In situ vaccination using unique TLR9 ligand K3-SPG induces long-lasting systemic immune response and synergizes with systemic and local immunotherapy

Although checkpoint inhibitors (CPIs) have changed the paradigm of cancer therapy, low response rates and serious systemic adverse events remain challenging. In situ vaccine (ISV), intratumoral injection of immunomodulators that stimulate innate immunity at the tumor site, allows for the development of vaccines in patients themselves. K3-SPG, a second-generation nanoparticulate Toll-like receptor 9 (TLR9) ligand consisting of K-type CpG oligodeoxynucleotide (ODN) wrapped with SPG (schizophyllan), integrates the best of conventional CpG ODNs, making it an ideal cancer immunotherapy adjuvant. Focusing on clinical feasibility for pancreaticobiliary and gastrointestinal cancers, we investigated the antitumor activity of K3-SPG-ISV in preclinical models of pancreatic ductal adenocarcinoma (PDAC) and colorectal cancer (CRC). K3-SPG-ISV suppressed tumor growth more potently than K3-ISV or K3-SPG intravenous injections, prolonged survival, and enhanced the antitumor effect of CPIs. Notably, in PDAC model, K3-SPG-ISV alone induced systemic antitumor effect and immunological memory. ISV combination of K3-SPG and agonistic CD40 antibody further enhanced the antitumor effect. Our results imply that K3-SPG-based ISV can be applied as monotherapy or combined with CPIs to improve their response rate or, conversely, with CPI-free local immunotherapy to avoid CPI-related adverse events. In either strategy, the potency of K3-SPG-based ISV would provide the rationale for its clinical application to puncturable pancreaticobiliary and gastrointestinal malignancies.


K3-SPG induces type I IFN and
Th1 type immune response more potently than conventional TLR9 ligands. First, the potential of K3-SPG as an effective vaccine adjuvant was compared with two conventional TLR9 ligands, K3 (K-type CpG ODN) and D35 (D-type CpG ODN). Type I IFN plays an essential role in the induction of cancer immunity. K3 is a weak type I IFN inducer, while D35 is a robust type I IFN inducer that is, however, difficult to translate into clinical application due to aggregation. In an in vitro human peripheral blood mononuclear cell (PBMC) experiment, enzyme-linked immunosorbent assay (ELISA) analysis showed that K3-SPG produced higher amounts of IFN-α than K3, which was comparable to that produced by D35 (Fig. 1A, left panel). IL-12, which is also an essential cytokine for antitumor immunity that skews to Th1 response, was produced by K3-SPG as did K3, while D35 only marginally produced it (Fig. 1A, right panel). Expression analysis of mRNA revealed that K3-SPG induced type I and II IFNs and their inducible genes (IFNA, IFNB, IFNG, MXA, and RANTES), DC activation markers (CD80 and CD86), and inflammatory cytokines (IL6 and TNFA) to a greater extent than K3, although not all genes with statistical significance. IL-12 induction was slightly stronger with K3 than K3-SPG (Fig. 1B). The observed discrepancy in the cytokine levels between ELISA and real-time PCR was presumably attributed to the difference in the degradation rates of protein and mRNA.
