|Home | About | Journals | Submit | Contact Us | Français|
Donor–recipient cell interactions are essential for functional engraftment after nonautologous cell transplantation. During this process, transplant engraftment is characterized and defined by interactions between transplanted cells with local and recruited inflammatory cells. The outcome of these interactions determines donor cell fate. Here, we provide evidence that lineage-committed embryonic stem cell (ESC)-derived vascular progenitor cells are the target of major histocompatibility complex (MHC) class I-dependent, natural killer (NK) cell-mediated elimination in vitro and in vivo. Treatment with interferon γ was found to significantly upregulate MHC class I expression on ESC-derived vascular progenitor cells, rendering them less susceptible to syngeneic NK cell-mediated killing in vitro and enhancing their survival and differentiation potential in vivo. Furthermore, in vivo ablation of NK cells led to enhanced progenitor cell survival after transplantation into a syngeneic murine ischemic hindlimb model, providing additional evidence that NK cells mediate ESC-derived progenitor cell transplant rejection. These data highlight the importance of recipient immune–donor cell interactions, and indicate a functional role for MHC-I antigen expression during successful ESC-derived syngeneic transplant engraftment.
Presently, one of the principal scientific barriers to the potential use of embryonic stem cells (ESCs) as a cell-based therapy for ischemic, cardiovascular, and other diseases is the very poor survival of donor cells after transplantation . Although many aspects of ESC graft survival remain to be fully understood, key factors thought to influence this process include poor cell viability after ex vivo expansion and preparation, an ischemic recipient milieu (as in the current study), a lack of appropriate trophic signals, and immune-mediated destruction .
Several reports have suggested an important role for major histocompatibility complex class I (MHC-I) and II (MHC-II) antigen expression in ESCs and their progeny. MHC-I, MHC-II, and nonclassical antigens are undetectable on murine embryonic stem cells and their early differentiated progenies despite the presence of their mRNA transcripts . The absence of this antigenic expression was suggested to play a role in the early evasion of immune destruction of transplanted allogeneic ESCs by natural killer (NK) and cytotoxic T-cells . However, it is known that cells with low MHC-I expression may be susceptible to NK cell-mediated rejection . Thus, it is perhaps surprising that ESCs devoid of MHC-I and -II expression were reported to be resistant to NK cell attack in vitro and in vivo [3, 5, 6]. Indeed, in vitro, it has paradoxically been suggested that upregulation of MHC-I expression by ESCs may heighten their susceptibility to cytotoxic T-cell killing . Adding further complexity at the systemic level, it would appear that by poorly understood mechanisms, transplanted ESCs may regulate host T-cell cytotoxicity and survival.
In this study, we sought to investigate the role of MHC-I expression by donor ESC-derived progenitor cells after syngeneic transplantation with respect to their susceptibility to host NK cell attack, survival, and subsequent physiologic engraftment. ESCs were differentiated into Brachyury+Flk-1+ (Bry+Flk-1+) cells (hemangioblasts)  and VE-Cadherin+ (VE-CAD+) endothelial progenitor cells. ESC-derived vascular progenitor cells, initially negative for MHC-I expression, were then selectively treated with interferon (IFN)γ to induce MHC-I expression. Two in vivo syngeneic models were used to evaluate the interaction between ESC-derived vascular progenitor cells and the host: a Matrigel plug transplantation model and a syngeneic mouse ischemic hindlimb model. These studies revealed that ESC-derived donor cell MHC-I antigen expression plays a critical role in dictating graft survival following transplantation.
ESCs were cultured and maintained in N2B27 media (Invitrogen, Carlsbad, CA, http://www.invitrogen.com) supplemented with 10 ng/ml Bone morphogenetic protein 4 (BMP-4) (R&D Systems, Minneapolis, MN, http://www.rndsystems.com) and 10 ng/ml leukemia inhibitory factor (Sigma, St. Louis, MO, http://www.sigmaaldrich.com/sigma-aldrich/home.html) according to the protocols of Ying and Nichols [8, 9] with minor modification. Mesodermal differentiation was induced by culturing ESCs as embryoid bodies at a cellular density of 2,500/cm2 in N2B27 media (0.3 ml/cm2) with 5 ng/ml BMP-4 and 10 ng/ml basic fibroblast growth factor (bFGF) (R&D Systems) applied for 3.25 days in suspension dishes. Bry+Flk-1+ double positive cells (hemangioblasts) were then isolated by fluorescence-activated cell sorting (FACS) sorting. For endothelial progenitor (VE-CAD+) cell differentiation, Bry+Flk-1+ cells were returned to culture for an additional 7 days in N2B27 media (0.3 ml/cm2) supplemented with 10 ng/ml vascular endothelial growth factor (VEGF) (R&D Systems) at a density of 20,000/cm2 on collagen IV-coated plates. MHC-I expression was induced by application of 10 −4 M IFNγ (R&D Systems) to the differentiation media for 2.5 days.
