Carbonic anhydrase‐9‐targeted near‐infrared photoimmunotherapy as a theranostic modality for clear cell renal cell carcinoma

Abstract

Carbonic anhydrase‐9 (CA9) is highly expressed in clear cell renal cell carcinoma (ccRCC) cells despite no expression in normal kidney tissues. Thus, CA9 has been proposed as a theranostic target for radioligand therapy (RLT). However, ccRCC tends to be radioresistant and may not effectively respond to RLT. Alternatively, CA9 can be targeted for near‐infrared photoimmunotherapy (NIR‐PIT) of ccRCC. Here, we sought to test NIR‐PIT using CA9 in a preclinical model of ccRCC to determine its potential as a therapeutic strategy. Tissue microarray analysis showed that membrane CA9 was expressed in the majority of ccRCC cases. In vitro, CA9‐targeted NIR‐PIT induced cell membrane damage and cell killing in all CA9‐expressing ccRCC cell lines specifically, UOK154, UOK220, and UOK122. In vivo, CA9‐targeted NIR‐PIT significantly inhibited tumor growth and prolonged survival in UOK154 and UOK220 subcutaneous xenograft models. Notably, 70%–80% of mice achieved complete remission after a single treatment of NIR‐PIT. Additionally, remaining tumors after the first NIR‐PIT persistently expressed CA9, suggesting that remaining tumors can be treated with repeated NIR‐PIT. Furthermore, CA9‐targeted NIR‐PIT induced significant cytoplasmic damages on ccRCC cells in UOK154 orthotopic xenograft models. In conclusion, CA9‐targeted NIR‐PIT, which allow for safe and repeated application on the same lesion, is a promising treatment for ccRCC, especially in the management of multiple primary ccRCC (e.g., von Hippel–Lindau syndrome) and oligometastatic ccRCC.

What's New?

The cell‐surface glycoprotein carbonic anhydrase‐9 (CA9) is highly expressed in clear cell renal cell carcinoma (ccRCC), suggesting that it may be leveraged as a therapeutic target. A promising and minimally invasive approach for CA9‐targeting is near‐infrared photoimmunotherapy (NIR‐PIT). Here, a novel CA9‐targeting strategy using NIR‐PIT was tested in ccRCC cell lines and xenograft models. CA9‐targeted NIR‐PIT induced cell killing in CA9‐expressing ccRCC cells in vitro and blocked tumor growth and increased survival in vivo. The findings demonstrate the clinical potential for NIR‐PIT directed against CA9 and suggest that its specific cell‐killing effects can potentially facilitate renal preservation in ccRCC.

1. INTRODUCTION

Kidney cancer is the 6th most common cancer among men and the 9th most common cancer among women in the US, comprising an estimated 81,610 new cases and 14,390 deaths in 2024. ¹ Renal cell carcinoma (RCC) constitutes 90%–95% of all kidney cancer cases. At diagnosis, nearly one‐third of patients present with metastatic disease, and an additional 20%–25% develop recurrent or metastatic lesions post‐curative surgery. ² Whereas metastatic RCC is treated with systemic treatment, cytoreductive nephrectomy, and metastasis‐directed therapy are considered therapeutic options in patients with oligometastatic disease. ³ Complete local treatment can yield a survival benefit while avoiding the adverse events of systemic treatment. ⁴ , ⁵ , ⁶ However, surgery and tumor ablation can be limited by lesion location.

Recently, theranostics, which combines diagnostics with targeted therapies to deliver personalized and effective treatment strategies, ⁷ , ⁸ has emerged as a promising approach in the identification and management of oligometastatic RCC. ⁹ , ¹⁰ Clear cell RCC (ccRCC), the predominant histological subtype of RCC, highly expresses carbonic anhydrase‐9 (CA9), a cell‐surface glycoprotein regulated in part by hypoxia‐inducible factor (HIF)‐1α. ¹¹ , ¹² , ^{13 89}Zr‐labeled girentuximab, an anti‐human CA9 monoclonal antibody (mAb), has been used in combination with positron emission tomography/computed tomography (PET/CT) to identify sites of CA9‐positive disease. ¹⁴ , ¹⁵ Although radioligand therapy (RLT) is the most common type of theranostics, ⁸ the radioresistant nature of ccRCC may hamper the application of RLT to ccRCC. ¹⁶ , ¹⁷ Therefore, alternative methods that exploit CA9‐expression may have a therapeutic role in the management of primary and oligometastatic ccRCC.

Near‐infrared photoimmunotherapy (NIR‐PIT) is a new, minimally‐invasive, anti‐cancer therapy that selectively eradicates cancer cells through NIR‐light‐triggered photochemical reactions within antibody‐photoabsorber conjugates (APCs). ¹⁸ , ¹⁹ To synthesize APCs, mAbs that target cancer‐specific antigens on the cancer cell membrane are conjugated with the photoabsorber IRDye700DX (IR700), a silica‐phthalocyanine dye. ²⁰ Upon intravenous administration, APCs bind to cancer cells. NIR light irradiation of the tumor alters the IR700 properties from highly hydrophilic to highly hydrophobic by inducing the release of water‐soluble axial ligands. ²¹ This photochemical transformation leads to the aggregation of APCs, inducing significant damage to the cancer cell membrane while sparing adjacent normal tissues. ²¹ In head and neck cancers, NIR‐PIT targeting epidermal growth factor receptor (EGFR) has been approved for clinical use in Japan and is currently being evaluated in a Phase III clinical trial in the US. NIR‐PIT could be applied to many tumor types depending on the available antibodies. In this study, we developed CA9‐targeted NIR‐PIT and evaluated it in ccRCC xenograft mouse models.

