This protocol presents a standardized method to grow VX2 cells in culture and to create an orthotopic VX2 model of endometrial cancer with retroperitoneal lymph node metastases in rabbits. Orthotopic endometrial cancer models are important for the pre-clinical study of novel imaging modalities for the diagnosis of lymph node metastases.
Endometrial cancer is the most common gynecologic malignancy in North America and the incidence is rising worldwide. Treatment consists of surgery with or without adjuvant therapy depending on lymph node involvement as determined by lymphadenectomy. Lymphadenectomy is a morbid procedure, which has not been shown to have a therapeutic benefit in many patients, and thus a new method to diagnose lymph node metastases is required. To test novel imaging agents, a reliable model of endometrial cancer with retroperitoneal lymph node metastases is needed. The VX2 endometrial cancer model has been described frequently in the literature; however, significant variation exists with respect to the method of model establishment. Furthermore, no studies have reported on the use of cultured VX2 cells to create this model as only cells propagated in vivo have been previously used. Herein, we present a standardized surgical method and post-operative monitoring method for the establishment of the VX2 endometrial cancer model and report on the first use of cultured VX2 cells to create this model.
Endometrial cancer, or cancer of the lining of the uterus, is the second most common gynecologic malignancy worldwide and the most common malignancy in developed nations1. The incidence of endometrial cancer has steadily increased, rising by 2.3% per year between 2005-2013 with a corresponding 2.2% increase in mortality1,2,3. The diagnosis of lymph node metastases is paramount as the presence of positive lymph nodes is a strong negative predictor of survival4,5,6,7 and can guide the administration of adjuvant therapy8,9,10,11,12,13. Lymph node metastases are currently diagnosed by surgically removing the lymphatic tissue overlying the major blood vessels in the pelvis and abdomen. This procedure, known as a lymphadenectomy, is controversial due to conflicting survival data from two large trials14,15,16,17,18 and the known risk of intra-operative15,19,20 and post-operative morbidity21,22,23. As current non-invasive imaging modalities do not have the required sensitivity and specificity to replace lymph node dissection24, there has been a push to develop new diagnostic imaging techniques. To test these novel techniques in a pre-clinical setting, a reliable model of endometrial cancer with retroperitoneal lymph node metastases is required.
The rabbit VX2 tumor model is a well-established model which has been used extensively to study multiple human solid organ tumors25 including lung26, head and neck27,28, liver29, kidney30, bone31,32, brain33, pancreas34 and uterus35,36,37. The VX2 model was originally developed in 1940 by Kidd and Rous38 by successfully transplanting a cottontail rabbit papilloma virus discovered by Shope in 193339. Since that time, the VX2 model has been maintained in vivo, requiring serial passage in the quadriceps muscle of White New Zealand rabbits40. More recently however, multiple groups have successfully grown VX2 cells in vitro40,41,42 and demonstrated the retained tumorigenecity of the cultured cell line31,42,43. VX2 tumors are histologically defined as anaplastic squamous cell carcinomas44 and contain glandular features which resemble adenocarcinoma26. Tumors are characterized by ease of implantation, rapid growth and hyper-vascularity44,45 and reliably metastasize, most commonly to regional and distant lymph nodes45. Similarities in uterine vascular and lymphatic anatomy46 as well as the orthotopic growth site ensure that the metastatic pattern of rabbit VX2 carcinoma mimics that of human endometrial cancer, making the VX2 model a reliable model for studying human metastatic disease. Furthermore, histologic features such as abnormal microvascular proliferation47 , as well as immunological48 and genetic similarities49,50 between humans and rabbits suggest that the tumor microenvironment may reflect that of human endometrial cancer.
Multiple groups have reported on the use of VX2 to create a model of endometrial cancer with retroperitoneal metastases with a high reported rate of success36,51,52; however, significant variation exists within the current literature with respect the method of model creation. Cell suspension doses as low as 4 x 105 cells/uterine horn51 and as high as 5 x 109 cells/uterine horn37,53 have been reported with no standard consensus on the required VX2 cellular dose. As well, a variety of inoculation methods have been reported including micro-surgical implantation of tumor into the uterine myometrium36, injection of VX2 cell suspension37,44,52,53 and in some cases, the addition of uterine horn suturing prior to innoculation52. Finally, no groups have reported the use of cultured VX2 cells to create this model. Thus, the purpose of this study is to demonstrate a successful standardized method of VX2 model creation and to report the first use of cultured VX2 cells to create a model of endometrial cancer with retroperitoneal metastases in a rabbit.
Deep abdominal wall closure: Identify the apex of the peritoneal incision and grasp the peritoneum, rectus muscle and fascia with tissue forceps.All animal studies were conducted in Animal Resource Center (ARC) approved facilities of the University Health Network and in accordance with approved animal use and care protocols (AUP #3994/#4299). VX2 cell line was obtained from Dr. Aken’s Lab at the University Health Network.
