We describe an in vivo murine model of perineural invasion by injecting syngeneic pancreatic cancer cells into the sciatic nerve. The model allows for quantification of the extent of nerve invasion, and supports investigation of the cellular and molecular mechanisms of perineural invasion.
Cancer cells invade nerves through a process termed perineural invasion (PNI), in which cancer cells proliferate and migrate in the nerve microenvironment. This type of invasion is exhibited by a variety of cancer types, and very frequently is found in pancreatic cancer. The microscopic size of nerve fibers within mouse pancreas renders the study of PNI difficult in orthotopic murine models. Here, we describe a heterotopic in vivo model of PNI, where we inject syngeneic pancreatic cancer cell line Panc02-H7 into the murine sciatic nerve. In this model, sciatic nerves of anesthetized mice are exposed and injected with cancer cells. The cancer cells invade in the nerves proximally toward the spinal cord from the point of injection. The invaded sciatic nerves are then extracted and processed with OCT for frozen sectioning. H&E and immunofluorescence staining of these sections allow quantification of both the degree of invasion and changes in protein expression. This model can be applied to a variety of studies on PNI given its versatility. Using mice with different genetic modifications and/or different types of cancer cells allows for investigation of the cellular and molecular mechanisms of PNI and for different cancer types. Furthermore, the effects of therapeutic agents on nerve invasion can be studied by applying treatment to these mice.
Nerves form a specific tumor microenvironment that stimulates cancer growth and migration1,2,3. Perineural invasion (PNI) is the process through which cancer cells invade in and around the nerves. It may be considered as a unique route of metastasis since cancer invasion extends away from the sites of origin along nerves. PNI is found in several cancer types including pancreatic, prostate, head & neck, salivary, cervical, and colorectal cancers with an incidence ranging from 22% to 100%1,2. PNI is associated with pain and correlates with poor prognosis and worse survival rates1,2.
Developing models of perineural invasion is essential to elucidate the cellular and molecular mechanisms of this process, and to test candidate therapeutic agents to decrease PNI. In vitro methods of studying interactions between cancers and nerves include the co-culture of cancer cells with nerve explants4, with dorsal root ganglions5,6,7, or with specific cells from the nerve microenvironment such as Schwann cells7. In vivo approaches, however, are more physiologically relevant, include the use of cancer mouse models in which cancer has been induced or transplanted and have the advantage of accounting for the entire nerve microenvironment. In orthotopic models of pancreatic or prostate cancer, PNI has been reported8,9,10 and the incidence of PNI may be recorded, but because of the small size of the nerves in those organs, it is difficult to see the entire nerve and therefore to quantify the extent of PNI. The model we describe here is an in vivo model of PNI in which cancer cells are injected into the sciatic nerve of mice through a simple surgical procedure11. The heterotopic transplant invades within the nerve toward the spinal cord. The length of the nerve invasion from the site of injection to the spinal cord may be measured, as well as the volume of the cancer within the nerve. Importantly, the invaded nerve can also be collected for a variety of assays including microscopic, and molecular analyses. A variety of cancer cells can be tested, and the host mice that have been genetically modified or treated with specific compounds may be used as well. This powerful assay allows for the cancer cells and the host microenvironment to be modified for investigation into the mechanisms of PNI.
All of the procedures with animal subjects were approved by the Institutional Animal Care and Use Committee at Memorial Sloan Kettering Cancer Center.
1. Preparation of the Cancer Cells
2. Preparation of the Mice and Surgery
Note: 8-week-old, male and female C57BL/6J mice are used in this study. The surgery conditions follows the IACUC rules of our institution. The instruments are sterilized, the surgical working surface is disinfected, the animal is disinfected and the surgeon wears sterile gloves.