2. METHODS

2.1. Patient cohort of tissue microarray analysis

A tissue microarray (TMA) of primary lesions of human ccRCC (serial no. BC07014b) was purchased from Biomax (Rockville, MD, USA). Cores of 63 ccRCC cases were analyzed by multiplex immunohistochemistry (IHC). Baseline demographics of ccRCC cases are shown in Table S1.

2.2. Conjugation of anti‐CA9 mAb with IR700

To synthesize anti‐CA9 mAb‐IR700 conjugates, 1 mg of girentuximab (6.7 nM; MedChemExpress, Monmouth Junction, NJ, USA), an anti‐human CA9 mAb, was incubated with five‐fold molar excess of IR700 NHS ester (10 mM in DMSO; LI‐COR Biosciences, Lincoln, NE, USA) in 100 mM Na₂HPO₄ solution (pH 8.5) for 1 h at room temperature. The mixture was purified with PD‐10 columns containing Sephadex G25 resin (Cytiva, Marlborough, MA, USA). The resulting APC was abbreviated as CA9‐IR700. Ultraviolet–visible spectroscopy (8453 Value System; Agilent Technologies, Santa Clara, CA, USA) was utilized to calculate the number of IR700 molecules bound to each mAb by measuring the absorbance at 280 and 689 nm. The number of IR700 per mAb was calculated to be approximately four. The quality of CA9‐IR700 was analyzed by sodium dodecyl sulfate‐polyacrylamide gel electrophoresis (SDS‐PAGE). In SDS‐PAGE, CA9‐IR700 was electrophoresed in a 4%–12% gradient polyacrylamide gel (Life Technologies, Gaithersburg, MD, USA). Unconjugated mAb was used as a control. After electrophoresis at 80 V for 2.5 h, the gel was imaged with the 700‐nm fluorescence of a PRISM in vivo imaging system (MediLumine, Montreal, Canada). The gel was then colored with Colloidal Blue staining to compare the molecular weight between the CA9‐IR700 and unconjugated mAb. In SDS‐PAGE analysis, CA9‐IR700 and unconjugated mAb had the same approximate molecular weight but 700‐nm fluorescent signal was detected only in CA9‐IR700 (Figure S1).

2.3. Cell culture

The human ccRCC cell lines, UOK154 (RRID: CVCL_B127), UOK220, and UOK122 (RRID: CVCL_B102), were utilized in this study. All cell lines were established in the Urologic Oncology Branch, National Cancer Institute, NIH, Bethesda, MD, USA. Cells were cultured in Dulbecco's Modified Eagle Medium (ATCC, Manassas, VA, USA) supplemented with 10% fetal bovine serum (Thermo Fisher Scientific, Rockford, IL, USA) and 100 IU/mL penicillin/streptomycin (Thermo Fisher Scientific). All cells were incubated in a humidified incubator at 37°C with an atmosphere of 95% air and 5% carbon dioxide. All cell lines were authenticated using short tandem repeat (STR) profiling within the last three years. All experiments were performed with mycoplasma‐free cells.

2.4. In vitro CA9 expression analysis

To assess in vitro CA9 expression on the cell surface of UOK154, UOK220, and UOK122 cells, 2 × 10⁵ cells were incubated with PE‐labeled anti‐human CA9 antibody (clone 303123; R&D Systems, Inc., Minneapolis, MN, USA) or its PE‐labeled mouse IgG2A isotype control (clone 20102; R&D Systems) as well as Fixable Viability Dye (Thermo Fisher Scientific) for 30 min at 4°C. After incubation, cells were analyzed by BD FACSLyric (BD Biosciences, San Jose, CA, USA) using FlowJo software (FlowJo LLC, Ashland, OR, USA).

2.5. In vitro cell binding assay

For in vitro CA9‐IR700 cell binding assay, cells (2 × 10⁵) were incubated in a medium containing CA9‐IR700 (10 μg/mL) for 30 min at 4°C. IR700 was detected by BD FACSLyric (BD Biosciences) and analyzed by FlowJo software (FlowJo LLC). To validate the specific binding of CA9‐IR700, 10‐fold molar excess of unconjugated girentuximab (MedChemExpress) was added 30 min before the incubation with CA9‐IR700. Dead cells were excluded from the analysis through staining with Fixable Viability Dye (Thermo Fisher Scientific).

2.6. In vitro fluorescence microscopy

Cells were seeded at 1 × 10⁵ on a 35‐mm dish. Twenty‐four hours later, cells were incubated in medium containing CA9‐IR700 (10 μg/mL) for 1 h at 37°C. Cells were stained with 10 μg/mL propidium iodide (PI; Life Technologies), followed by observation with Leica DM IL LED Inverted Laboratory Microscope (Leica Biosystems, Wetzlar, Germany). Transmitted‐light differential interference contrast (DIC) and PI fluorescence images were acquired prior to NIR light irradiation. Subsequently, cells were exposed to NIR light (690 nm, 150 mW/cm², 50 J/cm²) using an ML7710 laser system (Modulight, Tampere, Finland). DIC and PI fluorescence images were acquired immediately and 20 min after NIR light irradiation.