1. Creation of in vitro VX2 cell line
2. VX2 cell culture and creation of cell suspension
3. Surgical Model Establishment
4. Tumor Growth Monitoring
Twenty-eight rabbits were used for the creation of the endometrial cancer model. Rabbits had an average weight of 2.83 kg (2.71-3.58 kg) at the time of experiment. Uterine tumors successfully grew in 21 rabbits for an overall model success rate of 75%. Prior to the inclusion of uterine suturing in the protocol, the success rate was 57% compared to 81% after uterine suturing was added. Uterine suturing was added to the protocol after the 7th rabbit in response to the initial low model success rate. Five models were created from cultured VX2 cells (attempted in 8 rabbits, 63% success rate) and 16 rabbit models were created from in vivo propagated cells (attempted in 22, 73% success rate). In models created from in vivo propagated VX2 cells, a cell dose of 5 x 106 per uterine horn was used in all animals and the average number of days from inoculation to experiment was 29 days (range 24-31). In models created from cultured VX2 cells, an escalating dose protocol was used to determine the appropriate inoculation dose. Doses of 2.5 x 106 cells and 5 x 106 cells per uterine horn were unsuccessful and a dose of 10 x 106 cells per uterine horn was successful in one rabbit however, tumor growth was slow at 57 days from inoculation to experiment. A dose of 20 x 106 cells per uterine horn was successful in 4 rabbits in an average time of 45 days (range 36 – 51 days) from inoculation to experiment and was thus determined as the optimal injection dose. This data is summarized in Table 1. On PCR analysis after passage 5, the cultured VX2 cells were highly positive for both Rabbit LINE-1 and CRPV-E6 with only trace amounts of mouse LI NE-1 was identified (<0.01 pg/µL). Cells were subsequently grown in cell culture however, growth was slow with an average time of 7 days (6-9 d) to achieve flask confluency.
All models successfully resulted in the metastatic transformation of the retroperitoneal lymph nodes (Figure 2). Nineteen rabbits had pathologically confirmed lymph node metastases and 11 had pathologically confirmed extra-nodal abdominal metastases. One rabbit did not have distinct lymph nodes removed; however, it had a high burden of intra-abdominal disease in which lymph node metastases were assumed, and one rabbit died prior to the experiment. Tumors and metastatic lymph nodes from cultured VX2 cells appeared similar to tumors from in vivo propagated VX2 cells on histology (Figure 3), with dense hematoxylin stained cells invading muscle and forming glandular-like structures with many pathological mitotic figures.
Using a novel imaging agent, the Porphysome, 81 lymph nodes were identified intra-operatively and surgically removed for histologic analysis. 74 lymph nodes were left pelvic lymph nodes, 5 were right pelvic lymph nodes and 2 were right para-aortic lymph nodes. Lymph nodes removed from rabbits with in vivo propagated VX2 tumors were significantly larger and more necrotic than those removed from rabbits with cultured VX2 tumors with an average volume of 0.99 cm3 (range 0.12 – 3.89) versus 0.59 cm3 (0.01 – 2.92) (p=0.037). As well, rabbits with in vivo propagated tumors had larger more necrotic uterine tumors than rabbits with cultured cell tumors with an average length 5.6cm (4-6.8 cm) and average width of 5.2cm (3.3 – 9 cm) versus 3.6cm (2-5 cm) and of 4.56cm (3-7 cm) respectively. Finally, rabbit models made from in vivo propagated VX2 cells had more extra-nodal abdominal metastases than rabbit models made from cultured cells with 91% of all metastases found in in vivo propagated rabbits.
Figure 1: Uterine suturing. Black arrow = sutures, red arrow = uterine horns. Please click here to view a larger version of this figure.
Figure 2: VX2 tumor (A) Intra-uterine tumor. Black arrow = tumor, red arrow = uterine horns (B) Metastatic left pelvic lymph nodes. Black arrow = metastatic lymph nodes. Please click here to view a larger version of this figure.
Figure 3: Histology of cultured cell VX2 tumor (A) H&E staining demonstrates tumor infiltration of surrounding muscle (10x magnification, scale bar = 300 µm) (B) Pancytokeratin staining demonstrates densely staining tumor cells corresponding to areas of tumor on H&E (10x magnification, scale = 300 µm). Black arrow = VX2 tumor. Please click here to view a larger version of this figure.
