- Email: [email protected]
Reconstitution of B cell subsets in Rag deficient mice by transplantation of in vitro differentiated embryonic stem cells
Immunology Letters 57 (1997) 131 – 137
Reconstitution of B cell subsets in Rag deficient mice by transplantation of in vitro differentiated embryonic stem cells Alexandre J. Potocnik *, Gabi Nerz, Hubertus Kohler, Klaus Eichmann Max-Planck-Institute for Immunobiology, Stu¨beweg 51, 79108 Freiburg, Germany
Abstract In vitro differentiated embryonic stem (ES) cells contain a population which is similar to fetal liver pro/pre-B cells on the basis of cell surface antigens and cytoplasmic expression of immunoglobin heavy chain. This population was purified and transplanted into Rag-1 deficient recipients to characterize its developmental potential in vivo. Following intravenous transfer, these cells rapidly reconstituted the splenic B but not the T cell compartment. Reconstitution was transient, indicating the lack of long-term reconstituting capacity. Similar to fetal liver, B-1 type as well as conventional B cells were generated, accompanied by high serum IgM levels. Intraperitoneal injection generated high numbers of peritoneal B cells, predominately of the B-1a phenotype, with poor splenic repopulation and low serum IgM levels. These observations suggest the emergence of two different B lineage precursor populations during in vitro ES cell differentiation and define a possible role of the microenvironment in directing lymphoid development. © 1997 Elsevier Science B.V. Keywords: Embryonic stem cell; Lymphopoiesis; B cell development; Peritoneal B cells; CD5 (Ly-1) B cells
1. Introduction The hematopoietic system is generated and maintained throughout life by a small number of pluripotent hematopoietic stem cells (HSC). A number of differences have been observed between HSC during fetal and adult development, leading to the concept of a transition from a ‘fetal’ to an ‘adult’ stem cell type [1]. Most recently it has been reported that, in analogy to the avian system, HSCs during mouse development are generated independently at intra-embryonic and extraembryonic sites [2]. In addition, it has been proposed that lymphoid potential is restricted to an intra-embryonic site, the para-aortic splanchnopleura [3]. Embryonic stem (ES) cells represent an excellent model to study these developmental switches, since they recapitulate during in vitro differentiation, the formation of mesoderm, and subsequently develop into virtually all hemato-lymphoid lineages [4]. The potential of ES cells
* Corresponding author. Fax: + 49 761 5108534; e-mail: [email protected] 0165-2478/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 5 - 2 4 7 8 ( 9 7 ) 0 0 0 8 9 - 8
to generate B-lineage precursors in vitro was shown by the demonstration of B cell restricted genetic events, i.e. rearrangement of immunoglobin (Ig) loci [5], or by the emergence of mature B cells after co-culture on a stromal cell line [6]. So far, it has not been studied whether ES cell drived B lineage progenitors develop into B-1a (CD5 + IgM + IgDlow) and/or into conventional (CD5 − IgM + IgDhigh) B cell populations. During embryogenesis, the fetal omentum majus has been suggested as an early site of lymphoid differentiation [7]. Elegant cell transfer experiments demonstrated the presence of B cell progenitors, developing exclusively into B-1 type B cells [8]. Fetal liver contains precursor for B-1 B cells as well, but unlike fetal omentum also reconstitutes conventional B cells [9]. The restriction of pre-fetal liver progenitors to the B-1 lineage has been demonstrated also for the para-aortic splanchnopleura [10]. The developmental changes in the generation of these B cell lineages has been explained by different ‘layers’ of lymphoid development, reflecting the transition from the ‘fetal’ to the ‘adult’ HSC [11]. We thus decided to probe the potential of in vitro
132
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
differentiated ES cells to generate B-1 and/or conventional B cells. A pro/pre-B cell population was isolated from in vitro differentiated ES cells, transferred into Rag-1 deficient mice using two different routes of transplantation, and the generation and subtype of mature B cells were analyzed.
2. Materials and methods
2.1. Animals and cell lines Mice deficient for Rag-1 were backcrossed to C57/ Bl6 (Ly-9.1 − ) for at least four generations and maintained under specific pathogen-free conditions in the animal facility of our institute. ES cell lines, derived from C57/Bl6 (ES cell line Bl6-III), Balb/c (ES cell line Balb/c-I) or 129/Sv (D3/M and CCE/R), were routinely passaged and maintained in an undifferentiated state by culture on a monolayer of mitomycin C-inactivated embryonic fibroblasts except of the feeder-independent ES cell line CCE/R, which was propagated in leukemiainhibitory-factor (LIF) containing media.
sis by propidium iodide (1 mg/ml) counterstaining. To enrich the CD45 fraction, cells were incubated with mab M1/9.3.4.HL.2 followed by goat anti-rat Ig conjugated to paramagnetic microbeads (Miltenyi, Bergisch Gladbach, Germany) and passaged once over a MACS (Miltenyi, Bergisch Gladbach, Germany) system. The retained fraction was blocked using normal rat Ig; stained with FITC-labeled anti-B220 mab (RA3-6B2, Pharmingen, Hamburg, Germany) and biotinylated AA4.1 mab followed by streptavidin-PE (Southern Biotechnology, Heidelberg, Germany) in PBS/5% FCS on ice. Cells were separated in the indicated fractions using a FACStar Plus cell sorter (Becton—Dickinson, Heidelberg, Germany). Sorted cells were found to be 98% pure upon re-analysis. For phenotypic analysis, single cell suspensions were stained with mabs as indicated in the figure legends. Intracellular FACS staining was performed by staining 5× 105 cells after fixation in 4% paraformaldehyde and permeabilization in 0.2% Tween-20 with directly conjugated antibodies. For each diagram, at least 20 000 cells were analyzed on a logarithmic scale.
