Phototherapy with Cancer-Specific Nanoporphyrin Potentiates Immunotherapy in Bladder Cancer

Bladder cancer is one of the top 10 most common cancers. It is estimated that approximately 81,180 new cases will be diagnosed with 17,100 deaths from bladder cancer in the United States in 2022 (1). Immunotherapy with immune checkpoint inhibitors (ICI) plays major roles in the management of bladder cancer from early-stage non–muscle-invasive bladder cancer (NMIBC), to locally advanced and metastatic bladder cancer. However, the efficacy of ICIs is, in general, disappointing. For NMIBC unresponsive to Bacillus Calmette-Guérin (BCG) therapy, salvage anti-programmed cell death 1 (PD-1) antibody pembrolizumab intravenous treatment has a response rate of 40%, and less than 20% of patients retain response for 12 months or more (2). For patients with advanced or metastatic bladder cancer who do not have any disease progression after chemotherapy, anti-programmed death ligand-1 (PD-L1) antibody avelumab significantly improves the overall survival from 14 months of the best supportive care group to 21 months. Nevertheless, its median progression-free survival is only 3.7 months and objective response rate less than 10% (3). Anti-PD-1/PD-L1 antibodies are also used in patients who have disease progression after chemotherapy (if they do not receive ICI before), or cisplatin-ineligible patients whose cancer expresses PD-L1 or chemotherapy-ineligible patients. In all these treatments, ICIs as a single agent have the response rate of only around 20% (4–10). Hence, there is a dire unmet need for strategies to enhance the efficacy of ICIs to improve the overall treatment outcomes of bladder cancer of all stages.

Bladder cancer is an ideal disease for photodynamic diagnosis (PDD) and therapy (PDT), as it is already accessed by light and directly visualized during cystoscopy. PDD and PDT rely on cancer-specific accumulation of a photosensitizer that, when activated by light, emits fluorescence for PDD and generates reactive oxygen species (ROS) to kill cells for PDT. PDD with blue light cystoscopy and PDT have been widely studied for bladder diseases (11–14). However, current photosensitizers have poor selectivity, a low absorption band, poor bioavailability, and low efficiency (15). It has been shown that the difference of photosensitizer accumulation between cancer cells and normal cells is only two to three times (16), and many noncancerous conditions, such as infection and inflammation, can increase nonspecific uptake of photosensitizers. Hence, blue light cystoscopy has limited clinical applications, and PDT has not been approved for bladder cancer in the United States.

Recently we developed a bladder cancer–targeting peptide named PLZ4 (amino acid sequence: cQDGRMGFc, in which two small letters of “c” represent D-cysteine and capital letters represent L-amino acids) that binds to ανβ3 integrin (17, 18). It can specifically bind to human, dog, and mouse bladder cancer cells, but not to the confounding cells, such as normal urothelial cells, blood cells, fibroblasts, human vascular endothelial cells, and inflammation cells from a dog with chronic cystitis. We subsequently developed a porphyrin-based nanometer-scale micelles that are decorated with PLZ4 on the surface for cancer-specific targeting. This PLZ4-nanoporphyrin (PNP) can be used for PDD, PDT, and photothermal therapy in addition to many other applications (19–22). Compared with only a 2- to 3-fold difference of accumulation between cancer and noncancer cells of photosensitizers currently used in clinics, PNP achieves cancer-specific delivery 30 to 40 times more to adjacent normal urothelial cells in vivo. Furthermore, its potency increases by over 100-folds. PNP has already reached the Investigational New Drug (IND)-enabling pharmacology and toxicology studies supported by a Department of Defense grant.

Several studies showed that PDT can potentiate immunotherapy (23–28). PDT kills cancer cells, induces immunogenic cell death, releases tumor antigens, and stimulates innate immunity as well as adaptive immune response (reviewed in ref. 29). In this project, we aimed to determine whether PDT with PNP could potentiate ICIs in addition to its potential use of PDD and PDT in bladder cancer. If so, PNP can be developed to enhance the efficacy of ICIs along with PDD and PDT.

The pyropheophorbide, a (Por)-containing telodendrimer (PEG^5k-Por₄/CA₄) was synthesized via a solution-phase condensation reaction according to our established method (20). PLZ4-PEG^5k-CA₈ telodendrimer was synthesized by conjugation of alkyne-derivatized PLZ4 (CPC Scientific) to PEG^5k-CA₈ telodendrimer via click chemistry. PNP was obtained via a mixed micelle strategy. Briefly, 10 mg PLZ4-PEG^5k-CA₈ and 10 mg PEG^5k-Por₄/CA₄ were dissolved in chloroform and evaporated on a rotavapor to obtain a homogeneous dry polymer film. The film was reconstituted in 1 mL PBS with sonication to disperse into PNP solution.

The morphology and structure of PNP were examined by TEM (Talos L120C, FEI). Dynamic light scattering (DLS) was performed to measure the hydrodynamic particle size by Zetasizer NanoZS (Malvern). The UV-Vis spectra were measured with a spectrometer (UV-1800; Shimadzu). The absorbance range was set as 800 to 220 nm. The fluorescence spectra were collected by a fluorescence spectrometer (RF-6000; Shimadzu). The excitation was set as 412 nm, and the emissive band was set from 432 to 800 nm.

