Diphtheria Toxin-Epidermal Growth Factor Fusion Protein DAB₃₈₉EGF for the treatment of bladder cancer

Abstract

Purpose

The novel fusion protein, DAB₃₈₉EGF, is comprised of both the catalytic and translocation domains of diphtheria toxin that are fused to the human epidermal growth factor, providing a targeting and a toxicity component. We tested DAB₃₈₉EGF for anti-tumor activity in both in vitro and in vivo urinary bladder cancer models.

Experimental Design

Human bladder cancer lines were treated with DAB₃₈₉EGF and assessed for growth inhibition and clonogenic suppression. Using 6–8 week old female athymic nude mice implanted orthotopically with HTB9 cells, DAB₃₈₉EGF was administered intravesically twice weekly for two weeks. The response of the luciferase expressing HTB9 cells was monitored via bioluminescence as the primary endpoint..

Results

Treatment response with DAB₃₈₉EGF was specific and robust, with an IC₅₀ ranging from 0.5 to 15ng/ml in 8 tested bladder cancer cell lines, but greater than 50ng/ml in the EGFR-negative H520 control cell line. Simulating short duration intravesical therapy used clinically, a 2 hour treatment exposure of DAB₃₈₉EGF (10ng/ml) produced clonogenic suppression in three selected bladder cancer cell lines. In vivo, luciferase activity was suppressed in 5 of 6 mice treated with DAB₃₈₉EGF (70 μl (1ng/μL) per mouse), as compared to only 1 of 6 mice treated with a control DT fusion protein. Histologic assessment of tumor clearance correlated with the bioluminescent changes observed with DAB₃₈₉EGF treatment. Immunocompetent mice treated with intravesical DAB₃₈₉EGF did not demonstrate any non-specific systemic toxicity.

Conclusions

The intravesical delivery of targeted-toxin fusion proteins is a novel treatment approach for non-muscle-invasive urinary bladder cancer. With appropriate targeting, the treatments are effective and well tolerated in vivo.

Introduction

Urinary bladder cancer is common, with nearly 70,000 new cases and 15,000 deaths expected in 2011 in the United States.(1) The majority of these new cases are “superficial” or non-muscle invasive and treated with cystoscopic resection combined with intravesical medical therapy directly instilled into the bladder in those at higher risk of recurrence.(2) Bacillus Calmette-Guérin (BCG) is a live, attenuated Mycobacterium bovis strain with a long history of use as a tuberculosis vaccine globally. Intravesical BCG was introduced against early stage, non-muscle invasive bladder cancer in the 1970’s and continues to be regarded as the most effective intravesical therapy to date.(3) Despite BCG therapy, recurrent disease is common(4) and greater than 50% among high risk patients,(5) especially those with high-grade disease. Treatment options at the time of recurrence include additional intravesical therapy or even surgical cystectomy, recognizing the risk of lethal progression to more advanced bladder cancer in some cases. Therefore the clinical management of bladder cancer requires frequent and prolonged cystoscopic and laboratory assessments and retreatment, historically making bladder cancer a costly disease to manage.(6)

Targeted toxins are fusion proteins combining a targeting (ligand) and effector (toxin) component to selectively kill cells with specific cell membrane features. The Food and Drug Administration (FDA) has approved denileukin diftitox (ONTAK), a chimeric protein of modified diphtheria toxin (DT) and interleukin-2 (IL2), for the treatment of cutaneous T-cell lymphoma, confirming the translational relevance of toxic chimeric proteins in oncology.(7) We studied DAB₃₈₉EGF which links a modified DT to the epidermal growth factor (EGF) via His-Ala fusion that replaces amino acids 391–535, the native DT binding domain. Targeted toxins kill tumor cells through specific biochemical mechanisms, in this case through the inhibition of elongation factor 2, critical to translation. DAB₃₈₉EGF has previously been tested in the EGF receptor (EGFR)-expressing glioma cells, with good preclinical efficacy, but the complexity of delivering intra cranial therapy has limited its clinical testing.(8)

