H U M A N G E N E T H E R A P Y 2:323-326 (1991) Mary A n n Liebert, Inc., Publishers
C r y o p r e s e r v e d P r i m a t e B o n e M a r r o w Cells C a n B e U s e d for Retroviral-Mediated G e n e
Transfer
ROBERT WIEDER1'2
ABSTRACT Retro viral-mediated gene transfer into Rhesus monkey bone marrow cells, which were cryopreserved, stored, and then transduced at the time of thawing, was studied for potential application in gene therapy protocols. Albumin density gradient fractionation was used to define subpopulations of cryopreserved cells transduced by a murine retroviral vector. T h e transfer of the bacterial neomycin phosphotransferase gene into Rhesus m o n k e y bone m a r r o w that w a s cryopreserved and thawed w a s found to be preferential in a light-density population enriched for granulocyte-macrophage colony-forming units ( C F U - G M ) . These results are similar to results obtained with freshly harvested bone m a r r o w in which this population is enriched for C D 3 4 antigenpositive cells, as well as for cells that were undergoing cell division. This method m a y be useful in enriching for transduced precursors in future gene transfer experiments in primates.
OVERVIEW S U M M A R Y
1985; Dick etaL, 1985; Keller etaL, 1985;Capele/a/., 1989), primate experiments have only been partially successful (Kantoff et al., 1987). Although both myeloid and lymphoid lineages Some human gene transfer/therapy clinical protocols will require that the patient's cells be frozen and stored prior to were transduced in the freshly harvested bone marrow of some gene transfer. Wieder demonstrates that primate (rhesus of the monkeys in the published study, activity was low and monkey) bone marrow cells that have been stored in liquid invariably disappeared 6 months after transplant. This result nitrogen can be used successfully for retroviral-mediated was attributed to the transduction of a committed population that underwent terminal differentiation and senescence. Previgene transfer. ous data demonstrate that the in vitro transduction efficiency of early progenitor populations increased in primate bone marrow sampled during the recovery period following intravenous adINTRODUCTION ministration of 5-fluorouracil (Wieder et aL, 1991). Taking advantage of the increasing density of hematopoietic One OF THE potential applications of retroviral-mediated precursors with progressive differentiation, density fractiongene transfer will be for the treatment of genetic diseases ation (Wieder et al., 1991) was used to define and partially based on a single gene defect (Anderson, 1984). The scenario purify the populations that were transduced by N2, a retroviral may involve collection and freezing of the patient's own bone vector coding for a bacterial neomycin phosphotransferase gene marrow, followed by gene transfer and autologous bone mar- (NPT) (Armentano et aL, 1987). This population corresponded row transplantation after more conventional therapeutic modal- to light-density cells shown to be preferentially infected by ities have failed. The potential for transducing bone marrow naturally occurring retroviruses in wild mice (Buckheit et aL, that has been cryopreserved and subsequently thawed is a sub- 1988). Reconstitution experiments in primates have shown that ject of this study. While hematopoietic reconstitution with these light-density cells were up to 20-fold enriched for stem transduced marrow and long-term somatic expression of the cells (G. Wagemaker, personal communication). Others have transduced gene has been successful in mice (Eglitis et aL, demonstrated that lighter-density human bone marrow cell pop-
1 Molecular Hematology Branch, NHLBI, NIH, Bethesda, M D 20892. 2Present address: Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, N Y 10021.
323
324
WIEDER
ulations are enriched for early progenitors (Gabbianelli et aL, 1990; Sutherland et aL, 1990). In this study, albumin density gradients were used in conjunction with granulocyte-macrophage colony-forming units ( C F U - G M ) colony assays to demonstrate for the first time that cryopreserved bone marrow may be transduced by a retroviral vector. The goal was to determine if bone marrow progenitors that had been cryopreserved could be used as target cells in future gene transfer experiments.
MATERIALS A N D M E T H O D S
Colony assays Cells were plated in 1-ml volumes of 1% methylcellulose 3 0 % heat-inactivated fetal calf serum (HI F C S ) , 5 % lymphocyte-conditioned medium of a frequently phlebotomized hemochromatosis patient ( L C M ) , 1 0 % (wt/vol) B S A , 2 mmoles/liter glutamine and 12 mmoles/liter 2-mercaptoethanol (2-ME), 3.5 U/ml erythropoietin (Amgen, Thousand Oaks, C A ) with and without 0.8 mg/ml G 4 1 8 (a concentration shown to be lethal for nontransduced cells). C F U - G M and erythrocyte burst-forming units (BFU-E) containing an estimated 40 or more cells were counted at 14 ± 2 days (Eglitis etaL, 1985).
