We previously reported a method to generate dendritic cell (DC)-like antigen-presenting cells (APC) from human induced pluripotent stem (iPS) cells. However, the method is relatively complicated and laborious. In the current study, we attempted to establish a method through which we could obtain a large number of functional APC with a simple procedure. We transduced iPS cell-derived CD11b+ myeloid cells with genes associated with proliferative or anti-senescence effects, enabling the cells to propagate for more than 4 months in a macrophage colony-stimulating factor (M-CSF)-dependent manner while retaining their capacity to differentiate into functional APC. We named these iPS cell-derived proliferating myeloid cells ‘iPS-ML’, and the iPS-ML-derived APC ‘ML-DC’. In addition, we generated TAP2-deficient iPS cell clones by zinc finger nuclease-aided targeted gene disruption. TAP2-deficient iPS cells and iPS-ML avoided recognition by pre-activated allo-reactive CD8+ T cells. TAP2-deficient ML-DC expressing exogenously introduced HLA-A2 genes stimulated HLA-A2-restricted MART-1-specific CD8+ T cells obtained from HLA-A2-positive allogeneic donors, resulting in generation of MART-1-specific cytotoxic T lymphocyte (CTL) lines. TAP-deficient iPS-ML introduced with various HLA class I genes may serve as an unlimited source of APC for vaccination therapy. If administered into allogeneic patients, ML-DC with appropriate genetic modifications may survive long enough to stimulate antigen-specific CTL and, after that, be completely eliminated. Based on the present study, we propose an APC-producing system that is simple, safe and applicable to all patients irrespective of their HLA types.
An efficient means for the activation of cytotoxic T lymphocytes (CTL) reactive to tumor antigens is crucial for T-cell-mediated anti-tumor immunotherapy. Dendritic cells (DC) are specialized to control T-cell responses, and many clinical trials of anti-cancer therapies with DC have been conducted.1, 2, 3 In most cases, monocytes obtained from patients by apheresis have been used as a source of DC. However, the number of monocytes obtained from peripheral blood, the potential of the monocytes to differentiate into DC and the quality of the resulting DC vary depending on the patients. It is sometimes difficult to generate a sufficient number and quality of DC, especially when the donor is a patient with advanced cancer. In addition, isolation of monocytes and differentiation culture to generate DC need to be conducted individually for each patient. These limitations of cellular sources represent an obstacle that prevents wider application of DC therapies.
Embryonic stem (ES) cells and the recently developed induced pluripotent stem (iPS) cells4, 5 have the potential to differentiate into many types of cells, and also exhibit an unlimited propagation capacity. We expected that the above-described issues related to the supply of DC for therapeutic use may be resolved, if pluripotent stem cells could be used as the source for DC. Several groups, including ours, have established methods to generate DC from ES cells or iPS cells.6, 7, 8, 9, 10, 11, 12, 13, 14 In addition, using mouse models, we have demonstrated the usefulness of genetically modified pluripotent stem cell-derived DC in the induction of anti-cancer immunity.15, 16
However, our method for differentiation induction is relatively complicated and time-consuming, taking more than 1 month to obtain DC from human pluripotent stem cells.12, 17 The methods reported by other groups seem to be similar in this respect.18 In addition, established methods may be difficult or impossible to apply for mass production by automated cell processing systems. Furthermore, although the iPS cell-technology resolved the major obstacles associated with the clinical application of ES cells, that is, ethical issues and histoincompatibility between ES cell lines and patients, the generation of patient-specific iPS cells is also time-consuming and expensive.
In the current study, we attempted to develop a simpler and more efficient system to generate DC from human iPS cells. We made particular efforts to increase the yield of DC generated from human iPS cells. We found that myeloid cells derived from human iPS cells are different from human peripheral blood monocytes in that they become capable of long-term proliferation after the introduction of a few genes related to cell proliferation or the inhibition of senescence. Importantly, the resulting proliferating myeloid cell lines retained their capacity to differentiate into potent antigen-presenting cells (APC), similar to DC. In addition, to avoid the laborious and expensive generation of patient-specific iPS cells and subsequent differentiation cultures, we propose a system to generate APC from TAP-deficient iPS cells. The system will be applicable to all patients, irrespective of their HLA types.
Generation of proliferating myeloid cells from human iPS cells
We previously established a method of differentiation culture to generate DC from human iPS cells.17 Using this method, the differentiation of iPS cells into DC (iPS-DC) occurred through CD43+CD11b+ myeloid precursor cells (iPS-MC). These iPS-MC, which possess the capacity to differentiate into DC upon treatment with granulocyte-macrophage colony-stimulating factor (GM-CSF) and interleukin (IL)-4, could proliferate in the presence of GM-CSF and M-CSF for 7–14 days, but then ceased to grow. We expected that, if we could confer iPS-MC with long-term proliferative capacity, we would be able to generate a large number of DC from the cells. For this purpose, we introduced several genes with the potential to induce cell proliferation using lentiviruses. As a consequence, we found that simultaneous introduction of cMYC, together with BMI1, MDM2 or EZH2, resulted in the continuous proliferation of iPS-MC for more than 3 months, with a doubling time of 24–36 h. We named the iPS-MC-derived long-term proliferating cells iPS-ML (iPS cell-derived myeloid cell line).
Figure 1a shows the morphology of iPS-MC and iPS-ML. Both these cell lines were floating or weakly adherent, and expressed a leukocyte marker, CD45, and myeloid markers, including CD11b and CD33 (Figure 1b). They also expressed a monocyte marker, CD14. In contrast to iPS-MC, iPS-ML proliferated for more than 3 months, and their proliferation was dependent on the addition of M-CSF (Figure 1c).
Differentiation of iPS-ML into DC-like cells
iPS-ML expressed CD86, HLA class I and HLA class II (Figure 2a). When iPS-ML were cultured in the presence of IL-4 for 3 days, the expression of CD86, HLA class I and HLA class II increased. The cells exhibited DC-like morphology with protrusions, and many of the cells formed clusters (Figure 2b). We named the iPS-ML-derived DC-like cells ML-DC. Treatment of ML-DC with tumor necrosis factor (TNF)-α or OK432 (penicillin-killed Streptococcus pyogenes) for 2 days further increased the expression of HLA class II. The production of IL-12p70 by ML-DC was induced by stimulation with OK432, a potent inducer of DC maturation.19 Lipopolysaccharide (LPS) also induced the production of IL-12 by ML-DC, but TNF-α did not (Figure 2c). Collectively, iPS-ML could differentiate into ML-DC with DC-like morphology, appropriate surface molecules and the ability to produce IL-12.
