Clinically Relevant Transplantable Tumor Models for Studying Metastatic Ovarian Cancer and Therapeutic Response in Mouse

Summary

Here we present a step-by-step protocol for establishing the intraperitoneal and the intrabursal tumor models to study ovarian cancer cancer progression, metasatsis, and therapeutic response. Further, we describe how to use luciferase-based non-invasive in vivo imaging in these models to monitor tumor progression in mouse.

Abstract

Metastatic ovarian cancer is an extremely lethal gynecologic malignancy that lacks effective treatment options. Development of clinically translatable therapeutic approaches for ovarian cancer patients is contingent on understanding factors that drive ovarian cancer metastasis. While most solid tumors exhibit hematogenous metastasis, ovarian cancer cells spread primarily through the intraperitoneal route. This process involves detachment and shedding of cancer cells from the primary tumor into the peritoneal cavity, which then acquire properties that enable them to resist anoikis and form multicellular aggregates. These cell aggregates then attach to and invade the mesothelial lining of the peritoneum, forming metastatic lesions. Clinically relevant in vivo models are essential for understanding how cellular factors and the tumor microenvironment impact the different steps in the metastatic process and for testing novel therapeutic strategies. Two transplantable mouse tumor models widely used to study ovarian cancer metastasis and therapeutic response are the intraperitoneal and intrabursal models. In this report, we describe the step-by-step methodology for establishing these models and highlight their advantages and limitations. Additionally, we describe how to use bioluminescent imaging to non-invasively monitor tumor progression using these models.

Introduction

Despite significant progress in our understanding of factors that drive ovarian cancer progression, ovarian cancer remains the second most lethal gynecologic malignancy in the United States¹. Due to a lack of obvious symptoms and the absence of reliable early diagnostic markers, over 70 percent of ovarian cancer patients have advanced metastatic disease at diagnosis^2,3. Current therapeutics are not effective in treating metastatic ovarian cancer and consequently, the 5-year survival rates for patients with advanced-stage disease remain less than 30 percent⁴. A major cause for the lack of therapeutic success is the emergence of chemoresistance⁵. Therefore, there is a pressing need to identify factors that drive ovarian cancer metastasis and chemoresistance. Addressing this knowledge gap will be critical for developing more effective therapeutic strategies to target this disease. To achieve these goals, we need in vivo models that accurately mimic the key steps of ovarian cancer progression.

While ovarian cancer cell lines and three-dimensional in vitro models have contributed immensely towards our understanding of factors that impact metastasis and therapeutic response, these models have several limitations^6,7,8. For instance, these models do not account for the different non-cancerous cells that are part of the tumor microenvironment, nor capture the complex interactions that happen within the tumor microenvironment^6,7,8. Further, it is extremely challenging to study systemic factors and their impact on tumor progression and drug response using in vitro models^6,7,8. These limitations of in vitro models highlight the need for in vivo models that accurately mimic the various steps in ovarian cancer metastasis, including complex interactions with other cell types and factors within its microenvironment.

Genetically engineered mouse models (GEMMs) offer an in vivo platform wherein tumors arise spontaneously in a natural microenvironment with an intact immune system⁹. However, they are time-consuming, challenging, and costly to generate and maintain¹⁰. Most GEMMs of ovarian cancer also have a long latency period for tumor development, which makes working with them time-consuming and costly, even after they have been generated¹⁰. Transplantable models, such as xenograft and allograft models, offer a good balance between biological relevance and experimental ease, and are often much less expensive to generate than GEMMs¹¹. In addition to having lower latency than GEMMs, the ease with which tumor cells can be genetically modified in vitro makes these models convenient and suitable for studying how different genetic factors impact tumor progression¹⁰. Due to these benefits, transplantable tumor models have emerged as valuable tools for both basic and translational research in ovarian cancer.

In this report, we describe the protocol for establishing two different transplantable mouse tumor models that are widely used to study metastatic ovarian cancer: (1) the intraperitoneal model and (2) the orthotopic or intrabursal model. Both models can be established as either xenograft or allograft models. In the intraperitoneal model, cancer cells are injected into the peritoneal cavity of the mice. Introducing ovarian cancer cells into the intraperitoneal space mirrors the stage in ovarian cancer metastasis where metastatic ovarian cancer cells are shed from the primary tumor into the peritoneal cavity. Similar to the metastatic process in humans, these transplanted ovarian cancer cells will need to adapt to the hostile peritoneal microenvironment, attach and invade the mesothelial layer that lines the peritoneal cavity, and form metastatic lesions on intraperitoneal organs and tissues such as the omentum, mesentery, liver, and diaphragm. Depending on the cancer cell line used to generate the intraperitoneal model, metastatic tumors can be accompanied by ascites, as observed in late-stage disease in humans. Thus, the intraperitoneal model has been used widely to understand factors that drive ovarian cancer metastasis and their progression into late-stage disease¹². These models are also useful to determine the efficacy of novel therapeutic strategies for metastatic ovarian cancer¹². In the intrabursal or orthotopic model, cancer cells are implanted into the bursa, a membranous structure that surrounds the mouse ovary. This model mimics early stages of ovarian tumorigenesis and metastasis. At this stage cancer cells are still confined to the fallopian tube and the ovary. It is from these sites that cancer cells are shed into the peritoneal cavity, ultimately metastasizing to other organs. Therefore, the intrabursal model is useful in studying both early stages of ovarian tumorigenesis as well as events that lead to their subsequent transcoelomic metastasis through the peritoneal cavity¹².

