2.1.

Cell Culture and Xenograft Tumor Model

Male Balb/c nude mice, $6\text{to}8\text{weeks}$ of age, weighing an average of $24\text{to}25\mathrm{g}$ were obtained from the Animal Resource Centre (ARC), Western Australia. The epithelial human bladder carcinoma cell line (MGH-U1) was used to establish the xenograft model. The cells were cultured as a monolayer in RPMI-1640 medium supplemented with 10% fetal bovine serum (HyClone, Logan, Utah), 1% nonessential amino acids (Gibco, United States), 1% sodium pyruvate (Gibco, United States), and $100\text{units}∕\mathrm{ml}$ penicillin/streptomycin (Gibco, United States) and incubated at $37°\mathrm{C}$ , 95% humidity, and 5% $\mathrm{C}{\mathrm{O}}_2$ . Before inoculation, the cell layer was washed with phosphate-buffered saline (PBS), trypsinized, and counted using a hemocytometer. Approximately $3.0\times {10}^6$ MGH-U1 human bladder carcinoma cells suspended in $150\mu \mathrm{l}$ of Hanks’ balanced salt solution (Gibco, United States) was injected subcutaneously into the lower flanks of the mice. The tumors were allowed to grow to sizes of $6\text{to}7\mathrm{mm}$ in diameter before PDT treatment was carried out, and the tumors were measured three times a week to assess tumor regression. Tumor volume was estimated by using the formula, $\text{volume}=(\pi ∕6\times d1\times d2\times d3)$ , where $d1$ , $d2$ , and $d3$ are tumor dimensions in three orthogonal directions. All procedures were approved by the Institutional Animal Care and Use Committee (IACUC), Singapore Health Services, Singapore, in accordance with international standards.

2.2.

Photosensitizer

Hypericin was purchased from Molecular Probes, Oregon. A stock solution of $5\mathrm{mg}∕\mathrm{ml}$ hypericin was prepared by adding $200\mu \mathrm{l}$ of dimethyl sulfoxide (DMSO, Sigma Aldrich Inc, St Louis, Missouri) to $1\mathrm{mg}$ of hypericin. The stock solution was further diluted in DMSO and PBS (1:3 v/v) and injected intravenously into the tail vein based on the weight of the animal at a dosage of $5\mathrm{mg}∕\mathrm{kg}$ .

2.3.

Treatment Protocol

Animals were randomly assigned to four different groups $(n=10)$ for treatment: (1) control (mice with untreated tumors), (2) PDT only, (3) bevacizumab only (without PDT), and (4) PDT in combination with bevacizumab. PDT involved the intravenous injection of hypericin, followed by illumination with a light source consisting of filtered halogen light (Zeiss KL1500) fitted with a customized $560\text{to}640\text{-}\mathrm{nm}$ bandpass filter. Light irradiation was performed $6\mathrm{h}$ post hypericin administration. A light fluence of $120\mathrm{J}∕{\mathrm{cm}}^2$ and irradiance rate of $100\mathrm{mW}∕{\mathrm{cm}}^2$ was used to perform PDT. Bevacizumab was administered by multiple intraperitoneal injections $(10\mathrm{mg}∕\mathrm{kg})$ at time 0 (immediately after PDT), $24\mathrm{h}$ , $48\mathrm{h}$ , and then every other day up to $90\text{days}$ post PDT (as previously described in Ferrario ¹¹). Treated and control tumors were monitored three times per week to chart the growth rate.

2.4.

Dosage of Bevacizumab (Avastin)

A stock solution of concentration $25\mathrm{mg}∕\mathrm{ml}$ bevacizumab (Avastin^™; Genentech, Inc., South San Francisco, California) was diluted with saline to prepare a working solution. It was administered intraperitonially at a dosage of $10\mathrm{mg}∕\mathrm{kg}$ .

2.5.