Model Type | In vivo VX2 tumour model | Cultured VX2 tumour model |
Number of successful models | 16 | 5 |
Number of attempted models | 22* | 8 |
Model success rate | 73% | 63% |
Time from inoculation to experiment | 29 days (24-31) | 45 days (36-51) |
Successful injection dose | 1 x 107 | 40 x 106 |
(5×106 per uterine horn) | (20 x 106 per uterine horn) | |
Overall success rate | 75% | |
Overall Success rate prior to uterine suturing | 57% | |
Overall Success rate after uterine suturing | 81% |
Table 1: Model data including experimental conditions, the number of animals used, and success rate. 2 rabbits used initially for cultured cell tumors models in which growth was unsuccessful were subsequently used for in vivo tumor models
Herein, we have reported a standardized surgical method for the establishment of a VX2 endometrial cancer model and reported on the first use of cultured VX2 cells to create this model. The tumor take rate of 75% is lower than the 100% percent rate previously reported in the literature35,37,53,56; however, thw 90% rate of pathologically confirmed lymph node metastases is consistent with previous studies of this model35,53.
The inclusion of uterine suturing significantly increased the model success rate from 57% to 81% and we consider this step to be an integral part of the surgical protocol. Uterine suturing was not initially performed due to conflicting reports of the use of suturing in the literature and concerns regarding uterine horn devascularization from bilateral uterine artery ligation. Given the significant improvement in take rate with the addition of uterine suturing, we hypothesize that leakage of the cell suspension away from the injection site may have contributed to the initial low success rate. Anatomically containing the cell suspension in a small portion of the uterine horn ensures that a high local concentration of cells is exposed to the vasculature of the myometrium which likely improves tumor engraftment. Furthermore, no cases of uterine horn necrosis were noted in the experiment. Ensuring that the cell suspension is injected into the myometrium is also important as intra-uterine or extra-myometrial injections may also increase the loss of cell suspension. The uterine myometrium in rabbits is extremely thin and true intra-myometrial injection is difficult. Because of this, we hypothesize that the use of a cellular matrix scaffold such as Matrigel may improve the take rate of cells that are inadvertently injected into the uterine cavity. Despite these potential limitations, we believe that the cell suspension method is superior to previously reported microsurgical methods36 in which tumor blocks are grafted onto the uterine myometrium as this method is technically challenging. In comparison, the method here is simple, incorporates commonly used techniques and it is for these reasons, we believe it to be highly reproducible.
The in vivo propagated VX2 rabbit models were injected with a standard dose of 5 x 106 cells per uterine horn which was based upon a collaborator’s experience with VX2 rabbit models. This dose is significantly lower than reported in the literature in which cell doses as high as 1 x 108 by Harima et al.52 and 5 x 109 by both Huang37,53 and Xu35 were used. Given the short time frame in which the model was established and the high rate of both lymph node and extra-nodal metastases, we do not believe that the use of a higher dose would have improved the model. A higher dose may have promoted even more rapid and aggressive tumor spread which would impair the utility of the model. 83% of the in vivo propagated VX2 models had extra-nodal disease at the time of experiment and this it is surprising that other groups did not report the development of abdominal metastases at 30 days. A possible explanation for this difference could be inadvertent intra-vascular inoculation27 due to high pressure or high speed injection which can result in more rapid distant metastatic spread. We thus hypothesize that the speed of injection can be a factor in the rate of metastases which is why we recommend a slow injection speed in the protocol.
Comparatively, despite the higher injection dose (20 x 106 per uterine horn), only 40% of the cultured VX2 model rabbits developed extra-nodal disease and all metastatic deposits were noticeably smaller and less necrotic in these rabbits. We do not have any literature with which to directly compare the results as this is the first reported use of cultured VX2 cells to create this model. However, the findings are consistent with studies of other cultured VX2 tumors in which high inoculation doses were required and tumor growth was noted to be slow27,31. Through this experiment, we have identified the optimal injection dose of 20 x 106 cells per uterine horn which resulted in reliable growth and metastases in 80% of rabbits using an escalating dose protocol. It is possible that an even higher dose of cultured VX2 cells would result in quicker metastatic spread however as the VX2 cells grew slowly in culture, it was challenging to culture enough cells to attempt a higher dose. This is a limitation of the study and have identified this as a potential area for future investigation. However, we consider the slower growth rate in the cultured cell model to be advantageous as we believe it may replicate the clinical scenario of endometrial cancer more reliably, as endometrial cancer is generally a slow growing disease that metastasizes first to the pelvic lymph nodes and results in late distant metastases. The initial choice to propagate the VX2 cells in mice allowed for a faster turnaround, and less expensive maintenance costs; however alternatively, the cells could have been derived directly from the quadriceps muscle of rabbits.
