Objective: The in vivo immunogenicity of Embryonic Stem Cells is controversial. At present, there is only in vitro evidence of MHC I expression by this cell population but vivid speculation about their immune-privileged state. The immunology aspect of ESC transplantation deserves thorough investigation. Methods: We injected mouse ESC (expressing Green Fluorescent Protein, GFP) into injured myocardium of syngeneic, allogeneic and SCID recipients. Furthermore, we monitored host response for up to 4 weeks post cell transfer. We determined local response (CD 3, CD 11c expression by host cells), MHC I expression by donor cells, MHC-II expression within and around the graft, humoral response of allogeneic hosts using Flow Cytometry and evaluated the hosts' cytokine response using stimulated spleenocytes by means of ELISPOT. Cell survival was estimated by morphometry, by calculating the area of the GFP+ graft over the area of infarction at multiple sections of the harvested heart. Results: There was significant cellular infiltration into and around the graft consisting of T-lymphocytes (CD3+) and dendritic cells (CD 11c). Infiltration was detectable at 1 week and progressed through 4 weeks following cell transplantation. The humoral Ab response was moderate at 2 weeks but frank at 4 weeks. ELISPOT demonstrated a Th1 pathway of donor specific T-lymphocyte response with strong IFN-γ and Il-2 production (figure A). MHC I expression was significant within the graft and maximal in the allogeneic groups. Conclusions: An immune response against transplanted ESC was demonstrated and the future use of ESC will likely require the use of systemic immunosuppression.
Embryonic stem cells are viewed as a promising source of donor grafts for the treatment of a variety of diseases [1–4]. Their pluripotency and their ability to divide for multiple generations are crucial advantages over other types of stem cells, whose plasticity is questioned. There is vivid controversy with regards to their immunogenicity, since preimplantation-stage embryonic stem cells (ESC) were believed to be immune-privileged by the recipient organism [5,6]. This has been attributed to the lack, or low expression of MHC I molecules on their surface. More recent studies have provided evidence that ESC bear the potential of MHC I expression in the petri dish, which can be enhanced by administration of IFN . However, only speculations exist pertaining to the in vivo immunogenicity of ESC and whether this is extensive enough to undermine the efficacy of ESC transplantation.
ESC are derived form the inner cell mass of blastocyst-stage embryos [8–10]. The status of MHC I expression in the inner cell mass has been previously examined by immunostaining with the W6/32 Ab and gave contradictory results. Drukker et al.  showed that ESC express MHC I proteins albeit in very low levels in vitro. Addition of IFN-γ to the growth media of differentiated human ESC resulted in high levels of MHC I protein expression. The authors only speculated about their potential of expressing MHC I and inducing allorecognition responses following transplantation. Another intriguing finding, which will be addressed in the present work, is that MHC II was not expressed in human ESC and their derivatives. On the other hand, HLA expression could be detected and it was therefore assumed by the same group that antigen processing and presentation are intact and functional in undifferentiated and differentiated human ESC. Similar to these findings, Bradley et al.  report a fourfold expression of HLA class II molecules in human ESC upon differentiation in vitro. Furthermore, these authors confirmed the expression of the receptor for IFN-γ on human ESC, which in turn upregulates the expression of MHC I in response to IFN-γ It seemed certain to the authors that human ESC-derived tissues will be induced to express HLA class II molecules after transplantation if exposed to inflammatory cytokines. Moreover, it is believed that human ESC-derived grafts do not contain donor interstitial dendritic cells (DC), unless human ESC have been induced to differentiate along the hematopoetic pathway. The absence of donor DCs from non-hematopoetic hESC-derived tissue has important implications. Experimental strategies that deplete allografts of donor DCs have been found, in some models, to delay graft rejection [12–14]. The work presented here is the first to address some of the above controversial issues, such as the in vivo, potentially progressive, expression of significant amounts of MHC I upon differentiation, the ESC commitment to a dendritic cell progeny and the host response to ESC transplantation for the treatment of ischemic myocardial injury. A better understanding of these in vivo events and their consequences for donor cell survival are of utmost importance for the success of embryonic stem cell transplantation that aims at a robust and sustained restoration of target organ function and structure.
