X-ray Visualization of Intraductal Ethanol-based Ablative Infusion for Prevention of Breast Cancer in Rabbit Models

Breast cancer (BC) is the most common and the second-highest cancer-related death for women in the United States. Projections for 2025 estimate that there will be 316,950 new breast cancers, and 42,170 women will die from BC¹. Currently, bilateral prophylactic mastectomy is the most effective procedure to prevent BC. However, this is a highly invasive procedure that involves a complete removal of the epithelial cells, from which breast carcinoma arises, and the surrounding tissue. Due to its invasiveness as well as the psychological and social impact of this procedure, less than 50% of high-risk women undergo risk-reducing mastectomy². We, and others, have developed intraductal (ID) delivery procedures for primary prevention and/or local treatment of breast cancer in rodent models^2,3 as an alternative to the current preventions and treatments. Ethanol (EtOH) has a low toxicity and safety profile that is well established and is used in multiple clinical applications, such as sclerosing agents for treatment of venous malformations and as an ablative agent for local treatment of some cancers³. Typically, several milliliters of EtOH are infused or delivered at 90-100% concentration in these clinical procedures. In our previous work, delivery of 70% EtOH directly into the ductal tree system of mouse and rat models was effective at chemically ablating mammary epithelial cells with limited damage to adjacent normal tissue, and at preventing breast tumor formation^4,5,6,7. As this procedure is scaled up to the larger ductal tree system of a rabbit with a larger luminal volume to luminal epithelial cell surface area ratio, we explore the ablative properties of a solution with a lower percentage of EtOH (10% to 70%). With a lookout for clinical translation, we reason that the lowest percentage of ethanol that is effective at ablating epithelial cells will be the most well-tolerated and have the best safety profile.

Confirmation of complete ductal tree filling is necessary to guarantee that the ablative solution has come in direct contact with mammary epithelial cells. In our previous studies in rodent models, X-ray visualization of infused ductal tree(s) by microCT imaging was used after the procedure. Due to the required lapse of time to anesthetize, transfer, set, and position the animal for imaging, FDA-approved Omnipaque (iohexol) or similar iodine-containing fast-diffusing contrast agents were not suitable for ductal tree visualization in rodents^6,8. We found that nanoparticle-based contrast agents, especially those containing tantalum oxide nanocrystal, were slower diffusing and more suitable for ductal tree visualization in rodents^6,7,8,9. However, this posteriori confirmation by microCT imaging does not allow us to monitor or control the amount of infused volume and deviates from clinically established diagnostic procedures, such as ductography^10,11, for ductal tree visualization. Thus, a key step to establishing the technical feasibility of translating this ID procedure to humans is to demonstrate real-time fluoroscopy visualization of the infused ductal tree in an animal model of increasing size and complexity of its mammary glands. This protocol scales up this ablative procedure from rodents^4,5 to rabbit models. Evolutionarily, anatomically, and physiologically, rabbit mammary glands are more similar to human breasts than those of rodents or other large animal models, such as cows and sheep^12,13,14. Female rabbits have four pairs of mammary glands, each containing four ductal trees, whereas rodents have only one ductal tree per mammary gland. Rabbit teats can be cannulated^15,16 using a procedure similar to ID administration of contrast agent in clinical ductography in first-in-human clinical research. Therefore, rabbits provide a practical and relevant intermediate large-animal model for the translational application of this ID ablative procedure to humans. This protocol addresses technical challenges of ID delivery and in vivo imaging of a multi-ductal tree system that could not have been addressed in rodent models. This protocol uses instruments, reagents, and materials that are compatible with current clinical practice for visualization of ductal trees. Thus, the described procedure for fluoroscopy-guided infusion of iohexol-containing ethanol-based ablative solution could be readily implemented and evaluated in first-in-human clinical trials.