We believe that the close post-operative monitoring for signs and symptoms of tumor growth is an important aspect of the protocol. In our experience, once rabbits develop metastatic disease, they progress rapidly to being clinically unwell, most notably in the in vivo propagated group. This rapid, aggressive growth was highlighted by the death of one rabbit from metastatic disease after a delay of only 2 days from the planned experiment date. While the speed of VX2 model establishment has been considered a strength, as similar mouse models (i.e., HEC-1 endometrial carcinoma model with lymph node metastases in mice) can take up to twice as long to establish57, these findings also demonstrate that determination of the optimal experimental timing is paramount. The findings correlate with previously studies in which tumor growth and lymph node enlargement increased significantly after post-operative day 2137,53; however we believe that there will variability with respect to the VX2 cell line used and encourage groups to understand their specific experimental timing. This timeline does not hold true for our cultured VX2 models and have identified this as an area which requires further study. To be certain about tumor growth, we chose to use both non-invasive and invasive tumor monitoring during the protocol. However, another future direction may be to refine the post-operative imaging protocol to avoid the need for a second invasive procedure.
Overall, we have reported a simple, standardized method to create a model of endometrial cancer with retroperitoneal lymphadenopathy in rabbits. Through this protocol, we have addressed the significant variability within the VX2 literature with respect to cell dose, surgical technique and post-operative model monitoring. We recognize that a further limitation of the study is that in using a VX2 cell line instead of a human xenograft we are not completely mimicking the tumor biology and microenvironment of human cancer. However, we hope other groups will use cultured VX2 cells to create their models as we believe this cell type may model human endometrial cancer more reliably through its slower growth and decreased propensity to metastasize. We encourage other groups to this fast and easy model of uterine derived retroperitoneal lymph node metastases to study novel imaging therapies to help patients with metastatic endometrial cancer.
The authors have nothing to disclose.
This study was funded by the Terry Fox Research Institute (PPG#1075), the Canadian Institute of Health Research (Foundation Grant #154326), the Canadian Cancer Society Research Institute (704718), Natural Sciences and Engineering Research Council of Canada, Сanada Foundation for Innovation and Princess Margaret Cancer Foundation.
I would like to thank Dr. Marguerite Akens for providing the initial VX2 cells for the establishment of the initial VX2 model and the frozen VX2 tumor blocks. I would like to thank Marco DiGrappa for helping to perform initial VX2 cell culture experiments and Lili Ding for helping with VX2 cell culture.
11-blade scalpel, Sterile, Disposible | Aspen Surgical (VWR) | 80094-086 | |
22-gague ear vein catheter | CDMV | 14332 | |
3-0 absorbable poly-filament suture (Polysorb) | Covidien | 356718 | |
3-0 braided absorbable suture (Polysorb) | Covidien | 356718 | |
70uM cell strainer, Individually wrapped, Nylon | Falcon | 352350 | |
Acepromazine (Atravet) | CDMV | 1047 | |
Betadine soap (Poviodone iodine 7.5%) | CDMV | 4363 | |
Betadine solution (Poviodone iodine 10%) | UHN Stores | 457955 | |
Buprenorphine | McGill University | ||
Cefazolin | UHN in-patient pharmacy | No Cat # Needed | |
Chlorhexidine solution | CDMV | 119872 | |
Corning BioCoatCellware, Collagen Type I, 100mm dishes | Corning | 354450 | brand not important |
Corning BioCoatCellware, Collagen Type I, 24-well plates | Corning | 354408 | brand not important |
Corning BioCoatCellware, Collagen Type I, 6-well plates | Corning | 354400 | brand not important |
Corning Matrigel Basement Membrane Matrix, *LDEV-free, 10 mL | Corning | 354234 | |
DMEM/HAM F12 1:1 | Life Technologies | 11320 | brand not important |
DMSO | Caledon Lab Chem | 1/10/4100 | |
Enrofloxacin (Baytril injectable) | CDMV | 11242 | |
Falcon Tube | Corning Centri-Star | 430828 | |
Fetal Bovine Serum, Qualified, Canadian Origin, 500ml | Life Technologies | 12483020 | brand/source not important |
Isoflurane | UHN in-patient pharmacy | No Cat # Needed | |
Isohexol contrast | GE Healthcare | 407141210 | |
Meloxicam (Metacam 0.5%) | CDMV | 104674 | |
Normal Saline | House Brand (UofT Medstore) | 1011 | |
PBS | Multicell or Sigma | 331-010-CL or D8537-500mL | |
Penicillin/Streptomycin (100mL; 10000U Penicillin, 10000ug Streptomycin) | Corning-Cellgro | CA45000-652 | |
Sterile Hanks Balanced Salt Solution (-Ca++, -Mg++, -Phenol Red) | T.C.M.F (Dr Bristow) | 28-Jan-11 | |
Surgical Glue (Tissue Adhesive) | 3M Vetbond | 14695B | |
Trypsin (0.25%), Proteomics Grade | Sigma | T-6567-5X20UG | |
Trypsin-EDTA, 0.05%, 100ml | Wisent Inc | 325-542-EL | brand not important |