FACS was performed on a BD FACSCanto using Diva Software (BD, Franklin Lakes, NJ). Automatic compensation was performed based on single stained samples for the respective antibodies. Cells were isolated, washed, and then incubated with the following primary antibodies for 60 minutes: Flk-1 (#14-5821, eBioscience, San Diego, CA, http://www.ebio-science.com/), VE-CAD (#555289, BD Pharmingen, San Jose, CA, http://www.bdbiosciences.com/index_us.shtml), PE-conjugated anti-MHC-II antibody (#130-091-368, Miltenyi Biotec, Auburn CA), PE-conjugated rat anti-mouse inter-cellular adhesion molecule 1 (ICAM) (ab24869, abcam), PE-conjugated rat anti-mouse vascular cell adhesion molecule-1 (VCAM-1) (ab24853, abcam), and Biotin-conjugated H-2Db (#553572, BD Pharmingen). Cells were then washed and counter-stained by 60-minutes incubation with the following secondary antibodies as appropriate: donkey anti-rat IgG-Allophycocyanin (#712-136-153, Jackson ImmunoResearch, West Grove, PA, http://www.jacksonimmuno.com) and Streptavidin-Phycoerythrin (#554061, BD Pharmingen). Matching iso-type control antibodies were also used as appropriate.
Culture plates (24 well) were coated with Matrigel Matrix Basement Membrane (#356234, BD Biosciences, San Jose, CA) and incubated at room temperature for 2 hours. Cells in normal EBM-2 (#CC-3156, Lonza, Allendale, NJ, http://www.lonza.com/group/en.html) supplemented with EGM-2 SingleQuots (#CC-4176, Lonza) were plated on Matrigel at a density of 2 × 104/cm2 and cultured in regular cell culture incubator. Vascular tube formation was analyzed after 24 hours.
Matrigel plug assays were performed as previously published [10–12] with minor modification. Briefly, 8- to 12-week-old 129/ola female mice (Harlan, Indianapolis, IN, http://www.harlan.com/) were used as hosts for the Matrigel plug implants. ESC-derived progenitors were harvested, washed twice, and resuspended in Dulbecco's modified Eagle's medium (DMEM) basal medium at a concentration of 5 × 104 cells/μl. A 10 μl aliquot of cells was mixed with 0.5 ml of Matrigel on ice, and the liquid mixture was injected into the inguinal region of the recipient animals and allowed to gel at body temperature. Matrigel plugs were harvested and analyzed at days 7 or 14 after transplantation. A 200 μl volume of 50 mg/ml high-molecular weight Fluorescein isothiocyanate-dextran (125,000 molecular weight, Sigma) was injected intravenously 10 minutes before mouse euthanasia. Matrigel plug blood vessels were observed and photographed using a Leica MZ FL III Fluorescence Stereomicroscope (Leica Microsystems GmbH, Wetzlar, Germany, http://www.leica.com/). Matrigel plugs were either snap-frozen in liquid nitrogen and maintained at −80° C for subsequent immunohistochemistry staining or were used for quantification of FITC-dextran. Quantification of FITC-dextran in Matrigel plugs was achieved by incubating the plugs in 1 ml Dispase reagent (Invitrogen) for 16 hours at 37° C. The resulting mixture was centrifuged in a microfuge at 13,000 rpm for 30 seconds. The fluorescence of the resulting supernatants (excitation, 480 nm; emission, 520 nm) was measured using a fluorimeter (Spectra-Max GEMINI EM Dual-Scanning Microplate Spectrofluorometer, Sunnyvale, CA, http://www.moleculardevices.com) and quantitated against a standard curve for FITC-dextran (0.2– 25.6 μg/ml). Each test group used five Matrigel plug samples, and each sample used triplicate wells for analysis. The average of the triplicates was imputed for statistical analysis.
Eight- to ten-week-old 129/ola (Harlan) or non-obese diabetic (NOD) severe combined immunodeficiency (SCID) gamma (NSG) mice (Jackson Laboratory, Bar Harbor, ME, http://www.jax.org) underwent right femoral artery ligation as previously described, and served as the recipients for our cell transplantation experiments . These experiments were conducted according to the guidelines of the Animal Care and Use Committee of the National Heart, Lung, and Blood Institute and the guidelines for the Care and Use of Laboratory Animals published by the US National Institutes of Health (NIH Publication No. 85-23, revised 1996). Doppler scanning was used to monitor the ischemic status of the leg. Depletion of NK cells with AsGM1 antibody was performed as previously described [14, 15]. In brief, 50 μl of AsGM1 antibody (Cedarlane, Burlington, NC, http://www.cedarlanelabs.com/) was administered on days 0 (immediately prior to cell transplantation), 3, 6, and 9 after cell transplantation by tail vein injection (total dose 200 μg). Cell transplantation was performed by directly injecting 200,000 cells in 15 μl phosphate buffered saline (PBS) into the right tibialis muscle 24 hours after arterial ligation. Tissue harvest was performed at weeks 2 and 4 after cell transplantation. All experimental animal groups were controlled for sex mismatch grafting.
