Cells and cell cultures

Human bladder cancer cell lines (T24 and RT112) and the murine bladder carcinoma cell line MB49-luc (a kind gift from Prof. Gilad Bachrach, The Hebrew University-Hadassah School of Dental Medicine, Jerusalem, Israel) were cultured at 37 °C in 5% CO2 in Dulbecco's Modified Eagle Medium (DMEM; Gibco-ThermoFisher, Waltham, MA, USA) supplemented with 10% (v/v) fetal calf serum (FCS; HyClones; Utah, USA), penicillin (100IU/ml), streptomycin (100 µg/ml; Biological Industries, Beit Haemek, Israel), and 1% (v/v) sodium pyruvate (Biological Industries, Beit Haemek, Israel), denoted throughout the paper as “culture medium”.

Bacterial cultures

The M. bovis BCG was prepared after reconstitution in phosphate buffer saline (PBS) of freeze-dried preparation containing BCG (2–8 × 108 CFU/Vial; OncoTICE; Meck Sharp and Dohme, Petah-Tikva, Israel). The concentration was then adjusted as indicated in the figure legends.

Human peripheral blood eosinophil purification

Human eosinophils were purified as previously described [26] from the peripheral blood of asymptomatic mildly atopic volunteers (blood eosinophil levels 5–10%) not taking any medication. Written informed consent was obtained according to the guidelines of the Hadassah-Hebrew University Human Experimentation Helsinki Committee (0410-14-HMO). Venous blood (150 ml) was collected in heparinized syringes and left to sediment in 6% dextran (Sigma-Aldrich, Jerusalem, Israel). Leukocytes were loaded on lymphoprep solution (STEMCELL Technologies; Vancouver, Canada) and centrifuged (25 min; room temperature (RT); 1400 rpm with no brakes). Neutrophils and contaminating lymphocytes were tagged in the granulocyte-enriched pellet with micromagnetic beads bound to anti-CD16 and anti-CD3 (Miltenyi Biotec, North Rhine-Westphalia, Germany). Eosinophils were purified by passing the cell suspension through a magnetic column (LS-MACS). Eosinophils were evaluated by Kimura staining and by flow cytometry using APC-anti-human CCR3 (Abcam, Cambridge, UK) and PE anti-human Siglec-8 (BioLegend, California, USA) and collected at a purity of > 90%, with viability of > 98% (trypan blue staining). Then, eosinophils were re-suspended in eosinophil medium consisting of RPMI-1640 supplemented with 10% heat-inactivated FCS, penicillin–streptomycin solution (100 u/ml; Biological Industries), and granulocyte–macrophage colony-stimulating factor (GM-CSF; 5 ng/ml; Peprotech, Rocky Hill, NJ, USA).

Human cord blood derived mast cells purification

CBMCs were obtained as previously reported [27]. Human umbilical cord blood was obtained from normal births after informed consent and according to the guidelines of the Hadassah-Hebrew University Human Experimentation Helsinki Committee (7-14.01.05). CD34+ progenitor cells were enriched by lymphoprep centrifugation and grown in MEM-α medium supplemented with 2 mM L-glutamine, ribonucleases (10 µg/mL; Biological Industries), stem cell factor (SCF; 100 ng/ml; Peprotech), IL-6 (10 ng/mL; Peprotech), and PGE2 (0.3 µM; Sigma-Aldrich).

CBMCs were harvested for the experiments between 7 and 9 weeks of culture when > 95% of the cells were stained metachromatically with toluidine blue staining.

