We describe two methods for assessing transient vascular permeability associated with tumor microenvironment of metastasis (TMEM) doorway function and cancer cell intravasation using intravenous injection of high-molecular weight (155 kDa) dextran in mice. The methods include intravital imaging in live animals and fixed tissue analysis using immunofluorescence.
The most common cause of cancer related mortality is metastasis, a process that requires dissemination of cancer cells from the primary tumor to secondary sites. Recently, we established that cancer cell dissemination in primary breast cancer and at metastatic sites in the lung occurs only at doorways called Tumor MicroEnvironment of Metastasis (TMEM). TMEM doorway number is prognostic for distant recurrence of metastatic disease in breast cancer patients. TMEM doorways are composed of a cancer cell which over-expresses the actin regulatory protein Mena in direct contact with a perivascular, proangiogenic macrophage which expresses high levels of TIE2 and VEGF, where both of these cells are tightly bound to a blood vessel endothelial cell. Cancer cells can intravasate through TMEM doorways due to transient vascular permeability orchestrated by the joint activity of the TMEM-associated macrophage and the TMEM-associated Mena-expressing cancer cell. In this manuscript, we describe two methods for assessment of TMEM-mediated transient vascular permeability: intravital imaging and fixed tissue immunofluorescence. Although both methods have their advantages and disadvantages, combining the two may provide the most complete analyses of TMEM-mediated vascular permeability as well as microenvironmental prerequisites for TMEM function. Since the metastatic process in breast cancer, and possibly other types of cancer, involves cancer cell dissemination via TMEM doorways, it is essential to employ well established methods for the analysis of the TMEM doorway activity. The two methods described here provide a comprehensive approach to the analysis of TMEM doorway activity, either in naïve or pharmacologically treated animals, which is of paramount importance for pre-clinical trials of agents that prevent cancer cell dissemination via TMEM.
Recent advances in our understanding of cancer metastasis have uncovered that epithelial-to-mesenchymal transition (EMT) and the induction of a migratory/invasive cancer cell subpopulation are not, by themselves, sufficient for hematogenous dissemination1. Indeed, it was previously thought that metastasizing cancer cells intravasate through the entirety of cancer-associated endothelium as the tumor neovasculature is often characterized by low pericyte coverage, and as such, is highly permeable and unstable2,3,4. Although highly suggestive of defective functions within the tumor, vascular modifications during carcinogenesis do not provide evidence per se that tumor cells can penetrate blood vessels easily and in an uncontrolled fashion. Insights from intravital imaging (IVI) studies, in which tumor cells are fluorescently-tagged and the vasculature is labeled via the intravenous injection of fluorescent probes (such as dextran or quantum dots), show that, while tumor vessels are uniformly permeable to low molecular weight dextrans (e.g. 70 kD), high molecular weight dextrans (155 kD) and tumor cells can cross the endothelium only at specialized sites of intravasation which are preferentially located at vascular branch point5,6,7. Immunohistochemical (IHC) analyses using animal models and human patient-derived material have shown that these sites are "doorways" that specialize in regulating vascular permeability, locally and transiently, providing a brief window of opportunity for migratory/invasive cancer cells to enter the circulation. These doorways are called "Tumor Microenvironment of Metastasis" or "TMEM", and, quite expectedly, their density correlates with an increased risk of developing metastatic disease in breast cancer patients8,9,10.
Each TMEM doorway consists of three distinct types of cells: a perivascular macrophage, a tumor cell over-expressing the actin-regulatory protein mammalian enabled (Mena), and an endothelial cell, all in direct physical contact with each other1,5,9,10,11,12,13. The key event for the function of TMEM as an intravasation doorway is the localized release of vascular endothelial growth factor (VEGF) onto the underlying vessel by the perivascular macrophage14. VEGF can disrupt homotypic junctions between endothelial cells15,16,17,18,19, a phenomenon that results in transient vascular leakage, also known as "bursting" permeability as described in IVI studies 5. TMEM macrophages have been shown to express the tyrosine kinase receptor TIE2, which is required for VEGF-mediated TMEM function and homing of these macrophages to the perivascular niche5,20,21,22. In addition to regulating cancer cell dissemination and metastasis, TIE2+ macrophages have been shown to be central regulators of tumor angiogenesis21,22,23,24,25,26,27,28,29,30,31. As such, TIE2+ macrophages represent a critical constituent of the tumor microenvironment and the main regulator of the metastatic cascade.
