An oral cancer vaccine using Bifidobacterium vector augments combination of anti-PD-1 and anti-CTLA-4 antibodies in mouse renal cell carcinoma model

Recently, immune checkpoint inhibitor (ICI) based combination therapies, including anti-PD-1 antibody, nivolumab with anti-CTLA-4 antibody, and ipilimumab have become the primary treatment option for metastatic or unresectable renal cell carcinoma (RCC). However, despite the combination of two ICIs, 60–70% of patients are still resistant to first-line cancer immunotherapy. In the present study, undertook combination immunotherapy for RCC using an oral cancer vaccine (Bifidobacterium longum displaying WT1 tumor associated antigen (B. longum 420)) with anti-PD-1 and anti-CTLA-4 antibodies in a mouse syngeneic model of RCC to explore possible synergistic effects. We found that B. longum 420 significantly improved the survival of mice bearing RCC tumors treated by anti-PD-1 and anti-CTLA-4 antibodies compared to the mice treated by the antibodies alone. This result suggests that B. longum 420 oral cancer vaccine as an adjunct to ICIs could provide a novel treatment option for RCC patients. Our microbiome analysis revealed that the proportion of Lactobacilli was significantly increased by B. longum 420. Although the detailed mechanism of action is unknown, it is possible that microbiome alteration by B. longum 420 enhances the efficacy of the ICIs.


B. longum 420 and anti-PD-1 antibody suppressed Renca tumor growth. Tumor volume. The
B. longum 2012 strain with a plasmid carrying galacto-N-biose/lacto-N-biose I binding protein (GL-BP) gene only, constructed in our previous study, was used as a control 6 . GL-BP is a membrane protein in the ATP-binding cassette transporter on the wild-type B. longum cell wall, which we used as an anchor to display antigen on the bacterial cell surface. Oral administration of B. longum 420 combined with intraperitoneal injection of mouse anti-PD-1 antibody significantly suppressed the growth of Renca tumors compared to the PBS group at day 21 (p < 0.05) (Fig. 1a,b). In the other groups (n = 5), all mice died from tumor growth, with an average survival of 25.6 ± 5.8 (standard error) days with PBS control, 36.6 ± 7.6 days with B. longum 2012, 36.6 ± 5.7 with B. longum 420, and 43.4 ± 5.3 days with anti-PD-1 antibody alone. The survival curve for the combination of B. longum 420 with anti-PD-1 anti-body treatment showed a significant prolongation of survival of 71.2±29.9 days compared with the other treatment groups (Fig. 1c, p < 0.05).
TILs. We next collected tumor tissues from another set of mouse treatment groups after treatment to investigate the tumor infiltrating lymphocytes (TILs). Immunohistochemical staining for CD4 + and CD8 + T cells showed remarkably increased numbers of CD4 + and CD8 + T cells infiltrating into the tumor tissues in mice treated with the combination of B. longum 420 and anti-PD-1 antibody compared to the other treatment groups (Fig. 2a). The flow cytometric analysis of T cells isolated from tumor tissues revealed that the combination of B. longum 420 and anti-PD-1 antibody treatment increased tumor-infiltrating CD8 + , and CD107a + CD8 + T cells (p = 0.734, and 0.257, respectively), while anti-PD-1 antibody alone and combination with B. longum 420 increased CD4 + T cells compared to the other treatment groups (p = 0.55, Fig. 2b), even though there was not significant difference.

B. longum 420 enhanced the anti-tumor activity of combined anti-PD-1 and anti-CTLA-4 antibodies. Tumor volume.
To confirm the synergistic anti-tumor activity of adding the anti-CTLA-4 antibody to the combination of anti-PD-1 antibody and B. longum 420, we next performed further animal studies using the Renca-bearing BALB/c mouse RCC syngeneic tumor model. The combined anti-PD-1 and anti-CTLA-4 antibodies group and the combined oral administration of B. longum 420 with anti-PD-1 and anti-CTLA-4 antibodies group both showed significantly suppressed tumor growth compared to the PBS group at day 21 (p < 0.05) (Fig. 3b). The survival curve (Fig. 3c) for the combination of B. longum 420 with anti-PD-1 and anti-CTLA-4 antibodies showed a significant prolongation of survival of 78.7 ± 14.5 days compared with the other treatment groups (p < 0.05). In the other groups, an average survival of 29.5 ± 2.8 days with PBS control, 58.0 ± 9.9 days with anti-PD-1 and anti-CTLA-4 antibody alone.