3. Sciatic Nerve Extraction
4. Nerve Processing and Quantification
This method describes the surgical implantation of pancreatic cancer cells into the murine sciatic nerve to create an in vivo model of quantifiable nerve invasion. Figure 1 illustrates the anatomical location of the sciatic nerve and the site of injection. Figure 2 shows the two sciatic nerves of a nude mouse. A PBS injected nerve (left) may be compared to a nerve injected with MiaPaCa-2 cancer cells (right). The nerve injected with cancer cells is infiltrated by tumor, and appears both thicker and darker than the nerve injected with PBS.
Figure 3 shows the quantifiable differences of sciatic nerve invasion by pancreatic cancer cell Panc02H7 in wild type and NCAM KO mice. NCAM KO mice are mice in which the adhesion molecule neural-cell adhesion molecule 1 is deficient13. Length and area of invasion have been quantified using imaging software. Invasion was found significantly reduced in NCAM KO mice7.
Figure 1: The mouse sciatic nerve and the site of injection. Please click here to view a larger version of this figure.
Figure 2: The two dissected sciatic nerves of a mouse. A PBS injected nerve (left) and a nerve injected with MiaPaCa-2 cancer cells (right) are shown. Black arrows show the site of injection. Please click here to view a larger version of this figure.
Figure 3: H&E Longitudinal sections of sciatic nerves injected with Panc02H7 cancer cells in wild type and NCAM KO mice. White arrows indicate the site of injection and proximal side of spinal cord. The black arrows indicate the length of invasion with corresponding values (mm). Scale bar: 5 mm (top images) and 0.2 mm (bottom images). Quantification of nerve invasion, as measured by length and area of invasion. Data represent mean ± SD. n = 8 (NCAM1 WT), n = 7 (NCAM1 KO), **P < 0.005, ***P < 0.0005 using t test. (Results from Deborde et al., 20167). Please click here to view a larger version of this figure.
In this protocol we describe an in vivo murine model of perineural invasion that allows for the quantification of sciatic nerve invasion by pancreatic cancer cells. This model enables the study of molecular mechanisms of nerve invasion. Successful experiments using this technique require a careful approach to three critical steps in the process: 1) the injection of cancer cells (steps 2.7, 2.8), 2) the extraction of invaded nerves (step 3.4), and 3) processing of harvested nerves (step 4.1).
The injection should be done with great caution to avoid puncturing through the back wall of the nerve, which leads to a loss of cancer cells. Once the needle is in place, the injection should be done at a slow and steady pace over 5 s to prevent the hydraulic force from propelling the cancer cells too far forward from the point of injection. Leave the needle in the nerve after injection for an additional 3 s will minimize back flow and cancer cell leakage. During sciatic nerve extraction, dissection must be very gentle as the invaded sciatic nerve is very fragile. Cancer invasion disrupts the normal cellular organization of the sciatic nerve. Even though invaded nerves appear thicken in diameter compared to non-injected or non-invaded nerves, they are more prone to breaking with even minor stretching and manipulation. During embedding of the excised nerve in OCT for processing, it is very important to insure that the nerve is laid completely flat onto the mold. This step allows for the sectioning of the sample to capture the entire length of the nerve for proper analysis. Solid tumor on the nerve often needs to be trimmed to prevent the nerve segment proximal to the tumor from raising the nerve off the bottom of the mold. The entire length of the nerve must be laid completely flat against the mold.
Many modifications can be considered while using this protocol. When a small metal spatula is not available to be placed under the sciatic nerve, the surface of a pair of wide forceps can be used instead. A forceps may be used to hold the needle in place after insertion, although the downsides of this approach are a diminished view of the injection and potential to crush the nerve if excessive force is used. Another way is to use a fine forceps to lift the nerve, then opening the forceps to stretch the nerves. This method is easier to execute and provides both tension and stabilization during injection, but may damage the nerve with excessive stretching of the nerve. The injection can be done in either a proximal or distal direction, but we have found that in both cases, the direction of the cancer invasion is typically towards the spinal cord. Therefore, we prefer injecting distally to further minimize hydraulic force that may push the cancer cells towards the spinal cord, and confound later measures of length of invasion. We advise that it is important to be consistent in the choice of direction of injection within a set of experiments.
