In vitro and in vivo functions of T cells produced in complemented thymi of chimeric mice generated by blastocyst complementation

Blastocyst complementation is an intriguing way of generating humanized animals for organ preparation in regenerative medicine and establishing novel models for drug development. Confirming that complemented organs and cells work normally in chimeric animals is critical to demonstrating the feasibility of blastocyst complementation. Here, we generated thymus-complemented chimeric mice, assessed the efficacy of anti-PD-L1 antibody in tumor-bearing chimeric mice, and then investigated T-cell function. Thymus-complemented chimeric mice were generated by injecting C57BL/6 (B6) embryonic stem cells into Foxn1nu/nu morulae or blastocysts. Flow cytometry data showed that the chimeric mouse thymic epithelial cells (TECs) were derived from the B6 cells. T cells appeared outside the thymi. Single-cell RNA-sequencing analysis revealed that the TEC gene-expression profile was comparable to that in B6 mice. Splenic T cells of chimeric mice responded very well to anti-CD3 stimulation in vitro; CD4+ and CD8+ T cells proliferated and produced IFNγ, IL-2, and granzyme B, as in B6 mice. Anti-PD-L1 antibody treatment inhibited MC38 tumor growth in chimeric mice. Moreover, in the chimeras, anti-PD-L1 antibody restored T-cell activation by significantly decreasing PD-1 expression on T cells and increasing IFNγ-producing T cells in the draining lymph nodes and tumors. T cells produced by complemented thymi thus functioned normally in vitro and in vivo. To successfully generate humanized animals by blastocyst complementation, both verification of the function and gene expression profiling of complemented organs/cells in interspecific chimeras will be important in the near future.

Blastocyst complementation is a promising method of using pluripotent stem cells (PSCs) to generate humanized animals for the purpose of establishing novel in vivo models for drug development and the production of organs for transplantation. In this methodology, PSCs are injected into blastocysts that harbor spontaneous or genetically modified mutations causing the loss of target organs or cells. Injected donor PSCs take over the defective host's niche made vacant by the mutation, and they develop into normal organs or cells instead of those with mutation-related deficits. The following organs and cells have thus far been generated successfully through blastocyst complementation: T cells and B cells 1 , lens 2 , hematopoietic cells 3 , pancreas 4-7 , kidney [6][7][8][9] , thymus 6,10,11 , endothelial cells and hematopoietic cells 7,12 , lung 13 , forebrain (neocortex and hippocampus) 14 , and liver 7,15 .
For the successful application of blastocyst complementation to in vivo animal models for drug development and regenerative medicine, the complemented organs must have not only normal anatomical and cytohistological architecture but also normal physiological functions. In pancreas-complemented chimeras

Results
Generation of thymus-complemented mouse chimeras (B6 ESC CAG-EGFP → Foxn1 nu/nu and B6 ESC CAG-AG → Foxn1 nu/nu ). Chimeric mice were generated by injection of enhanced green fluorescent protein (EGFP)-and Azami green (AG)-positive C57BL/6 (B6) embryonic stem cells (ESCs) into KSN/Slc and CD1-Foxn1 nu/nu blastocysts, respectively. The mice varied in nude-hairy or black-white coat color chimerism (i.e., in the contribution of the donor) in appearance. At autopsy, chimeric mice had thymi regardless of the degree of chimerism. It seemed, however, that chimeric mice with lower chimerism in appearance had smaller thymi (n = 8 for B6 ESC CAG-EGFP → Foxn1 nu/nu ; n = 6 for B6 ESC CAG-AG → Foxn1 nu/nu ). Representative examples are shown in Suppl. Fig. 1. Figure 1a indicates the outward appearance and complemented EGFP + thymi of two B6 ESC CAG-EGFP → Foxn1 nu/nu chimeric mice with different coat color chimerism (mice nos. T1-2 and T1-3), compared with those of a nude (Foxn1 nu/nu ) mouse. Flow cytometry (FCM) analysis of TECs (CD45 − EpCAM + ) showed that almost all TECs of the chimeric mice were derived from EGFP + donors (Fig. 1b). In the periphery, we observed T cells (CD3 + ) with various types of chimerism in the chimeric mice. The higher chimerism in appearance (larger hairy and black coat-color area), the higher percentages of donor-derived (EGFP + or AG + ) peripheral T cells (Suppl. Fig. 2). In addition, they possessed CD8 single-positive and CD4 single-positive T cells, as in B6 mice, whereas there were almost no T cells in the nude mice (Fig. 1c). Peripheral T cell numbers of 11 B6 ESC CAG-EGFP → Foxn1 nu/nu chimeric mice were measured to examine relationships between chimerism and T cell numbers. Among 11 chimeras, one chimera showed extremely high number of T cells (~ 8500 /µL) (Suppl. Fig. 3a). Suppl. Fig. 3b depicts a regression line between percentages of EGFP + T cells and peripheral T cell numbers, after excluding the high value judged by Grubbs's test for outliers (P < 0.05). There was no positive or negative correlation between them. Peripheral T cell numbers of chimeras were not significantly different from those of B6 mice (Suppl. Fig. 3c). Age-dependent changes in chimerism of peripheral T cells were examined in 11 chimeric mice between 11 and 42 weeks of age. There were no significant differences in T-cell chimerism between the ages (Fig. 1d).

scRNA-seq analysis of thymic cells of B6 and B6 ESC CAG-EGFP → Foxn1 nu/nu chimeric mice.
To compare the gene expression patterns of complemented thymi of chimeric mice with those of normal mice, we conducted scRNA-seq analysis of whole thymi. After drawing a t-distributed stochastic neighbor embedding (t-SNE) chart, we plotted TEC marker gene expression, namely the expression of epithelial cell adhesion molecule (Epcam) for TECs, thymus-specific serine protease (TSSP; Prss16) for cortical TECs (cTECs), and FEZ family zinc finger 2 (Fezf2) for medullary TECs (mTECs), as well as Egfp expression (Fig. 2a). The three marker genes had specific distributions. Classification based on gene expression similarity resulted in a total of 20 clusters (Fig. 2b), among which cluster 14 was identified to be enriched with TEC gene markers, including interleukin 7 (Il7), thymoproteasome subunit β5t (Psmb11), and autoimmune regulator (Aire), in addition to the abovementioned three marker genes. We therefore regarded cluster 14 as the TEC cluster.
We extracted cells from the TEC cluster to create a heatmap of TEC-related gene expression (Fig. 2c). In the heatmap, the genes were classified into three categories on the basis of the patterns of common TEC, cTEC, and mTEC gene marker expression, along with Egfp expression. We observed correspondence of the origins of TECs (chimera or B6) with Egfp expression (presence or absence), reflecting the FCM results in Fig. 1b that almost all TECs in the chimeras were EGFP positive. In addition, this analysis revealed heterogeneity of TECs from adult chimeric and B6 mice, as reported in mouse embryos (E12.5-E18.5) and newborns (postnatal day 0) 20  In vitro T-cell-proliferation assays using splenic T cells of B6 ESC CAG-EGFP → Foxn1 nu/nu chimeric mice. For functional analysis of T cells produced by complemented thymi, splenic T cells isolated from three chimeric mice with different chimerism were stimulated with anti-CD3/CD28 antibody beads to evaluate the proliferation potential of CD4 + and CD8 + T cells by priming through costimulatory signals. Figure 3a is a representative histogram showing the proliferation of CD4 + and CD8 + T cells of B6 mice. We plotted the proliferation patterns of CD4 + and CD8 + T cells, categorized into EGFP − and EGFP + , in three chimeras (mice nos. T2-2, T2-9, and T2-10) (Fig. 3b) and in B6 mice. Both donor-and host-derived splenic T cells of the chimeras proliferated. The chimerism of splenic