Human Embryonic Stem Cells and Their ... - Stem Cells Journals

Download PDF
4 downloads 0 Views 618KB Size Report
Comment
A crucial step toward their successful clinical appli- cation is to examine the ... lymphocytes has a pivotal role in the development of immune response toward an ... In this study, by using immunocompetent mice and mouse strains with different ...
Embryonic Stem Cells–Characterization Series Human Embryonic Stem Cells and Their Differentiated Derivatives Are Less Susceptible to Immune Rejection Than Adult Cells MICHA DRUKKER,a HELENA KATCHMAN,b GIL KATZ,c SMADAR EVEN-TOV FRIEDMAN,b ELIAS SHEZEN,b ERAN HORNSTEIN,a OFER MANDELBOIM,c YAIR REISNER,b NISSIM BENVENISTYa a

Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, Jerusalem, Israel; Department of Immunology, Weizmann Institute of Science, Rehovot, Israel; cThe Lautenberg Center for General and Tumor Immunology, Hadassah Medical School, The Hebrew University, Jerusalem, Israel

b

Key Words. Human embryonic stem cells • Peripheral blood mononuclear cells • Embryoid bodies • Cytotoxic T lymphocytes

ABSTRACT Differentiated cell types derived from human embryonic stem cells (hESCs) may serve in the future to treat various human diseases. A crucial step toward their successful clinical application is to examine the immune response that might be launched against them after transplantation. We used two experimental platforms to examine the in vivo leukocyte response toward hESCs. First, immunocompetent and immunodeficient mouse strains were used to identify T cells as the major component that causes xenorejection of hESCs. Second, mice that were conditioned to carry peripheral blood leukocytes from human origin were used to test the human leukocyte alloresponse toward undifferentiated and differentiated hESCs. Using this model, we have detected only a minute immune response toward undifferentiated as well as differen-

tiated hESCs over the course of 1 month, although control adult grafts were repeatedly infiltrated with lymphocytes and destroyed. Our data show that the cells evade immune destruction due to a low immunostimulatory potential. Nevertheless, a human cytotoxic T lymphocyte clone that was specifically prepared to recognize two hESC lines could lyse the cells after major histocompatibility complex class I (MHC-I) induction. Although MHC-I levels in hESCs are sufficient for rejection by cytotoxic T cells, our data suggest that the immunostimulatory capacity of the cells is very low. Thus, immunosuppressive regimens for hESC-based therapeutics could be highly reduced compared with conventional organ transplantation because direct allorejection processes of hESCs and their derivatives are considerably weaker. STEM CELLS 2006;24:221–229

INTRODUCTION

tinocytes [18], osteoblasts [19], and hepatocytes [20, 21]. However, a major concern that may severely limit the use of hESCs in therapeutics is the immune processes that might be launched against the differentiated cells after transplantation into immunocompetent patients. The first attempts to determine the antigenicity (capacity to present antigens) of very early human embryonic cells were carried out in blastocysts. Using anti–major histocompatibility complex (MHC) antibodies, a weak expression of MHC class I

Human embryonic stem cell (hESC) lines are excellent candidates to serve as a valuable source of cells in transplantation medicine because they have the capacity to grow indefinitely in culture without losing pluripotency, as well as to differentiate to all cell types of the body upon proper induction [1–3]. For example, detailed differentiation protocols are available today for the derivation of neurons [4 – 8], cardiomyocytes [9 –14], endothelial cells [15], hematopoietic precursors [16, 17], kera-

Correspondence: Micha Drukker, Ph.D., Department of Genetics, Silberman Institute of Life Sciences, The Hebrew University, 91904 Jerusalem, Israel. Telephone: 972-2-6585183; Fax: 972-2-6584972; e-mail: [email protected] Received April 22, 2005; accepted for publication August 2, 2005; first published online in STEM CELLS EXPRESS August 18, 2005. ©AlphaMed Press 1066-5099/2006/$20.00/0 doi: 10.1634/stemcells.2005-0188

STEM CELLS 2006;24:221–229 www.StemCells.com

Human ES Cell Immunogenicity

222

(MHC-I) molecules on the inner cell mass (ICM) was noted [22]. We have previously addressed this issue by examining the expression of MHC antigens in hESCs, in embryoid bodies (EBs), and in hESC-induced teratomas [23]. We found that undifferentiated cells express low levels of MHC-I molecules, which were elevated in EBs and even to a higher extent in teratomas. MHC class II (MHC-II) molecules, however, were not expressed under these conditions. Importantly, we found that the undifferentiated as well as the differentiated cells upregulate MHC-I cell-surface expression by at least 10-fold in response to interferons (IFNs), although MHC-II expression was not induced. Although hESCs express relatively low levels of MHC-I, we could show that they are insensitive to human natural killer (NK) cell-mediated cytotoxicity [23]. However, the actual capacity of human T cells to recognize and attack hESCs and their differentiated derivatives has remained unknown. Similar to their central role in pathogen removal, the successful engagement of antigen-presenting cells (APCs) and T lymphocytes has a pivotal role in the development of immune response toward an allograft (HLA-mismatched transplant). It is widely accepted today that after transplantation, donor-derived APCs (especially dendritic cells) emigrate to regional lymph nodes, where they encounter and stimulate naive or memory allospecific host T cells to proliferate. Central to this process are the costimulatory molecules B7.1 (CD80) and B7.2 (CD86) that promote clonal expansion of T cells and their differentiation into effector cells, which can attack the graft upon return to circulation [24]. In contrast, immunosuppressive signals such as Fas ligand (FasL, CD95L) can limit alloresponsiveness toward foreign grafts by activating the Fas receptor (CD95) in lymphocytes [25]. For instance, it was recently suggested that FasL may protect rat embryonic stem cell-like cells (RESCs) from rejection in allogeneic rat hosts [26]. In this study, by using immunocompetent mice and mouse strains with different types of immunodeficiency, we attempted to define a precise mechanism of hESC xenorejection. To determine the human T cell alloresponse toward undifferentiated as well as differentiated hESCs in vivo and in vitro, we used the human/mouse Trimera model and a primary human cytotoxic T cell (CTL) line, respectively. Furthermore, we evaluated the capacity of the hESC-derived transplants to activate immune rejection using quantitative molecular analysis methods.