Fresh frozen slides were thawed, air-dried, and fixed by incubating in 4% paraformaldehyde for 10 minutes. Slides were then washed twice (5 minutes each wash) in × 1 PBS, and blocking solution (μ 1 PBS with 5% donkey serum and 0.1% Triton X-100) was applied for 1 hour. Blocking solution was then discarded and primary antibodies were applied (suspended in fresh blocking solution at appropriate dilutions) for 1 hour. Slides were then washed twice and secondary antibodies applied at 1:1,000 dilution (suspended in fresh blocking solution) for 1 hour in the dark. Depending on the specific antibody combinations, sequential staining with different primary and/or secondary antibodies was performed as required. Finally, slides were washed twice and mounting media containing 4′,6-diamidino-2-phenylindole (Vector Laboratories, Burlingame, CA, http://www.vectorlabs.com) was applied, followed by a cover slip. Anti-mouse primary antibodies were used against: Oct4 (SC-8628, Santa Cruz Biotechnology, Santa Cruz, CA, http://www.scbt.com), Nanog (ab5731, Chemicon, Charlottesville, VA, http://www.millipore.com/), CD3 (A0452, Dako, Fort Collins, CO, http://www.dako.com/), F4/80 (MCA497GA, Serotec, Raleigh, NC, http://www.abdserotec.com/), CD31 (#550274, BD Pharmingen), VE-CAD (#550548, BD Pharmingen); von-Willebrand Factor (vWF; A0082, Dako; sc-8068, Santa Cruz http://www.scbt.com/), α-smooth muscle actin (α-SMA; C6198 and F3777, Sigma), DsRed polyclonal antibody (#632496 Clontech, Mountain View, CA, http://www.clontech.com/; #sc-33353, Santa Cruz), smooth muscle myosin heavy chain (sc-79079 and sc-98705, Santa Cruz), and NKp46/NCR1 (AF2225, R&D Systems). Secondary antibodies were: donkey anti-rat IgG-Fluorescein (#712-136-153 and #712-096-153, Jackson ImmunoResearch), donkey anti-goat IgG-Fluorescein (#705-096-147, Jackson ImmunoResearch, http://www.jacksonimmuno.com/), and goat anti-rat IgG-Alexa Fluor 488: (A11006, Invitrogen). Images were acquired using either a Zeiss Axioskop plus light microscope with AxioVision V4.3 software or confocal microscopy, a Zeiss LSM 510 UV laser scanning confocal microscope system (Carl Zeiss GmbH, Germany, http://www.zeiss.com/). Cell counting was performed by a blinded member of our laboratory with extensive experience in both cell morphology and the techniques described earlier.
For α-SMA/Y-chromosome fluorescent in situ hybridization (FISH) costaining, fresh frozen slides were fixed in methanol:acetone (50:50) at −20° C for 10 minutes, washed in 1 × 1 PBS, and then blocked in 20% goat serum for 1 hour. Slides were then incubated in α-SMA Cy3- or FITC-conjugated antibody (C6198 and F3777, respectively; Sigma) at 1:100 dilution for 1 hour, washed, and then costained for Y-chromosome (see below). Alternatively, for CD31/Y-chromsome FISH costaining, following initial fixation (as aforementioned), slides were blocked in endogenous peroxidase with 1% H2O2 in methanol for 30 minutes, washed, blocked in 20% goat serum for 1 hour, washed again and blocked in avidin/biotin (#SP-2001, Vector Laboratories) for 15 minutes. Slides were then washed, incubated with biotin-rat anti-CD31 (ab24960, Abcam, Cambridge, MA) at 1:3,000 dilution for 1 hour, washed again, and incubated with Streptavidin HRP (#NEL750001EA, PerkinElmer, Waltham, MA, http://www.perkinelmer.com/) at 1:1,000 dilution for 1 hour. Following further washing with PBS and then 0.1 M Tris (pH 8.0), slides were incubated with tyramid amplification solution (FP1013, PerkinElmer) for 10 minutes (fluorescein at 1:330 dilution), washed, and then kept in × 1 PBS at 4° C overnight pending Y-chromsome FISH costaining. Y-chromosome FISH staining was performed according to the method of Mezey et al. [16, 17]. In brief, following staining for α-SMA or CD31, in situ hybridization was carried out using a digoxigenin-labeled riboprobe complementary to the satellite region of the mouse Y-chromosome, followed by staining with a Cya-nine 3/Cyanine 5 System (NEL752001KT, PerkinElmer). The riboprobe was constructed using an in vitro transcription kit (PerkinElmer) using template mouse genomic DNA kindly provided by Dr. Eve Mezey.