EGFR is frequently expressed in bladder cancer and is believed to play an important role in bladder cancer pathogenesis. EGF, the ligand of EGFR, has long been recognized for its role in stimulating the growth of bladder cancer cells in vitro, (9) with more recent transgenic in vivo data demonstrating urothelial hyperplasia with over-expression of EGFR.(10) In humans, EGFR is frequently over-expressed in bladder cancer cells, while this is uncommon in the normal urothelium.(11) In one large study of 245 bladder cancer patients, the majority (71%) of whom had non-muscle invasive disease, revealed that 72% were positive for EGFR expression and 27% of the total cohort was rated as having high expression(12). The cellular orientation of EGFR is important, considering the necessary interaction with DAB₃₈₉EGF - notably, EGFR is most frequently expressed on the luminal surface of bladder cancer cells, while it is frequently expressed on the basal (non-luminal) aspect of normal urothelium, and therefore not available to interact with EGF-directed intravesical therapy.(13, 14). Additionally, EGFR protein overexpression in bladder cancer is not limited to muscle-invasive tumors, but is also seen in low-grade, non-muscle invasive cases.(15) Taken together and based on these features, EGFR is a rational target for intravesical therapy with an EGF-targeted toxic fusion protein.

We hypothesize that an EGFR-targeting conjugated toxin protein will effectively treat bladder cancer in vivo without significant systematic toxicity. The aim of this present study is to examine DAB₃₈₉EGF’s effect in both cell culture and orthotopic mouse models of bladder cancer.

Material and Methods

Cell culture and reagents

Human bladder cancer cells J82, RT4, CRL1749 (CRL), T24, TCCSUP (SUP), HTB9 (American Type Culture Collection, Manassas, VA, USA) and lung cancer cell lines H520 and A549 were grown in OptiMEM (Gibco, Grand Island, NY, USA) with 3.75% fetal bovine serum (Gemini, Woodland, CA, USA) and 100 μg/ml streptomycin-100 IU/ml penicillin sulfate (Life Technologies, Grand Island, NY, USA). All cell lines were incubated at 37 °C with 5% CO₂. EFGR antibody and related secondary antibody were purchased from Cell Signaling Technology (Danvers, MA). The DAB₃₈₉EGF (DT-EGF) and DAB₃₈₈GMCSF (granulocyte-macrophage colony-stimulating factor (GM-CSF) fused to Diphtheria Toxin) were prepared as previously described (8, 16).

Western blot

For western blot assessment, the cells were plated in culture dishes until confluent and harvested by scraping and washing with PBS. Cells were collected by centrifuge at 1100 rpm and pellets were resuspended in lysis solution (10 mM Tris, 150 mM NaCl, 1% Triton X-100, 1 mM EDTA, 1 mM EGTA, 0.2 mM Sodium Orthovanadate, 0.5% NP-40, 0.3 mM PMSF, 10 ug/ml Aprotinin). Protein estimation was performed with Pierce protein dye, using a DU 800 Beckman Coulter spectrophotometer (Fullerton, CA). Fifty μg of protein was electrophoresed using NuPAGE 4–12% Bis-Tris Gel (Invitorgen, Carlsbad, CA). The protein was transferred to a nitrocellulose membrane (Invitorgen, Carlsbad, CA) using a wet method at 100 volts for 1 hour. The membrane was then blocked with 5% (w/v) milk and placed in a rotator for 1 hour at room temperature. The primary antibody was added in milk (2.5% w/v) and allowed to incubate overnight at 4° C after which it was washed with PBS/0.05% TWEEN-20 three times (15 minutes each) before the appropriate horseradish peroxidase-linked secondary antibody was added and incubated for an additional hour at room temperature. The membrane was again washed 3 × 15 minutes, before the addition of Pierce SuperSignal chemiluminescent substrate (Rockford, IL) and exposed. The films were scanned using EPSON PERFECTION V500 PHOTO and quantified by Image J ((NIH).

Cell proliferation assay

Cell proliferation was assessed using a tetrazolium-based assay (CellTiter 96® AQueous One Solution-Promega Corporation, Madison, WI). Three thousand cells in 50 μl of media per well were plated in 96-well plates in triplicate using the bladder cell lines. Twenty-four hours after plating, 50μl cell culture medium containing different concentrations of DAB₃₈₉EGF were added to give a total volume of 100 μl in each well. 50 μl cell culture media was used in the control condition. Seventy-two hours after the treatment, 20 μl of the AQueous One Solution was added to each well. Colorimetric analysis using a 96-well plate reader (V_max Kinetic Microplate Reader, Molecular Devices, Sunnyvale, CA) was performed between 2 and 4 hr (wavelength of 490 nm), depending on cell type and cell density. Cell proliferation assays were performed in triplicate.

