Vaccination with vascular progenitor cells derived from induced pluripotent stem cells elicits antitumor immunity targeting vascular and tumor cells

Abstract

Vaccination of BALB/c mice with dendritic cells (DCs) loaded with the lysate of induced vascular progenitor (iVP) cells derived from murine-induced pluripotent stem (iPS) cells significantly suppressed the tumor of CMS-4 fibrosarcomas and prolonged the survival of CMS-4-inoculated mice. This prophylactic antitumor activity was more potent than that of immunization with DCs loaded with iPS cells or CMS-4 tumor cells. Tumors developed slowly in mice vaccinated with DCs loaded with iVP cells (DC/iVP) and exhibited a limited vascular bed. Immunohistochemistry and a tomato-lectin perfusion study demonstrated that the tumors that developed in the iVP-immunized mice showed a marked decrease in tumor vasculature. Immunization with DC/iVP induced a potent suppressive effect on vascular-rich CMS-4 tumors, a weaker effect on BNL tumors with moderate vasculature, and nearly no effect on C26 tumors with poor vasculature. Treatment of DC/iVP-immunized mice with a monoclonal antibody against CD4 or CD8, but not anti-asialo GM1, inhibited the antitumor activity. CD8⁺ T cells from DC/iVP-vaccinated mice showed significant cytotoxic activity against murine endothelial cells and CMS-4 cells, whereas CD8⁺ T cells from DC/iPS-vaccinated mice did not. DNA microarray analysis showed that the products of 29 vasculature-associated genes shared between genes upregulated by differentiation from iPS cells into iVP cells and genes shared by iVP cells and isolated Flk-1⁺ vascular cells in CMS-4 tumor tissue might be possible targets in the immune response. These results suggest that iVP cells from iPS cells could be used as a cancer vaccine targeting tumor vascular cells and tumor cells.

Electronic supplementary material

The online version of this article (doi:10.1007/s00262-014-1531-1) contains supplementary material, which is available to authorized users.

Introduction

Induced pluripotent stem (iPS) cells were originally generated by transduction with the four genes associated with stemness and differentiation of the cell [1], and these cells have the capacity to differentiate into various cell types. Although iPS cells have a risk of tumorigenicity, they might be a promising tool for future tissue engineering and regenerative medicine, with the goal of preventing tumor development by implantation of iPS-derived cells [2, 3]. Additionally, iPS cells derived from patients with severe diseases arising from unknown internal dysfunction might contribute to the development of new treatment [4]. iPS cells have also attracted attention for application in cancer treatment and especially cancer immunotherapy. iPS-derived dendritic cells (DCs) and immune effecter cells for anticancer cell therapy have been reported [5, 6]. It has also been reported that iPS cells and embryonic stem (ES) cells have been utilized as a cancer vaccine in an experimental animal model [7].

Although cancer immunotherapy has been popular for decades and although significant advances in cancer immunotherapy have been achieved [8], this area has not yielded sufficient results in cancer therapy [9]. Attacking tumor cells using T-cell-mediated immune responses faces large obstacles. A loss of target molecules on the surface of tumor cells recognized by antigen-specific T cells or the expression of molecules that inhibit the immune response might be inevitably involved in the incompetence of T-cell-mediated antitumor immunity. Environmental factors, such as immunosuppressive cytokines produced by cancer cells and the induction of regulatory T cells and myeloid-derived suppressor cells, might be important factors that inhibit antitumor immune responses [10].

Given that the development of tumor vasculature is an essential factor in rapid tumor growth, vascular endothelial growth factor (VEGF)-targeting therapy has been widely applied for the treatment for various types of cancer [11]. Recently, targeting tumor vessels has also become a focus in cancer immunotherapy because tumor vascular cells are known to express unique antigen profiles, distinct from those of vascular cells in non-cancerous tissues [12, 13]. A cancer vaccine that activates a T-cell-mediated immune response to VEGF receptor 2 has been reported [14]. Immunotherapy targeting various tumor vasculature-specific antigens has been attempted using vaccination with xenogeneic vascular cells [15]. Although immunotherapy using tumor vascular cells as a vaccine might be an attractive approach to tumor vessel-specific immunotherapy, the collection of vascular cells from the tumor tissue of individual patients is not feasible.

