Magnetic Cell Sorting Purification of Differentiated Embryonic Stem Cells Stably Expressing Truncated Human CD4 as Surface Marker
http://www.100md.com
《干细胞学杂志》
University Clinic, Med I, Munich, Germany
Key Words. ES cells ? Magnetic cell sorting ? CD4 ? Cell transplantation
Correspondence: Wolfgang-Michael Franz, M.D., Ph.D, Klinikum Gro?hadern, Marchioninistra?e 15, 81377 München, Germany. Telephone: 49-89-7095-6095; Fax: 49-89-7095-6094; e-mail: wolfgang.franz@med.uni-muenchen.de
ABSTRACT
Embryonic stem (ES) cells offer a huge potential in the field of regenerative medicine as well as tissue engineering because of their capacity to produce every cell and tissue type in vitro. In the future, the treatment of human diseases may be revolutionized by the ability to generate any cell, tissue, or even organ in the laboratory. Yet several obstacles must be overcome to bring ES-derived cell or tissue types to a clinical application. One major hindrance is that no approach has yet yielded a highly pure population of a single functional transplantable cell type. This, however, is an essential prerequisite to avoid the risk of teratoma and tumor formation and further impairment of tissue function, which may result from the implantation of undifferentiated ES cells or of undesired differentiated cell types . Therefore, an efficient means to purify the desired population is required .
Methods such as fluorescence-activated cell sorting (FACS) or magnetic cell sorting (MACS) allow such purification but are dependent on the expression of a specific surface marker that can be recognized by a fluorescent or magnetic microbead-tagged antibody. To be fully effective, an endogenous marker needs to be absolutely cell-type specific. In many cases, however, such as cardiomyocytes, an appropriate endogenous marker is not known, and sorting methods have to rely on the introduction of a marker gene under the control of a lineage-specific promoter. In such an effort, we previously labeled and FACS purified ventricular cardiomyocytes expressing enhanced green fluorescent protein (EGFP) . However, cytometry is slow and typically capable of analyzing no more than 3,000 cells per second for sorting purities of greater than 95% with yields of 50%–70%. In cases of myocardial infarction with a 10% necrosis, approximately 40 g of viable myocardium may be necessary. With an average weight of 80 ng per single cardiomyocyte, more than 108 cells are required for transplantation, resulting in a theoretical purification period of more than 500 hours . Therefore, FACS does not seem to provide the capabilities to identify a rare population of cells or to separate large numbers of cells because of the excessive amount of analysis and sorting time.
Alternative approaches relying on the introduction of a drug-resistance gene rather than a fluorescent protein for antibiotic selection are critical because of the long incubation period with the hazard of resistance and possible harmful effects of the antibiotic on terminally differentiated cells themselves. Additionally, FACS based on fluorescent markers as well as antibiotic selection rely on the expression of nonhuman proteins that may cause additional immunological problems or even be toxic in patients.
To overcome the above described obstacles, we have established a protocol for the labeling and isolation of stably transfected ES cells based on MACS, which is currently regarded to be the gold standard for mild and time-sparing cell purification. Using MACS, up to 1011 cells can be analyzed in approximately 1 hour, making it possible to separate large cell numbers and to identify even rare populations of cells. Our method relies on the expression of an intracellular truncated human CD4 surface antigen, thereby making an immunogenic potential unlikely. It may become an important tool for high-yield selection of specific cell types, providing the basis for future cell transplantation therapy.
MATERIALS AND METHODS
To establish MACS for genetically engineered ES cells, we generated a CD4EGFP fusion construct simultaneously offering the advantages of in vivo detectability via green fluorescence. This construct consists of CD4 and EGFP (CD4EGFP), the latter thereby replacing the natural intracellular domain of the native CD4 molecule (Fig. 1A). The plasmid contains an SP-6 promoter for in vitro translation and a CMV promoter for expression in eucaryotic cells. Expression of CD4EGFP in vitro yielded a protein of the correct size of 74 kD (Fig. 2A), and injection into Xenopus embryos clearly led to the expected membrane-bound localization of the fluorescent fusion protein (Figs. 2B, 2C).
Figure 1. Constructs used for in vitro translation and expression of the CD4EGFP fusion protein in Xenopus embryos (A), for its stable expression in embryonic stem cells (B), and for expression of the CD4 protein in embryonic stem cells (C). Abbreviations: CMV, cytomegalovirus; EGFP, enhanced green fluorescent protein; MCS, multiple cloning site; PGK, phosphoglycerate kinase.