Preparation of Mice for In Vivo Fluorescence Imaging

Fluorescein isothiocyanate (FITC)-labeled dextran [molecular weight (MW) $150\mathrm{kDa}$ , Sigma Aldrich, United States] solution was prepared at a concentration of $25\mathrm{mg∕ml}$ . Then $10\mu \mathrm{l}$ FITC dextran per gram weight of the animal was injected intravenously. The nude mice were then anesthetized with a cocktail of Hypnorm ( $0.315\mathrm{mg}∕\mathrm{ml}$ fentanyl citrate and $10\mathrm{mg}∕\mathrm{ml}$ flanuisone, Janssen) and Dormicum ( $5\mathrm{mg}∕\mathrm{ml}$ midazolam HCl, David Bull Laboratories). The skin overlaying the tumor was carefully removed to expose the tumor.

2.6.

Laser Confocal Endomicroscopy

In vivo fluorescence microscopic imaging of tumor blood vessels was carried out using a laser confocal endomicroscope approximately $30\min$ after intravenous administration of FITC dextran. Experiments were performed on the animals from day 25 to day 90, based on the maximum tumor volume limit or the completion of the treatment. We previously described the design of a prototype of the laser confocal endomicroscope in detail.²⁰ The new Optiscan FIVE 1 system (Optiscan Pty Ltd., Victoria, Australia) features the same basic design with improved specifications. A schematic diagram of the system is shown in Fig. 2 . A $488\text{-}\mathrm{nm}$ laser provides excitation for in vivo fluorescence microscopy of tissue with a lateral resolution of $0.7\mu \mathrm{m}$ and an axial resolution of $7\mu \mathrm{m}$ within a field of view of $475\times 475\mu \mathrm{m}$ . The excitation light is coupled into a single optical fiber that acts as both a point source and a point detection pinhole for confocal imaging. A handheld rigid probe (model RBK6315A) with a probe shaft of length $150\mathrm{mm}$ and diameter $6.3\mathrm{mm}$ houses the miniaturized components of the $x\text{-}y$ scanning and $z$ axis actuator. A footswitch enables the imaging depth to be controlled along the $z$ axis range from the surface to $250\mu \mathrm{m}$ subsurface in $4\text{-}\mu \mathrm{m}$ steps, as well as to facilitate image capture. Serial $1024\times 1024\text{-pixels}$ images can be acquired at a frame rate of $0.7\text{frames}∕\mathrm{s}$ . For optimum image contrast, the laser power can be adjusted between $0\text{to}1000\mu \mathrm{W}$ at the $5.0\text{-}\mathrm{mm}$ -diameter distal tip of the probe in contact with the tissue. Fluorescence signals are collected via a $505\text{to}750\text{-}\mathrm{nm}$ emission filter.

Fig. 2

Schematic diagram of the confocal endomicroscopy system with a handheld rigid probe used for in vivo fluorescence imaging of tumor vasculature in animal models.

2.7.

Endomicroscope Image Processing and Analysis

Blood vessel analysis was carried out using the Cell D imaging software (Olympus, Soft Imaging Solutions GmbH, Germany). The measurement function was used to measure the cross-sectional areas of the blood vessels visible within the field of view of the endomicroscope images by drawing user-defined regions of interest along the vessels. The mean vessel area for each animal group was calculated from images captured from three locations with high vessel density and that are between 20 and $40\mu \mathrm{m}$ below the surface of the tumor.

2.8.