BBN963 murine bladder cancer cell line was kindly provided by William Kim, MD, PhD, at University of North Carolina (30). MB49 murine bladder cancer cell line (SCC148, RRID: CVCL_7076) and DC 2.4 murine dendritic cell line (SCC142, RRID: CVCL_J409) were purchased from Sigma. Once the cell lines were received, they were immediately expanded with freshly prepared medium, and 20 to 30 vials were cryopreserved in liquid nitrogen to ensure that each batch of cells was used within a few months. Mycoplasma test was performed with the PlasmoTest^TM Kit (InvivoGen; Catalog No.: rep-pt1) every 2 to 3 months. Because cell lines were used within 6 months of culture after arrival, no cell line authentication was performed. BBN963 cells were cultured with 90% DMEM + 10% FBS and MB49 cells were cultured with 90% RPMI + 10% FBS at a 37°C incubator with 5% CO₂. Anti-PD-1 antibody was purchased from BioXcell (Catalog No. BE0146; RRID: AB_10949053), and administered 200 μg/mouse, intraperitoneally twice a week. PNP (10 μg/mL) was administered 1 μg per 10 g body weight, intravenously once a week.

To study the cellular uptake of PNP, BBN963 cancer cells seeded in glass-bottom plates (5 × 10⁴ cells per well) were incubated with PNP (10 μg/mL) for 4 hours. Then the culture medium was discarded, and the cells were washed with PBS three times. The intracellular uptake of PNP was then observed on confocal laser scanning microscopy (CLSM; Leica).

Four groups (i.e., control, control + Laser, PNP, PNP + Laser) were used to study ROS production by PNP after laser treatment. BBN963 cells were incubated with PNP (10 μg/mL) for 4 hours. Then the culture medium was discarded and washed by PBS. Subsequently, the culture medium was replaced by containing DCFH-DA 10 μmol/L in 20,7′-dichlorofluorescein (DCF; Sigma, D6883) and laser treatment 4.2 J/cm² (Omnilux New-U LED panel with 635 nm light) was applied. After incubation for another 30 minutes, cells were washed with PBS three times and scanned by CLSM. ROS carbonylation of proteins was determined with Abcam Carbonyl Kit (Abcam, ab126287) per manufacturer's protocol.

DC 2.4 cells were seeded at the lower compartment and cultured overnight. A total of 3 × 10³ BBN963 cells received one of the following four treatments: PBS control, PBS control + Laser, PNP, or PNP + Laser, 24 hours prior to culture of DC 2.4 cells at the upper wells. After coculture for 24 hours, DC 2.4 cells were harvested and stained with anti-CD80 PerCP/Cyanine5.5 and anti-CD86 PE/Dazzle 594 followed by flow cytometry (BD Fortessa) analysis.

All in vivo experiments were performed following a protocol approved by University of California Davis IACUC (No. 17763). Eight-week-old female C57BL/6 albino mice (C57BL/6BrdCrHsd-Tyr^c) were purchased from Envigo. MB49 cells (1 × 10⁵) were injected into the right flank of the mice subcutaneously as a primary tumor. Seven days later, mice were randomized into four groups: control, anti-PD-1 (BioXcell, 200 μg i.p., twice per week), PNP (100 mg/kg in 50 μL, i.v., weekly) + Laser (L, 0.2w, 3 minutes), and anti-PD-1+PNP+L. Laser treatment was applied to PNP and combination groups at 24 hours post-injection of PNP. One week after the initial treatment, 1 × 10⁵ MB49 cells were injected into the left flank as a secondary site. In total, two rounds of treatments were applied to these mice. Tumors were harvested for analysis.

C57BL/6 mice carrying BBN963 syngeneic tumors were used for this study. PNP 100 mg/kg was intravenously administrated. The near-infrared fluorescence (NIRF) in tumor was monitored by animal imaging system at different time points. Mice were euthanized at 24, 48, 72, and 96 hours after PNP administration. Tumors and major organs were harvested for NIRF imaging.

For the lung metastasis model induced by tail vein injection, MB49 cells (1 × 10⁶) were injected into the right flank of C57BL/6 mice. When tumor sizes reached around 100 mm³, C57BL/6 mice were randomly divided into four groups (n = 4): control, anti-PD-1, PNP+L, and anti-PD-1+PNP+L. PNP was intravenously injected into the mice on days 1 and 8, followed by laser treatment 24 hours after. Anti-PD-1 treatment was applied twice a week, intraperitoneally. After the first round of treatment, MB49 cells (1 × 10⁶ per mouse) were injected to the left flank to determine abscopal effects. Left-side tumors were harvested on day 15 for analysis.