The capacity of human iPS-ML and ML-DC to stimulate naive T cells was examined based on their stimulation of allogeneic T cells (Figure 3a). Upon co-culture, iPS-ML and ML-DC induced significant proliferation of allogeneic T cells. TNF-α-treated ML-DC stimulated allogeneic T cells more potently than iPS-ML or untreated ML-DC.
The presentation of an antigenic peptide in the context of HLA class II by ML-DC was examined. The iPS cells we used were positive for the HLA-DRB4*01:03 gene encoding the β chain of the HLA-DR53 molecule, and thus, DR53-restricted antigen presentation by ML-DC was examined. We used a glutamic acid decarboxylase 65 (GAD65)-specific DR53-restricted human T-cell clone, SA32.5,20 as the responder T cells. As shown in Figure 3b, ML-DC pre-loaded with DR53-restricted glutamic acid decarboxylase 65 (GAD65) peptide specifically induced a proliferative response in the SA32.5 T cells.
The stimulation of antigen-specific autologous CD8+ T cells by ML-DC was also examined. ML-DC, positive for HLA-A*24:02, were loaded with an HLA-A*24:02-binding cytomegalovirus (CMV) pp65-derived peptide and co-cultured with autologous peripheral blood T cells for 9 days. To analyze the frequency of CMV peptide-specific T cells, the cultured T cells were stained with the tetramer of the HLA-A*24:02-CMV peptide complex. In T-cell samples before culture or after co-culture with ML-DC without the peptide, no T cells were detected with the HLA-peptide tetramer (Figure 3c). On the other hand, after co-culture with peptide-loaded ML-DC, 9.96% of the CD8+ T cells were positively stained with the HLA-peptide tetramer. In addition, the expanded CMV-specific CD8+ T cells were functional in that they produced IFN-γ upon stimulation with the peptide (Figure 3d). These results indicate that peptide-loaded ML-DC stimulated specific T cells in an autologous T-cell population and induced their expansion.
Therefore, the ML-DC generated through this method not only exhibited morphology and surface phenotype similar to DC, but also functioned as potent T-cell stimulators.
Targeted disruption of the TAP2 gene in iPS cells
As described above, the iPS-ML can propagate while retaining their capacity to differentiate into functional APC. Thus, we expected that iPS-ML can be used as an unlimited cell source to generate APC for vaccination therapy in the future. However, the generation of patient-specific iPS cells and subsequent generation of patient-specific iPS-ML may take a long time, presumably more than 2 months. In addition, individual cell processing, including the generation of patient-specific iPS cells, may be too laborious and costly to be applied for clinical medicine. We thus sought to resolve the issue of histoincompatibility between patients and allogeneic iPS cells.
As a means to overcome histoincompatibility, we considered modification of the cell surface HLA class I. TAP, a heterodimer composed of TAP1 and TAP2 molecules, transports antigenic peptides from the cytoplasm to the endoplasmic reticulum, and is essential for the cell surface expression of HLA class I.21 We decided to disrupt the TAP genes in human iPS cells by homologous recombination.
We attempted to generate TAP1- or TAP2-deficient human iPS cells by gene targeting. Although there have been some reports of targeted gene disruption in human ES cells,22 the efficiency of homologous recombination in human pluripotent stem cells is far lower than that in mouse ES cells. Our attempts to disrupt the TAP genes in human iPS cells with conventional gene targeting methods were unsuccessful. Therefore, to improve the efficiency of gene targeting, we used a zinc finger nuclease (ZFN), which can create a double-stranded DNA break at a specific genomic site.23 Figure 4a depicts our strategy used for TAP2 gene disruption in human iPS cells. We introduced the the targeting vector, together with ZFN constructs designed to cut exon 2 of the TAP2 gene, into the iPS cells. After the introduction of these vectors by electroporation, the iPS cells were cultured in the presence of a selection drug, G418. After culturing for 2 weeks, we isolated the iPS cell clones resistant to G418, extracted their DNA, and analyzed the TAP2 loci by PCR. PCR analysis revealed that 46 clones were TAP2+/− and 3 were TAP2−/− among the 153 isolated iPS cell clones, indicating that the efficiency of homologous integration of the targeting vector was 32% (Figure 4b).
Escape of TAP2-deficient iPS cells and iPS-MC from recognition by allo-reactive CD8+ T cells
We next analyzed the cell surface expression of HLA class I molecules in parental iPS cells and TAP2-deficient iPS cells. As shown in Figure 5a, parental iPS cells expressed HLA class I. On the other hand, the expression of cell-surface HLA class I was barely detected in TAP2-deficient iPS cells.
We also tested whether TAP2-deficient iPS cells could avoid recognition by allogeneic CD8+ T cells. To prepare an allo-reactive CD8+ T cell line, we stimulated peripheral blood CD8+ T cells obtained from an HLA-mismatched allogeneic donor with parental iPS cell-derived iPS-DC for 7 days. Then, we cultured the established allo-reactive CD8+ T cell line with parental (TAP2-intact) or TAP2-deficient iPS cells. The T-cell response to the iPS cells was assessed based on the production of IFN-γ by the T cells. As shown in Figure 5b, when parental iPS cells were used as stimulators, a significant amount of IFN-γ was produced by the T cells. On the other hand, very little IFN-γ was produced by the T cells upon co-culture with TAP2-deficient iPS cell clones. These results indicate that TAP2-deficient human iPS cells with a very low level of cell surface HLA class I evaded recognition by allo-reactive CD8+ T cells.