To better conceptualize TMEM-mediated vascular permeability (i.e. "bursting"), it is very important to distinguish it from other modes of vascular permeability that are not associated with the dissolution of endothelial cell-cell junctions. In an intact endothelium (one whose tight and adherens junctions are not disrupted), there are three main types of vascular permeability: (a) pinocytosis, which may, or may not, be coupled to transcytosis of the ingested material; (b) transportation of material through endothelial fenestrae; and (c) transportation of material through the paracellular pathway, which is regulated by endothelial tight junctions15,16,17,18,19,32,33,34. Although deregulated in many tumors, the aforementioned modes of vascular permeability have been described mostly in the context of normal tissue physiology and homeostasis, the extremes of which are tissues with either limited permeability (e.g., blood-brain barrier, blood-testis barrier), or abundant permeability (e.g., fenestrated capillaries of the kidney glomerular apparatus)34,35,36,37.
Using multiphoton intravital imaging and multiplexed immunofluorescence microscopy, we are able to distinguish between TMEM-mediated vascular permeability ("bursting") and other modes of vascular permeability in breast tumors. To achieve this, we perform a single intravenous injection of a high-molecular weight, fluorescently-labeled probe in mice. Spontaneous bursting events can then be captured using intravital imaging in live mice; or alternatively, extravasation of the probe can be quantified by co-localization studies with blood vasculature (e.g. CD31+ or Endomucin+) and TMEM doorways using immunofluorescence microscopy. The protocols presented here describe both of these techniques, which could be used either independently or in conjunction with one another.
All experiments using live animals must be conducted in accordance with animal use and care guidelines and regulations. The procedures described in this study were carried out in accordance with the National Institutes of Health regulations concerning the care and use of experimental animals and with the approval of the Albert Einstein College of Medicine Animal Care and Use Committee (IACUC).
1. Evaluation of "bursting permeability" using live animal imaging
2. Evaluation of extravascular dextran using fixed tissue analysis
The experimental procedures described in this protocol article are briefly summarized and illustrated in Figure 1A-C.
To measure TMEM-mediated vascular permeability ("bursting activity") and to reduce experimental noise from other modes of vascular permeability (i.e. transcellular and paracellular, as explained in the introduction), we performed intravenous (i.v.) injection of high molecular weight probes, such as 155 kDa Dextran, conjugated to tetramethyl rhodamine. Lower molecular weight dextrans did not distinguish the three modes of vascular permeability. Our prior experience has indicated that the systemic circulation of 155 kDa TMR-Dextran efficiently labeled the vasculature of both normal tissues, such as the mammary gland and the lungs (Figure 2A,B), as well as of neoplastic tissues, such as primary breast cancer and breast cancer metastases in lungs (Figure 2C,D). As also confirmed by prior studies5,46, 155 kDa TMR-Dextran is thus a suitable probe for measuring “bursting” in both primary and secondary tumor sites. A characteristic example of peak bursting activity (dotted yellow area), using multiphoton intravital imaging in a primary MMTV-PYMT breast tumor is illustrated as a series of still images (Figure 3A). Bursting activity was always noted in association with a TMEM doorway (dotted white circle), characterized by the spatial juxtaposition of a Dendra2+ tumor cell a CFP+ macrophage and endothelium (Figure 3A). As mentioned, each TMEM doorway consists of a perivascular macrophage, a Mena overexpressing tumor cell, and an endothelial cell, all in direct physical contact with each other. While Mena has not been explicitly labeled in the mouse models used in these experiments, it has previously been shown to become overexpressed in late stage PyMT carcinoma47. This simplifies the identification of TMEM doorways and eliminates the need to explicitly label Mena in the tumor cells.