MATERIALS

AND

METHODS

Animals, Immunization, and Transplantation Procedures Animals were maintained under conditions approved by the Institutional Animal Care and Use Committee at the Weizmann Institute. For transplantation studies in immunodeficient mice, SCID 3 CB6F1 mouse chimeras were generated. Briefly, CB6F1 mice (Harlan Olac, Oxon, U.K., http://www.harlan.com) were exposed to split-dose total body irradiation. Irradiated mice were radioprotected with bone marrow cells from nonobese diabetic severe combined immunodeficient (NOD/SCID) mice (Weizmann Institute Animal Breeding Center, Rehovot, Israel, http://www.weizmann.ac.il). One to 3 days later, approximately 100 ⫻ 106 human peripheral blood mononuclear cells (PBMCs) were injected intraperitoneally to reconstitute human T, B, and

NK cell populations. The engraftment of human lymphocytes was confirmed as described [27]. Immunization was carried out 1 week before transplantation by IP injection of approximately 5 ⫻ 106 irradiated target cells. Target cells or tissues were transplanted under the kidney capsule. Transplantation to immunocompetent (BALB/c, C57Bl, SJL, and CB6F1) and immunodeficient (NOD/SCID, C57BL/ 6J-Lystbg, Balb-nude, and CBA/CaHN Btkxid) animals was carried out in 10- to 12-week-old mice (Jackson Laboratory, Bar Harbor, ME, http://www.jax.org).

Transplantation Procedures All operations were performed under general anesthesia (2.5% Avertin in phosphate-buffered saline [PBS], 10 ml/kg). Host kidney was exposed through a left lateral incision. A 1.5-mm incision was made at the caudal end of the kidney capsule, and approximately 1 ⫻ 106 cells suspended in 50 to 100 ␮l PBS were injected using lacrimal cannula (catalog no. G-15161, Geuder, Heidelberg, Germany, http://www.geuder.de). Alternatively, 1- to 2-mm-diameter tissue fragments were grafted under the kidney capsule.

Transplant Growth and Differentiation Animals receiving implants were euthanized 3 to 5 weeks after transplantation, and their kidneys were then removed and fixed in 10% formaldehyde. Transplants were sectioned, mounted on slides coated with poly-L-lysine, and stained by hematoxylin and eosin (H&E).

Immunohistochemical and Immunocytometry Staining hESCs and their differentiated derivatives were trypsinized for fluorescence-activated cell sorter (FACS) analysis and stained with W6/32 and BBM1 monoclonal antibodies (mAbs) as previously described [23]. CD80 and CD86 antigens were detected by DAL-1 (immunoglobulin G1 [IgG1]) FITC-conjugated and B-T7 (IgG1) phycoerythrin (PE)-conjugated mAbs, respectively. FITC- and PE-conjugated MCG1 mAb clone was used as isotype control (IQ Products, Groningen, The Netherlands, http://www.iqproducts.nl). BB7.2 mAb (Serotec, Oxford, U.K., http://www.serotec.com) was used to verify the previously reported HLA-A2 isotype of hESC lines H9 and H13 [23]. Human specific NOK-1 mAb (ebioscience, San Diego, CA, http://www. ebioscience.com) was used for FasL detection. FITC-conjugated goat F(ab⬘) fragment (MP Biomedicals, Aurora, OH, http:// www.mpbio.com) was used as a secondary antibody. Determination of human T cell infiltration in tumors and tissues was performed using mouse anti-human CD3 and CD45 antibodies (DakoCytomation, Glostrup, Denmark, http://www. dakocytomation.com) on paraffin-embedded sections. Horseradish peroxidase-conjugated anti-mouse EnVision System (DakoCytomation) was used to detect the primary antibody.

Cells, Tissues, and Viruses hESCs and teratoma cell lines were cultured and treated with interferon (IFN)-␥ as previously described [23, 28]. Teratoma subclones were obtained from a 4-week-old hESC-induced teratoma. Class I MHC-negative human cell line 721.221, HLAA2, and Cw3 transfectants [29], Burkitt’s lymphoma cells (Raji), and KFL9 [30] cells (kindly provided by Dr. David

Drukker, Katchman, Katz et al. Kaplan) were cultured in RPMI-1640 medium, supplemented with 10% heat-inactivated fetal calf serum (Gibco, Grand Island, NY, http://www.invitrogen.com). Skin grafts were obtained from healthy individuals undergoing cosmetic surgery. The influenza A/Sydney (A/Sydney/5/97-like [H3N2]) was propagated as previously described [31]. The cells were infected by incubating 1 ⫻ 106 cells overnight in 3 ml complete medium at 37°C and 5% CO2 with 100 ␮l of virus preparation.

Peptides Influenza-A-matrix (IV/A) peptide GILGFVFTL was assembled by conventional solid-phase synthesis using an ABIMED AMS-422 automated solid-phase multiple-peptide synthesizer (Abimed, Langenfeld, Germany, http://www.abimed.de) on a 25-␮mol scale. Fmoc strategy was used throughout the peptide chain assembly following the company’s protocol. All protected amino acids, coupling reagents, and polymers were obtained from Nova Biochemicals (Novabiochem, Laufelfingen, Switzerland, http://www.emdbiosciences.com). Synthesis-grade solvents were obtained from Labscan (Dublin, Ireland, http://www. labscan.ie).