Thus, while the both intraperitoneal and intrabursal models are excellent tools to study ovarian cancer metastasis and therapeutic response, they differ in the specific stage along the metastatic process that they mimic. While the intrabursal models can mimic the progression of ovarian cancer from early stages to later stages in the metastastic process, due to the latency associated with this process, they are usually used to study early events in ovarian cancer metastasis. For studies focusing on later stages of ovarian cancer metastasis, the intraperitoneal model is more suitable. Further, the intrabursal model is also technically more challenging to establish compared to the intraperitoneal model¹². Establishing the intrabursal model requires access to a surgical suite equipped with microsurgical equipment that will allow precise injection of cancer cells into the ovarian bursa. The ability of cancer cells to establish tumors (“take rate” or “engraftment rate”) is also usually lower in intrabursal models as compared to intraperitoneal models. Moreover, as tumors cannot be easily observed after orthotopic or intraperitoneal injections, a non-invasive bioluminescent imaging technique is typically employed to monitor tumor progression. Therefore, both intraperitoneal and intrabursal models require additional equipment that will allow non-invasive imaging of tumors¹². Ultimately, the experimental goals, availability of resources, and technical skill set will determine the choice of model for each study.

This report describes the step-by-step protocol for establishing intraperitoneal and intrabursal mouse xenograft models using human ovarian cancer cell lines. These models can be established by transplanting the cells into immunocompromised mouse strains such as NSG (NOD.Cg-Prkdc^scid Il2rg^tm1Wjl/SzJ) or the Foxn1 nude (B6.Cg-Foxn1 nu/J). The same method can be adapted to generate allograft (syngeneic) models. While establishing allograft models, mouse ovarian cancer cells are implanted into immunocompetent mice of the same genetic background. As allograft models are developed using immunocompetent mice, they are useful for studying how different factors regulate the ovarian tumor microenvironment, the role of the immune system in ovarian cancer progression, and the effectiveness of immunotherapy. In addition to establishing the intraperitoneal and intrabursal xenograft mouse models, this report also describes how to use bioluminescent imaging to non-invasively monitor tumor progression in these models.

Protocol

All animal procedures described in this protocol have been approved by the Institutional Animal Care and Use Committee (IACUC) at the University of Maryland, Baltimore County (Principal Investigator - Achuth Padmanabhan; Protocol number - 1947).  The complete list of materials and equipment used in this protocol and their catalog numbers are provided in Table of Materials.

1. Preparing ovarian cancer cell lines for transplantation in mouse

Note:For all the in vivo models described in this protocol, ovarian cancer cells should be prepared for transplantation into mouse. Ovarian cancer cells should be prepared in a laminar flow tissue culture hood as described below at room temperature unless specified:

1. Grow ovarian cancer cells at 37 ^oC in a water jacketed CO₂ incubator maintained at 5% CO₂ using standard media as recommended in the literature. If non-invasive in vivo imaging of the tumor is desired, the ovarian cancer cells has to be engineered to stably express luciferase prior to transplantation.
2. When cells reach around 80%–85% confluency, harvest them under sterile conditions, as described below:
   1. Aspirate media from the tissue culture plate.
   2. Wash adherent cells twice with sterile 1X Phosphate-Buffered Saline (PBS) without calcium and magnesium. The presence of Calcium and Magnesium in PBS can interfere with trypsin activity and will promote cell adhesion. Aspirate the PBS from the tissue culture plate.
   3. Add 0.05% Trypsin-EDTA to the tissue culture plate. The amount of trypsin to be used will depend on the size of the tissue culture plate and the cell line being worked on. For most cell lines growing on a 100 mm plate, use 1 mL of 0.05% Trypsin-EDTA. For strongly adherent ovarian cancer cell lines such as OVCA420, 1 mL of 0.25% Trypsin-EDTA is recommended.
   4. Transfer the plate with cells back to the 37 ^oC water jacketed CO₂ incubator used for tissue culture and incubate for 1–5 min. The exact incubation time required will depend on the cell line. Some cell lines, such as TYK-Nu and HEYA8, detach very quickly (1–2 min), whereas others, like OVCA420, take longer (4–5 min).
   5. Once the cells are detached, add FBS-containing media to stop trypsin action. Add 6 mL FBS-containing media per 1 mL of trypsin. Gently pipette up and down to dislodge cells from the tissue culture plate and dissociate them into a single-cell suspension.
      Note: When viewed under a bright-field inverted microscope, detached cells will appear rounded (instead of flat/elongated) and float freely rather than being stuck to the bottom of the plate.
   6. Collect the cells in a conical tube and centrifuge at 400 ​​x g for 5 min at room temperature.
   7. Remove trypsin-containing supernatant and resuspend the cell pellet in fresh FBS-containing cell culture media.
3. Count cells using a cell viability dye such as trypan blue and a hemocytometer. Alternatively, an automatic cell counter can be used to count cells.
4. Reconstitute the cells in a defined volume of sterile 1X PBS (pH 7.4) such that the desired concentration of cells per unit volume is achieved. The recommended number of cells to inject for commonly used human ovarian cancer cell lines is shown in Table 1¹³.
   Note: The maximum volume that can be safely injected into mice depends on the model. For the intraperitoneal model, the maximum recommended volume is 500 µL (see section 2.1). For the intrabursal model, the maximum recommended volume is 5 µL (see section 2.2).
5. Keep the cell suspension on ice until it is time to inject it into mice. Inject cells into mice as soon as possible, preferably within 30–45 min of preparing the cells.