Figure 2. Functionality of the CD4EGFP fusion protein. (A): In vitro translation of CD4EGFP.(B, C): CD4EGFPshowscellmem-brane localization in Xenopus embryo cells compared with EGFP. (D, E): CD4EGFP shows cell membrane localization in GSES cells compared with EGFP. Abbreviation: EGFP, enhanced green fluorescent protein.
Next we planned to transfect murine GSES cells with CD4EGFP. Because viral promoters such as CMV often show extremely high yet unstable activity in ES cells, we generated the plasmid pPGK-CD4EGFP (Fig. 1B). After electroporation and selection, four stable G418-resistant clones were obtained. As expected, these clones displayed membrane-bound EGFP fluorescence, confirming the functionality of the fusion construct (Figs. 2D, 2E). In the various clones, 50.2%–63.1% of the cells were detectable in FACS via antibodies directed against the CD4 part of the fusion protein. Thereby, all of the EGFP-positive cells were contained within this fraction. However, a small proportion of the positively stained cells displayed no or weak EGFP fluorescence, demonstrating that detection of the fusion construct via the CD4 antibody is more sensitive than its EGFP fluorescence.
We now tested the feasibility of MACS using CD4EGFP-positive cells differentiated for 3 days. To reduce their percentage to a more realistic number with respect to future applications using specific promoters, we diluted the labeled cells to 5.9% with native differentiated GSES before MACS purification. This initial population of differentiated cells displaying the CD4 antigen by FACS was enriched to 96.2% (Fig. 3A). Again, all of the EGFP-positive cells were present within this fraction, and some positively stained cells showed a weak or even no fluorescence. Likewise, use of undifferentiated ES cells yielded results comparable to those described above with a purity of 97.3%, whereby the initial population consisted of 13.7% positive cells (data not shown).
Figure 3. (A): FACS analysis of 3-day differentiated CD4EGFP-expressing GSES cells after MACS plotting phycoerythrin-fluorescence against EGFP fluorescence. Enrichment from 5.9%–96.2% is shown. (B): FACS analysis of 3-day differentiated CD4-expressing GSES cells after MACS. Enrichment from 0.6%–98.5% is shown. (C): FACS analysis of 12-day differentiated CD4-expressing GSES cells after MACS. Enrichment from 10.7%–96.1% is shown. Abbreviations: EGFP, enhanced green fluorescent protein; FACS, fluorescence-activated cell sorting; FCS, forward scatter; MACS, magnetic cell sorting.
Our next goal was to verify the data using the nonimmunogenic CD4 molecule without an intracellular EGFP part as well as to extend our studies to lower percentages of CD4-positive cells. GSES cells were transfected with pPGK-CD4 (Fig. 1C), and after selection, four stable G418-resistant clones were obtained expressing CD4 on their cell surface as determined via FACS analysis. In the various clones, 37.7%–53.8% were CD4 positive.
Differentiated cells of clone 6, in which CD4 had been detected on 38.5% of the cells, were then subjected to MACS. We hereby investigated the feasibility of MACS using the stably CD4-expressing ES cells at early and late time points of differentiation. Again we compared different percentages of CD4-positive cells before the purification. Figure 3B shows that even 0.6% of positive cells within the whole population of ES cells after 3 days of differentiation (early time point) led to a 98.5% pure population of CD4-expressing cells. When the population before MACS contained 7.3% positive cells in a parallel experiment, a purity of 96.8% was achieved (data not shown). Equivalent results leading to a purity of greater than 96% were achieved at a later time point of differentiation (12 days), whereby the initial positive population was 10.7% (Fig. 3C). Likewise, when we used undifferentiated CD4-positive cells at initial percentages between 3.8% and 15.8%, purities between 97.6% and 98.4% of cells were achieved (data not shown).
To verify the viability of the MACS-selected cells, we recultivated those after the purification procedure, which was performed at day 3 of differentiation. Figure 4A shows single dissociated CD4EGFP-positive cells after the elution of the MACS column. These cells were able to reaggregate and form normal fluorescent EBs (Figs. 4B, 4C). Likewise, CD4-expressing cells formed normal EBs (data not shown). To verify the normal development of these reaggregated EBs, we subjected them to RT-PCR experiments. Figure 4D shows that CD4EGFP as well as CD4-expressing cells at day 4 after the MACS purification did express markers for all three germ layers compared with cells that were not MACS purified.
