| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
* Vascular Biology Program and Department of Surgery, Children’s Hospital Boston, Harvard Medical School, Boston, Massachusetts, USA;
Department of Radiation Oncology,
Department of Pediatric Oncology,
Department of Medical Oncology, Dana-Farber Cancer Institute, Harvard Medical School, Boston, Massachusetts, USA; and
|| Clinic for Hepatobiliary Surgery and Visceral Transplantation, University Hospital, Hamburg, Germany
1Correspondence: Vascular Biology Program, Children’s Hospital Boston, Harvard Medical School, Boston, MA 02115, USA. E-mail: judah.folkman{at}childrens.harvard.edu
ABSTRACT
The disease state of cancer appears late in tumor development. Before being diagnosed, a tumor can remain for prolonged periods of time in a dormant state. Dormant human cancer is commonly defined as a microscopic tumor that does not expand in size and remains asymptomatic. Dormant tumors represent an early stage in tumor development and may therefore be a potential target for nontoxic, antiangiogenic therapy that could prevent tumor recurrence. Here, we characterize an experimental model that recapitulates the clinical dormancy of human tumors in mice. We demonstrate that these microscopic dormant cancers switch to the angiogenic phenotype at a predictable time. We further show that while angiogenic liposarcomas expand rapidly after inoculation of tumor cells in mice, nonangiogenic dormant liposarcomas remain microscopic up to one-third of the normal severe combined immune deficiency (SCID) mouse life span, although they contain proliferating tumor cells. Nonangiogenic dormant tumors follow a similar growth pattern in subcutaneous (s.c.) and orthotopic environments. Throughout the dormancy period, development of intratumoral vessels is impaired. In nonangogenic dormant tumors, small clusters of endothelial cells without lumens are observed early after tumor cell inoculation, but the nonangiogenic tumor cannot sustain these vessels, and they disappear within weeks. There is a concomitant decrease in microvessel density, and the nonangiogenic dormant tumor remains harmless to the host. In contrast, microvessel density in tumors increases rapidly after the angiogenic switch and correlates with rapid expansion of tumor mass. Both tumor types cultured in vitro contain fully transformed cells, but only cells from the nonangiogenic human liposarcoma secrete relatively high levels of the angiogenesis inhibitors thrombospondin-1 and TIMP-1. This model suggests that as improved blood or urine molecular biomarkers are developed, the microscopic, nonangiogenic, dormant phase of human cancer may be vulnerable to antiangiogenic therapy years before symptoms, or before anatomical location of a tumor can be detected, by conventional methods.—Almog, N., Henke, V., Flores, L., Hlatky, L., Kung, A. L., Wright, R. D., Berger, R., Hutchinson, L., Naumov, G., Bender, E., Akslen, L., Achilles, E.-G., Folkman, J. Prolonged dormancy of human liposarcoma is associated with impaired tumor angiogenesis.
THE DISEASE STATE of cancer appears late in tumor development (1)
. Dormant tumors remain microscopic and do not expand in size over prolonged periods of time. These microscopic dormant tumors represent one of the earliest stages of human neoplasia. They are commonly found as occult lesions at autopsies of individuals who died of trauma, but who did not have clinically detectable cancer during life (1
2
3
4)
. Such dormant tumors also include in situ carcinomas, residual cancer cells following removal of primary tumors, and micrometastases that can remain clinically asymptomatic and undetectable for years (5
,6)
. While some types of tumor dormancy may be due to a preponderance of noncycling tumor cells, most dormant tumors consist of fully transformed, viable, proliferating tumor cells. Once these tumor cells undergo the switch to the angiogenic phenotype (7)
, an event often preceded by decreased expression of one or more endogenous angiogenesis inhibitors (8)
, these angiogenic tumors expand rapidly and kill their hosts.
Without neovascularization, even fully transformed cells, which have the capacity for uncontrolled proliferation, cannot form tumors of a clinically relevant size (7
, 8)
. Blocking various stages of tumor angiogenesis has therefore become a promising strategy for treating cancer (9)
.
