Seven to ten-week-old female C57BL/6 wild-type mice (Envigo) were used and all experiments were performed in accordance with Swiss law and with approval of the Cantonal Veterinary Office of Canton de Vaud, Switzerland (license VD-1046). The MB49 cell-line (kindly provided by Prof. A. Loskog, Uppsala University, Sweden) is derived from a carcinogen induced urothelial carcinoma in male C57BL/6 mice (32). Luciferase-expressing (MB49-luc) cells were generated by transfection with lentiviral vector encoding for firefly luciferase (kindly provided by Prof. D. Trono, EPFL, Lausanne, Switzerland). Female MyD88 KO (33) and TLR4 KO (34) mice were housed under specific pathogen-free conditions in the animal facility of Epalinges, Switzerland (license VD-H04).

Deeply anesthetized mice received at the indicated time point a single intravesical instillation of PBS (50µl) or bacterial solutions (3x10⁸ colony forming unit (CFU)) (11), using an Introcan 24G/3/4 (Braun, Melsungen, Germany) catheter inserted after lubrication with K-Y ^® gel. S. enterica serovar Typhi Ty21a lyophilized bacteria (Ty21a^LYO) in the format contained within the enteric-coated capsule used for oral immunization (Vivotif^® (Bavarian Nordic Berna, Thörishaus, Switzerland) including >2×10⁹ CFU Ty21a, sucrose (7.9–44 mg), lactose (max 176.4mg), ascorbic acid (E300), acid casein hydrolysate, magnesium stearate (E470)) were used within 1 h after reconstitution and dilution in PBS. For Ty21a^FR, Ty21a bacteria were grown in LB (Luria-Bertani) media at 37 °C to OD₆₀₀ = 0.6, concentrated by centrifugation and resuspended in PBS at the desired concentration for fresh administration (10, 35). Bacterial concentrations of Ty21a^FR were confirmed by serial dilution and plating on LB agar plates, followed by colony enumeration after overnight incubation at 37°C.

Mice were sacrificed by CO₂ inhalation. Bladder was intravesically instilled with an optimal cutting temperature (OCT) compound before mechanical dissociation and freezing into liquid nitrogen. Ten µm-thick transversal sections were stored at -80°C. For immunological staining, the bladder sections were quickly fixed 5 minutes in ice-acetone and dried for 5 minutes. The sections were then washed with PBS three times and each individual section was circled using a Dako-Pen. The cryosections of the bladder were incubated in blocking buffer (PBS, 5% BSA, 2.5% FCS) for 45min at room temperature (RT) in a humidity chamber and then incubated overnight at 4°C with one of the primary antibodies against Pnad (Meca79, Biolegend), CD31 (MEC 133, BD Pharmigen) or CD8 (10-0081-82, Bioscience), which were diluted in 1% BSA PBS solution. After 3-times washing with PBS under slight agitation, the bladder sections were incubated for 45 minutes at RT in a humidity chamber with corresponding Alexa-Fluor-488-conjugated secondary antibodies: anti-rat IgG (A11006) or anti-rat IgG (A21208, all from Life technologies). The sections were further washed three times with PBS and covered with PBS containing 50% of glycerol and DAPI diluted to 1:10 (Duolink; 82040 0005). After mounting, the slides were observed by fluorescence microscopy (Leica DMI3000B, DF345FX) with the respective software (Leica Application Suite). The whole tissue area of each bladder section was examined (8–10 microscopic fields) and the total count of CD31+ vessels or CD8 T cell numbers were shown in the graphs (i.e. numbers/bladder section). Pictures were analyzed by ImageJ.

Bladders were frozen and powdered in liquid nitrogen. The weighed bladder powder was homogenized in 100µl of 0.1% Brij L23 solution and incubated 2 minutes at RT. After centrifugation (5 minutes, 15’000 g, RT), 50µl of the supernatant was transferred to 450 µl of Drabkin’s reagent solution (Sigma-Aldrich, D5941-6VL). The absorbance was measured by spectrometry at 540 nm. The concentration of hemoglobin was then determined from the standard curve of cyan-methemoglobin.