2. Establishing transplantable ovarian cancer models in mouse

1. Establishing the intraperitoneal model of ovarian cancer
   Note: In this process, ovarian cancer cells stably expressing luciferase are injected into the peritoneal cavity of an adult female mouse. Using cells that stably express luciferase will enable longitudinal, non-invasive monitoring of tumor progression with the In Vivo Imaging System (IVIS).
   1. Prepare the cells as described in section 1. Resuspend the cells gently using a 1000 µL micropipette and ensure the cells remain dissociated.
      Note: Failure to dislodge the clumps will cause the needles to clog.
   2. Draw 500 µL of cell suspension into a 1 mL syringe and attach a 21G X 1 inch needle. Ensure that there are no air bubbles in the syringe.
      Note:The maximum volume that can be injected intraperitoneally into an adult female mouse (25–30 g) is 500 µL. Therefore, resuspend cells in a volume such that the required number of cells is contained in a maximum volume of 500 µL. The recommended number of cells to inject for commonly used human ovarian cancer cell lines is shown in Table 1¹³.
      Note: If the mouse weighs between 20–25 g, injecting a smaller volume (less than 400 µL) is recommended.
   3. Remove the animal from the cage and put it on a rough or wired surface to restrain it.
   4. Restrain the mouse by gently scruffing the loose skin at the base of the neck using the non-dominant hand as shown in Figure 1A.
   5. Gently pick the mouse up and turn the mouse such that its ventral side faces upwards. Secure the animal’s tail with fingers as shown in Figure 1B.
   6. Draw an imaginary line across the middle of the abdomen. To locate the injection site, imagine a triangle between the midline and the two nipples in the lower right quadrant. Ensure that the injection site is in the middle of this imaginary triangle. Refer to Figure 1B.
   7. Tilt the mouse’s head slightly towards the ground, making it lower than the hind end. Restraining the mouse in this position helps the abdominal contents shift cranially. This reduces the risk of puncturing internal organs during injection.
   8. Use a sterile alcohol swab to clean the injection area.
   9. Using the dominant hand, insert the needle into the injection site at an angle of 45° or lower. Advance only 0.5 inch and not more than half of the length of the needle. Ensure that the needle remains stable, as movement in the abdominal cavity can cause injury to internal organs.
   10. Dispense the cells at a slow and steady rate into the abdominal cavity of the mouse.
   11. Pull the needle out of the mouse at the same angle of insertion. Immediately place a sterilized gauge over the injection site to prevent any leakage.
   12. Dispose the needle and the syringe in the appropriate sharps container.
   13. Before putting the mouse in the cage, ear tag them using a preferred method. Ear tagging will allow indentification of the injected mouse later.
2. Establishing the intrabursal model of ovarian cancer
   Note:In the intrabursal model, ovarian cancer cells that stably express luciferase is injected directly into the ovarian bursa of an adult female mouse. Bursa is a membranous sac that surrounds the murine ovaries. Stable luciferase expression in ovarian cancer cells will allow non-invasive longitudinal monitoring of tumor progression using IVIS.
   1. Day 1 – Preparing equipment and mouse for the intrabursal transplantation surgery
      1. Autoclave the instruments needed for surgery. The list of instruments is (Figure 2A):
         Fine-tipped scissors, 5-inch curved fine-point forceps, Mayo-Heger Needle Holder,Dissecting jeweler microforceps, fine tip, Instruments Dissecting Forceps with Fine Points
      2. Administer pre-operative analgesia of carprofen (15 mg/Kg) every 24 h by subcutaneous injection in the mouse.
         Note: Administer the analgesic carprofen subcutaneously every 24 h, starting 1 day before surgery and continuing for 3 days post-surgery.
   2. Day 2 – Intrabursal transplantation of ovarian cancer cells in a mouse
      1. Prepare ovarian cancer cells for injection as described in section I. The maximum volume that can be safely injected intrabursally in a mouse is 5 µL. Therefore, it is preferable to have cells at a concentration of 10⁴ cells/µL. This will allow transplantation of 5 x 10⁴ cells.
         Note: The recommended number of cells to inject for commonly used human ovarian cancer cell lines is shown in Table 1¹³.
      2. Keep the cells on ice until use. Once harvested, it is recommended to inject cells as soon as possible, preferably within 30–45 min.
      3. Induce anesthesia by placing the mouse in an isoflurane chamber with an isoflurane flow rate of 4 percent and an oxygen flow rate of 3.5 L/min. Confirm the mouse is anesthetized by checking for immobilization and the absence of the toe pinch reflex.
      4. Transfer the anesthetized mouse onto a sterile surgical drape that is placed on top of a rodent heating pad maintained at 37 °C. Quickly place a nose cone and maintain anesthesia with 1.5% isoflurane and 98.5% oxygen. Ensure the mouse is anesthetized by confirming the absence of the toe pinch reflex.