Upon the initiation of differentiation into cardiovascular cells, iPS cells show the phenotype of vascular progenitor cells (induced vascular progenitor cells or iVP cells) [16]. These iVP cells could possibly be utilized as a cancer vaccine that could induce an immune response to tumor vascular cells and suppress the development of tumor vasculature. If so, a unique cancer vaccine could be generated from iPS cells because abundant iVP cells could easily be obtained from iPS cells, and lysates of iVP cells could be made available for clinical use.

In the present study, the potential use of iVP cells as a cancer vaccine that suppresses tumor vascular development was investigated using a mouse tumor model.

Materials and methods

Cells and reagents

The iPS cells used in the present study were iPS-MEF-Ng-20D-17 cells, which were established by Prof. Shinya Yamanaka [1]. CMS-4 fibrosarcoma cells [17, 18] were kindly provided by Dr. Tatsumi (Osaka University, Japan). bEnd.3 murine endothelial cells [19] were purchased from ATCC. Female BALB/c mice were purchased from Sakyo Labo Service (Tokyo, Japan). Monoclonal antibodies against mouse CD4 and CD8 were purified from ascites generated by the implantation of the hybridoma cell lines TIB120 and GK1.5, respectively, into severe combined immunodeficiency mice (Japan SLC, Hamamatsu, Japan).

Animal experiments

The animal studies were performed according to the recommendations in the Guide for the Care and Use of Laboratory Animals of the National Institutes of Health. The protocol was approved by the Committee on the Ethics of Animal Experiments of Jikei University (Permit Number: 21-050C2) and followed the ARRIVE guidelines. The condition of the mice was monitored twice per week, and the mice were humanely euthanized by treatment with an excess amount of ethyl ether when the tumor size reached more than 1,200 mm² (long diameter × short diameter). Vaccination with antigen-pulsed DCs and the implantation of the tumor cells were performed using a suitable mouse restrainer to minimize suffering.

Culture of iPS cells and induction of differentiation into iVP cells

Mitomycin C-treated or irradiated SNL cells (2 × 10⁶/dish) were seeded on gelatin-coated 100-mm tissue culture dishes. After overnight incubation, iPS cells (1 × 10⁶/dish) were seeded on the plates and cultured in EmbryoMax DMEM—high glucose, low bicarbonate w/o sodium pyruvate (Chemicon, Temecula, CA, USA)—containing 15 % EmbryoMax ES cell-qualified fetal bovine serum (Chemicon), 1,000 U/ml recombinant mouse (rm) leukemia inhibitory factor (LIF) (Chemicon), 0.1 mM EmbryoMax ES cell-qualified MEM with nonessential amino acids, and 0.1 mM EmbryoMax ES cell-qualified 2-mercaptoethanol (Chemicon).

Differentiation from iPS cells into iVP cells was induced according to the method described by Narazaki et al. [16]. iPS cells (2 × 10⁵/dish) were seeded on 100-mm tissue culture dishes coated with collagen type IV (BD Bioscience, San Jose, CA, USA) and cultured in MEM-alpha without feeder cells and LIF. After 4.5-day culture, the cells were detached by trypsin/EDTA treatment.

Collection of DCs and pulse with cell lysate

DCs were prepared according to the method described previously, with certain modifications [20]. Briefly, the adherent fraction of mouse bone marrow cells was cultured in medium containing granulocyte macrophage-stimulating factor (GM-CSF, PeproTech, Rocky Hill, NJ, USA) and interleukin (IL)-4 (PeproTech) for 5 days, and non-adherent and loosely attached cells were collected and used as DCs. Lysates of iPS cells, iVP cells, and CMS-4 cells (10⁷/ml in PBS) were generated by treatment with freezing and thawing (three times) and sonication. DCs were pulsed with the lysate of each cell type for 24 h and then treated with 0.01 U/ml OK-432 (Chugai Pharmaceutical, Tokyo, Japan) for 24 h for DC maturation.

