T lineage differentiation from human embryonic stem cells

Abstract

Harnessing the ability of genetically manipulated human embryonic stem cells (hESC) to differentiate into appropriate lineages could revolutionize medical practice. These cells have the theoretical potential to develop into all mature cell types; however, the actual ability to develop into all hematopoietic lineages has not been demonstrated. Using sequential in vitro coculture on murine bone marrow stromal cells, and engraftment into human thymic tissues in immunodeficient mice, we demonstrate that hESC can differentiate through the T lymphoid lineage. Stable transgene expression was maintained at high levels throughout differentiation, suggesting that genetically manipulated hESC hold potential to treat several T cell disorders.

Human embryonic stem cells (hESC) show promise to revolutionize treatment strategies for many diseases because of their potential to differentiate into all tissues and cell types in the body. However, signals required for proper directed differentiation of these cells are not well defined. Regarding hematopoietic differentiation, these cells have been directed toward myeloid and erythroid lineages by exposure to murine bone marrow stromal cells (1) or by induced formation of embryoid bodies in the presence of various cytokines (2), both conditions resulting in differentiation into CD34⁺ hematopoietic progenitor cells capable of forming hematopoietic colonies in vitro. Similarly, B lineage differentiation has been achieved by sequential exposure of hESC to two different murine bone marrow stromal cell lines, OP9 and MS-5 (3). Recently, functional dendritic cells (4), natural killer cells (5), and macrophages (6) were derived from hESC. Although T cells were successfully derived from murine embryonic stem (ES) cells (7), the differentiation toward the T lymphoid lineage from hESC has not been reported. Herein we use a combination of in vitro exposure to the OP9 stromal cell line, followed by implantation of the newly differentiated precursors into human thymic tissues growing in immunodeficient mice, to induce T lymphoid differentiation of these cells in vivo. Costimulation of the resulting T lineage cells by CD3 and CD28 resulted in expression of activation markers, indicating that these cells are functional. Furthermore, a lentiviral vector expressing EGFP under the control of the elongation factor 1α (EF1α) promoter, introduced at the hESC stage, continued to express the reporter gene at high frequency throughout thymopoiesis. Our results suggest that genetically manipulated hESC may hold promise for treatment of disorders of the T cell lineage.

Results

To obtain genetically marked ES cells, the hESC line H1 (8, 9) was transduced with the pSIN18.cPPT.hEF1α.EGFP.WPRE lentiviral vector, which allows constitutive expression of the EGFP reporter gene under the control of the EF1α promoter in hESC (9). To obtain colonies expressing EGFP at high frequency, ≈20 passages of manually isolating GFP⁺ cells were performed (Fig. 1b). However, these cells continued to express markers of ES cells, including Oct-4, alkaline phosphatase (Fig. 1 c and f), and SSEA-4 (not shown), for many more passages and maintained a normal karyotype past passage 120 (not shown), indicating that they remained undifferentiated. Aliquots of these cells, passaged up to 40 times after transduction, were placed on the OP9 murine bone marrow stromal cell line to direct differentiation toward hematopoietic progenitor cells. Before culture on the stromal cells, the ES cells expressed markers of pluripotent stem cells (Fig. 1) and did not express hematopoietic lineage markers (Fig. 2a). As described in ref. 3, over time a subset of these transduced ES cells began expressing markers characteristic of hematopoietic cells, including CD34, CD133, and CD117, while retaining expression of EGFP (Fig. 2 a and b). Interestingly, only minimal expression of CD45 was observed. After 7–14 days of culture on OP9, these hESC gained the ability to form myeloid and erythroid colonies in methylcellulose in response to erythropoietin, stem cell factor, granulocyte–macrophage colony-stimulating factor, and IL-3. Unsorted EGFP⁺ cells formed colonies at a frequency of ≈0.5% (data not shown). The resulting colonies retained the ability to express EGFP (Fig. 2c). These results establish that hESC, differentiated toward hematopoietic lineages after coculture with OP9 cells, maintain the expression of transgenes under the control of the EF1α promoter.