Peptide Loading on Acid-Stripped Cells

Modifying a previously published protocol [32], 1 ⫻ 106 suspension cultured cells or trypsinized adherent cultured cells were pelleted and treated with 50 ␮l of 1% bovine serum albumin (Amresco, Solon, OH, http://www.amresco-inc.com), 300 mM Glycine (Sigma-Aldrich, St. Louis, http://www. sigmaaldrich.com) at pH 2.4. After incubation at room temperature for 90 seconds, the suspension was neutralized by adding 5 ml of RPMI 1640 medium and centrifuged at 500g. Acidstripped cells were resuspended in 1 ml RPMI 1640 containing 2.5 ␮g recombinant human ␤2m (BD PharMingen, San Diego, http://www.bdbiosciences.com/pharmingen) and 50 ␮l influenza-A-matrix (IV/A) peptide GILGFVFTL solubilized in 100% dimethylsulfoxide (Merck, Hohenbrunn, Germany, http://www. merck.com) at a concentration of 1 mg/ml. Control cells were treated identically but with no peptide. The suspensions were left in 37°C incubator for 1.5 hours, washed in 10 ml PBS, and incubated with effector PBMCs.

Production of Cytotoxic Cells and Assay To prepare HLA-A2-restricted CTL line specific for nonameric peptide 58 – 66 (GILGFVFTL) derived from the influenza virus matrix A protein, 2 ⫻ 106 PBMCs from an HLA-A2⫹ donor were separated on Ficoll-Paque (Amersham Biosciences, Uppsala, Sweden, http://www.amersham.com) and incubated in 1 ml RPMI 1640 plus 10% human serum (Sigma-Aldrich) in the presence of 50 ␮M peptide. After 1 hour, the cells were washed, irradiated at 5,000 rad, and mixed with equal number of untreated PBMCs (responders) at a final concentration of 2 ⫻ 106 PBMCs in RPMI 1640 medium supplemented with human serum. Seven days later, the cells were resuspended in the same medium with 2.5 U/ml recombinant human interleukin-2 to induce T cell proliferation (Roche Diagnostics, Mannheim, Germany, http://www.roche-applied-science.com). After an additional 7 days, the cytotoxic activity of the CTL line was tested against 721.221/HLA-A2 and 721.221/Cw3 in the presence and absence of IV/A peptide. Cultures that specifically lysed at least www.StemCells.com

223

40% of peptide-pulsed cells and less than 10% of nonpulsed cells were maintained for future experiments. To measure cell lysis against various target cells, the cells were labeled overnight with 1 ␮Ci 35S methionine (Amersham Biosciences) per 1 ⫻ 106 cells. Effector cells were mixed with labeled target cells at different effector-to-target cell ratios (E/T) in U-bottomed microtiter plates. Assays were terminated by centrifugation at 1,000 rpm for 10 minutes at 4°C, and 100 ␮l of the supernatant was collected for liquid scintillation counting. Specific lysis was calculated as follows: % lysis ⫽ (cpm experimental well ⫺ cpm spontaneous release)/(cpm maximal release ⫺ cpm spontaneous release) ⫻ 100. Spontaneous release was determined by incubation of each group of labeled target cells with medium. Maximal release was determined by solubilizing target cells in 0.1 M NaOH. Each experiment was repeated three times. Each point represents the average of duplicate values. Error was ⬍5% of the mean of the duplicates.

DNA Microarray Analysis Total RNA was extracted from populations of undifferentiated and differentiated cell derivatives of the H9 hESC line. RNA extraction, hybridization to the U133A DNA microarrays, washing, and scanning were performed according to the manufacturer’s protocols (Affymetrix, Santa Clara, CA, http://www.affymetrix.com). The fluorescent signal was translated to expression value using the Affymetrix MAS5 program. All data were normalized to a mean of 1 and then log-transformed (using log base 2). DNA microarray data obtained from undifferentiated hESCs and 2-, 10-, and 30day-old EBs and 4-week-old induced teratomas (two to three replicas each) were compared with gene expression profiles of five normal bone marrow samples, three samples of sorted CD34⫹ cells [33], and two samples of CD56⫹ NK cells (GSM18875, GSM18876), CD4⫹ T cells (GSM18877, GSM18878), CD8⫹ T cells (GSM18879, GSM18880), CD14 monocytes (GSM18871, GSM18872), CD33 myeloid cells (GSM18869, GSM18870), BDCA4⫹ dendritic cells (GSM18873, GSM18874), whole blood (GSM18867, GSM18868), thymus (GSM18899, GSM18900), lymph node (GSM18903, GSM18904), lung (GSM18949, GSM18950), heart (GSM18951, GSM18952), liver (GSM18953, GSM18954), kidney (GSM18955, GSM18956), and skin (GSM19001, GSM19002) (available at http://www.ncbi.nlm.gov/ geo; accession no. GSE1133). Genes were selected on the basis of the GEArray S series human autoimmune and inflammatory responsegenearray(http://www.superarray.com/gene_array_product/ HTML/HS-602.3.html). Analysis of the results was performed using the Spotfire program (Spotfire, Somerville, MA, http:// www.spotfire.com). In addition, we used the following databases: Unigene, Entrez, LocusLink, and Geo from NCBI. The 210865_at (U133A DNA microarray) probe was used to measure FasL mRNA expression in the aforementioned tissues.

RESULTS Transplantation and Differentiation of hESCs in Immunocompetent and Various Strains of Immunodeficient Mice To test immune response evoked by hESCs in a xenogeneic environment, we transplanted undifferentiated hESCs under the kidney capsule of four different immunocompetent mouse strains (BALB/c, C57Bl, SJL, and CB6F1). In this assay, trans-

Human ES Cell Immunogenicity

224

plantation of the cells to NOD/SCID mice served as positive control. Without exception, 1 month after injection, hESCs failed to develop to teratomas in all strains of immunocompetent mice, in contrast to NOD/SCID mice, which had massive teratomas (Table 1). To address the mechanism of xenorejection, transplantation of hESCs into different immunodeficient mouse strains was undertaken. NOD/SCID mice served as a positive control and, as expected, teratoma development was noted in all transplanted animals (five out of five). Similarly, normal teratoma growth without any signs of rejection (five out of five) was observed in nude mice that are characterized by absence of T cells and the presence of normal levels of B and NK cells. In contrast, transplantation of hESCs to Lystbg (NK cell-deficient) and Btkxid (B cell-deficient) mice led to vigorous rejection and failure to develop teratoma (zero out of five in both groups). These experiments clearly show that xenorejection of hESCs is T cell-mediated, whereas NK cells or B cells play only a minor, if any, role in this process.