Clonogenic assays

Clonogenic survival was defined as the ability of the cells to maintain their clonogenic capacity and to form colonies. Depending on cell type, 200 to 500 cells were seeded into 6-well dishes in 2 mL of medium. The next day cells were treated with DAB₃₈₉EGF, which were then maintained for 7–10 days at 37 °C in a 5% CO₂ incubator. The cells were fixed with 12.5% acetic acid in 30% methanol and then stained with Brilliant Blue R. The plate was photographed or the colonies containing ≥50 cells were counted. The effect of short duration DAB₃₈₉EGF treatments was also examined. J82, SUP and T24 cells were seeded (300 cells/well) in a 6-well plate in 2 ml of medium as above. Twenty-four hours later, 2 ml of medium containing DAB₃₈₉EGF was used to replace the normal medium for 2 to 24 hours, following which the treatment medium was removed and quickly replaced with fresh medium. The clonogenic survival was examined according to the clonogenic assay method described above.

DAB₃₈₉EGF stability test

4×10⁵ J82 cells were seeded in 6 well plates (an empty well served as a no cell control) for 24 hours. The normal medium was replaced with 2ml 20ng/ml DAB₃₈₉EGF in medium for 2 hours in a 37 °C incubator. The treatment medium containing DAB₃₈₉EGF was then harvested and tested for cell killing and used to treat T24, J82 and SUP for 3 days, giving a final DAB₃₈₉EGF treatment concentration of 10 ng/ml, as described.

Preparation of HTB9-Luc cells

For imaging purposes, the human bladder cancer cell line HTB9 was infected with a lentivirus containing the firefly luciferase gene. Briefly, the full-length luciferase-coding DNA sequence was amplified by PCR reaction from the pGL-3 vector (Promega, Madison, WI, USA) and inserted at BamHI and NotI sites of the lentiviral vector pLEX. In addition, Kozak sequence GCCACC was added between BamHI and start codon ATG of luciferase to ensure the luciferase expressed correctly. To generate luciferase-expressing lentivirus (Lenti-luc), this vector was co-transfected using calcium-phosphate precipitation method with packaging plasmid psPAX2 and envelope plasmid pMD2.G into HEK-293T cells. Supernatant was collected and filtered with 0.45 μM micro-filter for infection. HTB9 cells were then infected with Lenti-luc virus (25 μL viral supernatant/mL medium) and mixed with polybrene (4 μg/mL medium). A single clone with strong luminescent intensity was selected with puromycin and subcultured. Approximately passages 3~4 HTB9-Luc cells were used for further work.

Animals

Female nu/nu and C57BL/6 mice were purchased from the Animal Production Area of the National Cancer Institute-Frederick Cancer Research and Development Center (Frederick, MD). The animals were housed four per cage in a specific pathogen-free animal facility and fed with regular chow diet with water ad libitum. The experimental protocol was reviewed and approved by the Institutional Animal Care and Use Committee.

Orthotopic implantation

HTB9-Luc cells were harvested from 70% to 80% confluent cultures by exposure to trypsin. Proteolysis was stopped with medium containing 10% fetal bovine serum, and the cells were washed once in serum-free medium and resuspended in PBS. Female athymic nude mice 6 to 8 weeks of age had their bladders catheterized with a 30 gauge plastic catheter, under sterile conditions and using lubricant, by separating the animal legs and exposing the external urethral orifice for catheterization. Anesthesia was accomplished with pentobarbital sodium 60mg/kg. The adequacy of anesthesia was monitored every 2–3 minutes by response to toe/skin pinch. The bladder was washed with PBS and scratched with the catheter tip before instilling 100 μL of 1.5 ×10⁶ HTB9-Luc cells. The mild physical irritation caused by the catheter scratching is believed to increase the tumor development rate and degree of scratching contributes to the tumor stage (i.e. superficial). The urethra was temporarily closed with a single, sterile suture at the distal portion of the urethra thus retaining the cells in the bladder for 3 hours. This single suture was placed in the distal portion of the urethra under sterile conditions and accomplished through a single needle pass. The suture was removed by cutting the suture with a scissor and pulling through the non-knotted end, without further invasive animal manipulation. Mice with xenograft tumor implantation confirmed by positive luciferase activity 7 days post human cancer cell implantation were used for subsequent DAB₃₈₉EGF treatment.