Fig. 1.

Genetic modification of undifferentiated hESC. The hESC line H1 was transduced with the EGFP expression lentiviral vector pSIN18.cPPT.hEF1α.EGFP.WPRE and selected for homogenously green colonies. Upper and Lower illustrate the same colonies in each column. (a and b) Phase-contrast (a) and fluorescent (b) images of an EGFP-labeled H1 colony at passage 71. (c, e, and f) The undifferentiated phenotype and transgene expression were maintained as shown by the expression of the early stem cell markers Oct-4 (c) and alkaline phosphatase (f), as well as EGFP (e), at passage 122. (d) DAPI stain.

Fig. 2.

In vitro differentiation of hESC toward hematopoietic lineage. (a) Changes in hematopoietic marker expression after in vitro culture of EGFP⁺ H1 cells on OP9 cells. Cells were analyzed by flow cytometry and gated on EGFP⁺ cells (which typically ranged between 70% and 90% of total live cells). Expression of each marker was determined at the indicated times. (b) Flow cytometric analysis of hematopoietic marker expression in the CD34⁺ cell population (typically <5% of total live cells). (c) hESC-derived hematopoietic progenitor cells give rise to erythroid (Upper) and myeloid (Lower) colonies that express EGFP, as viewed by phase-contrast (Left) and fluorescent (Right) microscopy after 10 days of differentiation on OP9 cells and 14 days of growth in methylcellulose containing erythropoietin, stem cell factor, granulocyte–macrophage colony-stimulating factor, and IL-3.

To provide a microenvironment for T lymphoid differentiation, in separate experiments 1 and 2, we used the SCID-hu (Thy/Liv) mouse, a chimeric model that is constructed by insertion of small pieces of human fetal liver and thymus under the renal capsule of severe combined immunodeficient (SCID) mice (10, 11). In this model, the human fetal liver provides hematopoietic stem cells and the thymus fragments provide stromal elements necessary for T lymphoid differentiation in the conjoint Thy/Liv organ. We and others have shown previously that direct injection of exogenous CD34⁺ human hematopoietic progenitor cells into Thy/Liv implants in sublethally irradiated SCID-hu mice results in engraftment and T lymphoid differentiation of the exogenous cells (12, 13). EGFP-transduced H1 cells were cultured for 7–14 days on OP9 cells in these two independent experiments. Differentiated cells were selected for expression of CD34 or CD133 in the absence of CD34 (CD133⁺/CD34⁻; see Fig. 3a). Some of these cells were assessed for colony-forming potential, the others were directly injected into Thy/Liv implants in SCID-hu mice previously sublethally irradiated (300 rads) to partially deplete endogenous thymocytes and fetal liver-derived progenitor cells. Enriched CD34⁺ cells formed hematopoietic colonies at a frequency of ≈1%, whereas CD133⁺ cells did not form colonies in these studies (not shown).

Fig. 3.

In vivo T lymphoid differentiation of H1-derived progenitor cells expressing EGFP. (a) Schematic representation of differentiation protocol. (b and c) Flow cytometry profiles of cells derived from irradiated SCID-hu Thy/Liv mice 3 weeks (b) (denoted week 3) and 5 weeks (c) (denoted week 5) after transplantation with no cells (Mock) (Upper Left), sorted CD34⁺ progenitor cells (Upper Center), or sorted CD133⁺ progenitor cells (Upper Right) that were derived from HI cells previously cultured on OP9 cells for 11 days (denoted by d11). SCID-hu mice were biopsied on two successive occasions, and thymocytes were analyzed for expression of EGFP, CD45, CD3, CD4, and CD8 by multicolor flow cytometry. Cells were analyzed by gating on live cells (forward vs. side scatter), then gating on CD45⁺ to identify human hematopoietic cells (not shown). (Upper) Live, CD45⁺ cells were then analyzed for CD3 and EGFP. Numbers in each corner represent the percentage of cells in each quadrant, and the gate drawn through the upper and lower right quadrants denotes EGFP⁺ cells, with the percentage of cells given within the gate. (Lower) CD4 vs. CD8 profiles of the EGFP⁺ population, determined by gating.