2.7. In vitro NIR‐PIT

Cells were seeded at 1 × 10⁵ onto 24‐well plates and incubated for 24 h at 37°C. Cells were next incubated in medium containing CA9‐IR700 (10 μg/mL) for 1 h at 37°C. The medium was changed into a phenol‐red‐free medium, followed by NIR light irradiation using an ML7710 laser system (690 nm, 150 mW/cm², Modulight, Tampere, Finland). Cell viability was assessed using 3‐(4,5‐dimethyl‐2‐thiazolyl)‐2,5‐diphenyl‐2 H‐tetrazolium bromide (MTT) assay and PI flow‐cytometric assay 1 h after NIR‐PIT. MTT assay was performed using a microplate reader (Synergy H1; BioTek, Winooski, VT, USA) according to the manufacturer's protocol (SIGMA Aldrich, St. Louis, MO, USA). For PI flow‐cytometric assay, cells were harvested and stained with 1 μg/mL PI (Life Technologies). PI‐stained cells were measured using BD FACSLyric (BD Biosciences) and FlowJo software (FlowJo LLC). Moreover, cell surface expression of calreticulin and heat shock protein‐70 (HSP70) was examined immediately after NIR‐PIT. Cells were incubated for 30 min at 4°C with anti‐mouse calreticulin antibody (rabbit poly; Bioss Antibodies, Woburn, MA, USA) or its rabbit IgG1 isotype control (Bioss Antibodies) for calreticulin staining and with anti‐mouse/human HSP70 antibody (clone REA349; Miltenyi Biotec, Gaithersburg, MD, USA) or its human IgG1 isotype control (clone REA293; Miltenyi Biotec) for HSP70 staining. The fluorescence of cells was measured using BD FACSLyric (BD Biosciences) and FlowJo software (FlowJo LLC).

2.8. Animal models

Six‐ to eight‐week‐old female homozygote athymic nude mice were purchased from Charles River Laboratories (Wilmington, MA, USA). For subcutaneous xenograft models, 5 × 10⁶ UOK154, UOK220, or UOK122 cells (100 μL of 100% Matrigel) were inoculated into the right dorsum. Tumor volumes were measured three times per week by caliper and calculated as follows: tumor volume (mm³) = length × width² × 0.5. Mice with tumors reaching approximately 100–200 mm³ in volume were randomized into groups using Tumorimager and TumorManager software (Biopticon, Princeton, NJ, USA). The mice were euthanized with carbon dioxide once the tumor volume reached 500 mm³. Surviving mice were censored 75 days post‐tumor inoculation. Complete remission (CR) was defined as tumor disappearance for four weeks or longer after NIR‐PIT. Orthotopic xenograft models were developed as follows. A 1 cm incision in the left flank of the mouse was made. After the left kidney was located through the peritoneum, 3 × 10⁵ UOK154 or UOK122 cells (in 50 μL of 100% Matrigel) were injected through the intact peritoneum into the center of the left kidney using a 1‐mL syringe and a 30‐gauge needle, followed by skin closure. ²² During the experiments, mice were anesthetized with inhaled 2%–3% isoflurane, with an intraperitoneal injection of sodium pentobarbital (37.5 mg/kg; Nembutal Sodium Solution, Ovation Pharmaceuticals Inc., Deerfield, IL, USA), or with an intraperitoneal injection of ketamine (80–100 mg/kg, covetrus, Portland, MD, USA)/xylazine (5–10 mg/kg, Rompun xylazine injection, Dechra, Overland Park, KS, USA) in addition to a subcutaneous injection of Ethiqa XR (3.25 mg/kg, Fidelis Animal Health, North Brunswick, NJ, USA).

2.9. In vivo fluorescence imaging

Mice with subcutaneous tumors in the right dorsum were injected with CA9‐IR700 (100 μg) via lateral tail vein. Dorsal fluorescence images were serially obtained with the 700‐nm fluorescence channel of a PRISM in vivo imaging system (MediLumine). The images were analyzed with Image J (NIH, Bethesda, MD, USA). Regions of interest (ROIs) were depicted on the tumor and the non‐tumoral region of the contralateral side. Target‐to‐background ratio (TBR) was calculated as mean fluorescence intensity (MFI) of the tumor/MFI of the non‐tumoral region of the contralateral side.

2.10. In vivo NIR‐PIT

For subcutaneous xenograft models, mice were randomized into three groups as follows: (1) no treatment (Control), (2) intravenous injection of CA9‐IR700 only (APC‐IV), and (3) intravenous injection of CA9‐IR700 followed by NIR light irradiation (NIR‐PIT). CA9‐IR700 (100 μg) was injected 13 d after tumor inoculation. Twenty‐four hours later, NIR light (690 nm, 150 mW/cm², 50 J/cm²) was transcutaneously irradiated into the tumor via a frontal light diffuser. When irradiating NIR light, mice were shielded by aluminum foil with a hole created so as to irradiate only the target tumor. For orthotopic xenograft models, mice were classified into two groups as follows: (1) no treatment (Control) and (2) intravenous injection of CA9‐IR700 followed by NIR light irradiation (NIR‐PIT). CA9‐IR700 (100 μg) was injected, and approximately 24 h later, a cylindrical light diffuser was subcutaneously inserted in the left dorsum, and NIR light (690 nm, 400 mW/cm, 100 J/cm) was applied to the left kidney. 700‐nm fluorescence and white light images were obtained before and after NIR‐PIT using a PRISM in vivo imaging system (MediLumine). A ROI was placed on the tumor and mean fluorescence intensity was calculated for each ROI.

2.11. Histological analysis

Mice with subcutaneous UOK154 tumors were euthanized 24 h post‐NIR light irradiation. Also, mice with orthotopic tumors in the left kidney were euthanized two and 24 h after NIR light irradiation. The tumor and left kidney were harvested, and subsequently formalin‐fixed, paraffin‐embedded (FFPE) sections were prepared and stained with hematoxylin and eosin (HE) or periodic acid‐Schiff (PAS) staining.