NK cells were isolated from splenocytes by magnetic bead negative depletion using a cocktail of biotin-conjugated antibodies against CD4, CD5, CD8a, CD19, Ly-6G, and Ter-119 (MACS NK Cell Isolation Kit, #130-090-864, Miltenyi Biotec, http://www.miltenyibiotec.com/). NK cells were then cultured in DMEM media (Invitrogen) supplemented with 10% fetal calf serum (FCS) (HyClone, Logan, UT, http://www.hyclone.com) and 500 U/ml recombinant human interleukin-2 (Roche, Nutley, NJ, http://www.roche.com/index.htm) for 4–7 days. Purity was typically >85% (by FACS using CD3− /DX5+ [#553067 and #558295 respectively, BD Pharmingen]). 51Chromium (Amersham Biosciences Inc., Piscataway, NJ, http://www.amersham.com) cytotoxicity assays were performed by plating 10,000 51Chromium-labeled target (T) cells per well (96-well plate) with varying ratios of effector (E) NK cells. After 4–16 hours incubation, supernatants were harvested onto Luma plates (PerkinElmer) and analyzed using a MicroBeta scintillation counter (PerkinElmer). Target cells analyzed with this assay were: VE-CAD+MHC-I+, VE-CAD+MHC-I− , H-2Db− YAC cells (ATCC, Manassas, VA, http://www.atcc.org) as a MHC-I− control population, and H-2Db+ Con A cells as a MHC-I+ control. H-2Db+ Con A cells were generated by culturing splenocytes in DMEM with 5 μg/mL concanavalin A (Sigma) and 10% FCS for 4–7 days.
Experimental data were analyzed by unpaired 2-tailed t-test or Mann Whitney test. Results are expressed as mean ± SEM. Differences were deemed significant when p < .05. Statistical analyses were performed using Prism, Version 4.00 (GraphPad Software, LA Jolla, CA, http://www.graphpad.com/welcome.htm). All FACS plots, histology, and immuno-staining images are representative of typical results.
For Additional “Materials and Methods” See Supporting Information Materials and Methods.
As early evasion of immune detection by transplanted ESCs has been reported to be reliant on an absence of MHC-I expression , we speculated that similar immune mechanisms may govern the fate of ESC-derived VE-CAD+ endothelial progenitors in our in vivo syngeneic models. To investigate this possibility, ESCs were cultured using defined serum-free media [7, 9, 18]. FACS characterization of undifferentiated ESCs revealed negligible expression of MHC-I, Bracyhury, Flk-1, or VE-CAD (Fig. 1A, and data not shown).
A two-step culture-differentiation process was used to derive Bry+Flk-1+ cells (hemangioblasts)  and VE-CAD+ endothelial progenitor cells as previously described [7, 19, 20]. By FACS, we observed that following this differentiation period Bry+Flk-1+ cells represented 27.8% ± 2.5% of all cells in culture (n = 7; Fig. 1Aa). To derive endothelial progenitor cells, Bry+Flk-1+ cells were isolated by FACS-sorting and returned to culture for further 7 days and with supplemental VEGF. Following this, we observed that VE-CAD+ cells constituted 30.9% ± 2.9% of all cells in culture (n = 7) (Fig. 1Ac). Immunohistochemistry revealed that in addition to being VE-CAD+ positive, these cells also expressed both CD31 (Fig. 1B) and vWF (data not shown). All cell populations remained MHC-I negative throughout differentiation (Fig. 1Ac).
MHC-I expression was induced using IFNγ during our culture-differentiation process. As anticipated, IFNγ treatment significantly increased MHC-I expression in both Bry+Flk-1+ (0.2% ± 0.17% to 41.3% ± 4.89%, both n = 7, p < .0001) (Fig. 1Af) and VE-CAD+ cells (0.18% ± 0.09% to 87.3% ± 5.24%, both n = 7, p < .0001) (Fig. 1Ad). IFNγ treatment did not induce MHC-II expression in ESC-derived vascular progenitor cell populations (Supporting Information, Fig. 1). In addition, IFNγ treatment and induced MHC-I expression did not interfere with cell differentiation, with IFNγ-treated Bry+Flk-1+ cells representing 25.8% ± 7.1% of all cells in culture (P = n.s., n = 7 compared with no IFNγ treatment) (Fig. 1Ae). VE-CAD+ cells constituted 26.8% ± 8.3% of all cells in culture (P = n.s., n = 7 compared with no IFNγ treatment) (Fig. 1Ad). Treatment with IFNγ did not change CD31 and VCAM-1 expression. However, ICAM-1 expression increased in the presence of IFNγ as compared with no IFNγ treatment (42.35% ± 5.43% and 5.50% ± 0.85%, respectively, n = 5; Supporting Information, Figs. 2 and 3).