Subsequently, we subjected TAP2-deficient iPS cells to differentiation culture to generate TAP2-deficient iPS-MC. The response of allo-reactive CD8+ T cells to TAP2-deificient iPS-MC was analyzed by ELISPOT analysis to detect IFN-γ-producing cells. The number of allo-reactive CD8+ T cells responding to TAP2-deficient iPS-MC was far lower than the number of cells responding to TAP2-intact iPS-MC (Figure 6a). We then transduced expression vectors for cMYC and EZH2 simultaneously into iPS-MC using lentiviruses in order to generate TAP2-deficient iPS-ML. The absence of TAP2 protein in TAP2-deficient iPS-ML was confirmed by western blot analysis (Figure 6b). Figure 6c shows the expression of HLA class I in wild-type and TAP2-deficient iPS-ML. Although the expression level was less than that detected in the wild-type iPS-ML, significant expression of HLA class I was observed, even in TAP2-deficient iPS-ML. Nevertheless, the response of allo-reactive CD8+ T cells to TAP2-deificient iPS-ML was far lower than the response to wild-type iPS-ML (Figures 6d and e).
The reduced response of allo-reactive T cells to TAP2-deficient iPS-ML may be owing to the reduced complexity of the peptides presented by HLA class I molecules as well as their reduced cell surface expression. In TAP-deficient cells, HLA class I molecules present only peptides that originated as signal peptides from membrane proteins or secreted proteins,24 thus resulting in very small variations in the peptides presented on HLA class I molecules. Therefore, the frequency of T cells that can react to TAP-deficient allogeneic target cells is far lower than those that are reactive to TAP-intact allogeneic cells, accounting for the reduced response of allo-reactive T cells to TAP-deficient iPS-ML, as shown in Figures 6d and e.
Stimulation of antigen-specific CD8+ T cells derived from allogeneic donors by TAP2-deficient ML-DC
In general, the stimulation of T cells specific to a certain antigenic peptide by using allogeneic APC is considered to be difficult, even if the APC and donor of the T cells share the restricting HLA class I. It is anticipated that the activation of antigen-specific T cells would be disturbed by a large population of allo-reactive T cells stimulated by the mismatched HLA expressed by allogeneic APC. As described above, the response of allo-reactive CD8+ T cells to TAP2-deificent APC was far lower in magnitude than the response to wild-type APC. Thus, by using TAP2-deficient iPS-DC, it may be possible to stimulate T cells specific to a particular peptide from CD8+ T cells prepared from the peripheral blood mononuclear cells obtained from allogeneic donors.
We attempted to establish an HLA-A2-restricted MART-1-specific CTL line from peripheral blood CD8+ T cells derived from HLA-A2-positive allogeneic donors by stimulation with TAP2-deificient ML-DC. Because the iPS cells were derived from an HLA-A2-negative donor (HLA-A*24:02/11:01), we introduced expression vectors for HLA-A*02:01 or 02:06 into TAP2-deficient iPS-ML using lentiviruses. We then induced the resulting TAP2-deficient iPS-ML with HLA-A*02:01 or 02:06 to differentiate into ML-DC.
Peripheral blood CD8+ T cells obtained from an HLA-A*02:01-positive individual were co-cultured with MART-1-peptide-loaded TAP2-deficient ML-DC expressing transgene-derived HLA-A*02:01. Peptide-loaded ML-DC were repeatedly added at 7-day intervals. After the co-culture, cells were recovered, and the proportion of CD8+ T cells specific to A*02:01/MART-1 was analyzed by staining with an HLA-peptide dextramer. As shown in Figure 7a, the number of T cells with a MART-1-specific T cell receptor (TCR) was almost negligible (0.06%) in unstimulated CD8+ T cells. The frequency increased to 0.98% after the second stimulation with peptide-loaded DC, and further increased (2.97%) after the third stimulation.
Next, T cells were analyzed for their reactivity to the MART-1 peptide. To do this, the T cells were co-cultured with MART-1 peptide-pulsed HLA-A*02:01-positive T2 cells for 16 h, and then T cells producing IFN-γ were counted using an ELISPOT analysis. As shown in Figure 7b, CD8+ T cells showed a significant MART-1-peptide-specific response, thus indicating that we successfully stimulated MART-1-specific T cells using TAP2-deficient, HLA-A*02:01-introduced ML-DC.
In addition, using TAP2-deficient, HLA-A*02:06-expressing ML-DC as stimulators, we were able to establish a MART-1-specific HLA-A*02:06-restricted CD8+ T cell line from peripheral blood CD8+ T cells isolated from an HLA-A*02:06-positive allogeneic donor (Figure 7c). These results indicate that we can generate CD8+ T cells specific for a certain peptide from the CD8+ T cells derived from allogeneic donors using TAP2-deficient ML-DC.
Our previous studies investigated the induction of anti-cancer immunity using semi-allogeneic ES cell-derived DC.25 The injection of mice with antigenic peptide-loaded semi-allogeneic DC sharing one MHC haplotype with the recipients resulted in priming of peptide-specific CTL restricted by the shared MHC class I molecule.25 However, the efficiency of immunization by semi-allogeneic DC was lower than that by DC genetically identical to the recipients, probably because of the rapid elimination of the transferred DC by allo-reactive CTL of the recipients in the semi-allogeneic setting.26 Loyer et al.27 studied the target histocompatibility antigens recognized by the immune system of recipients upon in vivo transfer of allogeneic APC. They concluded that elimination of the transferred allogeneic APC was mainly mediated by CTL reactive to allogeneic MHC class I, but not MHC class II or minor antigens. It has also been reported that the efficiency of immunization by DC in vivo is significantly reduced by pre-existing CTL reactive to the DC.28
On the basis of these findings, we expected that, if the expression of the intrinsic MHC class I by allogeneic DC is inhibited, DC inoculated into recipients could escape from the attack by allo-reactive T cells in the recipients, resulting in prolonged survival of the transferred DC and improved immunization efficiency. In a subsequent study, we evaluated the effects of disruption of the genes encoding β2M or TAP as a means of blocking the expression of MHC class I and evading recognition by allo-reactive CD8+ T cells.29 We generated DC from mouse ES cells deficient for β2M or TAP1, and observed that DC with modification in these MHC class I-related genes could avoid recognition by allo-reactive CD8+ T cells in vitro. Furthermore, upon transfer into allogeneic mice, TAP1-deficient and β2M-deficient DC showed a survival advantage compared with wild-type DC.
