We outline a protocol that implements both in vivo and ex vivo approaches to study ovarian cancer colonization of peritoneal adipose tissues, particularly the omentum. Furthermore, we present a protocol to quantitate and analyze immune cell-structures in the omentum known as milky spots, which promote metastases of peritoneal adipose.
High-grade serous ovarian cancer (HGSC), the cause of widespread peritoneal metastases, continues to have an extremely poor prognosis; fewer than 30% of women are alive 5 years after diagnosis. The omentum is a preferred site of HGSC metastasis formation. Despite the clinical importance of this microenvironment, the contribution of omental adipose tissue to ovarian cancer progression remains understudied. Omental adipose is unusual in that it contains structures known as milky spots, which are comprised of B, T, and NK cells, macrophages, and progenitor cells surrounding dense nests of vasculature. Milky spots play a key role in the physiologic functions of the omentum, which are required for peritoneal homeostasis. We have shown that milky spots also promote ovarian cancer metastatic colonization of peritoneal adipose, a key step in the development of peritoneal metastases. Here we describe the approaches we developed to evaluate and quantify milky spots in peritoneal adipose and study their functional contribution to ovarian cancer cell metastatic colonization of omental tissues both in vivo and ex vivo. These approaches are generalizable to additional mouse models and cell lines, thus enabling the study of ovarian cancer metastasis formation from initial localization of cells to milky spot structures to the development of widespread peritoneal metastases.
Unlike most solid tumors, metastases from high-grade serous ovarian cancer (HGSC) are limited to the peritoneal cavity 1. Thus, effective peritoneal therapies could potentially control or eradicate HGSC. Currently, a standard therapeutic approach is surgical cytoreduction combined with chemotherapy 1-3. Unfortunately, the vast majority of patients experience and succumb to complications of disease recurrence. These dismal statistics show the need for improved understanding of metastatic colonization, the process by which cancer cells localize to, utilize, and proliferate within host tissues to form metastases 2.
The omentum is a preferred early site of HGSC metastasis 4-7. Unlike other peritoneal adipose, omental adipose tissue contains unusual immune structures known as milky spots, which contain B, T, and NK cells and macrophages, which play key roles in peritoneal homeostasis8,9. In addition to their physiologic functions, we found that milky spots play an active role in ovarian cancer metastatic colonization 4. In experimental metastasis assays, SKOV3ip.1, CaOV3, and HeyA8 (human) and ID8 (murine; C57BL/6) ovarian cancer cells rapidly home to milky spots, suggesting that the cells are moving toward a secreted chemotactic factor. Interestingly, cancer cells do not colonize peritoneal adipose lacking milky spots (i.e., gonadal and uterine fat) 4.
In order to identify mechanisms regulating milky spot colonization, we have optimized xenograft models that enable the interrogation of cellular and molecular events over the time course of metastatic colonization 4. A specific advantage of the approaches described herein is its emphasis on tissue architecture and function, which enables users to test hypotheses in fully integrated in vivo and ex vivo models of metastatic colonization 4,10. By comparing cancer cell localization and growth in peritoneal fat depots which either contain or lack milky spots, investigators can test the relative contribution(s) of adipocytes and cells within milky spots to the lodging and progressive growth of ovarian cancer cells in physiologically relevant tissues.
All mice were housed, maintained and euthanized according to Institutional Animal Care and Use Committee (IACUC) guidelines and under the supervision of the University of Chicago Animal Resource Center.
1. Preparing Animals for Experimental Studies
2. Identification and Isolation of Peritoneal Fat Depots.
3. Milky Spot Identification Using Giemsa Staining
4. Propagation and Preparation of Ovarian Cancer Cells for In vivo and Ex vivo Studies
5. Intraperitoneal Injection of Cells
Note: In our experiments, mice are handled in a laminar airflow or biosafety cabinet within our barrier facility in order to limit the risk of pathogen exposure, In order to demonstrate this technique clearly, the procedures in the accompanying video were conducted in a laboratory approved for animal work. This particular technique does not require the animals to be under anesthesia. Under approved protocols, perform this technique on live animals. Use appropriate strains of mice for this study. For example: use immunocompromised, athymic nude mice for the study of human ovarian cancer cell (SKOV3ip.1, and HeyA8) colonization; use immunocompetent C57BL/6 mice for study of mouse ovarian cancer cells (ID8) colonization.