Despite the high prevalence of dormant tumors, their biology is poorly understood (10)
. This is mainly because of their microscopic size, which renders them invisible to conventional imaging technology and because of the lack of suitable experimental models. We previously developed an in vivo experimental system of human liposarcoma subclones, which after inoculation into immunodeficient mice, form either nonangiogenic, dormant tumors, or angiogenic, rapidly growing tumors (11)
. This system was based on the human liposarcoma cell line (SW872) that forms rapidly growing angiogenic tumors when inoculated into mice.
In this report we have selected two of these subclones. One clone generates angiogenic tumors that grow rapidly, reach a size of approximately 1500 mm3 in 30–35 days, at which time they are also lethal to their host mice. The other clone forms dormant tumors that remain microscopic for 3–4 mo.
We have defined "nonangiogenic" tumor cells as unable to induce or to sustain the induction of new microvessels from the host. Nonangiogenic tumor cells achieve a maximum tumor diameter of 
1 millimeter. The tumor cells themselves maintain a high proliferation rate balanced by a high apoptotic rate in vivo. Therefore, dormant microscopic tumors do not expand further until a median of 3 mo, when tumors switch to the angiogenic phenotype and undergo rapid growth. We have isolated nonangiogenic tumor cells from other human tumors, not shown here. They also switch to the angiogenic phenotype, but at different times and at different percentages than the liposarcoma (11)
. We chose liposarcoma for this study because it has a higher rate of switching to the angiogenic phenotype than other human tumors.
We also demonstrate that the switch to the angiogenic phenotype from the nonangiogenic microscopic dormant state, can be accurately quantified by bioluminescence of tumors that are less than 1 mm in diameter. We further show that dormant tumors contain fully transformed cells that continuously proliferate and undergo apoptosis. This dormant phenotype is characterized by minimal or significantly reduced intratumoral microvessel density which continues to decrease to virtually no microvasculature by a few weeks after tumor cell inoculation. The scant microvessels that are observed in the early stage of microscopic dormant liposarcomas, consist mainly of short clusters of endothelial cells that lack lumens. These cannot be sustained, and gradually disappear with time. Tumor cells from nonangiogenic dormant tumors secrete high levels of the angiogenesis inhibitors thrombospondin-1 (TSP-1), and tissue inhibitor of metalloproteinases (TIMP-1).
MATERIALS AND METHODS
Cell lines and culture
All cell lines were cultured in Dulbecco’s modified Eagle’s medium (DMEM) containing 5% inactivated FBS (Life Technologies, Gaithersburg, MD), 1% antibiotics (penicillin, streptomycin) and 0.29 mg/ml L-glutamine in a humidified 5% CO2 incubator at 37°C. For in vivo imaging of luciferase-labeled cells, tumor cells from each of the clones were infected with a virus containing the firefly luciferase (FL) gene as described previously (12)
. For the proliferation assay, 2500 cells per well were seeded in 24-well plates. Every 24 h, three wells per clone were trypsinized and the number of cells was determined using a cell counter (Coulter).
Animals and tumor cell inoculation
For analysis of s.c. growth of tumors, 6-wk-old male severe combined immune deficiency (SCID) mice from the Massachusetts General Hospital (MGH), Boston, MA. were used. All of these experiments were conducted in compliance with Boston Children’s Hospital guidelines. Protocols were approved by the Institutional Animal Care and Use Committee. For monitoring growth of tumors generated by luciferase infected cells, male nude mice were used (Taconic Farms, Germantown, NY). For inoculations of tumor cells into mice, confluent tumor cells were rinsed in PBS (PBS) (Sigma, St. Louis, MO), briefly trypsinized and suspended in DMEM (without serum). The cells were washed in DMEM twice and then adjusted to a final concentration, according to type of analysis. For monitoring s.c. growth of tumors, 25 x 106 viable cells/ml were used. Each mouse was injected s.c. on the dorsal surface with 5 x 106 cells (in 0.2 ml of DMEM without serum) from a single cell line. Each mouse received only one injection of tumor cells. Tumor volumes were recorded every 3–4 days. Tumor volume was calculated using the standard formula: length x width2x 0.52. For orthotopic injections, an incision of 5 mm was made in the right flank of anesthetized nude mice. The peritoneum was exposed and 0.5 x 106 luciferase infected cells in 0.1 ml of DMEM, which were inoculated into the renal fat pad. For histological analysis of s.c. tumors, each mouse was inoculated at 4 sites on the dorsal surface with 3.8–5 x 106 viable cells/inoculation.