In the present study, we examined the possibility of in vitro stimulation of antigen-specific human CD8+ T cells derived from allogeneic donors by TAP-deficient human iPS cell-derived DC. We disrupted the TAP2 gene in iPS cells from fibroblasts obtained from an HLA-A*24:02/11:01 donor. We selected the human TAP2 gene as the target for gene disruption in the present study, because the design of ZFN for the TAP1 gene was difficult. Subsequently, we generated a myeloid cell line (iPS-ML) from TAP2-deficient iPS cells, and then introduced the HLA-A*02:01 or 02:06 gene into these TAP-deficient iPS-ML using a lentivirus vector. We induced the differentiation of TAP2-deficient-HLA-A2 iPS-ML into DC (ML-DC), and we observed that the resulting TAP-deficient ML-DC expressing HLA-A*02:01 and 02:06, stimulated activation of MART-1-specific CD8+ T cells derived from allogeneic donors positive for HLA-A*02:01 and 02:06, respectively. These results suggest that we can use TAP-deficient ML-DC introduced with recipient-matched HLA class I for cellular vaccination to stimulate CD8+ T cells specific to a certain antigen restricted by the introduced HLA class I.
Even TAP-deficient cells express significant levels of HLA class I, as indicated in Figure 6c. However, HLA class I molecules in TAP-deficient cells present only TAP-independent peptides, and most of these are signal peptide-derived peptides. Therefore, the diversity of the peptides presented by the HLA class I in TAP-deficient cells is much smaller than that of TAP-intact cells, and the frequency of allogeneic CD8+ T cells reactive to TAP-deficient cells in the T-cell repertoire is far lower than the frequency of those reactiv to TAP-intact cells. In short, TAP-deficiency reduced the antigenicity of HLA class I, not only by lowering the level of cell surface expression of HLA class I, but also by reducing the diversity of peptides presented by HLA class I.
TAP-deficient ML-DC express intrinsic HLA class II and presumably cause an allo-reactive response of CD4+ T cells when administered to patients. It is possible that the activated allo-reactive CD4+ T cells may provide help in the stimulation of CTL. Regarding the effect of in vivo transfer of TAP-deficient DC, we have previously observed efficient priming of antigen-specific CD8+ T cells by TAP-deficient allogeneic ES cell-derived DC in vivo, using a mouse model system.29
Based on the results of the present study, we propose a novel DC-generating system applicable for patients with all types of HLA (Supplementary Figure s1). In this system, TAP-deficient iPS cells differentiate into CD11b+ myeloid cells (iPS-MC), and iPS-MC are transduced with expression vectors for cMYC, along with BMI1, EZH2, or MDM2, resulting in iPS-ML with the capacity for long-term propagation. Then, various alleles of HLA class I molecules can be introduced into TAP-deficient iPS-ML to establish a bank (or library) of iPS-ML carrying various HLA class I molecules. For vaccination therapy in patients, iPS-ML carrying appropriate HLA class I can be induced to differentiate into DC (ML-DC). The differentiation from iPS-ML to ML-DC takes only 2–3 days, and DC can therefore be supplied within a short period.
We found that mouse ES cell-derived myeloid cells could be immortalized by lentivirus-mediated introduction of cMYC (unpublished observation). Based on this finding, we tried to induce the proliferation of human iPS-MC by the introduction of cMYC. As a result, we observed that the forced expression of cMYC could induce the proliferation of human iPS-MC. However, this cMYC-induced proliferation of iPS-MC terminated in about 2 weeks. It is known that cMYC induces not only cell proliferation but also cell senescence,30 and we speculated that inhibition of the cell senescence-signaling pathway was necessary to allow for further proliferation beyond this 2-week period. To this end, we generated lentivirus vectors encoding genes with anti-senescence effects, and co-introduced them with cMYC into iPS-MC. We found that the co-expression of BMI-1, MDM2 or EZH2 with cMYC resulted in proliferation of these cells for more than 3 months.31, 32, 33
There were no significant differences in morphology or surface phenotypes among the iPS-ML generated using different factors. However, the optimal factors for generating iPS-ML differed depending on the individual iPS cells clones. Co-introduction of BMI-1 with cMYC was efficient for some iPS cell-derived iPS-MC, while EZH2 plus cMYC was suitable for another clone. The introduction of cMYC was always necessary. The triple-introduction of cMYC/BMI-1/MDM2 immortalized iPS-MC derived from all iPS cell clones so far tested. Moreover iPS-ML generated using all 3 factors tended to proliferate faster than those generated with either cMYC/BMI-1 or cMYC/EZH2.
Several genes expected to confer proliferative capacity or anti-apoptotic effects in these cells, including BCL2, cyclin D and telomerase reverse transcriptase, did not show any effects with regard to either inducing or enhancing the proliferation of iPS-MC both when used alone or when used in combination with other factors.
To generate the iPS cells used in the current study, we introduced reprogramming factors, OCT3/4, SOX2, KLF4 and cMYC, using lentivirus vectors into human fibroblasts, as described in our previous paper.17 However, these chromosomally integrated reprogramming factors, including cMYC, must have been transcriptionally silenced after the reprogramming was completed. In other words, proper silencing of exogenously introduced reprogramming factor genes is a requisite for the generation of iPS cells. If this did not occur, iPS cells could never differentiate. Although it is possible that the cMYC transgene is expressed at residual levels after the differentiation of iPS cells into iPS-MC, the level of cMYC expression would be too low to induce proliferation. This is why the re-introduction of the cMYC transgene was necessary for the generation of iPS-ML.
The long-term propagation capacity of iPS-ML may raise concerns about the occurrence of leukemia from ML-DC after administration to patients. However leukemia can be prevented by irradiation of the ML-DC before administration to patients. In addition, the survival and proliferation of iPS-ML depended on relatively high concentrations of M-CSF. Therefore, these cells should not proliferate after administration to patients, because the concentration of M-CSF in human plasma is less than 2 ng ml−1.34 Furthermore, in our system, ML-DC are allogeneic to their recipients. As described above, the magnitude of the rapid attack by the recipients’ allo-reactive CD8+ T cells against the transferred TAP-deficient DC is expected to be far lower than that against wild-type DC. Nevertheless, allo-reactive CD8+ T cells recognizing TAP-independent peptides may still be sufficient to eliminate the injected ML-DC, albeit less rapidly. In addition, the HLA class II molecules in TAP-deficient ML-DC are intact, and thus allo-reactive CD4+ T cells also attack the injected cells. Taken together, this indicates that the injected ML-DC will survive in the recipients long enough to stimulate antigen-specific CTL, but will be completely eliminated by the recipient immune system.