2.12. Multiplex IHC

Sections of TMA and FFPE specimens were utilized for multiplex IHC. Multiplex IHC was conducted as described previously, ²³ , ²⁴ using Opal Automation IHC Kit (Akoya Bioscience, Menlo Park, CA, USA) and Bond RXm autostainer (Leica Biosystems). The sections were stained with 4,6‐diamidino‐2‐phenylindole (DAPI) and the following antibodies: anti‐β‐actin (rabbit poly; 1:500; Abcam, Cambridge, UK), anti‐human CD45 (clone EP322Y; 1:500; Abcam), anti‐mouse CD45 (clone D3F8Q; 1:500; Cell Signaling Technology, Danvers, MA, USA), anti‐lactate dehydrogenase A (LDHA; clone C4B5; 1:500; Cell Signaling Technology), anti‐pan‐cytokeratin (pCK; rabbit poly; 1:250; Bioss Antibodies), anti‐digoxigenin (DIG; clone 9H27L19; 1:500; Thermo Fisher Scientific), and anti‐CA9 (rabbit poly; 1:500; Abcam). Coverslips were applied to the slide using ProLong Diamond Antifade Mountant (Thermo Fisher Scientific). Stained slides were analyzed with Mantra Quantitative Pathology Workstation (Akoya Biosystems) and inForm Tissue Finder software (Akoya Biosystems). To assess CA9 expression in cancer cells, inForm software was trained to detect tissue and cell phenotypes using machine‐learning algorithms based on the following criteria: areas with pCK expression = tumor, other areas = stroma, pCK⁺CD45⁻ cells = cancer cells, pCK⁻CD45⁺ = blood cells, and pCK⁻CD45⁻ = other cells, respectively. inForm software computed H‐scores based on membrane CA9 staining in cancer cells. The average H‐score was calculated from three images for each case, then specimens were classified into four groups: negative (H‐score 0–14), weak (H‐score 15–99), moderate (H‐score 100–199), and strong (H‐score 200–300), as previously described. ²⁵

2.13. Detection of DIG‐labeled mAb

Girentuximab (MedChemExpress) was labeled with DIG by incubating 1 mg of mAb and 50 μg of DIG Succinimidyl Ester (Thermo Fisher Scientific), in a similar manner to IR700 conjugation. The resulting DIG‐labeled mAb was abbreviated as CA9‐DIG. Tumor‐bearing mice were intravenously injected with CA9‐DIG (100 μg) and, 24 h later, tumors were harvested. CA9‐DIG distribution was evaluated in FFPE sections by multiplex IHC using anti‐DIG antibody (clone 9H27L19; Thermo Fisher Scientific).

2.14. Blood urea nitrogen assay

Peripheral blood samples were collected from mice with orthotopic tumors in the left kidney at indicated time points, and plasma was prepared. Blood urea nitrogen (BUN) was measured using BUN colorimetric detection kit (Thermo Fisher Scientific) according to the manufacturer's protocol.

2.15. Statistical analysis

Continuous data were compared using a one‐way analysis of variance (ANOVA) followed by Tukey's test among multiple groups. Categorical data were compared using chi‐square test. Fluorescence intensity, tumor volumes, and BUN were compared using repeated measures two‐way ANOVA followed by Tukey's or Sidak's test. Survival curves were illustrated using the Kaplan–Meier method, and the results were compared by the log‐rank test. When compared among groups of more than two, Bonferroni correction was performed. GraphPad Prism 10 (GraphPad Software, La Jolla, CA, USA) was utilized for statistical analysis. p < .05 was defined as statistical significance.

3. RESULTS

3.1. Validation of high expression of membrane CA9 in human ccRCC tissues

We evaluated membrane CA9 expression in the TMA of 63 ccRCC cases by multiplex IHC. Their baseline characteristics are shown in Table S1. All cases were categorized into four groups according to CA9 H‐scores (Figure 1A,B). Membrane CA9 expression was positive in 86% of ccRCC cases (Figure 1C), validating high expression of membrane CA9 in ccRCC. High CA9 expression was significantly associated with lower T stage (p < .0001) and lower grade (p < .0001).

FIGURE 1.

Immunohistochemical CA9 expression in human ccRCC tissue. (A–C) Immunohistochemical evaluation of membrane CA9 expression in human TMA specimens of ccRCC. CA9 H‐score was calculated based on membrane CA9 expression. CA9 staining was classified into one of four categories: Negative (H‐score: 0–14), weak (H‐score: 15–99), moderate (H‐score: 100–199), and strong (H‐score: 200–300). (A) Representative images of human ccRCC tissue are shown (images; ×200; scale bar, 20 μm). Antibody staining of CA9 is shown in brown. Nuclei are stained with DAPI and shown in blue. (B) Distribution of CA9 H‐score. Median and quartiles are shown as solid and dashed lines, respectively. (C) CA9 staining category.

3.2. Cell surface CA9 expression and CA9‐IR700 binding in human ccRCC cell lines

Cell surface CA9 expression was examined in vitro using three human ccRCC cell lines, UOK154, UOK220, and UOK122. Flow‐cytometric analyses showed positive CA9 expression in all three (Figure 2A). Positive CA9 expression was also confirmed in vivo (Figure S2). Furthermore, when CA9‐IR700 was incubated with the three cell lines, all of them showed fluorescent signals of IR700, representing the binding of CA9‐IR700 to cells (Figure 2B). These signals were completely lost by adding an excess of unconjugated girentuximab (Figure 2B). Thus, CA9‐IR700 specifically bound to CA9‐expressing ccRCC cells in vitro.

FIGURE 2.