Similarly, MB49 cells (1 × 10⁶) were injected into the right flank of C57BL/6 mice. When tumor sizes reached around 100 mm³, C57BL/6 mice were randomly divided into four groups (n = 5): control, anti-PD-1, PNP+L, and anti-PD-1+PNP+L. PNP was intravenously injected into the mice on days 1, 8, and 15 followed by laser treatment 24 hours after. Anti-PD-1 treatment was applied twice a week, intraperitoneally. After the first round of treatment, MB49 cells (1 × 10⁶ per mouse) were injected via mouse tail vein on day 7 to induce lung metastasis. Lungs were harvested on day 22 for hematoxylin & eosin (H&E) staining.

For the spontaneous lung metastasis model, MB49 cells (5 × 10⁶) were injected into the right flank of C57BL/6 mice. When tumors reached about 100 mm³, mice were randomly divided into four groups (n = 5): control, anti-PD-1, PNP+L, and anti-PD-1+PNP+L. PNP was intravenously injected into the mice on days 1, 8, and 15 followed by laser treatment 24 hours after. Anti-PD-1 treatment was applied twice a week, intraperitoneally. Lungs were harvested on day 22 for H&E staining.

BBN963 cells (5 × 10⁶) were subcutaneously injected into the right and left flanks of each C57BL/6 mouse. Fourteen days later, these mice were separated into four groups (n = 8 in each group): control, anti-PD-1, PNP+L, and anti-PD-1+PNP+L. PNP was intravenously injected into the mice on days 1, 8, and 15. Laser treatment was applied to the tumors at the right flank 24 hours after PNP administration. Anti-PD-1 treatment was applied twice a week intraperitoneally. The tumor size was recorded as length and width, whereas the tumor volume was calculated as (length × width²)/2. The effects on tumor growth at the right flank were considered direct antitumor activity whereas the effects on the left flank tumors (not treated with laser) were considered as abscopal effects).

Mice carrying heterozygous UPII-Ras* [Ras with codon 61 (Q > L), abbreviated as Ras*], and heterozygous UPII-SV40T transgene were obtained from Xue-Ru Wu, MD, at New York University. Mice at 7 to 8 weeks old were crossbred to produce double-transgenic mice. These mice were genotyped before day 21 with the primers (sequence 5′-3′): RAS-F: TCCCACTCCGAGACAAAATC; RAS-R: ATTCGTCCACGAAGTGGTTC; SV40T-F: GGAAAGTCCTTGGGGTCTTC; SV40T-R: CACTTGTGTGGGTTGATTGC. Because of technical difficulty of inserting an optical fiber into the bladder of male mice for PDT, only female double-transgenic mice were used and randomized into four groups: control, anti-PD-1, PNP+L, and anti-PD-1+PNP+L. PNP was intravenously injected into the mice on days 1, 8, and 15. Laser treatment followed after 24 hours. Anti-PD-1 treatment was applied on days 2, 9, and 16. Tumors were monitored by MRI.

BBN963 cells (5 × 10⁶) or MB49 cells (1 × 10⁶) were subcutaneously injected into the right and left flank of each C57BL/6 mouse. Right side was designed as primary tumor and treated with light in the PNP or anti-PD-1 plus PNP groups whereas the left side was designed as secondary site with no light treatment.

Mice were treated with PBS, anti-PD-1, PNP+L, or anti-PD-1+PNP+L for 2 weeks. Left tumors were collected to make single-cell suspension. The harvested cells were analyzed by flow cytometry or cytometry by time of flight (CyTOF) and/or immunofluorescence staining. Antibodies for the tumor tissue are Zombie Aqua (Biolegend, 423102, 1:400), APC-CD3 (Biolegend, 100312, 1:100), Brilliant Violet 711-CD4 (Biolegend, 100550, 1:100), PE/Cy7-CD8a (Biolegend, 100722, 1:100), and BUV395-CD45(BD, 564279, 1:100). CyTOF analysis was performed at the Human Immune Monitoring Center at Stanford University. Tumor samples were stained with antibody–polymer conjugate cocktail (Maxpar Mouse Sp/LN Phenotyping Panel Kit, 16 Marker, #201306; Fluidigm Corporation; each at 1 μL/test) following the manual procedure. ELISA kits (R&D Systems; Mouse Luminex Discovery Assay, LXSAMSM) were used to test cytokine levels.

IHC staining was performed according to the manufacturer's instruction (BioGenex). IHC primary antibodies include anti-CD8 (CST, 98941S, 1:100) and anti-Ki-67 (Abcam, ab279653, 1:1,000).

Independent experiments were performed for each analysis described in this article. Student t test was used to compare continuous parametric data between two groups, whereas one-way ANOVA was used to compare multiple groups. Overall survival of mice was analyzed using Kaplan–Meier survival curve and log-rank test. Statistical analysis was performed by GraphPad Prism 9 software.

The data generated in this study are available within the article and its Supplementary Data files. For raw data that are not included in the paper or supplementary files, please contact the corresponding author (C.-x. Pan) for additional information.

