iPSCs engrafted in allogeneic hosts without immunosuppression induce donor-specific tolerance to secondary allografts

Significance

The immunological properties of induced pluripotent stem cells (iPSCs) during allogeneic transplantation have not yet been fully elucidated. In this study, we found that iPSCs exhibit immunologically unique behavior in major histocompatibility complex (MHC)-compatible/minor antigen-mismatched allogeneic combinations. When injected subcutaneously, iPSCs were not rejected and, 40 d later, induced immune tolerance in secondary transplanted grafts in a donor-specific manner. Transforming growth factor (TGF)-β2 expression and Treg behavior was associated with this phenomenon. The results of this study provide insights into the relationship between pluripotent stem cells and immune tolerance, offering crucial insights into the advancement of regenerative medicine using iPSCs.

Abstract

Currently, most cell or tissue transplantations using induced pluripotent stem cells (iPSCs) are anticipated to involve allogeneic iPSCs. However, the immunological properties of iPSCs in an allogeneic setting are not well understood. We previously established a mouse transplantation model of MHC-compatible/minor antigen-mismatched combinations, assuming a hypoimmunogenic iPSC-setting. Here, we found that iPSCs subcutaneously inoculated into MHC-compatible allogeneic host mice resisted rejection and formed teratomas without immunosuppressant administration. Notably, when skin grafts were transplanted onto hosts more than 40 d after the initial iPSCs inoculation, only the skin of the same strain as the initial iPSCs was engrafted. Therefore, donor-specific immune tolerance was induced by a single iPSC inoculation. Diverse analyses, including single-cell RNA-sequencing after transplantation, revealed an increase in regulatory T cell (Treg) population, particularly CD25⁺ CD103⁺ effector Tregs within the teratoma and skin grafts. The removal of CD25⁺ or Foxp3⁺ cells suppressed the increase in effector Tregs and disrupted graft acceptance, indicating the importance of these cells in the establishment of immune tolerance. Within the teratoma, we observed an increase in TGF-β2 levels, suggesting an association with the increase in effector Tregs. Our results provide important insights for future applications of allogeneic iPSC-based cell or tissue transplantation.

Controlling alloimmune responses, including rejection, is an essential aspect of transplantation medicine using allogeneic induced pluripotent stem cells (iPSCs). Although the establishment of hypoimmunogenic iPSCs, such as HLA homologous donor-derived or HLA-knockout iPSCs, has recently been reported (1–4), minor antigen incompatibilities may elicit unwanted immune responses (5). To investigate these immune responses, we established a mouse allogeneic transplantation model using MHC-compatible, minor antigen-mismatched combinations (6). In this model, skin allografts from C57BL/6N (H-2^b/b) or CBA (H-2^k/k) mice were transplanted into C3129 F1 mice (H-2^k/b) generated by crossing C3H (H-2^k/k) with 129 (H-2^b/b) mice. Although the MHC haplotype matched in this model, a certain level of allograft rejection was observed (6). Thus, using this model, we demonstrated an optimized immunosuppressive strategy, including a costimulatory blockade for graft acceptance (7). Furthermore, we induced immune tolerance in the transplanted skin grafts using hematopoietic progenitor cells generated from donor-type iPSCs (8).

iPSCs are subject to immune rejection in allogeneic hosts (9). Rejection has been observed even in cases of MHC-compatible/minor antigen mismatch (10). To this end, we conducted an experiment in which iPSCs were implanted into the aforementioned model. Contrary to previous reports, we found that iPSCs inoculated subcutaneously without immunosuppression were not rejected and formed teratomas. Furthermore, we transplanted skin grafts from various mouse strains as secondary transplants into mice with teratomas. Importantly, the skin grafts were accepted without immunosuppression in a donor-specific manner. Other pluripotent stem cells, such as embryonic stem cells (ESCs), have immune-privileged characteristics (11, 12); however, whether immune tolerance is induced in secondary grafts in this setting is unknown.

In this study, we aimed to elucidate the mechanism underlying the tolerance induced by iPSCs inoculation. Our findings revealed that iPSCs have the ability to induce donor antigen-specific immune tolerance, which is mediated by TGF-β2 expression and induction of effector Treg cells. The immunological significance of this phenomenon and its impact on regenerative medicine are also discussed.

Results

Undifferentiated Pluripotent Stem Cells Overcome Immunological Barriers Caused by Minor Antigen and Form Teratomas in MHC-Compatible Allogeneic Mice.

