CD8α+ dendritic cells potentiate antitumor and immune activities against murine ovarian cancers

Dendritic cell (DC)-based immunotherapies have been shown to be a potential treatment option for various cancers; however, the exact strategies in ovarian cancer remain unknown. Here, we report the effectiveness of mouse CD8α+ DCs derived from bone marrow hematopoietic stem cells (BM-HSCs), equivalent to human CD141+ DCs, which have proven to be a highly superior subset. Mono-DCs from monocytes and stem-DCs from HSCs were characterized by CD11c+ CD80+ CD86+ and CD8α+ Clec9a+ expression, respectively. Despite a lower dose compared with Mono-DCs, mice treated with pulsed Stem-DCs showed a reduced amount of ascitic fluid and lower body weights compared with those of vehicle-treated mice. These mice treated with pulsed stem-DCs appeared to have fewer tumor implants, which were usually confined in the epithelium of tumor-invaded organs. All mice treated with DCs showed longer survival than the vehicle group, especially in the medium/high dose pulsed Stem-DC treatment groups. Moreover, the stem-DC-treated group demonstrated a low proportion of myeloid-derived suppressor cells and regulatory T cells, high interleukin-12 and interferon-γ levels, and accumulation of several tumor-infiltrating lymphocytes. Together, these results indicate that mouse CD8α+ DCs derived from BM-HSCs decrease tumor progression and enhance antitumor immune responses against murine ovarian cancer, suggesting that better DC vaccines can be used as an effective immunotherapy in EOC treatment. Further studies are necessary to develop potent DC vaccines using human CD141+ DCs.

. Workflow for the ovarian cancer orthotopic model using bone marrow-derived dendritic cells. We performed three independent experiments to evaluate the efficacy of mouse CD8α+ DCs originating from mouse BM stem cells, which correspond to human BDCA3+ (CD141+) DCs. An orthotopic model was established by intraperitoneally injecting ID8 mouse ovarian cancer cells. Six different treatments, including experimental groups of three doses of Ag-pulsed Stem-DCs, a comparative group of Ag-pulsed Mono-DCs, and two kinds of control groups treated with vehicle and unpulsed Stem-DCs, were analyzed. The immune responses before and after DC treatments and mouse reactions related to tumor formation and treatments were evaluated during and after DC treatments. www.nature.com/scientificreports/ Flow cytometry analysis of differentiated Mono-DCs revealed the expression of DC surface markers, including CD11c, CD80, CD86, MHC I, and MHC II. In addition, no expression of CD8α or Clec9a was observed, which are the intrinsic characteristics of mouse stem-DCs, indicating CTLs and myeloid lineage cells, respectively (Fig. 2B). Furthermore, immunostimulatory cytokines, such as IL-12 and IFN-γ, were secreted from mature Mono-DCs, and T cell proliferation was confirmed in proportion to the number of Mono-DCs (Fig. 2C,D). In contrast to Mono-DCs, the typical phenotype of mouse Stem-DCs induced by SCF and Flt3L was clearly confirmed as CD8α+ Clec9a+ cells (Fig. 2E). Similar to Mono-DCs, we identified not only the secretion of IL-12 and IFN-γ but also T cell proliferation induced by mature Stem-DCs (Fig. 2F,G).

DC treatments affect tumor growth and ascites formation. We harvested all mice treated with
Stem-DCs, Mono-DCs, and the vehicle control between 7 and 8 weeks after the intraperitoneal injection of 5 × 10 6 ID8 cells, although some mice in the vehicle control died earlier than we expected before the last DC injection because of tumor progression (Fig. 3). Most mice appeared to form ascites from 4 weeks after ID8 cell injection, which was confirmed by increasing abdominal circumferences. We observed significantly smaller body weights in the high-and medium-dose pulsed Stem-DC groups than in the vehicle group (P = 0.0323 and P = 0.0183, respectively). There were no differences in body weights according to the type of DC treatment among the DC-treated groups. When we compared the volume of ascitic fluid carefully collected at harvesting, all of the DC-treated mice showed a reduced volume of ascites. (F) IL-12 and IFN-γ secretion by differentiated stem-DCs with or without maturation factors was tested by ELISA immediately after culture (P = 0.0143 and P = 0.0286, respectively). (G) T cell proliferation (MLR assay) with stem-DCs in the presence or absence of maturation factors showed a significant difference compared with T cells only (P = 0.0012). The asterisks represent a statistically significant difference (*, P < 0.05; **, P < 0.01; ***, P < 0.001). DC treatments increase the survival of tumor-bearing mice. When survival was compared in six mouse groups, including the vehicle control, there was a significant difference in survival. (P = 0.0187) (Fig. 5). All mice injected with any type of DC demonstrated longer survival than the vehicle group. For instance, there were significant differences in the medium-and high-pulsed Stem-DC groups compared with the vehicle group (P = 0.0323 and P = 0.0183, respectively). There was no difference in comparisons between mice injected with DCs.