There are two dendritic oligomers or telodendrimers in PNP, which self-assemble to form nanometer-scale micelles. Both telodendrimers have a linear polyethylene glycol (PEG; MW: 5,000 Dalton, PEG^5k). One telodendrimer has PEG conjugated with 8 cholic acid (CA) units at one end and PLZ4 at the other end (PLZ4-PEG^5k-CA₈); the other one has four CA units and four porphyrin analog pyropheophorbide-a (Ppa) units at the same end (PEG^5k-CA₄/Ppa₄). These two telodendrimers were mixed at a 1:1 ratio in an aqueous solution which self-assemble into PNP (Supplementary Fig. S1A). PNP was well dispersed in the aqueous solution and showed spherical-like vesicles of about 20 nm in diameter (Supplementary Fig. S1B). Dynamic light scattering demonstrated PNP was around 18 nm in diameter (Supplementary Fig. S1C).

The UV absorbance of PNP showed a slightly red shift of the Q-band of the porphyrin, as shown in Supplementary Fig. S1D. The fluorescence of porphyrin in PNP was almost completely quenched due to the aggregation-induced quenching phenomenon (31). In comparison, dissolving of PNP in ionic detergent SDS or organic solvent methanol (MeOH) restored strong fluorescence around 680 nm, suggesting the successful self-assembly of the nanostructures (Supplementary Fig. S1E).

During PDT, laser activates porphyrin to produce ROS, which can modify macromolecules and induce immunogenic cell death to trigger an immune response. To verify ROS production induced by PNP, we first assessed the cell uptake of PNP in murine bladder cancer cell line BBN963. After incubating BBN963 with PNP for 6 hours, most of the PNP was internalized into cells (Fig. 1A). ROS production was subsequently detected by a ROS-detecting dye 2′-7′-dichlorofluorescein diacetate (DCFH-DA; ref. 32) in BBN963 in the presence or absence of PNP and laser. As shown in Fig. 1B, almost no ROS was detected in the control group. Very little ROS was observed in the presence of PNP without light treatment. BBN963 cells treated with tert-butyl hydrogen peroxide (TBHP) as positive control are shown in Supplementary Fig. S2. In the PNP with laser treatment (PDT) group, ROS production was much higher than that in the other three groups, and at a comparable level as TBHP. To determine the modification of proteins by ROS, a protein carbonylation assay was performed and showed that more carbonylation was found in the PNP plus laser treatment group whereas there was no difference between the PNP, laser treatment group, and the control group (Fig. 1C). These results suggested that PNP could accumulate in cancer cells and be evoked to produce ROS, which then modified proteins during PDT.

Dendritic cells (DC) play important roles in initiating and regulating adaptive immune responses (33). During DC maturation/activation, costimulatory molecules, such as CD80 and CD86, are upregulated (34). To determine whether PDT could stimulate and induce DC maturation, we performed a transwell chamber experiment. Mouse bladder cancer cell line BBN963 cells with different treatments were cultured in the upper chamber whereas mouse DC cell line DC 2.4 cells were incubated in the lower chamber (Fig. 1D). Flow cytometry was performed to determine CD80 and CD86 expression on DC 2.4 cells. Detailed gating information is shown in Supplementary Fig. S3. PDT with PNP plus light induced much more CD11c⁺/CD80⁺/CD86⁺ DC cells than the other three groups (Fig. 1E): 61.5% ± 2.04% in the PNP plus light versus 35.63% ± 1.89%, 36.23% ± 0.48%, and 51.27% ± 0.89% in the control, control plus light, and PNP groups, respectively. Representative flow cytometry results in each group are shown in Supplementary Fig. S4. Compared with the control and light treatment groups, more DC maturation was observed when DCs were incubated with PNP alone (Supplementary Fig. S4). It is possible that natural light and microscopy examination activated PNP to produce ROS and stimulated DCs.

We hypothesized that PDT induced immunogenic cell death and released damage-associated molecular pattern (DAMP) molecules to stimulate DCs. To test this hypothesis, we determined two DAMP molecules in the lower chamber, adenosine triphosphate (ATP), and high-mobility group box 1 protein (HMGB1). Both molecules increased in the PNP+L treatment group (Supplementary Fig. S5).

Next, we determined whether PDT could potentiate immunotherapy in treating spontaneous localized bladder cancer. If complete remission can be induced with this combination, some patients may be spared from radical cystectomy, a procedure associated with the worse health-related quality of life among all cancers. The UPII-Ras* and the UPII-SV40T double-transgenic mice were employed for this study as these mice consistently develop spontaneous localized bladder cancer within 1 month and die from urinary obstruction within 5 weeks. Detection of restriction fragments representing the UPII-Ras* transgene (1.7 kb), the UPII-SV40T transgene (2.7 kb), and the endogenous UPII gene (1.4 kb; Supplementary Fig. S6) at the tail was used to confirm that a mouse carries both activated Ras (Ras*) and SV40T genes. First, we determined in vivo PNP delivery and distribution. After the formation of bladder cancer was confirmed, PNP was administrated through tail vein injection (Fig. 2A). Twenty-four hours later, the bladder and major organs were harvested to determine drug delivery. Compared with a small bladder from a wild-type mouse, two bladders from double-transgenic mice showed significant enlargement secondary to tumor development inside of the bladder. PNP accumulated in the neoplastic bladder after 24 hours (Fig. 2B). No significant PNP signal was detected in other major organs, including heart, liver, kidney, lung, testis, and intestine except cecum (Supplementary Fig. S7).