T cells from chimeras nos. T2-2, T2-9, and T2-10 after 72 h of stimulation was 87.3%, 10.1%, and 98.8%, respectively-almost the same values of those of splenic T cells before the stimulation (see "Methods" section). Splenic T cells of chimeric mice proliferated upon stimulation with the antibodies in a manner similar to those isolated from B6 mice, regardless of the origin of the T cells (host or donor) (Fig. 3c). www.nature.com/scientificreports/ In vitro T-cell-activation assays using splenic T cells of B6 ESC CAG-AG → Foxn1 nu/nu chimeric mice. Next, to elucidate further the in vitro function of splenic T cells of chimeric mice, we used FCM analysis to investigate the production of cytokines (interferon-γ (IFNγ) and interleukin-2 (IL-2)) and a cytotoxic protease (granzyme B (GzB)) after stimulation with anti-CD3 antibody. In both B6 and chimeric mice, antibodyconcentration-dependent proliferation of CD4 + and CD8 + T cells was observed (Fig. 4a). The mean fluorescence intensity (MFI) of IFNγ + CD4 + , GzB + CD4 + , IFNγ + CD8 + , or GzB + CD8 + T cells was significantly increased in an anti-CD3-antibody-concentration-dependent manner. IL-2 + CD8 + T cells likewise showed a concentrationdependent increase, whereas the MFI of IL-2 + CD4 + T cells peaked at 0.01 or 0.03 μg/mL, showing a bell-shaped change (Fig. 4b).
Tumor transplantation experiments using B6 ESC CAG-AG → Foxn1 nu/nu chimeric mice. First, we monitored the growth of inoculated MC38 cells in B6 and CD1 mice for 21 days. In B6 mice, the tumors grew rapidly and their volumes in all transplanted mice exceeded 2000 mm 3 by day 21 (Fig. 5). In contrast, all CD1 Histograms of CellTrace Violet staining in EGFP − CD4 + and EGFP + CD4 + T cells and EGFP − CD8 + and EGFP + CD8 + T cells derived from B6 ESC CAG-EGFP → Foxn1 nu/nu chimeric mice (nos. T2-2, T2-9, and T2-10). Stimulated, red lines; unstimulated, gray shading. (c) Percentages of divided EGFP − CD4 + and EGFP + CD4 + T cells and EGFP − CD8 + and EGFP + CD8 + T cells derived from chimeric mice, as well as CD4 + and CD8 + T cells of B6 mice, as determined by using CellTrace Violet staining and FlowJo software. www.nature.com/scientificreports/ mice completely rejected the MC38 cells within 14 days (Fig. 5). As we had confirmed the normal functions of T cells produced by complemented thymi in vitro, we next investigated their functions in vivo. For this purpose, we evaluated the antitumor effects of anti-PD-L1 antibody in an MC38-transplantation model. B6 and chimeric mice were injected with MC38 cells (day 0) and then treated with anti-PD-L1 antibody or control IgG (days 7 and 10). Notably, anti-PD-L1 antibody administration seemed to suppress tumor growth in both B6 mice and the chimeras (Fig. 6a). Furthermore, we used FCM to investigate T-cell activation in the draining lymph nodes (dLNs), as well as T-cell-mediated target-cell-killing capacity in tumor-infiltrating lymphocytes (TILs), on day 14 after cancer cell injection. Anti-PD-L1 antibody treatment significantly enhanced T-cell activation by downregulating programmed cell-death 1 (PD-1) expression and upregulating IFNγ production by CD4 + and CD8 + T cells in the dLNs of both B6 and chimeric mice (Fig. 6b). In addition, in CD4 + and CD8 + TILs, we observed www.nature.com/scientificreports/ that IFNγ production was increased and PD-1 expression was decreased by anti-PD-L1 antibody in the chimeras (Fig. 6c). The similar effects of anti-PD-L1 antibody were also detected in B6 mice. In general, anti-PD-L1 antibody had similar effects in terms of both T-cell activation in dLNs and T-cell-killing capacity in TILs in chimeras and B6 mice. Taken together, these findings show that we successfully generated thymus-complemented chimeric mice. We demonstrated that thymus-complemented chimeric mice and normal B6 mice had similar genetic profiles and similar thymic functions, as well as similar thymus-derived peripheral T-cell responses.

