymium-iron-boron (NdFe-B) magnet. The collected protein A-BacMPs were washed with PBS at least five times by dispersion using ultrasonication and collection using the Nd-Fe-B magnet. To prepare protein A-BacMP binding with antibody, 500 µL of anti-mouse IgG antibody (1 mg/mL) was added to 500 µL of protein A-BacMP suspension (2 mg/mL). The amount of antibody bound onto BacMPs is estimated to be maximally 3% (w/w). The complexes were separated by Nd-Fe-B magnet and washed three times with PBS. The BacMPs were kept at 4 °C in PBS containing 0.1% sodium azide before use. Purification of Mononuclear Cells. Purification of mononuclear cells from peripheral blood was performed according to the method described previously.21 Blood samples were collected from six individuals in our laboratory. Among of them, we selected blood samples from one cell donor, which data could be clearly obtained by FACS analysis. Peripheral blood collected in tubes containing heparin was mixed with twice the volume of PBS. Then it was layered on 1/2 volume of Histopaque-1077 (Sigma) and centrifuged at 400g for 30 min to collect mononuclear cells at the plasma-Histopaque interface. The interface containing the mononuclear cell fraction was transferred to a sterile tube and washed twice with PBS containing 0.5% BSA and 2 mM EDTA. The mononuclear cells were incubated with FcR blocker (Miltenyi Biotech) at room temperature for 10 min to prevent Fc receptormediated phagocytosis. Then cells were washed with PBS containing 0.5% BSA and 2 mM EDTA. Magnetic Separation of Mononuclear Cells. A total of 2 × 107 mononuclear cells were used in the following experiments. B lymphocytes, T lymphocytes, and monocytes were reacted with mouse anti-CD19, anti-CD20, anti-CD8, and anti-CD14 monoclonal (20) Blakemore, R. P.; Maratea, D.; Wolfe, R. S. J. Bacteriol. 1979, 140, 720729. (21) Szuster-Ciesielska, A.; Tustanowska-Stachura, A.; Slotwinska, M.; Marmurowska-Michalowska, H.; Kandefer-Szerszen, M. Pol. J. Pharmacol. 2003, 55, 353-362.

antibodies, respectively. Mononuclear cells (2 × 107 cells) were incubated with various primary antibodies against cell surface antigens (CD8, CD14, CD19, CD20) at 4 °C for 15 min. After removal of excess primary antibodies, the cells with primary antibodies were further reacted with protein A-BacMP binding with anti-mouse IgG antibody at 4 °C for 10 min to separate specific cells magnetically and subsequently reacted with FITClabeled anti-mouse IgG antibody at 4 °C for 10 min to analyze by flow cytometer. Then, the cell suspension was transferred to a test tube (10 mm × 75 mm), and the cells were magnetically collected by applying the magnetic field using a columnar Nd-B magnet (diameter 22.5 mm, height 12.5 mm) that produced an inhomogeneous magnetic field (0.5 T at the surface). Magnetic separation was performed for 5 min. The magnetically separated cells were resuspended with PBS containing 0.5% BSA and 2 mM EDTA as the positive fraction. The positive fraction includes antimouse IgG antibody-protein A-BacMP complexes binding with cells labeled with mouse anti-CD antibody and stained with FITClabeled anti-mouse IgG. The supernatant in the washing process was analyzed as a negative fraction. The purity of collected cells was analyzed with a FACScan (Becton Dickinson, San Jose, CA). A total of 10 000 events were analyzed for each sample. The recovery ratio of target cells was estimated as follows; the initial number of target cells was calculated from FACS data analysis of total cells. Total cells were measured by direct cell count using a microscope. The number of target cells after magnetic separation was calculated in the same manner. The recovery ratio (%) of target cells was estimated as (the number of target cells after magnetic separation/the initial number of target cells) × 100. Hematopoietic Colony Assays. Hematopoietic colony assay was investigated as described previously.22 Approximately 1 × 105-2 × 105 of CD34+ cells separated from ∼1 × 108 cells were used for the hematopoietic colony assay. Hematopoietic colonies were demonstrated by growing cells in Methocult GF+ media (StemCell Technologies, Vancouver, Canada) consisting of 1% methylcellulose, 30% FBS, 1% BSA, 50 ng/mL stem cell factor, 20 ng/mL granulocyte-macrophage colony stimulating factor, 20 ng/ mL IL-3, 20 ng/mL IL6, 20 ng/mL granulocyte colony stimulating factor, and 3 units/mL erythropoietin. Stem cell marker (CD34) positive cells separated using protein A-BacMP binding with antibody were aliquoted in duplicate sample at (1-2) × 105 cells/ plate. After 14 days, colony forming was observed. Blood mononuclear cells before magnetic separation were used as controls. This experiment was duplicated. Immunostimulation Assay. To investigate the effect of magnetic cell separation on immunoresponse (expression of various cytokines) of cells, the CD14+ cells separated by mouse anti-CD14 antibody and protein A-BacMP binding with antimouse IgG antibody were cultured in the presence or absence of LPS, and then the cytokines (IL1β, TNFR, IL6) in culture supernatants of the CD14+ cells were measured by ELISA (Immumo-tech). Briefly, 5 × 105 cells/well of CD14+ cells were cultured for 7 days in 1 mL of RPMI-1640 containing 10% FCS (Sigma) and 1 ng/mL LPS, which is well known to cause an immunoresponse.23 Culture supernatants of nontreated cells and (22) Kaufman, D. S.; Hanson, E. T.; Lewis, R. L.; Auerbach, R.; Thomson, J. A. Proc. Natl. Acad. Sci. U.S.A. 2001, 98, 10716-10721. (23) Wright, S. D.; Ramos, R. A.; Tobias, P. S.; Ulevitch, R. J.; Mathison, J. C. Science 1990, 249, 1431-1433.