At Day 23 of age, PNP was administrated through intravenous injection followed by light treatment through an optic fiber inserted into the bladder cavity 24 hours later. Anti-PD-1 antibody was applied twice a week. These treatments were repeated weekly for up to three rounds (if mice were still alive). MRI was performed to evaluate spontaneous tumor growth (Fig. 2D). Even though immunotherapy with an anti-PD-1 antibody and PDT with PNP slightly improved the overall survival (Fig. 2C), there was no significant decrease in bladder cancer mass (Fig. 2D). In contrast, in the immunotherapy plus PDT combination group, tumor shrinkage was observed. In one mouse of the combination group that we performed serial MRI to monitor cancer progression, clearance and shrinkage to bladder cancer was observed; the bladder cavity effective volume increased which was filled with urine in the MRI T2 imaging, and that mouse died on Day 158 (Fig. 2D). Compared with the median overall survival of 33.7 days of the control, anti-PD-1 immunotherapy and PDT improved the overall survival by 11 and 19 days to 44.8 (P = 0.0123) and 52.6 days (P = 0.0054), respectively. In contrast, the combination of anti-PD-1 immunotherapy and PDT increased the overall survival by over 40 days (P = 0.0001) and only 1 out of 7 mice died when all mice in the other three groups died and the study ended at Day 75 (Fig. 2C). These results suggest that the PDT and anti-PD-1 immunotherapy synergistically enhanced the antitumor effect and improved the overall survival in vivo in the spontaneous bladder cancer model.

To accurately measure the antitumor activity and determine the abscopal effects, we established bilateral subcutaneous mouse bladder cancer models using the carcinogen-induced BBN963 cells implanted at the flanks (Fig. 3A). When tumor implants were around 100 mm³ bilaterally, mice received PBS, anti-PD-1, PNP, or combination through intravenous injection, followed by light treatment of the tumor located on the right flank. Tumor sizes at the right flank (treated tumor) and left flank (abscopal effects) were both measured and monitored. After intravenous injection, PNP started accumulating at the tumor sites 4 hours after administration and peaked at 24 hours. Starting at 8 hours, no significant PNP accumulation at other sites of the bodies was detected except at the tumors (Fig. 3B).

First, we investigated the direct antitumor activity of the tumors on the right flank, which received light treatment. Anti-PD-1 treatment alone showed moderate antitumor activity. The PNP with PDT treatment significantly delayed tumor growth at the right flank (direct anticancer effects) whereas the PDT with PNP plus anti-PD-1 combination treatment further reduced tumor growth. In 4 of 8 mice of the combination group, no tumor was palpable at the time of study termination and the remaining mice had tumors no bigger than at the study initiation whereas most of the control mice reached the study endpoint at Day 32 (Fig. 3C; Supplementary Fig. S8). At 25 days, mean tumor volume on the right (treated) side were 702.06 ± 135.54 mm³ in control group, 419.51 ± 94.06 mm³ in anti-PD-1 group, 135.34 ± 43.55 mm³ in PNP plus light group, and 48.91 ± 17.86 mm³ in combination treatment group. Compared with the control, the right-site-tumors (treated, direct antitumor effect) of the anti-PD-1, PNP, and combination groups had a 40.25% (P = 0.0003), 80.72% (P < 0.0001), and 93.03% (P < 0.0001) reduction, respectively (Fig. 3C).

Next, we determined the abscopal antitumor activity by monitoring tumor growth of tumors on the left flank that did not receive any laser treatment (Fig. 3D; Supplementary Fig. S8). At 25 days, tumor volume on the left side (untreated, abscopal effects) were 672.02 ± 78.38 mm³ in control group, 364.68 ± 96.01 mm³ in anti-PD-1 group, 302.93 ± 92.36 mm³ in PDT, and 161.58 ± 67.97 mm³ in combination treatment group. Compared with the control, the tumor reduction of the left-site tumors (measuring the abscopal effect) in the anti-PD-1, PDT, and combination groups was 45.73% (P = 0.0003), 54.92% (P < 0.0001), and 75.96% (P < 0.0001), respectively. The tumor reduction between the PDT with PNP and anti-PD-1 therapy groups was not statistically different (P = 0.8402).