The effects of TNF-α with regard to increasing the cell surface expression of CD40 and HLA class II by ML-DC were similar to that of OK432 (penicillin-killed Streptococcus) (Figure 2a). However, the production of IL-12 by ML-DC was induced by OK432 or LPS, but not by TNF-α (Figure 2d). Similar differences between a cytokine mixture and OK432 in the induction of the maturation of human monocyte-derived DC were reported by Hovden et al.35 However, as shown in Figure 3a, treatment with TNF-α enhanced the capacity of ML-DC to stimulate allogeneic T cells. In this experiment, it is possible that TNF-α-treated ML-DC also produced IL-12 by stimulation with activated T cells after co-culture with T cells.
In the experiments shown in Figure 3a, responder T cells were co-cultured with iPS-ML or ML-DC. These iPS-ML or ML-DC stimulated T-cell proliferation, and they also had an inhibitory effect on the proliferation of T cells. In Figure 3a, stimulation with 104 iPS-ML induced a lower magnitude of T-cell proliferation compared with 103 or 3.16 × 103 iPS-ML. This is probably because the inhibitory effect of 104 iPS-ML somehow exceeded the stimulatory effects of iPS-ML under these conditions.
Currently, for the induction and expansion of tumor antigen-specific T cells for therapies, DC to be used as stimulator cells are prepared from autologous monocytes. However, it is sometimes difficult to prepare a sufficient number and quality of DC from monocytes, especially when the donor is a cancer patient. This problem can be resolved by the TAP-KO iPS cell-based approach. As sufficient numbers of pre-made ML-DC expressing the proper HLA class I can be provided by this system, individual preparation of DC will be unnecessary. The cost and labor required for the preparation of DC will be tremendously reduced. The disadvantage of the TAP-deficient iPS cell-based approach is that we cannot use TAP-deficient allogeneic ML-DC for the induction of antigen-specific CD4+ T cells.
Materials and methods
Cell samples and donors
All experiments using human samples were conducted under the approval of the Institutional Review Board of Kumamoto University and written informed consent was obtained from the donors. Human iPS cells were previously established by lentivirus-mediated introduction of OCT3/4, SOX2, KLF4 and cMYC into human fibroblasts obtained from a donor positive for HLA-A*24:02/11:01 and DRB4*01:03. The detailed procedures used for the generation, maintenance and gene introduction by electroporation of human iPS cells have been reported previously.17 Allogeneic peripheral blood mononuclear cells were obtained from donors positive for HLA-A*02:01 or HLA-A*02:06.
Plasmid construction and generation of recombinant lentivirus
A cDNA fragment of human cMYC was obtained by PCR and cloned into the pENTR-TOPO vector (Invitrogen, Carlsbad, CA, USA). cDNAs for BMI1, EZH2 and MDM2 were provided by RIKEN BioResource Center (Tsukuba, Japan) and transferred to a lentiviral expression vector, pCSII-EF,36 by using the LR clonase system (Invitrogen). pCSII-EF and the plasmids for lentiviral vector packaging, pCMV-VSV-G-RSV-Rev and pCAG-HIVgp,36 were kindly provided by Dr H Miyoshi (RIKEN BioResource Center). Plasmid constructs were introduced into 293T cells by lipofection (Lipofectamine 2000, Invitrogen), and 3 days later, recombinant lentivirus was recovered from the culture supernatant by centrifugation (50 000 g, 2 h).
Generation of proliferating myeloid lineage cells from human iPS cells
iPS cells were induced to differentiate into myeloid cells according to a previously established procedure.17 In brief, iPS cells were cultured on feeder layers of mouse OP9 cells for 15–20 days in α-MEM/20% fetal calf serum. Then, the cells were recovered and dissociated into single cells by using trypsin/collagenase/DNase I solution (phosphate-buffered saline containing 0.25% trypsin, 0.1% collagenase type IV, 0.1% DNase I). The cells were incubated in culture dishes for 4–12 h, then the non-adherent cells were recovered and further cultured in the presence of GM-CSF (50 ng ml−1) and M-CSF (50 ng ml−1). After several days, floating or weakly adherent cells appeared, and these were named iPS-MC (iPS cell-derived myeloid cells). These iPS-MC were infected with lentivirus vectors expressing cMYC along with BMI1, EZH2 or MDM2 in the presence of polybrene (8 ng ml−1), and were cultured in the presence of M-CSF (50 ng ml−1) and GM-CSF (50 ng ml−1). After 5–10 days, proliferating cells appeared, and these cells named iPS-ML (iPS cell-derived myeloid cell line). To induce the differentiation into DC (ML-DC), the cells were cultured in the presence of the M-CSF (50 ng ml−1), GM-CSF (100 ng ml−1) and IL-4 (10 ng ml−1) for 2–4 days. To induce the maturation of ML-CD, ML-DC were stimulated with TNF-α (10 ng ml−1) or penicillin-killed Streptococcus pyogenes, OK432 (10 μg ml−1, Chugai Pharmaceutical, Tokyo, Japan) for 2–3 days.
Analysis of the dependence of iPS-ML proliferation on M-CSF and GM-CSF
The iPS-ML were plated into 96-well flat-bottomed culture plates (5 × 103 cells per well) in the presence of the indicated concentration of M-CSF and GM-CSF for 3 days. [3H]-methylthymidine was added to the culture (0.037 Mbq per well) in the last 16 h, and the incorporation of [3H]-thymidine was measured by scintillation counting.
The following monoclonal antibody (mAb) conjugated with fluorescein isothiocyanate or phycoerythrin were purchased from BD Pharmingen (San Diego, CA, USA) or eBioscience (San Diego, CA, USA): anti-CD45 (clone HI30, mouse IgG1), anti-CD11b (clone ICRF44, mouse IgG1), anti-CD33 (WM53, mouse IgG1), anti-HLA class II (clone TU39, mouse IgG2a), anti-HLA class I (clone G46-2.6, mouse IgG1), anti-CD80 (clone L307.4, mouse IgG1), anti-CD83 (clone HB15e, mouse IgG1), anti-CD86 (clone FUN-1, mouse IgG1), anti-CD40 (clone 5C3, mouse IgG1) and anti-CD14 (clone 61D3, mouse IgG1). Mouse IgG2a (clone G155-178) and mouse IgG1 (clone MOPC-21) were used as isotype-matched controls. The cell samples were treated with Fc-receptor-blocking reagent (Miltenyi Biotec, Bergish-Gladbach, Germany) for 10 min, stained with the fluorochrome-conjugated mAb for 30 min, and washed three times with phosphate-buffered saline /2% fetal calf serum. The stained cell samples were analyzed on a FACScan (BD Bioscience, Bedford, MA, USA) flowcytometer.