6. Milky spot colonization In vivo
7. Milky Spot Colonization Ex Vivo
Identification and Histologic Examination of Peritoneal Fat Depots
Gross anatomic dissection allows identification of four of the five primary sources of peritoneal fat (Figure 1A). Moving clockwise from the top center are: the omentum (OM; outlined) located over the stomach and spleen, the gonadal fat (GF) surrounding the left ovary (ov), the uterine fat (UF) attached to the uterine horns (uh) and the mesentery (MY) attached to the small intestine (si). The ovary (ov), uterine horn (uh), and small intestine (si) are indicated as points of reference. The splenoportal fat (SP) surrounds the splenic artery and connects the hilum of the spleen to the celiac artery (Figure 1B). Finally, the omentum is attached to the stomach and pancreas (Figure 1C). The gross structure and relative size of these tissues is shown in Figure 1D. Histologic examination of the tissues shows that splenoportal and omental fat contain milky spot [MS] structures at the tissue periphery, surrounded by adipocytes [A]; uterine, gonadal, and mesenteric fat do not (Figure 1E).
Milky spot identification using Giemsa Staining
To enable the comparison of overall milky spot content between different mouse strains, milky spot volume and area present in individual omenta was quantitated using a novel method. As described in Methods, a “digital whole mount” of the omenta was constructed by staining every third section of tissue in a 5% Giemsa solution and capturing images of the stained tissues (Figure 2A). Milky spots appear as regions staining dark blue (Figure 2B). The identity of these regions as milky spots was confirmed in serial sections by both H&E staining and IHC for cells positive for CD45, a common lymphocyte marker (Figure 2C and D, respectively). As described in Clark and Krishnan et al4, images (Figure 2A) were compiled and processed to construct a “digital whole mount” which could then be used to calculate the milky spot volume and area in individual omenta.
Quantitative and qualitative assessment of cancer colonization of omentum
A flow cytometry-based method was developed in order to quantitate the number of cancer cells present in adipose depots after intraperitoneal injection. This protocol is especially useful to study early colonization steps of the metastatic process. For example, in Figure 3A, ID8-tdTomato cells were prepared and injected intraperitoneally into C57BL/6 mice. At 7 days post injection (dpi), the adipose organs were harvested and dissociated into a single-cell suspension as described in the Methods. The number of tdTomato cells was quantified using flow cytometry (Figure 3A, left). The omentum showed an approximately 12-fold increase in the number of tdTomato-positive events, relative to PBS- injected controls (Figure 3A, right), but there was no significant increase in cell preparations from the gonadal fat, uterine fat, or mesentery. A complementary qualitative IHC-based approach also showed that 7 days after i.p. injection of SKOV3ip.1 cells, comparable cancer cell lesions were observed in both omental and splenoportal fat (Figure 3B). IHC for human pan-cytokeratin (pan-CK) confirms the presence of cancer cells within the milky spots. The specificity of IHC staining was confirmed using an IgG control for the pan-cytokeratin antibody (Figure 3B). ID8 ovarian cancer colonization of peritoneal fat depots was evaluated using C57BL/6 mice at 7 dpi. Large cancer cell lesions in the milky spots of both the omentum and splenoportal fat are outlined. Small clusters of cancer cells were occasionally seen in other fat depots, as illustrated in the panel inset. Thus, one can access both the relative number and location of cells within a given tissue.