TIL. Immunohistochemical staining for CD4 + and CD8 + T cells showed remarkably increased numbers of CD4 + and CD8 + T cells infiltrating into the tumor tissues in mice treated with the combination of B. longum 420 and anti-PD-1 and anti-CTLA-4 antibodies compared to the other treatment groups (Fig. 4a). Interestingly, the anti-PD-1 and anti-CTLA-4 antibodies alone showed the greatest increase in Foxp3+ T cells compared to the other groups (Fig. 4a). The flow cytometric analysis of T cells isolated from the tumor tissues revealed that the combination of B. longum 420, anti-PD-1 and anti-CTLA-4 antibodies showed the highest numbers of tumorinfiltrating CD8 + and CD107a + CD8 + T cells (p = 0.271, and 0.304, respectively), while the anti-PD-1 and anti-CTLA-4 antibodies alone group showed the highest number of CD4 + T cells among all the treatment groups (p = 0.285, Fig. 4b). Representative dot plots of each treatment group with gating for staining were shown in Supplementary Fig. S1.
Differences in bacterial community profiles before and after administration in each group.

Intracellular cytokine staining for splenocytes.
To determine the frequency of WT1-specific T cells induced by the oral vaccination of B. longum 420, we performed intracellular cytokine staining for splenocytes stimulated with mitomycin-C-treated Renca. The frequency of CD4 + T cells that were WT1-specifically producing interferon (IFN)-γ in the B. longum 420 treated mice group was significantly higher than in PBS group (Fig. 6, p < 0.05). Also, the frequency of TNF-α producing CD4 + T and CD8 + T cells in the B. longum 420 treated mice group was significantly higher than the other groups ( Fig. 6, p < 0.05, p < 0.01).

Discussion
The present study resulted in two critical observations. First, B. longum 420 significantly augmented the antitumor activity in mice loaded with Renca, mouse RCC tumors treated with an anti-PD-1 antibody, and mice treated with anti-PD-1 and anti-CTLA-4 antibodies. Second, the administration of B. longum 420 substantially changed the microbiome, and particularly increased the proportion of the Lactobacillaceae family. In our results, the group treated with a combination of anti-PD-1 antibody and B. longum 420 had significantly lower average tumor volumes compared to the PBS group ( Fig. 1b and significantly better survival compared to the other groups (Fig. 1c). Consistent with these results, the anti-PD-1 and anti-CTLA4 antibodies alone showed a significant tumor growth inhibitory effect compared to the PBS group, and the addition of B. longum 420 to these antibodies significantly enhanced the antitumor activity (Fig. 3b). Combination therapy with anti-PD-1 and anti-CTLA4 antibodies and B. longum 420 demonstrated a significant improvement in survival compared to www.nature.com/scientificreports/ the PBS group and the anti-PD-1 and anti-CTLA4 antibodies alone groups (Fig. 3c). These results suggested that even though B. longum 420 alone did not show significant antitumor activity, B. longum 420 could significantly augment the antitumor activity of ICI treatments in the mouse RCC model. In the clinical setting, the combination of the PD-1 inhibitor nivolumab with the CTLA-4 inhibitor ipilimumab is currently the only dual ICI combination approved in the first-line setting for patients with mRCC 17 . The clinical data show a reduction in risk of death by 35%, impressive long-term PFS (Progression Free Survival) plateauing at approximately 35% after 30 months, and superior quality of life versus sunitinib, making this combination attractive. Although it is highly likely that nivolumab plus ipilimumab has the highest CRR (Clinical Response Rate) 17 , most patients still either have primary resistance to these therapies or acquire resistance after an initial response 2 . Therefore, developing novel therapeutic strategies designed to overcome these resistance mechanisms is of paramount importance for patients with this disease. In this regard, the results of the present study show potential for improving ICI-ICI treatment outcomes further.