Quantitation of cytokine production
To analyze the production of IL-12 by ML-DC, the cells were cultured in 96-well flat-bottomed culture plates (1 × 105 cells per 200 μl of medium per well) in the presence of TNF-α (10 ng ml−1), LPS (2 μg ml−1), or OK432 (20 μg ml−1). Supernatants were collected after 60 h of culture, and the concentration of IL-12p70 was measured by ELISA (R&D Systems, Minneapolis, MN, USA).
T-cell proliferation assay
To analyze the stimulation of allogeneic T cells, T cells (4 × 104 cells per well) were purified from peripheral blood mononuclear cells obtained from an allogeneic donor using a magnetic bead-based T-cell isolation kit (Pan T cell isolation kit, Miltenyi Biotec) and were co-cultured with graded numbers of X-ray-irradiated (45 Gy) stimulator cells in RPMI-1640 medium supplemented with 10% human plasma in 96-well round-bottomed culture plates for 5 days. [3H]-Methylthymidine (247.9 Gbq mmol−1) was added to the culture (0.037 Mbq per well) in the last 16 h. The cells were harvested onto glass fiber filters (Wallac, Turku, Finland), and the incorporation of [3H]-thymidine was measured by scintillation counting. To stimulate SA32.5, a glutamic acid decarboxylase 65 (GAD65)-specific HLA-DR53-restricted human CD4+ T-cell clone was used,20 and ML-DC were pulsed with GAD65p111–131 (LQDVMNILLQYVVKSFDRSTK) (10 μM) for 3 h and X-ray-irradiated (45Gy). The SA32.5 T cells (4 × 104 cells per well) were cultured with the indicated numbers of peptide-pulsed ML-DC for 3 days. [3 H]-Methylthymidine was added to the culture (0.037 Mbq/well) in the last 16 h, and the incorporation of [3H]-thymidine was measured by scintillation counting.
Stimulation of CMV-specific autologous human CD8+ T cells
ML-DC stimulated with TNF-α were irradiated, incubated with an HLA-A24-restricted CMV pp65 341–349 peptide (QYDPVAALF, 10 μM)37 for 3 h, and plated into 48-well culture plates (4 × 104 cells/well). Autologous peripheral blood T cells were added to the wells (4 × 105 cells per well). The cells were then cultured in AIM-V medium (Life Technologies, Carlsbad, CA, USA) containing 10% autologous plasma and recombinant human IL-7 (5 ng ml−1), then human IL-2 (20 U ml−1) was added on day 2. On day 9 of the culture, the cells were harvested, stained with a phycoerythrin-labeled tetramer of the HLA-A*24:02/CMV peptide complex (MBL, Nagoya, Japan) in combination with a fluorescein isothiocyanate-labeled anti-human CD8 mAb (clone T8, Beckman Coulter, Brea, CA, USA), and analyzed by flow cytometry. The number of IFN-γ producing T cells upon stimulation with the peptide-pulsed C1R-A24 (HLA-A*24:02-expressing C1R lymphoma, provided by Dr M Takiguchi, Kumamoto University) cells was counted by ELISPOT assay (BD Bioscience).
Generation of TAP2-deficient iPS cells
The gene targeting vector for the human TAP2 gene included a 1069 bp left homologous arm, a neomycin-phosphotransferase expression cassette, and a 1059 bp right arm. The expression vectors for the ZFN designed to cut exon 2 of the human TAP2 gene, generated by Sigma-Aldrich (St Louis, MO, USA), were introduced simultaneously with the targeting vector by electroporation into iPS cells. The amino acid sequence of the DNA recognition helices of ZFN is proprietary information of Sigma-Aldrich. However, identical ZFN reagents used in this study are available from Sigma-Aldrich as lot D09021039MN. After electroporation, the iPS cells were culture in the presence of G418- (0.2 mg ml−1) and G418-resistant iPS cell colonies were picked on days 14–16. Genomic DNA was extracted from the isolated iPS cell clones using a DNA extraction kit (Qiagen, Maryland, MD, USA) and used for genotyping of the TAP2 locus by PCR. The PCR primers 1F (5′-GGGTGGGGTGGGATTAGATA-3′) and 1R (5′-CCAGTCCCAGCCTTATCAAA-3′) were used to detect the mutated allele, and the PCR primers 2F (5′-GCTGAGAAGGACGGATGAAG-3′) and 2R (5′-CCAGTCCCAGCCTTATCAAA-3′) were used to detect both the wild-type allele and the mutated alleles. TAP2 protein in iPS-ML was detected by western blot analysis, using anti-human TAP2 mAb (clone TAP2.17, MBL).
Analysis of the response of allo-reactive CD8+ T cells
Peripheral blood CD8+ T cells obtained from an allogeneic donor were stimulated with wild-type ML-DC for 7 days to generate an allo-reactive CD8+ T cell line. To analyze the recognition of undifferentiated iPS cells by the allo-reactive T cells, wild-type or TAP2-deficient iPS cells (1 × 104 cells per well) were co-cultured with the allo-reactive CD8+ T cell line (5 × 104 cells per well) in a 96-well flat-bottomed culture plate for 16 h, after which the concentrations of IFN-γ in the culture supernatants were measured by ELISA (R&D systems or Invitrogen). iPS-MC or iPS-ML (1 × 104 cells per well) were co-cultured with the allo-reactive CD8+ T cells for 16 h. Concentrations of IFN-γ in the culture supernatants were measured by ELISA, and the number of IFN-γ producing T cells was counted by ELISPOT assay (BD Bioscience).