Figure 4. (A): Dissociated CD4EGFP-expressing GSES cells after MACS performed at day 3 of differentiation. (B): Reaggregated EBs in suspension formed by the cells shown in (A) at day 4 after the MACS procedure. (C): Reaggregated EBs formed by the cells shown in (A) at day 12 after the MACS procedure. (E): Reverse transcription–polymerase chain reaction analysis of cells expressing CD4 () and CD4EGFP (E) grown without MACS versus the same cells MACS purified at day 3, reaggregated, and further grown until day 4 after MACS. MACS does not influence the expression of endodermal (HNF-4), mesodermal (Brachyury), and ectodermal (Involucrin, Neurogenin) markers. Abbreviations: EBs, embryoid bodies; MACS, magnetic cell sorting.
DISCUSSION
The great medical potentials of ES cells are impeded so far by the lack of an efficient nonimmunogenic way to isolate the desired cell types meant for transplantation. We have labeled ES cells stably expressing human CD4 truncated for its intracellular domain (CD4) for MACS. To track the labeled cells, CD4 was fused to an intracellular EGFP rest (CD4EGFP). We prove functionality of CD4EGFP and its suitability for MACS yielding purities greater than 97% when expressed via the PGK promoter. Likewise, expression of CD4 led to greater than 98% positive viable cells, which was not affected by their differentiation state. Additionally, purities were not influenced by the initial content of CD4-expressing cells, ranging from 0.6%–16%. After the MACS procedure, the cells were able to reaggregate and form normal EB-expressing markers of all three embryonic germ layers, proving their viability. The application of our technique in differentiated ES-derived cell types may allow the rapid purification of a desired cell type with high yields, thereby providing an important basis for tissue engineering and cell transplantation studies.
ACKNOWLEDGMENTS
Lawrenz B, Schiller H, Willbold E et al. Highly sensitive biosafety model for stem-cell-derived grafts. Cytotherapy 2004;6:212–222.
Wakitani S, Takaoka K, Hattori T et al. Embryonic stem cells injected into the mouse knee joint form teratomas and subsequently destroy the joint. Rheumatology (Oxford) 2003;42:162–165.
Conley BJ, Young JC, Trounson AO et al. Derivation, propagation and differentiation of human embryonic stem cells. Int J Biochem Cell Biol 2004;36:555–567.
Rippon HJ, Bishop AE. Embryonic stem cells. Cell Prolif 2004;37:23–34.
Muller M, Fleischmann BK, Selbert S et al. Selection of ventricular-like cardiomyocytes from ES cells in vitro. FASEB J 2000;14:2540–2548.
Kudriavtsev BN, Anatskaia OV, Nilova VK et al. Interconnection of parameters of the mitochondrial and myofibrillar apparatus of cardiomyocytes and ploidy and hypertrophy in certain mammalian species, differing in body mass . Tsitologiia 1997;39:946–964.
Klug MG, Soonpaa MH, Koh GY et al. Genetically selected cardiomyocytes from differentiating embryonic stem cells form stable intracardiac grafts. J Clin Invest 1996;98:216–224.
Zandstra PW, Bauwens C, Yin T et al. Scalable production of embryonic stem cell-derived cardiomyocytes. Tissue Eng 2003;9:767–778.
Zweigerdt R, Burg M, Willbold E et al. Generation of confluent cardiomyocyte monolayers derived from embryonic stem cells in suspension: a cell source for new therapies and screening strategies. Cytotherapy 2003;5:399–413.
Rupp RA, Snider L, Weintraub H. Xenopus embryos regulate the nuclear localization of XMyoD. Genes Dev 1994;8:1311–1323.
David R, Joos TO, Dreyer C. Anteroposterior patterning and organogenesis of Xenopus laevis require a correct dose of germ cell nuclear factor (xGCNF). Mech Dev 1998;79:137–152.
Miltenyi S, Muller W, Weichel W et al. High gradient magnetic cell separation with MACS. Cytometry 1990;11:231–238.
Siebenkotten G, Petry K, Behrens-Jung U et al. Employing surface markers for the selection of transfected cells. In: Recktenwald D, Radbruch A, eds. Cell Separation: Methods and Applications. New York: Marcel Dekker, 1998:271–281.