Consistent with these human PBMC results, in mouse splenocyte experiments, K3-SPG showed higher production of IFN-α and IL-12 than K3 in ELISA ( Supplementary Fig. 1A) and the upregulation of IFNs and their related genes, DC activation markers, and inflammatory cytokines at mRNA expression levels (Supplementary Error bars represent the mean ± SEM. Statistically significant differences were measured by one-way ANOVA followed by the Tukey-Kramer test. *p < 0.05; **p < 0.01; ***p < 0.001. K3-SPG-ISV induces intratumoral type I IFN and Th1 type immune response, suppresses tumor growth, and prolongs survival. We investigated the antitumor effect of K3-SPG-ISV using a syngeneic pancreatic cancer model. C57BL/6 mice were challenged with 2 × 10 6 mouse PDAC cell lines designated as KPC-N, established from Kras LSL-G12D/+ , Trp53 LSL-R172H/+ , and Pdx1-Cre mice (KPC), injected into the left flank, followed by intratumoral injection of either phosphate-buffered saline (PBS) or K3-SPG from 10 days after tumor implantation. K3-SPG-ISV (i.e., K3-SPG-it) significantly suppressed tumor growth and prolonged survival compared to the control treatment ( Fig. 2A). Recapitulating the results of in vitro stimulation experiments, as shown in Fig. 1, mRNA expression analysis of the TME after K3-SPG-ISV clearly revealed the upregulation of type I IFNs (Ifna and Ifnb), IFN-inducible gene (Isg56), IL-12, and DC activation markers (Cd80 and Cd86) (Fig. 2B). Tumor growth suppression and survival prolongation by K3-SPG-ISV were reproduced in another PDAC model using KPF-T cells ( Supplementary Fig. 2) and two CRC models using colon-26-and MC38-bearing mice (Fig. 2C,D, respectively). These results indicated the generality of the antitumor effect of K3-SPG-ISV, with repeated experiment results showing an overall tendency of greater potency on PDAC than CRC models. Next, we investigated the antitumor effect of K3-SPG-ISV compared to that of K3-ISV and intravenous administration of K3-SPG (K3-SPG-iv). K3-SPG-ISV significantly suppressed tumor growth and prolonged survival compared with K3-ISV (Fig. 2E). While our PDAC model of KPC-N-bearing mice was resistant to K3-SPG-iv, K3-SPG-ISV still remarkably suppressed tumor growth and prolonged survival compared with K3-SPG-iv (Fig. 2F). Taken together, these results indicate that K3-SPG-ISV has more potent therapeutic effects than conventional TLR9 ligand K3 or systemic administration of K3-SPG.

Scientific Reports
Antitumor activity of K3-SPG-ISV potentiates the effect of checkpoint blockade therapies. Systemic administration of CPIs, anti-PD-1 and anti-CTLA-4 antibodies, is the standard regimen for cancer immunotherapy. We evaluated the effect of K3-SPG-ISV on CPIs. In PDAC and CRC models, the anti-PD-1 antibody did not or partially suppressed tumor growth, respectively (Fig. 3). In both models, the combination of K3-SPG-ISV and PD-1 blockade enhanced the antitumor effect of each respective monotherapy (Fig. 3). Similarly, while anti-CTLA-4 antibody partially suppressed tumor growth in both models, the combination of K3-SPG-ISV and CTLA-4 blockade showed marked tumor growth inhibition, where the synergistic effect was clear particularly in the CRC model ( Supplementary Fig. 3).

K3-SPG-ISV induces immunological memory.
Motivated by the strong antitumor activity of K3-SPG-ISV, which frequently leads to the complete eradication of engrafted cells in the PDAC model, we investigated whether the mice cured after K3-SPG-ISV treatment developed immunological memory. PDAC mice that were cured after K3-SPG-ISV, followed by rechallenge with the same PDAC cells (i.e., KPC-N cells) 56 days after the first inoculation, completely rejected the replanted tumor cells, whereas all age-matched naïve control mice showed the growth of KPC-N cells (Fig. 4). Importantly, colon cancer MC38 cells, when inoculated into these cured mice that rejected replanted KPC-N cells, engrafted and expanded in all mice. These results clearly indicated that K3-SPG-ISV monotherapy induced immunological memory specific to PDAC cells. Although K3-SPG-ISV monotherapy failed to cure colon-26-bearing mice, the combination therapy of K3-SPG-ISV and PD-1 blockade cured half of the treated mice. All the mice cured by this combination rejected rechallenge with colon-26, while the same cells engrafted in the age-matched tumor-naïve control, indicating the establishment of immunological memory in these cured mice ( Supplementary Fig. 4).