Stimulation of MART-1-specific allogeneic human CD8+ T cells
iPS-ML generated from TAP2-deficient iPS cells were transduced with the HLA-A*02:01 or HLA-A*02:06 gene using lentivirus vectors. For stimulation of MART-1-specific CTL, ML-DC treated with OK432 for 2 days were irradiated (45Gy), incubated with MART-1 26-35 peptide (AAGIGILTV, 10 μM)38 for 3 h, and plated in 24-well culture plates (2 × 105 cells per well). Then, peripheral blood CD8+ T cells obtained from allogeneic donors positive for HLA-A*02:01 or HLA-A*02:06 were added (1.8 × 106 cells per well). The cells were cultured in AIM-V containing 5% human plasma and recombinant human IL-7 (10 ng ml−1), and human IL-2 (20 U ml−1) was added on day 2. The T cells were re-stimulated with MART-1 peptide-pulsed ML-DC on days 7 and 14. On days 14 and 21, to analyze the frequency of MART-1-specific T cells, T cells were recovered and stained with a phycoerythrin-labeled dextramer of HLA-A*02:01/MART1 peptide complex (Immundex, Copenhagen, Denmark) in combination with fluorescein isothiocyanate-labeled anti-human CD8 mAb (BD PharMingen), and analyzed by flow cytometry. To analyze reactivity to the MART-1 peptide, the T cells were cultured with T2 cells (HLA-A*02:01-positive) or T2 cells transduced with HLA-A*02:06 (1 × 104 cells per well) pre-pulsed with MART1 peptide. After culture for 16 h, the numbers of T cells producing IFN-γ were analyzed using an ELISPOT assay kit (BD Bioscience).
- ES cell:
embryonic stem cell
- iPS cell:
induced pluripotent stem cell
iPS cell-derived myeloid cells
iPS cell-derived myeloid cell line
iPS-ML-derived dendritic cell-like APC.
Eksioglu EA, Eisen S, Reddy V . Dendritic cells as therapeutic agents against cancer. Front Biosci 2010; 15: 321–347.
Robson NC, Hoves S, Maraskovsky E, Schnurr M . Presentation of tumour antigens by dendritic cells and challenges faced. Curr Opin Immunol 2010; 22: 137–144.
Schuler G . Dendritic cells in cancer immunotherapy. Eur J Immunol 2010; 40: 2123–2130.
Takahashi K, Yamanaka S . Induction of pluripotent stem cells from mouse embryonic and adult fibroblast cultures by defined factors. Cell 2006; 126: 663–676.
Takahashi K, Tanabe K, Ohnuki M, Narita M, Ichisaka T, Tomoda K et al. Induction of pluripotent stem cells from adult human fibroblasts by defined factors. Cell 2007; 131: 861–872.
Fairchild PJ, Brook FA, Gardner RL, Graca L, Strong V, Tone Y et al. Directed differentiation of dendritic cells from mouse embryonic stem cells. Curr Biol 2000; 10: 1515–1518.
Senju S, Hirata S, Matsuyoshi H, Masuda M, Uemura Y, Araki K et al. Generation and genetic modification of dendritic cells derived from mouse embryonic stem cells. Blood 2003; 101: 3501–3508.
Senju S, Haruta M, Matsunaga Y, Fukushima S, Ikeda T, Takahashi K et al. Characterization of dendritic cells and macrophages generated by directed differentiation from mouse induced pluripotent stem cells. Stem Cells 2009; 27: 1021–1031.
Zhan X, Dravid G, Ye Z, Hammond H, Shamblott M, Gearhart J et al. Functional antigen-presenting leucocytes derived from human embryonic stem cells in vitro. Lancet 2004; 364: 163–171.
Slukvin II, Vodyanik MA, Thomson JA, Gumenyuk ME, Choi KD . Directed differentiation of human embryonic stem cells into functional dendritic cells through the myeloid pathway. J Immunol 2006; 176: 2924–2932.
Su Z, Frye C, Bae KM, Kelley V, Vieweg J . Differentiation of human embryonic stem cells into immunostimulatory dendritic cells under feeder-free culture conditions. Clin Cancer Res 2008; 14: 6207–6217.
Senju S, Suemori H, Zembutsu H, Uemura Y, Hirata S, Fukuma D et al. Genetically manipulated human embryonic stem cell-derived dendritic cells with immune regulatory function. Stem Cells 2007; 25: 2720–2729.
Tseng SY, Nishimoto KP, Silk KM, Majumdar AS, Dawes GN, Waldmann H et al. Generation of immunogenic dendritic cells from human embryonic stem cells without serum and feeder cells. Regen Med 2009; 4: 513–526.
Fukushima S, Hirata S, Motomura Y, Fukuma D, Matsunaga Y, Ikuta Y et al. Multiple antigen-targeted immunotherapy with alpha-galactosylceramide-loaded and genetically engineered dendritic cells derived from embryonic stem cells. J Immunother 2009; 32: 219–231.
Matsuyoshi H, Senju S, Hirata S, Yoshitake Y, Uemura Y, Nishimura Y . Enhanced priming of antigen-specific CTLs in vivo by embryonic stem cell-derived dendritic cells expressing chemokine along with antigenic protein: application to antitumor vaccination. J Immunol 2004; 172: 776–786.
Motomura Y, Senju S, Nakatsura T, Matsuyoshi H, Hirata S, Monji M et al. Embryonic stem cell-derived dendritic cells expressing glypican-3, a recently identified oncofetal antigen, induce protective immunity against highly metastatic mouse melanoma, B16-F10. Cancer Res 2006; 66: 2414–2422.
Senju S, Haruta M, Matsumura K, Matsunaga Y, Fukushima S, Ikeda T et al. Generation of dendritic cells and macrophages from human induced pluripotent stem cells aiming at cell therapy. Gene Therapy 2011; 18: 874–883.
Choi KD, Vodyanik M, Slukvin II . Hematopoietic differentiation and production of mature myeloid cells from human pluripotent stem cells. Nat Protoc 2011; 6: 296–313.
Nakahara S, Tsunoda T, Baba T, Asabe S, Tahara H . Dendritic cells stimulated with a bacterial product, OK-432, efficiently induce cytotoxic T lymphocytes specific to tumor rejection peptide. Cancer Res 2003; 63: 4112–4118.