2.4. Reconstitution of Rag-1 deficient animals 2.2. Differentiation of ES cells To induce hematopoietic differentiation, ES cells were dissociated by gentle trypsinization and cultured in IMDM (Biochrom, Berlin, Germany) supplemented with 15% fetal calf serum (FCS) and 4.5 × 10 − 4 M monothioglycerol in gas-permeable 60 mm hydrophobic culture dishes (Heraeus, Hanau, Germany) at a concentration of 0.8–1.4 ×104 cells/ml in a total volume of 5 ml. For all ES cell lines, the initial concentration was adjusted to generate 60 – 100 embryoid bodies (EB) per dish. EB were passaged after 2 – 3 days on new gas-permeable culture dishes and refed with freshly prepared medium every second day. During the whole differentiation period, cultures were maintained at 37°C in an atmosphere of 7.5% CO2 and 5% O2 using an incubator with adjustable oxygen content (Heraeus, Hanau, Germany). After the indicated period of in vitro differentiation, EBs were harvested and dissociated by digestion with 0.1 U/mg collagenase and 0.8 U/mg dispase (Boehringer Mannheim, Mannheim, Germany) in PBS for 2 h at room temperature. Dead cells were removed by centrifugation through FCS.
2.3. Staining and cell separation Cells were labeled with purified CD45 mab (M1/ 9.3.4.HL.2), detected by goat anti-rat Ig-FITC (Jackson-Dianova, Hamburg, Germany) and mab TER-119 coupled to PE (Pharmingen, Hamburg, Germany) and analyzed on a FACScan (Becton-Dickinson, Heidelberg, Germany). Dead cells were excluded from analy-
Cell populations were injected either intravenously in the lateral tail vein or directly in the peritoneal cavity of 4–6 week old Rag-1 − / − mice in a final volume of 150 ml PBS. To allow lympho-hematopoietic development in vivo, animals were sublethally irradiated with 400 rads of a X-ray source prior to transfer. Throughout the whole period mice were housed in a positive pressure cabinet and received neomycin (0.1 g/100 ml) and Borgal® (10% final concentration, Hoechst, Frankfurt, Germany) for the first 2 weeks following irradiation. After the transfer, serum as well as cells were collected from the retro-orbital plexus of recipients by venipuncture. Reconstituted mice were sacrificed at the indicated timepoints and lymphoid populations prepared using standard protocol.
2.5. ELISAs Secreted IgM in the sera of transplanted mice was tested by standard ELISA. Serum samples were pre-diluted 1:10 or 1:20 and added in a serial dilutions on Nunc-Immuno plates (Nunc, Wiesbaden, Germany) coated with goat anti-mouse Ig F(ab)2 (Jackson-Dianova, Hamburg, Germany). Bound Ig was detected with alkaline-phosphatase conjugated goat anti-mouse IgM (Southern Biotechnology, Heidelberg, Germany). Absorption was read at 405/490 nm using p-nitrophenyl phosphate (Boehringer Mannheim, Germany) as a substrate. All results were quantified using external standards (all Southern Biotechnology, Heidelberg, Germany).
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
3. Results We developed a protocol for the enrichment and subsequent isolation of lymphoid progenitors from ES cells differentiated in vitro. As we reported previously [5], differentiating ES cells express genes restricted to the lymphoid lineage (Rag-1, Rag-2) and surface antigens either present exclusively on lymphoid cells, such as B220 (CD45R), or found on lymphoid progenitors as well as on various other cell types, such as CD44. Despite the presence of productively rearranged Ig heavy chains, no surface Ig positive cells were found following diffferentiation of ES cells. To purify populations containing lymphoid progenitors, we analyzed the expression of surface antigens on differentiating ES cells by flow cytometry and observed two non-overlapping populations: A compartment detected by mab TER-119 comprising various stages of erythroid differentiation and a subpopulation characterized by the expression of CD45. After 10 days of in vitro culture, nearly one-third (30.1%, range 20.2 – 48.6%) of all cells belonged to the erythroid lineage, whereas only 10.2% were positive for CD45 (Fig. 1a). These proportions were inverted on day 15, resulting in 11.8% (6.2– 25.3%) erythroid, but 20.1% (15.7 – 32.5%) CD45 + cells (Fig. 1a). To facilitate the detection of lymphoid progenitors, we enriched the CD45 + population by MACS, resulting in 18.7% (day 10) to 43.2% (day 15) CD45 + cells. These enriched preparations were next stained for B220 as a marker of B cell commitment and AA4.1, an antigen present on fetal hematopoietic stem cells [12], but also on early stages of B cell development [13]. As shown in Fig. 1a, B220 is absent on day 10 ES cells, but could be detected from day 15 onwards. The B220 + cells were subdivided by the expression of AA4.1 (Fig. 1a). The AA4.1 − B220 + cells (6.2%, range 4.1–9.7%) apparently represent an aberrant pathway of development, since they showed no developmental capacity in various in vitro and in vivo assays (data not shown). We therefore concentrated on the AA4.1 + B220 + fraction, a population of 1.8 –2.9% of all cells prior to enrichment. After purification by FACS, this population was negative for surface IgM or IgD (Fig. 1b), but contained a subset (18.9%, range 10.2 –31.8%) of cells which expressed cytoplasmic m chain (Fig. 1b). On the basis of these properties, it was likely that the AA4.1 + B220 + population contained a mixture of B cell differentiation stages extending into the pre-B cell stage. This pro/pre-B cell-like population was initially identified and purified from the 129/Sv derived ES cell line D3/M. Similar results were subsequently obtained with other 129/Sv, C57/Bl6 or Balb/c derived ES cell lines, with slightly different proportions for each of the subsets. To test the functional capacity of in vitro differentiated AA4.1 + B220 + ES cells, they were transferred
133