Three to 5 weeks after injection of hESC-derived cells into Thy/Liv implants, multicolor flow cytometric analysis was used to identify human (CD45⁺) cells that expressed EGFP. EGFP⁺ human cells were identified in Thy/Liv implants in animals receiving either CD34⁺ or CD133⁺ cells (Fig. 3 b and c). Engraftment was seen in 11 of 32 SCID animals and ranged from 0.1% to 6.2% of human cells (Table 1). Whereas the frequency of engraftment was low, no engraftment was seen in animals not previously irradiated (not shown). Analysis for CD3, CD4, and CD8 expression demonstrated that the EGFP-expressing cells could be found in various stages of thymocyte maturation, including immature CD4⁺/CD8⁺ cells and mature CD4⁺/CD8⁻ and CD8⁺/CD4⁻ cells. Successive biopsies of Thy/Liv implants from SCID-hu animals established that engraftment was stable for at least 5 weeks (Fig. 3c). We determined the frequency of T lineage engraftment of hESC cocultured for various times on OP9 cells. Although our data set is limited, it appears that under these conditions, 10 days of coculture on OP9 cells yielded the best engraftment frequency (50%; Table 2).

Table 1.

Engraftment of EGFP⁺ cells in SCID-hu and RAG-hu Thy/Liv implants

Cells	Time in coculture, days	% EGFP+
		Week 3	Week 4	Week 5	Week 6
Experiment 1
CD133	10	NT	0.44	NT	NT
CD133	10	NT	0.42	NT	5.6
CD133	10	2.6	NT	0.16	NT
CD34	10	2	NT	5.8	NT
CD34	10	1	NT	NT	NT
CD34	14	4	NT	0.6	NT
Experiment 2
CD133	7	0.33	NT	6.2	NT
CD133	11	0.1	1.8	2.89	NT
CD133	11	NT	3.45	NT	NT
CD34	7	NT	0.47	NT	NT
CD34	11	0.14	NT	0.71	NT
Experiment 3
CD133	10	NT	24.4	NT	NT
CD133	10	NT	NT	23.1	NT
CD133	10	NT	NT	18.6	NT
CD34	10	NT	3.1	NT	NT
CD34	10	NT	NT	15.6	NT
CD34	10	NT	NT	8.7	NT

Thy/Liv implants in SCID-hu (experiments 1 and 2) and RAG-hu (experiment 3) mice received direct injections of progenitor cells derived from EGFP-labeled H1 cells cocultured on OP9 cells for various times. Progenitor cells were enriched for expression of CD34 or for CD133⁺/CD34⁻ phenotype before injection as indicated. Implants were assessed for engraftment at the times indicated by flow cytometric analysis first by gating on human CD45 and subsequently by quantitating the percentage of CD45⁺/GFP⁺ cells. NT, not tested.

Table 2.

Engraftment frequency of mice to which hESC-derived CD34⁺ or CD133⁺ CD34⁻ cells had been transferred

Cells	Day 7	Day 10	Day 11	Day 14
Experiment 1
CD133+	NT	3/5	NT	0/3
CD34+	NT	2/5	NT	1/3
Experiment 2
CD133+	1/3	NT	2/7	NT
CD34+	1/4	NT	1/2	NT
Experiment 3
CD133+	NT	3/4	NT	NT
CD34+	NT	3/5	NT	NT

Two experiments with SCID-hu mice (experiments 1 and 2) and one experiment with RAG-hu mice (experiment 3) are shown. Those mice found positive for engraftment were determined to have CD45⁺, EGFP⁺ cells at the indicated time points. NT, not tested.