Immunohistochemistry
Tumor tissues from a mouse were fixed in 10% neutral buffered formalin at room temperature overnight. Tissues were embedded in paraffin, and sections (5 µm thick) were stained with hematoxylin and eosin (H&E). For immunohistochemical staining for CD31 (Pharmingen, San Diego, CA) and vascular endothelial growth factor receptor (Pharmingen), slides were treated with proteinase K at 37°C after deparaffinization, then blocked with TNT buffer for 30 min at room temperature. The slides were incubated with primary antibody (Ab) overnight at 4°C. Slides were then incubated in SA-HRB for 30 min at room temperature, biotinyl tyramide for 7 min at room temperature (PerkinElmer Life Science, Norwalk, CT), and alkaline phosphatase conjugated to avidin (Vector Laboratories, Burlingame, CA), for 30 min. The slides were then developed with the substrate Nova Red (Vector), according to the manufacturer’s instructions. Slides were counterstained for 2–4 min in Gill’s hematoxylin and mounted with a coverslip.
Apoptosis was assessed by terminal deoxynucleotidyltransferase-mediated dUTP nick end labeling (TUNEL) with ApopTag Apoptosis Detection System (Intergen, Purchase, NY) according to the manufacturer’s instruction. Tumor cell proliferation was analyzed by PCNA staining with the monoclonal anti-PCNA Ab PC10 (DAKO, Glostrup, Denmark). After deparaffinization, slides were treated with citrate buffer 0.1 M (Sigma) in a microwave oven to near boiling for 10 min and blocked with 10% goat serum in PBS with 0.3% triton-X (PBS-TX) for 30 min. The slides were then incubated with the primary Ab diluted 1:100 for 1 h. Slides were then incubated in biotinylated goat antimouse IgG (Vector) for 30 min (1:400). The slides were incubated with alkaline phosphatase conjugated to avidin (Vector), for 30 min. The slides were then incubated with substrate Nova Red (Vector), according to the manufacturer’s instructions. Slides were counterstained for 2–4 min in Gill’s hematoxylin, and mounted with a coverslip.
Imaging and determination of microvessel density
For imaging we used a Polyvar microscope (Leica Cambridge), a 3-CCD camera, and slides with high color contrast. Each slide was measured by individual frames of standard size. We measured the area (the total number of detected pixels within the field), as well as other features, including length, width, roundness, and bifurcation, using Quantimet 570 software in color detection mode (22)
.
The intensity of immunohistochemical staining of CD31 was quantified through evaluation of the Nova Red endpoint. The tumor area sampled was within the central area of the tumor. Peripheral areas of the tumor were excluded from analysis. The size of the sampled area, which varied with the size of the tumor, was divided into a grid of from 3 to 20 frames. Within each frame, quantification of the positively stained areas was achieved by operator-controlled threshholding, which summed all areas within the selected wavelengths. Five tumors from the nonangiogenic clone 4 were collected at each of the following time points: day 14, 30, and 60 after tumor cell inoculation. From these five tumors, 46, 12, and 25 frames, respectively, were imaged and quantified. Five large (over 1000 mm3) angiogenic tumors (originally from nonangiogenic clone 4) were collected following the angiogenic switch at 105–168 days (on average 138 days) after tumor cell inoculation. For quantification of microvessel density (MVD), a total of 134 frames from these tumors from clone 4 were imaged and quantified. Five large tumors (over 1000 mm3) from the angiogenic clone 9 were collected on days 31–37 after injection (day 34 on average). A total of 173 frames from tumors from clone 9 were imaged and quantified. For quantification of the percentage of proliferating cells (PCNA positively stained cells), a total of 80 frames (from 4 tumors) of clone 4 tumors were imaged and quantified. A total of 97 frames (from 5 tumors) of clone 9 tumors were imaged and quantified. For quantifying the branching sites, the incidence of two vessels arising from the same site on a vessel were automatically counted per a defined area of tumor. A total of 1138 frames from tumors on day 14 and 625 frames from tumors on day 138 (on average) were automatically counted.