      5. Ensure the mouse remains anesthetized throughout the procedure by confirming the absence of the toe-pinch reflex every 15 min. Monitor the mouse's breathing rate during anesthesia to ensure it is slow and smooth. If the breathing rate changes, adjust the isoflurane dosage. If the mouse exhibits labored breathing or gasping, decrease the dose. Conversely, if the mouse exhibits an increased breathing rate, increase the isoflurane dose slightly.
         Note: Avoid keeping the mouse under anesthesia for longer than 40 min. To achieve this, it is important to be efficient and quick during the surgical procedure.
      6. Position the mouse with the side in which the ovarian cancer cells are to be transplanted intrabursally, facing up as shown in Figure 2B.
      7. Apply eye ointment to both eyes of the mouse to protect them from drying out (corneal desiccation).
      8. If the mouse has fur, shave the hair on the lower half of the mouse on the side that is facing up. Clean the shaved area properly and place the mouse on a new sterile pad. If working with the Foxn1 nude (B6.Cg-Foxn1 nu/J) mouse strain, skip this step.
      9. Use a sterile surgical drape to cover the mouse in such a way that only the lower half of the mouse's body is revealed.
      10. Wash hands properly with soap before the start of the surgery and put on a fresh pair of sterile surgical gloves. Avoid touching any non-sterile areas or items after putting on surgical gloves.
      11. Use a sterile cotton tip applicator to apply betadine liquid at the surgical site in a clockwise circular motion from the inside out. Next, use a sterile alcohol swab to clean the same area in a similar fashion. Repeat this procedure two more times.
          Note: Do not touch non-sterile items or areas during the surgical procedure. Replace gloves immediately if contact occurs with a non-sterile surface. It is highly recommended to have another person to help while performing intrabursal injection. This person can help handle non-sterile items during surgery.
      12. Wait 1 min for the alcohol to dry.
      13. Locate the site that is approximately 1.5 cm above the hind leg and 5 mm from the rib cage of the mouse, as shown in Figure 2B. Lift the skin at this site with the curved forceps, then make a 5 mm vertical incision with a fine-tipped scissor.
          Note: Before making the incision, ensure that the mouse is completely under the influence of anesthesia by confirming the absence of the toe pinch reflex.
      14. Hold the skin on one side with fine-tip microforceps and make an incision in the subcutaneous tissue at the same site.
      15. At this point, the peritoneal layer under the subcutaneous tissue will become visible. Use a fine-tipped, curved, serrated forceps to hold the peritoneal layer, and make an incision about 2.5 mm long with fine-tipped scissors.
      16. Hold the side of the peritoneal incision flap with the serrated forceps and locate the ovarian fat pad using fine forceps. The ovary is located under the fat pad.
      17. Gently pull out a little bit of the ovarian fat pad along with the ovary through the incision.
      18. Use the dissecting forceps to hold the fat pad such that the ovary is facing up and is immobile.
          Note: Do not hold the ovary with forceps. Always use serrated forceps to hold the fat pad because the tissue is slippery.
      19. Ask the person assisting to get the syringe containing the cell suspension ready. It is recommended to use a 0.3 mL insulin syringe.
      20. Gently insert the needle with the bevel facing up into the ovarian bursa from the junction of the fallopian tube and ovary. Make sure the entire bevel goes inside the bursa before pressing on the plunger.
          Note: Be careful not to puncture the bursa at the opposite end of the insertion site.
      21. Gently push the plunger and transfer the contents of the syringe into the ovarian bursa. Remove the needle quickly.
      22. Swelling of the bursa surrounding the ovary will indicate successful inoculation of cells.
      23. Wait for 5 s and gently put the ovary back into the peritoneal cavity.
      24. Close the peritoneum using a 6-0 absorbable monofilament suture with one or two simple interrupted surgical stiches.
      25. Close the skin incision using the 6-0 absorbable monofilament suture with a simple running surgical stich. Ensure there is no bleeding from the incision site.
      26. Apply triple antibiotic ointment (Bacitracin, Neomycin, and Polymyxin B) at the incision site after closing it.
      27. Administer Carprofen 15 mg/Kg subcutaneously to the mouse for pain relief.
      28. Ear tag the mouse using a preferred method. Ear tagging will allow indentification of the injected mouse later.
      29. Transfer the mouse to a fresh cage. Place this cage with one-half on a heat pad for at least 60 min. Ensure that the effect of anesthesia has completely worn off by this time, and the mouse is awake and active.
          Note: Mice are prone to hypothermia under anesthesia. Hence, it is important to provide external heat to regulate body temperature.
      30. Check the mice after 4-5 h post-surgery for any signs of bleeding from the incision site or other discomfort.
      31. Give mice carprofen 15 mg/Kg subcutaneously every 24 h for the next 3 days.
      32. If the inoculated cells are genetically engineered to stably express luciferase, then tumor burden can be measured using IVIS from Day 1 post-injection as described in section 3.

