Preliminary efficacy of [90Y]DOTA-biotin-avidin radiotherapy against non-muscle invasive bladder cancer

Bladder cancer represents 3% of all new cancer diagnoses per year. We propose intravesical radionuclide therapy using the β-emitter 90Y linked to DOTA-biotin-avidin ([90Y]DBA) to deliver short-range radiation against non-muscle invasive bladder cancer (NMIBC). Image-guided biodistribution of intravesical DBA was investigated in an animal model by radiolabeling DBA with the 68Ga and dynamic microPET imaging following intravesical infusion of [68Ga]DBA for up to 4 h and post-necropsy γ-counting of organs. The antitumor activity of [90Y]DBA was investigated using an orthotopic MB49 murine bladder cancer model. Mice were injected with luciferase-expressing MB49 cells and treated via intravesical administration with 9.2 MBq of [90Y]DBA or unlabeled DBA 3 days after the tumor implantation. Bioluminescence imaging was conducted after tumor implantation to monitor the bladder tumor growth. In addition, we investigated the effects of [90Y]DBA radiation on urothelial histology with immunohistochemistry analysis of bladder morphology. Our results demonstrated that DBA is contained in the bladder for up to 4 h after intravesical infusion. A single dose of [90Y]DBA radiation treatment significantly reduced growth of MB49 bladder carcinoma. Attaching 90Y-DOTA-biotin to avidin prevents its re-absorption into the blood and distribution throughout the rest of the body. Furthermore, immunohistochemistry demonstrated that [90Y]DBA radiation treatment did not cause short-term damage to urothelium at day 10, which appeared similar to the normal urothelium of healthy mice. Our data demonstrates the potential of intravesical [90Y]DBA as a treatment for non-muscle invasive bladder cancer.


Introduction
Bladder cancer is the 10th most commonly diagnosed cancer worldwide; approximately 573,000 new cases were diagnosed in 2020 and 213,000 deaths were estimated [1,2]. Approximately 75% of bladder cancers are diagnosed as non-muscle invasive bladder cancer (NMIBC) [3] and 10-20% of the recurrent cases of NMIBC progress Robert N. Taub has retired. This article is part of the Topical Collection on Translational research.
to muscle-invasive bladder cancer (MIBC) [4]. The standard treatment for NMIBC is transurethral resection followed by adjuvant therapy to prevent recurrence and/ or progression according to the tumor grade classification [5]. Intravesical bacillus Calmette-Guérin (BCG) is the gold standard therapy for adjuvant treatment in patients with intermediate and high-risk urothelial NMIBC, while a single chemotherapy instillation is recommended for tumors classified as low risk [6,7]. However, these adjuvant treatments are accompanied by significant adverse events, and 30-50% of patients do not respond or relapse within 5 years after appropriate BCG therapy [8]. Given the high recurrence rates after BCG therapy, safe and effective adjuvant locoregional therapies are needed in order to limit the recurrence and/or progression of NMIBC.
In recent years, there has been a much greater emphasis on radionuclide therapy, which demonstrates increased selectively in delivering radiation doses to the desired molecular targets expressed by cancer cells and reduced cytotoxicity to normal tissue [9,10]. We hypothesize that infusion of radioactive agents directly into the bladder that can reach the surface tumor cells as well as those below the most superficial layer may offer a promising path for treating relapsed NMIBC [11], avoiding the disadvantages related to radical cystectomy, external beam radiation or systemic treatments.
Avidin is a 66 KDa tetrameric glycoprotein that binds to biotin with high affinity. Avidin is positively charged with isoelectric point of 10.5 and contains terminal N-acetylglucosamine and mannose residues able to bind lectins [12,13]. Lectins are carbohydrate-binding proteins expressed on the surface of tumor cells that act to internalize their ligands [12]. Therefore, glycoproteins like avidin recognized by lectins could be used as carriers of therapeutic drugs for different types of tumors including urinary bladder cancer [13][14][15]. Furthermore, cationized avidin binds rapidly to the cell surface and enters into cells by endocytotic uptake [16]. A previous study demonstrated that the intravesical administration of radiolabeled avidin resulted in a preferential accumulation in tumor tissue compared to normal urothelium [17]. Our goal is to develop an effective minimally invasive therapy that can reach below the urothelium of the bladder to selectively target cancer cells at or below the surface.
We propose an intravesical therapeutic application of [ 90 Y]-DOTA-biotin-avidin ([ 90 Y]DBA) complex in order to deliver short-range radiation up to 12 mm deep [10] to reach the surface and throughout the bladder wall, which is 3-5 mm thick in humans when distended. We, therefore, sought to test the distribution and efficacy of DBA-based radiotherapy in rodents using real-time PET imaging and an orthotopic bladder cancer model.

