Female C57BL/6 mice were purchased from Charles River Laboratories (Indianapolis, Indiana, USA). OT-I transgenic mice (C57BL/6-Tg (TcraTcrb) 1100Mjb/J) were purchased from The Jackson Laboratory. Mice were housed and bred in the Comparative Medicine Facility at Moffitt Cancer Center where they were monitored daily. When tumors reached 400 mm² (subcutaneous) or 250–300 mm³ (orthotopic) mice were humanely euthanized with CO₂ followed by cervical dislocation.

MB49 murine bladder cancer cells were a kind gift from Dr Jeffery Schlom (National Cancer Institute, Bethesda, Maryland, USA).27 An ovalbumin (OVA) expressing fluorescent MB49 cell line (MB49OVA) was generated by exposing cells to supernatants containing a lentiviral vector comprising a fluorescent ZsGreen (ZsG) protein and OVA. The viral supernatant was a kind gift from the Dr Brian Ruffell laboratory (H. Lee Moffitt Cancer Center, Tampa, Florida, USA). After transduction, ZsGreen^hi tumor cells were sorted by FACS using BD FACSAria. MB49OVA tumor cells were passaged in vitro 4 times. OVA expression was confirmed by staining for H2-K^b bound to SIINFEKL peptide (25-D1.16, BioLegend). Cell lines were cultured in Complete Medium (CM) consisting of RPMI 1640 supplemented with 10% heat-inactivated fetal calf serum, 0.1 mM non-essential amino acids, 1 mM sodium pyruvate, 2 mM fresh L-glutamine, 100 mg/mL streptomycin, 100 U/mL penicillin, 50 mg/mL gentamycin, 0.5 mg/mL fungizone (all from Life Technologies, Rockville, Maryland, USA) and 0.05 mM 2-ME (Sigma-Aldrich, St. Louis, Missouri, USA). After thawing, cells were passaged no more than 10 times and routinely tested negative for mycoplasma.

Mice were inoculated with 5.0×10⁵ MB49 or MB49OVA cells subcutaneously at the right ventral abdomen or 1.0×10⁵ MB49 or MB49OVA cells orthotopically. T cells from MB49 or MB49OVA tumors were isolated using EasySep Mouse CD90.2 Positive Selection Kit II (Stem Cell Technologies) according to the manufacturer’s protocol. Isolated cells were plated in CM+IL-2 (100 IU, Prometheus).

MB49OVA cells (5.0×10⁵) were injected subcutaneously at the right ventral abdomen. Once tumors were palpable, mice received intraperitoneal injections of cyclophosphamide (200 µg) followed by fludarabine 24 hours later (100 µg). One day after fludarabine, OT-I T cells were isolated from splenocytes using EasySep Mouse CD8⁺ T Cell Isolation Kit (Stem Cell Technologies) according to the manufacturer’s protocol and delivered intravenously in 250 µL phosphate-buffered saline (PBS). IL-2 (2.5e⁵ IU) was given intraperitoneally every 12 hours for 3 days following T cell injections. Control mice did not receive T cells or IL-2 injections. Tumor size was measured every 2–3 days.

MB49 or MB49OVA cells were injected orthotopically.28 Briefly, mice were put under anesthesia and catheters were placed. Poly-L-lysine (50 µL at 1 µg/mL) was injected into the bladder for 10 min. Bladders were washed with PBS (50 µL) and tumor cells were injected for 30 min (1.0×10⁵ cells in 100 µL of PBS). The presence of tumor was confirmed by ultrasound-guided imaging using the Vevo 2100 imaging system (FUJIFILM VisualSonics). Approximately 7 days after injection when tumors were detected, OT-I T cells or MB49 TILs (isolated as described above) were infused directly into the bladder while mice were under anesthesia with catheters in place for 3 hours. Mice were not given lymphodepleting chemotherapy prior to T cell delivery or IL-2 post-T cell delivery. Tumor size was measured by ultrasound every 3–4 days.