In experiment 3, we attempted to improve engraftment by using human Thy/Liv implants in RAG 2^−/− mice, which are less sensitive to irradiation than SCID mice. This mouse system thus allows a higher dose of irradiation (a total of 900 rads, delivered by two 450-rad exposures, 6 h apart) to deplete endogenous thymocytes and fetal liver-derived progenitor cells before engraftment of hESC. We reasoned that this treatment might allow a greater degree of reconstitution by ES-derived cells, because there would likely be less competition by renewed replication and differentiation of endogenous cells present in the Thy/Liv implants. ES cells were cocultured on OP9 feeder cells for 10 days, based on the initial studies shown above, and similarly sorted ES-derived progenitors were injected into these RAG-hu animals. These animals were also injected with 16-fold more hESC-derived CD34⁺ cells than in experiments 1 and 2. We again saw engraftment of these cells, with a total of six of nine animals containing EGFP-expressing cells (Table 2). These RAG-hu mice had generally smaller implants than did the SCID-hu mice, likely because of the higher irradiation dose. However, the overall percentage of human cells that expressed EGFP in these implants was considerably higher than in SCID-hu mice (Table 1; an example of this finding can also be seen in Fig. 4b). These results suggest that the modifications made in experiment 3 (increased irradiation and higher numbers of CD34 cells injected) together allowed for a greater percent of reconstitution by ES-derived precursor cells in the RAG-hu mice.

Fig. 4.

Phenotypic analysis of hESC-derived thymocytes. (a) To distinguish endogenous HLA-A2⁻ and hESC-derived HLA-A2⁺ cells, thymocytes from control mice (Left) and SCID-hu mice into which hESC-derived CD34⁺ cells had been transferred (Right) were stained and gated based on the expression of CD45 and EGFP. MFI, mean fluorescence intensity. Cells within each gate were subsequently analyzed for the expression of HLA-A2. (b) CD45⁺, EGFP⁺ cells from the Thy/Liv implants from the RAG-hu mice transferred with hESC-derived CD34⁺ cells (Upper) and CD45⁺ cells from the implants of the control animals (Lower) were gated and analyzed for the presence of the indicated T cell surface markers. Region markers are based on isotype controls.

Additional phenotypic studies were performed on cells from experiments 2 and 3. Specifically, we noted that in comparison to the endogenous fetal thymocytes, CD45 expression was dim on the majority of ES-derived thymocytes expressing EGFP (Fig. 4). In experiment 2, the fetal thymus used to generate the Thy/Liv implant was derived from a donor who did not express HLA-A2. Because H1 hESC express HLA-A2, thymocytes derived from these cells could be distinguished from endogenous fetal thymocytes on the basis of HLA-A2 staining (Fig. 4a). Levels of HLA-A2 on ES-derived CD45^dim thymocytes were similar to those on HLA-A2⁺ thymocytes derived from normal SCID-hu mice generated from HLA-A2⁺ fetal donors (not shown) and were ≈10-fold higher than on HLA-A2⁻ fetal liver-derived thymocytes within the same tissue (Fig. 4a). In animals in experiment 3, we assessed the expression of additional thymopoietic markers on these cells. We found levels of CD1a, T cell antigen receptor (TCR), CD3, CD7, and CD127 similar to those expressed on normal thymocytes (Fig. 4b). By gating on CD45^dim HLA-2⁺ cells in these implants that received ES-derived progenitor cells, we could further assess the relative percentage of these cells that maintained the ability to express the EGFP transgene during thymopoiesis. In three animals assessed in experiment 2, EGFP expression was 92%, 95%, and 100% of ES-derived cells (Fig. 5 and data not shown). Thus transgene expression is maintained at high efficiency through thymopoiesis.

Fig. 5.

hESC-derived thymocytes maintain expression of the reporter transgene in vivo. CD45^dim HLA-A2⁺ thymocytes from SCID-hu mice into which hESC-derived CD34⁺ cells had been transferred (Center) were analyzed for the expression of EGFP (Right). (Left) Lack of gated population in control mice receiving no hESC-derived progenitors.