Although we present a model of syngeneic grafting, a similar technique can be used for xenografts human cancer cells into immunodeficient mice, with only slight variations. Xenografts in nude mice may required longer periods of cancer invasion to yield similar lengths of nerve invasion as in syngeneic models. Different cancer cell lines may also be used. PNI is not only prevalent in the cancer of the pancreas, but also of the head and neck, prostate, and skin cancers. Our lab has successfully injected nude mice in the sciatic nerve with human pancreatic cancer cell lines MiaPaCa-2 and Panc17,14,15.
Possible adverse events include wound dehiscence, as the mice occasionally chew their stitches resulting in opening of the wound. In that case, the wound would need to be re-stitched. We check the wound daily for the first 3 post-operative days. The mice may also walk with a slight limp postoperatively, likely due to pain from both the surgical procedure as well as invasion of their sciatic nerve with cancer. This is usually not severe enough to require pain medication.
This model has limitations. Some amount of stretching and pinching of the nerve during the procedure might create damages in the nerve. In addition, it is difficult to control the exact number of cells going into the nerve. Furthermore, this model is limited to the study of cells that have already invaded nerves. It is not a model to study the interaction of nerves with cancer cells located in the primary tumor. The main disadvantage of using the sciatic nerves versus nerves of the gastrointestinal tract is that the sciatic nerve is not a metastatic site of pancreatic cancer. The difference in the composition and the nature of the nerve fibers could influence the invasion. However, the accessibility and very small size of the pancreatic nerves would not allow such a study.
Our key suggestions for the beginners would be to remove the hair in a wide area around the planned incision sites so that the landmarks can be easily identified and to make sure to have a steady position before beginning to inject. After many injections, the performer should be able to get 100% of animals with nerve invasion.
This model allows for the study the interactions between cancers and nerves in an animal model. It can be applied to a wide range of oncologic studies by using genetically modified mice, cancer cells, or both. Pharmacological treatments against tumor invasion along nerves can also be tested by comparing length and area of invasion between treatment and control groups. These different possible applications make this model an extremely versatile and powerful tool for studying PNI.
The authors have nothing to disclose.
The authors acknowledge the technical services provided by the molecular cytology facility and the animal facility of Memorial Sloan Kettering Cancer Center. This work was supported by NIH grants CA157686 (to R.J. Wong) and P30 CA008748 (Memorial Sloan Kettering Cancer Center support grant).
Mouse | Number and age variable depending on experimental needs | ||
Cell culture media (PBS, Trypsin, and DMEM+10% FBS) | Any | Steps 1.1, 1.2, 1.3. | |
Conical centrifuge tube, 50 mL | Falcon | 352098 | Step 1.1 |
Microcentrifuge tube 1.5 mL | Axygen | MCT-150-C-S | Step 1.2 |
Electric razor | WAHL | 9962 | Step 2.1. Can be substituted with commercial hair removal agent |
Isoflurane, 250 mL | Baxter | 1001936060 | Step 2.2 |
Hypoallergenic surgical tape | 3M Blenderm | 70200419342 | Step 2.3 |
Betadine Swapsticks | PDI | SKU 41350 | Step 2.4 |
Webcol Alcohol Preps | Covidien | 5110 | Step 2.4 |
Sterile surgical tools (scissors and forceps) | Steps 2.4, 2.5, 3.3, 3.4, 3.5 | ||
10 μL Hamilton syringe | Hamilton | 80308 | Steps 2.7, 2.8 |
Steel Micro spatula | Fisher Scientific | S50823 | Step 2.7 |
Dissecting microscope | Step 2.7 | ||
Bupivacine, 1 g | Enzo Life Sciences | BML-NA139-0001 | Step 2.9. Reconstitute to 0.5% |
5-0 Nylon suture | Ethicon | 698H | Step 2.9 |
Tissue-Tek O.C.T. Compound | VWR | 25608-930 | Step 4.1 |
Tissue-Tek Cryomold Molds | VWR | 25608-916 | Step 4.1 |