K3-SPG-ISV induces a systemic antitumor effect that is dependent on CD8 T cells. In addition to a durable memory response, another appealing point of immunotherapy is the potential for a systemic The arrows indicate the timing of therapy. Error bars represent the mean ± SEM. Statistically significant differences were measured by one-way ANOVA followed by Dunnett's post hoc tests (panel A, B, C and D) and the Tukey-Kramer test (panel E and F). *p < 0.05; **p < 0.01; ***p < 0.001. Survival curves were analyzed using log-rank tests. iv, intravenous.   Cancer vaccines aim to induce antitumor CD8 T cells that produce IFN-γ, a central player in cancer immunity. Immunohistochemistry analysis showed that K3-SPG-ISV increased CD8 T cell infiltration in both vaccinated and untreated tumors (Fig. 5B). Consistently, at the transcriptional level, upregulation of intratumoral Cd8 and Ifng expression was observed in the tumor on both sides, although their induction was stronger in the vaccinated than in the untreated side (Fig. 5C). Rantes, an IFN-γ-inducible chemokine, and TNFA were also upregulated, although Rantes was marginally expressed in the untreated side (Fig. 5C). Enzyme-linked immunospot (ELIS-POT) analysis showed that CD8 T cells purified from splenocytes of K3-SPG-ISV-treated KPC-N-bearing mice produced IFN-γ when cocultured with KPC-N as target cells but not with syngeneic control MC38 cells, indicating the induction of KPC-N-specific CTLs in the K3-SPG-ISV-treated mice (Fig. 5D). Under the condition that > 97% depletion of CD8 T cells was achieved in the splenocytes (Fig. 5E, right panel), CD8 T cell depletion canceled the antitumor activity of K3-SPG-ISV on both the vaccinated and untreated sides, indicating that ISV systemically suppressed tumor growth in a CD8 T cell dependent manner (Fig. 5E, left and center panels). CD8 T cell dependency of K3-SPG-ISV was also observed in the colon-26 model (Supplementary Fig. 5B).
Combined ISV strategy incorporating agonistic CD40 antibody in K3-SPG enhances the antitumor effect. The activation of CD40 on DCs through the interaction with CD40 ligand (CD40L) plays an important role in the induction of antitumor T cell responses. Thus, we reasoned that intratumoral injection of agonistic CD40 antibody (anti-CD40-ISV) might partner well with K3-SPG-ISV to achieve more potent ISV. Indeed, in the bilateral PDAC model, anti-CD40-ISV as well as K3-SPG-ISV suppressed tumor growth on both the treated and untreated sides, demonstrating the abscopal effect of each respective monotherapy. The combination of K3-SPG-ISV/anti-CD40-ISV resulted in enhanced tumor suppression and prolonged survival longer than each monotherapy (Fig. 6A). Flow cytometric analysis of splenic CD8 T cells revealed that either monotherapy with K3-SPG-ISV or anti-CD40-ISV shifted naïve (CD44-CD62L+) or memory (CD44+ CD62L+) phenotype at the baseline to effector (CD44+ CD62L-) phenotype and that the combination of K3-SPG-ISV/ anti-CD40-ISV further increased the frequency of effector cells, suggesting that this ISV combination enhanced T cell priming (Fig. 6B).
Finally, the systemic antitumor effect of the K3-SPG-ISV/anti-CD40-ISV combination was evaluated using a liver metastasis model of CRC that is closer to human clinical practice. Firefly luciferase-tagged colon-26 cells were injected subcutaneously into the spleen so that tumor cells would be engrafted in the liver, followed by a combination of K3-SPG-ISV/anti-CD40-ISV. Monitoring of bioluminescence imaging of luciferase-expressing cells clearly showed that this ISV combination suppressed both the vaccinated tumor and metastatic liver lesions, further confirming the induction of systemic response by ISV (Fig. 6C).

Discussion
As we have reported previously, K3-SPG, a second-generation TLR9 ligand, has superior properties to other previously known CpG ODNs in that it is a clinically translatable water-soluble nanoparticulate with the ability to induce robust type I IFN 9 . Accordingly, K3-SPG-iv has been demonstrated to act as a potential vaccine adjuvant for cancer and viral infection in mouse and NHP models 9,17-21 . In this study, focusing on the clinical feasibility and advantages of ISV as mentioned in the introduction section, we evaluated the mono-or combined therapies of K3-SPG-ISV using mouse models of PDAC and CRC, two representative cancer types in the fields of pancreaticobiliary and gastrointestinal oncology. Our results can be summarized as follows: K3-SPG-ISV (1) suppresses tumor growth and prolongs survival, (2) is more potent than K3-ISV or K3-SPG-iv, (3) synergizes with either systemic administration of CPIs (anti-PD-1 and anti-CTLA-4 antibodies) or intratumoral administration of agonistic-CD40 antibody, another innate immune stimulator, (4) induces, even when used as monotherapy, systemic and long-lasting memory responses, (5) induces an interferogenic immunostimulatory TME, and (6) expands effector CD8 T cells in the spleen, increases the number of tumor-infiltrating lymphocytes and suppresses tumor growth in a CD8 T cell dependent manner. The antitumor activity of K3-SPG-iv was shown to be dependent on both type I IFN, IL-12, and Batf3-positive cross-presenting DCs 21 . Therefore, although formal proof needs to be provided, we consider the same mechanisms to operate in K3-SPG-ISV. Regarding immune effectors induced by K3-SPG-ISV, the data presented herein only provide insight into the essential role of CD8 T cells. Future experiments are warranted to determine the role of other cell types including CD4 T cells.