Bladder cancer viability assay

As a marker of cell viability, bladder cells metabolic rate/mitochondrial activity was determined by 3-(4,5-dimethylthiazol-2-yl)-2,5-diphenyltetrazolium bromide (MTT; Sigma-Aldrich). T24 or RT-112 cells (5 × 103 cells/100 µl/well) were incubated overnight in a 96-well flat bottom plate (ThermoFisher, Nunc, Kiryat Shmona, Israel). The medium was then replaced with fresh culture medium containing CBMCs or eosinophils in 1:1, 1:5 and 1:10 ratio. The plate was then incubated at 37 °C, 5% CO2 for additional 48 h. Following the incubation period, CBMCs and eosinophils in suspension were removed, and the adherent cancer cells were washed once with PBS and incubated for additional 2.5 h in a fresh culture medium containing 10 µl MTT. Then, the plate was centrifuged (1250 rpm, 5 min, 4 °C), the medium was discarded and 100 µl dimethyl sulfoxide (DMSO; Sigma-Aldrich) was added to dissolve the crystals. Optical density (O.D.) was measured at 565 nm in an ELISA reader (BIO-TEK, 185 Winooski, VT, USA).

Chemotaxis assay

The chemotaxis assay was performed using a 5 µm transwell plate (Getter Biomed, Israel). The lower chamber was seeded with T24 cells (2 × 105/well/750 µl) in their medium and incubated overnight at 37 °C, 5% CO2 and the upper chamber was loaded with fluorescein isothiocyanate (FITC)-labelled CBMCs or pbEos (3 × 105/well/300 µl). Briefly, MCs and eosinophils were stained with 1 µl FITC/105 cells for 30 min at RT, washed twice with PBS and resuspended in 300 µl/2 × 105 cells. Subsequently, the plate was incubated in 37 °C, 5% CO2 for 24 h to allow migration across the membrane in response to full medium only or to cells with medium. Twenty-four hours later, the filter inserts were removed and 250 µl was aspirated from the lower chamber. Then the number of MCs and eosinophils migrated into the lower chamber was quantified by flow cytometry, by counting the number of FITC+ cells during 30 s. For chemotaxis assay with BCG and bladder cancer cells, the BCG (MOI 10) was added simultaneously with the CBMCs and eosinophils in the upper chamber.

Orthotopic bladder cancer model

The orthotopic bladder cancer model was performed according to the previously described protocol [28]. Nine to 14-week-old female wild type C57/BL6, MC (Kitw−sh; Sash)—and eosinophil—(ΔdblGATA; GATA) deficient mice (both from C57/BL6 background) were maintained at our animal care facility for 1 week prior to use. Mice were anesthetized by i.p injection of 85 mg/kg ketamine/15 mg/kg xylazine diluted in 1:1 with PBS.

To ensure an efficient laser coagulation, an area of 1 cm2 was shaved on the backs of the mice and a 24-gauge Teflon i.v. catheter (Becton Dickinson, New Jersey, United States) was inserted transurethrally into the bladder using paraffin. Mice were then placed with their backs on the ground plate of the cautery unit. To optimize contact, we used electrocardiogram electrode contact gel (Covidien; Dublin, Ireland). The soft tipped end of a spring-wire guide of a 24-gauge central venous catheter (Shuyou, Willich, Germany) was inserted into the bladder via the Teflon catheter and gently pushed forward until it reached the bladder wall. The guide wire was attached to the cautery unit and a monopolar coagulation was applied for 5 s at the lowest setting (5 W). After removal of the guide wire, 100 µl of PBS containing 3 × 104 murine bladder carcinoma cell line transfected with luciferase (MB49-luc) was instilled. The catheters were pinched off with a clamp, kept locked and left in place until the mice awakened (Supplementary Fig S1B left panel). Using this method, we ensured a dwell time of 2–3 h. Ten days after tumor inoculation, mice were injected i.p with D-luciferin (Sigma) and analyzed by bioluminescence imaging (IVIS). Mice in which no tumor was observed were removed from the experiment (Supplementary Fig. S1B right panel). After 40 days mice were sacrificed, and the bladder tumor was extracted and weighted. All experiments were performed according to the guidelines of the Hebrew university ethics committee for animal studies (MD-20-15950-5).