Ex vivo assay to study ovarian cancer colonization of milky spots
In order to address the limitations of in vivo conditions while maintaining tissue composition and architecture, we optimized an ex vivo approach to study ovarian cancer cell colonization of omental tissues. Whole omental tissues excised from mice are successfully cultured ex vivo to study ovarian cancer interaction with milky spots. SKOV3ip.1-GFP cells were seeded onto the ex vivo cultured omental tissues and maintained at 37 °C for a period of 6 hr. GFP-positive fluorescent cells were visible in discreet foci in both ex vivo and in vivo omenta harvested 6 hr post injection of SKOV3ip.1-GFP cells (Figure 4A). Under both conditions, histological verification using H&E staining confirmed cancer cell localization to milky spots (Figure 4B). Confirmatory stain using cytokeratin stain for cancer cells showed the presence of cancer cell foci within the milky spots (Figure 4C). These data demonstrate the feasibility of using ex vivo approaches for mechanism-based studies to complement in vivo findings.
Figure 1. The relative locations and structures of the main peritoneal adipose depots. (A) Gross anatomic dissection showing the relative location of four of the five primary sources of peritoneal fat. The omentum (OM; outlined), gonadal fat (GF), uterine fat (UF), mesentery (MY), ovary (ov), uterine horns (uh), and small intestines (si) are shown. (B) Splenoportal fat (SP; outlined) exposed by lifting the spleen with forceps. (C) The mouse omentum shown dissected free from the pancreas to improve visualization. D: Relative sizes and structures of excised peritoneal fat; the mesentery is shown attached to the mesenteric root. E. Histologic evaluation of peritoneal fat for the presence of milky spots (A) adipocytes, (MS) milky spot. Please click here to view a larger version of this figure.
Figure 2. Giemsa-based approach for quantifying milky spots in omental tissue. (A) Image of section of whole omentum stained with Giemsa. Serial sections of omental tissue evaluated by: (B) Giemsa staining; milky spots are indicated with black arrows. (C) H&E staining; and, (D) IHC using anti-CD45 antibody to identify lymphocytes within the milky spot structure. The scale bar is the same for B-D and denotes 100 µm. Please click here to view a larger version of this figure.
Figure 3. Ovarian cancer cells preferentially colonize peritoneal fat that contains milky spots. Flow cytometric analyses of omentum (OM), uterine fat (UF), gonadal fat (GF), and mesentery (MY) harvested from mice at 7 dpi of ID8-tdTomato cells (left). Data are presented as fold-change increase of tdTomato-postive cells over PBS-injected mice (right). Error bars indicate standard error of the mean. ** denotes p<0.01 compared to PBS controls. (B) Examination of tissues by both standard histology and IHC shows comparable colonization of milky spots in both omentum and splenoportal fat (after injection of 1×106 SKOV3ip.1 cells into Nude mice). Sections were evaluated by H&E staining. The presence of epithelial (cancer) cells within the lesions was confirmed by IHC detection of cytokeratin using a pan-cytokeratin (pan-CK) antibody. IHC using an IgG isotype antibody for pan-cytokeratin was used as a control for staining specificity. The scale bar is the same for all images and denotes 100 µm. (C) Evaluation of ID8 ovarian cancer colonization of peritoneal fat depots in C57BL/6 mice at 7 dpi. Please click here to view a larger version of this figure.
Figure 4. Ex vivo colonization of ovarian cancer cells to omentum.The columns represent a comparison of the three experimental conditions showing naïve omenta (left), omenta from mice 6 hr post i.p. injection of 1 x 106 SKOV3ip.1 cells (center) and omenta with SKOV3ip.1-GFP colonized under ex vivo conditions (right). Naïve omenta (Blank), omenta colonized by SKOV3ip.1-GFP from in vivo assay (attachment in vivo) or omenta colonized by SKOV3ip.1-GFP in ex vivo culture conditions (attachment ex vivo). Fluorescence imaging of whole omenta for GFP shows similar colonization pattern for both in vivo and ex vivo assays (A). H&E staining of serial sections of the omenta from panel A confirms the presence of cancer cells localized to the milky spots (B). Cytokeratin for ovarian cancer cells validates the H&E staining (C). Please click here to view a larger version of this figure.