This method has been implemented in our laboratory to successfully cannulate and sequentially infuse all four ductal trees of one or more mammary glands in a rabbit, in a single session, with an ethanol-based ablative solution containing a contrast agent (Figure 1, Figure 2, Figure 3). This method involves infusing the ablative solution directly into the cannulated teat opening with a 27 G blunt-tipped needle of a rabbit (4-month virgin) on a fluoroscopy table. This procedure is performed on an animal under general anesthesia (isoflurane) with peri- and post-procedure anti-inflammatory treatment (ketoprofen, non-steroidal anti-inflammatory drug). Fluoroscopy imaging allows us to monitor the filling of the ductal tree in real-time, to control the rate and amount of dispensed volume, and/or to determine how successful ID delivery is in each individual tree system (Figures 1, Figure 2, Figure 3). This fluoroscopy technique approximates more closely to the intended clinical application for image-guidance of the ablative treatment and can help limit the overall radiation dose imposed on the patient. This protocol demonstrates that FDA-approved Omnipaque (iohexol) is a suitable contrast agent to visualize the initial filling of the rabbit ductal tree (Figure 3). Observations by gross examination and histological analysis show that an ethanol concentration of 70% causes rapid tissue damage within and outside the ductal tree and extending beyond the mammary gland structure (Figure 3). Ethanol concentration in the 10-40% range provides adequate epithelial cell ablation with lower collateral tissue damage than 70% ethanol (Figure 4). Longitudinal studies using this procedure with appropriately powered group size per ablative solution and timed tissue collections will be required to establish optimal parameters of the ablative solution for its clinical evaluation in human patients.

All the described experiments were conducted under protocols approved by the Institutional Animal Care and Use Committee at Michigan State University. Rabbits (Oryctolagus cuniculus) were cared for in accordance with the Guide for the Care and Use of Laboratory Animals and the USDA Animal Welfare Act at an AAALAC-accredited facility.

NOTE: This method was conducted in virgin (nulliparous) and retired breeder (multiparous) New Zealand White animals ranging in age (4 months to > 1 year) and weight (2.6 to 4.2 kg) acquired from commercial sources. In our experience, the size of the animal determined by weight is more reliable than the age of the animal to predict the size of the teats. Generally, animals that weigh more than 3.3 kg present with suitable teats for cannulation. The protocol described below focuses on virgin animals of 4-5 months of age and a weight of more than 3.3 kg, since they are more appropriate for long-term efficacy, wound healing, toxicity, and safety studies.

1. Preoperative preparation

1. Acclimate animals in the new facility for at least 1 week upon arrival, especially for animals intended for recovery procedures and long-term studies. During this first week, monitor/check the rabbits daily and supply treats, nutritional enrichment as recommended by institutional guidelines, to help with the acclimation process.
2. Acquire the rabbit (~ 4 months old New Zealand White) from the approved housing facility. Record the body weight before the procedure.
   NOTE: The body weight can be recorded the day before the procedure to prepare the required calculations for the anesthesia. Retired breeders (> 1 year of age, > 3.5 kg) can also be used as they have larger teats and allow for easier cannulation of individual ducts (Figure 3). For these reasons, retired breeders can be used in initial experiments to become familiar with and optimize the intraductal procedure.
3. Inject 35 mg/kg of ketamine and 5 mg/kg of xylazine intramuscularly 20 min before isoflurane administration to sedate the animal.
   NOTE: Anesthesia is administered based on weight of the rabbit and the ranges for each drug are as follows: 15-35 mg/kg for ketamine, and 2-5 mg/kg for xylazine. Ensure the animal is sedated before moving on to hair removal and intubation. This is for the welfare and safety of the animals and staff. After confirmation of sedation, the rabbit can then be placed dorsally on the imaging/operating table.
4. Inject 5 mg/kg of ketoprofen subcutaneously for analgesia after clinical signs of sedation are shown (i.e., quiet demeanor, and partially closed and pink colored eyes).
   NOTE: Analgesia is administered based on the weight of the rabbit and the range for ketoprofen is 2-5 mg/kg.
5. Intubate the rabbit with the appropriate equipment (e.g., endotracheal tube or supraglottic airway device) and hook up to an isoflurane machine (1-2% isoflurane, 1.0 L/min oxygen) that has been adequately tested and certified to anesthetize the rabbit. Carefully monitor the animal's respiration to ensure that anesthesia is maintained at 1-2% of isoflurane. Monitor the rabbit for oxygen saturation via SpO₂, heart rate, respiratory rate, and temperature throughout the procedure.
   NOTE: The size of the intubation tube is based on the weight and size of the rabbit. However, the size range is not always accurate, so it is helpful to have a variety of sizes to see which one best fit that particular rabbit. A nosecone mask can also be used in place of an endotracheal tube for anesthesia purposes¹⁷, being aware that this mask does not provide protection of the animal's airway afforded by intubation. Warm-water circulating blankets (warming blankets) set to 37°C are placed under towels to maintain the rabbit's body temperature.
6. Place and secure a 25 G venous catheter into the marginal ear vein to allow for emergency drug administration.
   NOTE: A gauge range of 24 to 26 can be used depending on the size of the rabbit vein.
7. Apply eye lubricant to both eyes to prevent ocular irritation and corneal drying.
8. Shave the fur around the second and third pairs of teats with an electric razor. Use a cotton-tip applicator to spread hair removal cream onto the teat area. Allow the cream to contact the area for 15 s.
   NOTE: Extreme caution must be taken so as not to damage any teats with the razor. A cordless vacuum can also be used to aid in maintaining a clean procedure area.
9. Wet a gauze pad with sterile saline and use it to rinse the cream and loosen fur from the animal after 15 s of hair removal cream application. Confirm good visibility and access to the area of the teat from where the fur was removed. Repeat if necessary.
   NOTE: Cream should remain on the rabbit for the shortest possible interval, between 10 - 30 s and be removed completely to avoid chemical burns to the skin.