Materials and methods
Cell preparation and labeling
pEF-1 a-EGFP (Green Fluorescent Protein), which contains EGFP gene under the control of human EF1 a promoter and a neomycin resistance cassette, was constructed as follows: The promoter region of pEGFP-N3 (Clontech, Palo Alto, CA) was removed by removing the Ase I-Nhe I DNA fragment and joining of blunt-ended termini. Human EF1a promoter from a pEF-BOS (Hind III-Eco RI DNA) fragment was inserted into the Hind II-Eco RI site of the plasmid. D3 ES cells were transfected with pEF-1 a-EGFP and one clone, brightly expressing EGFP, was chosen and used for the experiments. The clone was adapted to feeder-free conditions and Leucemia Inhibitory Factor (LiF) was added to halt differentiation of the Embryonic Stem Cells (ESC). On the day of harvest the cells were trypsinized and one hour later injected intramyocardially in our murine myocardial ischemia model.
Group I: syngeneic 129sv animals (n=5), group II: allogeneic BALB/c recipients (n=5), group III: SCID recipients, (n=5). These animals were sacrificed 2 weeks after cell transfer. In a second series of experiments, timeline studies were performed, i.e., the host response to allogeneic ESC was studied: Group A: BALB/c mice were sacrificed 1 week following cell injection; in Group B, recipients were sacrificed 2 weeks following cell injection, and in Group C, recipient mice were sacrificed 4 weeks following cell injection into the area of myocardial injury. The heart, serum and spleen of these series of animals were harvested for histology, FACS analysis and ELISPOT studies of cytokine production, respectively. All donor ESC were of sv129 origin, thus, allogeneic (major mismatch). All mice were purchased from the Jackson laboratories, Bar Harbor ME). All animals of every single group underwent LAD-ligation before cell injection.
All surgical interventions and animal care were provided in accordance with the Laboratory Animal Welfare Act, the Guide for the Care and Use of Laboratory Animals (National Institutes of Health, publication number 78–23, revised 1978) and the Guidelines and Policies for the Use of Laboratory Animals for Research and Teaching of the Department of Comparative Medicine, Stanford University School of Medicine. The recipient mice were pre-anaesthetised with isoflurane and received an intraperitoneal injection of Ketanest/Xylazine (50mg/kg). The animals were then incubated and ventilated for the entire length of the procedure. The surgical approach involved a left lateral thoracotomy, pericardiectomy, and identification of the left anterior descending artery (LAD) for ligation. Following ligation of the LAD, 250,000 donor cells in 20μl medium were injected into the resulting area of infarction.
Histology and immunohistochemistry
Cryosections (5μm) were stained with hematoxylin and eosin, Masson's trichrome, or used for immunohistochemistry. Immunostaining was performed as previously described. The antibodies used included rabbit anti-connexin-43, mouse monoclonal anti-alpha-actinin, mouse monoclonal anti-smooth muscle actin (Sigma, St. Louis, MO), hamster anti-CD-3ε, hamster anti-CD11c, mouse anti-MHC-I-Ab and anti MHC-II-Ab (BD Pharmingen, San Diego, CA), goat anti-GFP antibody (Rockland, Gilbertsville, PA), and rabbit anti-GFP Alexa-488 conjugated antibody (Molecular Probes, Eugene, OR). Stained tissue was examined with a Leica DMRB fluorescent microscope and a Zeiss LSM 510 two-photon confocal laser scanning microscope.
To identify quality and quantity of cellular (inflammatory/immune) response to the injected cells, we have stained the same sections for CD 3 which indicates presence of T-lymphocytes and CD 11c, which is being expressed by activated dendritic cells which are antigen presenting cells (APCs). To assess the presence of Major Histocompatibility Complex molecules on the GFP positive cells we stained for MHC I, which is expressed by all adult cells, and MHC II which is expressed by antigen presenting cells. We also report on the distribution of the MHC signal (Texas red conjugated secondary antibody, BD Pharmingen, San Diego, CA) and its quantity over time.
For all morphometric evaluations, the focused microscopic field was photographed by an adapted camera (Diagnostic Instruments Inc, Sterling Heights, MI). To quantify the degree of expression of specific markers, five random sections of the GFP positive graft were photographed and evaluated using the Spot advanced software, version 3.4.2 (Diagnostic Instruments Inc, Sterling Heights, MI).