Animals
Female C57BL/6NTac mice (B6) and NTac:SD rats (SD) were purchased from Taconic (NY, USA) and maintained on a normal diet. All animal experiments were conducted according to protocols approved by the Institutional Animal Care and Use Committee of Columbia University Irving Medical Center. For free unbound 68 Ga infusion, 68 Ga was eluted from a 68 Ge/ 68 Ga generator, equilibrated to pH = 4.0 with 1 M sodium acetate buffer, and injected into the rat's bladder as described in the next section.

Biodistribution of [ 68 Ga]DBA
An INSYTE.w.22G catheter (BD Biosciences, CA, USA) was inserted into bladder of SD rats as previously described [18]. 44-52 MBq (1.2-1.4 mCi) of [ 68 Ga]DBA or free unbound 68 Ga in 0.5 mL of PBS was injected into bladder of SD rats using the catheter. Immediately after 68 Ga injection, rats were placed on a microPET scanner (Inveon, Siemens, Germany), and a 4-h dynamic scan was acquired. After the PET scan, radioactivity was removed from the rat's bladder and the bladder was washed 3 times with 1 mL of PBS each wash. Rats were dissected and blood, liver, bladder, kidneys, and muscle samples were collected and counted on HIDEX gamma counter (HIDEX, Finland).

MB49 cell implantation
B6 mice were anesthetized with isoflurane and placed in supine position. An INSYTE.w.24G catheter (BD Biosciences, CA, USA) was inserted into the bladder through the urethra [18], and the bladder was emptied by draining out the urine. To damage the urothelium and enable tumor cell implantation, 100 μL of 0.1 μg/mL poly-L-lysine solution (PLL, MW = 70,000 to 150,000) (Sigma, MO, USA) was injected into the bladder. After 15 min incubation, PLL solution was removed, 50,000 of MB49 cells in 50 µL of DPBS were injected into the bladder, and mice were kept for 45 min under anesthesia in a supine position to allow for MB49 cell implantation. After the incubation with MB49 cells, the catheter/syringe was removed and mice were allowed to recover from anesthesia.

Luciferase imaging
Mice were injected intraperitoneally with 0.25 mL of luciferin (Research Products International Corp, IL, USA) solution (30 mg/mL in PBS). 15 min after luciferin injection, mice were imaged for 20 min using a bioluminescent optical scanner (MILabs, Netherlands).

Quantification and statistical analysis
Statistical analysis was performed using Prism 8.0 (San Diego, CA, USA). Statistical p-values were calculated using Mann-Whitney tests.

Uptake of avidin-Cy7 by MB49 tumor cells
To demonstrate the feasibility of our approach, we studied the ability of MB49 cells to uptake avidin from cell medium by using the avidin labeled with Cy7. We found that 2 h after incubation of MB49 cells with 1 μg of avidin-Cy7 > 99% of the cells contained varies amounts of internalized avidin-Cy7 (Fig. 1a). Increasing dose of avidin-Cy7 from 1 to 5 μg considerably (p = 0.0001) elevated amount of internalized avidin in the MB49 cells after 2 h incubation (Fig. 1b). This confirms prior work demonstrating bladder cancer uptake of avidin based [ 99m Tc]N 4 -Lys-Biotin [17].

[ 68 Ga]DBA is contained in the bladder and does not significantly enter the systemic circulation
To study the distribution of the DBA complex using microPET, we radiolabeled DBA with 68 Ga, which binds DOTA similar to 90 Y. We hypothesized that linking 68 Ga to avidin through the DOTA-biotin bond will prevent its leakage through the urothelium and prevent it from getting into the circulation to expose systemic organs. To examine the distribution of [ 68 Ga]DBA after intravesical delivery, we infused rats (n = 4) with [ 68 Ga]DBA and scanned animals for 4 h using microPET. Following microPET and 3 washes with PBS, delayed microPET images were obtained, animals were sacrificed and systemic organs were counted for the presence of 68 Ga (Fig. 2 and Table 1). We found that the > 99% of the injected [ 68 Ga]DBA was contained within the bladder without detectable activity in systemic organs (Fig. 2a, b). Furthermore, > 99% of [ 68 Ga]DBA was washed out after 3 rinses of PBS (Fig. 2c). We confirmed the [ 68 Ga]DBA microPET distribution by post-necropsy counting ( Table 1) and did not detect any significant accumulation of [ 68 Ga]DBA in systemic organs outside the bladder (< 0.05% of injected dose per gram of body weight (%ID/g)) and only 0.1%ID/g in the bladder. These data demonstrate that the radiolabeled DBA complex was not significantly absorbed into circulation after intravesical administration and was nearly completely washed out after 3 washes with PBS, making this strategy viable for therapeutic applications using a 90 Y payload. In comparison, we found a 29-fold increase in blood 68 Ga activity when rats received intravesical infusion of free unbound 68 Ga (Table 1).