To assess TMEM-dependent vascular permeability in fixed tissue analysis, we performed the procedure described in this protocol in mice which were transplanted with tumor chunks taken from 14-week old MMTV-PyMT donor mice. When mice reached an appropriate tumor size, they received a single i.v. dose of 155-kDa TMR-Dextran, as described in Part 2 of this protocol, and were sacrificed after 1 hour. The tumor tissues were collected and subjected to IF analysis. The endomucin fluorescent signal was used as an exclusion mask to the dextran fluorescent signal, allowing for discrimination between highly-permeable and low, or non-permeable, blood vessels (Figure 3B). As previously described in Karagiannis et al.48, a sequential slide was also stained with the previously-established TMEM triple stain and aligned with the corresponding IF slide. Using this approach, we confirmed that the highly-permeable blood vessels within the breast tumor tissue have at least one associated TMEM doorway (Figure 3B).
Figure 1. Summary of procedure. (A) Dendra-2 labeled invasive carcinoma of the breast (green) taken from a transgenic animal [FVB/N-Tg(MMTV-PyMT)mul-Tg(MMTV-iCre)Jwp-Tg(loxP-stop-loxP-Pdendra2)Jwp] and cut into small pieces. One piece is then orthotopically transplanted into another transgenic animal whose macrophages were labeled by cyan fluorescent protein (cyan) [FVB/N-Tg(Cfms-gal4-vp16)-(UAS-eCFP)]. The orthotopically transplanted to tumor is then allowed to grow, after which the mouse is allocated to the appropriate treatment arm. (B) Once the treatment is finished, the mouse is then taken for intravital imaging. Once anesthetized, a skin flap surgery is performed then stabilized with a rubber backing. A shallow imaging window is then affixed over the tumor. Finally, the mouse is then placed on the microscope stage where the window fits into the custom-made stage plate and the images can then be acquired. (C) Alternatively, the mouse from A is administered high molecular weight dextran and sacrificed after 1 hour. The tumor is then removed, fixed, and processed for paraffin embedding. Sections of the tissue are cut, stained for TMEM in IHC and dextran and vessels in IF, and scanned on a digital whole slide scanner. Finally, images from the two scans are aligned and individual fields of view are chosen for analysis and quantification of vascular leakage. Please click here to view a larger version of this figure.
Figure 2. Visualization of TMR-Dextran labeled vasculature in healthy and diseased tissues. (A) Image of a healthy developing mammary gland within a transgenic animal whose vasculature is labeled by TMR-Dextran (red), mammary ductal epithelium labeled by the fluorescent protein Dendra-2 (green), and macrophages labeled by a cyan fluorescent protein (blue) [FVB/N-Tg(loxP-stop-loxP-Pdendra2)Jwp-Tg(Cfms-gal4-vp16)-(UAS-eCFP)]. (B) Healthy lung tissue visualized through a lung imaging window within a mouse whose vasculature is labeled by TMR-Dextran (red). Blue = SHG from collagen fibers. (C) Dendra2 labeled invasive ductal carcinoma taken from a transgenic animal [FVB/N-Tg(MMTV-PyMT)mulxTg(MMTV-iCre)Jwp-Tg(loxP-stop-loxP-Pdendra2)Jwp] and orthotopically transplanted into another transgenic animal whose macrophages are labeled by cyan fluorescent protein (blue) [FVB/N-Tg(Cfms-gal4-vp16)-(UAS-eCFP) and vessels labeled with TMR-Dextran (red). (D) Lung metastases eight days after iv injection of tumor cells (E0771; labeled by the fluorescent protein Clover) into a transgenic animal whose macrophages are labeled by cyan fluorescent protein (cyan) [FVB/N-Tg(Cfms-gal4-vp16)-(UAS-eCFP)] and vasculature labeled by TMR-Dextran (red). Please click here to view a larger version of this figure.