B o n e m a r r o w specimens
RESULTS
All animal studies were approved by the NHLBI Animal Care and Use Committee. Bone marrow cells for the cryopreservation experiments were obtained from a normal rhesus monkey that was sacrificed as a donor for a heart transplant in a separate experiment. Mononuclear cells were separated by centrifugation with Lymphocyte Separation Medium ( L S M , Litton Bionetics, Kensington, M D ) , program-frozen in 1 ml aliquots at 1°C per min to -40°C in 9 0 % autologous serum/10% D M S O using a Vari-Rate III programmable freezer with a West 2050 controller (Verdis Corporation, Gardner, N Y ) , and stored in liquid nitrogen. At the time of experimental use, cells were quick-thawed at 37°C and immediately washed with PBS. Cells excluding 0.2% trypan blue were counted for use in experiments.
Survival of cryopreserved marrow cells ranged from 26 to 7 0 % , and recovery of C F U - G M after gradient separation ranged from 54 to 128%. Survival after 24 hr cocultivation ranged from 41 to 8 5 % by 0.2% trypan blue exclusion. The C F U - G M plating efficiency of the intact transduced cryopreserved mononuclear cell population was 102 colonies/105 cells plated (Table 1). The plating efficiency of untransduced C F U - G M was 106 colonies/105 mononuclear cells, similar to that of transduced cells. The C F U - G M plating efficiency of cells fractionated by density centrifugation and assayed for colony formation was lower in layer 1 in the transduced population than in untransduced cells. The plating efficiency was less in layers 2 and 3 as well, but the difference was not statistically significant. The distributions of plating efficiency were similar in both populations. The plating efficiency of B F U - E was also similar in the two Albumin density gradients populations. The differences between 27 and 18 in the untransMononuclear cells, which were cryopreserved, were thawed, duced and transduced cells was not statistically significant in washed in P B S , and suspended in 1.5 ml of a 1 7 % (wt/vol) this study. Layer 2 of the transduced population however, had bovine serum albumin (BSA) solution, layered onto a pre- significantly fewer BFU-E/105 cells than the untransduced popformed step gradient consisting of 1.5-ml layers of 2 5 % , 2 3 % , ulation. The differences in layers 3 and 4 were not significant. 2 2 % , and 2 1 % (wt/vol) B S A in a Falcon polystryrene conical The transduction efficiency of all colonies was 1.4%. The tube, and centrifuged at 1,000 x g for 30 min at 20°C, as C F U - G M plating efficiency of density-fractionated cryoprepreviously described (Wieder et aL, 1991). The B S A solutions served mononuclear cells was greatest in layer 4 of the gradient, were a kind gift of Prof. G. Wagemaker (The T N O , The Hague, whereas the transduction efficiency was the highest in layer 3. The Netherlands). The solution interfaces and the bottom of the Layer 5 contained 4 6 % of the cells, 3 8 % of the C F U - G M , and tube, where cells segregated, were labeled as layers 1-5 in only 1 5 % of the G418-resistant (G418R) C F U - C in the transdescending order. After centrifugation, the cells were removed duced population, whereas layers 2, 3, and 4 contained 3 0 % of sequentially from the top using a 1-ml pipette. A parallel con- the cells, 6 1 % of the C F U - G M , and 8 3 % of the G418R C F U - C trol with Path-O-Cyte 4 B S A solutions (ICN ImmunoBiologi- (Fig. 1). The transduction efficiency in layers 1-3 was 5- to cals, Lisle, IL) yielded similar results. All albumin concentra- 7-fold higher than in unfractionated mononuclear cells. The tions were determined by refractive index measurements. C F U - G M in these layers were enriched for their transducible subpopulations of G418R colonies (Fig. 1). Retroviral gene transfer Experiments with retroviral vectors were conducted in compliance with N I H Recombinant D N A Advisory Committee guidelines. Cryopreserved bone marrow cells that had been thawed and separated by albumin gradient fractionation were transduced by 24 hr co-cultivation in 8 |xg/ml Polybrene with N2/PA317 (Miller and Buttimore, 1986; Armentano et al., 1987) producer cells gamma-irradiated with 2,000 cGy in a Gammacell 40 ,37Cs animal irradiator, as previously described (Eglitis era/., 1985;Kantoffe/a/., 1987).