Gaines P, Wojchowski DM. pIRES-CD4t, a dicistronic expression vector for MACS- or FACS-based selection of transfected cells. Biotechniques 1999;26:683–638.(Robert David, Michael Gro)
Key Words. ES cells ? Magnetic cell sorting ? CD4 ? Cell transplantation
Correspondence: Wolfgang-Michael Franz, M.D., Ph.D, Klinikum Gro?hadern, Marchioninistra?e 15, 81377 München, Germany. Telephone: 49-89-7095-6095; Fax: 49-89-7095-6094; e-mail: wolfgang.franz@med.uni-muenchen.de
ABSTRACT
Embryonic stem (ES) cells offer a huge potential in the field of regenerative medicine as well as tissue engineering because of their capacity to produce every cell and tissue type in vitro. In the future, the treatment of human diseases may be revolutionized by the ability to generate any cell, tissue, or even organ in the laboratory. Yet several obstacles must be overcome to bring ES-derived cell or tissue types to a clinical application. One major hindrance is that no approach has yet yielded a highly pure population of a single functional transplantable cell type. This, however, is an essential prerequisite to avoid the risk of teratoma and tumor formation and further impairment of tissue function, which may result from the implantation of undifferentiated ES cells or of undesired differentiated cell types . Therefore, an efficient means to purify the desired population is required .
Methods such as fluorescence-activated cell sorting (FACS) or magnetic cell sorting (MACS) allow such purification but are dependent on the expression of a specific surface marker that can be recognized by a fluorescent or magnetic microbead-tagged antibody. To be fully effective, an endogenous marker needs to be absolutely cell-type specific. In many cases, however, such as cardiomyocytes, an appropriate endogenous marker is not known, and sorting methods have to rely on the introduction of a marker gene under the control of a lineage-specific promoter. In such an effort, we previously labeled and FACS purified ventricular cardiomyocytes expressing enhanced green fluorescent protein (EGFP) . However, cytometry is slow and typically capable of analyzing no more than 3,000 cells per second for sorting purities of greater than 95% with yields of 50%–70%. In cases of myocardial infarction with a 10% necrosis, approximately 40 g of viable myocardium may be necessary. With an average weight of 80 ng per single cardiomyocyte, more than 108 cells are required for transplantation, resulting in a theoretical purification period of more than 500 hours . Therefore, FACS does not seem to provide the capabilities to identify a rare population of cells or to separate large numbers of cells because of the excessive amount of analysis and sorting time.
Alternative approaches relying on the introduction of a drug-resistance gene rather than a fluorescent protein for antibiotic selection are critical because of the long incubation period with the hazard of resistance and possible harmful effects of the antibiotic on terminally differentiated cells themselves. Additionally, FACS based on fluorescent markers as well as antibiotic selection rely on the expression of nonhuman proteins that may cause additional immunological problems or even be toxic in patients.
To overcome the above described obstacles, we have established a protocol for the labeling and isolation of stably transfected ES cells based on MACS, which is currently regarded to be the gold standard for mild and time-sparing cell purification. Using MACS, up to 1011 cells can be analyzed in approximately 1 hour, making it possible to separate large cell numbers and to identify even rare populations of cells. Our method relies on the expression of an intracellular truncated human CD4 surface antigen, thereby making an immunogenic potential unlikely. It may become an important tool for high-yield selection of specific cell types, providing the basis for future cell transplantation therapy.
MATERIALS AND METHODS
To establish MACS for genetically engineered ES cells, we generated a CD4EGFP fusion construct simultaneously offering the advantages of in vivo detectability via green fluorescence. This construct consists of CD4 and EGFP (CD4EGFP), the latter thereby replacing the natural intracellular domain of the native CD4 molecule (Fig. 1A). The plasmid contains an SP-6 promoter for in vitro translation and a CMV promoter for expression in eucaryotic cells. Expression of CD4EGFP in vitro yielded a protein of the correct size of 74 kD (Fig. 2A), and injection into Xenopus embryos clearly led to the expected membrane-bound localization of the fluorescent fusion protein (Figs. 2B, 2C).
Figure 1. Constructs used for in vitro translation and expression of the CD4EGFP fusion protein in Xenopus embryos (A), for its stable expression in embryonic stem cells (B), and for expression of the CD4 protein in embryonic stem cells (C). Abbreviations: CMV, cytomegalovirus; EGFP, enhanced green fluorescent protein; MCS, multiple cloning site; PGK, phosphoglycerate kinase.