3. Non-invasive imaging of a mouse tumor using the In Vivo Imaging System (IVIS)

An important objective in several studies using the transplantable tumor models described in this report is to obtain longitudinal data on tumor progression. While the volume of subcutaneous tumors can be estimated with a vernier caliper, this is not a viable option for tumors formed using the intraperitoneal and intrabursal models. To overcome this challenge and determine intraperitoneal tumor burden, most laboratories use luciferase-based, noninvasive bioluminescence imaging. To enable this, transplanted cancer cells are engineered to stably express luciferase, which produces bioluminescence upon administration of the substrate, D-luciferin. Bioluminescence is detected non-invasively using an In Vivo Imaging System (IVIS). This signal is then quantified to track tumor burden and reveal information on metastasis and drug therapeutic efficacy. Below, we describe the protocol for tracking tumor burden via luciferase-based, non-invasive bioluminescence imaging with IVIS in mice.

1. Preparation of the D-luciferin substrate
   1. Prepare 15 mg/mL D-luciferin solution in 1X PBS (pH 7.4) under sterile conditions.
   2. Filter-sterilize the solution using a 0.22 µm filter.
   3. Wrap the solution in aluminum foil to protect it from the light until use.
2. Initializing the IVIS imager for image acquisition
   1. Open the “living Image” software on the computer connected to the IVIS imager and log in.
   2. The imager needs to be initialized for image acquisition. This is done as follows:
      1. Upon logging in, a window of the IVIS control panel will open as shown in Figure 3A.
      2. Click on the "initialize" button located on the right lower side of the window (Figure 3A). It will take a couple of minutes for the system to initialize.
         Note: Initially, the temperature bar at the bottom of the window turns red, indicating that the imaging camera is in the process of reaching its target temperature of -90 °C (Figure 3B).
      3. When the imaging camera reaches the target temperature (-90 °C), the temperature bar will turn green, and the system status will say “Ready to use” (Figure 3C). Now the imager is ready for imaging the mouse.
         Note: The exposure time, binning, and f/stop can be adjusted as per user requirements.
3. Preparing the mouse for IVIS imaging
   ​Before imaging the mouse, D-luciferin must be injected intraperitoneally.
   1. Remove the mouse from the cage and place it on a rough or wired surface. Gently scruff the mice to restrain them properly as described in section 2.1.
   2. Pick the mouse up and turn it such that its ventral side is facing upwards.
      Note: If the mouse is properly restrained, all four legs will extend outwards upon picking up as shown in Figure 1B. Hold its tail between fingers as an extra measure of restraint during the procedure.
   3. Use a sterile alcohol swab to clean the site of injection as in the intraperitoneal injection protocol in section 2.1.
   4. Use a 1 ml TB syringe and a 30G, ½ inch needle to intraperitoneally inject an appropriate volume of D-luciferin from the stock solution to achieve a final dose of 150 mg/Kg D-luciferin.
      Note: Refer to the intraperitoneal injection protocol in section 2.1 for instructions on how to perform intraperitoneal injection. If using C57BL/6 mice (for syngenic allograft models), it is highly recommended to remove the fur before conducting bioluminescence imaging using IVIS. Black fur and pigmented skin significantly attenuate light, reducing the signal intensity.
   5. Gently place the mouse back in the cage. The exact wait time and exposure time will need to be empirically determined. This can be determined as follows.
      1. Anesthetize the mouse immediately after D-luciferin injection and image it as described in Section 3.4 every 2 min for 20 min.
         Note: When performing this experiment for the first time, it is recommended to determine how the luciferase signal varies over time after administering D-luciferin. This will help determine the most appropriate time to image the mouse after administering D-luciferin in subsequent experiments. The bioluminescent signal can vary depending on several factors, including the cell line, mouse xenograft model, level of luciferase expression, etc.
      2. At each interval take expsoures of 5, 10, and 15 s.
      3. Use this data to plot a standard curve. Based on the curve, choose the wait and exposure time at which the signal reaches a plateau while maintaining the highest intensity.
         Note: Usually this time is between 5–10 min but can vary depending on the model.
4. Non-invasive Imaging of mouse tumors using IVIS
   1. After the waiting period, place the mouse in an isoflurane chamber with 4% isoflurane rate and 3.5 flow rate of oxygen.
   2. Confirm the mouse is anesthetized by checking for the absence of the toe pinch reflex.
   3. Place the mouse on the stage inside the IVIS imager and slide its nose inside the nose cone for anesthesia maintenance during imaging. Make sure they stay within the marked lines of the camera focus and close the imager door.
      Note: There will be five positions on the stage inside the imager. Each position will also have their respective anesthesia nose cones. This allows imaging 5 mice simultaneously. Depending on the number of mice to image, adjust the camera field of view in the control panel box from “D”(broader field) to “A”(narrow focus). The mouse can be placed anywhere on the stage provided that their noses are kept inside the anesthesia nose cones through the entire duration of imaging. It is recommended to image the mice sequentially first on their dorsal sides and then on their ventral sides.
   4. Adjust the exposure time and click “Acquire” on the control panel (Figure 3C).
      Note: Make sure luminescence, photograph and overlay are “ON” in the control panel.
      Once the image has been captured, the image of the mouse with luciferase signal will show up on the screen.
   5. Open the imager and turn the mouse so that the ventral side is facing up and click “Acquire” after closing the imager door.
      Note: Ensure the mouse’s nose is inside the nose cone to ensure anesthesia maintenance during imaging.
   6. Once both dorsal and ventral images are acquired, open the imager door and place the mouse back in its cage.
   7. Wait for 5-10 min to ensure the mouse regains consciousness.
5. Analysis of IVIS image to determine tumor burden in the mouse
   1. Adjustment of the color scale of images
      1. Save the raw images to the appropriate folder by clicking on “file” followed by “Save as”.
         Note: The imager automatically adjusts the color scale based on signal intensity. It is advisable to keep the color scale constant across all images taken over the course of the experiment. The color scale can also be adjusted later by opening the raw image using this software.
      2. For manual adjustment, another window of the tool palette pops up in the upper right corner of the screen (Figure 3D) go to “Image adjust” and then choose “manual” for scale. Enter the minimum and maximum number desired for the scale value and click “Enter” (Figure 3E).
         Note: When adjusting the color scale of a previously acquired image, first open the saved raw infoclick.txt image file from the respective folder using “file” followed by “open”.
         3.5.1.3. Save this image separately from the raw image in the appropriate folder.
   2. Analysis of tumor burden in the mouse
      Note: For quantitative measurement of tumor burden, you can measure the signal intensity of each mouse across images taken over the entire period. You can place those values in Excel or GraphPad Prism for statistical analysis and to generate line graphs of tumor progression.
      1. Open the saved raw infoclick.txt file from the respective folder.
      2. Go to the ROI (Region of Interest) section of the tool palette window that pops up in the upper right corner (Figure 3F).
      3. Draw the ROI area on each mouse manually or automatically with the help of the “draw” function using any shape that works best for you. To do this manually, keep the area consistent for all the all images across all the mice that are being compared.
         ​Note:For Intraperitoneal inoculation, manual selection of the ROI for the entire chest and abdomen on the ventral image of the mouse works best. For intrabursal inoculation, manual selection of the ROI in the dorsal image area with signal works best.
         Note: Automatic ROI selection will only select regions with higher image signal intensity.
      4. Click on “Measure ROI” (Figure 3F). This will open another window with data on the total and average counts for each selected ROI.
      5. Save this file in the appropriate folder so that it can be accessed later. Use this data to plot the graph in either MS Excel or GraphPad Prism.