We performed skin transplantations from CBA/N (H-2^k/k) to B6C3F1 (H-2^b/k) mice, which are MHC-compatible allogeneic transplantations. CBA/N-skin grafts showed signs of rejection, such as epidermal atrophy, hyperplasia, and CD3⁺ cell infiltration 30 d after transplantation (Fig. 1A), and all grafts were rejected within 50 d (median survival time: MST = 39), whereas syngeneic grafts were accepted throughout the study (Fig. 1B). Next, we subcutaneously implanted iPSCs into the same mouse combination, and the survival of luciferase (Luc)-transfected CBA/N-iPSCs was evaluated based on their activity. The Luc signal was maintained for more than 60 d in both syngeneic and allogeneic recipients (Fig. 1 C and D). On day 40 after CBA/N-iPSC transplantation, we histopathologically evaluated the iPSC-derived tissue and found the three germ layers—neural tissue, cartilage, and ciliated epithelium—indicating that iPSC-derived teratomas were formed in both syngeneic and allogeneic immunocompetent mice (Fig. 1E). Teratoma volume increased over time in the allogeneic hosts (Fig. 1 F and G). Furthermore, the teratomas survived for 200 d without rejection (SI Appendix, Fig. S1 A and B). To exclude the notion that this phenomenon was unique to CBA/N-iPSCs, CBA/N-ESCs were also implanted into B6C3F1 mice, and teratoma formation was confirmed for more than 60 d (SI Appendix, Fig. S1 C–E). Furthermore, in a different combination of MHC-compatible allogeneic transplantations, BALB/c skin grafts (H-2^d/d) transplanted into B6D2F1 (H-2^b/d) mice were rejected (SI Appendix, Fig. S1 F and G), whereas BALB/c-iPSCs formed teratomas that survived for more than 60 d (SI Appendix, Fig. S1 H–J), suggesting that various PSC could be engrafted into MHC-compatible allogeneic hosts.

Fig. 1.

Engraftment of iPSC-derived teratoma in allogeneic immunocompetent mice. (A) Representative pathological images of the skin grafts. Macroscopic views and H&E-stained and immunohistochemical images (CD3) of B6C3F1 cells (Upper) and CBA/N-skin grafts (Lower) transplanted into B6C3F1 mice. (B) Survival of B6C3F1-skin and CBA/N-skin grafts in B6C3F1 mice. tx; transplantation. ***P < 0.001 (log-rank test). (C and D) Luc-CBA/N-iPSC subcutaneous transplantation. Representative bioluminescence images of Luc-CBA/N-iPSCs transplanted into CBA/N (Left) or B6C3F1 mice (Right) are shown (C), along with quantitative bioluminescence intensity (D). Luc; Luciferase. (E) H&E-stained images of CBA/N-iPSC-derived teratomas formed in CBA/N (Upper) or B6C3F1 mice (Lower). The neural tissue (ectoderm), cartilage (mesoderm), and ciliated epithelium (endoderm) are shown. (F and G) Representative macroscopic views (F) and weights (G) of the CBA/N-iPSC-derived teratomas engrafted with B6C3F1 cells. n = 20-33 per time point. **P < 0.01, ***P < 0.001 (Tukey’s HSD test). (H and I) Representative immunostaining images of CD3-positive cells infiltrating CBA/N-iPSC-derived teratomas engrafted in syngeneic (Upper) or allogeneic (Lower) (H), and quantitative infiltration scores (I). n = 3 to 5 per group. *P < 0.05 (Student’s t test). The experiment was repeated two times independently with similar results. (J) Activation markers (CD25 and CD69) on CD4⁺ (Left) and CD8⁺ T cells (Right) in the lymph nodes of B6C3F1 transplanted with CBA/N-iPSCs or CBA/N-skin. Analyses were performed 20 d after transplantation. n = 3 per group. **P < 0.01, ***P < 0.001 (Tukey’s HSD test). The experiment was repeated two times independently with similar results. [Scale bars: 25 µm in (A, E, and H).]

To elucidate the recipient immune response to MHC-compatible allogeneic teratomas, CBA/N-iPSCs were implanted into syngeneic CBA/N or allogeneic B6C3F1 mice, and 20 d later, CD3⁺ cells infiltrating the teratoma were histopathologically evaluated. The results showed that the infiltration score of CD3⁺ cells into allogeneic teratomas was comparable to that of syngeneic CBA/N mice (Fig. 1 H and I). In addition, CD25⁺ or CD69⁺ T cells in the lymph nodes of B6C3F1 recipients transplanted with CBA/N-skin or CBA/N-iPSCs were assessed 20 d after transplantation. iPSC implantation did not induce significant T cell activation compared to that seen with skin grafting (Fig. 1J).

We also investigated fully allogeneic combinations by transplanting CBA/N- or BALB/c-iPSCs into C57BL6/N mice. iPSCs were not accepted in this combination, and only adipose tissue was formed (SI Appendix, Fig. S1 K–M).

These results suggest that the potency of MHC-compatible allogeneic PSCs is lesser in inducing immunological rejection than that of the skin grafts and may be relatively easy to engraft.

Prior PSCs Implantation Induces Donor-Specific Allograft Tolerance.

To further clarify the immune response of the donor, we performed a second series of transplantation experiments. We inoculated CBA/N-iPSCs subcutaneously in B6C3F1 and then transplanted various donor-derived skin grafts at different sites and time points to evaluate skin graft acceptance (Fig. 2A). All CBA/N-skin grafts transplanted 10 d after iPSCs inoculation were rejected (MST = 38), as was the case without treatment. Notably, 40% of the CBA/N-skin grafts transplanted 20 d after iPSCs inoculation and 83% of those transplanted 40 d after iPSCs inoculation were engrafted over 100 d (Fig. 2B). In histopathological analyses of the accepted skin grafts (with iPSCs inoculation 40 d prior), no signs of rejection, such as epidermal atrophy, thickening, or CD3⁺ cell infiltration, were observed compared to those seen in the untreated group (Fig. 2 C and D). The absence of a rejection response lasted until day 100, with pathological findings comparable to those of syngeneic B6C3F1-skin (Fig. 2 E and F). Therefore, prior implantation of iPSCs into MHC-compatible allogeneic hosts induces tolerance to secondary transplanted allografts.