Black borders on bars indicate EGFP − T cells and green borders indicate EGFP + T cells. Open boxes indicate CD4 + T cells and color-shaded (gray or green) boxes indicate CD8 + T cells.
Blastocyst complementation technology will be invaluable in generating humanized animals for both research and commercial applications in regenerative medicine and drug development.

Discussion
The thymus is a primary lymphoid organ for the differentiation of thymocytes into mature T cells and maintenance of the pool of functional T cells. TECs are essential in the process of maturation of T cells. cTECs and mTECs play different roles, in positive and negative selection, respectively, in the T-cell maturation process. FOXN1 is a transcriptional factor that is a master regulator of TEC differentiation in the fetal period and of TEC homeostasis in the postnatal stage 16,17 . The gene responsible in nude mice (nu/nu) 18 and nude rats (rnu/rnu) 19 is Foxn1. In nude mouse fetuses, the thymus primordium forms normally and thymic epithelial precursor cells are present. Afterwards, however, the nude thymic anlage never develops to support T cells, and it remains as nonfunctional cystic thymic rudiments. As a result, nude mice are athymic and lack normal T-cell populations 21 .
In terms of the physiological roles of the thymus, thymus complementation by blastocyst complementation can be expected to generate an in vivo unique model for drug development, particularly in the field of cancer immunology. Determining whether or not T cells work functionally in chimeras is important for the application of such a model. Although intra-and interspecific thymus complementation has been reported 6,10,11 , there have been few detailed reports on the functions of T cells generated by complemented thymi. Therefore, we examined the in vitro and in vivo functions of T cells of B6 ESC → Foxn1 nu/nu chimeric mice. In addition, we conducted scRNA-seq analysis of complemented TECs, which control the development and maturation of T cells, to investigate whether their gene expression profile was comparable to that in TECs of B6 mice.
We generated thymus-complemented mouse chimeras in which peripheral T cells with varying chimerism were produced by blastocyst complementation. Previously, Müller et al. 10 indicated that mTECs and cTECs in mESCs → Foxn1 nu/nu chimeric mice were constituted from the progeny of engrafted mESCs by RT-PCR. However, the characteristics of complemented mTECs and cTECs are not proved yet. In this study, scRNA-seq analysis of total thymic cells showed similar t-SNE clustering patterns in B6 and chimeric mice, suggesting that there was normal development of complemented thymi, including T cells, at the gene expression level. The heatmap for cells in the TEC clusters of normal and chimeric mice demonstrated that the TECs of the chimera were derived from donor B6 ESCs. Therefore, we demonstrate that complemented TECs not only express TECs' marker genes, but also show heterogeneity comparable to wild-type mice. TECs express several genes that are prerequisites for T-cell selection and maturation. For example, Psmb11 and Prss16 are expressed for the positive selection of CD8 +22 and CD4 + T cells 23 , respectively, in cTECs. Aire is expressed for the negative selection of T cells in mTECs 24 . We confirmed that these TEC representative genes were expressed in similar ways in B6 and chimeric mice. TEC gene expression of the chimera, comparable to that in B6, could thus indicate normal differentiation of thymocytes and T cells in complemented thymi. As mentioned above, TECs control T-cell development from immature thymocytes to mature T cells. Conversely, thymocytes control TEC expansion and differentiation. These bidirectional interactions between thymocytes and TECs are referred to as thymic crosstalk 25,26 . For www.nature.com/scientificreports/ example, RANKL (receptor activator of nuclear factor kappa-Β ligand; tumor necrosis factor ligand superfamily member 11, Tnfsf11) provided by positively selected CD4 + thymocytes regulates mTEC differentiation via RANK (Tnfrsf11a) expressed on mTECs 27,28 . Therefore, the normal gene expression profiles of TECs of the chimera suggest that there is normal development of thymocytes and T cells.