CD14- cells were also analyzed as controls. Furthermore, the cytokine induction by using commercially available magnetite particles (artificial magnetic particles) (Miltenyi Biotec.) binding with anti-mouse IgG antibody was also performed as a control of the magnetic particle-based assay. CD14+ cells with artificial magnetic particles were separated by high-gradient magnetic sorting using the MACS technique.24 Magnetically captured cells were applied to magnetic columns and eluted by removal of columns from magnetic device. RESULTS Magnetic Separation of Target Cells Using Mouse AntiCD Antibodies and Protein A-BacMP Binding with AntiMouse IgG Antibody. Figure 1 shows the results of immunomagnetic cell separation using protein A-BacMP binding with antibody. As a result of flow cytometirc analysis of cells binding with mouse anti-CD19 or CD20 antibody and stained with FITClabeled anti-mouse IgG antibody, mononuclear cells contained 12.2 ( 2.1% CD19+ and 11.1 ( 2.4% CD20+ cells before cell separation (Figure 1). After magnetic separation using protein A-BacMP binding with mouse anti-CD19 and CD20 antibodies, the negative fraction contained 3.0 ( 0.9 and 2.8 ( 0.9% fluorescent cells. The positive fraction contained 97.5 ( 1.7 and 97.6 ( 1.4% fluorescent cells, respectively. Furthermore, more than 95% of the recovery ratio was obtained for CD19+ and CD20+cells. This indicates that CD19+ and CD20+ cells (B lymphocyte fraction) can be efficiently separated from mononuclear cells of peripheral blood using protein A-BacMP binding with anti-mouse IgG antibody. In the separation of CD8+ and CD14+ cells (T lymphocyte and monocyte fraction, respectively), the efficiencies was lower than that in CD19+ and CD20+ cells (B lymphocyte fraction) where the negative fractions contained 9.2 ( 0.6 (CD8+) and 3.9 ( 3.0% (CD14+) fluorescent cells and the positive fractions contained 91.2 ( 3.0 (CD8+) and 88.6 ( 2.2% (CD14+) fluorescent cells. When commercially available magnetic particles supplied by Miltenyi Biotech were used for the separation of CD8+ cells and CD14+ cells, the positive fractions contained 86.3% for CD8+ cells and 95.8% for CD14+ cells. These results indicate that higher purity of CD8+ cells and lower purity of CD14+ cells were obtained using BacMPs. Further investigation should be performed to analyze the difference between the purities obtained from both assays. However, the whole procedure for cell separation using BacMPs can be operated in a test tube using a Nd-Fe-B magnet, although a special magnetic column was required in collecting artificial magnetic particles. Therefore, the use of BacMPs allows a more simple and rapid cell separation process. Morphology of Immunomagnetically Separated Cells. To identify cells separated by using mouse anti-CD14 and anti-CD34 antibodies, and protein A-BacMP binding with anti-mouse IgG antibody, May-Grunwald Giemsa stain of monocyte marker (CD14) and stem cell marker (CD34) positive cells was performed. CD34 is the best identified surface antigen expressed on hematopoietic stem cells. The nuclei of the CD14+ cells were lobulated with a dense azurophilic chromatin in the May-Grunwald Giemsa staining (Figure 2B). Additionally, the staining of CD34+ cells showed high nuclear/cytoplasm (N/C) ratio, indicating the typical (24) Zhang, S. C.; Wernig, M.; Duncan, I. D.; Brustle, O.; Thomson, J. A. Nat. Biotechnol. 2001, 19, 1129-1133.