Recipient sera were used to evaluate the levels of circulating anti-donor antibodies. Blood was obtained—at 1, 2 and 4 weeks following ECS injection, and centrifuged for 10min at 2000rpm. Plasma was collected and stored at −80°C. We performed thymectomy on 2-month-old BALB/c and 129sv mice. Single cell BALB/c and 129sv thymocyte suspensions were prepared. 1×106 cells were then reconstituted in flow cytometry PBS buffer containing 0.5% BSA and 0.1% azide. The individual sera of the recipient mice were used as a source of primary anti-donor antibody at 1:100 dilution. Donor and recipient thymocyte suspensions were incubated with sera dilutions for 1h on ice; after washing, cells were incubated with secondary goat anti-mouse IgG antibody (Beckton Dickinson, San Jose, CA) labeled with FITC for 30min in the dark at 4°C. Cells were washed twice, and resuspended in 1% paraformaldehyde, and analyzed using a FACScan flow cytometer (Facscalibur, Becton Dickinson, San Jose, CA); fluorescence signals were visualized on logarithmic scales.
All materials used for ELISPOT were purchased from BD Biosciences Pharmingen, San Diego, USA. ELISPOT was carried out in accordance with previously published protocols .
All results are expressed as mean±SD. Multilineage differences between groups were analyzed by two-way analysis of variance (ANOVA). If a significant F-ratio was obtained, a Bonferroni post hoc test was used to specify pair-wise differences. Statistical analyses were performed with StatView 5.0 (SAS Institute, Cary, NC), and significance was accepted at p≪0.05. Comparison between two groups at a time was carried out using t-test for independent variables and significance was assumed when p≪0.05.
Engraftment and restorative effect
At 2 weeks following injection the cells formed large grafts in all animals (Fig. 1A). The overall size of the grafts was reduced by 4 weeks compared to 2 weeks following injection (data not shown). Cell-free gaps occurred within the grafts over the observation interval. The expression of Connexin 43 (Fig. 1B) was highest in the allogeneic group: 40.3±12.4% (p=0.04) vs. 26.4±14.2 in the syngeneic and 24.5±8.7 in the group of SCID mice (Fig. 1C) and increased over time. The overall expression of α-sarcomeric actin (Fig. 1D) ranged from 15 to 45% of the GFP graft area (33.5±11.2% in the allogeneic, 29.1±6.3% in the syngeneic and 27±5.8% in the group of SCID mice, Fig. 1E). The α-sarcomeric actin expression increased over time [3.94±1.12% at 1 week, 11.99±3.22% at 2 weeks (p=0.002), and 23.65±4.75 at 4 weeks (p=0.001) post cell injection]. Only a fraction of the cells which express Connexin 43 and α-sarcomeric actin ultimately adopts myofibrillar phenotype. Thus, the overall presence of the notorious spindle-like morphology is present in less than 20% of the cells. The majority of the cells show fibrocyte, chondrocyte, endothelial and adenocyte morphology. At 4 weeks, cellular dysplasia and nuclear polymorphism were evident, even though teratomas were not observed yet. Functional and oncological aspects of ESC transplantation are focus of parallel work in our laboratory.
Immune response to ESC in allogeneic hosts
Immunohistochemistry showed that CD3+ cells accumulated around the margins of allogeneic grafts in clusters (Fig. 2A). Their population was comparable at the first and second week following injection and increased more than twofold after 4 weeks (Fig. 2B). The numbers of CD3+T cells were highest in the allogeneic group and absent in the SCID group, as expected. Only a few CD3+ cells were present in the syngeneic group. Most of the CD 11c positive cells were dispersed within the graft without particular relationships to the margins (Fig. 2C). The presence of CD 11c positive cells displayed a linear progressive increase from week 1 to week 4 after cell transfer, (p=0.03, Fig. 2D). Their population was highest in the allogeneic group, negligible in the syngeneic group and absent in the SCID group. MHC class I expression could be ascribed to differentiated clusters of cells in corresponding sections, reaching the highest level at 2 weeks and followed by a decrease at 4 weeks after cell transfer (Fig. 2E). [40.96±10.12% of the grafted cells at 1w, 59.71±13.26% at 2w (p=0.044) and 28.84±7.56% at 4w (p=0.022) following cell injection]. MHC II expression was exclusive to the allogeneic group (36.14±12.27% of the grafted area). MHC class II was expressed by both donor and recipient cells concurrently (Fig. 2F). FACS analysis of circulating donor-specific antibodies in allogeneic recipients showed no B cell alloimmune response after 1 week in BALB/c animals transplanted with 129sv-derived ESC (Fig. 3A). However, alloantibody levels significantly increased after 2 weeks, reaching a maximum at 4 weeks after cell injection [naïve animals: 1.51±0.88%, 2.58±1.74% at 1w, 14.57±3.85% at 2w (p=0.015), 48.2±9.49% at 4w (p=0.003), Fig. 3A]. ELISPOT analysis demonstrated low frequency of IFN-γ-producing T-lymphocytes reactive to donor ESC at 1 week following cell implantation. However, at 4 weeks, the frequency of IFN-γ producing T-lymphocytes increased when compared to naïve animals and earlier post-transfer time points (Fig. 3B). Similarly, numbers of IL-2 producing T cells expanded dramatically 4 weeks following transplantation. (Fig. 3C). Interestingly, correlating with progressive accumulation of alloantibody in the sera of allogeneic recipients at 4 weeks, we observed expansion of IL 4-producing alloreactive T cells (293±88 spots compared to 54±7 spots in naïve recipients). IL-4 is a key Th2 cytokine involved in antibody production by B lymphocytes.