Fig. 2.

Prior PSCs implantation induces donor-specific allograft tolerance. (A) Experimental design. s.c.; subcutaneous. (B) Survival of CBA/N-skin grafts in B6C3F1 mice, with or without prior CBA/N-iPSC implantation. n = 5 to 12 per group. Three independent experiments were performed. ***P < 0.001 vs. nontreatment (log-rank test). (C–F) Representative immunostaining images of CD3⁺ cells infiltrating skin grafts (C and E), and quantitative infiltration scores (D and F). *P < 0.05 (Student’s t test). n = 5 to 7 per group. (G) Survival of 129x1 or BALB/c skin grafts in B6C3F1 mice with or without prior CBA/N-iPSC implantation. n = 5 to 12 per group. Three independent experiments were performed. (H) Survival of CBA/N-skin grafts in B6C3F1 mice with or without prior 129x1- or BALB/c-iPSCs implantation. n = 5 per group. (I) Survival of CBA/N-skin grafts in B6C3F1 mice, with or without prior CBA/N-ESC implantation. n = 5 to 6 per group. *P < 0.05 (log-rank test). (J) Macroscopic appearance of the B6C3F1 recipient with CBA/N-ESCs and skin grafts (day 143). (K) Macroscopic appearance of skin grafts. [Scale bars: 25 µm in (C and E).]

To investigate antigen specificity in this tolerance, we also transplanted 129x1- and BALB/c-skin grafts together with the CBA/N-skin grafts. However, although the CBA/N-skin grafts were accepted (Fig. 2B), survival of the third-party skin grafts was not prolonged at any time point (Fig. 2G). Conversely, subcutaneous implantation of 129x1- or BALB/c-iPSCs into B6C3F1 mice 40 d before transplantation of CBA/N-skin did not result in prolonged CBA/N-skin engraftment (Fig. 2H). By contrast, the induction of immune tolerance 40 d after iPSC implantation was also reproducible in CBA/N-ESCs using transplantation (Fig. 2 I–K). These results suggest that donor antigen-specific immune tolerance was induced 40 d after the implantation of donor-derived PSCs.

Teratoma Is the Site of Immune Response.

We analyzed the secondary lymphoid tissues of B6C3F1 mice transplanted with CBA/N-iPSCs to determine the percentages of Tregs, anergic T cells, and MDSCs that have been reported to contribute to immune tolerance. We examined the changes in immune cells over time using FACS, but no significant changes were observed (SI Appendix, Fig. S2 A, B, and D). We also examined whether the adoptive transfer of lymph node- and spleen-derived lymphocytes from B6C3F1 mice, wherein immune tolerance to CBA/N was induced by prior implantation of CBA/N-iPSCs, could induce immune tolerance in CBA/N mice. Lymphocytes derived only from CBA/N-skin-transplanted mice showed a significantly shorter engraftment period than that of the adoptive transfer of lymphocytes derived from wild-type mice (SI Appendix, Fig. S2E). By contrast, lymphocytes derived from tolerant mice showed no significant difference in the engraftment period compared with those derived from wild-type mice (SI Appendix, Fig. S2E). Furthermore, we performed adoptive transfer experiments with an enriched CD4⁺ or CD4⁻ cell fraction from lymph node- and spleen-derived lymphocytes of tolerant mice. However, the transfer of immune cells from secondary lymphoid tissues of tolerant mice did not significantly prolong skin graft survival (SI Appendix, Fig. S2 F and G). Therefore, we focused on the immune response inside the teratomas rather than in the secondary lymphoid tissues.

We investigated the contribution of teratomas to the induction and maintenance of immune tolerance by performing teratomectomies before and after CBA/N-skin transplantation (Fig. 3A). The teratomectomy performed 20 d before CBA/N-skin grafting prevented the induction of immune tolerance (Fig. 3B). When teratomectomy was performed 20 d after skin grafting, graft survival was prolonged; however, most of them were rejected by 120 d (Fig. 3B). These results indicated that iPSC-derived teratoma engraftment in allogeneic mice is essential for the induction and maintenance of immune tolerance.

Fig. 3.

Role of teratoma in the induction of immune tolerance. (A) Experimental design. (B) Survival of CBA/N-skin grafts in B6C3F1 mice after teratomectomy. **P < 0.01 (log-rank test). (C and D) Representative immunostaining images of CD3-, Foxp3-, and CD19-positive cells infiltrating CBA/N-iPSC-derived teratomas engrafted in B6C3F1 mice (C) and quantitative infiltration area (D). Three independent experiments were performed. Arrows indicate cell accumulation. n = 4 to 6 per group. *P < 0.05 (Tukey’s HSD test). (E) Representative immunostaining images of CD3-, Foxp3-, and CD19-positive cells infiltrating CBA/N-iPSC-derived teratomas engrafted into B6C3F1 mice on day 40. [Scale bar: 50 µm in (E).]

Subsequently, we performed immunostaining for CD3⁺, Foxp3⁺, and CD19⁺ cells infiltrating the CBA/N-iPSC-derived teratomas engrafted in B6C3F1 mice (Fig. 3C). Although there was no significant change in CD3-, Foxp3-, and CD19-positive areas (%) between days 10 and 40 (Fig. 3D), there was a trend toward the localization of each fraction at day 40 (Fig. 3C). Furthermore, CD3-, Foxp3-, and CD19-positive cells were found to colocalize and formed a tertiary lymphoid structure-like on teratomas of 40 d (Fig. 3E). These findings suggested the presence of an immune response within the teratoma.