In vivo detection of luciferase-labeled cells
Mice anesthetized by intraperitoneal (i.p.) injection of 155 mg/kg of ketamine and 2 mg/kg xylazine, were imaged using the in vivo imaging system (IVIS, Xenogen, Alameda, CA), as described previously (14)
. Bioluminescence was not normalized to bioluminescence at the initiation of experiment.
ELISA and Western blot for angiogenic factors from conditioned media
0.5 x 106 cells from each cell line were seeded in each well of a 6-well plate. The next day the medium was changed to DMEM without serum for ELISA analysis of TIMP-1 and TSP-1, and with 1% serum for ELISA analysis of vascular endothelial growth factor (VEGF) and basic fibroblast growth factor (bFGF), and the cells were incubated for 24 h. Conditioned medium was then collected and centrifuged to eliminate cell debris. VEGF, bFGF, and TIMP-1 levels were determined using the Quantikine ELISA kits for human VEGF and TIMP-1, and the Quantikine high-sensitivity ELISA kit for human bFGF (R&D System, Minneapolis, MN).
Soft agar assay
Assays to determine anchorage independence of tumor cells were performed as described previously (15)
.
Statistical methods
In vivo data were expressed as mean ± SE. In vitro data were expressed as mean ± SD Statistical significance was assessed using Student’s t test. P < 0.05 was considered statistically significant. All statistical tests were two-sided.
RESULTS
Analysis of s.c. tumor growth in immunodeficient mice
We previously developed an in vivo experimental system of human liposarcoma single cell-derived subclones, which form either nonangiogenic, dormant tumors or angiogenic rapidly growing tumors in immunodeficient mice (12)
. These subclones, which were not artificially genetically modified, generated tumors with distinctly different vasculature. For this study of a detailed analysis of the angiogenic potential and cellular characteristics of the tumor cells, clone 9, which forms angiogenic rapidly growing tumors, and clone 4, which most consistently forms dormant tumors, were chosen.
The angiogenic human liposarcoma clone 9 generated large, lethal tumors 30–40 days after the injection of the cells, while microscopic-sized tumors of up to 1 mm in diameter, generated from the nonangiogenic clone 4, remained undetectable by gross examination at these time points (Fig. 1
A). Cells from nonangiogenic clone 4 generated detectable tumors only after the angiogenic switch (which occurred at a range of 87–120 days after tumor cell inoculation, i.e., approximately one-third of the SCID mouse life span).
|
All of the nonangiogenic dormant tumors switched to the angiogenic phenotype and began rapid growth. These newly emerging angiogenic tumors grew as rapidly as tumors from clone 9 that were angiogenic when they were initially inoculated. In fact, angiogenic tumor cells from clone 9, grew to large sizes of approximately over 1500 mm3 diameter by 4–5 wk after tumor cell inoculation.
The most pronounced difference between sizes of tumors from the clones was seen 4–5 wk after injection. At that time, large and highly vascularized tumors developed from clone 9, while in mice injected with clone 4, tumors could be seen only on the inner side of the skin, sometimes requiring a hand-lens (Fig. 1B
). The presence of these microscopic tumors was confirmed by histochemical analysis (Fig. 1C
). Indeed, throughout the dormancy periods of tumors from clone 4, H&E staining revealed that all tumors were less than 1.5 mm diameter.
Four to five weeks after tumor cell inoculation, new functional blood vessels were abundant in angiogenic, rapidly growing tumors from clone 9. These vessels had large lumens that contained red blood cells. In contrast, these vessels were not found in the nonangiogenic dormant tumors. The only structures resembling microvasculature in the nonangiogenic dormant tumors were short clusters of endothelial cells, the majority of which had no lumens and contained no red blood cells (Fig. 1C
). Even these could not be sustained and gradually disappeared by 40–60 days (see below).