Our immunohistochemical study showed that administration of B. longum 420 increased the infiltration of both CD4 + T cells and CD8 + T cells into the tumor tissues (Figs. 2a, 4a). Although there was no statistically significant difference, flow cytometric analysis also showed that the combination of B. longum 420 and ICIs (anti-PD-1 anti-body w/wo anti-CTLA-4 antibody) increased the infiltration of CD8 + T cells and CD107a + CD8 + T cells, which are considered activated CD8 + T cells, into the tumor tissues (Figs. 2b, 4b). It is well known that CD8 + T cells play a critical role in the antitumor immune response, and these TIL findings strongly support the adjuvant effect of B. longum 420 with ICI treatments in RCC. Interestingly, the flow cytometric analysis showed that the numbers of CD4 + T cells in the ICI-alone groups were higher than the combination of B. longum 420 and ICIs (Figs. 2b, 4b). The detailed mechanism for this finding is unknown, but there have been several clinical reports that CD4 + Foxp3 + T cells are increased by ICIs in ICI-resistant tumors 18,19 . Tumor-infiltrating Tregs may promote tumor progression by suppressing the natural anti-cancer immune responses. The percentage of Tregs among CD4 + cells is significantly higher in tumors than in the immune organs [20][21][22] . The increased percentage of Tregs in many cancers is often associated with poor prognosis 23 . Therefore, the impact of Tregs should be considered for successful ICI therapies. Our immunohistochemical study showed that the number of Foxp3 + www.nature.com/scientificreports/ T cells was increased in the anti-PD-1 and anti-CTLA-4 antibodies alone group compared to the other groups (Fig. 4a). We also proved that oral administration of B. longum 420 induced WT1-specific CD4 + T and CD8 + T cells in BALB/c mice, by performing intracellular cytokine staining for splenocytes after vaccination. Our previous study using anti-PD-1 antibody in a poorly responsive MBT-2 mouse bladder cancer model revealed that B. longum 420 alone induced higher anti-tumor activity compared to the combination of anti-PD-1 antibody and B. longum 420. In that study, the anti-PD-1 antibody remarkably increased the number of regulatory T cells [23][24][25] .
Another interesting finding in this study is the modification of the intestinal microbiome by oral administration of B. longum 420. The proportion of Lactobacilli was significantly altered after treatment with ICI and B. longum 420 (Fig. 5). Since only one mouse in each group was analyzed, we must carefully interpret the results. However, this is an important finding because there have been no studies comparing gut microbiomes before and after the administration of ICI and probiotics including B. longum. The host microbiome is established as one of the critical parameters to improve or impair the efficacy of ICI therapies [24][25][26] . Increasing evidence indicates that some commensal bacteria modulate the tumor microbiome 27 and that intestinal commensal bacteria-induced IFN γ-producing CD8 + T cells augment anti-tumor responses in ICI 28 . For example, Kawanabe-Matsuda et al. reported that microbial exopolysaccharide produced by Lactobacillus delbrueckii subsp. bulgaricus OLL1073R-1 (EPS-R1) induced CCR6 + CD8 + T cells in mice and humans 29 . In mice, ingestion of EPS-R1 augmented the antitumor effects of anti-CTLA4 or anti-PD-1 monoclonal antibody against CCL20-expressing tumors. Although it is unclear whether the mechanism by which lactobacilli increase response to ICI is microbiota change or metabolite, there have been a number of reports in recent years that lactobacillus species can enhance reactivity to ICI 30,31 。Moreover, it was recently reported that Bifidobacterium pseudolongum produced inosine-activated Th1 cells with anti-tumor effects in the presence of IFN-γ or other costimulation 31   www.nature.com/scientificreports/ that repeated oral administration of enterobacteria alters the intestinal microbiota 32 , for instance the intense efficacy of B. longum BB536 (BB536Y group) in improving the intestinal environment, reducing harmful bacteria and improving defecation frequency and stool characteristics in humans [33][34][35] . Although the exact mechanism and the specific microbiota members associated with the enhanced clinical response of ICIs remains unknown, it is possible that microbiome alteration also contributed to antitumor activities, and the benefits of B. longum 420 as a probiotic-based cancer vaccine may provide a novel treatment option for RCC patients resistant to ICIs.    www.nature.com/scientificreports/ WT1 gene was fused to GL-BP, and then corresponding gene was ligated with the B. longum. B. longum 2012 expresses only GL-BP and is used as a control agent 34 . All three recombinant Bifidobacterium bacteria were anaerobically cultured in Gifu anaerobic medium (Nissui, Tokyo, Japan) with 50 mg/mL spectinomycin at 37 °C. After cultivation, these recombinant bacteria were heated for inactivation at 65 °C for 5 min.