We further investigated the dynamics of iPSC-derived cells in the host using Luc- or GFP-iPSCs to determine whether teratoma-derived cells moved to the host peripheral lymphoid tissues and played a key role in tolerance induction. We first confirmed that the fluorescence intensity of Luc-CBA/N-iPSCs increased with the cell count (SI Appendix, Fig. S3A). Luc signals in various lymphoid tissues of B6C3F1 mice transplanted with Luc-CBA/N-iPSCs 40 d previously were examined. However, no Luc signal was detected in any tissues (SI Appendix, Fig. S3 B and C). We also verified Luc signals 200 d after Luc-CBA/N-iPSC transplantation; however, a faint signal was detected only in the draining lymph nodes (SI Appendix, Fig. S3D). Next, GFP-transfected CBA/N-iPSCs (SI Appendix, Fig. S3E) were subcutaneously injected into eGFP B6C3F1 (Foxp3.eGFP-2A-DTR-2A-Luciferase C57BL/6N × C3H/He) mice, to avoid the possibility of rejection by the immunogenicity of GFP. Forty days after the iPSCs implantation, the percentages of donor haplotype (H-2K^k+H-2K^b-)-GFP⁺ cells in the peripheral lymphoid tissues were examined using FACS. In the teratoma derived from GFP-CBA/N-iPSCs, GFP was apparently detected (SI Appendix, Fig. S3F). However, donor haplotype-GFP⁺ cells were barely detected, similar to the group that received CBA/N-iPSC transplants (SI Appendix, Fig. S3 G and H). Hence, it is unlikely that systemic hematopoietic chimerism was induced by iPSC-derived cells.

Therefore, the presence of teratomas in allogeneic hosts is essential for the induction and maintenance of immune tolerance, suggesting the importance of the immune response inside teratomas.

Effector Treg Count Increases in Teratoma with Time Course.

Next, we performed single-cell RNA-sequencing (scRNA-seq) analysis to investigate the changes in immune cells infiltrating the teratoma in detail. CBA/N-iPSCs were subcutaneously injected into B6C3F1 mice. Teratomas were collected on days 10, 20, and 40, and scRNA-seq was performed on CD45⁺ cells within the teratomas (Fig. 4A). The proportion of neutrophils and B cells in the immune cells increased, and changes in macrophage and Treg phenotypes were observed (Fig. 4 B–E).

Fig. 4.

Single-cell RNA-sequencing revealed the increase of effector Treg. (A) Experimental design. TILs: teratoma-infiltrating lymphocytes. (B–D) Identification of cell types and their proportions in TILs. The cell types shown in UMAP (B) were identified based on the expression of marker genes (C), and their rates are shown in a bar plot (D). (E) Each plot shows the TIL density of TILs with time. (F) Gene ontology enrichment analysis of differentially expressed genes (upregulated) in Tregs (day 40 vs. day 10) from TILs. (G and H) UMAPs show the subclusters of Tregs in TILs (G) and the bar graph shows the percentage of each subcluster over time (H). (I) Heatmap showing the standardized expression values of the top five differentially expressed genes in the Treg subclusters. (J) Violin plot (Top) and kernel density estimate (Bottom) for Cd3e, Cd4, Foxp3, Il2ra, and Ctla4 expression in the Treg subcluster. (K and L) Percentage of Foxp3⁺CD4⁺ T cells in TILs over time. Representative plot patterns (K) and quantitative bar graphs (L) are shown. n = 6 per group. *P < 0.05 (Student’s t test). (M and N) Percentage of CD103⁺CD25⁺Foxp3⁺CD4⁺ T cells in TILs over time. Representative plot patterns (M) and quantitative bar graphs (N) are shown. n = 6 per group. *P < 0.05 (Student’s t test).

Changes in the count of Tregs over time were particularly pronounced. When 40-d and 10-d Tregs were compared based on gene set enrichment analysis, activation of 40-d Tregs was observed (Fig. 4F). Subclustering analysis revealed three major clusters, with an increased proportion of Treg-1 on 40-d Tregs (Fig. 4 G and H). Analysis of differentially expressed genes showed that Treg-1 expressed Lrig1 and Itgae, which have been reported to be expressed in effector Tregs (Fig. 4I) (13, 14). In addition, Il2ra and Ctla4, which are immunosuppressive molecules expressed by Tregs, were strongly expressed in Treg-1 cells (Fig. 4J). Indeed, when FACS was performed on Tregs within the teratoma, an increase in CD103⁺CD25⁺ cells among Foxp3⁺CD4⁺ T cells was observed, although the proportion of Foxp3⁺ cells among CD4⁺ T cells did not change (Fig. 4 K–N). Therefore, the count of effector Tregs is increased in teratomas associated with tolerance induction.

Furthermore, we examined Tregs in skin grafts. Immunostaining of Foxp3⁺ cells infiltrating the skin grafts was significantly higher in the iPSC-pre-implanted group than that in the untreated group 30 d after skin grafting (SI Appendix, Fig. S6 A and B). Furthermore, a comparison of Foxp3⁺ cells infiltrating 100 d after CBA/N skin grafting with syngeneic B6C3F1-skin grafts revealed a significantly higher count of Foxp3⁺ cells in the allogeneic grafts (SI Appendix, Fig. S6 C and D). Therefore, accompanied by tolerance induction with iPSCs implantation, the accumulation of Tregs was significant not only in the teratoma but also in secondary allografts.

iPSCs-Induced Immune Tolerance Is Treg-Dependent.