We also determined whether the T lineage cells arising from hESC were functional and able to respond to TCR-mediated signals. Thy/Liv cells from two animals in experiment 2 were cultured ex vivo in control medium or under costimulating conditions (plate-bound anti-CD3 plus soluble anti-CD28). We have shown previously that these costimulating conditions induce the expression of the high-affinity IL-2 receptor (CD25) on human thymocytes (14). As shown in Fig. 6, which illustrates results from one of the two animals tested, costimulated hESC-derived EGFP⁺ thymocytes similarly demonstrated dramatic increases in CD25 expression in comparison with nonstimulated control cells cultured in parallel. Similar responses were seen with cells obtained from the second animal tested (not shown). Thus the thymocytes resulting from transduced hESC are capable of receiving TCR-mediated signals. Together with the phenotypic data presented above, our studies indicate that normal human thymocytes developed from these transduced ES cells.

Fig. 6.

Costimulation of EGFP⁺ thymocytes after 5 weeks of differentiation in vivo. Cells were cultured in medium alone (unstimulated) or in the presence of anti-CD3 and anti-CD28 monoclonal antibodies (costimulated), and EGFP⁺ cells were analyzed for expression of CD25 by flow cytometry.

Discussion

Our results indicate that hESC can develop through the T lymphoid lineage when provided an appropriate microenvironment for differentiation. We observed very low levels of the hematopoietic marker CD45 on the ES-derived precursors that generated these thymocytes in the human thymic implants. However, it has previously been reported that hematopoietic progenitors in early mouse and human embryos lack CD45 expression (15, 16) and that CD45⁻ precursors can differentiate into various hematopoietic lineages in both human and mouse ES systems in vitro (17, 18). Interestingly, hESC-derived thymocytes did express CD45; however, the levels of this marker were lower than those found on typical thymocytes derived from fetal liver progenitors in the SCID-hu mouse. This CD45 expression may reflect the embryonic rather than fetal origin of these cells. However, virtually every other T-lineage marker assessed was similar on thymocytes derived from ES cells and fetal liver progenitors. In addition, ES-derived thymocytes responded normally to costimulation, suggesting that they are functional.

The availability of SCID-hu mice constructed from HLA-A2⁻ donors allowed us to discriminate HLA-A2⁺ ES-derived cells from endogenous human thymocytes. This discrimination further allowed us to determine the relative level of ES-derived thymocytes that retained the ability to express the EGFP transgene while undergoing differentiation in the human thymus. Importantly, we found that transgene expression was generally high, with ≈90% of ES-derived cells expressing detectable levels of EGFP under the control of the EF1α promoter. This finding suggests that genetic manipulations of these cells are feasible. With further optimization to increase efficiency and/or incorporate inducible promoters, this system could prove useful in providing a ready source of primary human T lymphocytes expressing genes of interest for basic mechanistic studies. Similarly, human T cells expressing inhibitory RNAs could be generated without manipulation of the T cell itself, to provide a “knockdown” phenotype. This type of strategy could prove quite useful, for example, for mechanistic studies of quiescent T cells or of T cell activation, because stimulation or perturbation of the T cell to introduce the vector would not be required. Thus, effects of the transgenes could be accurately assessed because responses of these cells to experimental stimuli would be otherwise similar to those of true quiescent cells.

Further studies are required to define the exact molecular signals needed to induce the appropriate hematopoietic differentiation pathway. However, these results have important implications for hematopoietic stem cell gene therapy approaches. Genetically manipulated hESC with well characterized vector integration sites could be expanded to large numbers. If these manipulated cells were intended to be used to therapeutically target the T lineage, our results suggest that they need only be differentiated as far as hematopoietic progenitor cells in vitro and that the thymus would then complete the differentiation program to CD4⁺ and CD8⁺ T cells in vivo. This type of approach might lead to improved therapeutic strategies to treat genetic hematopoietic disorders, such as X-linked SCID (19) or infectious diseases such as acquired immunodeficiency syndrome (20).

Materials and Methods

Cell Culture.