Characterization of orthotopic tumor growth
To determine whether the differences in tumor growth are influenced by the host’s s.c. microenvironment, we also studied orthotopic tumor growth. Tumor cells were labeled with firefly luciferase (FL), a cellular bioluminescence marker that allows convenient detection in vivo. These tumor cells were inoculated into the renal fat pad of nude mice. The fate of labeled cells was followed in vivo with a highly sensitive cooled charge-coupled camera (Xenogen). This method enabled us to pinpoint the location of the microscope-sized dormant tumors and to monitor the viability of the tumor cells and the precise time of the angiogenic switch in a quantitative, real time and noninvasive manner. (Fig. 2
).
|
The patterns of tumor growth in the renal fat pad were similar to those in the s.c. location (Fig. 2A
). While cells from the angiogenic clone 9 generated rapidly growing tumors, cells from the nonangiogenic clone 4 generated microscopic sized, dormant tumors that remained at the initial size of 1 mm diameter or less, for a prolonged period of 62–106 days after tumor cell inoculation. Nonangiogenic tumors from clone 4 could not be detected by gross examination of the mice, except by luciferase bioluminescence.
Angiogenic tumors from clone 9 generated large tumors (diameter of over 1 cm) by 27 days (Fig. 2B
). In contrast, nonangiogenic tumor cells from clone 4 formed dormant tumors (as quantified by luminescence). On examination of the peritoneum, large, vascularized tumors were evident in the peritoneal cavity of mice that were injected with angiogenic tumor cells from clone 9, whereas no tumors were detected following gross examination of the peritoneum of mice that received inoculations of nonangiogenic tumor cells from clone 4. The presence of microscopic, dormant tumors from clone 4 was revealed only after histological analysis of the renal fat pad (Fig. 2C
). The size of the orthotopic dormant tumors was very similar to those of dormant tumors formed in the s.c. space.
Tumor cell proliferation is not inhibited during tumor dormancy
To rule out that tumor dormancy of nonangiogenic tumors did not simply result from low or no tumor cell proliferation, we analyzed tumor cell viability and proliferation in vivo throughout the dormancy period. We collected tumors that were generated from nonangiogenic tumors at different time points following tumor cell inoculation, before and after the angiogenic switch and the subsequent tumor growth (Fig. 3
A). Tumors were harvested during the initial 2 mo after tumor cell inoculation, and also from 105 to 168 days postinoculation (the average day is 138), when after the angiogenic switch, tumor diameter had reached or exceeded 10 mm.
|
Throughout the dormancy period, tumors maintained a diameter of less than 1.5 mm (Fig. 3A
, top, H&E staining). However, in all tumors collected, numerous tumor cells staining positively for PCNA were evident (Fig. 3A
, bottom). Apoptotic cells were evident in all of the dormant tumors analyzed (data not shown). This result suggests that tumor cell proliferation is balanced by cell death. We further quantified and compared the percentage of proliferating cells in nonangiogenic tumors generated from clone 4 and angiogenic tumors from clone 9. Large tumors with a diameter >10 mm were collected 105–168 or 31–37 days post-tumor cell inoculation, respectively. 46.88% (±24.5%) of cells in nonangiogenic clone 4 tumors were positively stained for PCNA, compared to 81.3% (±18.12%) tumor cells positively stained for PCNA in angiogenic tumors from clone 9.
We confirmed the existence, location, and viability of tumor cells by following the fate of tumor cells from clone 4 that were labeled with luciferase (FL) and injected s.c. in nude mice and by monitoring the bioluminescence from the tumor cells (Fig. 3B
). The enzymatic activity of luciferase is rapid and transient and can be detected only following injection of the substrate. Throughout the dormant period (up to 120 days), tumor cells could be detected by the luciferase enzymatic activity, although no tumor was visible by gross examination on the dorsal surface of the mice. This implies that viable and metabolically active tumor cells persist and survive at the site of injection. Moreover, luminescence intensity increased 2 to 3-fold before any tumor growth could be detected by gross examination. An increase of luminescence was followed by growth of the tumor. Therefore, tumor dormancy in this model does not result from eradication, or from cell cycle arrest of the tumor cells.