Given that IFNγ may increase susceptibility to apoptosis , we verified that this treatment did not alter progenitor cell survival, proliferation, or differentiation. Bry+Flk-1+MHC-I+ cells were returned to culture for 7 days according to our endothelial progenitor differentiation protocol. These cells survived, proliferated, and exhibited unremarkable differentiation characteristics, as shown by FACS (Fig. 1Ah and data not shown) and immunohistochemistry staining (Fig. 1C and data not shown). Both VE-CAD+MHC-I+ and VE-CAD+MHC-I− cells were subjected to a Matrigel angiogenesis assay, revealing identical tube-forming ability and thereby further confirming that IFNγ treatment does not affect the endothelial properties of ESC-derived VE-CAD+ cells (Supporting Information, Fig. 4).
VE-CAD+MHC-I+ and VE-CAD+MHC-I− cells, mixed with liquid Matrigel, were injected subcutaneously into syngeneic recipient mice (129/ola) to form Matrigel plugs. Fluorescence stereomicroscopic analysis at day 14 postimplantation revealed significantly more vessel formation in VE-CAD+MHC-I+ cell Matrigel plugs than in VE-CAD+MHC-I− cell plugs (Fig. 2A and 2B). Fluorescence quantitation revealed significantly higher fluorescence in VE-CAD+MHC-I+ cell plugs than in VE-CAD+MHC-I− cell plugs (Fig. 2C, ***, p < .001, five plugs per test group). Immunostaining for DsRed/CD31 and DsRed/α-SMA on plugs derived from the VE-CAD+MHC-I+ cell group demonstrated that ESC-derived VE-CAD+ cells had functionally integrated into the new blood vessels (Fig. 2D and 2E). As a whole, these results suggest that ESC-derived VE-CAD+ cells are able to promote and functionally participate in neovascularization in a Matrigel plug model, and highlight the importance of MHC-I expression in this process.
To further elucidate the function of MHC-I and the mechanisms of interaction between transplanted cells and their syngeneic host (129/ola mice), we proceeded to investigate the fate of ESC-derived vascular progenitors with and without MHC-I expression in a syngeneic ischemic model. In control and PBS-injected mice, our ischemic murine hindlimb model was characterized by an initial inflammatory response in the ischemic musculature, in association with a degree of tissue degeneration/necrosis. Surprisingly, 2 weeks after femoral artery ligation, transplantation of ESC-derived VE-CAD+MHC-I− cells was associated with markedly increased tissue degeneration in tibialis muscle, while transplantation of VE-CAD+MHC-I+ cells was associated with only a minor increase in tissue degeneration when compared with PBS-injected control hindlimbs (Fig. 3A–3D). Consistent with this, compared with PBS-injected controls, immunohistochemistry revealed increased macrophage and T-cell infiltration in the VE-CAD+MHC-I− cell-injected tibialis muscle (Fig. 3G and 3K), with significantly less inflammatory cell infiltration observed in muscle injected with VE-CAD+MHC-I+ cells (Fig. 3H and 3L).
As our experimental protocol utilized male donor ESC-derived cells and recipient female mice, Y-chromosome expression was used to ascertain the fate of transplanted cells. Thus, Y-chromosome FISH of ischemic tissues in the MHC-I− group, but not MHC-I+ group, revealed a widespread atypical staining pattern, consistent with ESC-derived VE-CAD+MHC-I− cells undergoing active degeneration/necrosis (Fig. 3M–3O). Indeed, we detected very few Y-chromosome positive cells with a normal morphologic and FISH staining pattern, suggesting extensive rejection of transplanted VE-CAD+MHC-I− cells. Conversely, FISH staining performed on muscle injected with VE-CAD+MHC-I+ cells revealed typical staining, similar to that of normal male muscle, suggesting that MHC-I expression may modulate donor cell integrity and survival in this syngeneic transplantation model (Fig. 3P).