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

6209

Figure 1. Fluorescence histograms of mononuclear cells from peripheral blood before separation, cells of the unstained fraction (negative fraction), and cells of the stained fraction (positive fraction). Cells before magnetic separation, in negative fraction and positive fraction are analyzed by FACScan.

staining pattern of stem cells (Figure 2C). On the other hand, this staining pattern was hardly detected in the peripheral blood mononuclear cells before magnetic separation (Figure 2A), since CD34+ cells are very infrequently encountered in the mononuclear cell fraction of peripheral blood. These results show that magnetic separation of CD14+ cells and CD34+ cells by using mouse antiCD14 and anti-CD34 antibodies and protein A-BacMP binding with anti-mouse IgG antibody do not cause morphological change of separated cells. Hematopoietic Colony Assay of CD34+ Cells (Stem Cell Fraction) after Magnetic Separation. To analyze the proliferation and differentiation capacity of the stem cells separated by mouse anti-CD34 antibody and protein A-BacMP binding with anti-mouse IgG antibody, a hematopoietic colony assay of CD34+ cells was performed. When separated cells were incubated in Methocult GF+ media for 14 days, colony-forming granulocyte/ macrophage and burst-forming erythroid were observed (Figure 3A and B). These results show that CD34+ cells separated using BacMPs kept their capability of colony formation as hematopoietic stem cells without inhibition of the proliferation and differentiation 6210

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

abilities. In contrast, no colony forming was observed in blood mononuclear cells before magnetic separation used as a control. Immunostimulation of CD14+ Cells (Monocyte Fraction) after Magnetic Separation. To investigate the inhibitory effect of magnetic separation on CD14+ cells (monocytes), the cells separated by using mouse anti-CD14 antibody and protein A-BacMP binding with anti-mouse IgG antibody were cultured with or without LPS. Cytokines released in the culture supernatant of the CD14+ cells were measured by ELISA (Figure 4). The induction of IL1-β, TNFR, and IL6 was observed in the presence of 1 ng/mL LPS in all fractions. The higher induction of cytokines was obtained in the positive fraction as compared with cells in the negative fraction. A slight difference between cytokine inductions by cells before magnetic separation and cells in the positive fraction was observed. The induction of IL1-β in the absence of LPS was not detected in each fraction. This result indicates that CD14+ cells were efficiently separated enough not to observe the induction of IL1-β in the negative fraction. On the other hand, TNFR and IL6 were slightly detected in the absence of LPS in the positive fraction. This phenomenon suggests that the binding

Figure 2. Microscopic photographs of May-Grunwald Giemsastained cells. Human peripheral blood mononuclear cells were reacted with anti-CD14 antibody and anti-CD34 antibody as primary antibodies and protein A-BacMP binding with anti-mouse IgG antibody. Cells before magnetic separation (A), isolated CD14+ cells (B), and CD34+ cells (C) prepared by air-dried cytocentrifuge were stained with MayGrunwald-Giemsa. The arrows indicate the lymphocytes and the arrowheads indicate monocytes. Scale bar, 20 µm.