Transplantation and Differentiation of hESCs in the Trimera Model The in vivo intensity by which the human immune system may act to reject hESC-derived transplants was evaluated using the human PBMC-reconstituted Trimera mouse model [27, 34] (supplementary online Fig. 1). Target cells or tissues were transplanted under the kidney capsule of the animals either before intraperitoneal PBMC reconstitution or 1 week after immunization by the same type of cells to PBMC-reconstituted animals. Such immunization generated a more robust allospecific response [27]. Grafts were harvested 3 to 5 weeks after transplantation, and human Ig-specific ELISA tests were used to confirm human PBMC engraftment [35]. Three weeks after transplantation of human adult skin grafts under the kidney capsule of nonimmunized PBMC-reconstituted mice, tissue damage and massive infiltration by CD45⫹ human leukocytes in the graft were observed in all cases (nine out of nine) (Fig. 1A). In contrast, approximately 1 ⫻ 106 undifferentiated hESCs transplanted under the kidney capsule of PBMC-reconstituted mice developed into teratoma tumors (10 to 20 mm), exhibiting normal growth rate (compared with nonreconstituted mice) and no tissue damage or infiltration by CD45⫹ leukocytes (10 out of 10) (Fig. 1B). Similarly, small-differentiated tissue fragments derived from a 4-week-old hESC-induced teratoma developed into teratoma structures after transplantation and PBMC reconstitution. Although these fragments are composed of multiple cell types, in these tumors only minor scattered infiltration was noted, with no apparent slowed growth rate or graft destruction (nine out of nine) (Fig. 1C). To confirm these results, we stained the same grafts by anti-human CD3 antibody. The staining could demonstrate that most of the infiltrating leukocytes in the graft are human T cells (supplementary online Fig. 2). Similar results were obtained in the second set of experiments, in which grafts were transplanted 1 week after im-

Figure 1. Infiltration of human-derived grafts by CD45⫹ human leukocytes. Shown are histological sections of 3-week-old kidney subcapsular grafts stained by human-specific anti-CD45 antibody (original magnification ⫻10, counterstaining by hematoxylin-eosin). (A): After transplantation into immunodeficient mice (-PBMC), human skin grafts appear histologically normal (3 to 4 mm in length; marked are the two chief layers, dermis [open arrows] and epidermis [arrow heads]). In human leukocyte-reconstituted mice (⫹PBMC), the skin shows clear signs of rejection; the epithelium is degenerated and infiltrated with human leukocytes (2 to 4 mm in length) (arrows). (B): Transplantation of undifferentiated human embryonic stem cells resulted in teratoma formation in immunodeficient mice (-PBMC) as well as in immunocompetent mice (⫹PBMC), with no signs of growth interruption and infiltration by human leukocytes (10 to 20 mm in length for both). (C): Differentiated tissue fragments, derived from hESC-induced teratoma, continued to develop upon transfer to new hosts. No tissue damage was evident in the presence of human leukocytes (⫹PBMC), and only scattered leukocytes could be detected (arrows) (5 to 10 mm in length for both).

munization with irradiated target cells. Transplantation of undifferentiated hESCs, differentiated teratoma fragments, as well as teratoma-derived primary cell line resulted in similar tumor formation and cell growth regardless of the human PBMC reconstitution. In contrast, Burkitt’s lymphoma cells (Raji) develop into substantial tumors in the nonreconstituted mice but were completely eliminated in the PBMC-reconstituted animals (Table 2). These data indicate that hESCs and their differentiated derivatives have little potential to activate a direct allospecific response by HLA nonmatched human lymphocytes.

Table 1. Development of hESCs derived grafts in immunocompetent mice Undifferentiated hESCs

BALB/c

CB6F1

C57Bl

SJL

SCID/NOD

0 of 14

0 of 16

0 of 6

0 of 6

8 of 8

Abbreviations: hESC, human embryonic stem cell; SCID/NOD, nonobese diabetic severe combined immunodeficient.

Drukker, Katchman, Katz et al.

Restriction of Immune Response Toward hESCs Is Not Due to FasL Inhibition The restriction of immune processes in the immune-privileged sites of the body is attributed partially to the expression of FasL, a known inducer of T cell apoptosis [25]. Recently, it was suggested by Fandrish et al. [26] that FasL expression in RESC lines may act to protect these cells from rejection by alloreactive T cells after transplantation into MHC-mismatched rat hosts. To examine this possibility, the expression levels of FasL mRNA and cell-surface protein were determined in hESCs. Our analysis reveals that hESCs of various stages of differentiation do not express FasL mRNA in contrast to other somatic tissues (Fig. 2A). We could confirm these results using the mouse anti-

225 Table 2. Development of hESC-derived grafts in human/mouse radiation chimera PBMC reconstituted Nonreconstituted Undifferentiated hESCs Teratoma fragments* Teratoma derived cell line

35 of 35 4 of 4 5 of 5

12 of 12 6 of 6 3 of 3

*Immunization with embryoid body cells. Abbreviations: hESC, human embryonic stem cell; PBMC, peripheral blood mononuclear cell.

human FasL monoclonal antibody NOK-1 [36] (Fig. 2B). Therefore, we conclude that the resistance of hESCs to alloreactive T cells is not due to FasL expression, yet other mechanisms that might trigger activation-induced T cell death cannot be ruled out.