In vivo permeability assay. Mice were injected with 200 µl of 0.5% of Evans Blue solution (Sigma, E2129) in lateral tail vein (36). After 30 minutes, the mice were sacrificed through cervical dislocation. Different tissues such as the bladder, the lung and the liver were collected, weighed and incubated in 500 µl of formamide (Sigma, F9037) during 24h in water bath at 55 °C. After centrifugation (5 minutes, 15’000 g, RT), the absorbance was measured at 610 nm by spectrophotometry and Evans Blue content calculated with a standard curve.

Mice were sacrificed by CO₂ inhalation to collect the bladders. Single-cell suspensions were obtained by mincing in DL-dithiothreitol (Sigma, D9779) and by subsequent digestion with 1 mg/ml collagenase/dispase (Roche, 11097113001) and 0.1 mg/ml DNAse I (Sigma-Aldrich, D5025) with 20% fetal calf serum (Gibco, 10270). The recovered cells were stained and analyzed by flow cytometry. The following monoclonal anti-mouse antibodies were used: anti-CD3-PerCP/Cy5.5 (17A2) (Biolegend, 100218), anti-CD4 (RM4-5) FITC or eF450 (eBioscience, 11-0042–81 or 48-0041-82), anti-CD8 (53-6.7) APC (eBioscience, 17-0081-82) or PE-TXRD (Southern Biotech, 1550-10), anti-CD11b (M1/70) FITC (eBioscience, 101206) or APC (Biolegend, 101212), anti-CD88(C5aR)-PE-Cy7 (20/70) (Biolegend, 135809), anti-CXCR2-PE (242216) (R&D, FAB2164P). Uty-specific cells were stained using the H-2Db restricted dextramer Uty_246-254-PE (Immudex). Dead cells were excluded by a live/dead fixable kit: aqua dead cell stain kit (L34957, Invitrogen, Thermo Fisher Scientific). Cell acquisition and analysis were performed using Gallios Flow Cytometer (Beckman Coulter, Nyon, Switzerland) and FlowJo software (Tree Star, Ashland, OR), respectively.

The minor histocompatibility male antigen HY (Uty) is expressed by the MB49 bladder tumor cell line. The H-2Db-restricted epitope Uty_246–254 peptide (37) (WMHHNMMDLI) was synthesized by the Peptide and Tetramer Core Facility of the department of oncology (Lausanne University Hospital and University of Lausanne, Switzerland). Mice were immunized s.c. with 50 μg of Uty_256–254 adjuvanted with 0.4 μg heat labile enterotoxin (kindly provided by Berna-Biotech, Bern, Switzerland) and 10μg of CpG (#1826, 5’-TCCATGACGTTCCTGACGTT-3’, Coley Pharmaceutical Group). IFN-γ ELISPOT assays were performed as previously described (38) using Multi-screen-HA 96-well plates (MAHAS4510, Millipore) anti–IFN- γ mono-clonal antibody (R4-6A2, Beckton Dickinson PharMingen), biotinylated anti–IFN- γ monoclonal antibody (XMG1.2, Beckton Dickinson PharMingen), and Streptavidin-AP (Roche). In brief, 3x10⁴ bone-marrow-derived DCs (BMDCs)/well, used as antigen-presenting cells, were incubated for 1h in duplicate with 1mg/mL of Uty_256–254 peptide or medium alone (control wells) before addition of 10⁵ bladder cells and incubation for 16–24h. Uty-specific responses were defined as the number of IFN- γ spots/10⁵ cells in the Uty-stimulated wells minus the number of IFN- γ spots/10⁵ cells in the control wells (<3 spots/well). BMDCs were generated from bone marrow cells in the presence of 150 U ml⁻¹ of recombinant mouse granulocyte–macrophage colony-stimulating factor (R&D Systems, Abingdon, UK) as previously described (39).