Tumors were processed into a single cell suspension by enzymatic digestion using Hanks’ Balanced Salt Soltuion (HBSS) (Life Technologies) containing 1 mg/mL collagenase, 0.1 mg/mL DNAse I and 2.5 U/mL of hyaluronidase (all from Sigma-Aldrich). Samples were incubated at 37°C for 45 min with constant agitation. Samples were passed through a 70 µm strainer and washed with CM. Red blood cells (RBCs) were lysed using RBC lysis buffer (Biolegend) for 5 min. Spleens were processed by physical dissociation through a 100 µm strainer followed by RBC lysis.

Tumors and spleens were processed into single cell suspensions as described above. Cells (1×10⁶) were washed with 1 mL PBS. Zombie NIR viability dye (Biolegend, 1:1000) was added to each sample and incubated at room temperature for 30 min in the dark. Samples were washed with flow buffer (1.0 L PBS, 5% heat-inactivated fetal bovine serum, 1 mM EDTA, 0.1% sodium azide) and stained for OVA tetramer (MBL International, Woburn, Massachusetts, catalog# TB5001-1) at 4°C for 20 min in the dark followed by cell surface markers CD3 (catalog# 565643), CD8 (catalog# 558106), CD4 (catalog# 550954), CXCR3 (catalog# 562152) and CD45.1 (catalog# 560578, all from BD Biosciences, San Jose, California, USA) at 4°C for 20 min in the dark. Samples were run on a FACSCelesta flow cytometer (BD Biosciences) and analyzed using FlowJo V.10 software.

OT-I T cells were plated in a co-culture 10:1 with irradiated MB49 or MB49OVA tumor cells or OVA SIINFEKL peptide (1 µg/mL) as a positive control. Supernatants were collected after 48 hours. Interferon-gamma (IFN-γ) levels were analyzed using Mouse IFN-γ ELISA Kit II (RUO) (BD Biosciences) according to manufacturer’s protocol. Briefly, supernatants were diluted 1:10, 1:20 or 1:100 and plated along with the IFN-γ standard serial dilutions in wells containing 50 µL ELISA Diluent and incubated for 2 hours. Wells were washed 5 times. Working Detector (100 µL) was added to each well and incubated for 1 hour. Wells were washed 7 times. TMB One-step Substrate Reagent (100 µL) was added to each well and incubated for 30 min. Stop Solution (50 µL) was added to each well and absorbance was read at 450 and 570 nm.

OT-I T cells were isolated as described and incubated for 30 min on ice with either normal rat IgG isotype control or CXCR3 blocking antibody (500 µg, catalog# BE0249, BioXcell). In vitro, OT-I T cells were plated in normal CM in the upper well of a Corning Transwell insert with MB49OVA tumor digest conditioned media in the bottom. Cells were collected from the bottom chamber and counted 24 hours later. In vivo, mice were inoculated with orthotopic MB49OVA tumors as described above. Tumor-bearing mice were randomized and treated with antibody-coated OT-I T in 100 µL of PBS cells on day 7. Tumors were collected 3 and 24 hours after instillation, washed 3 times in PBS, digested and analyzed for CD45.1⁺ T cell infiltration by flow cytometry. An additional cohort was monitored for tumor growth by ultrasound-guided imaging every 2–3 days.

Two human bladder cancer TIL samples were screened for CXCR3 (catalog# 562558, BD Biosciences) expression by flow cytometry as described above. Informed consent was obtained prior to surgical resection under an IRB-approved study (MCC 18142). Both samples were from patients with muscle-invasive bladder cancer undergoing radical cystectomy. One patient had neoadujvent cisplatin-based chemotherapy and one patient did not receive chemotherapy. TIL was expanded as previously described.26

For comparison of in vitro studies with two experimental groups, a Student’s t-test was performed. For in vitro studies with more than two experimental groups, a one-way analysis of variance (ANOVA) was performed, followed by a Tukey’s post hoc comparison. For in vivo studies with two groups, the CGGC permutation test was used.²⁷ For in vivo studies with three or more treatment groups, a two-way ANOVA (Tukey’s post hoc) was performed at each time point. GraphPad Prism was used to perform statistical analysis. Data with a p value <0.05 were deemed statistically significant.