FACS analysis for DsRed was used to identify ESC-derived cells from total cells extracted from ischemic hindlimb muscle at 2 weeks post-transplantation. We identified that the survival of transplanted VE-CAD+MHC-I+ cells (Fig. 1Ad), or their progeny, was significantly greater than that of VE-CAD+MHC-I− cells (DsRed+ cells in tibialis muscle, 5.42% ± 0.61% vs. 0.97% ± 0.38%, respectively, both n = 7, p < .005; (Fig. 3Q). In addition to ESC-derived endothelial progenitor (VE-CAD+) cells, we also transplanted earlier stage ESC-derived hemangioblast Bry+Flk-1+MHC-I+ (Fig. 1Af) and Bry+Flk-1+MHC-I− cells (Fig. 1). Similar to VE-CAD+ cells, survival of transplanted Bry+Flk-1+MHC-I+ cells was significantly greater than that of Bry+Flk-1+MHC-I− cells (DsRed+ cells in tibialis muscle, 5.62% ± 0.69% vs. 0.68% ± 0.09%, respectively, both n = 7, p < .005) (Fig. 3Q).
We elected to further explore the pathobiology of the processes underlying the apparent rejection of ESC-derived MHC-I− cells. Initially, to exclude the possibility of poor graft viability or senescence, sorted VE-CAD+MHC-I− cells (as prepared for transplantation) were returned to culture for up to 3 weeks. These cells grew, proliferated, and were passaged normally throughout this additional culture period (data not shown). We next examined the nature of the host inflammatory cell infiltrate after donor cell transplantation. Immunohistochemistry revealed an abundance of infiltrating NK cells in VE-CAD+MHC-I− cell-injected tibialis muscle, whereas very few NK cells were observed in muscle injected with VE-CAD+MHC-I+ cells (7.32% ± 0.58% vs. 0.70% ± 0.27% of total cells respectively, both n = 7, p < .0001) (Fig. 4A and 4B).
We also employed an in vitro 51Chromium release assay to examine the susceptibility of various cell populations to syngeneic NK cell attack, including ESC-derived VE-CAD+MHC-I− progenitors, IFN-induced VE-CAD+MHC-I+ progenitors, MHC-I− control cells (YAC-1 cells, H-2Db− ), and MHC-I+ control cells (Concanavalin A blasts, H-2Db+). These studies revealed that MHC-I− cells, including ESC-derived VE-CAD+MHC-I− cells, were highly susceptible to NK cell-mediated killing. On the other hand, MHC-I+ control cells and VE-CAD+MHC-I+ progenitors evaded or were resistant to NK cell attack (Fig. 4C).
To confirm the importance of NK cell attack in vivo, a blocking antibody against NK cells was administered to abrogate NK cell activity at the time of donor cell transplantation. Two weeks after transplantation, survival of ESC-derived VE-CAD+MHC-I− cells was significantly greater in NK antibody-treated mice than in PBS-treated animals (5.37% ± 0.41% vs. 0.97% ± 0.38% of total single cells in tibialis muscle, p < .001; Fig. 4D).
We also investigated the survival of ESC-derived hemangioblast Bry+Flk-1+MHC-I+ and Bry+Flk-1+MHC-I− cells in the setting of absent NK cell function by using NSG mice, which lack functional NK cells . In agreement with our initial findings, following transplantation into the ischemic hindlimbs of NSG mice the survival of ESC-derived Bry+Flk-1+MHC-I− hemangioblasts was similar to that of ESC-derived Bry+Flk-1+MHC-I− cells (5.32 ± 0.76 vs. Five.51 ± 0.68 respectively, both n = 6, p = .61; Supporting Information, Figs. 5A and 5B).
Collectively, these in vitro and in vivo studies led us to conclude that NK cell attack is an important factor in the immune-rejection of transplanted ESC-derived vascular progenitor cells. IFNc-induced MHC-I expression promotes ESC-derived progenitor cell resistance to NK cell attack in vitro and enhances cell survival in vivo after syngeneic transplantation.
Given the improved early survival of MHC-I+ ESC-derived progenitor cells after syngeneic transplantation, we proceeded to investigate the intermediate-term survival and differentiation of these cells. The percentage of ESC-derived cells in total single cells extracted from tibialis muscle was assessed by FACS for DsRed expression at 4 weeks after MHC-I+ cell transplantation. At 2 weeks compared with 4 weeks post-transplantation, the numbers of Bry+Flk-1+MHC-I+ cells and VE-CAD+MHC-I+ cells (and/or their progeny) were unchanged (Bry+Flk-1+MHC-I+ cells: 5.62% ± 0.69% vs. 6.08% ± 0.28%, both n = 7, p = .41; VE-CAD+MHC-I+ cells: 5.42% ± 0.6132% vs. 5.58 ± 0.33, n = 7, p = .82; Fig. 5A). Although fluctuating levels of infiltrating inflammatory and local cells may have influenced these results (altered total number of cells), this data appears to suggest stable engraftment of transplanted MHC-I+ ESC-derived cells.