2. Intraductal Infusion

1. Prepare an ablative solution by mixing appropriate volumes from stock solutions under sterile conditions in a BSL2 tissue culture hood.
   NOTE: Iohexol (350 mg Iodine/mL) should be stored in a dark area due to light sensitivity. A range of EtOH concentrations was used during this experiment. For testing other percentages of EtOH, dilute the stock solutions to the necessary concentration of ablative solution. To maintain the same concentration of Iodine in ablative solution with different percentages of EtOH, PBS, or sterile water can be used to fill up the volume difference.
2. For this example, prepare a fresh ablative solution of 10% EtOH, 280 mg Iodine/mL iohexol , 1% food dye in a 5 mL tube. For a final volume of 5 mL, add 4 mL of iohexol stock (350 mg Iodine/mL), 500 μL of 100% EtOH (200 proof), 450 μL of PBS, 50 μL of stock blue food color.
   NOTE: Each ductal tree may be filled up with up to 400 µL, but typically 250-350 µL for animals under 3.5 kg. Evans Blue up to 0.2% may be used instead of food dye. Evans Blue may be a preferred option if whole mount or other tissue analyses are intended immediately after infusion(s).
3. Remove any dead skin that covers the ductal openings with fine pointed forceps.
   NOTE: Rabbits can have a keratinized plug protruding from the teat that can prevent successful cannulation if not removed. Topical lidocaine can also be applied around the teat to assist with minimizing irritation around the injection site (Table 1).
4. Wipe the infusion site with chlorhexidine gauze pads.
   NOTE: Chlorhexidine is used as a cleansing agent to disinfect the injection site before cannulation (Table 1).
5. Insert the bevel of a 28 G needle (length: 12.7 mm) into the side of the teat and slowly inject 200 µL of 0.9% saline at a rate of 200 µL/min. This allows for better visualization of the ductal openings.
   NOTE: Not all 200 µL of saline may need to be injected into the teat; stop injection once you see the saline coming out from one or more ductal openings.
6. Aspirate 1 mL of prepared ablative solution using a 1 mL Luer lock syringe. Attach the syringe to the female, "winged" end of the 12-inch male-female extension line. Carefully attach a 27 G blunt tip needle (length: 12.7 mm) to the male end of the extension line. Prime the line with the solution. Wipe the needle clean with an alcohol gauze pad. Also, be mindful not to tip the syringe with the ablative solution as this may cause air bubbles to form.
   NOTE: These are recommended volumes aimed at fully filling the ductal tree(s): up to 300 µL in each tree and up to 1.2 mL per cervical and/or inguinal mammary glands (1^st and 4^th pairs), up to 400 µL in each tree and up to 1.6 mL per thoracic and/or abdominal mammary glands (2^nd and 3^rd pairs). For other applications, it may be appropriate to use smaller or larger volumes based on experimental requirements and/or fluoroscopy guidance to avoid overfilling the ductal tree. The extension line allows for more control of the flow rate and for simultaneous infusion and live fluoroscopy imaging. For comparison, recommended volumes of intraductal infusion in 9-12 weeks of age mouse models⁴ are: up to 30 μL in cervical and inguinal and up to 50 μL in thoracic and abdominal mammary glands, and rat models⁵: up to 100 μL in cervical and inguinal and up to 300 μL in thoracic and abdominal mammary glands.
7. Use a 10x magnifying lamp to aid in locating the ductal openings. Gently hold the teat using fingers and cannulate the needle into the ductal opening. Gently continue inserting the 27 G blunt tip needle until the tip is fully inside the teat. To accommodate the needle in the teat, bring the teat up towards the needle instead of pushing the needle down into the teat. Take care to follow the path of the ductal opening.
   NOTE: In some rabbits, you may feel resistance when trying to insert the needle into the teat opening(s). Carefully apply slight pressure to break through the top epithelial cell layer. In our experience, a magnification device is required to clearly identify the ductal opening for cannulation. This may be a magnifying lamp, lens, loupe, or similar device.
8. Slowly infuse 300 µL of the solution at a constant rate of approximately 200 µL/min once the needle is completely inserted. Wait for 30 s after infusion to remove the needle from the cannulated tree; this ensures injected volume remains within the ductal tree and reduces the likelihood of leakage.
   NOTE: Typically, there is one researcher cannulating and holding the needle, while a second researcher holds the syringe and pushes the plunger at the desired rate. A syringe pump may be used to have a more controlled flow rate, as abrupt changes in infusion rate can burst or damage the ductal trees.
9. Clean off any spilled solution with moistened gauze or an EtOH wipe to avoid extraneous contrast solution in images.