Methods
Cell line. Renca, a renal cell carcinoma cell line expressing murine-WT1 protein derived from BALB/c, was purchased from ATCC and maintained in in Eagle's minimum essential medium supplemented with 10% fetal bovine serum (Sigma-Aldrich Japan, Tokyo, Japan) and 1% penicillin-streptomycin (Nacalai Tesque, Kyoto, Japan). The overexpression of WT1 protein in Renca was confirmed by western blotting analysis (data not shown).
Oral vaccination. Dose of oral vaccination and treatment schedules were determined based on our previous studies 4,5,13 . Male BALB/c mice were orally given 100 μL of PBS, 1.0 × 10 9 CFU of B. longum 420, or B. longum 2012 5 times a week for 4 weeks (days 7-11, 14-18, 21-25, and 28-32) with a feeding needle. After 3 weeks of vaccination, mice were euthanized and tumors were resected to investigate the local immune response.
Animal experiments for combination therapy. To explore the in vivo anti-tumor activity of the combination of B. longum 420 and anti-PD-1 antibody against renal cell carcinoma, we employed a Renca mouse RCC syngeneic subcutaneous tumor model. One million Renca cells were subcutaneously injected into male BALB/c mice at day 0. A total of 25 mice with Renca subcutaneous tumor were randomly assigned to 5 treatment groups (n = 5): B. longum 420 + anti-PD-1, B. longum 420, B. longum 2012, PBS + anti-PD-1, and PBS at day 7, and then oral administrations were carried out as described above (Fig. 1a). Anti-mouse PD-1 antibody (RMP1-14-derived mouse monoclonal antibody against murine PD-1, InvivoGen, San Diego, CA, USA) was used for the anti-PD-1 treatment, and saline was used as an isotype control. Two hundred milligrams of anti-PD-1 antibody or isotype control was intraperitoneally injected into mice at days 12, 16, 19, 23, and 26. Tumor volume was measured by the calculation formula of (longest diameter) × (shortest diameter) 2 × 0.5. Mice were euthanized when their tumors grew larger than 20 mm in diameter, and Kaplan-Meier survival curves were generated. Next, we prepared another set of mice to examine the in vivo anti-tumor activity of the combination of B. longum 420, anti-PD-1 antibody and anti-CTLA-4 antibody against renal cell carcinoma. A total of 45 mice with Renca subcutaneous tumor were randomly assigned to 3 treatment groups (n = 15): B. longum 420 + anti-PD-1 + anti-CTLA-4, PBS + anti-PD-1 + anti-CTLA-4, and PBS at day 7, and then oral administrations were carried out as in the previous study (Fig. 3a). Anti-mouse CTLA-4 antibody (9D9-derived mouse monoclonal anti-body against murine CTLA-4, InvivoGen, San Diego, CA, USA) was used for the anti-PD-1 + anti-CTLA4 treatment. 2.5 μg of anti-PD-1 antibody and anti-CTLA-4 were intraperitoneally injected into mice at days 12, 19, and 26. Immunohistochemical study. Another set of mice was injected with 1.0 × 10 6 Renca and treated by the same method described above. Tumors were resected and divided into two pieces, and half were fixed with  www.nature.com/scientificreports/ 4% paraformaldehyde-PBS and embedded in paraffin. Another piece of tumor was used for flow cytometry as described below. Paraffin embedded tumor tissue sections were deparaffinized and rehydrated. Antigen retrieval was performed in Bond epitope retrieval buffer (pH6.0 for CD8, pH9.0 for CD4; Leica Microsystems, Wetzlar, Germany) at 98C for 20 min. Immunohistochemical staining was performed in an automatic tissue processor (Leica Microsystems Bond) according to the manufacturer's standard protocol. Briefly, tissue sections were incubated at RT for 15 min with rabbit anti-mouse CD4 antibody (1:1,000, Abcam) or rabbit anti-mouse CD8 antibody (1:400, Cell Signaling Technology Japan, Tokyo, Japan). After washing, sections were incubated with horseradish peroxidase-conjugated secondary antibodies. After washing, sections were incubated with 3,30 diaminobenzidine and counterstained with hematoxylin. Anti-mouse CD4 antibody (1:1,000, Abcam, Cambridge, UK), anti-mouse CD8 antibody (1:400, Cell Signaling Technology Japan, Tokyo, Japan), and Foxp3 antibody 236A/E7 (Mouse monoclonal anti-body; Abcam, Cambridge, UK) were used in the immunohistochemical staining. The tissue slides were observed with a BZ-X710 microscope (Keyence, Osaka, Japan). Evaluation of immunostaining was performed by two of the authors independently (Ueki and Kitagawa). Target areas were selected for each specimen using free-form drawings, and the tissue was evaluated under 20-fold magnification. For regular sections, whole areas were screened under 20-fold magnification, and three representative highpower fields (HPF) were assessed for cell counts in every specimen (400-fold magnification).