The hESC line H1 was cultured on mouse embryonic fibroblast feeders in D-MEM/F12 medium containing 20% serum replacer (Invitrogen), 2 mM l-glutamine, a 100 μM concentration of each nonessential amino acid, and 8 ng/ml basic fibroblast growth factor. The cells were passaged on a weekly basis, and passages 88, 90, and 108 were used for the differentiation studies. All work with hESC was approved by the University of California Los Angeles Institutional Review Board. The mouse bone marrow stromal cell line OP9 (provided by Owen Witte, Howard Hughes Medical Institute, University of California, Los Angeles) was maintained in α-modified minimum essential medium (α-MEM) containing 20% FBS.

Differentiation Protocol.

For the differentiation studies, a protocol from Vodyanik et al. (3) was adapted. Briefly, gelatinized six-well plates were seeded with 2 × 10⁴ OP9 cells per well and grown for 5 days in OP9 medium, with the medium change on day 4. Undifferentiated H1-GFP hESC were harvested with collagenase IV (1 mg/ml) and dispersed into small clumps by scraping and pipetting. The resulting clumps were added to the OP9 layers, and the OP9/H1 cocultures were maintained for 5–14 days in α-modified minimum essential medium supplemented with 10% non-heat-inactivated defined FBS (HyClone) and 100 μM monothioglycerol (Sigma), with half-medium changes every second day. To obtain a single-cell suspension, the cultures were treated with collagenase IV for 20 min at 37°C, followed by treatment with trypsin/EDTA (0.05%) for 15 min at 37°C. The cell clumps were further disrupted by pipetting, and the resulting suspension was washed twice and filtered through a 70-μm strainer. The cells were then moved into T162 flasks (Corning) and incubated at 37°C for 1 h to allow OP9 cells to adhere to enrich for hESC-derived cells. This step removes ≈90% of OP9 cells. Nonadhered cells were collected and used for flow-based phenotypic analysis and further purification steps.

Virus Production.

The virus was produced in 293T cells by cotransfection of three plasmids: (i) pSIN18.cPPT.hEF1α.EGFP.WPRE, a vector expressing EGFP under control of the EF1α promoter (9), (a gift from M. Gropp and B. E. Reubinoff, Hadassah University Hospital, Jerusalem, Israel), (ii) the vesicular stomatitis virus G (VSVG) expression plasmid pHCMVG, and (iii) the packaging plasmid pCMVΔR8.2DVPR. In our protocol, 5 × 10⁶ 293FT human fibroblasts, plated on a 100-mm² dish the previous night, were transfected with 5 μg of pSIN18.cPPT.hEF1α.EGFP.WPRE, 5 μg of pCMVΔR8.2DVPR, and 2 μg of pHCMVG, by using Lipofectamine 2000 (Invitrogen) according to the manufacturer’s instructions. Culture supernatants that contained virus were collected on days 2 and 3, centrifuged to remove floating cells, passed through 0.45-μm filters, and subjected to centrifugation at 17,000 rpm for 90 min at 4°C in an L8-M ultracentrifuge (Beckman) to concentrate the virus. The viral pellets were resuspended in 1× PBS overnight at 4°C. Concentrated virus was titrated on the 293T fibroblast cell line.

Transduction of H1 hESC.

At the time of infection, undifferentiated H1 colonies were mechanically disrupted into smaller clumps and mixed with the virus in 200-μl final volume consisting of PBS and HES medium (1:1 ratio) in the presence of 4 μg/ml Polybrene (Sigma). The mixture was incubated for 2 h at 37°C with constant shaking, then washed and plated on fresh mouse fibroblast feeders. Newly grown colonies consisting of both transduced and nontransduced cells were analyzed under UV microscopy, and GFP-positive regions were selectively excised and passaged. After several such passages, H1 colonies were enriched for GFP expression to near-complete homogeneity.

Immunohistochemistry.