3. Fluoroscopy imaging

1. Take fluoroscopy images after each ductal tree has been infused. The parameters of the fluoroscopy are: 30 fps, 67 kV, and 17.3 mA on a fluoroscopy X-ray instrument. However, adjust these based on the experiment and imaging needs.
2. Use the fluoroscopy images to determine whether additional volume is required to fully fill the ductal tree.
   NOTE: Fluoroscopy imaging can take place live at the same time as infusion of the ablative solution. Metal or plastic forceps can be used to hold the teat while imaging is taking place to protect staff from harmful X-rays. This allows for monitoring of the filling of the ductal tree(s). Live fluoroscopy can guide when to stop infusion based on the increased volume at the alveoli ends. Fluoroscopy after infusion can confirm if the ductal tree was fully filled or if there was some leakage. A confirmatory fluoroscopy is typically performed after infusion of each duct within the same mammary gland.

4. Postoperative care and recovery

1. Discontinue the flow of isoflurane after the last intraductal infusion.
2. Inject 0.5 mg/kg of atipamezole intramuscularly.
   NOTE: The recovery time varies between animals, but the rabbit should start showing signs of recovery 5-20 min after injection.
3. Provide continued heat support to the animal on a warming blanket until fully recovered from anesthesia. Maintain oxygen flow for up to 5 min before removing from anesthesia.
   NOTE: Signs of recovery include mouth movements such as chewing, coughing, nose twitching, and/or eye movement. Rabbits should have a righting reflex and be able to maintain themselves in sternal positioning prior to placement back in recovery carrier. Recovery agent is administered based on the weight of the rabbit with a range of 0.1-1 mg/kg atipamezole.
4. Inject 5 mg/kg of ketoprofen subcutaneously.
5. Remove the intravenous catheter once the rabbit can maintain themselves in a sternal position. Hold a gauze pad to where the catheter was removed to stop any excess bleeding.
   NOTE: The removal of the catheter can take place when the rabbit is in the transport carrier.
6. Transport the rabbit back to the appropriate housing facility.
7. Continue injections of 5 mg/kg of ketoprofen subcutaneously for at least 3 d post-procedure.
8. Monitor the rabbit for signs of discomfort, distress, pain, and self-mutilation once daily for at least 3 d post-procedure. If the rabbit exhibits any of these clinical signs, treatment with ketoprofen can be extended. Record and monitor the body weight to assess for signs of anorexia.
   NOTE: Ketoprofen is administered based on the weight of the rabbit with a range of 2-5 mg/kg. It can be administered every 24 h for up to 5 days after intraductal infusion to reduce inflammation and minimize scarring. To minimize adverse events of skin ulceration or other wound healing related issues, apply lidocaine topically on the injection site. Any animal showing persistent signs of discomfort, distress, pain, or injury after ketoprofen treatment should be euthanized.