Eosinophil degranulation assay

PbEos (1 × 105/100 µl) were re-suspended in the specific medium/buffer according to the degranulation assay performed (see below). The cells were incubated with BCG (MOI of 1, 5, 10) in 96-well plates (37 °C, 45 min, 5% CO2). As positive control, eosinophils were incubated with granulocyte–macrophage colony-stimulating factor (GM-CSF; 100 ng/ml) at 37 °C for 20 min.

For EPX release, eosinophils were re-suspended in 0.1% BSA/PBS and incubated with BCG. EPX release was measured by a colorimetric assay using a standardized assay with purified human EPX (a kind gift of Prof. Gerald J. Gleich, University of Utah Health Sciences Center, Salt Lake City, UT, USA) and freshly prepared peroxidase substrate solution containing o-phenylenediamine (OPD; ThermoFisher). Data are expressed as EPX concentration, based on a standard curve.

At all-time points, an aliquot of infected cells was harvested and counted. This step allowed an exact quantification of cells as well as the determination of cellular viability by trypan blue exclusion. Recovery of cells was 80% in all experiments, with cell viability regularly exceeding 90% of total cells.

Mast cell degranulation assay

For activation of CBMCs, human myeloma IgE (0.3 µg/mL, Sigma‐Aldrich, #401152) was used for sensitizing the cells in the presence of recombinant human IL‐4 (Gibco, PeproTech, #200-04; 10 ng/mL) for 3 days. Subsequently, cells were washed twice and resuspended in Tyrode's buffer (consisting of 137 mM NaCl, 5.5 mM glucose, 2 mM KCl, 12 mM NaHCO3, and 0.3 mM Na2HPO4, supplemented with 1.8 mM CaCl2 and 0.9 mM MgCl2; pH 7.3. CBMCs were then incubated for 1 h with polyclonal rabbit anti-human IgE Ab (5 µg/mL, Dako, #A0094) as positive control or with BCG at MOI of 1, 5, and 10.

To measure the degranulation of CBMC, we measured the release of β-hexosaminidase as a readout. The supernatant was collected by centrifugation and the remaining cell pellet was lysed in Tyrode’s buffer containing 10% NP-40. Ten microliters of pellet and supernatant were incubated with 50 µL of substrate solution (1.3 mg/mL p-Nitrophenyl-N-acetyl-β-D-glucose in 0.1 M sodium citrate, pH 4.5).

The measurement of p-nitrophenol, generated by β-hexosaminidase was done by a spectrophotometric reader (BioTek Eon) at a wavelength of λ = 405 nm. The amount of degranulation in percent was determined as follows:

$$\text[\%]= \frac}\times \text} \times 100$$

At all-time points, an aliquot of infected cells was harvested and counted. This step allowed an exact quantification of cells as well as the determination of cellular viability by trypan blue exclusion. Recovery of cells was 80% in all experiments, with cell viability regularly exceeding 90% of total cells.

Evaluation of TNF-α release

PbEos and CBMCs (1 × 105/100 µl) were incubated with BCG (MOI of 1, 5, and 10) for 4 h (37 °C, 5% CO2) in 96-well plates. Following centrifugation at 250 g supernatants were collected and TNF-α levels were determined using commercial enzymatic immunoassay kits according to the manufacturer's instructions (Cat. No: 900-T25, Peprotech, Rocky Hill, NJ, USA).

Primary NMIUC patient samples

Formalin fixed paraffin embedded (FFPE) tissue sections and H&E-stained specimens obtained from each patient diagnosed with primary NMIUC of the bladder were obtained from the department of Urology of Hadassah Ein-Kerem Hospital (Israel). All the patients included in the study had completed 6 weekly intravesical injection with BCG following TURBT resection. The patients were followed up in the Urology Clinic according to the best treatment practice on a routine basis for at least 2 years after the first TURBT surgery. Patients who received immunosuppressive drugs were excluded from the study.