To further determine whether PDT and immunotherapy could induce abscopal anticancer immunity, we assessed whether treatment of local MB49 tumors could prevent the formation of distant tumors after injecting cancer cells to the left flank (Fig. 4A). For this experiment, we applied two rounds of treatment (PNP injection on Day 1 and Day 8 followed by PDT 24 hours later. Anti-PD-1 treatment was applied twice a week for the primary tumor site at the right flank. One week later (Day 7), MB49 cells were injected into the left flank to mimic distant metastasis. On day 15, left tumors were harvested and analyzed. The gross image for tumors was shown in Supplementary Fig. S9. For the right (treated) site, tumor weight was 1.83 ± 0.12 g in the control group, 0.72 ± 0.13 g in the anti-PD-1 group, 0.78 ± 0.14 g in the PNP+ PDT group, and 0.21 ± 0.13 g in the combination treatment group (Fig. 4B, left). The combination group had smaller tumors than any other groups. For the left (distant, abscopal) site, tumor weight was 0.2 ± 0.02 g in the control group, 0.07 ± 0.04 g in the anti-PD-1 group, 0.05 ± 0.01 g in the PNP+PDT group, and 0.02 ± 0.001 g in the combination treatment group (Fig. 4B, right). All three treatment groups had smaller tumors than the control (P < 0.05), but were not statistically different among themselves. Single-therapy treatment, anti-PD-1 or PDT with PNP, considerably decreased tumor growth in both primary and secondary tumors. Tumors of the combination group were significantly smaller than the control (P < 0.0001 for the right-treated tumors and P = 0.0013 for the left-untreated abscopal tumors). In the combination treatment group, one mouse showed a rejection at the distant site. These results indicate the combination treatment could significantly reduce distant tumor growth.

Next, we determined whether treatment of a local MB49 tumor can prevent the development of lung metastasis from an intravenous injection of MB49 cancer cells. Because intravenous injection of diluted cancer cells required a longer time to develop lung metastasis, we applied three rounds of treatment (PNP injection on Days 1, 8, and 15, followed by PDT 24 hours later. Anti-PD-1 treatment was applied twice a week) on the right flank. The representative appearance of each group was shown in Fig. 4C and the gross appearance of all lungs in each group is shown in Supplementary Fig. S10. There were many large nodules in control groups; the anti-PD-1 treatment group had a considerably reduced number whereas the PNP+PDT group had numerous but much smaller metastatic nodules. The combination treatment group had the least lung metastatic nodules (Fig. 4D). The quantification of the tumor to total lung areas are shown in Fig. 4E. The tumor/lung ratio was 7.61% in the combination group, compared with 38.69% (P = 0.0383), 28.25% (P = 0. 0492), and 37.96% (P = 0.119) of the control, anti-PD-1, and PNP plus light treatment groups, respectively.

It has been shown that subcutaneous MB49 models could develop spontaneous lung metastasis (35). To confirm, MB49 tumor implants were first established at the right flank of C57BL/6 mice. Three weeks later, PNP was injected into the tail vein. Fluorescence from PNP concentrated at the tumor sites and lungs in a time-dependent manner (Fig. 4F), suggesting that spontaneous lung metastases had developed. To determine how treatments affected spontaneous lung metastasis (Fig. 4G), when MB49 tumor implants reached the size of 100 mm³, C57BL/6 mice were treated with PBS control, anti-PD-1 immunotherapy, PDT, and combination weekly for three rounds as described in Fig. 4G. These mice did not receive intravenous injection of MB49 cells through the tail vein. On Day 22, lungs were collected in all these mice and H&E staining was performed (Fig. 4H). There were metastases in the control and anti-PD-1 treatment groups whereas both PDT and combination treatments almost completely prevented metastasis in the lung (Fig. 4I).

To analyze the immune response induced by PNP with PDT plus anti-PD-1 treatment, we tested immune cell infiltration in different mouse models.

For BBN963 mouse model, we repeated the same treatment strategy as described in Fig. 3A for three rounds of PDT to the tumors at the right flank. Twenty-four hours after last treatment, the untreated tumors at the left flank were harvested for CyTOF. Detailed gating information for CyTOF is shown in Supplementary Fig. S11A. As shown in Fig. 5A, the CD3e⁺TCRb⁺ T cells in the combination treatment group was 2.81% compared with 1.06%, 1.08%, and 0.79% in the control, anti-PD-1, and PDT groups, respectively. Of the T cells in the tumors, CD3e⁺TCRb⁺ CD8⁺ T cells in the combination treatment group was 65.2% of CD3⁺ cells compared with 57.5%, 37.2%, and 64.8% in the control, anti-PD-1, and PDT groups, respectively (Supplementary Fig. S11B). Dendritic cells were also higher in the combination treatment group at 6.62% compared with 3.90%, 4.04%, and 5.25% in the control, anti-PD-1, and PNP PDT groups, respectively (Supplementary Fig. S12). IHC staining confirmed that in the combination treatment group, more CD8⁺ T cells infiltrated into the untreated tumor site (Fig. 5B). Statistical analysis of IHC staining is shown in Supplementary Fig. S13.

The increased CD8⁺ T-cell infiltration led to tumor cell death and low proliferation in the mice that received combination treatment as determined by H&E staining, Ki-67 IHC staining, and TUNEL staining (Supplementary Figs. S14 and S15). IL6, IL12p70, and IL1α/IL1F levels in serum also markedly increased in the combination treatment group (Supplementary Fig. S16).

We also investigated MB49 tumors for tumor immune microenvironment changes. Two weeks after control, anti-PD-1, PDT, and combination treatment of mice carrying bilateral MB49 implants, the contralateral untreated tumors were harvested and stained with CD8. As shown in Fig. 5C, there was no significant CD8⁺ T-cell infiltration in tumors when mice were treated with IgG control, anti-PD-1, and PDT. In contrast, the PDT with PNP plus anti-PD-1 treatment induced CD8⁺ T-cell infiltration in the perivascular areas (P = 0.035 when compared with the control; Supplementary Fig. S17).