Multiple Factors Are Associated with the Reduced Immunogenicity of hESCs

Figure 2. Expression of immune-modulating genes in human embryonic stem cells (hESCs). (A): Mean expression of FasL mRNA as measured by DNA microarray in undifferentiated hESCs (ES), 2-day, 10-day, and 30-day differentiated EB cells and teratoma (red), leukocytes (CD8⫹ T cells, whole blood, CD56⫹ natural killer cells; green), and solid organs (thymus, lymph node, lung, heart, liver, kidney, skin; blue). Shown are standard error bars of two to five independent experiments. *Absent by Affymetrix p value analysis. Cells were stained by human-specific monoclonal antibodies against FasL (B) CD80 (B7.1) and CD86 (B7.2) (C) and evaluated by fluorescence-activated cell sorter. KFL9 cells and B cell line (721.221) were used as a positive control for FasL and CD80 and CD86 staining, respectively. Broken line indicates background fluorescence.

www.StemCells.com

The weak immune response toward hESCs also might be explained by T cell ignorance or anergy [37]. The two most important signals that must be delivered to T cells to induce their activation are engagement of the foreign MHC molecules with the T cell receptor and costimulation delivered mainly by costimulatory molecules, such as CD80 and CD86 molecules [38]. As demonstrated in our previous study, hESCs do not express MHC-II molecules, even after in vivo differentiation, but show low levels of MHC-I molecules, which are slightly induced by differentiation [23]. To test whether hESCs can also deliver costimulatory signals to T cells, we examined the expression of CD80 and CD86 on their cell surface by FACS (Fig. 2C). We found that the two proteins were not expressed on undifferentiated hESCs in contrast to control cells (B cell line). Moreover, application of IFN-␥, a known inducer of costimulatory molecules [38], did not induce CD80 and CD86 expression, although its receptor is expressed on the cells (not shown). Furthermore, a differentiated primary cell line derived from an hESC-induced teratoma (Tu cell line) was also found to be CD80 and CD86 negative, indicating that hESCs have a poor costimulatory capacity even after their differentiation and IFN-␥ treatment. To extend this analysis further, the transcription pattern of 306 genes that are known to be involved in immune responses was examined in hESCs and their differentiated progeny using 445 DNA microarray probe sets. This analysis included genes that encode for cytokines, chemokines and their receptors, leukocyte adhesion molecules, cell-surface receptors, signal transduction proteins, and transcription factors that mediate immune responses. Expression profiles were prepared from three independent samples of undifferentiated hESCs, 2-day-old EBs, 10-day-old EBs, 30-day-old EBs, and two 4-week-old teratomas [39] and compared with the expression profile of nine different leukocyte subsets and seven organs. Hierarchical clustering of the gene expression profiles shows that hESCs and their derivatives were clustered together in one group, which was separated from the other tissues (Fig. 3A). The most prominent distinction between the gene expression patterns of hESCs and the organs involves approximately 34 genes, which are highly expressed in the latter but were almost undetectable in hESCs

226

Human ES Cell Immunogenicity reasonable that using other differentiation protocols, at least a part of the population could develop a costimulatory capacity [43]. Thus, it is important to determine whether MHC-I cellsurface levels are sufficient for recognition by effector T cells, which develop after stimulation from naive and memory T cells. To this end, we artificially generated a human CTL line that specifically recognizes the HLA-A2 antigen expressed on hESC lines H9 and H13 (Fig. 4A). It is also crucial to test T cell response toward IFN-stimulated hESCs because these cytokines are released naturally during the course of many immune responses. Stimulatory leukocytes were produced by exogenously loading the HLA-A2-specific influenza virus type A (IV/A) matrix synthetic peptide GILGFVFTL on irradiated PBMCs from an HLA-A2⫹ donor. Coculture of these cells with untreated PBMCs from the same donor led to sensitization and expansion of HLA-A2-restricted and influenza peptide-specific CTLs. To measure the susceptibility of IV/A peptide-loaded target cells to lysis by effector CTLs, we used a lysis assay in which different amounts of CTLs are incubated with a constant population of labeled target cells. Percentage of killing was plotted against the ratio of CTL to target cell (E-to-T ratio). The restriction of the CTL line to the HLA-A2 isotype was demonstrated using 721.221 transfectants that express the HLA-A2

Figure 3. Hierarchical clustering of immune-related gene expression data. (A): The 50 tissue samples were divided into three categories based on difference in gene expression: undifferentiated human embryonic stem cells (hESCs) and in vitro and in vivo differentiated hESCs (red); bone marrow cells, myeloid cells, and lymphocytes (green); and adult tissues (blue). The dendrogram provides a measure for the relatedness of gene expression data between the samples. (B): Depicted are ⬃1700 measurements obtained from the 50 tissue samples that were hybridized to Affymetrix U133A DNA microarray. Expanded view of the gene cluster that was upregulated in leukocytes and all other tissues in comparison with hESCs and their derivatives. We applied hierarchical clustering both on the columns (tissues) and rows (genes) using the UPGMA algorithm, with average linkage and Euclidean distance function. Gene names are indicated on the right. Green represents transcript levels less than the average, black represents transcript levels equal to the average, and red represents transcript levels greater than the average. Color saturation represents the magnitude of the ratio relative to the average.

and their derivatives (Fig. 3B). This gene expression signature includes diverse genes that are known to be especially important for T cell activation, such as the intercellular adhesion molecules-2 and -3 [40], interleukin-1␤ [41], and CCL3 (MIP-1␣) [42]. This pattern of gene expression clearly favors the ignorance hypothesis: hESCs and their differentiated derivatives only rarely differentiate to cells of the immune system that harbor the potential to properly stimulate immune responsiveness of alloreactive T cells. This notion is also supported by the fact that the genes that were upregulated in hESCs compared with the other samples do not possess known immune suppressive capacities (data not shown).