into sublethally irradiated Rag-1 deficient recipients. 50 000 cells isolated at day 15 of culture were either injected intravenously (i.v.) or intraperitoneally (i.p.), to compare the influence of the route of injection on the further development of the transplanted population. The amount of lymphoid reconstitution was monitored by the emergence of IgM in the serum of the hosts. As shown in Fig. 2a, IgM could be detected 7–10 days post i.v. injection and reached a peak 2–3 weeks after cell transfer. The level of IgM decreased rapidly thereafter (Fig. 2a), and was undetectable after 8 weeks. When the cells were injected i.p., serum IgM appeared with delayed kinetics (Fig. 2a). In some animals injected i.p., serum Ig could be demonstrated at low levels up to 12 weeks post reconstitution (two out of six mice). In all mice injected i.p., the levels of serum IgM were significantly lower than in the group which received the cells i.v. Analyses of the IgM + cells in the peripheral blood showed an even more pronounced difference between i.v. and i.p. injection than that between secreted IgM (Fig. 2a and b). In no case did we observe T cells in primary or secondary lymphoid organs after transfer of AA4.1 + B220 + ES derived lymphoid progenitors. Next we investigated the B cell populations generated in more detail. Reconstituted Rag-1 − / − mice were sacrificed 6 weeks post transfer and peritoneal cells were analyzed for the presence of IgM, IgD and CD5 (Fig. 3a and b). Only 17.89 7.4% (n= 7) of the peritoneal cells were IgM + in case of i.v. injection, compared to 65.99 22.1% (n= 8) surface IgM + cells after i.p. transfer. In addition, the intraperitoneal population of i.p. injected animals showed a strong bias towards a IgM + IgDlow B-1 phenotype (Fig. 3a). This is further corroborated by the strong predominance of CD5 + IgM + B-1a type B cells (59.0 9 12.1%) in the peritoneal cavity as well as in spleen (77.1 910.2%) following i.p. transfer (Fig. 3b and c). This high frequency of B-1a cells derived from ES cells is not only a direct consequence of local factors at the site of transplantation, since also after intravenous transfer approximately 40% (39.8910.2%, n= 7) of the splenic B cells were CD5 + . We never found evidence for donorderived (Ly-9.1 + ) pro-B cells (B220 + c-kit + ) in bone marrow or spleen. Another striking difference between i.v. and i.p. transfer of lymphoid progenitors is revealed by the absolute numbers of lymphocytes present at the target sites. Following intravenous transfer, the number of B cells found in the spleen was 15 times greater than that after intraperitoneal injection of the same number of progenitors (Fig. 4). Vice versa, i.p. injection generated a peritoneal B cell compartment comparable in size to the spleen B cell population i.v. transfer (Fig. 4). Interestingly, the presence of lymphoid cells increased the number of non-lymphoid endogenous peritoneal
134
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
Fig. 1. Generation of hemato-lymphoid cells by in vitro differentiated ES cells. Appearance of hematopoietic cells was assessed by two-color flow cytometry for the expression of CD45 (T200) versus the erythroid marker TER-119 (a, upper part). Numbers in the individual quadrants or gates indicate percentages. Lymphoid progenitors were further enriched for CD45 + by MACS and restained for CD45R (B220) and AA4.1 (a, lower part). The AA4.1 + B220 + population was isolated by FACS and analyzed for the presence of IgM or IgD on the surface or in the cytoplasm (ic) respectively.
cells. When ES derived progenitors were injected i.v., a total of 1.25× 106 mononuclear cells were found in the peritoneal cavity, compared to 0.5 ×106 cells in Rag-1 − / − mice injected with PBS. The majority (68.2910.2%) was found to be host-derived myeloid cells. A similar accumulation was observed in recipients injected i.p., suggesting in both cases the recruitment of a non-lymphoid population into the peritoneal cavity.
4. Discussion In the present study, we report a novel approach to isolate an irreversibly lymphoid-committed cell type derived from in vitro differentiated ES cells. These lymphoid precursors emerged between days 10 and of in vitro differentiation and are characterized by the expression of B220 and AA4.1. This population, repre-
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
senting initially about 2% of all cells, contained a significant proportion of cells expressing immunoglobin heavy chain proteins in their cytoplasm. The presence of cytoplasmic m indicates a developmental stage similar to the pre-B cell stage in bone marrow. The AA4.1 + B220 + population could therefore be regarded as an inhomogenous pool of cells at various levels of pro/preB cell development. It is noteworthy to mention that, in parallel to the first appearance of these cells around day 15, transcripts for Rag-1 and Rag-2 and first complete heavy chain VDJ rearrangements were also observed at this time [5]. The transfer of these lymphoid precursors into Rag-1 deficient mice revealed three major features of this population: (i) The B cell compartment was rapidly restored, suggesting the presence of relatively advanced lymphoid progenitors within the transferred population; (ii) this lymphoid reconstitution was only transient, indicating the absence of a long-term repopulating activity. (iii) in addition to the lack of long-term reconstituting activity, the complete absence of T cell
Fig. 2. Kinetics of peripheral lymphoid reconstitution in Rag-1 deficient mice using ES-derived lymphoid progenitors. At indicated timepoints post transfer, blood was analyzed for serum IgM (a), as well as for peripheral B cells (b). Results are given either as total amount of IgM (a) or as numbers in percent relative to total (CD45 + ) leukocytes in blood (b), comparing the intravenous transfer ( ) to the intraperitoneal injection () of 5× 105 ES derived progenitors. Dashed lines indicate wild-type levels of IgM and B cells in C57/Bl6 for comparison. Both figures represent results (mean 9 S.D.) from three independent experiments, each with 4–8 animals per group
135