To investigate the functions of T cells generated by complemented thymi in chimeras, we first assessed splenic T-cell proliferation in chimeras after stimulation with anti-CD3/CD28 antibodies, in comparison with that in B6-derived T cells. We observed comparable proliferation in chimera-and B6-derived T cells, as well as in donor (EGFP + Foxn1 +/+ )-and host-derived T cells (EGFP − Foxn1 nu/nu ), suggesting that Foxn1 genotypes do not affect the function of T cells. There was also comparable production of IFNγ, IL-2, and GzB by splenic T cells stimulated with anti-CD3 antibody in chimeras and B6 mice. Notably, anti-CD3 antibody induced T-cell-receptor-mediated proliferation and activation in a concentration-dependent manner. Collectively, these findings indicated that the development of thymus-derived T lymphocytes works normally in thymus-complemented chimeric mice created by blastocyst complementation. The host immune system is one of the most important elements for protection from tumor development and control of tumor growth. Tumor growth is controlled mainly by T lymphocytes. CD8 + cells are regarded as being a uniform population of cells that largely and quickly secret IFNγ and a protease, GzB, which can kill tumor cells. In addition, antigen-specific CD4 + T cells showing a Th1 IFNγ-producing phenotype were detected against tumor growth; this was identified recently as an additional mechanism for controlling tumor growth 29 . T cells become activated through both antigen-receptor signaling and CD28 costimulatory signaling. During activation, these T cells express PD-1, which interacts with its ligand PD-L1, resulting in the inhibition of T-cell proliferation and activation. In the tumor microenvironment (TME), PD-1 is expressed mainly on activated T cells, whereas PD-L1 is expressed on several types of tumor cells and antigen-presenting cells. Under these Percentages of PD-1 + CD4 + and PD-1 + CD8 + TILs and IFNγ + CD4 + and IFNγ + CD8 + TILs were compared between the control and anti-PD-L1 antibody treatments. Values were analyzed by F-test, followed by an unpaired two-tailed Student's t-test or two-tailed Welch's t-test between treatments. *, **, and *** indicate P < 0.05, P < 0.01, and P < 0.001, respectively. Means ± S.E.M. n = 4 for each group of chimeras; n = 5 for each group of B6 mice. www.nature.com/scientificreports/ circumstances, tumor cells escape from host immunity by utilizing the immune checkpoint system. Preclinical and clinical studies have demonstrated the efficacy of immune checkpoint inhibitors (that is, anti-PD-1 antibody and anti-PD-L1 antibody) against several cancers. Currently, several antibodies have been marketed (e.g., nivolumab and pembrolizumab against PD-1; atezolizumab and avelumab against PD-L1), and antibodies against the PD-1 pathway are revolutionizing cancer immunotherapy 30,31 .
To assess the tumor-inhibitory effect of T cells in thymus-complemented chimeric mice, we evaluated the pharmacological efficacy of anti-PD-L1 antibody on B6-derived tumors, syngeneic to the donor, implanted into B6 ESC → Foxn1 nu/nu chimeras, in consideration of the future application of thymus-complemented models to cancer immunology. Anti-PD-L1 antibody administration suppressed tumor growth in the chimeras and resulted in the increased abundance of IFNγ + CD8 + and IFNγ + CD4 + T cells and decreased abundance of PD-1 + CD8 + and PD-1 + CD4 + T cells in dLNs. The same phenomena were also observed in TILs. We postulate that the immune checkpoint interaction between PD-1 on T cells generated by complemented thymi and PD-L1 on tumor cells worked normally. In addition, the activation of CD4 + T cells by anti-PD-L1 antibody in the chimeras suggested that TSSP (PRSS16) in complemented TECs functions normally, because TSSP is crucial for the antitumoral role of CD4 + T cells 32 . This speculation was supported by our finding of normal expression of Prss16 in cTECs of the chimera by the scRNA-seq analysis described above.