Development of therapies to target disseminated cells requires a mechanistic understanding of metastatic colonization, the critical first step in the development of peritoneal disease. To address these issues we report approaches that can be used to discern how the omentum’s unique tissue composition and architecture promote ovarian cancer metastatic colonization. Distinguishing features of our approach are: 1) the focus on early events in the colonization process; 2) comparison of milky spot-containing and deficient adipose depots; 3) availability of a reproducible technique to quantitate milky spot volume in tissues; 4) availability of complementary in vivo and ex vivo approaches to evaluate ovarian cancer cell colonization of milky spot structures.
In conducting these studies we came to realize that investigators frequently are unaware of the location of the mouse omentum and the existence of the splenoportal fat band, which serves as a second source of milky spot-containing peritoneal fat. Although the techniques described herein enable studies of steps in metastatic colonization, they only offer “snapshots” of these processes, and not assessment of colonization in real time. If followed appropriately, the protocols as described should yield reproducible results to a range of investigators. Critical to the success of these techniques are: 1) Identifying and dissecting out specific peritoneal adipose tissues as contamination by other tissue types will skew quantitative data obtained from flow cytometry; and 2) Implementing appropriate strains of transgenic mice for the study in question. It should be noted that the time course studies should be determined for individual cell lines. Empirically as it depends on a variety of factors including proliferation rate, time to tumor take and metastatic potential.
In addition to dissecting the interactions of well-established ovarian cancer cell lines and their omental microenvironment that are important to metastasis formation, the technique described herein can also be used to evaluate the malignant potential of clinically derived cell lines that recapitulate early molecular alterations in serous tubal intraepithelial carcinomas in the fimbriae of the fallopian tubes 12-17. Although a pathway for the molecular progression of these lesions is emerging, in vivo studies are needed to determine the biological significance of specific alterations. Complementary in vivo and ex vivo assays can be used to assess specific steps in metastatic colonization of omental tissue by such model precursor cell lines Information gleaned from studies may be key to the generation of therapeutic hypotheses and the design of further preclinical studies.
Identifying mechanisms that control metastatic colonization may enable the prevention or control of metastasis formation. For example, women who are at increased risk for developing ovarian cancer 18-22 (i.e., BRCA1 or BRCA2 gene mutation carriers) often undergo prophylactic salpingo-oophorectomy to reduce the risk of developing peritoneal disease 23,24. Unfortunately, patients often still develop HGSC due to the outgrowth of ovarian cancer cells present within the omentum and other metastatic sites at the time of surgery. Identification of agents that disrupt omental tissue microenvironment could prevent cancer cell localization and/or growth within the omentum. Such approaches could potentially be used in combination with therapeutic agents that target ovarian cancer cells to prevent metastasis formation or suppress cancer regrowth (after surgical debulking).
The authors have nothing to disclose.
Supported by grants from the Department of Defense (W81XWH-09-1-0127), the NIH (2-R01-CA089569), the Elsa U. Pardee Foundation, a Marsha Rivkin Center for Ovarian Cancer Research Pilot Study Award, and generous philanthropic support from Section of Urology and Section of Research in the Department of Surgery, University of Chicago.
Dissection Tools | Fine Science Tools | NA | |
Geimsa | Fluka/Sigma Aldrich | 48900 | |
5% Formalin | Sigma | HT501320 | |
DMEM | Corning | 10-013-CV | |
Trypsin | Gibco | 25200-056 | |
PBS | Corning | 21-040-CV | Without calcium and Magnesium |
26 gauge needle | BD | 329652 | |
BSA | Sigma | A7906 | |
Collagenase | Worthington | LS004196 | |
Stomacher | Seward Labsystems | Stomacher 80 Biomaster | |
Microstomacher bag | Stomacher Lab Systems | BA6040/Micro | |
ACK Lysis Buffer | Gibco | A10492-01 | |
Millicell culture plate insert | Millipore | PICM01250 | |
Cell-Tak | BD | 354240 |