In vitro CA9 expression and efficacy of in vitro CA9‐targeted NIR‐PIT in human ccRCC cell lines. (A) Flow‐cytometric analysis of in vitro CA9 expression on the cell surface of human ccRCC cell lines. (B) Detection of CA9‐IR700 bound to each cell line in vitro by flow‐cytometric analysis. Ten‐fold molar excess of unconjugated anti‐CA9 antibody was used to validate the specific binding of CA9‐IR700. (C) Microscopic observation of UOK154 cells before and after in vitro CA9‐targeted NIR‐PIT (images, ×100; scale bar, 20 μm). Immediately after NIR‐PIT, cells showed ballooning, bleb formation (white‐filled arrowhead), and propidium iodide (PI) positivity. DIC, differential interference contrast. (D) Cell viability of UOK154 cells after in vitro CA9‐targeted NIR‐PIT measured by MTT assay (n = 5; mean ± SEM; one‐way ANOVA followed by Tukey's test). The value of absorbance in the treated group was normalized to the untreated control. ***, p < .001; ****, p < .0001 versus the untreated control. (E) Cell membrane damage of UOK154 cells induced by in vitro CA9‐targeted NIR‐PIT was assessed with the dead cell count using PI staining (n = 5; mean ± SEM; one‐way ANOVA followed by Tukey's test). ****, p < .0001 versus untreated control. (F) and (G) Cell surface expression of calreticulin (F) and heat shock protein (HSP) 70 (G) in UOK154 cells after in vitro CA9‐targeted NIR‐PIT was measured by flow cytometry (n = 5; mean ± SEM; one‐way ANOVA followed by Tukey's test. RFI was calculated as the ratio of the median fluorescence intensity of the anti‐calreticulin or HSP 70 antibody to that of the isotype control. ****, p < .0001.

3.3. Cytotoxic effects of CA9‐targeted NIR‐PIT in vitro

We evaluated the cell‐killing efficacy of CA9‐targeted NIR‐PIT in vitro using CA9‐expressing human ccRCC cell lines. First, cell morphological changes were examined following CA9‐targeted NIR‐PIT in vitro. Microscopically, UOK154 cells showed cellular swelling and bleb formation within 20 min after NIR‐PIT (Figure 2C). Additionally, PI staining gradually increased in the nuclei (Figure 2C). Similar results were obtained in UOK220 and UOK122 cells (Figure S3A), indicative of cell membrane damage‐based cytotoxicity due to CA9‐targeted NIR‐PIT. Next, MTT and PI flow‐cytometric assays were performed after CA9‐targeted NIR‐PIT in vitro. Cell viability by MTT was significantly decreased after CA9‐targeted NIR‐PIT in a light dose‐dependent manner using UOK154 (Figure 2D), UOK220, and UOK122 (Figure S3B) cells. In PI flow‐cytometric assay, the percent of PI‐stained cells was significantly increased after CA9‐targeted NIR‐PIT in a light dose‐dependent manner using UOK154 (Figure 2E), UOK220, and UOK122 (Figure S3C) cells. These results corroborate the cytotoxic mechanism of CA9‐targeted NIR‐PIT based on cell membrane damage. Additionally, the expression of damage‐associated molecular patterns (DAMPs) was evaluated immediately after CA9‐targeted NIR‐PIT. Cell surface calreticulin and HSP70 expressions were significantly upregulated in UOK154 (Figure 2F,G), UOK220, and UOK122 (Figure S3D,E) cells. This suggests that CA9‐targeted NIR‐PIT can induce immunogenic cell death of ccRCC cells.

3.4. In vivo fluorescence imaging after intravenous administration of CA9‐IR700

Fluorescence imaging of CA9‐IR700 was serially performed in vivo after its intravenous administration using UOK154, UOK220, and UOK122 tumor‐bearing mice (Figure S4). In all models, the fluorescence intensity of CA9‐IR700 at the tumor site peaked 12 h after its injection and then decreased gradually over the following days. TBR of CA9‐IR700 increased up to 12 h after injection and was maintained thereafter in all models. These results allow NIR light irradiation at approximately 24 h after injecting CA9‐IR700, which is the treatment schedule used in clinical practice of NIR‐PIT.

3.5. In vivo therapeutic efficacy of CA9‐targeted NIR‐PIT in subcutaneously inoculated ccRCC tumor models

Therapeutic efficacy of in vivo CA9‐targeted NIR‐PIT was evaluated using subcutaneously inoculated UOK154 xenograft models. Treatment schedule is shown in Figure 3A. Tumor fluorescence was clearly detected before NIR light irradiation. The first NIR light irradiation immediately decreased this signal, followed by a further decline of fluorescence after the second irradiation (Figure 3B,C), due to photobleaching by NIR light irradiation. NIR‐PIT significantly suppressed tumor growth compared to the Control and APC‐IV groups (Figure 3D). Of note, 70% of mice achieved CR in the NIR‐PIT group (Figure 3E). Additionally, NIR‐PIT significantly prolonged survival compared to the Control and APC‐IV groups (Figure 3F). Moreover, NIR‐PIT induced cytoplasmic vacuolation in most cancer cells (Figure S5), which is a unique histological change in cancer cells after NIR‐PIT. ²⁵ , ²⁶ We further performed CA9‐targeted NIR‐PIT in vivo using subcutaneously inoculated UOK220 xenograft models (Figure S6). Akin to the results in UOK154 models, tumor growth was significantly inhibited in the NIR‐PIT group compared to the Control and APC‐IV groups (Figure 3G). CR was observed in 80% of mice in the NIR‐PIT group (Figure 3H). The NIR‐PIT group had significantly longer survival compared to the Control and APC‐IV groups (Figure 3I). Therefore, CA9‐targeted NIR‐PIT was highly effective in vivo in CA9‐expressing ccRCC tumor models.