Here, we generated thymus-complemented chimeric mice by blastocyst complementation-that is, by injection of B6 ESCs into Foxn1 nu/nu blastocysts-and characterized the mice's T cells in vitro and in vivo. We also performed gene expression profiling of their complemented TECs. We showed that T cells produced by complemented thymi worked normally, as they did in B6 mice. We speculate that the normal function of the chimeras' T cells was reflected by the normal gene expression patterns of the TECs, which were the targets of complementation in the chimeras. It was reported that T cell lymphopenia and reduced proliferative responses of T cells to mitogens in patients with thymic hypoplasia, such as DiGeorge syndrome 33,34 . As we observed that a tendency that chimeras with lower chimerism had smaller thymi, it is likely that the chimerism has an effect on T cell numbers and functions. Our results showed, however, that the chimerism did not affect peripheral T cell numbers. The chimerism of the chimeric mice used as a tumor-engrafted model was extremely low, but T cells of the chimeras functioned normally comparable to those of B6 mice in vivo. Taken together with these findings, it is shown that the chimerism of thymus-complemented mice affects the thymus size, but not T cell numbers or functions.
For oncogenesis research and anticancer drug development, humanized mice (e.g., human hematopoietic stem cell (HSC)-engrafted immunodeficient mice) have been utilized and bettered (e.g., transgenic for human leukocyte antigen (HLA)). However, there are some limitations of the models, such as occurrence of graftversus-host disease after HSC engraftment, difficulty of construction of the TME 35 . In the principle of blastocyst complementation, it has the potentials to generate novel types of humanized mouse models which can break the current limitations. Thus, we could generate a novel model in which cells involved in the TME (e.g., endothelial cells, macrophages, and adipocytes 36 ) are also complemented, in addition to T cells and TECs, by using hosts harboring Foxn1 nu/nu and (an)other mutant gene(s). For example, it is presumed that we may produce chimeric mice which have not only donor-derived vascular endothelial cells and myeloids, but also donor-derived T cells selected by donor-derived TECs, by using double Foxn1 nu/nu and Flk-1 −/− mutant hosts 12 . The results of this study are fundamental for the future application of blastocyst complementation to cancer immunology.
Blastocyst complementation technology will advance the generation of humanized animals for research and commercial applications in regenerative medicine and drug development. To achieve this, validation of the function and gene expression patterns, which are influenced by xenogeneic barriers (e.g., ligand-receptor incompatibility, cell adhesion incompatibility), of complemented organs and cells of interspecific chimeras through comparison with normal donor animals or humans will be important in the future.

Methods
Mice. KSN/Slc (Foxn1 nu/nu ) and C57BL/6NCrSlc mice were obtained from Japan SLC (Hamamatsu, Shizuoka, Japan). C57BL/6NCrl, C57BL/6J, Crlj:CD1-Foxn1 nu (CD1-Foxn1 nu/nu ), and Crl:CD1 (CD1) mice were purchased from Charles River Laboratories Japan (Yokohama, Kanagawa, Japan). MCH/ICR mice were purchased from CLEA Japan Inc. (Tokyo, Japan). The C57BL/6 substrains used in this study are abbreviated as B6 in the text. Animal care and experimental procedures were performed in animal facilities accredited by the Health Science Center for Accreditation of Laboratory Animal Care and Use of the Japan Health Sciences Foundation. All protocols were approved by the Institutional Animal Care and Use Committee of Eisai Co., Ltd. (Tokyo, Japan) and KAN Research Institute, Inc. (Kobe, Hyogo, Japan), and of the Institute of Medical Science, University of Tokyo (Tokyo, Japan). All methods were performed in accordance with relevant guidelines and regulations, and the study complied with ARRIVE guidelines (Animal Research: Reporting of In Vivo Experiments).

Generation of mouse ESC lines, and cell culture. A B6 ESC CAG-EGFP line was derived from blastocysts
obtained from C57BL/6N female mice mated with C57BL/6N-Tg male mice (CAG-EGFP) (Japan SLC) and maintained as previously described 5 .
