Induced pluripotent stem cell-derived dendritic cell vaccine therapy genetically modified on the ubiquitin-proteasome system

We previously reported that dendritic cells (DCs) transduced with the full-length tumor-associated antigen (TAA) gene induced TAA-specific cytotoxic T lymphocytes (CTLs) to elicit antitumor responses. To overcome the issue of quantity and quality of DCs required for DC vaccine therapy, we focused on induced pluripotent stem cells (iPSCs) as a new tool for obtaining DCs and reported efficacy of iPSCs-derived DCs (iPSDCs). However, in clinical application of iPSDC vaccine therapy, further enhancement of the antitumor effect is necessary. In this study, we targeted mesothelin (MSLN) as a potentially useful TAA, and focused on the ubiquitin-proteasome system to enhance antigen-presenting ability of iPSDCs. The CTLs induced by iPSDCs transduced with MSLN gene (iPSDCs-MSLN) from healthy donors showed cytotoxic activity against autologous lymphoblastoid cells (LCLs) expressing MSLN (LCLs-MSLN). The CTLs induced by iPSDCs transduced ubiquitin-MSLN fusion gene exhibited higher cytotoxic activity against LCLs-MSLN than the CTLs induced by iPSDCs-MSLN. The current study was designed that peripheral T-cell tolerance to MSLN could be overcome by the immunization of genetically modified iPSDCs simultaneously expressing ubiquitin and MSLN, leading to a strong cytotoxicity against tumors endogenously expressing MSLN. Therefore, this strategy may be promising for clinical application as an effective cancer vaccine therapy.


INTRODUCTION
Dendritic cells (DCs) are the most potent professional antigenpresenting cells. The presentation of peptides generated by the ubiquitin-proteasome pathway on MHC class I molecules, known as cross-presentation, is essential for the initiation of antigenspecific CD8+ T cells responses [1]. The cross-presentation is an extremely useful feature for antitumor immune responses. We previously reported that DCs transduced with the full-length tumor-associated antigen (TAA) gene induced TAA-specific cytotoxic T lymphocytes (CTLs) to elicit antitumor responses [2][3][4][5]. However, the DC vaccine therapy has not reached breakthrough in the cancer treatment, it requires a large quantity of DCs generated from peripheral blood monocytes of patients with cancer. Frequent apheresis must be performed, which is burdensome upon such patients [6]. Moreover, the function of DCs such as maturation, antigen-presenting ability, and migration is diminished in patients with cancer [7]. To overcome these problems of DC vaccine therapy, we focused on induced pluripotent stem cells (iPSCs) as a new tool for obtaining DCs. Several reports have highlighted methods for differentiating DCs from iPSCs (iPSDCs) in mouse and human studies [8][9][10][11][12]. Our previous study reported that mouse iPSDCs transduced with TAA gene have a capacity for TAA-specific antitumor immunity equal to that of bone marrow-derived DCs [13]. In another study, we reported that iPSDCs transduced with carcinoembryonic antigen (CEA) gene have the capacity for CEA-specific antitumor immunity in our mouse and human models [14]. Furthermore, in vitro CTLs obtained by transfecting tumor-derived RNA into iPSDCs from patients with colorectal cancer showed potent tumor-specific killing effect [15]. However, we have not been able to create an iPSDC vaccine that exerts a strong antitumor effect. In clinical application of iPSDC vaccine therapy, further enhancement of the antitumor effect is necessary.
In the current study, we targeted mesothelin (MSLN) as a potentially useful TAA, and focused upon the ubiquitinproteasome system to enhance antigen-presenting ability of iPSDCs. MSLN is a CA125 (Carbohydrate Antigen 125)-binding protein. CA125 might contribute to the peritoneal metastasis of ovarian cancer by initiating cell attachment to the mesothelial epithelium via binding to MSLN [16]. Co-expression of MSLN and MUC16 (Mutin16, CA125) is related to the invasion process in pancreatic ductal adenocarcinoma [17]. MSLN is a predictive factor for peritoneal recurrence and correlated with poor outcomes in gastric cancer [18]. MSLN is known as an attractive target for cancer immunotherapy, owing to its low expression on normal mesothelial cells and high expression in several solid tumors including gastrointestinal tumors [19]. For cancer immunotherapy targeting MSLN, there are reports of dendritic cell vaccine [5], a peptide vaccine [20,21], and a DNA vaccine [22] in basic research. Miyazawa et al. reported that CTLs induced by monocytes-derived DCs (MoDCs) transfected with whole MSLN gene elicited MSLNspecific cytotoxicity against pancreatic cancer cell line [5]. In clinical practice, phase I/II trials of antibody therapy have been conducted and some efficacy has been shown [23].
The ubiquitin was originally identified as the signal to target proteins for proteasomal degradation [24]. The addition of a ubiquitin monomer sequence to an adenoviral transgene as a source of antigen increased both in vivo specific CD8 immunogenicity and in vitro MHC class I-restricted presentation by DCs of the encoded peptides [25]. In tumor immunity, the use of ubiquitinfused epitopes reportedly enhanced the antitumor effect in mouse DNA vaccines [26][27][28] and in human DC vaccines [29].
We investigated whether genetically modified human iPSDCs expressing MSLN could induce MSLN-specific CTLs. Furthermore, we assessed whether the induction of ubiquitin-MSLN fusion gene into iPSDCs enhanced the antigen-presenting ability through the ubiquitin-proteasome system and whether it enhanced the cytotoxic ability of CTLs.