development could be regarded as an irreversible restriction of the ES cell derived population to the B cell lineage. We cannot rule out the possibility that the absence of T cells could be also explained by a low frequency of pro-T cells in the initial preparations or the inability of T cell precursors to reach an environment which supports their differentiation. Future experiments will directly address the T cell potential of the AA4.1 + B220 + population. Taken together, these points argue for the irreversible commitment of this ES-derived population to the B cell lineage. The exclusive presence of a B cell committed precursor at the time of transplantation clearly distinguishes these experiments from previous transfer experiments of differentiated ES cells into immunodeficient recipients, reported by several groups [14–16]. In all experimental setups, unseparated ES populations were transferred into hosts and T and B cells were generated. In the case of Nisitani and colleagues [16], B cell development was restored even in the bone marrow. Thus, these results do not prove the successful lymphoid reconstitution by an advanced lymphoid precursor, since in all experiments, the transfer of a primitive multilineage progenitor was not excluded. Our study also addressed the question which particular B cell lineage is generated by ES derived lymphoid progenitors. Interestingly, the generation of B-1 versus conventional B cells seems to be strongly connected to the route of transplantation. When injected intraperitoneally, the reconstituted B cell compartment is predominantly of the B-1a phenotype and is present in high numbers mainly at the initial site of transplantation. These animals had surprisingly low titers of serum IgM, which is unexpected since the B-1a subset is thought to strongly contribute to serum IgM levels [8]. The predominance of the B-1a lineage was less pronounced following intravenous transfer of lymphoid progenitors where they contribute to the B-1 as well as to the conventional B cell lineage, quite similar to fetal liver [9]. We also observed a disconcordance between the serum IgM titer and the number of circulating and splenic B cells 6–8 weeks after intravenous injection. The serum of reconstituted animals contained only low to undetectable levels of IgM, whereas IgM + cells were still present. Whether this might reflect an aberrant functional state of the transferred population is currently under investigation. Where in the ontogeny of the lymphoid system could the ES derived lymphoid precursor be placed? During mouse development, the omentum majus harbors at day 13 progenitors, which exclusively develop into B-1 type B cells [8]. Fetal liver already contains progenitors both for B-1 and conventional B cells [9]. In addition, transplantation of the para-aortic splanchnopleura under the kidney capsule restored exclusively the B-1a
136
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
Fig. 3. Analysis of ES derived B cell populations in reconstituted Rag-1 − / − mice. Six weeks post transfer, recipients were sacrificed and peritoneal cells as well as splenocytes analyzed by flow-cytometry for the expression of IgM and IgD (a) or IgM and CD5 (b). For the analysis of splenic B cells (c), preparations were stained additionally with the donor-specific marker Ly-9.1. Note the dominance of the B-1 (IgM + IgDlowCD5 + ) B cell lineage in case of intraperitoneal transfer.
compartment [10]. Our results might therefore indicate a fetal liver-like stage of hemato-lymphoid development, as defined by the appearance of both B cell lineages after transfer. The intriguing finding that the microenvironment at the initial site of transplantation contributes to the resulting proportion of cells from the two B cell lineages might suggest that the microenvironment directs the developmental outcome, possibly via
cell-to-cell contacts and/or soluble factors. The different proportions of B-1 versus conventional B cells after i.v. and i.p. transfer could also be explained by selective outgrowth of two different types of lymphoid progenitors or differences in homing potential. Following i.v. injection, a more ‘adult’ progenitor be selected, whereas the peritoneal cavity would strongly favor a ‘fetal’ precursor. The hypothetical presence of two
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
137
types of lymphoid progenitors—both induced during in vitro differentiation of ES cells—might offer an attractive system to study developmental switches in the lymphoid lineage.
References
Fig. 4. Total number of B cells generated by ES derived progenitors in spleen and peritoneal cavity. Purified AA4.1 + B220 + ES cells were injected intravenously or intraperitoneally into Rag-1 − / − mice. After 6 weeks, cells were isolated from spleen and peritoneal cavity, counted and analyzed for IgM, IgD or IgM, CD5 by flow-cytometry (Fig. 3). Stacked histograms represent the total number of B cells as well as the respective subsets as indicated in the legend.
[1] K. Ikuta, N. Uchida, J. Friedman, I.L. Weissman, Ann. Rev. Immunol. 10 (1992) 759 – 783. [2] A. Medvinsky, E. Dzierzak, Cell 86 (1996) 897 – 906. [3] A. Cumano, F. Dieterlen-Lie`vre, I. Godin, Cell 86 (1996) 907–916. [4] G.M. Keller, Curr. Op. Cell Biol. 7 (1995) 862 – 869. [5] A.J. Potocnik, P.J. Nielsen, K. Eichmann, EMBO J. 13 (1994) 5274 – 5283. [6] T. Nakano, H. Kodama, T. Honjo, Science 265 (1994) 1098–1101. [7] L. Kubai, R. Auerbach, Nature 301 (1983) 154 – 156. [8] N. Solvason, A. Lehuen, J.F. Kearney, Int. Immunol. 3 (1991) 543 – 550. [9] R.R. Hardy, K. Hayakawa, PNAS 88 (1991) 11550 – 11554. [10] I.E. Godin, J.A. Garcia-Porrero, A. Coutinho, F. DieterlenLie`vre, M.A. Marcos, Nature 364 (1993) 67 – 70. [11] L.A. Herzenberg, A.B. Kantor, L.A. Herzenberg, Ann. New York Acad. Sci. 651 (1992) 1 – 9. [12] C.T. Jordan, J.P. McKearn, I.R. Lemischka, Cell 61 (1990) 953 – 963. [13] J.P. McKearn, J. McCubrey, B. Fagg, PNAS 82 (1985) 7414– 7418. [14] A.M. Mu9 ller, E.A. Dzierzak, Development 118 (1993) 1343–1351. [15] J.C. Gutierrez-Ramos, R. Palacios, PNAS 89 (1992) 9171– [16] S. Nisitani, T. Tsubata, T. Honjo, Int. Immunol. 6 (1994) 909–916.