Results

Representative data from an intraperitoneal ovarian cancer model using luciferase-tagged OVCAR8 cells in immunocompromised Foxn1 nude mice is shown in Figure 4. Stable luciferase expression in these cells enabled longitudinal monitoring of tumor progression by non-invasive bioluminescent imaging using the IVIS imager as described in section III. Intraperitoneal injection of 5 million OVCAR8-luciferase cells into Foxn1 nude mice resulted in the establishment of metastatic tumors in the peritoneal cavity with a 100 percent “take rate” (Figure 4). The tumors progressed rapidly resulting in large tumor burden and ascites (Figure 4A,B). In vivo imaging revealed a high tumor burden within 29 days of intraperitoneal inoculation of 5 million OVCAR8-luciferase cells in Foxn1 nude mice (Figure 4C). To understand how differences in the number of intraperitoneally injected ovarian cancer cells impact the kinetics of tumor progression and survival in mice, we compared tumor burden over time in mice that had 5 million OVCAR8-luciferase cells intraperitoneally injected to mice that had 1 million OVCAR8-luciferase cells intraperitoneally injected. In comparison to mice that had 5 million OVCAR8-luciferase cells injected, the tumor burden was significantly lower in mice injected with 1 million OVCAR8-luciferase cells(Figure 4C,D). In fact, the tumor burden of mice that had 1 million cells 49 days post-injection was lower than the tumor burden observed on day 29 post injection in mice injected with 5 million cells (Figure 4C,D). Consistent with higher tumor burden in mice injected with 5 million cells, we also saw that these mice survived for fewer days as compared to mice injected with 1 million cells (Figure 4E).

Representative images from an intrabursal ovarian cancer model using luciferase-tagged control and ZNF217 overexpressing OVCA420 cells is shown in Figure 5. For intrabursal model, 50,000 cells in 5 µL PBS was injected into the ovarian bursa as described in section 2.2. Stable luciferase expression in these cells enabled longitudinal monitoring of tumor progression by non-invasive bioluminescent imaging using the IVIS imager as described in section 3. The transcription factor Zinc finger protein 217 (ZNF217) was recently shown to act as a potent oncogene and drive ovarian cancer progression¹⁴. IVIS imaging on day 1 post-injection confirmed successful intrabursal transplantation of cells (Figure 5A). Consistent with ZNF217’s role as an oncogene, in vivo imaging at day 28 post intra-bursal injection demonstrates a higher tumor burden in mouse innocluated with ZNF217 overexpressing OVCA420 cells compared to control (Figure 5B,C).

Taken together, these data demonstrate the utility of the intraperitoneal and intrabusal models to mimic ovarian cancer progression and metastasis. As the number of cells injected is an important criteria that determines the rate of tumor progression and survival, this parameter has to be optimized first for different cell lines to suit individual experimental goals.

Figure 1: Intraperitoneal injection of ovarian cancer cells in mouse. (A) An investigator demonstrating the correct technique to scruff mouse. (B) A representative Foxn1 nude mouse with site for intraperitoneal injection shown. Please click here to view a larger version of this figure.

Figure 2: Intrabursal injection of ovarian cancer cells in mouse. (A) Surgical tools needed for intraperitoneal injection of ovarian cancer cells in mouse. (B) A representative anesthetized Foxn1 nude mouse with incision site for surgery to access the ovarian bursa shown. Please click here to view a larger version of this figure.

Figure 3: Steps involved in setting up the IVIS to image luciferase bioluminescence in mouse using the Living Image Software. (A) In the initial window of the living Image Software, press the “initialize” button. The camera temperature bar turns red (B) and takes 5–7 min to turn green. (C) Green camera temperature bar indicates the instrument is ready. At this time, place the D-luciferin injected, anesthetized mice on the stage and press “acquire” (highlighted by red box). (D,E) Use the tool palette to access the ROI and other parameters. Indicate the desired ROI and adjust the image scale according to the experimental goals. (F) Click the “Measure ROI” button to measure tumor burden with the ROI. Please click here to view a larger version of this figure.

Figure 4: Representative data from Foxn1 nude mice showing data obtained using an intraperitoneal mouse model of metastatic ovarian cancer. (A) 5 million OVCAR8-luciferase cells were injected intraperitoneally into 6-week-old female Foxn1 nude mice. The mice were euthanized when they met the criteria, and a representative image of a mouse shows a swollen belly due to high tumor burden and ascites accumulation. (B) Dissection of the mouse injected with 5 million OVAR8-luciferase shows large tumors within the peritoneal cavity at sites including omentum, mesentery, and diaphragm. Tumors are indicated by white arrows. (C) IVIS images of Foxn1 nude mice injected with 5 million OVCAR8-luciferase cells show rapid tumor progression over time. (D) IVIS images of Foxn1 nude mice injected with 1 million OVCAR8-luciferase cells show that tumor progression rate is slower than in mice injected with 5 million cells. (E) Kaplan-Meir curve reveals that Foxn1 nude mice injected with 5 million OVCAR8-luciferase cells have lower overall survival compared to mice injected with 1 million cells. Please click here to view a larger version of this figure.