Bioluminescence measurement

The primary endpoint of the in vivo study was bioluminescent activity after 4 intravesical treatments. Each mouse received 150 mg luciferin/kg body weight (e.g. for a 20 g mouse, we injected 200 μL to deliver 1.5 mg of luciferin). The luciferin was injected intra-peritoneally (i.p.) 12 minutes before imaging. Luciferin-injected mice were anesthetized and monitored using the XG-8 isofluorane device. The mice were first placed into an induction chamber and anesthetized with a high initial dose of isofluorane (8%) and then maintained with a lower dose of isofluorane (2.5%) through a manifold located on the imaging stage. Imaging of the mice was then conducted under continuously anesthetized conditions. The mice were imaged with In Vivo Imaging System (IVIS) 200 Imaging System (Xenogen Corp., Alameda, CA USA). Finally, 4 weeks post tumor cell implantation, the mice were sacrificed and bladders were taken for hematoxylin and eosin (H & E) staining and immunohistochemistry stain using whole-mount bladder step-sections.

In vivo treatment

One week after implantation, the degree of bioluminescence was measured and used to equally distribute the mice into balanced treatment and control groups. Drugs were stored at −80°C and thawed on ice. 70μls of 1ng/μL drug in PBS per mouse was instilled into the animal bladder following a similar procedure as cancer cell implantation procedure. This intravesical treatment was maintained for two hours each time, roughly simulating the short duration of intravesical therapy used in current clinical practice. The mice were treated twice per week for two weeks. Tumor growth was monitored weekly by the IVIS Imaging System. Following sacrifice, the bladders were harvested, and formalin fixed and paraffin embedded for immunohistochemistry and fluorescent in situ histology analysis.

Histological Analysis

Tissue processing and H&E staining of 5 micron tissue sections were performed either by the Diabetes and Endocrinology Research Center Histology Core Facility at the Barbara Davis Center for Childhood Diabetes or by the Prostate Diagnostic Laboratory, University of Colorado Anschutz Medical Campus. The slides were reviewed by a pathologist, Francisco G. La Rosa, MD. The pathology evaluation of urinary bladder specimens was done to confirm the presence or absence of tumor as observed in the bioluminescent measurement. Examination of bladder, liver and kidney specimens were also done to evaluate for toxic related changes in these tissues.

FISH analysis

Formalin-fixed, paraffin-embedded mouse tissue sections were subjected to a dual-color fluorescence in situ hybridization (FISH) assay using a human/mouse probe mix, to allow for unambiguous identification of the human xenograft in murine bladder when required. Human and mouse Cot-1 DNA (Invitrogen) were labeled with SpectrumRed and SpectrumGreen conjugated dUTPs respectively, (Abbott Molecular) using the Nick Translation Reagent Kit (Abbott Molecular Catalog#32-801300), according to manufacturer’s instructions. FISH assays were performed according to standard protocol (17). Initially the slides were incubated for 2 h at 56°C, deparaffinized in Citri-Solv (Fisher) and washed in 100% ethanol. The slides were sequentially incubated in 2xSSC at 75°C for 15 min, digested in 0.50mg/ml Proteinase K/2xSSC at 45°C for 15 minutes, washed in 2xSSC for 5 min, and dehydrated in ethanol. Probe mix was applied to the selected hybridization area, which was covered with glass coverslip and sealed with rubber cement. DNA denaturation was performed for 15 min at 85°C and hybridization was allowed to occur at 37°C for 15 h. Post-hybridization washes were performed sequentially with 2xSSC/0.3%NP40 (pH 7.0–7.5) at 72 °C for 2 minand 2xSSC at room temperature for 2 min, followed by dehydration in ethanol. Chromatin was counterstained with DAPI (0.3 μg/ml in Vectashield mounting medium, Vector Laboratories). The work was performed by University of Colorado Cancer Center Cytogenetics Core Program directed by Dr. Marileila Varella-Garcia.

Toxicology Studies

C57BL/6 mice were treated intravesically with 70 μL at concentrations of 1ng/μL and 43ng/μL (yielding a total dose administered for 70 ng and 3μg) of DAB₃₈₉EGF per mouse every other day for four doses. Additionally, systemic tolerance was assessed by giving 70 μL (43ng/μL) DAB₃₈₉EGF per mouse intravenously every other day for two doses, based on the use of this dose in previous in vivo work with this agent. (8). The mice were euthanized at the completion of the experiments and their tissues of liver, kidney and bladder were collected and formalin fixed. H&E stains were performed for these paraffin-embedded mouse tissue sections.