To confirm the involvement of Tregs in iPSC-induced immune tolerance, Treg depletion experiments were performed (Fig. 5). Administration of the anti-CD25 antibody (Ab) reduced Foxp3⁺ cell counts in CD4⁺ T cells in the peripheral blood and teratoma (Fig. 5 A–E). Teratoma size was not significantly changed by antibody treatment on day 40 (Fig. 5 F and G). However, the engraftment rate of the secondarily transplanted CBA/N skin grafts was significantly lower in the anti-CD25 Ab-treated group (Fig. 5H). Therefore, it was suggested that Tregs contribute to the induction of tolerance. To further verify this, we conducted another Treg depletion experiment using Foxp3^DTR B6C3F1 (Foxp3.eGFP-2A-DTR-2A-Luciferase C57BL/6N x C3H/He) mice (Fig. 5I). Tregs in the peripheral blood and teratoma were significantly decreased 4 d after administration of diphtheria toxin (DT) (Fig. 5 J–M). Teratoma size was not significantly changed by the DT administration (Fig. 5 N and O). However, the secondarily transplanted CBA/N skin grafts in the DT-treated Foxp3^DTR group were completely rejected within 20 d (Fig. 5P). To determine whether the rejection of skin grafts was simply promoted by Treg depletion, we transplanted CBA/N- skin grafts onto Foxp3^DTR B6C3F1 mice without teratoma formation and depleted Tregs by administering DT on the day before and the day of transplantation. As a result, the CBA/N skin grafts were rejected at approximately the same time, regardless of whether Tregs were depleted (SI Appendix, Fig. S7). Therefore, it was considered unlikely that Treg depletion via DT administration nonspecifically enhanced host rejection responses to shorten graft survival in this mouse transplantation model. These results strongly suggest that Foxp3⁺ Tregs contribute to the iPSCs-induced immune tolerance.

Fig. 5.

iPSCs-induced immune tolerance is Treg-dependent. (A) Experimental design for CD25⁺ depletion. (B and C) Percentage of Foxp3⁺CD4⁺ T cells in PBMC on day 40 after CBA/N-iPSC s.c. Representative plot patterns (B), and quantitative bar graphs (C). n = 5 per group. ***P < 0.001 (Student’s t test). (D and E) Percentage of Foxp3⁺CD4⁺ T cells in iPSC-derived teratomas on day 40 after CBA/N-iPSC s.c. (D) Representative plot patterns and (E) quantitative bar graphs. n = 5 per group. *P < 0.05 (Student’s t test). (F and G) The volume of CBA/N-iPSC-derived teratomas 40 d after implantation of CBA/N-iPSC into B6C3F1. Representative macroimages (F) and quantitative teratoma weights are shown (G). n = 5 per group. *P < 0.05 (Student’s t test). (H) Survival of CBA/N-skin grafts in CBA/N-iPSC-implanted B6C3F1 mice treated with the control IgG i.p. (blue), or anti-CD25 Ab i.p. (orange). *P < 0.05 (log-rank test). (I) Experimental design using Foxp3^DTR B6C3F1 mice. (J and K) Percentage of GFP⁺CD4⁺ T cells and Foxp3⁺CD4⁺ T cells in PBMC before (day 38) and after (day 42) DT injection. Representative plot patterns (J), and quantitative bar graphs (K). n = 4-6 per group. *P < 0.05, ***P < 0.001 (Tukey’s HSD test). (L and M) Percentage of Foxp3⁺CD4⁺ T cells in iPSC-derived teratomas before (day 38) and after (day 42) DT injection. (L) Representative plot patterns and (M) quantitative bar graphs. n = 6 per group. **P < 0.01 (Tukey’s HSD test). (N and O) The volume of CBA/N-iPSC-derived teratomas 42 d after implantation of CBA/N-iPSCs into Foxp3^DTR B6C3F1. Representative macroimages (N) and quantitative teratoma weights are shown (O). n = 6 per group. *P < 0.05 (Student’s t test). (P) Survival of CBA/N-skin grafts in CBA/N-iPSCs-implanted host mice treated with DT or saline. *P < 0.05, **P < 0.01 (log-rank test).

To determine whether the Tregs that increase in the skin graft are derived from the Tregs expanded within the teratoma, we transplanted a teratoma derived from CBA/N-iPSCs, formed in Foxp3^DTR B6C3F1 mice, subcutaneously into wild-type B6C3F1 mice on day 30 after the teratoma formation. Two days later, additional skin grafts were transplanted into the same recipients. We evaluated whether Foxp3^DTR B6C3F1-derived hematopoietic cells contained in the teratoma migrated to the skin grafts based on Luc signals (SI Appendix, Fig. S6 E and F). As a control for the signal detection, a Luc signal could be detected in skin grafts from Foxp3^DTR C57BL/6N mice (physiologically containing Luc⁺ Tregs) transplanted in syngeneic hosts (SI Appendix, Fig. S6F). With this condition, Luc signals were detected in the teratoma and in regions corresponding to lymph nodes (axillary and inguinal), but no signals were detected in the skin grafts not only from syngeneic but also allogeneic donors (SI Appendix, Fig. S6F). Since this experimental model differs from the original transplantation model, these results alone do not definitively rule out the migration of Tregs from the teratoma to the skin graft, but the possibility appears to be low. These findings suggest that, in immune tolerance induced by iPSC implantation, Foxp3⁺ Tregs do not directly regulate immunity within the secondary skin graft by migrating from the teratoma. Rather, they may exert indirect effects within the teratoma or lymph nodes.