Immunohistochemistry

Processing of the samples was done using a tissue processor (Leica TP 1020, Germany). Briefly the tissue samples were fixed in 10% formalin for $24\mathrm{h}$ , and then processed in an ascending series of ethanol and subsequently cleared with xylene and embedded in paraffin. The paraffin-embedded samples of bladder tumors were cut at a thickness of $4\mu \mathrm{m}$ using a microtome (Leica RM 2135, Germany). The sections were mounted on superfrost/plus slides (Fischer Scientific, United States) and air-dried. On the day of staining, the slides were placed in $60°\mathrm{C}$ oven for $1\mathrm{h}$ and immersed in zylene for $10\min$ before rehydration in ethanol series. Sections were incubated in hydrogen peroxide for $10\min$ to block endogenous peroxidase activity. After which, the sections were incubated with VEGF (BD Biosciences Pharmingen, United States) primary antibody (1:100) for $1\mathrm{h}$ . To confirm the specificity of binding, normal mouse serum immunoglobulin G $\left(\mathrm{Ig}{\mathrm{G}}_1\right)$ (1:500) was used instead of primary antibody to serve as the negative control. Following extensive washing, sections were incubated for $30\min$ in the secondary biotinylated antibody followed by DAB Chromogen (Dako REAL EnVision Detection System, United States) for $10\min$ . Sections were then counterstained with Harris’s hematoxylin and dehydrated in ascending grades of ethanol before being cleared in xylene and mounted under a cover slip. Images from the slides were captured using an image processing software (AxioVision v.4.6.3.0, Carl Zeiss Imaging Solutions, GmbH). The National Institutes of Health (NIH) Image J v1.62 software was used to analyze the images and quantify the expression of VEGF. Briefly, the percentage of positively stained cells was calculated by obtaining the area of the immunostained regions divided by the total area of the image. IHC scoring was performed based on the prevalence of CD31 staining within the tumor (no $\text{staining}=0$ , $<10\%$ $\text{staining}=1$ , $10\text{to}25\%=2$ , $26\text{to}50\%=3$ , $51\text{to}75\%=4$ , and $76\text{to}100\%=5$ ).

2.9.

Immunofluorescence

Fresh frozen tissue sections were fixed in 2% paraformaldehyde for $2\min$ . The specimens were blocked for $1\mathrm{h}$ with normal goat serum in Triton X-100 (BDH, United Kingdom). After blocking, sections were incubated overnight with VEGF mouse monoclonal antibody (Abcam, United Kingdom) at $4°\mathrm{C}$ . Nonimmune IgG was used as a control. After rinsing in PBS, the specimens were stained with FITC-conjugated secondary antibody for $2\mathrm{h}$ at room temperature in the dark. The slides were then rinsed with PBS and stained with DAPI for $30\min$ . Finally, the slides were rinsed and mounted with Vectashield^® Mounting Medium (Vector Laboratories Inc., United States). IF images were captured using a laser confocal fluorescence microscope (Meta LSM 510, Carl Zeiss, Germany) and image analysis was performed using the ImageJ software (1.41o, W. Rasband, NIH, United States).

2.J.

Statistical Analyses

Statistical analysis was performed using GraphPad Prism, version 4.0 (GraphPad Software, Inc., San Diego, California). A one-way analysis of variance (ANOVA) was used to analyze the variance. Dunnett’s multiple comparison test was used to analyze the difference in mean vessel area between control and the other groups. Bonferroni’s multiple comparison test was used to compare the significance between all the groups for tumor regression experiments. A $p$ value of $<0.05$ was considered to be significant.

3.1.

Tumor Regression

To investigate the long-term effectiveness of PDT in combination with bevacizumab, we induced MGH-U1 xenograft tumors in athymic nude mice models. Tumors were allowed to grow to sizes of $6\text{to}7\mathrm{mm}$ in diameter before PDT treatment was carried out and were measured three times a week and charted for $90\text{days}$ (Fig. 3 ). A control IgG group was also evaluated. However, since the results were similar to the control tumors, they are not described in this paper. The mean tumor volume for each group of 8 to 10 animals is shown against the number of days post PDT. Tumor inhibition was calculated on day 29, when the control tumors reached a maximum volume of $2{\mathrm{cm}}^3$ . Tumor inhibition of $85.8\pm 1.5\%$ $(p<0.001)$ was observed in tumors treated with the combination therapy of PDT and bevacizumab in comparison to control tumors. A week after treatment, accelerated tumor growth was noticed in the combination therapy group, but there was a decrease thereafter in tumor size, resulting in a permanent and complete tumor regression. The tumors treated with PDT only and bevacizumab only, exhibited $57.8\pm 1.2\%$ $(p<0.001)$ and $76.5\pm 1.8\%$ $(p<0.001)$ inhibition, respectively, compared to the control group. Compared to the control, the overall tumor response was greater in the monotherapy groups of PDT only and bevacizumab only. The treatment modalities in our study did not induce any signs of toxicity such as excessive weight loss, diarrhea, or vomiting in the animals. No treatment-related death occurred.