In order to study antigen-specific T cell response using ACT, we developed a model using a strong antigen and target T cell. MB49 cells were transduced to express OVA which is recognized by OT-I T cells. OT-I mice are transgenic mice containing TCRα-V2 and TCRβ-V5 specific for OVA peptide residues 257–264 in the context of H2K^b. Figure 1A,B show successful incorporation of ZsGOVA and surface expression of H2K^b SIINFEKL in MB49OVA cells as determined by flow cytometry. To confirm OT-I T cells are specific for MB49OVA cells, a co-culture was performed with the parental MB49 cell line, MB49OVA cell line and OVA SIINFEKL peptide as a positive control. IFN-γ production was measured in the co-culture supernatants by an ELISA assay. OT-I T cells specifically recognize MB49OVA cells and not the parental MB49 cells that lack OVA (figure 1C). TILs were isolated from MB49OVA subcutaneous tumors in C57BL/6 mice and a co-culture assay was performed with the same conditions (figure 1D). TIL produced IFN-γ after co-culture with MB49OVA cells and to a lesser degree with the parental MB49 cells. These data validate our MB49OVA/OT-I model, which will be used in subsequent experiments.

To investigate the potential for systemic ACT to treat bladder cancer, MB49OVA cells (5.0×10⁵) were injected subcutaneously into female C57BL/6 mice. Twenty days later when tumors were established, mice were randomized and treated with NMAC (cyclophosphamide/fludarabine) followed by intravenous infusion of OT-I CD8⁺ T cells (5.0×10⁶ cells/mouse) and intraperitoneal injection of IL-2. ACT-treated mice had a reduction in tumor growth compared with control-treated mice (figure 2A). On day 50, tumors and spleens were processed, and immune cell populations were examined by flow cytometry. CD45.1⁺ T cells, a congenic marker used to distinguish the transferred OT-I T cells from the recipient mice T cells, were detected both in the spleen and tumors of ACT-treated mice and not detected in control-treated mice (figure 2B,C). In addition, there was an increase in CD3⁺ and CD4⁺ and CD8⁺ OVA tetramer⁺ T cells in the tumors of ACT-treated mice (figure 2C). An increase in CD8⁺ OVA tetramer⁺ T cells was also measured in the spleens of treated mice (figure 2C), with no difference in total T cells (online supplemental figure 1A). ACT-treated mice that had complete tumor regression were subsequently challenged with the parental MB49 subcutaneous tumors 100 days later. All mice rejected MB49 tumors compared with naïve mice (online supplemental figure 2A). Splenocytes taken from re-challenged mice were able to produce IFN-γ in a co-culture with either MB49 or MB49OVA tumor cells (online supplemental figure 2B). Lastly, flow cytometric analysis of the splenocytes confirmed the presence of both CD45.1⁺ CD8 T cells and OVA tetramer⁺ CD8 T cells over 100 days after adoptive transfer of OT-1 T cells. These data suggest that systemic ACT can decrease tumor burden in our model, and that adoptively transferred T cells can persist systemically. In addition, epitope spreading can occur as mice were able to reject rechallenge with the parental MB49 cell line lacking OVA.

To determine the potential to treat bladder tumors with intravesical TIL, we first determined how long it took T cells to migrate into the tumor. Mice were inoculated with MB49OVA cells (1.0×10⁵) orthotopically. The presence of tumor in the bladder was confirmed by ultrasound-guided imaging. One-week post-tumor inoculation, tumor-bearing mice were treated with OT-I CD8⁺ T cells (4.0×10⁶ cells/mouse) label with CellTrace Violet (CTV). Mice were put under anesthesia for 1–3 hours with catheters in place to prevent the cells from leaking out of the bladder. Bladder tumors were collected and washed with PBS. Tumors were digested and analyzed by flow cytometry for the presence of CTV-labeled T cells. After 3 hours, CTV⁺ cells were detectable within the tumor digest (figure 3A). These data suggest that 3 hours is sufficient for T cells delivered intravesically to migrate into the tumor.