Cytotoxic T Cells Have the Capacity to Recognize hESCs Although we show that immediate differentiated derivatives of hESCs have low potential to stimulate alloreactive T cells, it is

Figure 4. Lysis of human embryonic stem cells (hESCs) by influenzaA-matrix (IV/A) synthetic peptide GILGFVFTL-restricted CTL line. (A): The level of HLA-A2 antigen expression by the cell lines that were used in the killing assay was examined by HLA-A2-specific monoclonal antibody and evaluated by fluorescence-activated cell sorter. Cell types: major histocompatibility complex class I-negative B cell line (721.221) transfected to express the HLA-Cw3 and HLA-A2 antigens (negative and positive controls, respectively); naive and interferon (IFN)-␥-treated undifferentiated hESC line H9. Note the similar levels of HLA-A2 expression in IFN-␥-treated H9 hESCs and HLA-A2 721.221 transfectants. Broken line indicates background fluorescence. Lysis of IV peptide-loaded (B) and IV-infected (C) cells. Cell lines tested: 721.221/ HLA-A2 and 721.221/HLA-Cw3 transfectants; naive and IFN-␥-treated hESCs; naı¨ve (only B) and IFN-␥-treated primary teratoma (Tu) cell line. E:T ratios are plotted against percentage of lysis. Results presented represent one of the three experiments performed in A–C.

Drukker, Katchman, Katz et al. protein alone versus 721.221 cells that express the HLA-Cw3 protein [29]. Whereas the HLA-A2-expressing cells were efficiently killed by CTLs, HLA-Cw3-expressing cells were only minimally recognized (Fig. 4B). In this test, synthetic IV/A peptide-loaded undifferentiated hESCs were not efficiently recognized by CTLs. Furthermore, even if undifferentiated hESCs were grown for 2 days in the presence of IFN-␥ that induces MHC-I expression in these cells by 50-fold [23], killing efficiency was still very low. This is despite the fact that HLA-A2 levels on IFN-␥-treated hESCs and 721.221/A2 are similar (Fig. 4A). Similarly, when naive and IFN-␥-treated differentiated Tu cells were used as target cells in this assay, the cells were not killed by CTLs (Fig. 4B). Next, we tested whether IV infection of hESCs and Tudifferentiated cell line enhances CTL-mediated cytolysis. Infection of target cells was confirmed by human anti-IV sera (data not shown). As expected, the control cell line HLA-A2/721.221 was killed effectively, whereas IV-infected undifferentiated hESCs were not killed (Fig. 4C). However, when the cells were infected and treated by IFN-␥ simultaneously, effective killing response of at least 50% was noted (compared with the positive control). Surprisingly, this effect was less prominent when Tu cells were used. IV-infected IFN-␥-treated Tu cells show clear but low lysis sensitivity (Fig. 4C). Taken together, these results demonstrate that hESCs might be recognized by human CTLs after strong stimulation of MHC-I expression. Also, killing efficiency by activated CTLs suggests that a very efficient peptide-loading method, such as a viral infection, is required to carry out cytotoxic assays against hESCs.

DISCUSSION Since hESCs were first isolated, it has been widely accepted that these cells may hold the potential to change the face of medicine in light of their potential capacity to differentiate to every cell type in the human body. If so, efforts must be carried out to understand the possible immune attack that might be launched against the cells by the patients’ immune systems. In this study, we aim to determine the extent and possible mechanisms of immune response toward hESCs and their differentiated progeny in the setting of xenotransplantation and allotransplantation. Although it was recently reported that hESCs fail to elicit immune response in immunocompetent mice during the first 48 hours after intramuscular injection [44], we found that 1 month after transplantation, hESCs were totally eliminated in various strains of immunocompetent mice. In the experiments undertaken in mouse strains with different types of immune deficiency, T cell-deficient animals failed to reject hESC-derived graft, whereas the lack of NK cells or B cells did not interfere with vigorous hESC rejection. These findings clearly show the pivotal role of T cells in xenorejection of hESCs and their differentiated derivatives. Using the humanized (Trimera) mouse model that enables evaluation of the direct immune rejection pathway [45], we investigated here for the first time allorejection of hESCs in vivo. Transplantation of hESCs and their differentiated derivatives to the human PBMC-reconstituted mice revealed that the cells uniformly escaped rejection and developed normally into teratomas. Similarly, hESC-derived primary cell lines and EBs were not rejected in this model, in contrast to human adult skin grafts or a B cell line, which were heavily infiltrated by leukowww.StemCells.com