Statistical Analysis

A student’s t-test was used for comparisons of cell proliferation results of two groups (A540 vs. H520) and Fisher statistical test was performed between in vivo results of DAB₃₈₉EGF vs. DAB₃₈₈GMCSF treatment groups. A one-way analysis of variance (ANOVA) was used to determine whether the means were significantly different overall for the remainder of the statistical analysis. If the overall means were significantly different, we carried out a pair-wise comparison. Tukey’s method for multiple comparison was used to compare differences between groups. SE of the mean was indicated for each value by a bar. All analyses were carried out using GraphPad Prism version 5.0c for Windows (GraphPad Software, San Diego) and SAS statistical software (SAS, Cary, NC).

Results

DAB₃₈₉EGF treatment inhibited the growth of the human bladder cells and is correlated with EGFR expression

As shown in Figure 1A, DAB₃₈₉EGF significantly inhibited cellular growth in all tested human bladder cancer cells. To account for differences in the observed level of IC₅₀, we quantified the EGFR protein expression level of the tested cell lines via Western blotting, as shown in Figure 1B. HTB9 and T24 cells exhibited higher EGFR expression compared to J82 and CRL cells, which is consistent with their sensitivity to DAB₃₈₉EGF (Figure 1A).

Fig. 1. Correlation between cell killing effect and EGFR expression of DAB₃₈₉EGF.

A: Cell proliferation was assessed with three day DAB₃₈₉EGF treatment at concentrations ranging from 0 to 50 ng/ml of six human bladder cancer cell lines, CRL, J82, HTB9, RT4, SUP and T24 was assessed using a tetrazolium-based assay in triplicate. Cell killing effect is represented as the ratio of the optical density reading of the treated cells compared to the no-treatment control at 490 nm (normalized to 1 in no treatment control). B: EGFR expression of these six cell lines was determined by Western blot. The ratio of EGFR to β-actin was calculated by the band density of Western blots of each cell line using Image J software. C: Clonogenic assessment was additionally performed after 10 to 15 days of treatment with 1.5 and 15 ng/ml DAB₃₈₉EGF in 3 human bladder cancer cell lines, J82, SUP and T24 to characterize the dose-dependent effect of DAB₃₈₉EGF, with nearly complete suppression at 15 ng/ml dose, shown in the third column for each cell type. The results are represented as the percentage of colony numbers compared to no-treatment controls (normalized to 100%) (n=4,* p<0.05 versus corresponding untreated controls).

To further investigate the relationship between cell killing efficacy and EGFR expression, we performed Cell proliferation (3 day treatment) and clonogenic assay (10 day treatment) using two well-characterized lung cancer cell lines with (A549) and without (H520) EGFR expression. Cell proliferation results demonstrate that the A549 cell line has much higher sensitivity to DAB₃₈₉EGF compared with H520 (Supplemental Figure 1A). Furthermore, clonogenic assay results demonstrate that at the concentration of 25ng/ml, DAB₃₈₉EGF almost completely inhibits the colony formation in A549 cells with little influence on H520 cell growth (Supplemental Figure 1B).

After 10 days of treatment, clonogenic assay for selected four human bladder cancer cell lines was performed (Figure 1C), showing that a higher concentration of DAB₃₈₉EGF had greater inhibition of colony formation, indicating that this suppression is dose dependent.

Short duration DAB₃₈₉EGF treatment suppressed clonogenicity

In order to prepare for in vivo intravesical testing, short duration treatments with DAB₃₈₉EGF treatment were used in vitro for the clonogenic assay. As shown in Figure 2A, a 2 hour exposure with DAB₃₈₉EGF treatment demonstrated growth inhibition in three selected bladder cancer cell lines (J82, SUP and T24). Meaningful reductions in clonogenic formation were seen with 2 hours of exposure and large or even complete suppression was seen with a 24 hour treatment. To further explore the stability of DAB₃₈₉EGF, considering that in a clinical application that the dwell time would be less than 3 hours, we pre-incubated this agent with J82 cells for 1, 2 and 3 hours and then applied the collected medium containing DAB₃₈₉EGF to 3 selected cell lines (T24, J82 and SUP). As shown in Figures 2B~D, the media with and without cells exhibits similar cell killing efficiency in a manner comparable to the non-incubated (control) treatment. The efficacy of DAB₃₈₉EGF did not vary despite pre-treatment incubation with J82 cells for various time course (1–3 hours) prior to the treatment phase when we examined the cell killing effects of three cell lines T24, J82 and SUP. All together, these results reveal that DAB₃₈₉EGF exhibits significant cell killing effect with a 2hour exposure and is very stable in cell culture, providing a strong rationale for studying clinically applicable intravesical treatment in vivo.