of BacMPs on CD14+ cells will potentially induce TNFR and IL6 release. To investigate the effect of BacMP binding on TNFR and IL6 induction, TNFR and IL6 concentrations in the culture supernatant of CD14+ cells separated by BacMPs were compared with those by artificial magnetic particles. TNFR and IL6 concentrations in the culture supernatant of CD14+ cells separated by BacMPs were a similar level with that by artificial magnetic particles (data not shown). This result suggests that BacMPs have as little immunopotentiation to cells as that of artificial magnetic particles. DISCUSSION The development of efficient separation of undifferentiated cells is required for regenerative medical techniques. In research on human hematopoiesis, typically, cell surface antigens (such as CD34) have been used to separate putative hematopoietic stem cell populations within the mixed cell population, and cell sorting methods are used to enrich the cell of interest.25,26 In this study, (25) Sutherland, H. J.; Lansdorp, P. M.; Henkelman, D. H.; Eaves, A. C.; Eaves, C. J. Proc. Natl. Acad. Sci. U.S.A. 1990, 87, 3584-3588. (26) Andrews, R. G.; Singer, J. W.; Bernstein, I. D. J. Exp. Med. 1989, 169, 17211731.

Figure 3. Microscopic photographs of hematopoietic colonies. Stem cells purified from human peripheral blood mononuclear cells were reacted with anti-CD34 antibody and protein A-BacMP binding antimouse IgG antibody. Cells of stem cell fraction formed colonies in semisolid media for 14 days. (A) colony-forming granulocyte/ macrophage. (B) Burst-forming erythroid. Scale bar, 200 µm.

target cells incubated with primary antibody against CD34 were successfully separated using protein A-BacMP binding with antimouse IgG antibody. This result suggests that BacMPs are useful for the separation of hematopoietic and differentiated stem cells in regenerative medical techniques. Magnetic beads technology allows for the simple, rapid, and efficient enrichment of target cell populations. In general, nanosized magnetic particles, rather than microsized ones, are preferred for cell separation because separated cells using nanosized magnetic particles are subsequently useful for flow cytometric analysis,27,28 and microsized magnetic particles probably have an inhibitory effect on cell growth and differentiation after magnetic separation. Several cell separation systems using nanosized magnetic particles, such as 50-nm iron oxide particles with polysaccharide (Miltenyi Biotech) and dextran-coated superparamagnetic nanoparticles (StemCell Technologies Inc.), are commercially available. Because these particles are superparamagnetic and are preferred in high-gradient magnetic separation, especially designed magnetic columns are required for cell separation to produce a high magnetic field gradient.29,30 BacMP used in this study consists of ferrimagnetic iron oxide covered with a lipid (27) Graepler, F.; Lauer, U.; Gregor, M. J. Biochem. Biophys. Methods 1998, 36, 143-155. (28) Simon, J. C.; Dittmar, H. C.; de Roche, R.; Wilting, J.; Christ, B.; Schopf, E. Exp. Dermatol. 1995, 4, 155-161. (29) Miltenyi, S.; Muller, W.; Weichel, W.; Radbruch, A. Cytometry 1990, 11, 231-238. (30) Miltenyi, S. U.S. Patent 5411863, 1995.

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

6211

>95,5 >96,4 96,6 and 98.57%,7 respectively. They indicate that BacMPs are useful for a simple, rapid, and efficient cell separation which is applicable to an automated system. On the other hand, several membrane proteins were shown to be specifically localized onto the BacMP membrane.16,18,33,34 Therefore, BacMPs have potentially immunogenicity and antigenicity against mammalian cells. Because it is hypothesized that the BacMP membrane is derived from the cytoplasm through an invagination process,16 the immunogenicity and antigenicity would be much lower than those of the bacterial outer membrane. In this study, BacMPs did not influence cell proliferation and differentiation of CD34+ and had little immunopotentiation to CD14+ cells. Further analysis of the immunogenicity and antigenicity of BacMPs should be investigated, and BacMP variants with reduced immunogenicity and antigenicity should be developed. Until now, preparation of protein-free BacMPs has been examined35 to eliminate the potential antigenicity. In the most recent work, an in vitro integration technique of useful protein on BacMPs has been developed using anchor molecules. Membrane protein, MagA, or amphiphilic peptides act as molecular anchors for assembling useful proteins. Luciferase-MagA fusion proteins were successfully integrated into BacMP membrane in vitro.36 Furthermore, a spontaneous integration of amphiphilic peptides as anchor molecules enables us to assemble streptavidin on the BacMP membrane in a few minutes.35 These techniques will enable us to eliminate the antigenicity of BacMPs although they may be further investigated and optimized in future work.