5. Tissue analysis

1. Administer euthanasia solution (pentobarbital sodium and phenytoin sodium) intravenously at 100 mg/kg. After 60 s check for signs of life by toe/ear pinching, signs of breathing or heartbeat, corneal reflex, and/or pupil stimulation.
2. Perform a necropsy to obtain mammary gland(s) tissue and process for the routine paraffin embedding procedure after 24-36 h in 10% Neutral Buffered Formalin¹⁸. Then, perform standard Hematoxylin and Eosin (H&E) staining and/or immunohistochemical stain with cell type-specific marker to assist in desired analysis readouts¹⁸. Dispose of the carcass through the proper disposal protocol (e.g., incineration).
3. Analyze mammary gland tissues in consultation with a pathologist. Use a computer software program to assist in the quantification of ablation rate and collateral tissue damage.
   NOTE: Tissue analysis was performed on 4-month-old New Zealand White rabbits within one hour after infusions (Figure 4) with QuPath Open Software for Bioimage Analysis (https://qupath.github.io/). This analysis is based only on H&E-stained tissues. QuPath or similar computer software, requires input and calibration by a pathologist. Some cells may be misclassified using just morphological features (Figure 4). The use of cell type specific markers such as cytokeratins and α-smooth muscle actin can be used to improve computer-assisted classification⁶. Ultimately, cell classification analysis must be curated and validated by a pathologist.

Each of the 8 mammary glands of a female rabbit contains 4 ductal trees that open at independent teat orifices (Figure 2). Due to the difference in size and number of ductal trees per mammary gland between rodents (only 1 duct per mammary gland), rabbits are a good intermediate model for human translation. We can infuse up to 400 µL of 10-70% EtOH solution to fill the entire ductal tree of any mammary gland of 4-month-old New Zealand White rabbits (Figure 1, Figure 2, Figure 3, Figure 44,8,9). We can infuse up to 4 ductal trees in up to 8 mammary glands with the ablative solution in a single session. A typical experimental design consists of infusing 2-3 ductal trees within a single mammary gland in up to 4 mammary glands with a particular ablative solution containing Iodine-based X-ray contrast agent (Figure 2, Figure 3). For iohexol-containing (90-300 mg of Iodine/mL) ablative solution, fluoroscopy is performed during and/or after each infusion to determine the individual success of infusing each ductal tree with partial or full amount of infused solution (Figure 2, Figure 3). Collection of the mammary gland tissue enables assessment of how changes in formulation affects destruction of mammary epithelial cells (Figure 4). These imaging analyses provide information to understand the most suitable solution to achieve maximal ablation while minimizing surrounding tissue damage. We determined that 10% EtOH solution provides a comparable ablative rate to ablative solutions containing a higher percentage of EtOH (Figure 4).

Figure 1: Workflow of intraductal procedure. Key steps of the ID procedure are highlighted. Please see the video for more details. Please click here to view a larger version of this figure.

Figure 2: Key steps of intraductal cannulation and infusion. (A) Injection of saline perpendicular to the teat to dilate the ductal openings for cannulation (median plane view). (B) Cannulation and filling of a ductal tree (D1) can be tracked with blue dye in the ablative solution (median plane view). (C) Real-time fluoroscopy imaging offers precise and high-resolution monitoring of ductal tree filling (D1) with iohexol in the ablative solution (dorsal plane view). Ductal tree openings are numbered left to right, starting on the upper quadrant (D1, left upper quadrant) and finishing on the lower quadrant (D4, right lower quadrant). Please click here to view a larger version of this figure.