We performed double immunofluorescence staining on ischemic tibialis muscle transplanted with VE-CAD+MHC-I+ cells. Using DsRed to identify the ESC-origin of transplanted cells, we readily detected ESC-derived cells that were double positive for CD31 and DsRed (Fig. 5B), as well as cells double positive DsRed and vWF (Supporting Information, Fig. 6), indicating that MHC-I+ ESC-derived endothelial progenitor cells can contribute functionally to the endothelial vascular layer in vivo. We also identified ESC-derived cells (and/or their progeny) that expressed α-SMA or vascular smooth muscle cell (VSMC)-specific myosin heavy chain (SM-MHC) (Fig. 5B and Supporting Information, Fig. 6). Particularly, as SM-MHC expression is indicative of a mature VSMC phenotpye, and consistent with our Matrigel plug experiments, these data suggest that transplanted ESC-derived VE-CAD+MHC-I+ cells can give rise to VSMCs.
To confirm these results, we further investigated the in vivo fate of transplanted ESC-derived MHC-I+ cells using Y-chromosome FISH to locate donor cells that exhibited either endothelial (costaining for CD31) or smooth muscle (costaining for α-SMA) characteristics. Similar to the immunohistochemistry results, we identified frequent Y-chromosome+CD31+ and also Y-chromosome+α-SMA+ double positive cells (Fig. 5Cii) in VE-CAD+MHC-I+ cell-transplanted muscle. We also identified frequent Y-chromosome+CD31+ and Y-chromosome+ α-SMA+ double positive cells in the Bry+Flk-1+MHC-I+ cell-transplanted group (Fig. 5Ci). Importantly, Bry+Flk-1+MHC-I+ cells were negative for both CD31 and a-SMA prior to transplantation (data not shown).
Finally, FACS analysis was used to identify CD31 and VSMC-specific marker Sm22 positive donor DsRed+ cells in tibialis muscle at 2 and 4 weeks after transplantation. Consistent with our results above, FACS analysis revealed a significant number of donor-derived endothelial (CD31+) and smooth muscle (Sm22+) cells at both time-points (Fig. 5D and Supporting Information, Fig. 7). Interestingly, we observed that transplantation of Bry+Flk-1+MHC-I+ hemangioblasts, as compared with VE-CAD+MHC-I+ endothelial progenitor cells, gave significantly higher numbers of Sm22+ cells at both the 2 and 4 week time-points. Conversely, transplantation of VE-CAD+MHC-I+ endothelial progenitor cells, rather than Bry+Flk-1+MHC-I+ hemangio-blasts, was associated with greater numbers of CD31+ cells at both time-points (Fig. 5D and Supporting Information, Fig. 7).
As a whole, these results suggest that transplanted Bry+Flk-1+MHC-I+ and/or VE-CAD+MHC-I+ ESC-derived cells are able to engraft in a syngeneic model of hindlimb ischemia, and that a significant proportion of these cells are able to survive over the intermediate term. Both cell populations appeared capable, in vivo, of giving rise to both smooth muscle and endothelial progeny. However, the respective ESC-derived progenitor populations exhibited a clear bias toward an in vivo fate that was consistent with their anticipated in vitro phenotype.
Significant obstacles remain to be overcome before ESCs might realistically undergo clinical evaluation as a potential cell-based therapy for ischemic, cardiovascular, and other diseases . Indeed, as shown in the current study, even during syngeneic transplantation where donor and recipient share a very high degree of genetic homogeneity, the survival of transplanted ESC-derived progenitor cells is highly dependent on the host immune response. Here, we provide direct evidence to show that in both a Matrigel plug syngeneic model and a more complex syngeneic rodent ischemic hindlimb model, the host immune response and subsequent survival of transplanted ESC-derived progenitor cells is critically dependant on their MHC-I expression. In the absence of donor MHC-I expression, we observed virtually complete graft destruction; a phenomenon principally mediated by host NK cell attack. We show that in contrast to MHC-I− cells, a significant number of transplanted MHC-I+ ESC-derived progenitor cells (5%–6% of total single cells population in the muscle) survived out to at least 4 weeks post administration. Furthermore, detailed analysis using a robust genetic marking system indicated that transplanted MHC-I+ ESC-derived cells underwent differentiation into smooth muscle and endothelial progeny.