Recombinant adenoviral vector construction and gene transfer into DCs
The recombinant adenoviral vector AxCAMSLN, encoding MSLN, was generated by the cosmid-terminal protein complex (COS-TPC) method, as previously described [5]. AxCALacZ, encoding β-gal, was also generated by the COS-TPC method. The ubiquitin gene was amplified by PCR from PBMC genomic DNA, using the 5′ primer (5′ AGT CCG CTA GCC GCC ACC ATG CAG ATC TTC GTG AAG ACC 3′) and the 3′ primer (5′ TAG TCC GTC GAC GTA TTT AAA TCG ACC CCC CCT CAA GCG CAG GAC 3′). The 5′ primer contained Nhe I restriction site, the Kozak sequence (CGCCACC), and the ATG start codon. The 3′ primer added an arginine to the N terminal extremity of ubiquitin protein and contained Swa I restriction site [25,28]. The monomer ubiquitin cDNA was inserted into the cosmid vector pAxCAwt to yield the ubiquitin-expressing recombinant cosmid. The MSLN gene was inserted into the Swa I restriction site to prepare the ubiquitin-MSLN gene expression recombinant cosmid. AxCAubiquitin-MSLN, encoding ubiquitin-MSLN, was generated by the COS-TPC method ( Fig. 1). Immature DCs were transfected with each recombinant adenoviral vector using the centrifugal method at 100 multiplicity of infection (MOI) [2,4]. The genetically modified DCs were placed at 1.5 × 10 6 cells/well in 6-well Ultra-Low Attachment Surface plate in the presence of 100 ng/mL rhIL-6, 10 ng/mL rhIL-1β, 10 ng/mL rhTNF-α, and 1 µg/mL PGE2 to induce final maturation for 48 h [14].

Assay for IFN-ɤ secretion
The MSLN-specific CTLs were adjusted to a concentration of 1 × 10 5 cells/ well and DCs-LacZ, DCs-MSLN, or DCs-UbMSLN were adjusted to a concentration of 1 × 10 4 cells/well and co-cultured on 48-well plate for 48 h in AIM-V medium (1 mL/well), respectively. The supernatants were then harvested, and the human IFN-ɤ levels were measured using an IFN-ɤ ELISA kit (Thermo Fisher Scientific).

Statistical analysis
JMP version 16 (SAS institute Inc., Cary, USA) was used for all statistical analyses. Statistical comparisons between the groups were performed using Mann-Whitney U test. P < 0.05 was considered to be significant.

Expression of co-stimulatory molecules in MoDCs and iPSDCs
The expression of CD86 was comparable in iPSDCs and MoDCs, however the expression of CD40, CD80, and CD83 among costimulatory molecules was lower in iPSDCs than in MoDCs. We found that introduction of the MSLN gene or the UbMSLN fusion gene did not alter the expression of co-stimulatory molecules in MoDCs and iPSDCs, respectively (Fig. 2).

Expression of MSLN in genetically modified DCs
Immature DCs were transduced with recombinant adenoviral vector by centrifugal method at 100 MOI. The genetically modified DCs were analyzed using intracellular MSLN-staining flow cytometry to compare the transfection efficiency of the MSLN gene between MoDCs-MSLN and iPSDCs-MSLN. The percentage of positively stained cells among the MoDCs and iPSDCs was 83% and 82%, respectively (Fig. 3a).

Cytotoxic activity of CTLs induced by MoDCs-MSLN and iPSDCs-MSLN
CTLs induced by MoDCs-MSLN from healthy donors 1, 2, and 3 showed cytotoxic activity against autologous LCL-MSLN in a concentration-dependent manner. CTLs induced by iPSDCs-MSLN from healthy donor A, B, and C also showed cytotoxic activity against autologous LCL-MSLN in a concentration-dependent manner (Fig. 3b).

Secretion of IFN-ɤ from CTLs and DCs
The levels secreted from CTLs and DCs-UbMSLN were higher than CTLs and DCs-MSLN ( Supplementary Fig. 2).