Rosa26-CAG-fNeo-H2B-AG-dNeo, a Cre-reporter mouse ESC line, was generated by gene targeting. A reporter gene cassette containing the following components was constructed in a Bluescript SK + (Stratagene, Santa Clara, CA, USA) backbone: CAG promoter-LoxP-Neo-polyA (as a stop cassette)-LoxP-histone H2B (H2B)-AG-polyA. The Rosa26 5ʹ and 3ʹ arms (1.7 kb and 6 kb, respectively) were amplified with the following primers and then assembled with the reporter gene cassette to construct the targeting vector: www.nature.com/scientificreports/ GG-3ʹ/5ʹ-GAG GGT ACC CCC TGA CAA AAG GGA TGC CCA ATT CC-3ʹ. A knock-in allele was generated by homologous recombination in the B6J-S1 UTR ESC line 37 in accordance with standard procedures. To generate the Rosa26-CAG-H2B-AG line constitutively expressing H2B-AG (ESC CAG-AG ), a Cre expression vector was transfected into Rosa26-CAG-fNeo-H2B-AG ESCs to delete the LoxP-flanked Neo cassette. The ESC CAG-AG cell line was cultured in accordance with reference 37 .

Generation of chimeric mice by blastocyst complementation. To generate B6
ESC CAG-EGFP → Foxn1 nu/nu chimeric mice, three to five B6 ESC CAG-EGFP cells were injected into each KSN/Slc morula embryo. These embryos were cultured in N2B27 medium for 24 h up to the blastocyst stage and then transferred into the uteri of pseudopregnant ICR female mice (2.5 days post-coitum). Female CD1-Foxn1 nu/nu mice were superovulated by intraperitoneal injection of HyperOva (0.1 mL; Kyudo, Tosu, Saga, Japan), followed by intraperitoneal injection of 7.5 IU human chorionic gonadotrophin (hCG) 48 h later. After being injected with the hCG, the female mice were mated with male CD1-Foxn1 nu/nu mice. Foxn1 nu/nu mouse embryos were collected in M2 medium (Sigma-Aldrich, St. Louis, MO, USA) at the two-cell stage. They were then transferred into KSOM medium (ARK Resource, Kumamoto, Japan) and cultured for 48 h. Chimeric embryos were generated by microinjection of B6 ESC CAG-AG cells into Foxn1 nu/nu blastocysts. An Preparation of transcript count data. The single-cell RNA-seq data derived from FASTQ files were processed with CellRanger software (10× Genomics, version 3.1.0) in accordance with the manufacturer's instructions. Mouse genome data (refdata-cellranger-mm10-3.0.0.tar.gz, corresponding to GRCm38.93 in NCBI; https:// suppo rt. 10xge nomics. com/ single-cell-gene-expre ssion/ softw are/ relea se-notes/ build# mm10_3. 0.0) and EGFP sequence data (http:// radia tion-japan. info/ pdf/ green_ mouse) were used as the reference genome to generate matrix files containing cell barcodes and transcript counts.
Normalization and visualization of transcript count data. Normalization and visualization of the transcript count data were performed by using R (version 3.6.0; R Foundation for Statistical Computing, Vienna, Austria) and the Seurat package (https:// satij alab. org/ seurat/, version 3.1.4). The transcript data were normalized to obtain nUMI (normalized unique molecular identified) counts and were scaled by using the NormalizeData and ScaleData functions of the Seurat package. Highly variable genes and significant principal components across the normalized count data were identified and extracted for subsequent analysis by using either the Find-VariableFeatures or the RunPCA function of the Seurat package. To illustrate the similarities between cells, dimensionality reduction was performed on the data with the t-SNE algorithm implemented in the RunTSNE function of the Seurat package. The single-cell data were plotted in two-dimensional space on the basis of the reduced matrix generated by the t-SNE algorithm and the expression levels of tissue-specific marker genes.

