Bladder cancer tissue microarrays (TMAs) BL244, BL244a and T124a were purchased from US Biomax Inc. TMAs BL244 and BL244a contain samples of 12 bladder tumors and their matching normal tissues. TMA T124a contains samples of 10 bladder tumors and two normal bladder tissues (all different from those in TMAs BL244 and BL244a). More details regarding the tissue samples are available on official US Biomax website (BL244: https://www.biomax.us/p_BL244_1298628261, BL244a: https://www.biomax.us/p_BL244a_221856716, T124a: https://www.biomax.us/p_T124a_407306661).

UPEC CFT073 mutants for fimbriae expression UPEC Fim ON and UPEC Fim OFF, were used in this study. In UPEC, expression of the type 1 fimbria is regulated at the transcriptional level by a promoter situated on an invertible element, which can exist in one of two different orientations. The orientation of the invertible element that allows the expression of the fimbria is defined as “on”, and the opposite orientation in which no transcription occurs, is defined as “off” (Gunther et al., 2002). UPEC Fim ON is a UPEC CFT073 mutant in which the promoter-containing invertible element is locked in the on orientation resulting with overexpression of the type 1 fimbria (Gunther et al., 2002). UPEC Fim OFF is a UPEC CFT073 mutant in which the invertible element is locked in the off orientation, preventing expression of type 1 fimbria (Gunther et al., 2002). Both strains were kind gifts from Professor H. Mobley. These UPEC strains were transformed with the GFP expressing pCM18 plasmid (Hansen et al., 2001). GFP expressing bacteria were grown in Luria Broth (LB) for 18 hr at 220 rpm and 37°C in the presence of erythromycin (300 μg/mL).

TMA tissue deparaffinization and hydration was performed as follows: xylene (Gadot-group) twice for 5 min, xylene 1 min, xylene/ethanol (50%/50%) 1 min, 100% ethanol (Gadot-group) twice 10 min, 90% ethanol 2 min, 80% ethanol 2 min, 70% ethanol 2 min and then washed in PBS for 2 min. Slides were then blocked with PBS supplemented with 10% BSA, 10% FBS and 0.5% Triton X-100 for 2 h at room temperature. Next, the slides were incubated overnight at 4°C with 50 μg/ml in PBS of FITC–conjugated Peanut Agglutinin (PNA) (Sigma-Aldrich, cat. No. L7381), anti-uroplakin Ia antibody (Abcam, cat. No. ab95207) or ConA (Sigma-Aldrich, cat. No. C7642) for the detection of Gal-GalNAc, uroplakin Ia and mannose respectively. Next day, the slides were washed three times with PBS for 10 min each, and then incubated with Hoechst 33258 (Sigma-Aldrich, cat. No. 94403) diluted 1:5,000 in PBS for 15 min at room temperature.

TMA slides went through deparaffinization and hydration as follows: xylene twice for 5 min, xylene 1 min, xylene/ethanol (50%/50%) 1 min, 100% ethanol twice 10 min, 90% ethanol 2 min, 80% ethanol 2 min, 70% ethanol 2 min and then washed in PBS. Next, sections were blocked with TBS (0.05 M Tris-HCl [pH 7.8], 0.1 M NaCl) supplemented with 20% BSA, 20% FBS and 5% Triton X-100 for 6 hours at room temperature, followed by incubation with 5×10⁷ GFP expressing Fim ON per ml blocking solution, overnight at 4°C. Slides were then washed three times with PBS for 10 min each, and incubated with Hoechst 33258 (Sigma-Aldrich, cat. No. 94403) diluted 1:5000 in PBS for 15 min at room temperature.

For competition assay, Fim ON was pre-incubated with 400 mM mannose or 400 mM galactose or 0.001 mM unconjugated ConA (Sigma-Aldrich, cat. No. L7647) in PBS for 30 min prior to incubation with sections.

Pictures were taken using the Nikon Eclipse Ti microscope.

Several cores were found to have undergone distortion, assumingly caused by improper handling of the slides, and were excluded from the analysis.

Fluorescence analyses were performed using NIS Elements microscope imaging software version 5.0 (Zaccard et al., 2021). H&E staining of the sections containing tumor and normal samples is available at US Biomax website (links provided above in “Tissue microarray analysis”). For tumor samples, the whole core was determined as region of interest (ROI). For normal bladder samples, only the lumen epithelial area was determined as ROI. The ROI in each sample quantified by the software in mm².

For analysing fluorescence levels of the conjugated lectins, the software calculated the fluorescent intensities that were above the predetermined threshold, and then, these values were divided by the ROI area to get the final ratios (fluorescence intensity/area (mm²)).

The bacteria were counted by measuring the area of florescence (higher than the predetermined threshold) in the ROI, divided by mean fluorescence of the area of a single bacteria, which was determined in three arbitrary pictures. The bacterial numbers were divided by the ROI area to get the final ratios (bacteria/mm²).

Mouse bladder cancer cell lines, MBT2 (derived from C3H mouse) and MB49 (derived from C57BL/6 mouse), were cultured with RPMI with 10% fetal bovine serum, 1% L-Glutamine, penicillin–streptomycin and amino acids (Biological Industries).

FITC-labeled ConA (Sigma-Aldrich, cat. No. C7642) was incubated at a final concentration of 160 nM (in PBS) per 2.5x10⁵ MB49 cells in 96 well plates, for 30 min at room temperature and gentle shaking. Cells were then washed twice with PBS, prior to flow cytometry (Accuri, C6 flow cytometer, BD, USA). For competition experiments, mannose or galactose was added to wells, to a final concentration of 100mM or 400mM. Data analyses were performed using Flowjo 10.0.8 software (Tree Star, Ashland, OR, USA).