Figure 3. Visualization of TMEM-mediated vascular permeability using intravital imaging and fixed tissue analysis. (A) Stills from a time-lapsed intravital imaging movie of 155-kDa TMR-Dextran labeling of neo-angiogenic vessels (red) within a Dendra-2 labeled invasive carcinoma of the breast (green) taken from a transgenic animal [FVB/N-Tg(MMTV-PyMT)mul-Tg(MMTV-iCre)Jwp-Tg(loxP-stop-loxP-Pdendra2)Jwp] and orthotopically transplanted into another transgenic animal where the macrophages were labeled by cyan fluorescent protein (cyan) [FVB/N-Tg(Cfms-gal4-vp16)-(UAS-eCFP)]. The dotted yellow line denotes the outline of the transient vascular leakage area before (t = 0'), during (t = 17') and after (t = 52') the leakage (bursting) event. The dashed white circle points to a TMEM doorway, as captured in live imaging. (B) Multichannel immunofluorescence of Endomucin (first column), 155-kDa dextran-TMR (second column), their merged image along with DAPI (third column), the thresholded blood vessel and extravascular dextran masks (fourth column), and the corresponding sequential section of TMEM IHC (fifth column) in MMTV-PyMT mice. Top row: Vascular profile away from TMEM, appearing as a "non-leaky" vascular profile. Bottom row: TMEM-associated vascular profile, appearing as a "leaky" vascular profile. Please click here to view a larger version of this figure.
Here, we outline two protocols that can be applied to visualize and quantify a specific type of vascular permeability which is present at TMEM doorways and is associated with the disruption of vascular tight and adherens junctions. This type of vascular permeability is transient and controlled by the tripartite TMEM cell complex, as explained above5. The ability to identify and quantify TMEM-associated vascular permeability is crucial for the assessment of a pro-metastatic cancer cell microenvironment, as well as for pre-clinical studies that examine the effect of conventional cytotoxic therapies on tumor microenvironment as well as anti-metastatic potential of targeted and non-targeted therapies. Importantly, we demonstrated here that this assessment can be done by either intravital imaging or immunofluorescence in fixed tissues. Both approaches have their advantages and disadvantages. Intravital imaging allows direct visualization of the dynamic process of vascular permeability, but requires specialized equipment and training. On the other hand, immunofluorescence performed in fixed tissues offers only a static image, but can be routinely performed in most laboratories.
Cancer patients are commonly treated with some form of systemic therapy, and the success of systemic treatment is usually assessed by the evaluation of parameters related to cancer cell survival such as apoptosis, proliferation, or gross tumor size. However, it is equally important to understand how these systemic therapies affect cancer cell dissemination given that most cancer related mortality occurs due to metastasis, a process which involves both cancer cell proliferation as well as cancer cell dissemination1. For example, recent studies have demonstrated that chemotherapy can induce a significant increase in the density of TMEM doorways in mouse and human breast cancer48,49. Moreover, it was shown that chemotherapy induces TMEM activity and results in hematogenous dissemination of cancer cells, increased circulating cancer cells, and increased lung metastasis48. The chemotherapy-induced increase in TMEM activity can be blocked by systemic administration of drug that inhibits TIE2 function and therefore TMEM activity, called rebastinib14,48. Thus, the protocols described here can offer valuable endpoint measurements for the effect of various systemic treatments on TMEM activity through the comparison of treated to non-treated cohorts of mice.
In recent years, the concept of antiangiogenic therapy in cancer has been revisited, suggesting that a vascular "normalization" strategy (i.e. the transient reconstitution of the abnormal structure and function of blood vessels), may be preferable to targeting blood vessels for destruction, as it makes the neovasculature more effective for drug delivery50,51,52,53,54. In view of this emerging hypothesis, other groups have also developed assays for assessing vascular permeability and for testing the efficiency of vascular normalization strategies55. It should be mentioned that, in principle, these methods are utilized to assess TMEM-independent modes of vascular permeability. Therefore, the selection of the most appropriate vascular permeability assay depends on the scope of a study, and researchers must ensure that they comprehend the applicability of each of the published vascular permeability assays, before applying them to their studies.