FIGURE 3.

Efficacy of in vivo CA9‐targeted NIR‐PIT in ccRCC subcutaneous xenograft models. (A) Treatment schedule. (B) Diagram of NIR light irradiation and representative 700‐nm fluorescence images before and after NIR‐PIT. Subcutaneous UOK154 tumors were created at the right dorsum. The orange circle represents where NIR light was irradiated with a frontal light diffuser. A.U., arbitrary units. (C) Changes in 700‐nm fluorescence intensity at the tumor site before and after NIR‐PIT using UOK154 tumors (n = 9–10; mean ± SEM; repeated measures two‐way ANOVA followed by Sidak's test); ****, p < .0001. (D) Tumor volume curves for UOK154 tumors (n = 9–10; mean ± SEM; repeated measures two‐way repeated measures ANOVA followed by Tukey's test); ****, p < .0001 versus the Control group. (E) Tumor volume curves of individual mice in each group with UOK154 tumors. CR, complete remission. (F) Survival curves for UOK154 tumors (n = 9–10; log‐rank test with Bonferroni correction); **, p < .01; ns, not significant. (G) Tumor volume curves for UOK220 tumors (n = 9–10; mean ± SEM; repeated measures two‐way repeated measures ANOVA followed by Tukey's test); ****, p < .0001 versus the Control group. (H) Tumor volume curves of individual mice with UOK220 tumors. CR, complete remission. (I) Survival curves for UOK220 tumors (n = 9–10; log‐rank test with Bonferroni correction); *, p < .05; **, p < .01; ns, not significant.

3.6. Persistent CA9 expression in remaining tumors after CA9‐targeted NIR‐PIT

To evaluate whether ccRCC tumors can be repeatedly treated with CA9‐targeted NIR‐PIT, we examined CA9 expression in three UOK154 and two UOK220 tumors remaining after initial CA9‐targeted NIR‐PIT (Figure 3). As shown in Figure 4A, tumors were harvested (Figure S7) and analyzed by multiplex IHC 75 days after tumor inoculation (60 days after NIR light irradiation). Moderate CA9 expression was seen in the three remaining UOK154 tumors (Figure 4B). For the two remaining UOK220 tumors, one showed moderate CA9 expression, and the other showed weak expression (Figure 4C). These results suggest that the remaining tumors after the first CA9‐targeted NIR‐PIT can be treated with repeated CA9‐targeted NIR‐PIT.

FIGURE 4.

Immunohistochemical CA9 expression in remaining tumors after in vivo CA9‐targeted NIR‐PIT using ccRCC subcutaneous xenograft models. (A) Treatment schedule. Immunohistochemical assessment was performed 75 d after tumor inoculation. (B) CA9 expression in remaining UOK154 tumors after in vivo CA9‐targeted NIR‐PIT (n = 3; images; ×200; scale bar, 20 μm). CA9 staining category was determined according to CA9 H‐score. (C) CA9 expression in remaining UOK220 tumors after in vivo CA9‐targeted NIR‐PIT (n = 2; images; ×200; scale bar, 20 μm). CA9 staining category was determined according to CA9 H‐score.

3.7. CA9‐targeted NIR‐PIT in ccRCC orthotopic xenograft models

We developed a UOK154 orthotopic xenograft model to preclinically simulate CA9‐targeted NIR‐PIT against primary ccRCC lesions. Figure 5A shows the treatment schedule. Seven hundred nanometers fluorescence was not clearly detected in the kidney before NIR light irradiation because it is deeply located beneath the skin and the light does not penetrate (Figure 5B). NIR light was applied to the tumors using a cylindrical diffuser (Figure 5B). Tumors were harvested and histologically analyzed two and 24 h after CA9‐targeted NIR‐PIT. Twenty‐four hours after NIR‐PIT, most cancer cells showed cytoplasmic vacuolation (Figure 5C). This histological change was observed in some of cancer cells 2 h after NIR‐PIT (Figure 5C). Similarly, cytoplasmic vacuolation was also found in UOK122 orthotopic xenograft models treated with CA9‐targeted NIR‐PIT (Figure S8). Moreover, leakage of LDHA into extracellular space after NIR‐PIT was detected by multiplex IHC (Figure 5D), suggestive of necrotic cell death. ²⁵ Additionally, CA9‐targeted NIR‐PIT induced aggregated actin cytoskeleton in most cancer cells, in contrast to their uniform distribution across the entire cytoplasm of cancer cells in the control (Figure 5E). ²⁶ Taken together, these results indicate that CA9‐targeted NIR‐PIT caused significant histological damage to cancer cells in ccRCC orthotopic xenograft models.

FIGURE 5.

Histological evaluation after in vivo CA9‐targeted NIR‐PIT in a UOK154 orthotopic xenograft model. (A) Treatment schedule. (B) Diagram of NIR light irradiation and representative 700‐nm fluorescence images before and after NIR‐PIT. Orthotopic UOK154 tumors were created in the left kidney. The orange rectangle represents where NIR light was irradiated with a cylindrical light diffuser. A.U., arbitrary units. (C) HE staining of the tumor 2 and 24 h after NIR‐PIT (images; ×100 and ×400; scale bar, 20 μm). The inset displayed on the right side shows a representative image of cytoplasmic vacuolation (black‐filled arrowhead). (D) Immunohistochemical evaluation of lactate dehydrogenase‐A (LDHA) expression in UOK154 orthotopic tumors 2 h after NIR‐PIT. Representative pictures of LDHA expression show LDHA leakage into the extracellular space, which suggests necrotic cell death (white‐filled arrowhead; images; ×400; scale bar, 20 μm). Antibody staining of LDHA and CA9 is shown in orange and cyan, respectively. Nuclei are stained with DAPI and shown in white. (E) Immunohistochemical evaluation of actin cytoskeleton in UOK154 orthotopic tumors 24 h after NIR‐PIT. Representative pictures of β‐actin expression (images; ×400; scale bar, 20 μm). Antibody staining of β‐actin is shown in magenta. Nuclei are stained with DAPI and shown in white.