Bladder tumors were established in deeply anesthetized mice that were uretherally catheterized using Introcan 24Gx3/4 catheters (Braun, Melsungen, Germany) as previously described (12). A 15 minutes pre-treatment with 100 μl 22% ethanol was performed before instillation of 500’000 MB49-luc cells in 50μl of Hank’s balanced salt solution (HBSS) (Gibco, 14025092). MB49-luc tumor growth was monitored by bioluminescence 15 minutes after intraperitoneal (i.p.) injection of D-luciferin (Promega, L8220, 150 μg/g of body weight) in the Xenogen imaging system (Xenogen/IVIS Caliper Life Science, kindly provided by cellular imaging facility, CIF/UNIL, Lausanne, Switzerland). All mice will develop bladder tumors and monitoring of MB49-luc tumors establishment and growth can be efficiently assessed during the first 3 weeks. Uncontrolled loss of luminescence of the growing tumors can then often appear (40), requiring additional monitoring by palpation, hematuria and overall health status of the mice, that were euthanized when they reached human endpoints.

Bladders were recovered 24h after intravesical immunostimulation with 3x10⁸ CFU Ty21a^FR or Ty21a^LYO and homogenized in 500-1000 μL PBS with protease inhibitors (10 μg/mL Aprotinin from bovine lung, 10 μg/mL Leupeptin hemisulfate salt, and 10 μg/mL Pepstatin A; all from Sigma-Aldrich). TritonX-100 (final concentration 1%, Sigma-Aldrich) was added after homogenization, and after two freeze-thaw cycle’s samples were centrifuged at 10’000 g for 5 minutes to remove debris. Protein concentration was assessed using BCA protein assay (Thermo scientific). Chemokines were detected using the Proteome profiler array: mouse chemokine array kit (R&D Systems), according to the manufacturer instructions. Briefly, 150 μg of protein (pooled from three bladders, 50 μg each) of each condition were used for the assay. Detection of chemokine levels was performed using ImageJ software (NIH) and expressed as mean pixel density. Increased chemokine levels between Ty21a^FR and Ty21a^LYO were considered significant when ≥ to the 99% confidence interval of the mean fold-increases (i.e. ≥ 1.5 fold).

Statistical analyses were performed using GraphPadPrism 10 for Windows (GraphPad software). Multiple comparisons were performed using one-way ANOVA and Tukey’s or Sidak post-test or log-rank test as indicated in the figure legends.

Female C57 BL/6 mice that were intravesically instilled with PBS or 3x10⁸ CFU of Ty21a^FR were sacrificed 72h later and their bladder analyzed by immunohistochemistry, as well as for hemoglobulin content. As expected from our previous experiments showing increased immune cell infiltration upon Ty21a^LYO (12), and despite high variability, a significant (ca. 10-fold) increase of the number of CD8⁺ T cells upon Ty21a^FR was observed (n= 3 mice) (Figures 1A, B). However, the number of blood vessels (CD31⁺, Figure 1C, D) in the same bladder, as well as the total blood content (as measured by the hemoglobin content, Figure 1E) were not increased by Ty21a^FR. Moreover, no HEV (Pnad⁺) were observed in the bladder with or without Ty21a^FR treatment (Figure 1F, right panel), while they were detected in LNs (Figure 1F, left panel). These results suggest that the rapid increase of CD8⁺ T cells upon Ty21a instillation is not associated with neoangiogenesis.

We then examined whether vessel permeability was altered using an Evans Blue assay. The data showed that vessel permeability was significantly increased in the bladder one day after Ty21a^FR instillation, but not later (Figure 1G). Similarly, elevated vessel permeability was observed with Ty21a^LYO treatment while heat-killed bacteria had no effect (Figure 1H). No increased permeability was observed at distant organs, such as the lung or the liver (Supplementary Figures 1A, B) and thus the increased vascular permeability was limited to the instillation site. Vessel permeability in the bladder was similarly increased by Ty21a^FR and Ty21a^LYO which may in part explain the increased T-cell infiltration observed in independent experiments with either of the two formulations (11, 12, 19). To address this hypothesis, we therefore directly compared the ability of Ty21a^FR and Ty21a^LYO to increase immune cell infiltration in the bladder.