DISCUSSION This study demonstrates that primate bone marrow cells that had been cryopreserved were capable of being transduced by a murine retroviral vector at the time of thawing. This information is important in considering human gene therapy clinical protocols using bone marrow progenitor cells as the target of gene transfer. Albumin density gradient centrifugation was used to enrich bone marrow progenitor cell populations preferentially trans-
RETROVIRAL GENE TRANSFER INTO PRIMATE BONE M A R R O W
325
result of freeze/thawing. The effect is likely one secondary to an increase in the water-to-lipid ratio after thawing and not one due to preferential loss of light-density precursors. O n the contrary, % of total the C F U - G M frequency in cryopreserved cells is slightly increased due to preferential survival of earlier progenitors over more differentiated myeloid cells. The recovery of C F U - G M was 9 1 % (range 5 4 - 1 2 8 % ) with identical patterns of density distribution regardless, of the percentage of recovery in any one particular experiment. C F U - G M transduced by N 2 belonged to a less dense population of C F U - G M . This population m a y be a slightly less differentiated subset in which a larger fraction of cells are undergoing cell cycling than in the whole C F U - G M population. The C F U - G M and B F U - E plating efficiencies of the lower layer 1 layer 2 layer 3 layer 5 layer 4 density populations in cocultivated bone marrow cells appear to be less than those in uninfected cells. This phenomenon m a y be I Cells S CFU-GM C D G418 CFU-GM due to preferential adherence of early progenitors to the supF I G . 1. Distribution of cryopreserved, transduced bone mar- porting monolayer in these low-density, progenitor-rich popurow mononuclear cells, C F U - G M , and G 4 1 8 R C F U - G M in an lations (Dumenil et aL, 1989), and m a y have accounted for albumin density gradient. Values represent the arithmetic prod- hematopoietic nonengraftment in primates transplanted with uct (normalized over the five layers) of the percentage of mono- autologous bone marrow after retroviral gene transfer by coculnuclear cells in a particular layer and the C F U - G M / 1 0 5 cells and tivation(Kantoff
C r y o p r e s e r v e d P r i m a t e B o n e M a r r o w Cells C a n B e U s e d for Retroviral-Mediated G e n e
Transfer
ROBERT WIEDER1'2
ABSTRACT Retro viral-mediated gene transfer into Rhesus monkey bone marrow cells, which were cryopreserved, stored, and then transduced at the time of thawing, was studied for potential application in gene therapy protocols. Albumin density gradient fractionation was used to define subpopulations of cryopreserved cells transduced by a murine retroviral vector. T h e transfer of the bacterial neomycin phosphotransferase gene into Rhesus m o n k e y bone m a r r o w that w a s cryopreserved and thawed w a s found to be preferential in a light-density population enriched for granulocyte-macrophage colony-forming units ( C F U - G M ) . These results are similar to results obtained with freshly harvested bone m a r r o w in which this population is enriched for C D 3 4 antigenpositive cells, as well as for cells that were undergoing cell division. This method m a y be useful in enriching for transduced precursors in future gene transfer experiments in primates.