Identification of thymus epithelial cluster and visualization of marker expression. Cells were
classified into 20 clusters according to the similarity of their expression patterns by using the FindNeighbors and FindClusters functions of the Seurat package. We mapped the expression of TEC gene markers (Epcam, Prss16, and Fezf2) into the t-SNE chart and identified a cluster enriched with TECs. The expression data of the cells included in the thymus epithelial cluster were extracted, and the gene expression values of either TEC markers (Ciita, Epcam, Foxn1, H2-Aa, H2-Ab1, and H2-Eb1), or cTEC markers (Ccl25, Cd83, Ctsl, Cxcl2, Dll4, Enpep, Il7, Ly75, Prss16, Psmb11, Psmb8, and Tbata), or mTEC markers (Aire, Ccl19, Ccl21a, Cd40, Cd80, Cldn3, Cldn4, Fezf2, Il4, Krt5, Skint1, Tnfrsf11a, and Xcl1), as well as of Egfp, were chosen for visualization by heatmap. Gene expression levels were indicated in scaled nUMI counts, which were standardized to a mean of 0 and standard deviation of 1 in individual genes. Hierarchical clustering was performed by using the hclust function implemented in the fastcluster module (https:// cran.r-proje ct. org/ web/ packa ges/ fastc luster/, version 1.1.25) with Ward's linkage algorithm and the Manhattan distance metric. The heatmap figure was illustrated by using the heatmap function in the NMF (Non-negative Matrix Factorization) package (https:// cran.r-proje ct. org/ web/ packa ges/ NMF/, version 0.22.0). Transplantation of B6-derived cancer cells into B6 and CD1 mice. Ten CD1 and six C57BL/6NCrl mice were injected subcutaneously into the right flank with 1 × 10 6 MC38 cells in 100 μL of HBSS (Gibco). Tumor sizes were measured 6, 10, 14, and 21 days after the injection; their volumes were calculated as length × width 2 × 0.5 (mm 3 ). When the tumor volume was > 2000 mm 3 , the mice were euthanized by cervical translocation.

FCM.
To analyze peripheral cells, blood was collected from the caudal vein, mixed with saline (Otsuka Pharmaceutical, Tokyo, Japan) containing heparin (10 units/mL; AY Pharmaceuticals, Tokyo, Japan), and treated with BD Pharm Lyse Lysing Buffer to remove erythrocytes. Thymi were minced and digested at 37 °C for 30 min with RPMI 1640 (FUJIFILM Wako) containing 0.2 mg/mL Liberase TM (Sigma-Aldrich) and 0.2 mg/mL DNase I (Roche). After filtration through a 70-µm-mesh cell strainer, cells were incubated in 1 mL of Pharm Lyse Lysing Buffer to remove red blood cells. For FCM of TILs, tumor tissues were minced into 2-to 3-mm pieces and digested by using a Tumor Dissociation Kit (mouse) (Miltenyi Biotec) and a gentleMACS Octo Dissociator (Miltenyi Biotec). After the removal of red blood cells by using BD Pharm Lyse Lysing Buffer, TILs were isolated with CD45 (TIL) MicroBeads (mouse) (Miltenyi Biotec) and a MiniMACS Separator (Miltenyi Biotech). Isolated 1.5 × 10 6 TILs were treated with 13 ng/ mL phorbol 12-myristate 13-acetate (Sigma-Aldrich), 333 ng/mL ionomycin (Sigma-Aldrich), and BD GolgiStop (BD Biosciences) at 37 °C for 4 h. After this stimulation, TILs were incubated with Mouse BD Fc Block (BD Biosciences) for 10 min, followed by staining with antibodies against cell-surface markers. Next, TILs were fixed with 4% paraformaldehyde phosphate buffer solution (FUJIFILM Wako), permeabilized by using BD Perm/Wash Buffer (BD Biosciences), and then stained with antibodies against intracellular cytokines. In addition, cells were isolated from dLNs for FCM. They were blocked and stained as described above. Stained cells were analyzed by using an SH800 cell sorter (Sony, Tokyo, Japan), a BD LSRFortessa Flow Cytometer (BD Biosciences), or a BD FACSymphony Flow Cytometer (BD Biosciences) with FlowJo software. Peripheral T cell numbers in blood were calculated by using CountBright Absolute Counting Beads for flow cytometry (ThermoFisher Scientific) according to its protocol. Antibodies used are listed in Supplementary Table 1. Statistical analysis. Graph drawing and statistical analysis were performed by using GraphPad Prism version 9.0.2 (GraphPad Software, San Diego, CA).