Tgfb2 Expression Was Upregulated in Teratoma and Induced Smad2/3 Phosphorylation in Local T Cells.

Tgfb is involved in the proliferation and maintenance of Tregs (15, 16). Therefore, we compared the expression of Tgfb family members within teratomas. The results showed that the expression of Tgfb1, 2, and 3 was upregulated, with Tgfb2 expression significantly higher by day 40 (Fig. 6A). The amount of Tgfb2 protein in the teratoma and serum of teratoma-bearing mice was also measured, and only that of the teratoma increased (Fig. 6 B and C). scRNA-seq of immune cells within the teratoma showed no significant increase in the expression of Tgfb family members; that is, Tgfb2 or 3 were not detected (Fig. 6D). Therefore, increased expression of Tgfb2 could be attributed to teratoma parenchymal cells. In fact, immunostaining for Tgfb2 on 40-d teratomas showed that it was positive in areas with a muscle or neural appearance (Fig. 6E). The phosphorylation of smad2 in CD4⁺ and CD8⁺ T cells infiltrating the teratoma was also assessed using FACS, and a significant increase was detected in 40-d teratoma (Fig. 6 F and G). Simultaneously, p-smad2 levels in T cells in the spleen remained unchanged on days 10, 20, and 40 after iPSC implantation (SI Appendix, Fig. S2C).

Fig. 6.

Tgfb2 expression was upregulated in teratoma and induced Smad2/3 phosphorylation in local T cells. (A) Expression of Tgfb family members in teratomas. Tgfb1 (Left), Tgfb2 (Middle), and Tgfb3 (Right) expression in iPSC and teratomas is shown. n = 5 per group. *P < 0.05 (Tukey’s HSD test). (B and C) Quantification of Tgfb2 (pg/mL): Tgfb2 expression was measured in iPSCs and teratomas at 10, 20, and 40 d after iPSC implantation (Left) and in mouse serum (Right). n = 4 to 6 per group. *P < 0.05, **P < 0.01, ***P < 0.001 (Tukey’s HSD test). (D) Expression of Tgfb family members in TILs. Gene expression of the Tgfb family in TILs (Left) and in each cluster of TILs (Right). (E) Representative Tgfb2 immunostained images of the teratomas on day 40 after implantation. Overall (Left) and magnified (Right) views are shown. Muscle-like tissue (a), nerve-like tissue (b), and unknown cells (c) are positive for Tgfb2. (F and G) Phosphorylation analysis of smad2 in T cells expressing TILs. Representative histograms are shown (F), and MFI is shown (G). n = 6 per group. *P < 0.05 (Student’s t test). (H and I) MFI of p-smad2 in T cells expressing TILs. Representative histograms (H) and quantitative bar graphs (I) are shown. n = 3 per group. *P < 0.05 (Student’s t test). (J and K) Percentage of CD103⁺CD25⁺Foxp3⁺CD4⁺ T cells in teratomas on day 40 after CBA/N-iPSC s.c. (J) Representative plot patterns and (K) quantitative bars n = 3 per group. *P < 0.05 (Student’s t test). (L and M) Volume of the CBA/N-iPSC-derived teratomas in B6C3F1 cells. Representative macroimages (L) and quantitative weights of teratomas with control IgG i.p. (Left) or anti-Tgfb Ab i.p. (Right) are shown (M). n = 3 per group. *P < 0.05 (Student’s t test). (N and O) Representative pathological images of skin grafts. Macroscopic views and H&E-stained and immunohistochemical image (CD3) of CBA/N-skin grafts with control IgG i.p. (Upper) or anti-Tgbf Ab i.p. (Lower) (N) and quantitative infiltration area (O) are shown. n = 3 to 4 per group. *P < 0.05 (Student’s t test). [Scale bar: 25 µm in (N).]

Next, we investigated the effects of neutralizing Tgfb using antibodies in an iPSC-induced tolerance model. On day 40, anti-Tgfb Ab treatment significantly inhibited the induction of p-smad2 in the T cells of the teratomas (Fig. 6 H and I). Furthermore, the anti-Tgfb Ab-treated group showed a lower proportion of Foxp3⁺ CD4 T cells in the teratomas than that of the control IgG group, and the proportion of CD103⁺CD25⁺ effector Tregs was also lower (Fig. 6 J and K). Although the administration of anti-Tgfb Ab had no effect on the volume of the 40-d teratoma (Fig. 6 L and M), the condition of skin allografts was significantly worse in the antibody-treated group, as estimated by pathological findings (Fig. 6 N and O).

Thus, the increase in Tgfb2 in the teratoma seems to be associated with effector Treg induction. Taken together with the results of teratoma removal (Fig. 3B) and Treg depletion (Fig. 5) experiments, the Tgfb2–effector Treg axis in the teratoma may be crucial for the induction of immune tolerance by iPSCs implantation.