Fig. 2. Rapid cell killing efficiency and stability of DAB₃₈₉EGF in cell culture medium.

Three human bladder cancer cell lines J82, SUP and T24 were seeded for 24 hours and then incubated with 10ng/ml of DAB₃₈₉EGF for either 2 or 24 hours. The media was then changed to fresh culture medium, they were incubated for another 10 days and then the colony counts were performed (2A). A short 2 hour treatment duration yielded significant cell growth inhibition, while 24 hour treatment had greater inhibitory efficiency, with nearly complete suppression observed, shown in the third column of each cell type. The results are represented as the percentage of colony numbers compared to no-treatment controls (normalized to 100%) (2A, n=3, * p<0.05 versus corresponding untreated controls). The medium containing DAB₃₈₉EGF which was previously incubated in J82 cells for 1, 2 and 3 hours was collected and then used to treat T24, J82 and SUP for 3 days, giving a final concentration of DAB₃₈₉EGF at 10 ng/ml, normalized to the no treatment condition (2B~D, n=3, *p<0.05 versus corresponding untreated controls). Cell killing effect was expressed as the ratio of optical density reading of no treatment control to respective treatments at 490 nm (normalized to 1 in no treatment control).

Establishment of orthotopical xenograft bladder cancer model

Using human bladder cancer HTB9 cells, we have successfully established an orthotopic, xenograft bladder cancer model confirmed by bioluminescent imaging, H&E stain and FISH. First, HTB9 cells were transfected so that they expressed Luciferase and single luciferase-bearing cell clones were selected based on bioluminescent expression after luciferin exposure (Figure 3A). Their Luciferase activity was further confirmed after 3~4 passages in culture (Figure 3B). Bioluminescent imaging provides for in vivo determination of bladder cancer cell implantation and tumor formation in the nude mice. Using the technique described, an implantation rate of approximately 75% was obtained, which was assessed at 1 week after implantation (Figure 3C). Once tumor implantation was confirmed at day 7, we selected and followed a cohort of 4 mice with positive Luciferase activity for 2 months to establish the stability of long-term, orthotopic implantation with this method and found that these mice have positive but relatively decreased Luciferase activity (Figure 3D). Based on this information, we selected mice for intravesical DAB₃₈₉EGF treatment on the basis of Luciferase activity 7 days after the cells were administered and established the relative stability of this model over several months, with a low spontaneous tumor loss rate. Unlike a syngeneic tumor model, these mice bearing the positive Luciferase activity do not necessarily develop palpable tumors and the tumor implants do not appear to be fatal with longer term follow-up. Based on serial FISH and histologic assessments with this model, bioluminescent findings appear to be the most reliable marker of tumor presence and this was the primary endpoint of this study.

Fig. 3. Establishment of orthotopical xenograft bladder cancer model.

After transfection with Lenti-luc virus, HTB9 cells were seeded into 96 well plate and single clones with positive Luciferase activity were selected (3A). These cells were cultured and Luciferase activity was further confirmed (3B). The cells were implanted intravesically into the mouse bladders with approximately 75% tumor-take rate (3C). Mice with positive tumor take confirmed by bioluminescent imaging were followed for two months with the mice demonstrating persistent tumor implantation (3D), although we have observed occasional spontaneous tumor loss over time, after initial confirmation.

Treatment of DAB₃₈₉EGF

Mice were selected for DAB₃₈₉EGF treatment based on confirmation of tumor implantation, via the presence of luminescent activity 7 days after the cell administration procedure, with the treatment schema shown in Figure 4A. As a control treatment, the related agent DAB₃₈₈GMCSF was used and seemed suitable, as the urothelium is not known to express GMCSF receptors(18). Six mice were treated with intravesical DAB₃₈₉EGF and 6 with control (DAB₃₈₈GMCSF). After 1 week, a subjective decrease in luminescent activity was observed in the DAB₃₈₉EGF treated mice, with nearly uniform loss of luminescent activity after 2 weeks treatment (Figure 4B). In contrast, there was little decrease of luminescent activity with DAB₃₈₈GMCSF treatment in all but one of the mice after the 2 week treatment phase (Figure 4B). At the conclusion of the treatments, luciferase activity was lost in 5 of the 6 mice treated with DAB₃₈₉EGF compared to 1 of the 6 mice treated with control (Figure 4B and Table 1). Fisher statistical test confirmed that the difference between DAB₃₈₉EGF and DAB₃₈₈GMCSF treatment groups is significant (p<0.05). The bladders were then examined histologically and the findings were in agreement with the luciferase results, with loss of bladder cancer tissue observed in 5 of the DAB₃₈₉EGF-treated mice as compared to 1 of the DAB₃₈₈GMCSF-treated mice. The presence of tumor xenograft was additionally confirmed in selected mice by using species-specific dual-color FISH assay (human versus murine) as noted in Figure 4C. This assay is highly sensitive and specific and is able to detect a very small number of xenograft (human) cells in the murine background.