3.8. Effect of CA9‐targeted NIR‐PIT on normal renal parenchyma

We evaluated whether CA9‐targeted NIR‐PIT could cause cytotoxic effects in the normal parenchyma of kidneys implanted with UOK154 orthotopic xenografts. The BUN assay showed no significant difference between pre‐ and post‐NIR‐PIT BUN levels (Figure S9A), indicating no significant damage to kidney function after CA9‐targeted NIR‐PIT. Histological changes in normal parenchyma of the irradiated kidney were examined after CA9‐targeted NIR‐PIT. Although there were no histological changes in Bowman's capsules, cytoplasmic vacuolation was observed in some proximal tubules near the tumor (Figure S9B). Meanwhile, such damages were not observed in proximal tubules far from the tumor (Figure S9C). Thus, proximal tubule injury was likely to have occurred only in the area where NIR light was irradiated. To clarify why NIR light irradiation induced cytotoxic effects on proximal tubules, the distribution of CA9‐DIG, DIG‐labeled girentuximab, in the kidney was evaluated after its intravenous administration (Figure S9D). CA9 expression was detected in the cell surface of cancer cells. Non‐specific CA9 expression was also detected in the cytoplasm of renal tubules. In addition to the cancer cell surface, CA9‐DIG was also distributed in stromal tissues surrounding proximal tubules which do not have CA9 expression. However, CA9‐DIG was not detected on the cell surface of proximal tubules. These results suggest that NIR light activation of non‐specifically accumulated CA9‐IR700 in stromal tissues may lead to minor proximal tubule injury.

4. DISCUSSION

In this study, we developed CA9‐targeted NIR‐PIT as a treatment for human ccRCC expressing CA9. Up to 86% of human ccRCC TMA specimens demonstrate CA9 expression. CA9‐targeted NIR‐PIT was effective in all CA9‐expressing ccRCC cell lines in vitro. CA9‐targeted NIR‐PIT significantly inhibited tumor growth and improved survival in ccRCC subcutaneous xenograft models. Notably, 70%–80% of mice successfully achieved CR after only a single treatment with NIR‐PIT. Given persistent CA9 expression in remaining cancer cells after the first NIR‐PIT, repeated NIR‐PIT may improve the overall treatment efficacy. Furthermore, the effectiveness of CA9‐targeted NIR‐PIT was histologically verified using ccRCC orthotopic xenograft models. Considering the minimal cytotoxicity of NIR‐PIT in normal tissues, CA9 targeted NIR‐PIT offers a safe and repeatable treatment for renal lesions. Theoretically, NIR light can be applied to tumors anywhere in the body amenable to needle placement of cylindrical light diffusers. ²⁷ Taken together, this study paves the way for employing CA9‐targeted NIR‐PIT as a potential therapy for ccRCC.

In prostate cancer, prostate‐specific membrane antigen (PSMA)‐targeted RLT has been established as a theranostic treatment. ⁸ PSMA‐targeted RLT is preceded by PSMA PET/CT for patient selection and monitoring of therapeutic efficacy. ⁶⁸Ga‐PSMA PET/CT is useful in monitoring the therapeutic response to ¹⁷⁷Lu‐PSMA RLT in metastatic prostate cancer. ²⁸ For ccRCC, a recent paper has reported a diagnostic role of ⁸⁹Zr‐labeled girentuximab PET/CT. ¹⁴ The combination of ⁸⁹Zr‐labeled girentuximab PET/CT and conventional CT detected a significantly higher number of lesions than CT alone (91% vs. 56%). ¹⁴ Therefore, ⁸⁹Zr‐labeled girentuximab PET/CT may be a useful diagnostic tool for selecting patients for treatment and monitoring the therapeutic effect of CA9‐targeted NIR‐PIT. In addition, inactivation of the von Hippel–Lindau (VHL) tumor suppressor gene causes the accumulation of HIF‐1α in ccRCC. ²⁹ Because HIF‐1α regulates CA9 expression, HIF‐1α could be used as a biomarker to select patients and predict the efficacy of CA9‐targeted NIR‐PIT.

CA9‐targeted NIR‐PIT has a potential to treat both primary and metastatic lesions. It is possible that in patients with bilateral, multifocal lesions such a therapeutic approach, in contrast with ablative approaches such as radiofrequency or cryotherapy, will have minimal effect on the surrounding normal parenchyma. This could have significant benefits for patients at risk for recurrent, bilateral multifocal tumors and/or those with renal insufficiency. While we observed histological damages in a subset of proximal tubules in the irradiated normal parenchyma, this may be caused by the non‐specific accumulation of CA9‐IR700 in stromal tissues surrounding proximal tubules. In humans, the cytotoxic side effects are considered negligible because the penetration of light can be more strictly limited to cancer tissue. Future studies will determine the optimal timing of NIR light irradiation to ensure adequate accumulation of CA9‐IR700 in ccRCC tumors but minimal accumulation in the normal kidney.