Discussion

In this study, we found that iPSCs were not rejected and formed teratomas when subcutaneously inoculated into MHC-compatible but minor antigen-mismatched allogeneic hosts. Importantly, engraftment was induced without the administration of immunosuppressants. Furthermore, in the immunocompetent allogeneic hosts, once iPSCs were accepted, secondarily transplanted skin grafts were also accepted in a donor-specific manner without any use of immunosuppressants. Therefore, prior iPSC implantation can induce immune tolerance to donor antigens. We also demonstrated the Treg-dependent mechanisms of tolerance induction. These findings provide insights for future research on the relationship between stem cells and immune response. Furthermore, they provide important information for the development of safe and efficient regenerative medicine (transplantation of iPSC-derived cells and tissues).

Mouse ESCs have been shown to be engrafted in MHC-mismatched allogeneic mice via the production of Tgfb by ESCs (17). In addition, several in vitro experiments have reported immune-privileged characteristics of PSCs, such as low MHC expression and expression of immunosuppressive factors, such as Arg1 and MFG-E8, in mouse and human ESCs (18–20). By contrast, MHC-matched allogeneic iPSCs are not engrafted in some mice and monkeys, reflecting the differences in the characteristics of each cell line and immunogenicity among donor recipients (9, 10). Moreover, the injection of donor-derived ESCs into the portal vein induces systemic tolerance (21, 22). However, tolerance may also be induced by the injection of antigens into the portal vein alone (23, 24), suggesting the involvement of a different mechanism. Therefore, to the best of our knowledge, our study is the first to report that iPSCs engraft in allogeneic immunocompetent mice without treatment and induce systemic tolerance by the engraftment of teratomas.

We further identified Tgfb2 as one of the factors that induces Tregs, which may contribute to this phenomenon in teratomas. The three isoforms of Tgfb (Tgfb1, Tgfb2, and Tgfb3) are transcribed from different genes, but all direct the transcription of TGF-β-modulated genes by the canonical TgfbR signaling (ALK5 phosphorylates SMAD2 and SMAD3, which then partner with SMAD4, and translocate to the nucleus) (25, 26). Tgfb2 is mainly expressed in the nerves and eyes (27–31). It has been considered an important factor in promoting the immune privilege of those tissues by contributing to the differentiation of CD4⁺CD25⁻ T cells into Tregs (15, 16, 32) and the generation of Tregs via the reduced antigen-presenting capacity of antigen-presenting cells (28, 32, 33). We also considered that the immune-privileged properties of the nerves derived from living mice and ESCs, which do not trigger an allogeneic lymphocyte response (34, 35), as well as the Tgfb2 expression in ESC-derived nerves that is enhanced in the inflammatory environment after transplantation (36), are related to the findings of our study.

We also observed the formation of tertiary lymphoid organ (TLO)-like tissue inside the allogeneic iPSC-derived teratoma (Fig. 3E), which forms in inflammatory tissues and many allogeneic grafts. However, whether the formation of TLOs is beneficial or deleterious for graft engraftment is contested (37). Nevertheless, recent reports have shown that TLO favorably contributes to graft viability via Tregs (38–41). Based on the study by Li et al. tolerance induction after lung transplantation is associated with the induction of Foxp3⁺ T cell–rich bronchus-associated lymphoid tissue (BALT), and depletion of graft-resident Foxp3⁺ T cells has been shown to abolish the maintenance of lung transplant tolerance (39, 41). Similarly, the research by Rosales et al. demonstrated that Tregs are critical for kidney graft survival in a mouse renal transplantation model with induced spontaneous tolerance, where Treg-rich organized lymphoid structures (TOLS) were formed in the tolerized kidneys (40). These studies also suggest that TLOs contribute to the differentiation and maintenance of Tregs (39–41). A clinical report has also shown that a decrease in Th17 in TLOs correlates with graft engraftment (42); our results show a decrease in IL-17⁺ cells within teratomas (SI Appendix, Fig. S8 A and B) and, also in tolerated skin grafts (SI Appendix, Fig. S8 C and D), are consistent with this finding. Studies focusing on the TLO-like tissues within allo-iPSC-derived teratomas may provide different insights.

This study shows that implantation of PSCs induces donor antigen-specific immune tolerance. On the other hand, there have been various reports on the induction of immune tolerance using donor-derived cells and organs, such as through hematopoietic chimerism due to donor-derived hematopoietic stem cell transplantation (43), and spontaneous tolerance after liver (44) or kidney (45) transplantation. Various mechanisms have been proposed in each experimental system, but there is no common unified view. However, by conducting further analysis, including the experimental system we have shown in this study, it may be possible to reveal insights to understand the mechanisms of transplant immune tolerance.

As a limitation of this study, the phenomenon of immune tolerance induction by PSCs was demonstrated only in transplantation experiments using specific mouse-to-mouse combination. In this iPSC-induced immune tolerance model, where the removal of effector Treg-rich teratoma disrupts immune tolerance, elucidating the mechanisms how Tregs induce donor-specific tolerance to secondary allografts is of significant importance. In our investigation, we observed that more Tregs infiltrated the tolerized allogeneic skin grafts compared to syngeneic skin grafts. However, it is unlikely that this was a result of effector Tregs expanded within the teratoma migrating to the grafts (SI Appendix, Fig. S6). Rather, Foxp3⁺ Tregs are considered to contribute to iPSC-induced immune tolerance at different locations from the secondary skin grafts, such as the teratoma or lymph nodes. Previously, a tumor study demonstrated a mechanism of antigen-specific immune regulation where Foxp3⁺ T cells eliminate dendritic cells in tumor-draining lymph nodes (46). A similar mechanism may operate in our experimental system. For such analyses, it will be necessary to utilize reporter mice capable of tracking Tregs within the teratoma to perform detailed investigations. Furthermore, research using intrateratoma Treg-specific depletion experiments or Tgfb2-KO iPSCs would also be informative. It should be noted that neutralizing Tgfb itself may induce transplant rejection, not just in the PSCs-mediated tolerance induction. Therefore, it is important to be careful when interpreting experimental data.