Immune responses contribute to the significant effect of DC treatments. To understand changes in the immune environment after DC treatments that are associated with tumor growth and survival, we investigated immune reactions in blood samples, ascites, and tumor tissues obtained after treatments with vehicle, high-dose Stem-DCs, and Mono-DCs (Fig. 6). The proportion of CD11b+ Gr1+ MDSCs was low in the serum samples of the pulsed Stem-DC group compared with the Mono-DC group (P = 0.0141), and the proportion of Tregs was significantly lower in the serum of the pulsed Stem-DC group compared with the vehicle and Mono-DC groups (P = 0.0434 and P = 0.0132, respectively). The immunosuppressive marker IL-10 was secreted at lower levels in the serum of the pulsed Stem-DC group (P = 0.04199), whereas immunostimulatory cytokines, includ- www.nature.com/scientificreports/ ing IL-12 and IFN-γ, showed no difference in the serum of mice according to the type of DC treatment. Unlike in serum samples, the levels of IL-10, IL-12, and IFN-γ were significantly different in the ascitic fluid of the pulsed Stem-DC group (P = 0.0347, 0.0073, and 0.0005, respectively). We did not observe different immune reactions in the spleen according to DC treatments (Fig. S2). Finally, we performed IHC staining to confirm the distribution of immune cells that are recruited in response to the tumor. The expression levels of CD3, CD4, CD8, and CD11c were significantly higher in the pulsed Stem-DC group (P = 0.0008, 0.0006, < 0.0001, and < 0.0001, respectively), and Ki67 expression was significantly lower in the pulsed Stem-DC group (P = 0.0028).

Discussion
We evaluated the antitumor and immune responses of murine CD8α+ DCs. Among the characterized human DC subpopulations, cDC1s have emerged as a highly desirable tool for enhancing antitumor immunity, but existing approaches exploring cDC1s for antitumor immunotherapies have been limited to a few cancer types, such as melanoma, colon cancer, and lung cancer 16 . In this sense, our results are very encouraging and suggest a potentially effective new immunotherapy for ovarian cancer. The mouse CD8α+ Clec9a+ DCs (or stem-DCs) produced in this study demonstrated antitumor effects characterized not only by a reduced volume of ascites and number of tumor implants but also high levels of immunostimulatory cells and cytokines and low levels of immunosuppressive cytokines. Of course, all DC treatments have shown significant effects on the survival of tumor-bearing mice, but these antitumor and immune responses were significant even at relatively low therapeutic doses of Stem-DCs compared with Mono-DCs. DC vaccines are the preferred approach to be implemented as a personalized cell-based therapy for cancer treatment, with more than 300 completed or ongoing registered clinical trials conducted to evaluate their application for boosting antitumor immunity 16 . Our authors have previously reported a phase I/II trial of a therapeutic DC vaccination with IL-2 as a consolidation therapy for ovarian cancer patients 17 . In that phase I/II study, 10 EOC patients with minimal residual disease after initial debulking and chemotherapy were treated with two subcutaneous doses of autologous monocyte-derived DCs pulsed with autologous tumor lysates and IL-2 to evaluate the safety and feasibility of this therapeutic strategy and characterize the antigen-specific immune alterations induced by this treatment. As a result, this DC vaccination was well tolerated and induced tumor-related www.nature.com/scientificreports/ immunity potentially associated with long-term clinical responses against EOC. In the three patients with longterm survival for 83, 81, and 38 months after DC vaccination without disease relapse, significant immune alterations were observed, including increased natural killer (NK) activity, IFN-γ-secreting T cells, immunostimulatory cytokine secretion, and reduced immunosuppressive factor secretion. Based on the DC studies of EOC published to date, many studies have focused on the selection of the antigen source. Hydrostatic pressure lysate preparation appears to be a crucial factor for DC vaccines to improve their efficacies, especially in ovarian cancer. Fucikova et al. reported that the high hydrostatic pressure used for preparing tumor lysates could induce immunogenic cell death, enhance DC uptake and