Flow cytometry of single-cell suspension from the whole tumors showed that the percentage of CD45⁺ leukocytes in the PDT with PNP plus anti-PD-1 treatment group was 22.60% in the distant tumors of the left flank that were not treated with light (abscopal effects), compared with 8.63% (P = 0.025), 9.80% (P = 0.014), and 9.97% (P = 0.006) in the control, anti-PD-1, and PNP groups, respectively. Meanwhile, the percentage of CD45⁺CD3⁺CD8⁺ cells in the PDT with PNP plus anti-PD-1 treatment group was 1.463% of all cells in tumor, compared with 0.408% (P = 0.0008), 0.826% (P = 0.0351), and 0.598% (P = 0.0231) in the control, anti-PD-1, and PNP PDT groups, respectively (Fig. 5D and E).

We next determined the impact of PNP with PDT on systemic immune response. An IFNγ ELISPOT assay showed spleen cells from the combination treatment group had much higher numbers of cells producing IFNγ (Supplementary Fig. S18). Positive and negative control of spleen cells were shown in Supplementary Fig. S19.

In our previous study, we confirmed bladder cancer–specific PNP could specifically deliver photosensitizer to cancer cells 30 to 40 times higher than to adjacent normal urothelial cells in the same bladder after intravesical instillation (20), compared with only two to three times with other photosensitizers commonly used in clinics during photodynamic diagnosis (16). In addition, PNP is over 100 times more potent than 5-aminolevulic acid (5-ALA) for PDT (20). In this series of experiments, we show that PDT with PNP can not only potentiate immunotherapy in treating locally advanced bladder cancer (laBC), but also in treating distant tumors, inducing abscopal effects and diminishing lung metastasis. This potentiation is mainly mediated through inducing immunogenic cell death, promoting dendritic cell maturation, attracting immune cell infiltration, and creating a pro-immune tumor microenvironment. Our data suggest that PNP has the potential to be developed as a multifunctional theranostic platform, which combines PDD, PDT, and potentiation of immunotherapy in a single platform to be performed in a single procedure to address multiple unmet needs of NMIBC to locally advanced and metastatic bladder cancer.

For NMIBC, some of the major unmet needs are incomplete resection during TUR, which occurs in over one third of patients (36), cancer recurrence in over 70% of patients, and cancer progression to advanced stages in one third of these patients (37). Upon cancer recurrence after BCG, immunotherapy with pembrolizumab is at most moderately effective with a short-term response rate of 40%, but a durable (12 months or longer) response of less than 20%. Most of the patients who failed pembrolizumab undergo radical cystectomy (2). Our previous study already showed the great potential of PNP in identifying cancer and guiding the complete resection of TUR, which can potentially decrease cancer recurrence (20). PDT with PNP is especially suitable for diffuse cancer in situ (CIS), where TUR has a limited role, but, through intravesical instillation, CIS can be directly accessed and treated by PNP and PDT. Here we describe that PDT potentiates immunotherapy. PNP has tremendous potential to be used from upfront diagnosis and treatment of NMIBC to salvage therapy, especially in combination with an ICI.

Some of the major issues associated with laBC are low efficiency of neoadjuvant chemotherapy, highly morbid cystectomy, and high cancer recurrence in 30% to 80% of patients depending on the stage of cancer at the time of cystectomy. Because cystectomy is associated with the worst health-related quality of life, most patients would tend to avoid or delay cystectomy as much as possible. Cancer immunosurveillance achieved from immunotherapy has the potential of inducing long-term remission or even complete cure, thus providing a favorable alternative to cystectomy. Neoadjuvant immunotherapy alone has similar efficacy as neoadjuvant cytotoxic chemotherapy (38, 39). Our studies here showed that the combination of PDT and immunotherapy is highly effective in treating laBC. The median survival for the double-transgenic mice with laBC is 33.7, 44.8, and 52.8 days for the control, anti-PD-1, and photodynamic therapy, respectively, whereas the survival for the combination group is over 75 days. When the study was terminated on Day 75, only one out of seven mice died in the combination group whereas all mice in the other three treatment groups had died. Longitudinal MRI was done in one mouse to determine tumor progression (Fig. 3). That mouse died at Day 158. In addition, this combination therapy can prevent both induced and spontaneous lung metastasis (Fig. 4). With this promising preclinical data, it is rational to develop less toxic PDT together with immunotherapy as an option for neoadjuvant therapy for laBC.

PDT is traditionally considered a local therapy. However, when combined with immunotherapy, it can induce abscopal antitumor activity, treat distant tumors, and decrease both spontaneous and induced lung metastasis (Fig. 4). These suggest that PDT in combination with immunotherapy has the potential to be developed for metastatic bladder cancer. Cytotoxic chemotherapy is the standard of care for newly diagnosed bladder cancer with distant metastasis with immunotherapy used as salvage therapy or for chemotherapy-ineligible patients or those with PD-L1 expression. Targeted therapy and antibody–drug conjugates are other options for advanced bladder cancer (40–44), but resistance always develops. Even though immunotherapy can induce durable response, it, as a single agent, has a disappointing response rate of around 20% (4, 5, 7, 9, 10, 45). The response rate is even lower, at less than 10%, as maintenance therapy after chemotherapy (3). On the basis of our study results, the addition of PDT with PNP to immunotherapy can potentially treat both local cancer and target distant cancers.