Intravesical [ 90 Y]DBA therapy is effective against orthotopic MB49 tumors
To study the effects of intravesical [ 90 Y]DBA therapy on bladder cancer, we treated mice bearing orthotopic syngeneic MB49 bladder cancers with 9.2 MBq (250 µCi) of [ 90 Y]DBA for 2 h (Fig. 3a). We chose the dose and treatment time to result in an estimated exposure equivalent of not exceeding12.6 Gy to the bladder wall, which is a less than half of the published maximum tolerated dose 50% rate within in 5 years (TD50/5) for the bladder (30 Gy). We found that tumors grew below the urothelial surface 72 h post-implantation (Fig. 3b). To test the efficacy of our approach, we infused the bladders of tumor-bearing mice with [ 90 Y]DBA or with unlabeled DBA using urethral catheters to deliver [ 90 Y]DBA directly into the bladder. [ 90 Y]DBA was retained in the bladder for 2 h, removed via the catheter, and subsequently washed 3 times with PBS. We evaluated tumor growth in the [ 90 Y]DBA and control groups using bioluminescent imaging 10-days posttumor implantation. We found that tumors developed in the bladders of 100% of control mice. In contrast, 50% of the treated mice had no detectable bladder tumors (p = 0.02) per bioluminescent imaging. As a group, [ 90 Y]DBA-treated mice (n = 8) had significantly less (p = 0.01) tumor burden compared to mice treated with unlabeled DBA (n = 8), demonstrated by significantly lower bioluminescence signal (Fig. 3c, d).

Intravesical [ 90 Y]DBA treatment does not demonstrate any short-term adverse effects on treated mice
We examined animal weight and urothelium histology after treatment. We did not observe any significant difference in  weights between control and [ 90 Y]DBA treated groups (Fig. 4).
To study the effect of the [ 90 Y]DBA treatment on the bladder urothelium, we dissected bladders from [ 90 Y]DBA-treated mice and compared them to normal (no tumor) and control (untreated) tumor bearing mice. H&E staining revealed typical urothelium morphology in normal mice (Fig. 5a). Large tumors (asterisks on Fig. 5a) were found in the bladders of control mice. In contrast, mice implanted with orthotopic tumors and subsequently treated with [ 90 Y]DBA demonstrated normal bladder morphology and had no tumors in the treatment area, indicating successful locoregional treatment (Fig. 5a). We further analyzed the urothelial integrity by staining sections for keratin 5 (Krt5) and uroplakin 3 (Urp3). Normal bladder urothelium is comprised of 3 cell layers: (1) Krt5+ /p63+ /Urp-basal cells,   (Fig. 5b). As previously reported [20,21], we did not find Urp3+ cells in the bladder urothelium of tumor-bearing untreated mice (Fig. 5b). Instead, urothelium of these mice consisted of multiple layers of Krt5+ cells. In contrast, the bladder urothelium of [ 90 Y]DBA-treated mice had a normal basal layer of Krt5+ cells and intermediate/top layers of Urp3+ cells that closely resemble urothelium of normal untreated mice (Fig. 5b). Adult undamaged urothelium is quiescent epithelium containing very few proliferating cells. Damage induces cell proliferation and regeneration of the urothelium [22]. We stained bladder sections with anti-Ki67 antibody and found that both the bladder urothelium of the normal mice as well as tumor-bearing mice treated with [ 90 Y]DBA (10 days after treatment) did not have Ki67+ cells, indicating a quiescent urothelium (Fig. 5c). This suggests that [ 90 Y]DBA did not irreversibly damage bladder urothelium as the urothelium is not substantially different from the normal untreated mice. In contrast, we detected multiple Ki67+ cells in the urothelium of untreated tumor-bearing mice (Fig. 5c).