Comparison of the gut microbiota composition. To observe whether B. longum 420 administration caused changes in the intestinal microbiota, fecal samples were collected before (day 7) and after (day 22) B. longum 420 administration. From each group, n = 1 was selected for microbiome analysis (TechnoSuruga Laboratory Co, Shizuoka, Japan). The mouse with the largest tumor size in the PBS and ICI monotherapy group was selected, and the mouse with the smallest tumor size in the B. longum 420 combination group was selected. DNA extraction was conducted according to a previously described method 39 using an automated DNA isolation system (GENE PREP STAR PI-480 KURABO, Japan). The V3-V4 regions of Bacterial and Archaeal 16S rRNA were amplified using the Pro341F/Pro805R primers and dual-index method 39,40 . Barcoded amplicons were paired-end sequenced on a 2 × 301-bp cycle using the MiSeq system with MiSeq Reagent Kit version 3 (600 Cycle). The primer sequences on paired-end sequencing reads were trimmed by Cutadapt ver 1.18 with default settings 41 . Paired-end sequencing reads were merged using the fastq-join program with default settings 42 . Only joined-reads that had a quality value score of ≥ 20 for more than 99% of the sequence were extracted using the FASTX-Toolkit 43 . The chimeric sequences were deleted with usearch61 44,45 . Nonchimeric reads were submitted for 16S rDNA-based taxonomic analysis using the Ribosomal Database Project ver 2.13 (RDP) and the Techno-Suruga Lab Microbial Identification database ver 16.0 (DB-BA, TechnoSuruga Laboratory, Japan) with homology for ≥ 97% 46,47 . Intracellular cytokine staining for splenocytes. Another male BALB/c mice were orally given 100 μL of PBS, 1.0 × 10 9 CFU of B. longum 420, or B. longum 2012 5 times a week for 4 weeks (days 0-4, 7-11, 14-18, and 21-25) as described above. After the last oral vaccination, spleens were resected and single cell suspensions were collected. The splenocytes (2.0 × 10 6 ) were cultured and re-stimulated with 2.0 × 10 5 mitomycin-C-treated Renca cells in vitro. GolgiStop (BD Biosciences, San Jose, CA) was added to the medium after 26 h of the cell cultivation, and then the cells were cultured for additional 12 h. The cells were collected and processed using a BD Cytofix/CytopermTM Plus Fixation/Permeabilization Kit (BD Biosciences) for intracellular cytokine staining assay according to our previous studies 4,14 . As for the intracellular staining, PE-anti-IFN-γ or PE-anti-TNFα (BD Biosciences, respectively) were used in this study. The stained cells were counted and analyzed by using Guava flow cytometer (Luminex, Austin, TX).
Statistical analysis. One-way ANOVA followed by the Tukey-Kramer method was employed for the comparisons between multiple groups. The log-rank test on Kaplan-Meier curves was employed for the statistical analysis of survival between groups. Differences among experimental groups were considered significant when p < 0.05.

IRB.
All experiments and methods were performed in accordance with the relevant guidelines and regulations, and all experimental protocols, including animal experimental designs and procedures, were reviewed and approved by the institutional ethics and animal welfare committees of the Kobe University Graduate School of Medicine.

Data availability
The datasets generated and/or analyzed during the current study are available from the corresponding author on reasonable request.