Allogeneic, preimplantation-stage ESC trigger a significant and progressive immune response, which might be a causative factor of cell loss following cell transplantation. The host response towards the donor cells is identifiable at three different levels. Locally, by accumulation of cells of the immune system, humorally, by increasing amounts of donor-specific antibodies, and functionally, by production of increasing amounts of cytokines by activated spleenocytes of the host organism [16,17]. These processes are progressive, i.e. the findings are increasing over time following transplantation of the ESC into the area of ischemic injury of the recipient heart. This progression of host response to the injected ESC coincides with increasing expression of MHC I and increasing expression of markers of ESC differentiation toward the organ-specific progeny of cells. Hence, we have reason to believe that ESC become increasingly immunogenic in the process of differentiation in vivo, and the theories that, first, ESC are capable of only minor expression of MHC I and, second, that ESC escape the hosts' immune response can be refuted. The impact of host immune response against the donor ESC is visible in form of reduction of graft size over time. This reduction is the highest in the syngeneic group, suggesting that ensuing postoperative cell death is largely due to antigen-independent, innate immune responses or under-perfusion in the area of lesion. However, a considerable portion of the observed cell death can well be attributed to ischemic cell death due to severe undersupply in the area of injury. Latter may lead to accumulation of free radicals, cytotoxic products of the purine metabolism and the unfavorable microsctructural changes in the process of remodeling of heart muscle. The present study was not designed to discriminate between the various modalities of cell death, and it is noteworthy that ischemic cell death, apoptosis and immune-mediated cell loss share common pathways at the histological level.
According to our findings, the rise of a hematopoetic progeny, which carries MHC II markers on its cellular surface cannot be excluded. Donor cells expressed MHC II molecules as well, as shown by colocalization of GFP and Texas-red-coded MHC II. MHC II positive donor cells were found only in allogeneic recipient mice. It may therefore be possible that the microenvironmental niche of the injected cells exercises selective pressure for the undifferentiated cells to commit to the phenotype of antigen presenting cells, most probably driven by co-stimulation of infiltrating NK cells and other immune-effector cells of the host. This speculation is supported by findings of Pan et al. , who showed expansion of stimulatory NK cells by Interferon-γ. Interferon-γ was the prominently produced cytokine by activated recipient splenocytes and increased in correlation to the increase of humoral response. The host dendritic cells at the site of injection (CD-11 positive) assume reporter function within the recipients' spleen following allorecognition at the site of injection, and mediate lymphocytic infiltration in the area of ESC engraftment. The somewhat delayed CD 3 peak within the donor graft (infiltrating T-lymphocytes) is in compliance with this mechanism.
The presence of MHC I expression in the graft area and the demonstrated allogeneic immune response [19,20] poses a clinical challenge with regards to transplantation of ES cells or cardiac progenitors derived from them. These data suggest the inflammatory/immune barriers to successful ESC transplantation are indeed similar to those for tissue allografts [16,21]. It is noteworthy that we observed colocalization of CD11c and GFP signals (Fig. 2C) within the allografts. This shows that mouse ESC allografts give rise to donor interstitial dendritic cells, which are known to be potent stimulants of alloresponse.