Our study showed the mechanisms of potentiation of immunotherapy by PDT is multifactorial. PDT produces ROS that can induce carbonylation of proteins and make them more immunogenic. The transwell study showed that it could induce dendritic cell maturation and upregulation of CD80 and CD86, two molecules that provide costimulation signals during T-cell activation. Analyses of tumor tissues showed that PDT increased immune cell infiltration into tumors. This can have tremendous impact on managing bladder cancer as both The Cancer Genome Atlas (TCGA, 408 cases) and the European Bioinformatics Institute (EBI, 476 cases) databases show over 60% of bladder cancers have no or little immune cell infiltration at the tumor sites (immune “cold” tumors; ref. 46). Furthermore, PDT treatment induces systemic immune response as demonstrated by the increased production of pro-inflammatory systemic cytokines (Supplementary Fig. S12). PDT treatment also enhances the recognition of major histocompatibility complex-I (MHC-I): antigen complexes and tumor-associated antigens by immune cells (28).

One of the major strengths of this project is that we used multiple clinically relevant animal models, which consistently showed the potentiation of immunotherapy by PDT, and that the mechanistic studies supported the findings. One of the potential drawbacks is the lack of human immune system in this study. As immunotherapy research requires a competent immune system, immunocompetent mice are commonly used. We previously reported using humanized mice to study immunotherapy in which immune cells were developed from human CD34 hematopoietic stem cells and cancers from human patient-derived xenografts (47). Because of the defective thymus and lymphoid tissue in immunocompromised mice used to generate humanized mice, the immune system in those humanized mice can be significantly different from human patients and mouse immune systems.

In summary, PDT with bladder cancer–specific PNP can potentiate immunotherapy in treating laBC, induce abscopal antitumor activity for distant cancer, and decrease cancer metastasis. These synergistic antitumor effects are secondary to multiple underlying mechanisms of action, such as modification of macromolecules to activate dendritic cells, increase of immune cell infiltration into tumors, and alteration of tumor immune microenvironment. These findings can have a tremendous impact on the management of bladder cancer, one of the top 10 most common cancers, for which immunotherapy has disappointing efficacy. We have already held a pre-IND consultation with the FDA (PIND No.: 132838). Currently, the IND-enabling pharmacology and toxicology studies of PNP are ongoing, which was supported by Department of Defense. Hence, this study can be translated into a phase I clinical trial in the near future to address multiple unmet needs in bladder cancer.

H. Zhang reports a patent for PLZ4 licensed to LP Therapeutics Inc. T.-Y. Lin reports a patent for porphyrin-based micelle issued and licensed to Lamno Therapeutics Inc. Y. Li reports a patent for US Patent 11,219,692 pending. K.S. Lam reports grants from University of California Davis during the conduct of the study, as well as other support from LP Therapeutics and LamnoTherapeutics outside the submitted work; in addition, K.S. Lam has a patent for PLZ4 issued and licensed to LP Therapeutics and a patent for Micellar nanoplatform issued and licensed to LamnoTherapeutics. C.-x. Pan reports grants from Department of Veterans Affairs and Department of Defense during the conduct of the study; in addition, C.-x. Pan has a patent for PLZ4 issued and licensed to LP Therapeutics Inc., and a patent for Porphyrin-based micelles issued and licensed to LamnoTherapeutics. No disclosures were reported by the other authors.

Z. Zhu: Data curation, formal analysis, validation, writing–original draft. A.-H. Ma: Data curation, formal analysis. H. Zhang: Supervision. T.-Y. Lin: Data curation, supervision. X. Xue: Data curation, methodology. H. Farrukh: Writing–review and editing. S. Zhu: Data curation. W. Shi: Data curation. R. Yuan: Data curation. Z. Cao: Data curation. V.C.S.R. Chittepu: Data curation. R. Prabhala: Data curation, software, visualization. Y. Li: Supervision. K.S. Lam: Conceptualization, supervision. C.-x. Pan: Conceptualization, supervision, funding acquisition, investigation, writing–review and editing.

Thanks to Mingyue Li and Lutian Yao for analyzing flow cytometry data. This project was supported by Merit Review (Award Nos. 1I01 BX003840 and 1I01BX001784, PI: Pan) from the U.S. Department of Veterans Affairs Biomedical Laboratory Research and Development Program, DoD (Award No. W81XWH-19–1-0204 and W81XWH-13-1-0490, PI: Lam and Pan), NIH (Award No. U54CA233306, PI: Pan). The contents do not represent the views of the U.S. Department of Veterans Affairs or the United States Government.

The publication costs of this article were defrayed in part by the payment of publication fees. Therefore, and solely to indicate this fact, this article is hereby marked “advertisement” in accordance with 18 USC section 1734.