cytokine release, and activate T cells in ovarian cancer 18 . In a recently reported phase I trial of a DC vaccine generated by the differentiation of autologous monocytes pulsed with oxidized autologous whole-tumor cell lysates in platinum-treated, immunotherapy-naïve, recurrent ovarian cancer patients, the DC vaccine was administered intranodally either alone, in combination with bevacizumab, or in combination with bevacizumab and low-dose intravenous cyclophosphamide until disease progression or vaccine exhaustion 19 . This vaccine induced T cell responses (increased IFN-γ production), and the antitumor immune response was associated with significantly prolonged survival, which was similar to our previous phase I trial 17 . Taken together, there are limited results of DC vaccinations for EOC. As a result, DC vaccines have not yet been approved as a standard treatment.
Although DC vaccination was introduced as a promising therapy after Sipuleucel-T approval in 2010, DCbased immunotherapies for the treatment of malignancies have generally shown limited clinical benefit. Among reported clinical trials with DC vaccines, the most common approach relies on the use of ex vivo differentiated DCs from leukapheresis-isolated monocytes cultured in the presence of GM-CSF and IL-4. However, although these DC vaccines are well tolerated, responses are achieved in less than 15% of patients 12,16 . These unsatisfactory outcomes are linked to immunosuppressive tumor microenvironments, such as checkpoint receptor signaling (PD-1/PD-L1, CTLA-4), and immune suppressive cells, such as Tregs and MDSCs. Moreover, in addition to TME characteristics, there are several limitations to the production of effective DC vaccines, including inadequate antigenic stimulation, suboptimal cell maturation 20 , lack of proliferative potential of DCs 21 , and expensive and complex manufacturing processes 22 . Therefore, the identification of a specific DC subset functionally skilled to achieve an effective antitumor immune response is essential.

Figure 5.
Comparative survival analysis according to DC treatments. We identified significant differences in survival in the vehicle group (P = 0.0187). All groups injected with DCs demonstrated longer survival than the vehicle group. For instance, there were significant differences in the medium-and high-pulsed Stem-DC groups compared with the vehicle group (P = 0.0323 and P = 0.0183, respectively). However, there was no difference according to the type of DC injected. This survival probability was analyzed by the log-rank (Mantel-Cox). www.nature.com/scientificreports/ However, cDC1s are very rare and account for < 0.03% of peripheral blood cells in humans 15 . Therefore, although there are no clinical studies with isolated cDC1 populations to date, CD141+ DCs are worth investigating for the development of potential therapeutic vaccines. There are several approaches to perform ex vivo differentiation of CD141+ DCs from human CD34+ progenitors using Flt3L, GM-CSF, SCF, thrombopoietin (TPO), IL-6, IL-3, and IL-4 15,16,28,29 . Moreover, preclinical studies have explored the use of cDC1s in some types of cancers using syngeneic mouse models and mouse CD8α+ DCs that are putative equivalents to human CD141+ DCs 30 . This present study is the first preclinical study to successfully show the possibilities of human cDC1 populations as an immunotherapy against ovarian cancer. We performed these experiments with well-differentiated mouse CD8α+ DCs supported by Flt3L, GM-CSF, SCF, and IL-4 in addition to pulsing by tumor cell lysates, which has turned out to be the most effective approach to produce CD141+ cDC1s. As a result, mice injected with pulsed Stem-DCs showed a smaller amount of ascites and tumor implants compared with those of the vehicle group, even at a relatively low injected DC dose (P < 0.05). This finding is supported by the absence of differences in pulsed Mono-DC injected mice. Furthermore, reduced tumor formation was associated with high expression of CD3, CD4, CD8, and CD11c in the IHC assay (all P < 0.001) and a reduced proportion of MDSCs and Tregs. In addition, low IL-10 and high IL-12 and IFN-γ levels were observed, which were significant in ascitic fluid. The immune response and tumor response after stem-DC treatments demonstrated the effect of murine CD8α+ DCs against ovarian cancer.
