New models of hematogenous ovarian cancer metastasis demonstrate preferential spread to the ovary and a requirement for the ovary for abdominal dissemination

Abstract

Emerging evidence suggest that many high-grade serous "ovarian" cancers (HGSOC) start in the fallopian tube. Cancer cells are then recruited to the ovary and then spread diffusely through the abdomen. The mechanism of ovarian cancer spread was thought to be largely due to direct shedding of tumor cells into the peritoneal cavity with vascular spread being of limited importance. Recent work challenges this dogma, suggesting hematogenous spread of ovarian cancer may play a larger role in ovarian cancer cell metastasis than previously thought. One reason the role of vascular spread of ovarian cancer has not been fully elucidated is the lack of easily accessible models of vascular ovarian cancer metastasis. Here, we present 3 metastatic models of ovarian cancer which confirm the ability of ovarian cancer to hematogenously spread. Strikingly, we observe a high rate of metastasis to the ovary with the development of ascites in these models. Interestingly, oophorectomy resulted in a complete loss of peritoneal metastases and ascites. Taken together, our data indicate that hematogenously disseminated HGSOC cells have a unique tropism for the ovary and that hematogenous spread in ovarian cancer may be more common than appreciated. Furthermore, our studies support a critical role for the ovary in promoting HGSOC cell metastasis to the abdomen. The models developed here represent important new tools to evaluate both the mechanism of cancer cell recruitment to the ovary and understand and target key steps in ovarian cancer metastasis.

Figure 1

Murine intravenous injection of SKOV3 ovarian cancer cells result in a high rate of ovarian metastasis. A) percent of mice with organ specific metastasis after IV SKOV3 tumor cell injection Bi) representative macroscopic ovarian metastasis ii) H&E stain of metastatic ovary tumor iii) human nuclear antigen stain (green) of metastatic ovary tumor demonstrating human origin of tumor iv) human PAX8 stain of metastatic ovary tumor.

Figure 2

Intravenous injection of multiple ovarian cancer cell lines result in ovary metastasis. A) percent of mice with organ specific metastasis after IV tumor cell injection with multiple ovarian cancer cell lines B) representative macroscopic ovarian metastasis C) dilated murine bowel/bowel obstruction after A2780 IV injection D) H&E stain of metastatic OVCAR3 ovary tumor E) human nuclear antigen stain (green) of metastatic OVCAR3 ovary tumor demonstrating human origin of tumor.

Figure 3

Murine subcutaneous ovarian tumors metastasis to the ovary. A) Summary of metastatic sites after subcutaneous and intraperitoneal murine ovarian cancer cell injection B) representative picture of macroscopic ovarian metastasis from the subcutaneous model. C) H&E stain of ovarian tumor metastasis from the subcutaneous model.

Figure 4

Subcutaneous primary human ovarian cancer cells in combination with HemSC and CA-MSCs form spontaneous ovarian cancer metastasis. A) tabular display of percent of mice which developed metastasis per axillary xenograft composition B) p53 stain of pt366+HemSC+CA-MSC xenograft C) CD34 stain demonstrating CD34+ vasculature surrounding tumor cells in flank pt366+CA-MSC+HemSC tumors indicating lymphovascular space invasion (arrows).

Figure 5

Oophorectomy decreases the burden of abdominal metastasis in IV tail vein injection model. A) Prevalence of metastatic sites following IV tail vein injection of OVCAR3 cells demonstrates no ascites or kidney/adrenal metastasis in mice with oophorectomy vs mock surgery group which developed ovarian tumors, ascites and kidney metastasis. B) Quantification of the number of gross liver lesions identified in control vs oophorectomy mice demonstrating significant decrease in burden of liver lesions in oophorectomy mice. C) Representative liver of (i) control mouse (ii) oophorectomy mouse demonstrating fewer liver lesions in the oophorectomy mice. D) (i) image of dissected uterine horn from oophorectomy mouse demonstrating nodularity corresponding with tumor deposits demonstrated in (ii, iii) H&E sections of the paraffin embedded uterine horn.