227

cytes or completely eliminated. These results indicate that hESCs and their differentiated progeny have reduced immunogenicity (capacity to evoke immune response) compared with adult tissues and do not elicit direct allorejection. A special emphasis should be given to the fact that teratomas contain multiple terminally differentiated cell types, as in conventional grafts. The resistance of hESCs to direct T cell-mediated allorejection may emanate from reduced potential to properly stimulate alloreactive T cells, from immune protective mechanisms, or from both. Clearly, hESCs and their derivatives fail to properly deliver the right stimulatory signals: they express only low levels of MHC-I molecules, MHC-II molecules are absent [23], and, as shown here, CD80 and CD86 proteins are not expressed. In contrast, control grafts that were used as a positive control contain MHC-II-expressing cells and are therefore likely to be rejected by initial activation of CD4⫹ T cells. Raji cells are MHC-II-expressing cells, and human skin grafts contain APCs that express MHC-II molecules, both of which can migrate into regional lymph nodes and stimulate alloresponse against the graft. These findings suggest that despite multilineage differentiation of hESCs, observed in teratomas, substantial development of MHC-II-expressing cells such as APCs does not occur in the grafts. This highlights the ability to transplant purified hESC-derived MHC-II-negative populations that have reduced capacity to activate immune response. As shown recently, donor endothelial cells in human grafts can function as APCs and/or targets for T-cell-mediated cytotoxicity and trigger acute rejection [46]. Yet, in this study we did not see any significant sign for rejection of hESC-derived teratomas, although it was shown recently that hESC-derived endothelial structures formed in teratomas [47]. Hence, it is reasonable to assume that hESC-derived vasculature does not considerably contribute to activation of allogeneic immune responses. It remains to be determined whether more mature hESC-derived endothelial cells will acquire APC functions after prolonged differentiation. Although FasL expression was suggested to be involved in protecting RESC-like populations from rejection by allogeneic hosts [26], this seems not to be the case for hESCs. As demonstrated in this study, FasL mRNA and its protein could not be detected in hESCs and EBs; thus, it seems that the protective capacity of FasL is species-specific and is absent in hESCs. It is more likely that the capacity to survive in an allogeneic environment is due to ignorance by T cells [37]. It was recently suggested that immunological maturity is a rather late event during the gestational period of human embryos [34]. Our DNA microarray data support this notion. The expression of approximately half of the immune-related genes that were expressed in hematopoietic cells, lymphoid organs, and other tissues was not upregulated during in vitro and in vivo differentiation of hESCs, indicating that using most of the currently available differentiation protocols, the cells do not reach immunological maturity. These data suggest that immunosuppressive regimes for hESC-based therapeutics could be highly reduced compared with conventional organ transplantation. Still, it is important to note that whether or not hematopoietic differentiation take place and T cell sensitization occurs, the MHC-I levels in hESCs are sufficient for T cell recognition, as shown by the CTL assays. Thus, any hematopoietic cells or

Human ES Cell Immunogenicity

228

progenitors should be eliminated before transplantation if they are therapeutically unnecessary. Also, it is possible that after transplantation the cells would further mature and start to express higher levels of transplantation antigens. To examine this possibility, the immunological properties of hESC-derived tissues should be evaluated in long-term studies. Finally, because cross priming of human T cells by mouse APCs is weak in the Trimera model [34], it is likely that rejection of hESCs and their differentiated progeny in immunocompetent mice is mediated indirectly upon triggering of mouse T cells by mouse APC-presenting human antigens originating from the implant. Dekel et al. [34] have recently demonstrated that early embryonic pig kidney precursor tissues from E28 in contrast to E42 tissues evade rejection in the Trimera model by human lymphocytes. The same tissues are rejected in normal immunocompetent mice, but this rejection can be overcome by

REFERENCES 1 Thomson JA, Itskovitz-Eldor J, Shapiro SS et al. Embryonic stem cell lines derived from human blastocysts. Science 1998;282:1145–1147. 2 Reubinoff BE, Pera MF, Fong CY et al. Embryonic stem cell lines from human blastocysts: Somatic differentiation in vitro. Nat Biotechnol 2000; 18:399 – 404. 3 Itskovitz-Eldor J, Schuldiner M, Karsenti D et al. Differentiation of human embryonic stem cells into embryoid bodies comprising the three embryonic germ layers. Mol Med 2000;6:88 –95. 4 Carpenter MK, Inokuma MS, Denham J et al. Enrichment of neurons and neural precursors from human embryonic stem cells. Exp Neurol 2001; 172:383–397. 5 Reubinoff BE, Itsykson P, Turetsky T et al. Neural progenitors from human embryonic stem cells. Nat Biotechnol 2001;19:1134 –1140. 6 Schuldiner M, Eiges R, Eden A et al. Induced neuronal differentiation of human embryonic stem cells. Brain Res 2001;913:201–205. 7 Zhang SC, Wernig M, Duncan ID et al. In vitro differentiation of transplantable neural precursors from human embryonic stem cells. Nat Biotechnol 2001;19:1129 –1133. 8 Schulz TC, Palmarini GM, Noggle SA et al. Directed neuronal differentiation of human embryonic stem cells. BMC Neurosci 2003;4:27. 9 Kehat I, Kenyagin-Karsenti D, Snir M et al. Human embryonic stem cells can differentiate into myocytes with structural and functional properties of cardiomyocytes. J Clin Invest 2001;108:407– 414. 10 Kehat I, Gepstein A, Spira A et al. High-resolution electrophysiological assessment of human embryonic stem cell-derived cardiomyocytes: A novel in vitro model for the study of conduction. Circ Res 2002;91:659 – 661. 11 Mummery C, Ward D, van den Brink CE et al. Cardiomyocyte differentiation of mouse and human embryonic stem cells. J Anat 2002;200: 233–242. 12 Xu C, Police S, Rao N et al. Characterization and enrichment of cardiomyocytes derived from human embryonic stem cells. Circ Res 2002; 91:501–508. 13 Mummery C, Ward-van Oostwaard D, Doevendans P et al. Differentiation of human embryonic stem cells to cardiomyocytes: Role of coculture with visceral endoderm-like cells. Circulation 2003;107:2733–2740. 14 He JQ, Ma Y, Lee Y et al. Human embryonic stem cells develop into multiple types of cardiac myocytes: Action potential characterization. Circ Res 2003;93:32–39. 15 Levenberg S, Golub JS, Amit M et al. Endothelial cells derived from human embryonic stem cells. Proc Natl Acad Sci U S A 2002;99:4391– 4396.

mild immunosuppression with costimulatory blockade agents such as CTLA4. The potential of such agents in overcoming rejection of hESCs and their differentiated derivatives should be further investigated in preclinical large-animal models.

ACKNOWLEDGMENTS M.D. and H.K. contributed equally to this work. We thank Professor Joseph Itskovitz-Eldor at the Rambam Medical Center for kindly providing us with the hESCs as collaboration. We thank Eran Maryuma for his assistance in the bioinformatic analysis. This research was partially supported by funds from the Herbert Cohn Chair and the Israel Science Foundation.

DISCLOSURES The authors indicate no potential conflicts of interest.