.
Reconstitution of B cell subsets in Rag deficient mice by transplantation of in vitro differentiated embryonic stem cells Alexandre J. Potocnik *, Gabi Nerz, Hubertus Kohler, Klaus Eichmann Max-Planck-Institute for Immunobiology, Stu¨beweg 51, 79108 Freiburg, Germany
Abstract In vitro differentiated embryonic stem (ES) cells contain a population which is similar to fetal liver pro/pre-B cells on the basis of cell surface antigens and cytoplasmic expression of immunoglobin heavy chain. This population was purified and transplanted into Rag-1 deficient recipients to characterize its developmental potential in vivo. Following intravenous transfer, these cells rapidly reconstituted the splenic B but not the T cell compartment. Reconstitution was transient, indicating the lack of long-term reconstituting capacity. Similar to fetal liver, B-1 type as well as conventional B cells were generated, accompanied by high serum IgM levels. Intraperitoneal injection generated high numbers of peritoneal B cells, predominately of the B-1a phenotype, with poor splenic repopulation and low serum IgM levels. These observations suggest the emergence of two different B lineage precursor populations during in vitro ES cell differentiation and define a possible role of the microenvironment in directing lymphoid development. © 1997 Elsevier Science B.V. Keywords: Embryonic stem cell; Lymphopoiesis; B cell development; Peritoneal B cells; CD5 (Ly-1) B cells
1. Introduction The hematopoietic system is generated and maintained throughout life by a small number of pluripotent hematopoietic stem cells (HSC). A number of differences have been observed between HSC during fetal and adult development, leading to the concept of a transition from a ‘fetal’ to an ‘adult’ stem cell type [1]. Most recently it has been reported that, in analogy to the avian system, HSCs during mouse development are generated independently at intra-embryonic and extraembryonic sites [2]. In addition, it has been proposed that lymphoid potential is restricted to an intra-embryonic site, the para-aortic splanchnopleura [3]. Embryonic stem (ES) cells represent an excellent model to study these developmental switches, since they recapitulate during in vitro differentiation, the formation of mesoderm, and subsequently develop into virtually all hemato-lymphoid lineages [4]. The potential of ES cells
* Corresponding author. Fax: + 49 761 5108534; e-mail: [email protected] 0165-2478/97/$17.00 © 1997 Elsevier Science B.V. All rights reserved. PII S 0 1 5 - 2 4 7 8 ( 9 7 ) 0 0 0 8 9 - 8
to generate B-lineage precursors in vitro was shown by the demonstration of B cell restricted genetic events, i.e. rearrangement of immunoglobin (Ig) loci [5], or by the emergence of mature B cells after co-culture on a stromal cell line [6]. So far, it has not been studied whether ES cell drived B lineage progenitors develop into B-1a (CD5 + IgM + IgDlow) and/or into conventional (CD5 − IgM + IgDhigh) B cell populations. During embryogenesis, the fetal omentum majus has been suggested as an early site of lymphoid differentiation [7]. Elegant cell transfer experiments demonstrated the presence of B cell progenitors, developing exclusively into B-1 type B cells [8]. Fetal liver contains precursor for B-1 B cells as well, but unlike fetal omentum also reconstitutes conventional B cells [9]. The restriction of pre-fetal liver progenitors to the B-1 lineage has been demonstrated also for the para-aortic splanchnopleura [10]. The developmental changes in the generation of these B cell lineages has been explained by different ‘layers’ of lymphoid development, reflecting the transition from the ‘fetal’ to the ‘adult’ HSC [11]. We thus decided to probe the potential of in vitro
132
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
differentiated ES cells to generate B-1 and/or conventional B cells. A pro/pre-B cell population was isolated from in vitro differentiated ES cells, transferred into Rag-1 deficient mice using two different routes of transplantation, and the generation and subtype of mature B cells were analyzed.
2. Materials and methods
2.1. Animals and cell lines Mice deficient for Rag-1 were backcrossed to C57/ Bl6 (Ly-9.1 − ) for at least four generations and maintained under specific pathogen-free conditions in the animal facility of our institute. ES cell lines, derived from C57/Bl6 (ES cell line Bl6-III), Balb/c (ES cell line Balb/c-I) or 129/Sv (D3/M and CCE/R), were routinely passaged and maintained in an undifferentiated state by culture on a monolayer of mitomycin C-inactivated embryonic fibroblasts except of the feeder-independent ES cell line CCE/R, which was propagated in leukemiainhibitory-factor (LIF) containing media.
sis by propidium iodide (1 mg/ml) counterstaining. To enrich the CD45 fraction, cells were incubated with mab M1/9.3.4.HL.2 followed by goat anti-rat Ig conjugated to paramagnetic microbeads (Miltenyi, Bergisch Gladbach, Germany) and passaged once over a MACS (Miltenyi, Bergisch Gladbach, Germany) system. The retained fraction was blocked using normal rat Ig; stained with FITC-labeled anti-B220 mab (RA3-6B2, Pharmingen, Hamburg, Germany) and biotinylated AA4.1 mab followed by streptavidin-PE (Southern Biotechnology, Heidelberg, Germany) in PBS/5% FCS on ice. Cells were separated in the indicated fractions using a FACStar Plus cell sorter (Becton—Dickinson, Heidelberg, Germany). Sorted cells were found to be 98% pure upon re-analysis. For phenotypic analysis, single cell suspensions were stained with mabs as indicated in the figure legends. Intracellular FACS staining was performed by staining 5× 105 cells after fixation in 4% paraformaldehyde and permeabilization in 0.2% Tween-20 with directly conjugated antibodies. For each diagram, at least 20 000 cells were analyzed on a logarithmic scale.