Figure 5: Representative data from Foxn1 nude mice showing data obtained using an intrabursal mouse model of ovarian cancer. (A) 50,000 luciferase-tagged control and ZNF217 overexpressing OVCA420 cells was resuspended in 5 µL PBS and injected into the ovarian bursa of 8-weeks old female Foxn1 nude mice. The mice were imaged using IVIS 1 day after intrabursal transplantation of cells to confirm successful implantation. (B) IVIS images of mice post intrabursal injection of luciferase-tagged control and ZNF217 overexpressing OVCA420 cells at days 14, 21, and 28 show how tumor burden changes over time. ROI used to measure signal intensity is also shown. (C) Graph showing how luciferase signal intensity changes over time in the experiment described in Figure 5B. Please click here to view a larger version of this figure.

Cell line	p53 status	Intraperitoneal model	Intrabursal model
HEYA8	WT	106 cells	2.5 x 105 cells
OVCAR8	Splice‐site mutation (c.376-1G>A) → leads to deletion Y126–K132	106 cells	2.5 x 105 cells
OVCAR420	R273H	107 cells	Does not form tumors
		(forms small metastatic tumors)
TYK-Nu	R175H	107 cells	5 x 104 cells
OVCAR3	R248Q	107 cells	5 x 104 cells
OVCAR4	L130V	106 cells	2.5 x 105 cells
SKOV3	Null	106 cells	2.5 x 105 cells

Table 1: Recommended number of cells to transplant for commonly used human ovarian cancer cell lines. The number of cells to inject for commonly used ovarian cancer cell lines to generate intraperitoneal and intrabursal models is listed. These numbers are the general recommendations based on data from published literature^13,14. The appropriate number for an experiment will depend on respective experimental goals and will need to be experimentally determined.

Discussion

The ability to mimic complex interactions between tumor cells, microenvironment, and systemic factors makes mouse models an excellent resource to study ovarian cancer progression¹². In addition to advancing our understanding of factors that drive ovarian cancer progression and metastasis, in vivo mouse models also play an important role in pre-clinical evaluation of small molecules and treatment strategies for metastatic ovarian cancer^12,15. Given these advantages, several in vivo mouse models have been developed to study metastatic ovarian cancer^12,15. Understanding the advantages and limitations of each model will be critical in determining which model is appropriate for specific research questions.

While GEMMs have contributed immensely to our understanding of factors driving ovarian cancer progression, they are challenging and more expensive to develop¹⁰. In addition to this, the latency in tumor development in these models make working with them time and resrourse intensive¹⁰. Due to these reasons, transplantable tumor models have become highly popular to understand factors driving ovarian cancer progression and therapeutic response. Although sub-cutaneous models are easy to generate and convenient to monitor tumor progression, they are not recommended to study ovarian cancer progression¹⁶. A major reason for this is the profound differences in the sub-cutaneous microenvironment and the ovarian-fallopian-peritoneal microenvironment¹⁶. Due to these differences the sub-cutaneous models are not clinically relevant and not suitable to evaluate how factors in the tumor microenvironment impact ovarian cancer progression¹⁶. The two transplantable tumor models that are most frequently used in ovarian cancer research are the the intraperitoneal model, and the intrabursal model. Their ability to mimic critical steps in ovarian cancer progression as well as the tumor microenvironment, make these models suitable to study how cell intrinsic and extrinsic factors impact ovarian cancer biology and therapeutic response.

The site of tumor cell implantation differs in intrabursal and intraperitoneal models. This difference has implications on both the tumor microenvironment and the specific steps that they mimic along the metastatic process. In the intrabursal model, tumor cells are injected into the ovarian bursa. Therefore, the intrabursal model is suitable to study how different factors impact the early events in metastasis, such as the ability of the tumor cells to colonize the ovary as well as early events that enable them to shed from the primary tumor and invade the peritoneal cavity. In comparison, in the intraperitoneal model the tumor cells are injected directly into the peritoneal cavity. Therefore, the intraperitoneal model mimics the stage in disease progression post-shedding of cancer cells into the peritoneal space. The intra-peritoneal model is often characterized by widespread tumor dissemination and formation of ascites, which is a hallmark of late-stage ovarian cancer¹⁷. Therefore the intraperitoneal model is suitable to understand how different cancer cell intrinsic and extrinsic factors impacts advanced stages in the metastatic process. As over 75 percent of ovarian cancer patients have advanced metastatic diasease at diagnosis, the intraperitoneal model is also an excellent platform to determine the efficacy of novel therapeutic strategies¹⁸.

A major advantage of the transplantable models described in this report is the ease with which they can be used to study the impact of different genetic factors on ovarian cancer progression. Unlike the time and effort required to generate GEMMs, cell lines can be engineered to either overexpress or deplete (knockdown or knockout) a protein within a month with much lesser effort and resources. Multiple genetic alterations can be incorporated in a cell line to study functional interaction between different factors to drive ovarian cancer progression. Overexpression or depletion of proteins can be achieved both constitutively or in an inducible manner, allowing flexibility with experimental design. Moreover, the protocols described in this report can be used to develop both xenograft models as well as syngeneic allograft models. In xenograft models, human ovarian cancer cells are transplanted into immunodeficient mouse strains such as the NSG mouse and Foxn1 nude mouse. In contrast, in allograft models, mouse cancer cells transplanted into a immunecompetent mouse strain of the same genetic background such as the C57BL6/J or FVB. The allograft models are useful to study how immune cells interact with ovarian tumor cells, and to understand the effect of different cancer cell intrinsic and extrinsic factors on ovarian tumor microenvironment. These models are also extremely valuable to test novel immunotherapeutic agents in a preclinical setting.