16 Kaufman DS, Hanson ET, Lewis RL et al. Hematopoietic colonyforming cells derived from human embryonic stem cells. Proc Natl Acad Sci U S A 2002;98:10716 –10721. 17 Chadwick K, Wang L, Li L et al. Cytokines and BMP-4 promote hematopoietic differentiation of human embryonic stem cells. Blood 2003;102:906 –915. 18 Green H, Easley K, Iuchi S. Marker succession during the development of keratinocytes from cultured human embryonic stem cells. Proc Natl Acad Sci U S A 2003;100:15625–15630. 19 Sottile V, Thomson A, McWhir J. In vitro osteogenic differentiation of human ES cells. Cloning Stem Cells 2003;5:149 –155. 20 Rambhatla L, Chiu CP, Kundu P et al. Generation of hepatocyte-like cells from human embryonic stem cells. Cell Transplant 2003;12:1–11. 21 Lavon N, Yanuka O, Benvenisty N. Differentiation and isolation of hepatic-like cells from human embryonic stem cells. Differentiation 2004;72:230 –238. 22 Jurisicova A, Casper RF, MacLusky NJ et al. HLA-G expression during preimplantation human embryo development. Proc Natl Acad Sci U S A 1996;93:161–165. 23 Drukker M, Katz G, Urbach A et al. Characterization of the expression of MHC proteins in human embryonic stem cells. Proc Natl Acad Sci U S A 2002;99:9864 –9869. 24 Rogers NJ, Lechler RI. Allorecognition. Am J Transplant 2001;1:97– 102. 25 Green DR, Ferguson TA. The role of Fas ligand in immune privilege. Nat Rev Mol Cell Biol 2001;2:917–924. 26 Fandrich F, Lin X, Chai GX et al. Preimplantation-stage stem cells induce long-term allogeneic graft acceptance without supplementary host conditioning. Nat Med 2002;8:171–178. 27 Segall H, Lubin I, Marcus H et al. Generation of primary antigen-specific human cytotoxic T lymphocytes in human/mouse radiation chimera. Blood 1996;88:721–730. 28 Schuldiner M, Yanuka O, Itskovitz-Eldor J et al. Effects of eight growth factors on the differentiation of cells derived from human embryonic stem cells. Proc Natl Acad Sci U S A 2000;97:11307–11312. 29 Cohen GB, Gandhi RT, Davis DM et al. The selective downregulation of class I major histocompatibility complex proteins by HIV-1 protects HIV-infected cells from NK cells. Immunity 1999;10:661– 671. 30 Sieg S, Smith D, Yildirim Z et al. Fas ligand deficiency in HIV disease. Proc Natl Acad Sci U S A 1997;94:5860 –5865. 31 Achdout H, Arnon TI, Markel G et al. Enhanced recognition of human NK receptors after influenza virus infection. J Immunol 2003;171:915– 923. 32 Smith KD, Mace BE, Valenzuela A et al. Probing HLA-B7 conformational shifts induced by peptide-binding groove mutations and bound

Drukker, Katchman, Katz et al.

229

peptide with anti-HLA monoclonal antibodies. J Immunol 1996;157: 2470 –2478.

ICAM-1 and ICAM-2 in immune responses. J Exp Med 1994;179:619 – 629.

33 Valk PJ, Verhaak RG, Beijen MA et al. Prognostically useful geneexpression profiles in acute myeloid leukemia. N Engl J Med 2004;350: 1617–1628.

41 Sims JE. IL-1 and IL-18 receptors, and their extended family. Curr Opin Immunol 1994;14:117–122.

34 Dekel B, Burakova T, Arditti FD et al. Human and porcine early kidney precursors as a new source for transplantation. Nat Med 2003;9:53– 60. 35 Lubin I, Segall H, Marcus H et al. Engraftment of human peripheral blood lymphocytes in normal strains of mice. Blood 1994;83:2368 – 2381. 36 Smith D, Sieg S, Kaplan D. Technical note: Aberrant detection of cell surface Fas ligand with anti-peptide antibodies. J Immunol 1998;160: 4159 – 4160. 37 Walker LS, Abbas AK. The enemy within: Keeping self-reactive T cells at bay in the periphery. Nat Rev Immunol 2002;2:11–19. 38 Sharpe AH, Freeman GJ. The B7-CD28 superfamily. Nat Rev Immunol 2002;2:116 –126. 39 Dvash T, Mayshar Y, Dar H et al. Temporal gene expression during differentiation of human embryonic stem cells and embryoid bodies. Hum Reprod 2004;19:2875–2883. 40 de Fougerolles AR, Qin X, Springer TA. Characterization of the function of intercellular adhesion molecule (ICAM)-3 and comparison with

42 Menten P, Wuyts A, Van Damme J. Macrophage inflammatory protein-1. Cytokine Growth Factor Rev 2002;13:455– 481. 43 Zhan X, Dravid G, Ye Z et al. Functional antigen-presenting leucocytes derived from human embryonic stem cells in vitro. Lancet 2004;364: 163–171. 44 Li L, Baroja ML, Majumdar A et al. Human embryonic stem cells possess immune-privileged properties. STEM CELLS 2004;22:448 – 456. 45 Dekel B, Bocher WO, Marcus H et al. Acute cellular rejection of human renal tissue by adoptive transfer of allogeneic human peripheral blood mononuclear cells into chimeric rats: Sequential gene expression of cytokines, chemokines and cytolytic effector molecules, and their regulation by CTLA-4-Ig. Int Immunol 1999;11:1673–1683. 46 Kreisel D, Krupnick AS, Gelman AE et al. Non-hematopoietic allograft cells directly activate CD8⫹ T cells and trigger acute rejection: An alternative mechanism of allorecognition. Nat Med 2002;8:233–239. 47 Gertow K, Wolbank S, Rozell B et al. Organized development from human embryonic stem cells after injection into immunodeficient mice. Stem Cells Dev 2004;13:421– 435.

See www.StemCells.com for supplemental material available online.

www.StemCells.com