CPIs targeting PD-1 and CTLA-4, the current standards of cancer immunotherapy, have the limitation of a low response rate 2,3 . Accordingly, most efforts in this field are directed toward developing CPI-based combinations with other classes of immunotherapy that have different mechanisms of action from CPIs 31 . The desirable effects of ISV in that the tumor itself can be turned into a vaccine and that the immunosuppressive shield is broken in the TME might increase the likelihood of response to CPIs. Indeed, there are several preclinical and clinical studies of ISV utilizing TLR9 agonists, IMO-2125 (B type), SD-101 (C type), MGN1703 (C type) and CMP-001 (D type), in combination with pembrolizumab, atezolizumab, or ipilimumab for melanoma, lung cancer, pancreatic cancer and colorectal cancer 13,[32][33][34][35][36][37][38][39] . Considering the advantages of K3-SPG over C-, D-, and B-types of TLR9 agonists 9 , the synergism of K3-SPG-ISV with CPIs encourages the use of K3-SPG as an adjuvant to CPIs to improve their response rate. On the other hand, another critical limitation of CPIs is the problem of serious systemic adverse effects 40,41 . A challenging but promising strategy to overcome this shortcoming is to develop an immunotherapy that is sufficiently effective even without CPIs. The remarkable finding of this study  42,43 . Together with the observation that additional local innate immune stimulation by anti-CD40-ISV further enhanced the antitumor activity of K3-SPG-ISV (Fig. 6), the CPI-free ISV strategy that utilizes either K3-SPG alone or in combination with anti-CD40-ISV might be a promising option for achieving effective antitumor activity without serious systemic adverse effects caused by CPIs. In any scenario, the K3-SPG-based ISV could satisfy unmet needs in the current field of cancer immunotherapy.
In contrast to preclinical studies where the protocol of tumor puncture can be optimized to provide proof of concept, multiple repeated tumor punctures are not practical in clinical settings. Therefore, it is important to maximize the efficacy of ISV in a single puncture to minimize the number of punctures. The activation of CD40 on DCs through interaction with its cognate ligand CD40L licenses DCs to boost CD8 T cell priming 30 . Our observation that the addition of anti-CD40-ISV-enhanced K3-SPG-ISV encourages the development of ISV incorporating other classes of innate immune stimulators, such as CD40 agonists, to maximize efficacy. Besides targeting CD40, intratumoral injection of fms-related tyrosine kinase 3 ligand or granulocyte-macrophage colony-stimulating factor might also be promising options, because they recruit and activate DCs at the tumor site [44][45][46] . Cancer immunity is activated by ICD that releases tumor antigens from dead tumor cells and local innate immune activation, the latter of which is mainly aimed at by ISV. Thus, an alternative maximization strategy for ISV should be incorporation of local therapeutic interventions that accelerate tumor antigen release. ISV could be combined with radiotherapy 47 , photodynamic therapy 48 , thermal ablation 49 , irreversible electroporation 37,50 , cryoablation 51 or the recently developed near-infrared photoimmunotherapy. As our previous results showed that ICD is induced by K3-SPG itself 21 , these additional interventions would synergize with K3-SPG. Our group is currently conducting studies to identify the best partner of K3-SPG-based ISV.
The high incidence of recurrence after surgical resection is problematic for patients with PDAC and CRC 52,53 . The perioperative immunotherapy would prevent tumor recurrence by overcoming postoperative immunosuppression and enhancing antitumor immunity 54 . While our results encourage the clinical application of K3-SPGbased ISV for therapeutic purposes, the vaccine strategy presented here should also operate in the context of neoadjuvant immunotherapy that utilizes the preoperative tumor as a vaccine, thereby inducing systemic and memory responses for surveying and destroying the micrometastatic lesions causing postoperative recurrence. Although a clinical trial of intratumoral injection of TLR9 ligand (CMP-001) along with nivolumab for resectable melanoma has been conducted (NCT03618641), to the best of our knowledge, published data are only available for neoadjuvant immunotherapy using only CPIs 55 . Future studies are warranted to investigate the usefulness of K3-SPG-based ISV as neoadjuvant immunotherapy.
