Quantitation of Intra-peritoneal Ovarian Cancer Metastasis

Summary

Ovarian cancer metastasis is characterized by numerous diffuse intra-peritoneal lesions, such that accurate visual quantitation of tumor burden is challenging. Herein we describe a method for in situ and ex vivo quantitation of metastatic tumor burden using red fluorescent protein (RFP)-labeled tumor cells and optical imaging.

Abstract

Epithelial ovarian cancer (EOC) is the leading cause of death from gynecologic malignancy in the United States. Mortality is due to diagnosis of 75% of women with late stage disease, when metastasis is already present. EOC is characterized by diffuse and widely disseminated intra-peritoneal metastasis. Cells shed from the primary tumor anchor in the mesothelium that lines the peritoneal cavity as well as in the omentum, resulting in multi-focal metastasis, often in the presence of peritoneal ascites. Efforts in our laboratory are directed at a more detailed understanding of factors that regulate EOC metastatic success. However, quantifying metastatic tumor burden represents a significant technical challenge due to the large number, small size and broad distribution of lesions throughout the peritoneum. Herein we describe a method for analysis of EOC metastasis using cells labeled with red fluorescent protein (RFP) coupled with in vivo multispectral imaging. Following intra-peritoneal injection of RFP-labelled tumor cells, mice are imaged weekly until time of sacrifice. At this time, the peritoneal cavity is surgically exposed and organs are imaged in situ. Dissected organs are then placed on a labeled transparent template and imaged ex vivo. Removal of tissue auto-fluorescence during image processing using multispectral unmixing enables accurate quantitation of relative tumor burden. This method has utility in a variety of applications including therapeutic studies to evaluate compounds that may inhibit metastasis and thereby improve overall survival.

Introduction

Epithelial ovarian cancer (EOC) is the most common cause of death from gynecologic malignancy, with an estimated 21,290 new diagnoses in the U.S. in 2015 and an estimated 14,180 deaths¹. The vast majority (> 75%) of women are diagnosed with late-stage disease (stage III or IV) characterized by diffuse intra-peritoneal metastasis and poor prognosis. Disease recurrence in the peritoneal cavity following first-line chemotherapy is also common and represents a major cause of mortality^2,3. EOC metastasizes by an unique mechanism involving both direct extension from the primary tumor to neighboring peritoneal organs as well as by dissociation or shedding of cells from the primary tumor surface as single cells or multi-cellular aggregates. Cells are shed into the peritoneal cavity, wherein they resist detachment-induced apoptosis⁴. Buildup of peritoneal ascites is common, as shed tumor cells block peritoneal lymphatic drainage and tumors produce growth factors that alter vascular permeability. A portion of shed tumor cells attach to the surface of peritoneal organs and structures including intestine, liver, omentum and mesentery, whereupon they anchor and proliferate to produce multiple widely disseminated secondary lesions^3,5. Hematogenous metastasis is uncommon. Thus, clinical management commonly consists of cytoreductive surgery including "optimal debulking", defined as resection of all visible tumor (no matter how small). Complete cytoreduction is associated with a significant increase in overall survival^6,7 and is associated with the challenge of identification and removal of lesions < 0.5 cm.

Small animal models have proven utility in ovarian cancer research in improving our understanding of disease progression as well in identification of prognostic biomarkers and testing of novel chemotherapies or combination therapy approaches. As the primary site of ovarian cancer incidence and metastasis is the peritoneal cavity, orthotopic models of EOC metastasis involve the analysis and characterization of intraperitoneal disease. Although there have been recent improvements in the ability to image tumor cells, even at the single cell level, there still exist significant difficulties in quantifying the metastatic tumor burden of EOC. These challenges arise due to the number, size and anatomic location of metastatic lesions. Furthermore there exists a need to label cancer cells to distinguish them from normal host cells. Previous studies have utilized antibody-based labeling protocols or transfection of tumor cells with luciferase^8,9. Direct fluorescent labeling of cancer cells was first reported by Chishima and coworkers in 1997¹⁰. Fluorescent labels do not require addition of exogenous substrate and provide exquisite tumor cell specificity, providing a more effective means to track cancer metastasis^11,12.