Vascular structure in microscopic tumors
To examine the angiogenic potential of the nonangiogenic dormant tumors, we analyzed the presence, size, and structure of vessels in tumors by immunohistochemical staining for CD31 and VEGF receptor (VEGFR).
At least until 60 days after inoculation, the majority of the vessels inside of tumors appear as small clusters of endothelial cells that lack open lumens (Fig. 4
A). At the periphery of the dormant tumors at an early stage (day 14), abundant vessels with open lumens that contain red blood cells are observed (Fig. 4A
). This type of vessel is rare inside the dormant tumor. After the angiogenic switch, when tumors enlarge and expand in mass, many large vessels are evident. Only then, intratumoral vessels often reveal large lumens and contain red blood cells. In addition, these vessels have more branching sites (number of branching sites in tumors on day 14 was 0.3±0.85 while in tumors on day 138 there were 0.55±1.61). This pattern of vasculature is similar to that found in large tumors generated from the angiogenic clone 9 (Fig. 1C
).
|
We analyzed the microvessel density (MVD), as well as the average length of intratumoral vessels throughout the dormancy period. Interestingly, in tumors generated from nonangiogenic clone 4, the microvessel density, as well as the average length of intratumoral vessels, decreased between day 14 to 60 post inoculation (Fig. 4B
, as well as data not shown). The decrease is most pronounced in the second month post inoculation (from day 30 to day 60). The microvessel density and average intratumoral vessel length, perimeter, and area increase after the angiogenic switch (Fig. 4B
as well as data not shown). This suggests that tumor dormancy is associated with inhibition of angiogenesis. The inhibition is relieved following the angiogenic switch.
MVD was measured also in large tumors (with diameters over 10 mm). The MVD in large tumors from clone 4 (collected after the angiogenic switch at 105–168 days postinoculation) was significantly higher (3.63±2.6, P<0.01) than that in similar sizes of clone 9 angiogenic tumors at (1.98±1.51).
Comparison of angiogenic potential
We further analyzed the relative expression and secretion of angiogenic factors that are known to affect tumor angiogenesis, focusing on VEGF and bFGF.
Expression of these factors was higher in the tumor cells that formed nonangiogenic dormant tumors (Fig. 5
A and B). However, these tumor cells also released significant quantities of thrombospondin-1 (TSP-1) into their conditioned medium (Fig. 5D
). The thrombospondin-1 levels in cells that form nonangiogenic dormant tumors were consistently higher than in cells that generated angiogenic tumors (Fig. 5D
). High levels of TSP-1 were previously reported to result from elevated levels of phosphorylated Myc (15)
. Consistent with these observations, significantly higher levels of Myc protein and its phosphorylated form were found in the nonangiogenic tumor cells (clone 4) than in angiogenic tumor cells (clone 9) (Watnick, R., personal communication).
|
In the search for markers of nonangiogenic dormant tumors, we compared the protein profiles in the conditioned medium of the cell lines that generate the different tumor types. We used two-dimensional gel electrophoresis to separate proteins and bioinformatic analysis to isolate proteins that are highly expressed in the dormant tumors and not in the angiogenic rapidly growing tumors (data not shown). Using this approach, we found that the dormant nonangiogenic liposarcoma cells secrete a relatively high concentration of tissue inhibitor of metalloproteinase (TIMP-1). This result was further confirmed by ELISA for TIMP-1 (Fig. 5C
) and by a radiometric MMP/TIMP activity assay (personal communication by M. Moses and G. Louis).
Human liposarcoma cells from both nonangiogenic and angiogenic clones are fully transformed
To analyze the tumorigenicity of cells, we compared their ability to grow in an anchorage-independent manner. Tumor cells from both nonangiogenic and angiogenic clones were seeded in soft agar and the presence of colonies was recorded 34 days later. Both clones consist of fully transformed cells that are able to grow and form colonies in soft agar (Fig. 6
A).
|
Although both clones were anchorage independent, colonies generated from angiogenic tumor cells (clone 9) appeared larger and more abundant than colonies generated from nonangiogenic tumor cells (clone 4). In vitro, the proliferation of the angiogenic tumor cells from clone 9 was slightly increased compared to nonangiogenic tumor cells from clone 4 (Fig. 6B
). This is consistent with the differences of the percentage of proliferating cells in vivo and may explain, in part, the ability of angiogenic clone 9 tumor cells to form somewhat larger colonies in soft agar.