Discussion
NMIBC patients who have not responded to BCG have an urgent unmet need for improved therapeutic options. The standard of care (SOC) for NMIBC consists of transurethral resection with post-operative intravesical infection with BCG immunotherapy [23,24]. However, an estimated 40% of patients fail to respond to this SOC. Of patients who do initially respond, 30-50% relapse within 5 years [25,26]. Following BCG failure, there is no clear SOC. 90 Y is routinely used in clinical practice for locoregional radioablation of liver cancer with 90 Y labeled microspheres [27]. This approach has proven to be safe, even if up to 10% of radioactivity leaks into circulation. In contrast, local radioablation of refractory NMIBC using an intravesical approach can potentially near completely sequester radioactivity in the bladder and treat the entire urothelial surface as well as the entirety of the distended bladder wall. Previously, the α -emitter 213 Bi coupled to a monoclonal antibody (mAb) targeting EGFR has demonstrated therapeutic efficacy in BCGrefractory carcinoma of the bladder [28]. However, antigenic and vascular heterogeneity of tumors combined with the extremely short range of the α-emitter 213 Bi (~ 0.1 mm) or medium energy β-emitter 177 Lu (~ 2 mm) [29] might limit the efficacy if the biomarker is not expressed or if there are nascent tumor cells below the surface layer. These limitations might be overcome with intravesical infusion of high energy β-emitter ( 90 Y) that has a range of up to 12 mm. This range permits the so-called "crossfire" effect, where β particles can access several cell layers adjacent to the targeted cell in the event that not all cancer cells are successfully targeted. Furthermore, 90 Y β particles can traverse the entire distended bladder wall, depriving tumor cells from seeking refuge below the surface layer. Crossfire is critical to β particles emitter therapy to improve tumor dose homogeneity and to ensure sufficient dose to each cell [30]. Therefore, in our study, we tested the efficacy of [ 90 Y]DBA intravesical therapy using an orthotopic MB49 murine bladder tumor model. We hypothesized that the range of the high-energy β-emitter 90 Y will be effective at killing tumor cells both at the surface and below the urothelium. Nevertheless, while the 12-mm range of high energy β-emitters might be enough for the treatment of bladder cancers at earlier stages, this range might not reach late-stage larger bladder cancers.
To minimize 90 Y leakage through the bladder urothelium, we attached it to avidin. Avidin has a positive charge that facilitates cell uptake and has been shown to bind lectins expressed on cancer cells [31]. Our labeling approach using [ 90 Y]biotin coupled with avidin takes advantage of the high biotin-avidin affinity (Kd ~ 10 −15 M) and can be easily converted into a kit format for clinical use. Our real-time PET study of the DBA complex demonstrated that it is indeed retained in the bladder for up to 4 h and can be nearly completely washed out with 3 saline washes. These characteristics are critical to successfully translate our approach to clinical use, as they indicate that the DBA treatment will not significantly distribute and damage normal organs and can be removed in a controlled manner that will not cause any radiation safety concerns when the therapy is over.
To test the preliminary feasibility, safety, and efficacy of [ 90 Y]DBA, we used a syngeneic MB49 model implanted below the urothelial surface to test whether [ 90 Y]DBA can effectively reach beyond the surface urothelium of the bladder. Our results demonstrated significant efficacy after a 2-h intravesical treatment with [ 90 Y]DBA measured by bioluminescent signal from the luciferase-expressing tumor cells. Importantly, we chose an initial estimated treatment dose that exposes normal bladder tissue to less than 50% of the exposure that has been shown to cause longer-term adverse events from external beam radiation.
A shortcoming of the MB49 model is that it is highly aggressive and metastatic, which cannot be completely treated by a locoregional approach. Thus, we designated a 10-day endpoint in this preliminary study to evaluate the safety of our approach and to obtain an initial efficacy readout. We found that 50% of [ 90 Y]DBA-treated animals did not have detectable tumor after treatment (compared to 100% having tumors in the control group). Importantly, the 50% of treated animals in which tumors were present, demonstrated disease in the urethra, liver, and lungs, which were outside the field of treatment. Thus, while the MB49 model is good for preliminary efficacy studies, we are currently performing longer-term survival studies using other orthotopic tumor models, such as the BBN model [32], which better represent NMIBC as they do not typically metastasize but are less consistent and take longer to establish.
Damage to the bladder urothelium initiates a regenerative response which repairs most of the urothelium within 72 h [22]. Our study demonstrated that [ 90 Y]DBA treatment did not irreversibly damage urothelium and we observed a completely restored urothelium 10 days after treatment. While we did not observe any indication of acute radiation cystitis in treated mice, shorter-term and longer-term follow-up studies are being performed to evaluate for more acute and delayed adverse events [33].

Conclusion
We demonstrated that radiotherapy using a single administrated dose of intravesical [ 90 Y]DBA (9.8 MBq) significantly treated bladder tumors in an orthotopic model and did not cause irreversible damage to the urothelium at the experimental endpoint (day 10). [ 90 Y]DBA was contained within the bladder, which prevents radiation toxicity to the rest of the body. These preliminary results support the further study of [ 90 Y]DBA as a promising treatment for refractory NMIBC that can potentially spare patients from the morbidity related to bladder resection, external beam radiation, or systemic therapies.