Spontaneous Reversion of the Angiogenic Phenotype to a Nonangiogenic and Dormant State in Human Tumors

Published OnlineFirst February 26, 2014; DOI: 10.1158/1541-7786.MCR-13-0532-T

Molecular Cancer Genomics Research

Michael S. Rogers1,3,4, Katherine Novak1,3, David Zurakowski2, Lorna M. Cryan1,3,4, Anna Blois1,3,7, Eugene Lifshits1,3, Trond H. Bø, Anne M. Oyan5, Elise R. Bender1,3, Michael Lampa1,3, Soo-Young Kang1,3, Kamila Naxerova4, Karl-Henning Kalland5,6, Oddbjorn Straume7,8, Lars A. Akslen7, Randolph S. Watnick1,3,4, Judah Folkman†1,3,4, and George N. Naumov1,3,4

Abstract The angiogenic switch, a rate-limiting step in tumor progression, has already occurred by the time most human tumors are detectable. However, despite signiﬁcant study of the mechanisms controlling this switch, the kinetics and reversibility of the process have not been explored. The stability of the angiogenic phenotype was examined using an established human liposarcoma xenograft model. Nonangiogenic cells inoculated into immunocompromised mice formed microscopic tumors that remained dormant for approximately 125 days (vs. <40 days for angiogenic cells) whereupon the vast majority (>95%) initiated angiogenic growth with second-order kinetics. These original, clonally derived angiogenic tumor cells were passaged through four in vivo cycles. At each cycle, a new set of single-cell clones was established from the most angiogenic clone and characterized for in vivo for tumorigenic activity. A total of 132 single-cell clones were tested in the second, third, and fourth in vivo passage. Strikingly, at each passage, a portion of the single-cell clones formed microscopic, dormant tumors. Following dormancy, like the original cell line, these revertant tumors spontaneously switched to the angiogenic phenotype. Finally, revertant clones were transcriptionally proﬁled and their angiogenic output determined. Collectively, these data demonstrate that the angiogenic phenotype in tumors is malleable and can spontaneously revert to the nonangiogenic phenotype in a population of human tumor cells.

Implications: Leveraging the rate of reversion to the nonangiogenic phenotype and tumor dormancy may be a novel anticancer strategy. Mol Cancer Res; 12(5); 754–64. 2014 AACR.

Introduction thus cease macroscopic growth—is rare (2). Nevertheless, In cancer, a tumor's switch to angiogenesis is a rate- there are a limited number of reports of tumors that cease limiting step in its progression from microscopic to macro- macroscopic growth, suggesting that such a reversion of the scopic size (1). As a result, small, occult tumors are a common angiogenic switch may occur (3, 4). finding on autopsy of individuals who die of nonneoplastic Regulation of angiogenesis is typically viewed as a switch causes (primary studies summarized in ref. 2). In contrast, the whose state is governed by the relative local concentration of converse finding—of tumors that switch off angiogenesis and angiogenesis stimulators and inhibitors. The switch meta- phor is common in the field because of the strong biphasic nature of angiogenesis. In experimental models of pathologic conditions, it is rare to observe small, stepwise accumulation Authors' Affiliations: Departments of 1Surgery and 2Anesthesia; 3the Vascular Biology Program, Boston Children's Hospital; 4Harvard Medical of additional vessels. Rather it is common to observe that School, Boston, Massachusetts; 5Department of Microbiology, Haukeland once angiogenesis has been initiated, vessel growth proceeds University Hospital; 6Section for Microbiology, The Gade Institute; 7Centre for Cancer Biomarkers CCBIO, Department of Clinical Medicine; and throughout the pathologic process. In addition, variation in 8Section of Oncology, Institute of Internal Medicine, University of Bergen, intensity of angiogenesis has been observed, and seems to be Bergen, Norway a major rate-limiting factor in tumor growth. For example, Note: Supplementary data for this article are available at Molecular Cancer substantial variation in microvessel density is observed Research Online (http://mcr.aacrjournals.org/). throughout tumors, with regions exhibiting the highest † Died January 14, 2008. density predicting the overall growth rate of a tumor, metastatic status, and patient survival (5–7). Corresponding Author: Michael S. Rogers, Boston Children's Hospital, 11.211 Karp Family Research Bldg., 300 Longwood Avenue, Boston, MA In contrast with the tumor as a whole, we and others have 02115. Phone: 617-919-2252; Fax: 617-730-0231; E-mail: shown that individual tumor cells can exhibit significant [email protected] variation in their ability to induce angiogenesis. For example, doi: 10.1158/1541-7786.MCR-13-0532-T when individual cell clones derived from a primary human 2014 American Association for Cancer Research. liporsarcoma are implanted in immunocompromised mice,