An obvious limitation of ISV is the technical hurdle of intratumoral injection. This study is from the standpoint of emphasizing ISV, which argues against safety concerns regarding systemic therapy. Nonetheless, the therapeutic value of K3-SPG-iv is not discounted, particularly for tumors that are not accessible, because its adjuvant effect has been proven 21 , and an undesirable systemic inflammatory response has been shown to be weaker in K3-SPG-iv than in TLR3 ligand poly(I:C)-iv in NHP experiments 19 . Careful pretreatment evaluation of the accessibility of tumors and selection of patients for ISV are crucial. Our observation that K3-SPG-ISV remains effective in PDAC model which is resistant to K3-SPG-iv presumably due to insufficient intratumoral accumulation of K3-SPG (Fig. 2F) might encourage the indication of K3-SPG-ISV for refractory but accessible cancer types.
In summary, even K3-SPG-ISV monotherapy has remarkable potential for inducing both systemic and memory immune responses. The present results also demonstrated that K3-SPG-ISV synergizes with systemic administration of CPIs or local administration of agonistic CD40 antibodies. Taken together, K3-SPG-ISV can  www.nature.com/scientificreports/ be combined with CPIs to improve their response rate or, conversely, applied as CPI-free local immunotherapy with or without an additional innate immune stimulator to avoid CPI-related adverse events. Our results provide a strong rationale for clinical translation of K3-SPG-based ISV to gastrointestinal and hepatopancreatobiliary malignancies for which endoscopy-, ultrasound-, or EUS-guided puncture is a routine clinical technique.

Methods
This study was carried out in accordance with relevant guidelines and regulations.
Mice. C57BL/6 and BALB/c wild-type mice were purchased from Charles River Laboratories (Yokohama, Japan). All the mice were maintained under specific pathogen-free conditions. No specific sex selection was used in this study. The protocols of all mouse experiments were approved by the Institutional Animal Care and Use Committee and the Ethics Committee of Kyoto University Graduate School of Medicine (Med Kyo 20315). All animal experiments were carried out in accordance with ARRIVE guidelines.
In vivo mouse studies.   Immunohistochemistry. Extracted tumor tissues were fixed with 10% neutral phosphate-buffered formalin and embedded in paraffin. For immunohistochemical staining, antigen retrieval was performed by incubating the sections in citric acid buffer (pH 6.0) for 15 min at 98 °C. Then, endogenous peroxidase was quenched with 0.3% hydrogen peroxide in methanol at room temperature for 30 min. Blocking was performed by incubating the sections with a blocking solution (Dako). After blocking, the sections were incubated at 4 °C overnight with the following primary diluted antibodies: anti-CYP2E1 (dilution, 1:500; Abcam). Primary antibody incubation was performed at 4 °C overnight in a humidified chamber. The primary antibody used was a rabbit anti-CD8α antibody (dilution 1:100; Cell Signaling Technology, Cat.# 98941). Subsequently, the sections were incubated with peroxidase-labeled polymer conjugated secondary antibody (Dako) for 60 min at room temperature. Immunoreactivity was detected with diaminobenzidine substrate kit (Dako), and the sections were counterstained with hematoxylin.

RNA isolation and quantitative real-time PCR analysis. Total RNA was isolated from human
PBMCs, mouse splenocytes, and explanted tumors using RNeasy kit (QIAGEN) according to the manufacturer's instructions. Complementary DNA was synthesized from 500 ng of input RNA using the ReverTra Ace qPCR RT Master Mix (Toyobo) and subjected to quantitative real-time PCR (qPCR) with a SYBR Green-based gene expression assay using a LightCycler 480 System (Roche) as described previously 45 . The expression levels were standardized by comparing the levels of mouse 18S rRNA or human GAPDH as reference genes. All reactions were performed in triplicate. Primer sequences are listed in the Supplementary Table.