Herein we describe an optical imaging method for quantitative analysis of metastatic disease using a syngeneic orthotopic xenograft model comprised of red fluorescent protein (RFP)-tagged murine ID8 ovarian cancer cells¹³ and immuno-competent C57/Bl6 mice. We demonstrate a novel method of relative tumor burden quantification combining in vivo and ex vivo imaging with removal of tissue auto-fluorescence. This approach has potential utility in studies designed to evaluate the effect of specific genetic, epigenetic or micro-environmental modifications and/or treatment modalities on organ-specific metastasis of ovarian cancer.

Protocol

All in vivo studies were approved by the University of Notre Dame Animal Care and Use Committee and used female C57/BL6J mice.

1. Murine Ovarian Cancer Cell Culture

1. Make the ID8 murine ovarian cancer cell culture medium as follows: 1 L of Dulbecco's Modified Eagle Medium (DMEM) supplemented with 4% Fetal Bovine Serum (FBS), 1% Penicillin/Streptomycin, 5 µg/ml Insulin, 5 µg/ml Transferrin and 5 ng/ml Sodium Selenite.
2. Transduce ID8 murine ovarian cancer cells¹³ to express Red Fluorescent Protein (RFP) using commercially procured lentiviral particles expressing RFP and a selection marker (blasticidin gene). Note that green fluorescent protein can be used, but provides a weaker fluorescence signal.
   1. Immediately prior to transduction, culture cells to 50 - 70% confluence and add fresh medium (without penicillin/streptomycin) that includes 1 ml polybrene (5 µg/ml final concentration).
   2. Gently pipette RFP-lentivirus (20 µl) dropwise into the dish and incubate at 37 ^oC for 72 hr. Remove medium and incubate with fresh medium for 24 hr. Begin selection of transduced cells by adding blasticidin to the culture medium (5 µg/ml) and monitor for red cells using fluorescence microscopy (Figure1).
3. Culture the RFP-tagged cells in ID8 medium by seeding approximately 10⁶ cells into a tissue culture dish and visualizing daily using a light microscope until the cell monolayer is confluent.
4. When confluent, harvest the cells using trypsin (0.25%, 3 min) and return the flask to the 37 ^oC incubator until the cells are detached.
5. Neutralize the trypsin with 6 ml of ID8 culture medium containing serum. Pipette up and down against the bottom of the dish, then transfer to a 15 ml conical tube. Centrifuge at 1,200 RPM for 2 min, then aspirate the medium using a glass pipette.
6. Wash the cells with Phosphate Buffered Saline (PBS) by gently resuspending in 6 ml warm PBS and centrifuging as in 1.5 above. Re-suspend in PBS to a final concentration of 5 x 10⁶ cells/ml. Keep cells warm (37 ^oC) until injection.

2. Intra-peritoneal Injection of ID8 Cells and in vivo Imaging

1. Inject each mouse with 2 mL of the warm ID8-PBS solution for a total of 10 million cells through intra-peritoneal (i.p.) injection. Allow tumor cells to grow in vivo for approximately 10 weeks with weekly live imaging.
2. Immediately following injection and each week thereafter, weigh the mice and then anesthetize each mouse using isofluorane (2.5% flow rate in 0.5 L/min O₂ flow). Note: The mice were weighed, not as a quantitative measure of tumor burden, but rather a more general measure of individual mouse physical progression throughout the study.  Mice weights were compared with their own previous weight values.  
   1. Anesthetize mice by placing in an isofluorane chamber. Maintain constant anesthesia while scanning by positioning the mouse head in the nose-cone for exposure to 2.5% flow rate isofluorane throughout the procedure.
3. Use depilatory cream to remove the hair on the torso and abdomen of each mouse, as the black hair will cause attenuation of the fluorescence signal. Mice are bathed with warm water after application of the depilatory cream and dried with RT paper towels.  
4. While anesthetized, scan the entire body of each mouse using a small animal optical imaging system. Scan all mice in a cohort at each time point. Use the following two step protocol (Figure 2A):
   1. In Step 1, to observe fluorescently labeled (RFP) tumor cells, execute multispectral acquisition to collect a total of 5 images using excitation filters of 440 nm, 460 nm, 480 nm, 520 nm and 540 nm, and an emission filter of 600 nm. Use the following acquisition parameters: a standard exposure of 15 sec, 2 x 2 binning, FOV of 160 mm, and f_stop of 1.1.
   2. In Step 2, to observe the whole animal, acquire a reflectance image using an open excitation filter (white light), an open emission filter, a standard exposure of 0.2 sec with 2 x 2 binning and an FOV of 16 cm.  Note: The actual imaging time is ~1 min. 