OVERVIEW S U M M A R Y
1985; Dick etaL, 1985; Keller etaL, 1985;Capele/a/., 1989), primate experiments have only been partially successful (Kantoff et al., 1987). Although both myeloid and lymphoid lineages Some human gene transfer/therapy clinical protocols will require that the patient's cells be frozen and stored prior to were transduced in the freshly harvested bone marrow of some gene transfer. Wieder demonstrates that primate (rhesus of the monkeys in the published study, activity was low and monkey) bone marrow cells that have been stored in liquid invariably disappeared 6 months after transplant. This result nitrogen can be used successfully for retroviral-mediated was attributed to the transduction of a committed population that underwent terminal differentiation and senescence. Previgene transfer. ous data demonstrate that the in vitro transduction efficiency of early progenitor populations increased in primate bone marrow sampled during the recovery period following intravenous adINTRODUCTION ministration of 5-fluorouracil (Wieder et aL, 1991). Taking advantage of the increasing density of hematopoietic One OF THE potential applications of retroviral-mediated precursors with progressive differentiation, density fractiongene transfer will be for the treatment of genetic diseases ation (Wieder et al., 1991) was used to define and partially based on a single gene defect (Anderson, 1984). The scenario purify the populations that were transduced by N2, a retroviral may involve collection and freezing of the patient's own bone vector coding for a bacterial neomycin phosphotransferase gene marrow, followed by gene transfer and autologous bone mar- (NPT) (Armentano et aL, 1987). This population corresponded row transplantation after more conventional therapeutic modal- to light-density cells shown to be preferentially infected by ities have failed. The potential for transducing bone marrow naturally occurring retroviruses in wild mice (Buckheit et aL, that has been cryopreserved and subsequently thawed is a sub- 1988). Reconstitution experiments in primates have shown that ject of this study. While hematopoietic reconstitution with these light-density cells were up to 20-fold enriched for stem transduced marrow and long-term somatic expression of the cells (G. Wagemaker, personal communication). Others have transduced gene has been successful in mice (Eglitis et aL, demonstrated that lighter-density human bone marrow cell pop-
1 Molecular Hematology Branch, NHLBI, NIH, Bethesda, M D 20892. 2Present address: Department of Medicine, Memorial Sloan-Kettering Cancer Center, New York, N Y 10021.
323
324
WIEDER
ulations are enriched for early progenitors (Gabbianelli et aL, 1990; Sutherland et aL, 1990). In this study, albumin density gradients were used in conjunction with granulocyte-macrophage colony-forming units ( C F U - G M ) colony assays to demonstrate for the first time that cryopreserved bone marrow may be transduced by a retroviral vector. The goal was to determine if bone marrow progenitors that had been cryopreserved could be used as target cells in future gene transfer experiments.
MATERIALS A N D M E T H O D S
Colony assays Cells were plated in 1-ml volumes of 1% methylcellulose 3 0 % heat-inactivated fetal calf serum (HI F C S ) , 5 % lymphocyte-conditioned medium of a frequently phlebotomized hemochromatosis patient ( L C M ) , 1 0 % (wt/vol) B S A , 2 mmoles/liter glutamine and 12 mmoles/liter 2-mercaptoethanol (2-ME), 3.5 U/ml erythropoietin (Amgen, Thousand Oaks, C A ) with and without 0.8 mg/ml G 4 1 8 (a concentration shown to be lethal for nontransduced cells). C F U - G M and erythrocyte burst-forming units (BFU-E) containing an estimated 40 or more cells were counted at 14 ± 2 days (Eglitis etaL, 1985).
B o n e m a r r o w specimens
RESULTS
All animal studies were approved by the NHLBI Animal Care and Use Committee. Bone marrow cells for the cryopreservation experiments were obtained from a normal rhesus monkey that was sacrificed as a donor for a heart transplant in a separate experiment. Mononuclear cells were separated by centrifugation with Lymphocyte Separation Medium ( L S M , Litton Bionetics, Kensington, M D ) , program-frozen in 1 ml aliquots at 1°C per min to -40°C in 9 0 % autologous serum/10% D M S O using a Vari-Rate III programmable freezer with a West 2050 controller (Verdis Corporation, Gardner, N Y ) , and stored in liquid nitrogen. At the time of experimental use, cells were quick-thawed at 37°C and immediately washed with PBS. Cells excluding 0.2% trypan blue were counted for use in experiments.