There are limitations to the present study that prevent the prediction of human CD141+ DCs as an effective EOC treatment. Most importantly, although our data indicate that CD8α+ DCs induce immune responses against ovarian cancer in the mouse model, the extrapolation to the human setting seems too far stretched. Therefore, further studies to investigate mouse CD8α+ DCs and human CD141+ DCs are necessary. Second, Unlike in the serum, ascites was significantly different in the pulsed Stem-DC group in terms of the secretion of cytokines, including IL-10, IL-12, and IFN-γ (P = 0.0347, 0.0073, and 0.0005, respectively). (E) We performed IHC staining to compare the number of representative immune cells. The expression levels of CD3, CD4, CD8, and CD11c were significantly higher in the pulsed Stem-DC group (P = 0.0008, 0.0006, < 0.0001, and < 0.0001, respectively). In contrast, the Ki67 level was significantly lower in the pulsed Stem-DC group (P = 0.0028). The asterisks represent a statistically significant difference (*, P < 0.05; **, P < 0.01; ***, P < 0.001). www.nature.com/scientificreports/ although we performed three independent experiments, there were only six mice in each treatment group. Last, because of the nature of the orthotopic model, tumor progression and immune monitoring after DC treatment were evaluated to a limited extent. It was feasible but not objective to compare the quantitative weight or size of tumors because our experimental model was intraperitoneal, so numerous small tumor implants inside the entire peritoneal cavity could not be counted exactly, and immune responses related to DC treatments were evaluated in restricted methods. Nevertheless, to the best of our knowledge, this is the first study to demonstrate the effectiveness of murine CD8α+ DCs in an ovarian cancer mouse model, suggesting that cDC1 vaccines represented by CD141+ Clec9a+ cells may be an effective immunotherapy in patients with ovarian cancer.
In conclusion, mouse CD8α+ DCs derived from BM-HSCs decrease tumor progression and enhance antitumor immunogenicity against murine syngeneic ovarian cancer. These findings suggest that a better understanding and development of DC vaccines can increase the effectiveness of ovarian cancer immunotherapy. Our results may provide useful information for the future development of more effective DC vaccines using human CD141+ DCs.

Methods
Study approval. All experimental protocols were approved by the Institutional Animal Care and Use Committee (IACUC) of the Asan Institute for Life Sciences. All methods were carried out in accordance with the guidelines and regulations of the IACUC of the Asan Institute for Life Sciences (protocol code 2017-12-082).

Ex vivo generation of stem-DCs and mono-DCs. Isolation of BM-MNCs. Two different subtypes of
DCs were generated from bone marrow mononuclear cells (BM-MNCs). BM-MNCs were obtained from the tibia and femur of C57BL/6 mice sacrificed by cervical dislocation. All experimental procedures were performed in accordance with the regulations of the Institutional Animal Care and Use Committee at the Asan Institute for Life Sciences.
Tumor antigen pulsing and DC maturation. For both stem-DCs and mono-DCs, ID8 ovarian cancer cell lysates as a tumor antigen were added to the culture media 2 days before the DC harvest. One day after tumor cell lysate pulsing, ginsenoside Rg3 (Sigma-Aldrich, St Louis, MO, USA) for stem-DCs and lipopolysaccharide (LPS) (Sigma-Aldrich, USA) for mono-DCs were added to the culture media as maturation factors for 24 h before the harvesting of DCs. Tumor cell lysates were prepared from cultured ID8 cells by a freeze-thaw process that was repeated six times in liquid nitrogen (− 180 °C) and an incubator (37 °C). Protein quantification of the tumor cell lysate was performed using centrifuged (1800 rpm, 10 min) supernatant by the Bradford method (Bio-Rad Laboratories, Hercules, CA, USA). Cultured cells were characterized as DCs by specific marker phenotyping, naïve T cell proliferation induced by DCs, and cytokine secretion.