Several factors can determine the successful establishment of these models. One major factor is the inherent tumorigenic potential of the cell line being used. Some cell lines do not form tumors in vivo and are hence not suitable for either of these approaches. For example ALST ovarian cancer cells do not form tumors using both the intraperitoneal and the intrabursal models (data not shown). The ability of a cell line to form tumors and the kinetics of tumor growth can also vary between the two transplantable models described in this report. The number of cells that are transplanted is another critical factor that will determine both the engraftment rate (take rate) as well as the tumor kinetics. Some ovarian cancer cell lines do not establish successful tumors when transplanted in low numbers. For example, we have observed that a commonly used ovarian cancer cell line, TYK-Nu, has to be injected at a high cell number (5–10 million) to successfully establish intraperitoneal tumors in Foxn1 nude mice (data not shown). On the other hand, some cell lines such as HEYA8 and OVCAR8 are highly tumorigenic and can establish metastatic tumors with much lesser number of cells (50,000–1 million). Using higher cell numbers (5–10 million) in these cell lines will result in extremely aggressive metastatic tumors in an intraperitoneal model. These tumors progress very rapidly to euthanasia criteria within 2–4 weeks. Such rapid tumor progression might not be suitable for certain experimental goals. Therefore, injecting the appropriate number of cells is critical and will need to be determined based on experimental goals and the tumorigenic potential of the cells. In xenograft models, the tumor cell engraftment rate (take rate), and kinetics of tumor formation can vary depending on the strain of the immunodeficient mouse that is used. For instance, NSG mouse exhibits increased both primary tumor growth and metastatic dissemination as compared to Foxn1 nude mouse¹⁹. In addition to the mouse strain, the kinetics of tumor formation, pattern of tumor cell dissemination, and the ability to form ascites in both xenograft and allograft models depend heavily on the cell line being injected. Using correct needle guage is critical for both intraperitoneal as well as intrabursal injection of cancer cells in mice. Using a needle that is too thin (e.g., 30G) can create high shear stress and pressue, significantly decreasing cell viability. On the other hand using a needle that is too large can cause excessive tissue damage or leakage. While using high concentrations of cells (> 10⁷ cells/ml) there are also risks with clogging if a thin syringe is used. Another common issue is over-penetration and injury to internal organs. This can be overcome by using the needle of appropriate length (1/2 to 5/8 inch) and a proper injection technique as described in this protocol. For the intraperitoneal model, injecting at a 30–45° angle in the lower right quadrant of the abdomen minimizes the risk of organ damage.

The intraperitoneal model is technically easier to develop compared to the intrabursal model. In comparison to the intrabursal model, intraperitoneal injection is easier to learn and does not require anesthesia. Developing the intrabursal model involves performing a survival surgery, which requires specific equipments as described in section 2.2. Successful transplantation of cells should be confirmed by imaging the mouse either the same day or on day 1 post-injection. The luciferase signal intensity will drop during the initial days as some of the transplanted cells will die. The cells that survive that survive and are able to adapt to the new microenvironment will subsequently proliferate, leading to a subsequent increase in the bioluminescence signal. During intrabursal injection, there is also a risk of leakage of cells from the bursa during the injection of cancer cells into the peritoneal cavity. Mastering the surgical skills needed to consistently perform intrabursal inoculation of ovarian cancer cells can take time and practice. It is also recommended to have another person to assist the individual performing the surgery during intrabursal transplantion. A limitation of both intrabursal and intraperitoneal approaches is that advanced non-invasive imaging is required to monitor tumor progression and metastatic burden. The cost associated with the equipment needed for visualization of the tumors can be prohibitive for some researchers. In addition to these, it is also important to keep the time between cell line preparation and injection to ensure cell viability and for achieving consistent experimental results. Overall, it is recommended to factor in specific experimental goals and optimize the conditions to accommodate these goals while using these models.

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
0.22 um syringe sterile filter, 33mm	Fisher Scientific	09-720-004
1 ml TB syringe	BD	309659
21G, 1 inch needle	BD	305167
30G, ½ inch needle	BD	305106
5 inch Curved fine point forceps	Fisher Scientific	1631
6-0 Absorbable monofilament suture	MaxonTM	6532-11
Alcohol swab(70% isoprolyl alcohol)	Dukal Corp	80003156
Betadine (1% iodine solution)	Purdue Products	367618150085
Carprofen Inj, 50mg/ml	Levafen	RXLEVA50-20INJ
Centrifuge	Eppendorf	5702
D-luciferin	Syd Labs	MB000102-R70170
Dissecting jeweler microforceps, fine tip	Fisher Scientific	08-953E
Eye ointment	Medvet	11897
Fine tipped scissors	Fisher Scientific	8940
Heat pad	Sunbeam	731-500
Instruments Dissecting Forceps with Fine Points	Fisher Scientific	S08096
Insulin Syringe 3/10ml	BD	328438
Isofurane	Fluriso	Fluriso250
IVIS	Perkin Elmer	IVISLMIII
Mayo- Heger Straight Needle Holder	Fisher Scientific	08-966
Sterile 1X PBS, without calcium and Magnesium	Corning	21-040-CV
Sterile Cotton tip applicators	Mckeesson	24-106-1S
Sterile Gauge	Kendall	3033
Sterile Pad/Towel/Drape, non-fenestrated	Henry-Schein	100-9687
Triple Antibiotic Ointment	Globe	4006-4SL
Trypsin-EDTA (0.25%), phenol red	Gibco	25200-056