Survival of cryopreserved marrow cells ranged from 26 to 7 0 % , and recovery of C F U - G M after gradient separation ranged from 54 to 128%. Survival after 24 hr cocultivation ranged from 41 to 8 5 % by 0.2% trypan blue exclusion. The C F U - G M plating efficiency of the intact transduced cryopreserved mononuclear cell population was 102 colonies/105 cells plated (Table 1). The plating efficiency of untransduced C F U - G M was 106 colonies/105 mononuclear cells, similar to that of transduced cells. The C F U - G M plating efficiency of cells fractionated by density centrifugation and assayed for colony formation was lower in layer 1 in the transduced population than in untransduced cells. The plating efficiency was less in layers 2 and 3 as well, but the difference was not statistically significant. The distributions of plating efficiency were similar in both populations. The plating efficiency of B F U - E was also similar in the two Albumin density gradients populations. The differences between 27 and 18 in the untransMononuclear cells, which were cryopreserved, were thawed, duced and transduced cells was not statistically significant in washed in P B S , and suspended in 1.5 ml of a 1 7 % (wt/vol) this study. Layer 2 of the transduced population however, had bovine serum albumin (BSA) solution, layered onto a pre- significantly fewer BFU-E/105 cells than the untransduced popformed step gradient consisting of 1.5-ml layers of 2 5 % , 2 3 % , ulation. The differences in layers 3 and 4 were not significant. 2 2 % , and 2 1 % (wt/vol) B S A in a Falcon polystryrene conical The transduction efficiency of all colonies was 1.4%. The tube, and centrifuged at 1,000 x g for 30 min at 20°C, as C F U - G M plating efficiency of density-fractionated cryoprepreviously described (Wieder et aL, 1991). The B S A solutions served mononuclear cells was greatest in layer 4 of the gradient, were a kind gift of Prof. G. Wagemaker (The T N O , The Hague, whereas the transduction efficiency was the highest in layer 3. The Netherlands). The solution interfaces and the bottom of the Layer 5 contained 4 6 % of the cells, 3 8 % of the C F U - G M , and tube, where cells segregated, were labeled as layers 1-5 in only 1 5 % of the G418-resistant (G418R) C F U - C in the transdescending order. After centrifugation, the cells were removed duced population, whereas layers 2, 3, and 4 contained 3 0 % of sequentially from the top using a 1-ml pipette. A parallel con- the cells, 6 1 % of the C F U - G M , and 8 3 % of the G418R C F U - C trol with Path-O-Cyte 4 B S A solutions (ICN ImmunoBiologi- (Fig. 1). The transduction efficiency in layers 1-3 was 5- to cals, Lisle, IL) yielded similar results. All albumin concentra- 7-fold higher than in unfractionated mononuclear cells. The tions were determined by refractive index measurements. C F U - G M in these layers were enriched for their transducible subpopulations of G418R colonies (Fig. 1). Retroviral gene transfer Experiments with retroviral vectors were conducted in compliance with N I H Recombinant D N A Advisory Committee guidelines. Cryopreserved bone marrow cells that had been thawed and separated by albumin gradient fractionation were transduced by 24 hr co-cultivation in 8 |xg/ml Polybrene with N2/PA317 (Miller and Buttimore, 1986; Armentano et al., 1987) producer cells gamma-irradiated with 2,000 cGy in a Gammacell 40 ,37Cs animal irradiator, as previously described (Eglitis era/., 1985;Kantoffe/a/., 1987).
DISCUSSION This study demonstrates that primate bone marrow cells that had been cryopreserved were capable of being transduced by a murine retroviral vector at the time of thawing. This information is important in considering human gene therapy clinical protocols using bone marrow progenitor cells as the target of gene transfer. Albumin density gradient centrifugation was used to enrich bone marrow progenitor cell populations preferentially trans-
RETROVIRAL GENE TRANSFER INTO PRIMATE BONE M A R R O W
325
result of freeze/thawing. The effect is likely one secondary to an increase in the water-to-lipid ratio after thawing and not one due to preferential loss of light-density precursors. O n the contrary, % of total the C F U - G M frequency in cryopreserved cells is slightly increased due to preferential survival of earlier progenitors over more differentiated myeloid cells. The recovery of C F U - G M was 9 1 % (range 5 4 - 1 2 8 % ) with identical patterns of density distribution regardless, of the percentage of recovery in any one particular experiment. C F U - G M transduced by N 2 belonged to a less dense population of C F U - G M . This population m a y be a slightly less differentiated subset in which a larger fraction of cells are undergoing cell cycling than in the whole C F U - G M population. The C F U - G M and B F U - E plating efficiencies of the lower layer 1 layer 2 layer 3 layer 5 layer 4 density populations in cocultivated bone marrow cells appear to be less than those in uninfected cells. This phenomenon m a y be I Cells S CFU-GM C D G418 CFU-GM due to preferential adherence of early progenitors to the supF I G . 1. Distribution of cryopreserved, transduced bone mar- porting monolayer in these low-density, progenitor-rich popurow mononuclear cells, C F U - G M , and G 4 1 8 R C F U - G M in an lations (Dumenil et aL, 1989), and m a y have accounted for albumin density gradient. Values represent the arithmetic prod- hematopoietic nonengraftment in primates transplanted with uct (normalized over the five layers) of the percentage of mono- autologous bone marrow after retroviral gene transfer by coculnuclear cells in a particular layer and the C F U - G M / 1 0 5 cells and tivation(Kantoff