3. Mouse Dissection and Image Acquisition

1. When live imaging shows the presence of extensive intra-peritoneal disease or animals exhibit accumulation of ascites (as evidenced by abdominal distension), loss or gain of body weight greater than 20%, lethargy or other signs of distress, euthanize each mouse using CO₂ anesthesia followed by cervical dislocation.
2. Cut the skin anteriorly along the ventral midline of the mouse, using pointed dissection scissors, from the top of the thighs to the bottom of the rib cage. Take extreme care to cut through only the skin layer (Figure 2Bi). Gently separate the skin from the peritoneal tissue, being sure to not puncture the peritoneal wall.
3. Using pointed dissection scissors, cut laterally both along the bottom of the rib cage and the top of the thighs to create two flaps of skin (Figure 2B ii,iii).
4. Place each mouse on its ventral side (peritoneal cavity facing downward) with the skin flaps pulled open and scan each mouse with its organs in situ using the small animal optical imaging system (Figure 3A,B).
   1. Use the same image acquisition parameters as described in 2.4.1.-2.4.2.
5. Continue dissecting the mouse by extracting individual peritoneal organs including liver, omentum/pancreas, stomach, diaphragm, small intestine, large intestine, left and right peritoneum, ovaries, kidneys, mesentery and fat pad.
   1. Observe, enumerate, and record visible tumors on each organ. Use a caliper to measure the diameter of each tumor.
   2. Designate tumors less than 2 mm in diameter as small, in between 2 and 5 mm as medium and greater than 5 mm as large.
6. Place the organs on the organ scanning template (Figure 4) that identifies a predetermined location for each organ. Use PBS to keep the organs moist.
7. Scan each sheet of organs using the small animal optical imaging system (Figure 3C,D).
   1. Use the same image acquisition parameters as described in step 2.4.
8. After scanning, place organs in 10% formaldehyde to enable subsequent processing for histology.