Next, to test the efficacy of intravesical ACT, mice were inoculated with MB49OVA orthotopic tumors. One-week later, tumor-bearing mice were randomized into two groups. Mice were placed under anesthesia for 3 hours and treated with OT-I CD8⁺ T cells (4.0×10⁶ cells/mouse) or PBS. Tumor volume was monitored by ultrasound-guided imaging and quantified using Vevo Lab V.3.1.0 software. When mice reached end point, tumors and spleen were collected and processed. Flow cytometry analysis was performed to determine T cell infiltration as well as detect OVA tetramer⁺ and CD45.1⁺ adoptively transferred T cells. Intravesical ACT prevented tumor growth and improved survival compared with PBS-treated mice (figure 3B,C, representative ultrasound images figure 3E). Treated mice had an increase in CD3⁺, CD8⁺ and OVA tetramer⁺ T cells within the tumor (figure 3D). No differences were seen in bulk CD3⁺, CD4⁺, CD8⁺ or CD8⁺ OVA tetramer⁺ cell populations in the spleens, which was expected since this was a local therapy (online supplemental figure 1B-C). Results from this experiment confirmed intravesical T cell delivery can control tumor growth without lymphodepleting chemotherapy or IL-2.

While an OVA-based tumor model can be used to optimize treatment strategies, OVA is a highly immunogenic foreign antigen and does not represent the antigens found in patient tumors. To more accurately model ACT using TILs, orthotopic MB49 tumors (1.0×10⁵) were established in mice. Tumors were collected after 2 weeks and TILs were isolated. TILs were plated in 100 IU IL-2. Over the span of 25 days, TILs increased in number from 1.5×10⁶ (1 well) to 39.2×10⁶ (24 wells) as seen in figure 4A. IFN-γ production was determined by ELISA after co-culturing TILs with MB49 tumor cells after 1 or 7 days in culture (figure 4B). The fold increase of IFN-γ was highest after 7 days in culture compared with the CM condition. TILs were predominately CD3⁺CD8⁺ T cells (90.2%, figure 4C). Next, TILs isolated from 20 MB49 orthotopic tumors and expanded in culture for 7 days was adoptively transferred intravesically into mice bearing orthotopic MB49 tumors on day 8. Tumor growth was abrogated in mice treated with TILs compared with PBS control-treated mice (figure 4D, online supplemental figure 3). A representative image of PBS and TILs treated tumors −2 days prior to TILs delivery and 17 days after treatment started can be seen in figure 4E. These data confirm intravesical TILs delivery can decrease tumor growth.

Next, to determine a potential mechanism by which adoptively transferred T cells migrate from the intravesical space into bladder tumors, we first screened orthotopic MB49 and MB49OVA bladder tumors for the chemokines CXCL9, CXL10, CXCL11 by PCR after tumor digestion and RNA isolation (figure 5A). We then confirmed expression of T cell activation and migration marker CXCR3, which binds CXCL9/10/11, on OT-I T cells, MB49 TIL and two patient TIL samples by flow cytometry (figure 5B,C). Blocking CXCR3 decreased migration of OT-I T cells in vitro (online supplemental figure 4). Next, we adoptively transferred OT-I T cells (3.5×10⁶) into the bladders of mice bearing orthotopic MB49OVA tumors. T cells were coated with a CXCR3 blocking antibody or control normal rat IgG prior to intravesical delivery (online supplemental figure 4). Two mice per group were euthanized at 3 and 24 hours after transfer. Tumors were collected and screened for CD45.1⁺ T cells (figure 5D). Mice treated with T cells coated with αCXCR3 antibodies had a reduction in CD45.1⁺ T cells within the tumor at both time points. In a second cohort of mice, tumor growth was monitored by ultrasound after treatment. CXCR3 antibody-coated OT-I T cells were not able to delay tumor growth compared with control rat IgG-coated OT-I T cells (figure 5E). These data suggest that CXCR3 is at least in part responsible for T cell infiltration into bladder tumors from the intravesical space.