2.5x10⁶ GFP-expressing UPEC Fim ON or Fim OFF were incubated in PBS with MBT2 or MB49 cells at a multiplicity of infection (MOI) of 10 in 96 well plates, for 3 hours at room temperature and gentle shaking. The cells were then pelleted by centrifugation at 4°C, 1000 rpm, for 5 min and washed twice with PBS prior to flow cytometry (Accuri, C6 flow cytometer, BD, USA). For the competition experiments, mannose (100mM, 400mM), galactose (100mM, 400mM) or unconjugated ConA (80, 160, 320, 640 and 1280 nM) were added together with bacteria and cancer cells.

Flow cytometry analysis were performed using Flowjo 10.0.8 software (Tree Star, Ashland, OR, USA).

GraphPad Prism software (Version 6.0) used for figures graphics and statistical analysis as described in the figure legends. The various groups were compared against each other in matched pairs, we therefore used Wilcoxon test or t-test to calculate p values. The multiple comparisons were adjusted via Bonferroni corrections. Differences were considered statistically significant at p < 0.05.

To compare mannose level in a large number of cancer and normal bladder samples under identical conditions, we used the BL244a human tissue microarray (TMA). This TMA contains samples of 12 bladder tumors and their matching normal tissues. Duplicate slides from this TMA were screened; one for mannose levels using a fluorescently conjugated, mannose-specific ConA lectin, and one for the level of uroplakin Ia, the mannose-enriched bladder glycoprotein, using fluorescently conjugated anti-uroplakin Ia. Uroplakin Ia levels in bladder cancer samples were found to be similar to those in the matching normal bladder control tissues ( Figure 1A ). In contrast, higher mannose levels were detected in bladder cancer samples compared to the matching normal bladder uroepithel ( Figure 1B ), as represented by the intense fluorescent staining shown in Figure 1C .

It was previously reported that Gal-GalNAc level is elevated in certain cancers (Abed et al., 2017), but increase of Gal-GalNAc was not detected in bladder cancer (Neal et al., 1987; Abed et al., 2017). To confirm these results, and quantify Gal-GalNAc level in bladder cancer, the BL244a TMA was stained with fluorescently labeled PNA, a Gal-GalNAc specific lectin. Unlike mannose, no significant differences in Gal-GalNAc levels could be detected between tumor and normal bladder tissue ( Figure 1D ).

We next used flow cytometry and competition assays to verify mannose expression in bladder cancer cells. We measured attachment of FITC-conjugated ConA to the MB49 mouse bladder cancer cell line in the presence or absence of two different concentrations of soluble mannose, or galactose used as control ( Figures 1E, F ). In agreement with previous results (Huo et al., 2022), we detected ConA binding to the MB49 cells ( Figures 1E, F ). Importantly, ConA attachment to the MB49 cells was inhibited by mannose in a statistically significant, dose-dependent manner ( Figure 1F ). Despite the structural similarity between mannose and galactose ( Figure 1G ), we were unable to detect any changes in ConA attachment to MB49 cells when galactose was used as a competitor ( Figures 1E, F ). Taken together, these results support the early ones (Huo et al., 2022) reporting that mannose is overexpressed in the surface of bladder tumors of humans and mice.

The type 1 fimbria of uropathogenic E. coli (UPEC) binds mannose residues present on uroplakin Ia in the uroepithelium (Zhou et al., 2001). In particular, the mannose-binding adhesin FimH at the tip of the fimbria, mediates this bacterial-mannose interlinkage (Subashchandrabose and Mobley, 2015). In accordance with the results showing higher mannose levels in bladder cancer cells compared to non-cancerous ones ( Figures 1B, C ), we expected an increased attachment of type 1 fimbria - expressing bacteria to bladder cancer cells. To investigate this, we compared attachment of the GFP-expressing Fim ON and Fim OFF mutants of UPEC CFT073 (see Materials and Methods) to bladder cancer cell lines. Flow cytometry analysis of the attachment of the fluorescent bacteria to the MBT2 ( Figure 2A ) and MB49 ( Figure 2B ) mouse bladder cancer cell lines, revealed that type 1 fimbria plays a key role in UPEC attachment to mouse bladder cancer cells ( Figures 2A, B ). This type 1 fimbria mediated attachment is mannose sensitive and unaffected by galactose used as control ( Figure 2B ). Lastly, we tested whether the mannose lectin ConA, can compete with the attachment of GFP-expressing Fim ON to MB49. As expected, ConA (similar to mannose used as control), impaired the attachment of Fim ON to the bladder cancer cells ( Figure 2C ).

Due to its mannose-specific binding (McLellan et al., 2021; Mohammad Zadeh et al., 2021), we used type 1 fimbria expressing UPEC to confirm higher mannose levels in human bladder cancer tissue, compared to non-cancerous bladder uroepithelium. Attachment of GFP-expressing Fim ON to human tumor or normal bladder tissue was monitored in the absence or presence of soluble mannose, galactose ( Figure 2D ), or ConA ( Figure 2E ). Fluorescence-microscopy analysis demonstrated a significantly higher attachment of Fim ON to bladder cancer tissue, compared to normal bladder surfaces. Mannose, as opposed to galactose used as control, significantly impaired Fim ON attachment to bladder tissues (both normal and cancer) and eliminated the increased binding of Fim ON to the cancer tissues ( Figure 2D ). ConA, the mannose lectin, also competed with Fim ON, and significantly reduced its binding to bladder cancer tissues (similar to mannose used as control) ( Figure 2E ).

These data indicate that mannose is overexpressed in human and mouse bladder cancer, and that type 1 fimbriated UPEC binds to bladder cancer cells in a preferred manner compared to normal bladder tissue.