Figure 3: Teat size and successful delivery of ablative solution to multiple ducts. Typical presentation of teat sizes in New Zealand White rabbits. Teat size varies based on the weight and age of the rabbit. Mammary glands are numbered from top left (L1, left cervical) to bottom right (R8, right inguinal). All images are shown in the dorsal plane. (A) 2.8 kg virgin rabbit (top) with smaller teats, difficult to cannulate, 3.5 kg virgin rabbit (middle) with suitable teats for cannulation, and 4.1 kg multiparous rabbit (bottom) with larger teats, much easier to cannulate. (B) Blue food color in the infused solution may be used as in vivo evidence of intraductal delivery and ductal tree filling. Unsuccessful infusion is indicated with a red outline (fat pad delivery, top), and successful infusions with a blue outline (intraductal delivery, middle and bottom). A 70% EtOH solution causes more damage to the skin (erythema) minutes after infusion (dark blue, middle panel) compared to a 10% solution (light blue, lower panel). (C) Fluoroscopy provides in vivo evidence of intraductal delivery. Unsuccessful infusion (fat pad delivery, top panel). Successful sequential infusion of D1 ductal first and D2 ductal tree second (bottom left panel). Live fluoroscopy provides image guidance for filling (white arrows) of D3 ductal tree (bottom right panel); the extension line filled with iohexol-containing ablative solution and forceps to hold the teat are also visible. Scale bars correspond to 1 cm in images at different magnification. Please click here to view a larger version of this figure.

Figure 4: Tissue analysis of mammary glands in New Zealand White rabbits after intraductal procedure with ethanol-based ablative solution. (A-B) Representative H&E staining of a right inguinal mammary gland of 4 month old animal with no ablative treatment compared to a right inguinal mammary gland of another animal with 10% EtOH ablative treatment. Tissue slices are cut along the median plane, so D1 and D3 (left ductal trees) are represented on the same tissue sections. Whole-tissue view (A) and high-magnification view (B) display morphological and chromatic effects of EtOH ablation on H&E stain (upper panels) and deduced epithelial and stromal cell classes based on computer-assisted trained classifier (lower panels). Black scale bar corresponds to 1 mm in A and white scale bar to 100 μm in B. (C) Graph bar displays cell class distribution in ductal trees (n > 4 per group) treated with different concentrations of EtOH or left untreated. Asterisks indicate p-value of unpaired Welch's t-test of each cell class per group compared to its matching cell class in the 10% EtOH-treated group (* <0.05, ** < 0.01, **** <0.0001). Please click here to view a larger version of this figure.

In our previous studies, we have shown that an ID delivery of 70% EtOH ablates the mammary epithelial cells while causing minimal damage to the surrounding tissue in rodents^6,7. In this procedure, we have demonstrated that an ID infusion of an ablative solution can be scaled up to a rabbit model. In particular, we demonstrate in this larger model with a multi-ductal tree system that sequential infusion of all ducts within a mammary gland can be successfully infused. This represents the next phase in translating this ablative procedure into a viable alternative to prophylactic mastectomy for the primary prevention in high-risk individuals.

The FDA-approved Iodine-based X-ray contrast agent Omnipaque (iohexol) allows us to assess the success of the infusion, by live visualization of the solution delivery process through the ductal tree. Using fluoroscopy to visualize the infused mammary gland corresponds closely with what is likely to be the equipment and instruments deployed to bring this image-guided procedure to the clinic^10,11. This imaging modality will allow us to know when to cease infusion, thus making live fluoroscopy a key component of the clinical implementation of this ablative procedure. We established the scalability of this procedure from rodent to rabbit ductal trees, but further research is needed to determine the percentage of EtOH that best ablates the mammary epithelial cells while minimizing surrounding tissue damage in this larger animal model. This protocol focused on the feasibility of intraductal infusion and tissue analysis immediately after the procedure. Troubleshooting strategies and useful tips are detailed in Table 1. While the main focus of this protocol is to deliver ethanol-based ablative solutions, this protocol can be used in other applications to locally deliver other imaging reagents, gene-modifying systems such as CRISPR/Cas9, lentiviral vectors, siRNAs, growth factors, and/or carcinogens^2,3 to study other aspects of mammary gland development and oncogenesis in rabbits.

In conclusion, this protocol provides the research tools needed to perform a systematic and longitudinal study of the local and systemic effects of different ablative solutions. This information will determine the safety of this procedure and pinpoint any concerns related to further development and clinical evaluation in first-in-human clinical trials.