Tumor prevention by vaccination with the lysate-pulsed DCs

Female BALB/c mice at an age of 8–10 weeks were inoculated subcutaneously with the lysate-pulsed DCs (10⁶/mouse) on Day −14 and Day −7. On Day 0, tumor cells (10⁶/mouse) were implanted under the dorsal skin. The tumor size was measured once per week and expressed as (long diameter in millimeters) × (short diameter in millimeters). In certain experiments, mice were injected intraperitoneally with 0.2 mg of mAb to mouse CD4 (GK1.5) or CD8 (TIB120), anti-asialo GM1 antibody (30 μl/mouse, Wako Pure Chemical Industries, Osaka, Japan), or 0.2 mg of control rat IgG (Sigma-Aldrich, St. Louis, MO, USA) on Days −4, −1, 0, 3, 7, and 10.

Immunohistochemistry and tomato-lectin perfusion study

For CD31 immunohistochemical observation, deparaffinized slides were treated in a microwave oven for 15 min and then treated with rat anti-mouse CD31 mAb (SZ 31, Histonova, Hamburg, Germany) at 4 °C overnight. After washing with PBS, the cells were treated with biotinylated anti-rat IgG at room temperature for 30 min. Next, the cells were treated with the ABC mixture of the VECTASTAIN Elite ABC KIT STANDARD (Vector Laboratories, Burlingame, CA, USA). The cells were treated with ImmPACT DAB (Vector Laboratories) at room temperature for 10 min and then washed, dried, and mounted.

For alpha-smooth muscle actin (ASMA) immunohistochemical observation, deparaffinized slides were treated with mAb to antihuman ASMA (1A4, Dako, Glostrup, Denmark) at a 1:200 dilution. Immunohistochemical staining was performed using an iVIEW DAB Detection Kit (Ventana, Tucson, AZ, USA) and BenchMark XT (Ventana) according to the manufacturer’s instructions.

For the tomato-lectin perfusion study, anesthetized tumor-bearing mice were injected intravenously with fluorescein-labeled Lycopersicon esculentum (tomato) lectin (Vector Laboratories) (50 μg/250 μl) via the tail vein. Frozen sections of tumor tissue were observed under a fluorescent microscope.

Cytotoxicity of splenic CD8⁺ T cells

Spleen cells were obtained from each immunized mouse and incubated with lysate and rhIL-2 (25 U/ml) for 4 days. After treatment with DNase, splenic CD8⁺ T cells were collected by negative selection using magnetic sorting according to the manufacturer’s instructions (Miltenyi Biotec, Bergisch Gladbach, Germany). A cytotoxicity assay was performed using a standard 4-h culture ⁵¹chromium (⁵¹Cr) release assay, as described elsewhere [21].

Isolation of Flk-1⁺ cells from CMS-4 tumor tissue

Established CMS-4 tumors were resected; cut into small pieces; and then digested by treatment with 0.05 % collagenase (Worthington, Lakewood, NJ, USA), dyspase (Sankojunyaku, Tokyo, Japan), and DNase (Sigma-Aldrich). After washing with PBS, the dyspase-treated cells were treated with anti-Flk-1 mAb labeled with phycoerythrin (PE). The cells were further treated with anti-PE antibody conjugated to microbeads. Flk-1⁺ cells were positively isolated by passage through a column in a magnetic field according to the manufacturer’s instructions. More than 90 % of the cells obtained were PE-positive cells.

DNA microarray analysis

Total RNA from cultured cells was isolated using an RNeasy Plus Mini Kit (Qiagen, Hamburg, Germany) according to the manufacturer’s protocol. Pooled RNAs (500 ng) were subjected to cRNA synthesis using an Illumina TotalPrep RNA Amplification Kit (Life Technologies, Carlsbad, CA, USA) according to the manufacturer’s instructions. Hybridization of the cRNA (1,500 ng) to an Illumina MouseWG-6 version 2.0 Expression BeadChip (Illumina, Inc. SanDiego, CA, USA) was performed for 16 h. Washing and staining after hybridization were performed according to the protocol of the Illumina Direct Hybridization Assay. Data collection was performed using iScan (Illumina, Inc), and signal data were analyzed by GenomeStudio (Illumina). The fold change between two samples had to be at least 3 to identify a transcript as altered.