Tabata H, Kanai T, Yoshizumi H, Nishiyama S, Fujimoto S, Matsuda I et al. Characterization of self-glutamic acid decarboxylase 65-reactive CD4+ T-cell clones established from Japanese patients with insulin-dependent diabetes mellitus. Hum Immunol 1998; 59: 549–560.
Neefjes JJ, Momburg F, Hammerling GJ . Selective and ATP-dependent translocation of peptides by the MHC-encoded transporter. Science 1993; 261: 769–771.
Zwaka TP, Thomson JA . Homologous recombination in human embryonic stem cells. Nat Biotechnol 2003; 21: 319–321.
Kim YG, Cha J, Chandrasegaran S . Hybrid restriction enzymes: zinc finger fusions to Fok I cleavage domain. Proc Natl Acad Sci U S A 1996; 93: 1156–1160.
Bacik I, Cox JH, Anderson R, Yewdell JW, Bennink JR . TAP (transporter associated with antigen processing)-independent presentation of endogenously synthesized peptides is enhanced by endoplasmic reticulum insertion sequences located at the amino- but not carboxyl-terminus of the peptide. J Immunol 1994; 152: 381–387.
Fukuma D, Matsuyoshi H, Hirata S, Kurisaki A, Motomura Y, Yoshitake Y et al. Cancer prevention with semi-allogeneic ES cell-derived dendritic cells. Biochem Biophys Res Commun 2005; 335: 5–13.
Wells JW, Cowled CJ, Darling D, Guinn BA, Farzaneh F, Noble A et al. Semi-allogeneic dendritic cells can induce antigen-specific T-cell activation, which is not enhanced by concurrent alloreactivity. Cancer Immunol Immunother 2007; 56: 1861–1873.
Loyer V, Fontaine P, Pion S, Hetu F, Roy DC, Perreault C . The in vivo fate of APCs displaying minor H antigen and/or MHC differences is regulated by CTLs specific for immunodominant class I-associated epitopes. J Immunol 1999; 163: 6462–6467.
Hermans IF, Ritchie DS, Yang J, Roberts JM, Ronchese F . CD8+ T cell-dependent elimination of dendritic cells in vivo limits the induction of antitumor immunity. J Immunol 2000; 164: 3095–3101.
Matsunaga Y, Fukuma D, Hirata S, Fukushima S, Haruta M, Ikeda T et al. Activation of antigen-specific cytotoxic T lymphocytes by beta2-microglobulin or TAP1 gene disruption and the introduction of recipient-matched MHC class I gene in allogeneic embryonic stem cell-derived dendritic cells. J Immunol 2008; 181: 6635–6643.
Lee S, Schmitt CA, Reimann M . The Myc/macrophage tango: oncogene-induced senescence, Myc style. Semin Cancer Biol 2011; 21: 377–384.
Jacobs JJ, Kieboom K, Marino S, DePinho RA, van Lohuizen M . The oncogene and Polycomb-group gene bmi-1 regulates cell proliferation and senescence through the ink4a locus. Nature 1999; 397: 164–168.
Bracken AP, Kleine-Kohlbrecher D, Dietrich N, Pasini D, Gargiulo G, Beekman C et al. The Polycomb group proteins bind throughout the INK4A-ARF locus and are disassociated in senescent cells. Genes Dev 2007; 21: 525–530.
Kamminga LM, Bystrykh LV, de Boer A, Houwer S, Douma J, Weersing E et al. The Polycomb group gene Ezh2 prevents hematopoietic stem cell exhaustion. Blood 2006; 107: 2170–2179.
Weinberg JB, Haney AF, Xu FJ, Ramakrishnan S . Peritoneal fluid and plasma levels of human macrophage colony-stimulating factor in relation to peritoneal fluid macrophage content. Blood 1991; 78: 513–516.
Hovden AO, Karlsen M, Jonsson R, Aarstad HJ, Appel S . Maturation of monocyte derived dendritic cells with OK432 boosts IL-12p70 secretion and conveys strong T-cell responses. BMC Immunol 2011; 12: 2.
Miyoshi H, Blomer U, Takahashi M, Gage FH, Verma IM . Development of a self-inactivating lentivirus vector. J Virol 1998; 72: 8150–8157.
Kuzushima K, Hayashi N, Kimura H, Tsurumi T . Efficient identification of HLA-A*2402-restricted cytomegalovirus-specific CD8(+) T-cell epitopes by a computer algorithm and an enzyme-linked immunospot assay. Blood 2001; 98: 1872–1881.
Kawakami Y, Eliyahu S, Sakaguchi K, Robbins PF, Rivoltini L, Yannelli JR et al. Identification of the immunodominant peptides of the MART-1 human melanoma antigen recognized by the majority of HLA-A2-restricted tumor infiltrating lymphocytes. J Exp Med 1994; 180: 347–352.
The cDNA clones for MDM2, BMI1, and EZH2 and 293T cells were provided by RIKEN BRC, which is participating in the National Bio-Resources Project of the MEXT, Japan. The pCSII-EF, pCMV-VSV-G-RSV-Rev and pCAG-HIVgp constructs were kindly provided by Dr H Miyoshi (RIKEN BioResource Center). This work was supported in part by Grants-in-Aid Nos 18014023, 19591172 and 19059012 from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), Japan, Research Grant for Intractable Diseases from Ministry of Health and Welfare, Japan and grants from Japan Science and Technology Agency (JST).
The authors declare no conflict of interest.
Supplementary Information accompanies the paper on Gene Therapy website
About this article
Cite this article
Haruta, M., Tomita, Y., Yuno, A. et al. TAP-deficient human iPS cell-derived myeloid cell lines as unlimited cell source for dendritic cell-like antigen-presenting cells. Gene Ther 20, 504–513 (2013). https://doi.org/10.1038/gt.2012.59
- dendritic cells
- iPS cells
- tumor immunity and vaccination
Type I Interferon Delivery by iPSC-Derived Myeloid Cells Elicits Antitumor Immunity via XCR1+ Dendritic Cells
Cell Reports (2019)
Trends in Molecular Medicine (2019)
Microphysiological systems meet hiPSC technology – New tools for disease modeling of liver infections in basic research and drug development
Advanced Drug Delivery Reviews (2019)
Immunological Medicine (2018)
Journal of General Virology (2018)