Figure 4. Cytokine induction in the culture supernatants of CD14+ cells (monocytes fraction) obtained by magnetic cell separation. CD14+ cells (monocytes fraction) purified from human peripheral blood mononuclear cells were reacted with anti-CD14 antibody and protein A-BacMP binding with anti-mouse IgG antibody and then separated by the magnet into positive and negative subpopulation. These cells were cultured with or without LPS and the concentrations of TNFR (top panel), IL6 (middle panel), and IL1β (bottom panel) in the culture supernatant were measured by ELISA. Cytokines induced by cells before magnetic separation were also measured as controls.

bilayer, which provides good dispersion of BacMPs in aqueous solutions.14 Consequently, BacMPs behave as paramagnetic particles in aqueous solutions despite ferromagnetic particles; i.e., BacMPs are easily separated from cell suspensions using a permanent magnet and require no special magnetic columns. A Nd-Fe-B magnet, which is the strongest permanent magnet commercially available, was used to collect BacMPs in this study. These properties have enabled us to apply BacMPs to a fully automated assay system.14,31,32 In this study, little effect of BacMPs on light-scattering properties of cells was observed during flow cytometric analysis. By using a BacMP-based separation system, the purity of target cells was routinely about 97.5 ( 1.7% for CD19+, 97.6 ( 1.4% for CD20+, 91.2 ( 3.0% for CD8+, and 88.6 ( 2.2% for CD14+ with the positive selection procedure. These results are similar to the previous reports by FACS analysis, where the purity of sorted CD19+, CD20+, CD8+, and CD14+ cells were (31) Tanaka, T.; Matsunaga, T. Biosens. Bioelectron. 2001, 16, 1089-1094. (32) Tanaka, T.; Maruyama, K.; Yoda, K.; Nemoto, E.; Udagawa, Y.; Nakayama, H.; Takeyama, H.; Matsunaga, T. Biosens. Bioelectron. 2003, 19, 325-330.

6212

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

CONCLUSIONS Magnetic separation using BacMPs expressing protein A binding with anti-mouse IgG antibody has been demonstrated. Specific separation of target cells using BacMPs was achieved by simple magnetic separation from cell suspensions using a permanent magnet and with no special magnetic columns. Furthermore, little effect of BacMPs on light-scattering properties of cells was observed during flow cytometric analysis. These properties enable us to apply BacMPs to a fully automated system with a magnetic cell separation process and subsequent FACS analysis process. In addition, the proliferation and differentiation of stem cells separated using BacMPs was observed. The use of BacMPs had no inhibitory effect on the cells. To attain more highly specific separation of CD8+ and CD14+, optimization of immunoassay conditions and use of monoclonal antibody binding with protein A-BacMPs will be required. The interaction between protein A and IgG is not equivalent for all species or all antibody subclasses. For example, protein A is scarcely able to bind to rat immunoglobulin G. Further study to express a single-chain fragment variable (scFv), composed of the variable heavy chain and light chain, on BacMPs by a gene fusion technique may solve these problems. (33) Nakamura, C.; Burgess, J. G.; Sode, K.; Matsunaga, T. J. Biol. Chem. 1995, 270, 28392-28396. (34) Okamura, Y.; Takeyama, H.; Matsunaga, T. Appl. Biochem. Biotechnol. 2000, 84-86, 441-446. (35) Tanaka, T.; Takeda, H.; Kokuryu, Y.; Matsunaga, T. Anal. Chem. 2004, 76, 3764-3769. (36) Matsunaga, T.; Arakaki, A.; Takahoko, M. Biotechnol. Bioeng. 2002, 77, 614-618.

ACKNOWLEDGMENT This work was funded in part by a Grant-in-Aid for Specially Promoted Research, No. 13002005 from the Ministry of Education, Culture, Sports, Science and Technology.

Received for review April 28, 2004. Accepted August 18, 2004. AC0493727

Analytical Chemistry, Vol. 76, No. 21, November 1, 2004

6213