Statistical analysis

A Student’s t test was performed using Microsoft Office Excel 2007 (Microsoft Corporation, Redmond, WA, USA). When p values were 0.05 or less, differences were considered statistically significant.

Results

Suppression of tumor incidence by immunization of mice with DC/iVP

Immunization of mice with DC/iVP showed a significant prophylactic effect against implanted CMS-4 cells (Fig. 1a). Although control mice implanted with CMS-4 cells showed 100 % (8/8) tumor incidence in 2 weeks, mice immunized with DC/iVP showed only 25 % (2/8) tumor incidence during a 7-week observation period (p < 0.05, Student’s t test). Importantly, prevention by immunization with DC/iVP was more potent than by immunization with DC/iPS or DC/CMS-4. Furthermore, immunization with DC/iVP cells prolonged the survival of CMS-4-implanted mice (Fig. 1b). In total, 50 % of DC/iVP-immunized mice were still alive without tumor development 6 months after tumor implantation. Although 50 % of mice immunized with DC/CMS-4 were tumor-free during the 7-week observation period, all of the mice showed tumor development after that and finally died. Similarly, 50 % of mice immunized with DC/iPS showed tumor development in the 7-week observation period, but 25 % ultimately survived after 6 months of observation. All of the tumor-implanted mice that were not treated died. It is likely that no mice died as a result of the intervention, as all of the surviving mice were healthy, and no bleeding tendency or delayed wound-healing was observed.

Fig. 1.

DC/iVP-immunized mice showed preventive antitumor activity and longer survival. a Female BALB/c mice were untreated or immunized with DCs (10⁶/mice) alone or DCs loaded with the lysate of iPS, iVP, or CMS-4 cells, as described in the “Materials and methods” section (n = 8). CMS-4 cells (10⁶/mouse) were implanted subcutaneously, and the tumor incidence in each treated group was examined thereafter. (*p < 0.05 vs. control, Student’s t test). b Six-month survival of the CMS-4-implanted mice in each treated group (n = 8). The experiments were repeated three times, and similar results were obtained

Marked inhibition of tumor vasculature in DC/iVP-immunized mice

Histological characteristics were compared between tumors that developed in untreated mice (6 weeks after tumor cell implantation) and tiny tumors that tardily developed in DC/iVP-immunized mice (3 months after tumor cell implantation). The tumors that developed in the immunized mice showed markedly impaired tumor vasculature (Fig. 2a). The number of tumor vessels in the immunized mice was found to be significantly decreased, as assessed by counting the number of vessels in 10 randomly chosen microscopic fields (×100) of tumor tissue (Fig. 2b). Furthermore, the cellularity of the tumor tissue decreased, and the tumor cells seemed to be more spindle-shaped compared to the original tumor cells (Fig. 2a).

Fig. 2.

a Development of tumor vessels was impaired in CMS-4 tumors by the DC/iVP vaccination. Left histology of CMS-4 tumors that developed in untreated mice 6 weeks after CMS-4 cell implantation (hematoxylin–eosin, ×100). Right histology of CMS-4 tumors that developed in the DC/iVP pre-immunized mice 3 months after CMS-4 cell implantation (hematoxylin–eosin, ×100). The yellow arrows show representative images of tumor vessels. b Number of vessels in one microscopic field (×100) of the CMS-4 tumor tissue from mice untreated, treated with DC alone or DC/iVP (n = 10, *p < 0.05)

Suppression of tumor vasculature composed of CD31⁺ endothelial cells (Fig. 3a) and ASMA⁺ pericytes (Fig. 3b) was also observed in the tumor tissue of the DC/iVP-immunized mice. The tomato-lectin perfusion study demonstrated a significant decrease in the volume of tumor vessels in the DC/iVP-immunized mice (Fig. 3c) compared with the control mice.

Fig. 3.