As mentioned above, each of the two methods described here has certain advantages over the other. For example, measuring "bursting permeability" in vivo offers valuable kinetic information, but since bursting is a rare event, prolonged imaging sessions of several hours each may be required to be able to obtain sufficient data. In addition, intravital imaging requires specialized equipment which may not available to all investigators. On the other hand, evaluation of TMEM activity in fixed tissues is relatively easy to perform and requires only standard laboratory equipment. In addition, the analysis of TMEM activity in fixed tissues is easier to interpret, but lacks kinetic information. Moreover, the fixed tissue analysis is less specific, since it may capture cumulative leakage of dextran over time from transcellular and paracellular vascular permeability of intact endothelia, or even a sudden leakage event due to spontaneous vascular injury. Thus, the synergistic use of the two methods would make the interpretation of the TMEM-mediated vascular permeability more specific and more sensitive. While we have not explicitly tested other applications of the techniques presented here, they may prove useful for investigating induced vessel permeability, for example by infection or by pharmaceutical therapies.
In summary, the measurement of TMEM-associated vascular permeability, as outlined by two independent methodologies here, provide useful tools for assessing the pro-metastatic tumor microenvironment and its associated TMEM activity, and can be used to some extent by any laboratory. These methods are particularly useful in the pre-clinical assessment of the effect of systemic therapies on cancer cell dissemination. In addition, they can be used to assess the effect of agents that can block chemotherapy-induced TMEM activity, and lastly, these methods can be used to assess the best combination therapy preceding phase I patient trials.
The authors have nothing to disclose.
We would like to thank the Analytical Imaging Facility (AIF) in the Albert Einstein College of Medicine for imaging support. This work was supported by grants from the NCI (P30CA013330, CA150344, CA 100324 and CA216248), the SIG 1S10OD019961-01, the Gruss-Lipper Biophotonics Center and its Integrated Imaging Program, and Montefiore’s Ruth L. Kirschstein T32 Training Grant of Surgeons for the Study of the Tumor Microenvironment (CA200561).
GSK co-wrote the manuscript, performed imaging for figure 1C and 3B, developed fixed tissue analysis protocol, and analyzed and interpreted all data; JMP co-wrote the manuscript, and performed the surgery and intravital imaging for Figure 1B,2C and 3A; LB & AC performed the surgery and intravital imaging for Figure 2B; RJ performed the surgery and intravital imaging for Figure 2A; JSC co-wrote the manuscript and analyzed and interpreted all data; MHO co-wrote the manuscript and analyzed and interpreted all data; and DE performed the surgery and intravital imaging for Figure 2D, co-wrote the manuscript, developed fixed tissue analysis and intravital imaging protocols, and analyzed and interpreted all data.
Anti-rabbit IgG (Alexa 488) | Life Technologies Corporation | A-11034 | |
Anti-rat IgG (Alexa 647) | Life Technologies Corporation | A-21247 | |
Bovine Serum Albumin | Fisher Scientific | BP1600-100 | |
Citrate | Eng Scientific Inc | 9770 | |
Cover Glass Slips | Electron Microscopy Sciences | 72296-08 | |
Cyanoacrylate Adhesive | Henkel Adhesive | 1647358 | |
DAPI | Perkin Elmer | FP1490 | |
Dextran-Tetramethyl-Rhodamine | Sigma Aldrich | T1287 | |
DMEM/F12 | Gibco | 11320-033 | |
Endomucin (primary antibody) | Santa Cruz Biotechnology | sc-65495 | |
Enrofloxacin | Bayer | 84753076 v-06/2015 | |
Fetal Bovine Serum | Sigma Aldrich | F2442 | |
Fish Skin Gelatin | Fisher Scientific | G7765 | |
Insulin Syringe | Becton Dickinson | 309659 | |
Isofluorane | Henry Schein | NDC 11695-6776-2 | |
Matrigel | Corning | CB40234 | Artificial extracellular matrix |
Needle (30 G) | Becton Dickinson | 305128 | |
Phosphate Buffered Saline | Life Technologies Corporation | PBS | |
Polyethylene Tubing | Scientific Commodities Inc | BB31695-PE/1 | |
Pulse Oximeter | Kent Scientific | MouseOx | |
Puralube Vet Ointment | Dechra | NDC 17033-211-38 | |
Quantum Dots | Life Technologies Corporation | Q21561MP | |
Rubber | McMaster Carr | 1310N14 | |
TMR (primary antibody) | Invitrogen | A6397 | |
Tween-20 | MP Biologicals | TWEEN201 | |
Xylene | Fisher Scientific | 184835 |