2.4. Reconstitution of Rag-1 deficient animals 2.2. Differentiation of ES cells To induce hematopoietic differentiation, ES cells were dissociated by gentle trypsinization and cultured in IMDM (Biochrom, Berlin, Germany) supplemented with 15% fetal calf serum (FCS) and 4.5 × 10 − 4 M monothioglycerol in gas-permeable 60 mm hydrophobic culture dishes (Heraeus, Hanau, Germany) at a concentration of 0.8–1.4 ×104 cells/ml in a total volume of 5 ml. For all ES cell lines, the initial concentration was adjusted to generate 60 – 100 embryoid bodies (EB) per dish. EB were passaged after 2 – 3 days on new gas-permeable culture dishes and refed with freshly prepared medium every second day. During the whole differentiation period, cultures were maintained at 37°C in an atmosphere of 7.5% CO2 and 5% O2 using an incubator with adjustable oxygen content (Heraeus, Hanau, Germany). After the indicated period of in vitro differentiation, EBs were harvested and dissociated by digestion with 0.1 U/mg collagenase and 0.8 U/mg dispase (Boehringer Mannheim, Mannheim, Germany) in PBS for 2 h at room temperature. Dead cells were removed by centrifugation through FCS.
2.3. Staining and cell separation Cells were labeled with purified CD45 mab (M1/ 9.3.4.HL.2), detected by goat anti-rat Ig-FITC (Jackson-Dianova, Hamburg, Germany) and mab TER-119 coupled to PE (Pharmingen, Hamburg, Germany) and analyzed on a FACScan (Becton-Dickinson, Heidelberg, Germany). Dead cells were excluded from analy-
Cell populations were injected either intravenously in the lateral tail vein or directly in the peritoneal cavity of 4–6 week old Rag-1 − / − mice in a final volume of 150 ml PBS. To allow lympho-hematopoietic development in vivo, animals were sublethally irradiated with 400 rads of a X-ray source prior to transfer. Throughout the whole period mice were housed in a positive pressure cabinet and received neomycin (0.1 g/100 ml) and Borgal® (10% final concentration, Hoechst, Frankfurt, Germany) for the first 2 weeks following irradiation. After the transfer, serum as well as cells were collected from the retro-orbital plexus of recipients by venipuncture. Reconstituted mice were sacrificed at the indicated timepoints and lymphoid populations prepared using standard protocol.
2.5. ELISAs Secreted IgM in the sera of transplanted mice was tested by standard ELISA. Serum samples were pre-diluted 1:10 or 1:20 and added in a serial dilutions on Nunc-Immuno plates (Nunc, Wiesbaden, Germany) coated with goat anti-mouse Ig F(ab)2 (Jackson-Dianova, Hamburg, Germany). Bound Ig was detected with alkaline-phosphatase conjugated goat anti-mouse IgM (Southern Biotechnology, Heidelberg, Germany). Absorption was read at 405/490 nm using p-nitrophenyl phosphate (Boehringer Mannheim, Germany) as a substrate. All results were quantified using external standards (all Southern Biotechnology, Heidelberg, Germany).
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
3. Results We developed a protocol for the enrichment and subsequent isolation of lymphoid progenitors from ES cells differentiated in vitro. As we reported previously [5], differentiating ES cells express genes restricted to the lymphoid lineage (Rag-1, Rag-2) and surface antigens either present exclusively on lymphoid cells, such as B220 (CD45R), or found on lymphoid progenitors as well as on various other cell types, such as CD44. Despite the presence of productively rearranged Ig heavy chains, no surface Ig positive cells were found following diffferentiation of ES cells. To purify populations containing lymphoid progenitors, we analyzed the expression of surface antigens on differentiating ES cells by flow cytometry and observed two non-overlapping populations: A compartment detected by mab TER-119 comprising various stages of erythroid differentiation and a subpopulation characterized by the expression of CD45. After 10 days of in vitro culture, nearly one-third (30.1%, range 20.2 – 48.6%) of all cells belonged to the erythroid lineage, whereas only 10.2% were positive for CD45 (Fig. 1a). These proportions were inverted on day 15, resulting in 11.8% (6.2– 25.3%) erythroid, but 20.1% (15.7 – 32.5%) CD45 + cells (Fig. 1a). To facilitate the detection of lymphoid progenitors, we enriched the CD45 + population by MACS, resulting in 18.7% (day 10) to 43.2% (day 15) CD45 + cells. These enriched preparations were next stained for B220 as a marker of B cell commitment and AA4.1, an antigen present on fetal hematopoietic stem cells [12], but also on early stages of B cell development [13]. As shown in Fig. 1a, B220 is absent on day 10 ES cells, but could be detected from day 15 onwards. The B220 + cells were subdivided by the expression of AA4.1 (Fig. 1a). The AA4.1 − B220 + cells (6.2%, range 4.1–9.7%) apparently represent an aberrant pathway of development, since they showed no developmental capacity in various in vitro and in vivo assays (data not shown). We therefore concentrated on the AA4.1 + B220 + fraction, a population of 1.8 –2.9% of all cells prior to enrichment. After purification by FACS, this population was negative for surface IgM or IgD (Fig. 1b), but contained a subset (18.9%, range 10.2 –31.8%) of cells which expressed cytoplasmic m chain (Fig. 1b). On the basis of these properties, it was likely that the AA4.1 + B220 + population contained a mixture of B cell differentiation stages extending into the pre-B cell stage. This pro/pre-B cell-like population was initially identified and purified from the 129/Sv derived ES cell line D3/M. Similar results were subsequently obtained with other 129/Sv, C57/Bl6 or Balb/c derived ES cell lines, with slightly different proportions for each of the subsets. To test the functional capacity of in vitro differentiated AA4.1 + B220 + ES cells, they were transferred
133