Tumor vasculature was impaired in CMS-4 tumor tissue by the immunization of mice with DC/iVP. Tumors that developed in untreated control or DC/iVP pre-immunized mice were examined by immunohistochemical staining. a CD31, b alpha-smooth muscle actin (ASMA). Left untreated mice. Right DC/iVP-immunized mice. c Tumor-bearing mice that were untreated or pre-immunized with DC/iVP were injected intravenously with tomato lectin, as described in the “Materials and methods” section. The tumor tissue of each mouse was observed by fluorescent microscopy. Left untreated mice. Right DC/iVP-immunized mice

DC/iVP vaccination induced potent antitumor activity against vascular-rich tumors

The antitumor activity induced by DC/iVP immunization to the tumors with different vasculatures was investigated. CMS-4 tumors showed the highest level of vascular development among the three tumor types examined. BNL tumors had less vasculature than did the CMS-4 tumors, and C26 tumors had much less. DC/iVP immunization significantly suppressed CMS-4 tumors (Fig. 4a), but the suppression of BNL tumors was not significant (Fig. 4b). Tumor suppression by DC/iVP immunization was very weak in C26 tumors (Fig. 4c).

Fig. 4.

DC/iVP vaccination induced potent antitumor activity against vascular-rich tumors. Left CMS-4 fibrosarcoma (a), BNL hepatoma (b), and C26 colon cancer (c) cells (10⁶/mouse) were implanted subcutaneously into untreated or DC/iVP-immunized mice. The size of the tumors (long diameter x short diameter, mm²) in each group was then measured. Right vascularity of each tumor. Tumors developed by implantation with CMS-4 fibrosarcoma (a), BNL hepatoma (b), or C26 colon cancer (c) cells were stained with hematoxylin–eosin, and the number of vessels observed in one microscopic field per tumor was counted. The number of tumor vessels was counted in 10 randomly chosen fields (×100) in each type of tumor tissue, derived from three animals. *p < 0.05, ** not significant (n = 4)

Treatment of the DC/iVP-immunized mice with mAb against CD4 or CD8 suppressed antitumor activity

Tumor suppression by the DC/iVP immunization was inhibited by the treatment of the immunized mice with mAb against CD4 or CD8, but not anti-asialo GM1 Ab (Fig. 5), indicating that the tumor suppression induced by DC/iVP immunization was T-cell-mediated.

Fig. 5.

Depletion of CD4⁺ and CD8⁺ T cells inhibited the antitumor activity induced by vaccination with DC/iVP. Mice were untreated or immunized with DC/iVP on Days −14 and −7, and CMS-4 cells (10⁶/mouse) were implanted on Day 0. Anti-CD4 mAb, anti-CD8 mAb, or anti-asialo GM1 Ab was injected intraperitoneally on Days −4, −1, 0, 3, 7, and 10. The size of the tumors in each group was measured (n = 4). This experiment was repeated twice, and similar results were obtained. *p < 0.05

CD8⁺ T cells from the DC/iVP-immunized mice showed cytotoxic activity against murine endothelial cells and CMS-4 tumor cells

As shown in Fig. 6a, CD8⁺ T cells from the DC/iPS-immunized mice did not show significant cytotoxic activity against endothelial cells or CMS-4 cells. In contrast, CD8⁺ T cells from the DC/iVP-immunized mice showed highly cytotoxic activity against murine endothelial cells (Fig. 6b) and significant cytotoxic activity against CMS-4 cells.

Fig. 6.

CD8⁺ T cells from the DC/iVP-immunized mice showed cytotoxicity against CMS-4 tumor cells and bEnd.3 endothelial cells. Spleen cells were obtained from the DC/iPS- or DC/iVP-immunized mice and incubated with the same lysate and rhIL-2 (25 U/ml) for 4 days. After treatment with DNase, splenic CD8⁺ T cells were collected by negative selection using magnetic sorting. Cytotoxic activity against bEnd.3 murine endothelial cells or CMS-4 fibrosarcoma cells was examined using a standard 4-h culture ⁵¹Cr release assay, as described in the “Materials and methods” section. a Cytotoxic activity of CD8⁺ T cells from the DC/iPS-immunized mice. b Cytotoxic activity of CD8⁺ T cells from the DC/iVP-immunized mice. The experiments were repeated twice, and similar results were obtained