Immune cell infiltration in the bladder of mice was examined 24h and 72h after intravesical instillation of 3x10⁸ CFU of Ty21a^FR or Ty21a^LYO, or in untreated naïve mice. The data showed that CD4⁺ (Figure 2A) and CD8⁺ T cells (Figure 2B), as well as myeloid cells (CD11b^high, Figure 2C), were all significantly increased (ca. 2-fold) by Ty21a^FR as compared to Ty21a^LYO, at both time points, despite considerable interindividual variability at 72 hours. Overall, Ty21a^FR instillation induced a significantly augmented (5-10-fold) immune cell infiltration compared with naïve controls, consistent across all cell types and time points. More importantly, in the presence of bladder tumor, Ty21a^FR was significantly more efficient than Ty21a^LYO at improving short-term survival of mice bearing large orthotopic bladder tumors, while at long-term only a trend was observed (Figures 2D, E). Whether higher immune cell infiltration induced by Ty21a^FR at 72 hours correlates with the improved short-term survival of mice warrants further investigation. In naïve mice, a combinatory treatment setting involving subcutaneous vaccination with the Uty antigen followed by intravesical instillation of either formulation, Ty21a^FR was superior to Ty21a^LYO in enhancing the numbers of total CD4⁺ and CD8⁺ T cells, as well as vaccine-specific CD8⁺ T cells (DexUty⁺) and IFN-γ–secreting CD8⁺ T cells in the bladder (Supplementary Figure 2). Altogether, the data show that despite similar increased vessel permeability, the two formulations differed in the magnitude by which they increase immune cell infiltration. We thus further examined whether differences in chemokine secretion in the bladder could underlie this discrepancy.

The levels of 25 chemokines were analyzed in the bladders of mice 24h after intravesical instillation of 3x10⁸ CFU of Ty21a^FR or Ty21a^LYO using a chemokine array. Relative changes of these chemokines between the two formulations (Figure 3A) showed that C5a, CXCL5 and CXCL2 were significantly increased (> 1.5-fold) in bladders treated with Ty21a^FR compared to those treated with Ty21a^LYO. Their cognate receptors (C5aR for C5a and CXCR2 for both CXCL5 and CXCL2), however, were either undetectable or lowly expressed on T cells from naïve mice and were not modulated by either Ty21a^FR or Ty21a^LYO (Figures 3B, C), precluding their direct effect on T-cell infiltration. In contrast, C5aR⁺CD11b^high myeloid cells were significantly more numerous 24h after instillation in bladders treated with Ty21a^FR as compared to Ty21a^LYO (Figures 3D, E), while CXCR2⁺CD11b^high cells were less impacted (Figure 3F). In addition, Ty21a^FR significantly increased C5aR^-CXCR2^-CD11b^high myeloid cells (Figure 3G) as compared to Ty21a^LYO, suggesting that other analytes not present in our array may be involved and, that different types of myeloid cells may indirectly participate in T-cell attraction.

In the early oral typhoid vaccine field trials, freshly harvested Ty21a was more effective than enteric capsules containing lyophilized Ty21a, which may be related to the expression of a more active form of LPS in the fresh Ty21a (30, 31, 41–43). To examine whether TLR4-dependent LPS sensing was involved in the infiltration of immune cells in the bladder, we compared the effects of intravesical Ty21a^FR in wild-type (WT), TLR4-KO and MyD88-KO mice. MyD88-KO mice were included to evaluate the role of MyD88, an essential adaptor for signaling through all TLRs (except TLR3) and IL-1 receptors. CD8⁺ T cells were significantly decreased upon Ty21a^FR in both TRL4- and MyD88-KO mice compared to WT (Figure 4A) while the levels of CD4⁺ T cells upon Ty21a^FR were also decreased in TLR4- and MyD88-KO mice, though not significantly (Figure 4B). In contrast, the levels of CD11b^high myeloid cells were significantly reduced only in MyD88-KO mice. These data suggest that T-cell infiltration induced by Ty21a is at least partially mediated by LPS, while myeloid cell infiltration occurs by TLR4-independent but MyD88-dependent mechanisms.