into sublethally irradiated Rag-1 deficient recipients. 50 000 cells isolated at day 15 of culture were either injected intravenously (i.v.) or intraperitoneally (i.p.), to compare the influence of the route of injection on the further development of the transplanted population. The amount of lymphoid reconstitution was monitored by the emergence of IgM in the serum of the hosts. As shown in Fig. 2a, IgM could be detected 7–10 days post i.v. injection and reached a peak 2–3 weeks after cell transfer. The level of IgM decreased rapidly thereafter (Fig. 2a), and was undetectable after 8 weeks. When the cells were injected i.p., serum IgM appeared with delayed kinetics (Fig. 2a). In some animals injected i.p., serum Ig could be demonstrated at low levels up to 12 weeks post reconstitution (two out of six mice). In all mice injected i.p., the levels of serum IgM were significantly lower than in the group which received the cells i.v. Analyses of the IgM + cells in the peripheral blood showed an even more pronounced difference between i.v. and i.p. injection than that between secreted IgM (Fig. 2a and b). In no case did we observe T cells in primary or secondary lymphoid organs after transfer of AA4.1 + B220 + ES derived lymphoid progenitors. Next we investigated the B cell populations generated in more detail. Reconstituted Rag-1 − / − mice were sacrificed 6 weeks post transfer and peritoneal cells were analyzed for the presence of IgM, IgD and CD5 (Fig. 3a and b). Only 17.89 7.4% (n= 7) of the peritoneal cells were IgM + in case of i.v. injection, compared to 65.99 22.1% (n= 8) surface IgM + cells after i.p. transfer. In addition, the intraperitoneal population of i.p. injected animals showed a strong bias towards a IgM + IgDlow B-1 phenotype (Fig. 3a). This is further corroborated by the strong predominance of CD5 + IgM + B-1a type B cells (59.0 9 12.1%) in the peritoneal cavity as well as in spleen (77.1 910.2%) following i.p. transfer (Fig. 3b and c). This high frequency of B-1a cells derived from ES cells is not only a direct consequence of local factors at the site of transplantation, since also after intravenous transfer approximately 40% (39.8910.2%, n= 7) of the splenic B cells were CD5 + . We never found evidence for donorderived (Ly-9.1 + ) pro-B cells (B220 + c-kit + ) in bone marrow or spleen. Another striking difference between i.v. and i.p. transfer of lymphoid progenitors is revealed by the absolute numbers of lymphocytes present at the target sites. Following intravenous transfer, the number of B cells found in the spleen was 15 times greater than that after intraperitoneal injection of the same number of progenitors (Fig. 4). Vice versa, i.p. injection generated a peritoneal B cell compartment comparable in size to the spleen B cell population i.v. transfer (Fig. 4). Interestingly, the presence of lymphoid cells increased the number of non-lymphoid endogenous peritoneal
134
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
Fig. 1. Generation of hemato-lymphoid cells by in vitro differentiated ES cells. Appearance of hematopoietic cells was assessed by two-color flow cytometry for the expression of CD45 (T200) versus the erythroid marker TER-119 (a, upper part). Numbers in the individual quadrants or gates indicate percentages. Lymphoid progenitors were further enriched for CD45 + by MACS and restained for CD45R (B220) and AA4.1 (a, lower part). The AA4.1 + B220 + population was isolated by FACS and analyzed for the presence of IgM or IgD on the surface or in the cytoplasm (ic) respectively.
cells. When ES derived progenitors were injected i.v., a total of 1.25× 106 mononuclear cells were found in the peritoneal cavity, compared to 0.5 ×106 cells in Rag-1 − / − mice injected with PBS. The majority (68.2910.2%) was found to be host-derived myeloid cells. A similar accumulation was observed in recipients injected i.p., suggesting in both cases the recruitment of a non-lymphoid population into the peritoneal cavity.
4. Discussion In the present study, we report a novel approach to isolate an irreversibly lymphoid-committed cell type derived from in vitro differentiated ES cells. These lymphoid precursors emerged between days 10 and of in vitro differentiation and are characterized by the expression of B220 and AA4.1. This population, repre-
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
senting initially about 2% of all cells, contained a significant proportion of cells expressing immunoglobin heavy chain proteins in their cytoplasm. The presence of cytoplasmic m indicates a developmental stage similar to the pre-B cell stage in bone marrow. The AA4.1 + B220 + population could therefore be regarded as an inhomogenous pool of cells at various levels of pro/preB cell development. It is noteworthy to mention that, in parallel to the first appearance of these cells around day 15, transcripts for Rag-1 and Rag-2 and first complete heavy chain VDJ rearrangements were also observed at this time [5]. The transfer of these lymphoid precursors into Rag-1 deficient mice revealed three major features of this population: (i) The B cell compartment was rapidly restored, suggesting the presence of relatively advanced lymphoid progenitors within the transferred population; (ii) this lymphoid reconstitution was only transient, indicating the absence of a long-term repopulating activity. (iii) in addition to the lack of long-term reconstituting activity, the complete absence of T cell
Fig. 2. Kinetics of peripheral lymphoid reconstitution in Rag-1 deficient mice using ES-derived lymphoid progenitors. At indicated timepoints post transfer, blood was analyzed for serum IgM (a), as well as for peripheral B cells (b). Results are given either as total amount of IgM (a) or as numbers in percent relative to total (CD45 + ) leukocytes in blood (b), comparing the intravenous transfer ( ) to the intraperitoneal injection () of 5× 105 ES derived progenitors. Dashed lines indicate wild-type levels of IgM and B cells in C57/Bl6 for comparison. Both figures represent results (mean 9 S.D.) from three independent experiments, each with 4–8 animals per group
135