DNA microarray analysis indicated genes possibly encoding a T-cell target antigen expressed in tumor vascular cells

As the antitumor immunity induced by the DC/iVP immunization was significantly higher than that induced by the DC/iPS immunization, the gene products induced by differentiation from iPS cells into iVP cells should have been closely associated with the molecules recognized by the induced T cells. Common genes (Fig. 7, area B) between genes enhanced by differentiation from iPS cells into iVP cells (Fig. 7, area A plus B) and common genes between iVP cells and isolated tumor vascular cells (Fig. 7, area B plus C) were investigated by DNA microarray analysis. In total, 29 genes were identified in area B in the gene ontology of “vascular.” Catenin-beta; platelet-derived growth factor receptor (PDGFR)-beta peptide; fibroblast growth factors 8, 10, and 18; and LIF were listed as candidate target molecules of the DC/iVP-induced immune response. Lists of the genes in areas A, B, and C are shown as supplementary data (supplementary Tables 1, 2, and 3, respectively).

Fig. 7.

Two Venn diagrams of DNA microarray analysis, indicating possible T-cell target gene products. Area A + B shows genes enhanced by differentiation from iPS cells into iVP cells, and area B + C shows genes shared between iVP cells and isolated vascular cells from CMS-4 tumor tissue

Discussion

iVP cells, which are Flk-1⁺ and E-cadherin⁺ common progenitors of vascular cells derived from undifferentiated iPS cells, have the potential to differentiate into various vascular cell types [16]. In response to VEGF, these cells can differentiate into vascular endothelial cells. Alternatively, iVP cells that lose Flk-1 expression can differentiate into vascular mural cells in response to platelet-derived growth factor (PDGF) [16]. The cells can further differentiate into cardiac progenitor cells exhibiting pulsation or primitive hemangioblastic cells that generate CD45⁺ blood cells. Accordingly, iVP cells might have characteristics similar to those of vascular progenitor cells and presumably express antigens shared with vascular progenitor cells. Vaccination of mice with iVP cells significantly suppressed the development of CMS-4 tumors and prolonged the survival of tumor-bearing mice. CMS-4 cells have vigorous proliferative activity and well-developed tumor vasculature. Tumors that were tardily generated in the DC/iVP-vaccinated mice exhibited markedly impaired tumor vasculature. Importantly, the DC/iVP vaccination induced stronger antitumor activity against CMS-4 cells than did the DC/iPS immunization. These results suggest that an immune response against vascular progenitor cells might have been induced by vaccination with iVP cells, and consequently, the development of tumor vasculature might have been suppressed by the immune response. Coarse cellularity and loose cell–cell contact, with more spindle-shaped tumor cells in the CMS-4 tumors, developed in the DC/iVP-immunized mice. This finding suggests that epithelial–mesenchymal transition-like changes might have been induced by the hypoxic conditions caused by the inhibition of the vasculature [22].

Vasculogenesis (the generation of vasculature) and angiogenesis (the development of vasculature) are essential factors in tumor growth. Endothelial progenitor cells (EPCs) are known to be a rare population that circulates in the blood and has the ability to differentiate into endothelial cells [23], suggesting that iVP cells have similar characteristics and roles as EPCs. EPCs also play important roles in vasculogenesis and angiogenesis for tumor development [24, 25]. EPS-like cells in the CMS-4-implanted mice might have been damaged by the cytotoxic T lymphocytes (CTLs) induced by the DC/iVP immunization. Notably, marked lymphocyte infiltration was not observed in the tumors that slowly developed in DC/iVP-immunized mice, suggesting that vasculogenesis at an initial stage of tumor development, but not angiogenesis, was suppressed by DC/iVP immunization.