Fig. 4. In vivo study of DAB₃₈₉EGF.

Schematic representation (4A) of experimental protocol described in Materials and Methods. After instillation of luciferase-overexpressing HTB9 in the nude mice, bioluminescent imaging was performed at day 7 and luminescent positive mice were randomly divided into two groups (n=6 each group). Groups I and II were intravesically treated with 70 μl (1 ng/μl) DAB₃₈₈GMCSF or DAB₃₈₉EGF respectively in PBS per mouse, twice per week for two weeks, starting at day 10. The luminescent intensity was measured at day17 and day 24 (4B). Mice were sacrificed at day 27 and bladder tissues were collected and fixed. Histological sections from these tissues were subjected to hematoxylin and eosin (H&E) stain with fluorescent in situ hybridization (FISH) assessment used to confirm the presence or absence of human xenograft in the treated animals. In Figure 4C, H&E staining demonstrates loss of tumor in the actively treated (a), but not the control treated mice (d). FISH assessment of the control-treated mice confirm the persistence of both human (nuclei stained in red) and mouse (nuclei stained in green)-originated cells (e), while the DAB₃₈₉EGF-treated mice do not demonstrate any hybridization with the human (red) probe (b), consistent with the H&E staining of tumor clearance. Nuclear staining with DAPI in both settings is provided for reference (c and d).

Table 1.

Luciferase activity results of DAB₃₈₉EGF treatment

	Total mice	Luciferase activity loss	Luciferase activity presence
DAB389EGF	6	5	1
DAB388GMCSF	6	1	5

P<0.05 DAB₃₈₉EGFvs. DAB₃₈₈GMCSF

Preliminary Toxicity Results

Previous work has demonstrated that DAB₃₈₉EGF at doses of ≥5μg given intravenously in BALB/c mice leads to listlessness, lethargy, dehydration, and reduced activity.(19) In the current experiments, we used the immunocompetent C57BL/6 mice to examine the toxic effect of DAB₃₈₉EGF when given intravesically compared to intravenous treatment. In these mice, there was no obvious behavioral toxicity or reduction in activity when given 4 treatments intravesically, twice per week for two weeks, with either 70ng or 3μg DAB₃₈₉EGF. Histologic examination of organ tissue from the mice given intravesical treatment showed no changes or toxicity as evidenced in figure 5. In contrast, in the intravenously treated group, histologic examinations of kidney, liver and bladder reveal the presence of severe renal tubular necrosis and rare apoptotic cells in the kidney (Figure 5A) and liver (Figure 5B) with two treatments of 3μg DAB₃₈₉EGF every other day, although the toxicity in the urothelium (Figure 5C) was not evident. As a result of subjective changes in the intravenously-treated mice, the intravenous treatment was stopped in all mice after 2 treatments, with the observation of reduced activity and general weakness in the mice, consistent with systemic toxicity.

Fig. 5. Toxicity information of DAB₃₈₉EGF in immune-competent mice.

C57BL/6 mice were treated with intravesically a total administration dose of 70 ng or 3 μg DAB₃₈₉EGF per mouse twice per week for two weeks with a separate cohort treated with 3 μg DAB₃₈₉EGF intravenously every other day for two treatments (four mice per group, total 12 mice). After three days from the last intravesical treatment or one day for last intravenous treatment, the mice were euthanized and their tissues were collected and formalin fixed. H&E stains were performed for these paraffin-embedded mouse tissue sections (5A, kidney; 5B, liver; 5C, bladder). Histologic evidence of cellular damages was only noted in the in intravenously-treated mice with the observation of severe renal tubular necrosis and rare apoptotic cells in the kidney (Figure 5A) and liver (Figure 5B), although no toxicity was noted in the urothelium of the intravenously-treated mice (Figure 5C).