DISCUSSION
We show here that dormant, nonangiogenic human liposarcoma contains fully transformed, proliferating, neoplastic cells throughout the dormancy period (of approximately one-third of the life span of a SCID mouse) but is harmless to the host because of impaired angiogenesis. In this model, tumor dormancy is associated with the inability of tumor cells to sustain the induction of new capillary blood vessels. Proliferating tumor cells in the dormant tumor appear to be balanced by tumor cells undergoing apoptosis. Importantly, in contrast to rapidly growing tumors, which induce angiogenesis and quickly expand in size, the process of tumor angiogenesis in dormant tumors appears to be impaired. In dormant liposarcomas presented here, vessels rarely have lumens or contain red blood cells. This suggests that these dormant tumors are unable to successfully complete the process of developing functional vessels.
Throughout this study we inoculated mice s.c. with 5 x 106 tumor cells to obtain viable tumor take for both nonangiogenic and angiogenic tumors. The temporary induction of a few scant microvessels without lumens, as well as the inward migration of scattered endothelial cells into the nonangiogenic, microscopic, dormant (clone 4) tumors was at first puzzling. At the completion of the study, we realized that the initial temporary burst of impaired neovascular sprouts in the nonangiogenic tumors was possibly due to an excess of inoculated tumor cells. We now know that the excess of tumor cells in an inoculum of 5 million tumor cells contributed a temporary excess of pro-angiogenic proteins. For future experiments with this human liposarcoma we will use an inoculation size of between 500,000 and 1 million tumor cells, and we would recommend this to other investigators.
Nevertheless, the artificially high tumor cell inoculum of 5 million cells used in the present study demonstrated that even with this robust proangiogenic challenge, the nonangiogenic tumors could not sustain the temporary neovascularization contributed to them by the excess of proangiogenic factors carried with tumor cells in the inoculum. In fact, microvessel density steadily decreased after inoculation of nonangiogenic tumor cells. In other words, even given an artificial burst of proangiogenic proteins that likely accompanied the inoculation of 5 million nonangiogenic tumor cells, these cells could not sustain the neovascularization. This effect was at least, in part, due to high levels of thrombospondin-1 produced by the nonangiogenic tumor cells and by their stromal fibroblasts and other stromal cells.
At a predictable time, nonangiogenic, dormant tumors switch to the angiogenic phenotype. The molecular events that initiate the switch are not yet known but are being studied in our laboratory. Once dormant tumors undergo the angiogenic switch and initiate growth and expansion of mass, the tumor growth kinetics are similar to those of the rapidly growing angiogenic tumors. This further supports the concept that inhibition of angiogenesis, rather than slower proliferation rate of the dormant tumor cells, is the mainstay of tumor dormancy in this model. Additional mechanisms that contribute to tumor growth, other than angiogenesis, could still play a role in tumor dormancy, including interactions with stromal cells (17)
.
The angiogenic switch is regulated by several angiogenic stimulators and inhibitors produced by the tumor and host cells. In vitro the dormant liposarcoma cells secrete relatively high levels of the angiogenic proteins VEGF and bFGF, but also high levels of the endogenous inhibitors of angiogenesis thrombospondin-1 and tissue inhibitor of metalloproteinases 1 (TIMP-1). The inhibition of angiogenesis by thrombospondin and TIMP-1 is further enhanced by indirect effects. Thrombospondin-1 reduces the bioavailability and function of bFGF (18)
and VEGF (19)
. Both thrombospondin-1 and TIMP-1, whose expression is elevated in dormant tumors, interfere with MMP-9 activity (20
, 21)
, which has been reported to help mediate the angiogenic switch (22
, 23)
.