These results refute the theory of non-immunogenicity of ESC in vivo. Future experimentation with ESC and their derivatives should therefore consider approaches aimed at suppression of these immune responses. Whether this would be achieved with state-of-the-art immunosuppressive strategies can only be speculated upon at the present time. In that case, adverse effects of immunosuppressive drugs may be encountered. These might include an increased infection risk, organ malfunction, decreased seizure threshold, neuro- and nephrotoxicity, hirsutism, and others. It is however possible that short term immunosuppression will be just enough to retain cell lives as long as it is necessary to prevent propagation of remodeling in the early, critical post-injury phase. In addition to classical immunosuppression protocols, it is conceivable that the efficiency of transplantable hematopoetic stem cells (HSC) from ES cells will improve. It has been demonstrated elsewhere that cotransplantation of HSC and hearts from the same donor into conditioned hosts results in lifelong transplantation tolerance without post-transplant immunosuppression . In order to eliminate tumorigenicity from ES cells, committed populations of cardiomyocyte progenitors should be transplanted. It is encouraging that high yield harvesting and purification of immature cardiomyocytes from ESC cultures have been obtained recently . For the reasons above and various, locally diverse bioethical considerations, ESC research has not progressed so far towards to bedside, as did autologous stem cell transfer. Provided the 100% purified population of non-tumorigenic, hESC-derived, committed cell population will be obtained, a powerful tool in organ restoration will be at hand.
To the favor of ESC, it is of note, that many of the currently ongoing clinical studies which involve autologous adult stem cells (some have gone far into clinical application) are poorly designed. They are not double-blind, prospectively randomized, and often they are not placebo-controlled. We view these results with skepticism. Ourselves have published negative reports on myocardial restoration using bone marrow stem cells, which lack transdifferentiation potential . Furthermore these cells are not robust enough and die rapidly following intramyocardial injection (75% of the cells dead after 48h). They do not assume myofibrillar phenotypes . When questioned about their mechanism of myocardial repair, one is frequently confronted with the all-to-trivial argument of reactive angiogenesis. The survival and robustness of ESC grafts is unique, according to our data. Parallel studies from our lab demonstrate a more significant restorative effect on myocardial architecture and function compared to studies published on adult stem cells (data in press).
Finally, the possibility of producing human pluripotent stem cell lines from predefined donors by nuclear transfer technology could provide almost histocompatible populations of cardiac progenitors and tolerance-inducing HSC.
Appendix A. Conference discussion
Mr M. Yacoub (Middlesex, United Kingdom): I just want to congratulate you on tackling such an important and difficult question. I have several questions. The first is, were these cells cultured on feeder cells, which are xenogenic, and would interfere with your results? And the second is, did you try and differentiate them or you put them as embryonic stem cells? And the third is you looked at only MHC Class I. What about Class II coreceptors which can modify your immune response in a major way, I mean, C D40,B7, et cetera? But were they grown in feeder cells?
Dr Kofidis: The mouse embryonic stem cells were grown in feeder-free free media as did the human embryonic stem cells.
Mr Yacoub: The human ones, yes.
Dr Kofidis: Human and the mouse. We did both options with the mice and then turned to use feeder-free cells.
Mr Yacoub: But were they on feeder cells, the human embryonic stem cells?
Dr Kofidis: Feeder-free.
Mr Yacoub: Feeder-free.
Dr Kofidis: Yes. I've been provided with these cells by Geron, which generally works with feeder-free cells.
We did not try to differentiate the mouse embryonic stem cells. At Bobby Robbins' lab back in Stanford, a transgenic model of connexin 43 overexpression has been designed to try to make them commit to the cardiac phenotype.
Now, with regards to the last leg of your question, we did not check for MHC II in a regular manner. But we found out within the same conglomerate that there is a couple of cells that do express MHC II. And these cells, particularly, and that is a very interesting fact to mention at this point, are also green which is donor cells. The immunologist in our department, Eugenia Fedoseyeva, is now trying to identify if these cells possibly commit to the hematopoietic phenotype within the host issue. I think this is a very important point.
Mr Yacoub: How was antigen presentation then if you think this was only MHC Class I, by the indirect pathway?
Dr Kofidis: I would content myself in presenting the preliminary data we obtained on MHC I; but again, I'm sure that a small portion of these cells might turn to the hematopoietic phenotype which would express MHC II as well. But we did not go further into the detail.
- embryonic stem cells
- immune response
- flow cytometry
- cell survival
- cell transplantation
- dendritic cells
- precipitating factors
- severe combined immunodeficiency
- tissue transplants
- therapeutic immunosuppression
- natural immunosuppression
- immunity, humoral
- enzyme linked immunospot assay
- transfer technique
- host (organism)
- european society of cardiology