The role of MHC-I in NK cell-dependent killing has been extensively investigated and debated. Originally described as the concept of “missing self” and subsequently “revised missing self,” MHC-I is known to play a central role in the initiation and extent of NK cell attack on target cells [24, 25]. Nevertheless, it is accepted that an absence of MHC-I is not sufficient to initiate NK cell attack and that further activating ligands are required on target cells, with their expression ratio of inhibitory MHC-I to activating ligands greatly influencing this process. In conventional clinical syngeneic and allogeneic transplantation, this may not be relevant as this ratio may be relatively balanced and the relevant cells/tissues do not undergo extensive ex vivo manipulation or tissue culture adaptation. However, for potential cell therapy approaches that may subject cell populations to extensive tissue culture adaptation before transplantation, this may become a central problem. In fact, in many syngeneic animal models and initial clinical trials using adult stem/progenitor cell populations it is generally accepted that the survival of transplanted cells is very low. As an example, it was recently described that multi-potent adult progenitor cells (MAPCs) undergo functional elimination following transplantation in syngeneic and immune-compromised animal models. MAPCs are considered to exhibit low or even negative MHC-I expression, and subsequent inhibition of NK cells greatly improved MAPC survival in these models . These findings are consistent with our data and reinforce the paradigm that MHC-I negative donor cells may be vulnerable to NK cell destruction and poor graft survival, and that upregulation of donor cell MHC-I expression improves in vivo graft survival. Importantly, NK cell-mediated killing of targeted cells is an immediate response that does not require immune priming or extensive immune system activation. Unlike naïve T-cells, NK cells are “ready to go” with their lytic response able to be triggered within minutes of initial stimulation . Nevertheless, it is important to consider other likely contributors to graft destruction, including a hypoxic donor milieu, and insufficient environmental adaptation of the transplanted cells. However, it is possible that NK cell-mediated killing may also contribute to these latter processes.
The fact that IFNγ is able to upregulate MHC-I expression, including ESCs, is widely known [2, 3, 27]. However, to our knowledge, the use of this cytokine in the context of in vivo ESC transplantation has not been systematically studied. Our results strongly suggest, at least in a syngeneic setting, that the induction of MHC-I expression by the adjunctive use of IFNγ during in vitro ESC differentiation may improve the subsequent survival and immune integration of these cells after transplantation. In addition, in contrast to other reports , the use of IFNγ according to the protocols and dosages described in our study did not cause any significant detrimental effects on ESC survival or functionality. Apart from MHC-I, although VCAM-1 expression was unchanged, the expression of ICAM-1 was noted to increase after IFNγ treatment. Given that ICAM-1 is known to play a role in neutrophil and lymphocyte adhesion [28, 29], we speculate increased ICAM-1 expression may exert a further immune-modulatory role during ESC-derived cell transplantation.
The implications of our results potentially extend beyond ESCs. At the current time intense research is being undertaken to explore the feasibility of reprogramming adult somatic cells to become ESC-like induced pluripotent stem (iPS) cells [30, 31]. Given that one of the fundamental attributes of iPS cells is that they closely resemble ESCs, our findings suggest that due attention to the MHC antigenic status of iPS cells during their differentiation and possible application to the treatment of clinical diseases will be required. Similarly, pluripotent cells created by nuclear transfer might also be anticipated to exhibit MHC-I-dependant survival after transplantation into an immune-competent recipient. Underscoring the relevance of these assertions, it is also important to note that our syngeneic rodent model, using highly inbred animals, is likely to be a reasonably similar immunologic environment to that into which autologous iPS-derived cells might be administered.
Using both a Matrigel plug and a rodent ischemic hindlimb model, we identified that the MHC-I antigen expression status of ESC-derived vascular progenitor cells plays a critical role in determining the host immune response following in vivo syngeneic transplantation. In this setting, IFNγ-induced expression of MHC-I by ESC-derived progenitor cells was associated with attenuated host NK cell attack, enhanced graft survival, and the further differentiation of transplanted ESC-derived cells into VSMCs and endothelial cells. Understanding the complex regulatory mechanisms that govern donor–recipient immune interactions will undoubtedly be an obligatory hurdle that will need to be overcome for the successful implementation of cell-based regenerative therapies.
We thank Dr. Gordon Keller for kindly providing Brachyury-GFP ESCs and Dr. Eve Mezey for providing template mouse genomic DNA. We also thank the staff of the Laboratory of Animal Medicine and Surgery and Ms. Robin Schwartzbeck for their assistance with the mice surgery. We acknowledge the professional skills and advice of Drs. Christian A. Combs and Daniela Malide (Light Microscopy Core Facility, National Heart, Lung and Blood Institute) and Dr. J. Philip McCoy (Flow Cytometry Core Facility, National Heart, Lung and Blood Institute). This research was supported by the Intramural Research Program of the National Heart, Lung, and Blood Institute. C.B. is supported by the Faculty of Medicine, University of Glasgow.
Disclosure of Potential Conflicts of Interest
The authors indicate no potential conflicts of interest.