Finally, we verified the phenomenon in which iPSCs (and ESCs) can induce allo-immune tolerance found in various mouse combinations and elucidated part of its mechanism. It is possible to further elucidate which cell populations in the teratoma contribute to Treg formation and use them as immune tolerance-inducing cells through in vitro differentiation, which may lead to the development of an innovative, mild treatment to induce immune tolerance without teratoma formation.

Materials and Methods

Animals.

Male B6C3F1 (haplotype H2b/k) (cross between female C57BL/6N and male C3H/He), B6D2F1 (H-2b/d) (cross between female C57BL/6N and male DBA/2), CBA/N (H-2 k/k), BALB/c (H2d/d), and 129X1/SvJ (H-2b/b) mice were purchased from Japan SLC, Inc. (Shizuoka, Japan). The use of Foxp3.eGFP-2A-DTR-2A-Luciferase (Foxp3^DTR) C57BL/6N mice (47) was permitted by Günter Hämmerling [German Cancer Research Center (DKFZ)] and kindly provided by Shohei Hori (Tokyo University) and Takeshi Kawabe (Tohoku University, Japan). All animal experiments were approved by the Hokkaido University Animal Care Committee (approval number: 18-0004). All animal experiments were performed in accordance with the Animal Research: Reporting of In Vivo Experiments guidelines.

PSC Transplantation.

PSCs were seeded at a density of 2.5 × 10⁵ cells/6-cm dish in the maintenance medium described above, trypsinized on day 4, washed twice with PBS, and subcutaneously injected into mice at 2.5 × 10⁶ cells/50 µL Matrigel (Corning). In some experiments, a teratoma derived from CBA/N-iPSCs was formed in Foxp3^DTR B6C3F1 mice and transplanted subcutaneously into wild-type B6C3F1 mice on day 30 after its formation, followed by skin grafting.

Skin Transplantation.

For mouse skin transplants, recipient mice were anesthetized by intraperitoneal injection of a three-drug mixture of medetomidine (Nippon Zenyaku Kogyo Co., Ltd.), midazolam (Takeda Pharmaceutical Co., Ltd.), and butorphanol (Meiji Seika), and the dorsal side of the auricle skin or tail skin from donor mice was transplanted onto the dorsal thoracic region of the recipients. The recipient mice were wrapped with bandages for 7 d to protect the skin grafts and warmed until they moved freely. Graft diameter was measured to assess rejection. The day of graft rejection was defined as the day on which the graft reached less than 30% of its initial diameter. Graft survival was calculated using the following formula:

$$\begin{array}{c}Graft survival\hfill \\ =\left(\frac{total number of transplanted grafts-number of rejected grafts}{total number of transplanted grafts}\right)\times 100.\hfill \end{array}$$

All mice were killed by cervical dislocation at the end of the experiment.

scRNA-seq Analysis.

Teratomas from each mouse were collected at 10 (n = 33), 20 (n = 20), and 40 d (n = 20) after subcutaneous administration of CBA/N iPS into B6C3F1 mice and processed collectively at each time point. Single-cell suspensions were obtained using BD Horizon™ Dri Tumor & Tissue Dissociation Reagent (BD Biosciences), and infiltrated immune cells were collected through sorting of CD45-positive cells by FACSAria II (BD Biosciences; SI Appendix, Fig. S5A). We followed the 10× Genomics protocol for Next GEM Single Cell 3′ Gene Expression v3.1 (CG000315 Rev C) and used cell suspensions with ≥70% viability to generate complementary DNA and prepare scRNA-seq libraries. The libraries were sequenced using an Illumina NovaSeq 6000 (Takara Bio Inc., Japan). Raw sequencing data were processed using the Loupe Cell Browser (10× Genomics) and further processed using the R package Seurat (v4.0.1) to filter low-quality cells (SI Appendix, Fig. S5B). Normalization, scaling, and variable gene selection were performed using the SCTransform function in Seurat, and principal component analysis was performed on the resulting gene expression matrices. Uniform Manifold Approximation and Projection (UMAP) allowed us to assign cells to thirteen clusters. The cluster’s DEGs were identified with Findmarker and manually annotated with reference to CellKB and the R package Celldex (48, 49). Cd45⁻ clusters in teratoma-infiltrating lymphocytes (TILs) and Cd4⁻Foxp3⁻ subcluster in Treg clusters were removed (SI Appendix, Fig. S5 C and D). The cell types were annotated based on known cell markers and differential gene expression analyses. The results were visualized using R (v4.0.1) and the Seurat, ggplot2, and dplyr packages.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

Raw and processed scRNA-seq data have been deposited in the NCBI GEO database (GSE279441) (50). All other data are included in the manuscript and/or SI Appendix.