Immunohistochemistry (IHC) staining of MCs

Twenty six bladder sections were deparaffinized by immersing the slides through the following solutions: (1) xylene; two washes for 5 min each and a third wash for 1 min. (2) 100% ethanol, two washes, 10 min each; (3) 95% ethanol, 1 wash for 2 min; (4) 80% ethanol, 1 wash for 2 min; (5) 70% ethanol, 1 wash for 2 min; (6) Immersion in double distilled water (DDW).

To perform antigen retrieval, slides were treated with citrate buffer, pH 6 (HIER Citrate buffer pH = 6; Zytomed Systems; Germany) by microwave for 15 min (P100 for 3 min, P10 for 12 min). Slides were then cooled down and equilibrated to RT for 40 min and subsequently washed twice with wash buffer (Zytomed systems) for 4 min with gentle agitation.

Then the tissue on each slide was blocked using a hydrophobic barrier pen (Pink PAP pen, Zytomed system). For MCs staining, primary anti-tryptase antibody (Monoclonal Mouse Anti-Human Mast Cell Tryptase Clone AA1 Cat. No. M7052, Dako, USA) was diluted to the optimal concentration of 1:500 (determined via calibration experiments) in a blocking solution (Background blocker, BioSB, Germany), and applied to the tissue. The slides were then incubated in a humidity chamber overnight at 4 °C. Four control slides were incubated with mouse IgG1 isotype control (Cat. No. 14471482, eBioscience, ThermoFisher USA) at the same dilution as the anti-tryptase primary antibody (1:500).

The following day, slides were allowed to equilibrate to RT for 30 min. Then, they were washed twice with wash buffer for 4 min, with gentle agitation. Secondary antibody polymer conjugated to alkaline phosphatase (AP) (Zytochem Fast (AP) One-Step Polymer anti-Mouse/Rabbit/Rat, Cat. No. ZUC068-006 Zytomed) was added to the slides and incubated for 30 min at RT. Slides were then washed twice with wash buffer for 4 min with gentle agitation.

To visualize the enzymatic reaction ALK magenta chromogen (Polydetector Alk magenta, Goleta, CA, USA) was added to the slides and incubated for 20 min at RT. Slides were then submerged in DDW to stop the enzymatic reaction. Subsequently, slides were dipped quickly in hematoxylin (Gills Hematoxylin, Sigma Aldrich, Israel) and then dehydrated in the following solutions: (1) 95% ethanol for 2 min; (2) 100% ethanol twice for 1 min; (3) 50:50 ethanol:xylene for 1 min; (4) absolute xylene, tree times for 1 min each. Slides were then mounted using xylene based mounting medium (xylene based mounting medium; Bar Naor, Petah Tikva, Israel) and cover glass.

Immunohistochemistry (IHC) staining for eosinophils

Eosinophils were stained using the same protocol described above for MCs with different primary and secondary antibodies. For eosinophils staining, anti-EPX primary antibody (mouse monoclonal anti-human/mouse EPX, IgG2a, Clone MM25-82.2, J. Lee labs, Mayo Clinic, Arizona, USA) was diluted to the optimal concentration of 1:200 (determined via calibration experiments) in a blocking solution (Background blocker, BioSB, Germany), was applied to the tissue, and then the slides were incubated in a humidity chamber overnight at 4 °C. Control slides were incubated with mouse IgG2a isotype control (Negative Control Mouse IgG2a Cat.No. X 0943, Dako, USA) at the same dilution as the anti-EPX primary antibody (1:200). The following day, slides were allowed to equilibrate to RT for 30 min. Then, they were washed twice with wash buffer for 4 min with gentle agitation. Goat anti–mouse AP conjugated secondary antibody (Goat anti–mouse IgG H&L Alkaline Phosphatase; ab97020, Abcam, Israel) was added to the slide at optimal concentration of 1:200 and incubated for 60 min at RT. Slides were then washed twice with wash buffer for 4 min with gentle agitation.