Figure 2. Functionality of the CD4EGFP fusion protein. (A): In vitro translation of CD4EGFP.(B, C): CD4EGFPshowscellmem-brane localization in Xenopus embryo cells compared with EGFP. (D, E): CD4EGFP shows cell membrane localization in GSES cells compared with EGFP. Abbreviation: EGFP, enhanced green fluorescent protein.
Next we planned to transfect murine GSES cells with CD4EGFP. Because viral promoters such as CMV often show extremely high yet unstable activity in ES cells, we generated the plasmid pPGK-CD4EGFP (Fig. 1B). After electroporation and selection, four stable G418-resistant clones were obtained. As expected, these clones displayed membrane-bound EGFP fluorescence, confirming the functionality of the fusion construct (Figs. 2D, 2E). In the various clones, 50.2%–63.1% of the cells were detectable in FACS via antibodies directed against the CD4 part of the fusion protein. Thereby, all of the EGFP-positive cells were contained within this fraction. However, a small proportion of the positively stained cells displayed no or weak EGFP fluorescence, demonstrating that detection of the fusion construct via the CD4 antibody is more sensitive than its EGFP fluorescence.
We now tested the feasibility of MACS using CD4EGFP-positive cells differentiated for 3 days. To reduce their percentage to a more realistic number with respect to future applications using specific promoters, we diluted the labeled cells to 5.9% with native differentiated GSES before MACS purification. This initial population of differentiated cells displaying the CD4 antigen by FACS was enriched to 96.2% (Fig. 3A). Again, all of the EGFP-positive cells were present within this fraction, and some positively stained cells showed a weak or even no fluorescence. Likewise, use of undifferentiated ES cells yielded results comparable to those described above with a purity of 97.3%, whereby the initial population consisted of 13.7% positive cells (data not shown).
Figure 3. (A): FACS analysis of 3-day differentiated CD4EGFP-expressing GSES cells after MACS plotting phycoerythrin-fluorescence against EGFP fluorescence. Enrichment from 5.9%–96.2% is shown. (B): FACS analysis of 3-day differentiated CD4-expressing GSES cells after MACS. Enrichment from 0.6%–98.5% is shown. (C): FACS analysis of 12-day differentiated CD4-expressing GSES cells after MACS. Enrichment from 10.7%–96.1% is shown. Abbreviations: EGFP, enhanced green fluorescent protein; FACS, fluorescence-activated cell sorting; FCS, forward scatter; MACS, magnetic cell sorting.
Our next goal was to verify the data using the nonimmunogenic CD4 molecule without an intracellular EGFP part as well as to extend our studies to lower percentages of CD4-positive cells. GSES cells were transfected with pPGK-CD4 (Fig. 1C), and after selection, four stable G418-resistant clones were obtained expressing CD4 on their cell surface as determined via FACS analysis. In the various clones, 37.7%–53.8% were CD4 positive.
Differentiated cells of clone 6, in which CD4 had been detected on 38.5% of the cells, were then subjected to MACS. We hereby investigated the feasibility of MACS using the stably CD4-expressing ES cells at early and late time points of differentiation. Again we compared different percentages of CD4-positive cells before the purification. Figure 3B shows that even 0.6% of positive cells within the whole population of ES cells after 3 days of differentiation (early time point) led to a 98.5% pure population of CD4-expressing cells. When the population before MACS contained 7.3% positive cells in a parallel experiment, a purity of 96.8% was achieved (data not shown). Equivalent results leading to a purity of greater than 96% were achieved at a later time point of differentiation (12 days), whereby the initial positive population was 10.7% (Fig. 3C). Likewise, when we used undifferentiated CD4-positive cells at initial percentages between 3.8% and 15.8%, purities between 97.6% and 98.4% of cells were achieved (data not shown).
To verify the viability of the MACS-selected cells, we recultivated those after the purification procedure, which was performed at day 3 of differentiation. Figure 4A shows single dissociated CD4EGFP-positive cells after the elution of the MACS column. These cells were able to reaggregate and form normal fluorescent EBs (Figs. 4B, 4C). Likewise, CD4-expressing cells formed normal EBs (data not shown). To verify the normal development of these reaggregated EBs, we subjected them to RT-PCR experiments. Figure 4D shows that CD4EGFP as well as CD4-expressing cells at day 4 after the MACS purification did express markers for all three germ layers compared with cells that were not MACS purified.