H1-GFP cells were grown on coverslips on a feeder layer of mouse embryonic fibroblasts. For alkaline phosphatase (AP) staining, cells were first fixed in 0.2% formaldehyde for 1 h at 37°C, then washed with 1× PBS. Staining for AP was performed with the Vector Red alkaline phosphatase substrate kit I (Vector Laboratories), following the manufacturer’s instructions. For Oct-4 staining, cells were first fixed in cold methanol for 15 min followed by permeabilization in PBS/0.1% Triton X-100 for 20 min at room temperature. The coverslips were blocked in 0.5% goat serum (Calbiochem) and 0.3% BSA (Sigma) for 1 h followed by incubation with Oct-3/4 antibody (Santa Cruz Biotechnology). Secondary staining was performed with goat anti-mouse IgG2b-phycoerythrin (Santa Cruz Biotechnology). Isotype control antibody was purchased from Serotec. The coverslips were mounted on a glass slide in VECTASHIELD mounting medium with DAPI (Vector Laboratories) and observed under bright and fluorescent fields.

In Vivo T Cell Development.

SCID-hu mice were generated as described in ref. 21. Briefly, small pieces of human fetal liver and thymus were inserted beneath the renal capsule of SCID mice and allowed to develop into a thymus-like organoid called a Thy/Liv implant. RAG-2^−/− mice were used in experiment 3 instead of SCID mice. All immunodeficient mouse work was approved by the University of California Los Angeles Animal Research Committee. SCID-hu mice were irradiated with 300 rads and injected with either 5 × 10⁴ purified CD34⁺ or 10⁶ purified CD133⁺ CD34⁻ cells directly into Thy/Liv implants. RAG-hu mice were exposed to a total of 900 rads, given in two doses of 450 rads administered 6 h apart, and were reconstituted with 2 × 10⁶ mouse bone marrow cells. The implants of these animals were injected with either 8 × 10⁵ purified CD34⁺ or 10⁶ purified CD133⁺ CD34⁻ hESC-derived cells. The Thy/Liv implant biopsies were performed at various time points as specified in the text. Single-cell suspensions were obtained, and the cells were analyzed for the expression of EGFP and T cell developmental markers and used in the functional assay.

Cell Sorting and Flow Cytometry.

H1-GFP hESC cultured in vitro and cells derived from Thy/Liv implants were stained with monoclonal antibodies to CD3, CD4, CD8, CD19, CD25, CD34, CD45, CD56, CD127 (Coulter); CD1a, CD7, CD117 (Bioscience, San Diego, CA); SSEA-4 (R & D Systems); HLA-A2 (Serotec); TCR (Pharmingen); and CD133 (Miltenyi Biotec, Auburn, CA), conjugated to phycoerythrin-cyanin 7, electron-coupled dye, allophycocyanin, or PC7. Cells were analyzed for fluorochrome and EGFP expression by flow cytometry with a Coulter FC500 flow cytometer and flojo (Tree Star, Ashland, OR). At the indicated times after differentiation in vitro, H1-GFP hESC were sorted by magnetic cell sorting (MACS). Cells were initially magnetically labeled with anti-CD34 microbeads (Miltenyi Biotec), and the positive fraction was collected after two-column sorting (possel_d2 program) on an AutoMACS cell sorter (Miltenyi Biotec). The CD34⁻ fraction was then labeled with anti-CD133 microbeads (Miltenyi Biotec), and the CD133⁺ fraction was collected after another AutoMACS sort (possel_d2 program). Typically, there was >85% enrichment for the respective CD34⁺ and CD133⁺ populations as determined by postsort flow cytometry analysis.

Cell Stimulation.

Cells derived from biopsied Thy/Liv implants were cultured in RPMI medium 1640 containing penicillin (100 units/ml), streptomycin (100 μg/ml) (Sigma), and 10% human AB serum (Gemini Bioproducts, Sacramento, CA). A fraction of the cells were costimulated with anti-CD3 (1 μg/ml) (Ortho Biotech, Bridgewater, NJ) and cross-linked to the plate with goat anti-mouse IgG and soluble anti-CD28 (100 ng/ml) (Coulter). Cells were then removed after 3 days of costimulation and assessed for phenotype by flow cytometry for CD25 expression.

Abbreviations

EF1α

elongation factor 1α

TCR

T cell antigen receptor.