Discussion

In this work we present a novel use of an EGFR-directed fusion protein with a modified DT. Bladder cancer frequently over-expresses EGFR on the luminal surface, while such expression is uncommon on the normal urothelium. Using an orthotopic, xenograft model of bladder cancer, we showed a difference in the elimination of the implanted tumor after 2 weeks of intravesical treatment with DAB₃₈₉EGF, as compared to the control DT chimeric protein (DAB₃₈₈GMCSF). Importantly, in immunocompetent mice, there was clear, generalized toxicity with the intravenous DAB₃₈₉EGF treatment, but there was no evidence of histologic or generalized toxicity in the intravesically-treated mice at equal doses.

While targeted agents have made significant inroads into the clinical management of most common cancer types, there is no integration of such agents into the routine treatment of urothelial carcinoma. Importantly, in most cases, the utility of inhibiting a particular molecular pathway is dependent on “pathway addiction”, frequently denoted by an activation mutation, such as the EGFR mutations in non-small cell lung cancer (20), which predicts clinical benefit with the systemic use of EGFR inhibitors. Another weakness of selective pathway inhibition via a systemic agent is the common “cross talk” that occurs, in which the neoplastic cells utilize a secondary growth pathway in response to the inhibition of one pathway by a targeted therapy (21).

In contrast to other targeted agents, currently in clinical practice and given systemically toxic fusion proteins do not require pathway addiction for activity. Rather, they simply require differential expression of the receptor in question between the targeted tumor cells and the normal background; the receptor merely serves as a doorway, allowing entry of the chimeric, therapeutic protein component. EGFR in bladder cancer serves as a very favorable “targeting” receptor or portal in this light, in that is commonly and differentially expressed. Importantly, bladder cancer also offers a direct and unique method of delivery, via simple catheterization and intravesical administration. There are also available clinical assays to assess EGFR expression, allowing for the selection of patients for DAB₃₈₉EGF therapy, based on EGFR expression levels in tumor biopsies. BCG and other medical therapies for bladder cancer are currently given in this manner and the use of DAB₃₈₉EGF in this way would not represent a change in general practice. Of note, denileukin diftitox (ONTAK), a chimeric protein of DT and IL2, is currently in clinical use, in addition to other analogous agents currently in clinical evaluation (22), establishing the realistic translational potential of this approach in the novel intravesical setting.

Another advantage of intravesical therapy is the protective effect of the physiologic space provided by the bladder itself. As noted in our experiments, the intravesical delivery of DAB₃₈₉EGF at more than 40 times the effective therapeutic levels (70 ng compared to 3000 ng) was very well tolerated, while similar high doses of DAB₃₈₉EGF given intravenously led to toxicity. Simply, the bladder physically contains the fusion protein when delivered intravesically, limiting systemic exposure. In a clinical application, this will improve the clinical tolerability and safety of such an approach. Of note, we also did not see any toxicity in the mice with orthotopic bladder xenograft in pace, and therefore an altered bladder wall from tumor growth, when treated with 70 ng of intravesical DAB₃₈₉EGF during the therapeutic treatment experiments.

In summary, non-muscle invasive bladder cancer is common, frequently recurs after therapy, with little change in the medical therapy over 2 decades. The intravesical administration of DAB₃₈₉EGF yielded a high rate of tumor clearance after 2 weeks of therapy using an orthotopic, xenograft model of bladder cancer, in contrast to control-treated mice. In addition, intravesical treatment limits systemic exposure of the therapy, reducing non-specific toxicities. The use of targeted, toxic fusion proteins represents a novel therapeutic advantage, building on currently principles of intravesical treatment.

A Statement of Translational Relevance.

The first-line use of intravesical Bacillus Calmette-Guérin (BCG)for non-muscle invasive bladder cancer has not significantly changed in 2 decades. Despite this persistent utilization, intravesical BCG is associated with frequent, symptomatic bladder irritation and recurrent bladder cancer after BCG is common. The intravesical administration of a target-toxin protein represents a new treatment approach for early-stage bladder cancer. DAB₃₈₉EGF, which targets any epidermal growth factor receptor (EGFR)-expressing cell with a modified diphtheria toxin, is well suited as a superficial bladder cancer treatment. EGFR is differently expressed in bladder cancer, with frequent expression observed on superficial tumors, but not on the luminal surface of the normal urothelium. Additionally, since the bladder represents a physiologic space, systemic exposure and toxicity from DAB₃₈₉EGF would be expected to be minimal with intravesical use. In a future clinical application, EGFR expression is utilized to rationally selected bladder cancer patients for intravesical DAB₃₈₉EGF treatment.