The nonangiogenic dormant tumors as well as the angiogenic rapidly growing tumors investigated here were generated from clones that were derived from the same human liposarcoma cell line, without any directed genetic modification. The patterns of tumor growth were similar at both s.c. and orthotopic sites. To our knowledge, this is the first time that dormant tumors have been monitored in vivo in a noninvasive method for such prolonged periods of time. We wish to emphasize that in addition to liposarcoma, other tumor types are likely to generate a spectrum of dormant phenotypes that differ in the sequential morphological steps during the angiogenic switch. In other words, progression toward the angiogenic switch in a dormant tumor appears to proceed at different rates for different tumor types.
This experimental system is a reproducible animal model of dormant human cancer, which recapitulates clinical dormant tumors found in humans. The dormant phase of a nonangiogenic tumor may be a potential target for antiangiogenic therapy aimed at preventing tumor recurrence, for example after initial treatment of breast cancer or colon cancer. The ongoing development, in many laboratories, of molecular biomarkers (24)
sufficiently sensitive and specific to detect the presence of microscopic dormant tumors, before or just after the angiogenic switch, without necessarily having to know anatomical location, will be an important step for the clinical application of preventive antiangiogenic cancer therapy.
ACKNOWLEDGMENTS
This work was supported by a grant from the Breast Cancer Research Foundation to J.F. by the Department of Defense Innovator Award W81XWH-04-1-0316, by a postdoctoral training fellowship from the Institute of Cancer Research/Oliver R. Grace, Jr. Fellowship and by NIH Grant CA-78496 (for L.H.), and by NIH Grant CA-64481 to J.F. We thank Dr. Vania Nose from the department of Pathology at the Brigham and Women’s Hospital, Boston, for reviewing the histological slides and for very helpful comments. We are very grateful to Kristin Gullage for photography and to David Zurakowski for reviewing the statistical calculations. We thank Carolyn Cooney for excellent technical assistance with imaging, Dr. Gil Blander, for providing iRNA constructs and generating stable infected clones, and Dr. Dipak Panigrahy for helping with orthotopic injections of tumor cells.
Received for publication June 3, 2005. Accepted for publication December 23, 2005.
REFERENCES
This article has been cited by other articles:
|
|
 |
|
  G. Lakka Klement, T.-T. Yip, F. Cassiola, L. Kikuchi, D. Cervi, V. Podust, J. E. Italiano, E. Wheatley, A. Abou-Slaybi, E. Bender, et al. Platelets actively sequester angiogenesis regulators Blood, March 19, 2009; 113(12): 2835 - 2842. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 N. Almog, L. Ma, R. Raychowdhury, C. Schwager, R. Erber, S. Short, L. Hlatky, P. Vajkoczy, P. E. Huber, J. Folkman, et al. Transcriptional Switch of Dormant Tumors to Fast-Growing Angiogenic Phenotype Cancer Res., February 1, 2009; 69(3): 836 - 844. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Ma and D. J. Waxman Combination of antiangiogenesis with chemotherapy for more effective cancer treatment Mol. Cancer Ther., December 1, 2008; 7(12): 3670 - 3684. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 Y. Cao and R. Langer A review of Judah Folkman's remarkable achievements in biomedicine PNAS, September 9, 2008; 105(36): 13203 - 13205. [Full Text] [PDF]
|
|
|
|
 |
|
 C. M. Becker, N. Rohwer, T. Funakoshi, T. Cramer, W. Bernhardt, A. Birsner, J. Folkman, and R. J. D'Amato 2-Methoxyestradiol Inhibits Hypoxia-Inducible Factor-1{alpha} and Suppresses Growth of Lesions in a Mouse Model of Endometriosis Am. J. Pathol., February 1, 2008; 172(2): 534 - 544. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
D. Cervi, T.-T. Yip, N. Bhattacharya, V. N. Podust, J. Peterson, A. Abou-Slaybi, G. N. Naumov, E. Bender, N. Almog, J. E. Italiano Jr, et al. Platelet-associated PF-4 as a biomarker of early tumor growth Blood, February 1, 2008; 111(3): 1201 - 1207. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 J. Folkman Tumor Suppression by p53 Is Mediated in Part by the Antiangiogenic Activity of Endostatin and Tumstatin Sci. Signal., September 26, 2006; 2006(354): pe35 - pe35. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 L. Rush NOTEWORTHY Vet. Pathol., July 1, 2006; 43(4): 587 - 588. [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