Fig. 3

Tumor volume charted against number of days posttreatment, to assess the tumor response in various treatment groups. The combination therapy group of PDT and bevacizumab (Bmab) exhibited the greatest tumor response in comparison with all other groups. Each group represents the mean response [bars, standard error (SE)] of 10 animals.

To further confirm the results from the tumor regression experiments, investigations were conducted at the vascular and molecular levels to determine the mechanism of the angiogenesis process following treatment.

3.2.

Visualization of Tumor Vessel Using Laser Confocal Endomicroscopy

Tumor blood vessels were visualized using confocal endomicroscopy approximately $30\min$ after intravenous administration of FITC-dextran. Control [Fig. 4 ] blood vessels were well developed, dense, and intact while those in treated groups [Figs. 4, 4, 4] showed varying degrees of vessel damage following treatment. Tumors treated with bevacizumab only exhibited lower vessel density; however, normalized morphology was noticed in the remaining blood vessels that could possibly allow blood flow to tumors. The PDT and bevacizumab combination treatment group showed the most posttreatment damage, with shrunken vessels and leaky vessel walls resulting in a diffuse fluorescence background.

Fig. 4

Tumor blood vessels visualized approximately $30\min$ after intravenous administration of FITC-dextran in (a) control, (b) PDT only, (c) bevacizumab only, and (d) PDT and bevacizumab treated groups. Imaging was performed on the animals from day 25 to day 90, based on the maximum tumor volume limit or the completion of the antiangiogenesis treatment. Control blood vessels were well developed, dense, and intact while those in treated groups showed varying degrees of vessel damage following treatment. The PDT and bevacizumab combination treatment group showed the most posttreatment damage, with shrunken vessels and leaky vessel walls resulting in a diffuse fluorescence background.

3.3.

Analysis of Mean Vessel Area

Mean vessel areas were calculated from the endomicroscope images of the tumors. Compared to the control group, a 1.3-fold decrease in mean vessel area $(p<0.05)$ was observed in the bevacizumab-only group and a three-fold reduction of vessel area $(p<0.01)$ was noticed in the PDT and bevacizumab combination therapy group (Fig. 5 ).

Fig. 5

Graph depicting the mean vessel area in the different treatment groups. A 1.3-fold decrease of vessel area was observed in bevacizumab only group, while the PDT and bevacizumab (Bmab) combination therapy group exhibited threefold reduction in mean vessel area compared to the control group.

3.4.

Detection of VEGF in Tumor Tissue Using Immunohistochemistry and IF

VEGF expression was assessed in the xenograft bladder tumor model using immunohistochemistry and IF techniques. All of the tumors exhibited cytoplasmic staining for VEGF. Maximum VEGF staining of 27 to 31% (IHC score 3) was noticed in PDT only and control tumor tissue. Around 16 to 20% (IHC score 2) scattered staining in certain regions of the cytoplasm was observed in the bevacizumab-only treated group and light staining of 6 to 8% (IHC score 1) was found in the combination therapy group (Figs. 6 and 7 ).

Fig. 6

VEGF expression assessed in tumor sections using immunohistochemistry. The brown-colored regions in the images indicate VEGF staining in (a) control, (b) PDT only, (c) bevacizumab only, and (d) PDT and bevacizumab treated tumors. Maximum VEGF staining of 28 to 30% (IHC score 3) was noticed in the cytoplasm of PDT and control tumors. Around 16% (IHC score 2) concentrated staining was observed in the bevacizumab only group, and light staining of 6% (IHC score 1) was found in the combination therapy group.

Fig. 7

IF was performed to confirm the immunohistochemistry results. In the confocal images, the green FITC fluorescence staining indicates the expression of VEGF. Intense staining of 27 to 31% was observed in (a) control and (b) PDT only groups; and 8% staining were recorded in the (c) bevacizumab only and (d) PDT and bevacizumab groups, respectively. The results obtained from IHC and IF staining techniques were comparable for all the groups.