In the clinical application of NIR‐PIT to ccRCC, EGFR‐targeted NIR‐PIT is another option because membrane EGFR expression was high in human ccRCC tissues, with a 92% positive rate. ³⁰ However, Dorđević et al. showed that membrane EGFR expression in ccRCC cells was heterogeneously distributed and the median percentage of membranous staining was 28.9%, ³¹ suggesting that EGFR‐targeted NIR‐PIT may have cytotoxic effects on only a subset of ccRCC cells. Meanwhile, previous studies reported that CA9 expression positive rates were more than 90% in ccRCC tissues, with diffuse strong expression in most cases. ³² , ³³ This suggests that CA9‐targeted NIR‐PIT would eradicate most ccRCC cells. Importantly, high CA9 expression was associated with lower grade and more favorable prognosis in patients with ccRCC. ³⁴ The association between low CA9‐expressing ccRCC and poorer prognosis would be caused by the simultaneous EGFR overexpression that activates the Akt/mammalian target of rapamycin (mTOR) pathway. ¹¹ Therefore, CA9‐targeted NIR‐PIT might be combined with EGFR‐targeted NIR‐PIT in the treatment of ccRCC.

There are several limitations in this study. First, since immunodeficient mice were used, we could not examine whether CA9‐targeted NIR‐PIT can activate host immunity. Our previous studies showed dendritic cell maturation and migration and intratumoral infiltration of CD8⁺ T cells after NIR‐PIT. ³⁵ , ³⁶ Because CA9‐targeted NIR‐PIT significantly increased the expression of DAMPs in vitro, CA9‐targeted NIR‐PIT is expected to stimulate anti‐cancer immune responses, which may affect untreated lesions. Second, we did not evaluate the therapeutic efficacy of CA9‐targeted NIR‐PIT using mice with metastatic ccRCC tumors. Given the clinical application of CA9‐targeted NIR‐PIT in oligometastatic ccRCC, it should be tested using various metastatic mouse models including lung and bone metastasis. Third, the therapeutic efficacy of repeated CA9‐targeted NIR‐PIT was not examined in this study. Remaining cancer cells after CA9‐targeted NIR‐PIT may be potentially resistant to the rechallenge of CA9‐targeted NIR‐PIT. Our previous study showed that multiple‐sessions of Nectin‐4‐targeted NIR‐PIT were more significantly effective in bladder cancer models compared to one‐time Nectin‐4‐targeted NIR‐PIT, ²⁵ suggesting the anti‐tumor effects of repeated NIR‐PIT. Fourth, we did not evaluate the inhibitory effect of CA9‐targeted NIR‐PIT on tumor growth of orthotopic xenograft models. Tumor growth following CA9‐targeted NIR‐PIT in orthotopic models can be evaluated by bioluminescence imaging in future analyses. Finally, the therapeutic efficacy of CA9‐targeted NIR‐PIT was not compared with that of the gold standard treatments including partial nephrectomy and cryoablation for small renal masses in this study. Potentially, CA9‐targeted NIR‐PIT can be expanded to the treatment of small renal masses of ccRCC.

5. CONCLUSION

We demonstrate that CA9‐targeted NIR‐PIT is highly effective against ccRCC. CA9‐targeted NIR‐PIT showed significant in vivo tumor control in ccRCC subcutaneous xenograft models, with 70%–80% CR rates following a single treatment of CA9‐targeted NIR‐PIT. Additionally, the cancer cell‐killing efficacy of CA9‐targeted NIR‐PIT was histologically confirmed using ccRCC orthotopic xenograft models. Therefore, CA9‐targeted NIR‐PIT can be a promising treatment for ccRCC, especially in the management of multiple primary ccRCC (hereditary renal cancer) and in patients with oligometastatic CA9‐expressing ccRCC.

AUTHOR CONTRIBUTIONS

Hiroshi Fukushima: Conceptualization; methodology; investigation; writing – original draft; formal analysis; validation. Seiichiro Takao: Investigation; writing – original draft; validation. Aki Furusawa: Investigation; writing – original draft. Motofumi Suzuki: Investigation. Youfeng Yang: Investigation. Christopher J. Ricketts: Investigation. Makoto Kano: Investigation. Shuhei Okuyama: Investigation. Hiroshi Yamamoto: Investigation. Miyu Kano: Investigation. Mark W. Ball: Investigation; writing – review and editing. Peter L. Choyke: Supervision; writing – review and editing. W. Marston Linehan: Conceptualization; supervision; writing – review and editing. Hisataka Kobayashi: Conceptualization; methodology; investigation; validation; supervision; funding acquisition; writing – review and editing; project administration; resources.

FUNDING INFORMATION

This work was supported by the Intramural Research Program of the National Institutes of Health, National Cancer Institute, Center for Cancer Research (award recipient: Hisataka Kobayashi, grant number: ZIA BC 011513).

CONFLICT OF INTEREST STATEMENT

The authors declare no potential conflicts of interest.

ETHICS STATEMENT

All animal experiments were performed in compliance with the Guide for the Care and Use of Laboratory Animals and approved by the local Animal Care and Use Committee (MIP‐003‐4‐Y).

Supporting information

Appendix S1: Supporting information.

Fukushima H, Takao S, Furusawa A, et al. Carbonic anhydrase‐9‐targeted near‐infrared photoimmunotherapy as a theranostic modality for clear cell renal cell carcinoma. Int J Cancer. 2025;156(12):2377‐2388. doi: 10.1002/ijc.35364

DATA AVAILABILITY STATEMENT

The data generated in this study are available within the article and Appendix S1 or from the corresponding author upon reasonable request.