development could be regarded as an irreversible restriction of the ES cell derived population to the B cell lineage. We cannot rule out the possibility that the absence of T cells could be also explained by a low frequency of pro-T cells in the initial preparations or the inability of T cell precursors to reach an environment which supports their differentiation. Future experiments will directly address the T cell potential of the AA4.1 + B220 + population. Taken together, these points argue for the irreversible commitment of this ES-derived population to the B cell lineage. The exclusive presence of a B cell committed precursor at the time of transplantation clearly distinguishes these experiments from previous transfer experiments of differentiated ES cells into immunodeficient recipients, reported by several groups [14–16]. In all experimental setups, unseparated ES populations were transferred into hosts and T and B cells were generated. In the case of Nisitani and colleagues [16], B cell development was restored even in the bone marrow. Thus, these results do not prove the successful lymphoid reconstitution by an advanced lymphoid precursor, since in all experiments, the transfer of a primitive multilineage progenitor was not excluded. Our study also addressed the question which particular B cell lineage is generated by ES derived lymphoid progenitors. Interestingly, the generation of B-1 versus conventional B cells seems to be strongly connected to the route of transplantation. When injected intraperitoneally, the reconstituted B cell compartment is predominantly of the B-1a phenotype and is present in high numbers mainly at the initial site of transplantation. These animals had surprisingly low titers of serum IgM, which is unexpected since the B-1a subset is thought to strongly contribute to serum IgM levels [8]. The predominance of the B-1a lineage was less pronounced following intravenous transfer of lymphoid progenitors where they contribute to the B-1 as well as to the conventional B cell lineage, quite similar to fetal liver [9]. We also observed a disconcordance between the serum IgM titer and the number of circulating and splenic B cells 6–8 weeks after intravenous injection. The serum of reconstituted animals contained only low to undetectable levels of IgM, whereas IgM + cells were still present. Whether this might reflect an aberrant functional state of the transferred population is currently under investigation. Where in the ontogeny of the lymphoid system could the ES derived lymphoid precursor be placed? During mouse development, the omentum majus harbors at day 13 progenitors, which exclusively develop into B-1 type B cells [8]. Fetal liver already contains progenitors both for B-1 and conventional B cells [9]. In addition, transplantation of the para-aortic splanchnopleura under the kidney capsule restored exclusively the B-1a
136
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
Fig. 3. Analysis of ES derived B cell populations in reconstituted Rag-1 − / − mice. Six weeks post transfer, recipients were sacrificed and peritoneal cells as well as splenocytes analyzed by flow-cytometry for the expression of IgM and IgD (a) or IgM and CD5 (b). For the analysis of splenic B cells (c), preparations were stained additionally with the donor-specific marker Ly-9.1. Note the dominance of the B-1 (IgM + IgDlowCD5 + ) B cell lineage in case of intraperitoneal transfer.
compartment [10]. Our results might therefore indicate a fetal liver-like stage of hemato-lymphoid development, as defined by the appearance of both B cell lineages after transfer. The intriguing finding that the microenvironment at the initial site of transplantation contributes to the resulting proportion of cells from the two B cell lineages might suggest that the microenvironment directs the developmental outcome, possibly via
cell-to-cell contacts and/or soluble factors. The different proportions of B-1 versus conventional B cells after i.v. and i.p. transfer could also be explained by selective outgrowth of two different types of lymphoid progenitors or differences in homing potential. Following i.v. injection, a more ‘adult’ progenitor be selected, whereas the peritoneal cavity would strongly favor a ‘fetal’ precursor. The hypothetical presence of two
A.J. Potocnik et al. / Immunology Letters 57 (1997) 131–137
137
types of lymphoid progenitors—both induced during in vitro differentiation of ES cells—might offer an attractive system to study developmental switches in the lymphoid lineage.
References
Fig. 4. Total number of B cells generated by ES derived progenitors in spleen and peritoneal cavity. Purified AA4.1 + B220 + ES cells were injected intravenously or intraperitoneally into Rag-1 − / − mice. After 6 weeks, cells were isolated from spleen and peritoneal cavity, counted and analyzed for IgM, IgD or IgM, CD5 by flow-cytometry (Fig. 3). Stacked histograms represent the total number of B cells as well as the respective subsets as indicated in the legend.
[1] K. Ikuta, N. Uchida, J. Friedman, I.L. Weissman, Ann. Rev. Immunol. 10 (1992) 759 – 783. [2] A. Medvinsky, E. Dzierzak, Cell 86 (1996) 897 – 906. [3] A. Cumano, F. Dieterlen-Lie`vre, I. Godin, Cell 86 (1996) 907–916. [4] G.M. Keller, Curr. Op. Cell Biol. 7 (1995) 862 – 869. [5] A.J. Potocnik, P.J. Nielsen, K. Eichmann, EMBO J. 13 (1994) 5274 – 5283. [6] T. Nakano, H. Kodama, T. Honjo, Science 265 (1994) 1098–1101. [7] L. Kubai, R. Auerbach, Nature 301 (1983) 154 – 156. [8] N. Solvason, A. Lehuen, J.F. Kearney, Int. Immunol. 3 (1991) 543 – 550. [9] R.R. Hardy, K. Hayakawa, PNAS 88 (1991) 11550 – 11554. [10] I.E. Godin, J.A. Garcia-Porrero, A. Coutinho, F. DieterlenLie`vre, M.A. Marcos, Nature 364 (1993) 67 – 70. [11] L.A. Herzenberg, A.B. Kantor, L.A. Herzenberg, Ann. New York Acad. Sci. 651 (1992) 1 – 9. [12] C.T. Jordan, J.P. McKearn, I.R. Lemischka, Cell 61 (1990) 953 – 963. [13] J.P. McKearn, J. McCubrey, B. Fagg, PNAS 82 (1985) 7414– 7418. [14] A.M. Mu9 ller, E.A. Dzierzak, Development 118 (1993) 1343–1351. [15] J.C. Gutierrez-Ramos, R. Palacios, PNAS 89 (1992) 9171– [16] S. Nisitani, T. Tsubata, T. Honjo, Int. Immunol. 6 (1994) 909–916.
.