Samples analyses

The integrity of the staining was examined and compared to isotype control samples by a pathologist. H&E samples were analyzed manually at ×400 magnification (high power field; hpf) using Nikon TL microscope (Core Research Facility, Faculty of Medicine, The Hebrew University of Jerusalem, Israel) whilst being blinded to the status of the patients. IHC staining of eosinophils and MCs was analyzed manually by an independent researcher blinded to the status of the patients at ×400 magnification using Olympus light microscope (School of Pharmacy, Faculty of Medicine, The Hebrew University of Jerusalem). Numerical values were assigned to each level of MCs and eosinophils count in the tissue as shown in Table 2. The total score for each sample was calculated as the average score from 5 fields with the maximal cell presence.

Table 2 Numerical values assigned to each level of tumor Eos/MCs count in each hpfRNA extraction

RNA extraction was performed using standard RNeasy kit for RNA purification (Qiagen, Hilden, Germany). Using a blade, tissue sections were scraped in tubes and immersed in 500 µl mineral oil for 10 min at 80 °C. Samples were then centrifuged (14,000 g, 1 min), washed with 1 ml ethanol and vortex for 15 s. Ethanol was then removed and samples were incubated in 180 μl protease K digestion buffer and 20 μl Proteinase K for 3-12 h at 55 °C. Then, 750 μl Trizol was added directly to the samples and incubated for 5 min and RT for lysis.

Then after, RNA was precipitated by adding 200 μl chloroform for 2–3 min at RT, and centrifuged at 12,000 g for 15 min, 4 °C. The upper aqueous phase was then collected to a new tube and an equal volume of 100% isopropanol was added. Samples were incubated for 10 min, centrifuged (12,000 g, 10 min, 4 °C) and the supernatant was discarded. The RNA pellet was then washed with 700 µl of 70% ethanol, centrifuged (12,000 g, 5 min, 4 °C), and the RNA pellets resuspended in 50 µl RNase-free water.

RNA-sequencing of patient’s samples

Quality-trimming and trimming of poly-A sequences and adapters was done with Cutadapt [29] (version 2.10). Then, reads were filtered using the fastq_quality_filter program of the FASTX package, to keep only reads with a quality threshold of at least 20 at 90% of the read. Then these were mapped to the Homo sapiens transcriptome and genome (GRCh38 with annotations from Ensembl release 99) using TopHat [30] (version 2.1.1). Quantification of raw counts was done using htseq-count [31] (version 0.12.4). Genes were filtered, retaining only those genes for which a minimum of three of the samples had an expression value of 1 count per million. Subsequent differential expression analysis was done with the DESeq2 package [32]. Default parameters were used, except for not using independent Filtering algorithm. Significance threshold was taken as padj < 0.1 (default).

RNA sequencing data were deposited in GEO database (GEO accession numbers: GSE291192).

Gene set enrichment analysis (GSEA)

Whole differential expression data were subjected to gene set enrichment analysis using GSEA [33] to determine whether a priori defined set of genes showed statistically significant concordant differences between two biological states. We used the hallmark gene set collection, taken from the molecular signatures database (MSigDB [33]), as well as the REACTOME and the NAKAJIMA pathways database and a manually curated gene set collection related to immune cells and functions. Gene sets with FDR (corrected p value < 0.05) were considered significant.

Statistical analysis

All statistical analyses were performed in GraphPad Prism software with data from three or more biologically independent replicates using two-tailed Student’s t-test or one-way ANOVA with Tukey’s multiple comparisons. Results are represented as mean ± standard deviation (SD). Statistical analyses were performed using GraphPad Prism 10.2.3.403 (GraphPad, San Diego, CA). The p values are *p < 0.05, **p < 0.01, ***p < 0.001, and ****p < 0.0001, NS-non-significant (p > 0.05).