Formation and follow-up of the ovarian cancer orthotopic model and DC treatments. The orthotopic model of ID8 cancer cells was generated by injecting 5 × 10 6 cells into the peritoneal cavity of C57BL/6 mice. Four weeks after cell injection, mice were categorized into six groups: the vehicle, low-dose pulsed Stem-DC, medium-dose pulsed Stem-DC, high-dose pulsed Stem-DC, pulsed Mono-DC, and unpulsed Stem-DC groups. Mice were subcutaneously injected with 5 × 10 6 DCs for the high Stem-DC group, 5 × 10 4 DCs for the low Stem-DC group, and 5 × 10 5 DCs for the medium Stem-DC group around the inguinal lymph nodes. DCs were injected once a week for three weeks. We continued to assess the tumor status in mice by measuring body weight and abdominal circumference every two days after ID8 cell injection. One week after the last DC injection, mice were sacrificed, and tissues, such as the bilateral ovaries, peritoneum, diaphragm, and spleen, and several tumor implants were harvested and processed for histological analysis. At the time of harvest, we first carefully collected total ascitic fluid and blood samples from the abdominal vein.
Immune response confirmation of generated DCs and DC treatments. Immunophenotypic analysis. Immunophenotypic analyses were performed to assess the characteristics of DCs and immune status after DC treatments by flow cytometry. The lymphocytes from mouse blood samples were prepared and purified by density gradient separation using lymphocyte separation medium (Mediatech, Manassas, VA, USA). Subsequently, the lymphocytes were stained with fluorescently labeled antibodies for 40 min at 4 °C. The following antibodies from eBioscience (San Diego, CA, USA) were used to confirm the characteristics of cultured DCs: In addition to immune cell markers, we investigated tumor markers using anti-Ki67 (ab16667) and anti-p53 (ab131442) primary antibodies purchased from Abcam.
Analysis of IFN-γ producing cells using enzyme-linked immunospot (ELISPOT) assays. The ELISPOT assay was used to detect and count individual cells that secreted IFN-γ protein in vitro upon exposure to an antigen. ELIS-POT IFN-γ assay kits were purchased from AID (Strassberg, Germany) and used according to the manufacturer's instructions. Splenic lymphocytes from tumor-bearing mice with or without DC treatment were stimulated in vitro with tumor lysates on a precoated 96-well plate. The plate was incubated for 20 h at 37 °C with 5% CO 2 . After washing, a detection antibody was added to each well and incubated for 2 h at room temperature. The plate was incubated with an alkaline phosphatase conjugate and developed with a BCIP/NBT substrate solution. Visible spots were counted using an automated AID ELISPOT reader (AID, Strassberg, Germany) and the default program.
Statistical analysis. The mean value ± standard deviation was determined from at least three samples from different mice. All experiments were performed three times with DCs produced at different times. To compare means between two groups, the data were analyzed using the Mann-Whitney U test and considered statistically significant when the P value was < 0.05. Multiple group comparisons were analyzed by one-way analysis of variance, followed by a Friedman test and corrected using Dunn's multiple comparison test. Survival was analyzed by the log-rank (Mantel-Cox) test. All statistical analyses were performed using GraphPad Prism 5.0 software (GraphPad software, San Diego, CA, USA).
Ethics approval and consent to participate. The animal study protocol was approved by the Institutional Animal Care and Use Committee (IACUC) of Asan Institute for Life Sciences, and all experiments were performed in accordance with the regulations of IACUC of Asan Institute for Life Sciences (protocol code 2017-12-082 approved on May 02, 2017). All methods are reported in accordance with ARRIVE guidelines (https:// arriv eguid elines. org) for the reporting of animal experiments under ethics approval and consent to participate.

Data availability
The datasets supporting the conclusions of this article are included within the article.