DISCUSSION
Our results showed that the CTLs induced by genetically modified DCs expressing MSLN could induce MSLN-specific cytotoxic activity. MSLN is the optimal cancer vaccine target for many gastrointestinal cancers [5,[19][20][21][22][23]. The strategy of this study using DCs-MSLN are expected to be applied to the treatment of gastrointestinal cancer expressing MSLN.
Not only the genetically modified naïve DCs-MSLN but iPSDCs-MSLN could induce MSLN-specific CTLs with cytotoxic activity against autologous LCL-MSLN. In addition, CTLs induced by iPSDCs-MSLN also showed cytotoxic activity against MSLNspecific epitope peptide-pulsed LCL in the HLA-A type restricted manner. CTLs induced by iPSDCs transfected with whole MSLN gene were indicated to recognize multiple MSLN-derived epitope peptides in an HLA-A type restricted manner. According to 11th International Histcompatibility Workshop data, the highest gene frequency of HLA-A antigen in the Japanese population was A24 at 37.1%, followed by A2 at 24.1%. MSLN-specific epitope peptides have been identified in HLA-A24 and A2. Therefore, HLA-A24 and A2 donors were selected in current study. We previously reported a gene-based vaccination strategy using iPSDCs transduced with the whole TAA gene [13,14]. This study also clearly showed that  Fig. 2 Expression of co-stimulatory molecules in MoDCs and iPSDCs. Histograms show the staining results of specific antibodies and isotype-matched controls (thin lines). We found that introduction of the MSLN gene and the UbMSLN fusion gene did not alter the expression of co-stimulatory molecules in MoDCs and iPSDCs, respectively. Results were shown the mean in ± SD. *Significant increase of the lysis was shown (p < 0.05). c Cytotoxic activity of CTLs generated from iPSDCs-MSLN against LCLs pulsed MSLN-specific epitope peptides and WT1 specific peptide as a control. CTLs induced by iPSDCs-MSLN showed cytotoxic activity against MSLN-specific epitope peptide-pulsed LCL in the HLA-A type restricted manner. iPSDCs had a capacity equal to that of naïve DCs as antigenpresenting cells. However, the expression of some co-stimulatory molecules was lower in iPSDCs than in MoDCs, and further studies on the differentiation protocols and maturation methods are required.
In the current study, iPSCs-derived DCs were adenovirally transduced simultaneously with ubiquitin gene and TAA gene, and it was examined whether the in vitro CTLs immunized with these genetically modified DCs induce a strong cytotoxic activity against LCLs expressing TAA in healthy volunteers. Autologous LCLs were used as a target with a completely matched HLA type. Autologous cells should be targeted because DCs transfected with the entire TAA gene will produce CTLs for many peptides. We will also examine targeting cancer cell lines overexpressing MSLN and in vitro studies in mice transplanted with cancer cell lines overexpressing MSLN. Our result showed that the in vitro immunization of iPSDCs-UbMSLN elicited a more potent cytotoxic activity than the in vitro immunization of iPSDCs-MSLN. This is the first report of production of the recombinant adenovirus expressing the ubiquitin-MSLN fusion gene. To ensure that ubiquitin functions properly, the primers were provided with Kozak sequences (CGCCACC) that are important during eukaryotic translation. In addition, arginine was placed at the N-terminus of ubiquitin. The use of arginine as a linker enhances proteolysis of the fusion gene. This is known as the N-end rule [28]. To investigate whether the ubiquitin-MSLN fusion gene was functioning normally, MSLN protein expression in DCs was assessed by immunohistochemistry and intracellular FACS. In the DCs transduced ubiquitin-MSLN fusion gene, MSLN is forced into the ubiquitin-proteasome system and MSLN proteolysis is enhanced because the MSLN protein is thought to be expressed in a form in which ubiquitin is pre-bound to the MSLN protein. As a result, MSLN protein expression in DCs-UbMSLN is thought to decrease. In contrast, inhibition of proteasome function with MG132, proteasome inhibitor, did not reduce MSLN protein expression in DCs-UbMSLN. The ubiquitin-MSLN fusion gene transfer enhanced MSLN proteolysis in DCs-UbMSLN, suggesting that proteolysis is proteasome-dependent. Pentamer assay showed that increased MSLN proteolysis enhanced the antigen-presenting capacity of MSLN-derived epitope peptides in DCs-UbMSLN and induced more MSLN-specific CTLs. Assay for IFN-ɤ secretion showed that in DCs-UbMSLN, not only the number of MSLNspecific CTLs induced increases, but also their reactivity. In addition, it was suggested that ubiquitin-MSLN fusion genetransfected DCs may also enhance cytotoxic activity. To solve the problem of insufficient cytotoxic activity of iPSDC vaccines, it may be useful to introduce ubiquitin-MSLN fusion gene into iPSDCs.
In conclusion, the current study was designed that peripheral T-cell tolerance to MSLN could be overcome by the immunization of genetically modified iPSDCs simultaneously expressing ubiquitin and MSLN, leading to a strong cytotoxicity against tumors endogenously expressing MSLN. Therefore, this strategy may be promising for clinical application as an effective cancer vaccine therapy.

DATA AVAILABILITY
The data are available upon reasonable request.