Figure 4. (A): Dissociated CD4EGFP-expressing GSES cells after MACS performed at day 3 of differentiation. (B): Reaggregated EBs in suspension formed by the cells shown in (A) at day 4 after the MACS procedure. (C): Reaggregated EBs formed by the cells shown in (A) at day 12 after the MACS procedure. (E): Reverse transcription–polymerase chain reaction analysis of cells expressing CD4 () and CD4EGFP (E) grown without MACS versus the same cells MACS purified at day 3, reaggregated, and further grown until day 4 after MACS. MACS does not influence the expression of endodermal (HNF-4), mesodermal (Brachyury), and ectodermal (Involucrin, Neurogenin) markers. Abbreviations: EBs, embryoid bodies; MACS, magnetic cell sorting.
DISCUSSION
The great medical potentials of ES cells are impeded so far by the lack of an efficient nonimmunogenic way to isolate the desired cell types meant for transplantation. We have labeled ES cells stably expressing human CD4 truncated for its intracellular domain (CD4) for MACS. To track the labeled cells, CD4 was fused to an intracellular EGFP rest (CD4EGFP). We prove functionality of CD4EGFP and its suitability for MACS yielding purities greater than 97% when expressed via the PGK promoter. Likewise, expression of CD4 led to greater than 98% positive viable cells, which was not affected by their differentiation state. Additionally, purities were not influenced by the initial content of CD4-expressing cells, ranging from 0.6%–16%. After the MACS procedure, the cells were able to reaggregate and form normal EB-expressing markers of all three embryonic germ layers, proving their viability. The application of our technique in differentiated ES-derived cell types may allow the rapid purification of a desired cell type with high yields, thereby providing an important basis for tissue engineering and cell transplantation studies.
ACKNOWLEDGMENTS
Lawrenz B, Schiller H, Willbold E et al. Highly sensitive biosafety model for stem-cell-derived grafts. Cytotherapy 2004;6:212–222.
Wakitani S, Takaoka K, Hattori T et al. Embryonic stem cells injected into the mouse knee joint form teratomas and subsequently destroy the joint. Rheumatology (Oxford) 2003;42:162–165.
Conley BJ, Young JC, Trounson AO et al. Derivation, propagation and differentiation of human embryonic stem cells. Int J Biochem Cell Biol 2004;36:555–567.
Rippon HJ, Bishop AE. Embryonic stem cells. Cell Prolif 2004;37:23–34.
Muller M, Fleischmann BK, Selbert S et al. Selection of ventricular-like cardiomyocytes from ES cells in vitro. FASEB J 2000;14:2540–2548.
Kudriavtsev BN, Anatskaia OV, Nilova VK et al. Interconnection of parameters of the mitochondrial and myofibrillar apparatus of cardiomyocytes and ploidy and hypertrophy in certain mammalian species, differing in body mass . Tsitologiia 1997;39:946–964.
Klug MG, Soonpaa MH, Koh GY et al. Genetically selected cardiomyocytes from differentiating embryonic stem cells form stable intracardiac grafts. J Clin Invest 1996;98:216–224.
Zandstra PW, Bauwens C, Yin T et al. Scalable production of embryonic stem cell-derived cardiomyocytes. Tissue Eng 2003;9:767–778.
Zweigerdt R, Burg M, Willbold E et al. Generation of confluent cardiomyocyte monolayers derived from embryonic stem cells in suspension: a cell source for new therapies and screening strategies. Cytotherapy 2003;5:399–413.
Rupp RA, Snider L, Weintraub H. Xenopus embryos regulate the nuclear localization of XMyoD. Genes Dev 1994;8:1311–1323.
David R, Joos TO, Dreyer C. Anteroposterior patterning and organogenesis of Xenopus laevis require a correct dose of germ cell nuclear factor (xGCNF). Mech Dev 1998;79:137–152.
Miltenyi S, Muller W, Weichel W et al. High gradient magnetic cell separation with MACS. Cytometry 1990;11:231–238.
Siebenkotten G, Petry K, Behrens-Jung U et al. Employing surface markers for the selection of transfected cells. In: Recktenwald D, Radbruch A, eds. Cell Separation: Methods and Applications. New York: Marcel Dekker, 1998:271–281.
Gaines P, Wojchowski DM. pIRES-CD4t, a dicistronic expression vector for MACS- or FACS-based selection of transfected cells. Biotechniques 1999;26:683–638.(Robert David, Michael Gro)