4. Quantification of Tumor Burden in the Fluorescence Images

1. In order to reduce the amount of auto-fluorescence in each multispectral scan, first unmix the files using software available with the small animal imaging system.
   1. Load first the RFP: In Vivo spectral file, followed by the Autofluor: In Vivo Red file in the Unmixed Images window and select Unmix.
2. Export the .bip files as 16 bit .tiff (unscaled) format .
3. Use ImageJ software (ex vivo organs and then the full mouse body with organs in situ).
   1. Use File > Open to open the 16 bit .tiff files of the organ scanning template files in ImageJ.
   2. Under Image > Adjust > Brightness/Contrast, choose to Set the Minimum displayed value to 0 and the Maximum displayed value to +35353 and then under Image > Lookup Tables, select Red Hot. Note: Various animal models will require different brightness levels, but it is important to keep the brightness consistent throughout a given study.
   3. Using the Freehand Selection tool, draw a free-form region of interest selection (ROI) around each organ and use the Measure tool (CTRL+M) to calculate the surface area of each organ; record the surface Area of each organ in a spreadsheet.
   4. With the ROI drawn around each organ, right click within the ROI and select Duplicate to include only the organ.
   5. While looking at the individual organ, under Image > Adjust > Threshold, choose to Set the threshold of the image to a Lower Threshold Level of + 1200 and an Upper Threshold Level of + 1700 to select only the most brightly fluorescing regions and check to select Dark Background; select OK. Note: The images may require manual manipulation of the threshold sliders in ImageJ so that the previously brightest areas are selected for the analysis step, as shown above.  Different disease models will require different threshold constraints, but it is important to keep the threshold levels consistent throughout a given study.
   6. Under Analyze > Analyze Particles, change the Size (pixels ^ 2) to 10-Infinity, choose to Show: Outlines, check to select only "Display results" and click OK to measure both the areas and raw integrated densities of these bright regions, deemed tumors.
   7. Record the values of the surface Area and the Raw Integrated Density of each tumor selection displayed in the Results table in the same spreadsheet containing the organ surface areas.
   8. Under Analyze > Tools > Calibration Bar, add a calibration scale bar to the images of the organs. Note: In this example, this was done by changing the Location to Upper Right, the Fill Color to Black, the Label Color to White, the Number of Labels to 5, the Decimal Places to 0, the Font Size to 12 and the Zoom Factor to 4.0 and enabling Bold Text.
   9. Under File > Save As…, save the montage as a .Tiff and .Jpeg.
   10. Use File > Open to open the .jpeg file of the organs and then under Insert > Text Box, add organ labels by creating a text box below each of the three rows of organs.
   11. Use File > Open to open each of the 16 bit .tiff files of the full body scans in ImageJ in order of mouse number.
   12. Under Image > Adjust > Brightness/Contrast, choose to Set the Minimum displayed value to 0 and the Maximum displayed value to + 35353 and then under Image > Lookup Tables, select Red Hot.
   13. Under File > Save As…, save the montage as a .Tiff and .Jpeg.

5. Data Analysis

1. Add the tumor selection areas and raw integrated densities, respectively, for each organ of each mouse and record the total organ tumor area for each mouse.
2. Normalize the recorded area and raw integrated density values for the size of each organ by dividing the total tumor area and total raw integrated density values by the corresponding organ surface area (Table 1).
3. Calculate and plot the average of both the normalized tumor surface areas and the normalized raw integrated density values (Figure 5A,B).
4. Compare these average values to those obtained by visually counting the tumors that were visible to the naked eye during the inspection of each organ for metastatic lesions.

Results

The metastatic mechanism of ovarian cancer is characterized by highly diffuse intra-peritoneal metastasis comprised of numerous lesions of varying size, including multiple small (< 2mm) lesions. Thus, use of RFP-labelled tumor cells (Figure 1) and optical imaging provides an alternative method to manual counting and measurement of lesion size. The development of tumor burden over time can be determined by weekly weighing of mice and measurement of abdominal girth to g...

Discussion

In contrast to studies using human ovarian cancer cells that must be conducted in immunocompromised mice, the protocol described above utilizes immunocompetent C57/Bl6 mice and syngeneic murine ovarian cancer cells. While this enables assessment of the potential role of immune infiltrates in tumor progression and metastasis, the presence of dark hair on the abdominal surface renders imaging less sensitive. Use of a depilatory to remove hair prior to imaging enhances image acquisition, but is time-consuming, particularly ...

Materials

Name	Company	Catalog Number	Comments
Dulbecco's Modified Eagle Medium	Corning	10-014-CM
Fetal bovine serum	Gibco	10437-028
penicillin/streptomycin
Insulin-transferrin-sodium selenite media supplement	Sigma	I-1884
Bruker Xtreme small animal imaging system	Bruker Corp.
Bruker Multispectral software	Bruker Corp
lentiviral particles with Red fluorescent protein	GenTarget, Inc.	LVP023
trypsin for cell culture	Corning	25-053-CI
PBS	Corning	21-040-CM
depilatory cream (such as Nair Hair Remover Lotion)	purchases from drugstore	n/a
ImageJ software	http://imagej.nih.gov/ij/		free download
dissecting tools (forceps)	Roboz Surgical Instrument	RS 5130
dissecting tools (Scissors)	Roboz Surgical Instrument	RS 5910