In a large variety of tumor types, the vasculature is an independent prognostic factor for overall survival and disease-free survival [26], indicating that vascular-rich tumors have more malignant phenotypes. Trans-catheter arterial embolization therapy for hepatocellular carcinoma treatment is significantly more effective against hypervascular tumor types than against hypovascular tumors [27]. These results indicate that hypervascular tumors receive sufficient blood flow to fuel rapid growth and that their malignant phenotypes might be closely associated with rich vasculature. CMS-4 tumors showed more rapid growth than did the other two tumor types and induced more abundant vessels than did BNL and C26 tumors, suggesting that the survival of CMS-4 cells might be more dependent on blood flow. BNL and C26 tumors growing in a hypovascular environment might be able to survive under more hypoxic conditions compared with CMS-4 tumors. Accordingly, it is conceivable that CMS-4 tumors are more vulnerable to immune-related, vascular damage-causing hypoxic conditions than are BNL and C26. Thus, an iVP-based cancer vaccine might be suitable for the prevention of the recurrence of hypervascular tumor types, such as sarcomas, in possible clinical use. On the other hand, if iVP cells express various vascular antigens to induce specific immunity to vascular progenitor cells, this therapy might be effective to many kinds of highly vascular tumors unrelated to the antigenicity of the tumor cells. This might be one of the advantageous points of iPS-derived vaccine.

A cytotoxic assay using CD8⁺ T cells from iVP-immunized mice demonstrated that the CTLs induced by the iVP vaccination exhibited cytotoxic activity against not only endothelial cells but also CMS-4 tumor cells. Although not identified in the present study, there should be common gene expression between CMS-4 and iVP cells. It is well known that common gene expression, such as VEGFR and PDGFR, is observed between tumor cells and tumor vascular cells [28]. Recently, it was reported that tumor vascular cells could be generated from tumor cells [29, 30], suggesting that tumor vascular cells and tumor cells might have similar antigenicity. Tumor suppression by DC/iVP vaccination was potent and even stronger than DC/CMS-4 immunization, indicating that iVP vaccination induced both anti-vascular and antitumor immune responses. However, we do not know how much an anti-vessel immune response was involved in this antitumor activity. It is not excluded that tumor cells with low tumor vessel induction were not affected by the immune response. Furthermore, it should particularly be noted that the iVP cells expressed the alloantigen H-2^b, which derived from B6/129 mice and was incorporated during generation of the iPS cells. Expression of alloantigen by the iVP cells might have stimulated a T-cell response and might have been involved in the induction of antitumor immunity.

As suppression of the antitumor activity by anti-CD4 mAb treatment was stronger than suppression by anti-CD8 mAb treatment, the response of CD4⁺ T cells from iVP-immunized mice to the antigens of iVP cells and CMS-4 cells was examined by ELISPOT assay. The CD4⁺T-cell response to CMS-4 cells was stronger than the response to iVP cells (data not shown), suggesting that DC/iVP-primed CD4⁺T cells might have responded more actively to antigens of CMS-4 cells than that of tumor vessels in vivo.

Flk-1⁺ cells have the characteristics of vascular progenitor cells [31]. Gene expression profiles were compared between Flk-1⁺ cells in CMS-4 tumor tissue and iVP cells for the identification of the target antigens of the iVP vaccination. DNA microarray analysis demonstrated several candidate antigens that might be targets of the iVP-induced immune response. In the gene ontology of the vasculature, 125 genes shared between iVP cells and isolated Flk-1⁺ cells were found, indicating that these cells types have considerably similar vasculature-related gene expression profiles. Among these genes, 29 genes were upregulated by differentiation from iPS cells into iVP cells. Among the 29 candidate genes, PDGFR-beta peptide is an interesting candidate because PDGFR is an important factor in the recruitment of pericytes, which is a crucial process in the formation and stabilization of mature blood vessels [32]. Intriguingly, Kaplan et al. reported that a recombinant DNA vaccine encoding PDGFR-beta promoted the immune-mediated loss of pericytes in PDGFR-beta-negative tumor tissue and prolonged survival [33]. The tumors that developed in the iVP-vaccinated mice showed a decreased number and disarrangement of pericytes, suggesting that immune-related damage to pericytes might be involved in the failure of vascular development. As it has been reported that the expression of cell surface protein is similar between pericytes and endothelial precursor cells [34], vaccination with iVP cells might have activated an immune response against both pericytes and endothelial cells.

Electronic supplementary material

Below is the link to the electronic supplementary material.