754 Mol Cancer Res; 12(5) May 2014

Reversion of Angiogenesis

the timing of the transition to macroscopic growth varies the second, third, and fourth in vivo passage, respectively. In from 7 to >160 days (8). In addition, several commercially vitro growth curves were generated using a Coulter counter. available tumor cell lines, originally derived from human tumors that were macroscopic in size, exhibit extended Subcutaneous tumor growth periods of preangiogenic growth, ranging from a few weeks, Male SCID mice, 6 to 8 weeks of age (MGH), were cared to years, to the lifespan of the animal (9). However, an for in accordance with the standards of the Institutional experimentally induced increase in the angiogenic output of Animal Care and Use Committee, and used under an these tumors (e.g., by transfection with VEGF; refs. 10, 11) approved protocol. Tumors were generated by injecting 5 results in early and rapid macroscopic growth. Similarly, 106 cells in 0.2 mL PBS subcutaneously into the shaved alterations occurring during extended evolution of dormant lower right flank. Injection sites were palpated weekly for tumors in mice can result in increased net angiogenic output. tumors (typically detected at 50 mm3) and measured every In at least one system, this was accompanied by a decrease in 3 to 4 days thereafter, with volumes calculated using the expression of angiogenic inhibitors (e.g., thrombospondin- formula 0.52W2L. Mice were euthanized when tumor 1), rather than an increase in angiogenic stimulators (e.g., volume reached approximately 1,000 mm3 or mice showed VEGF; ref. 12). Importantly, in all of these experiments in signs of discomfort. vitro growth rates for the angiogenic sublines (which are derived from nonangiogenic parental lines) did not differ Immunohistochemistry and histology significantly from the growth rate of nonangiogenic sublines. Excised tumors were rinsed in ice-cold PBS and fixed in These findings exclude differences in cell division rate as a 4% paraformaldehyde. Paraffin-embedded tissue was sec- mechanism for the observed differences in macroscopic tioned (4 mm) and representative sections (5/tumor) were growth. stained with H&E, proliferating cell nuclear antigen (PCNA; Finally, experiments in which angiogenic cells were PC10; 1:150; DAKO), and CD31 (PECAM; 1:250; BD admixed with nonangiogenic cells before inoculation in mice Biosciences). have demonstrated that even a minority of proangiogenic cells is sufficient to induce growth (and metastasis) in the entire Quantitation of angiogenesis regulators tumor (11). Nontransformed (i.e., stromal) cells have also Serum-free medium (15 mL) was incubated on day-old been shown to play a critical role in the induction of cultures (5 106 cells/15-cm plate) for a further 24 hours. angiogenesis in some tumors (13). These observations lead Concentrations of human VEGF165 (R&D), bFGF (R&D), to the notion that the angiogenic switch may be an ensemble and Tsp-1 (CYT Immune Science) were measured by ELISA property comprised of contributions from all the cells in the kits using the manufacturers' protocols with control values tumor, rather than an obligate property of only the tumor- (serum-free media) subtracted and results normalized per igenically transformed cells in a tumor (10). Therefore, it is 104 cells. Protein lysates from angiogenic and revertant cell possible that individual tumor cells, although derived from an clones were western blotted as previously described (13) for angiogenic tumor, may not possess the angiogenesis-inducing Tsp-1 (Ab11; LabVision), and b-actin (Abcam), using potential of that tumor. We sought to test that possibility by horseradish peroxidase–conjugated goat anti-mouse (Jack- serial in vivo passage, cloning, and quantitative analysis of the son Immunoresearch Labs) for detection. growth phenotypes of such individual tumor cells. In addition to assessing the stability of the switch to an angiogenic Gene expression analysis phenotype in individual cells, these experiments allowed us to Independent angiogenic (two) and nonangiogenic assess the nature of the events that lead to the angiogenic (three) cell lines from the second cycle of cloning were switch, and determine that it is not comprised of a single step. analyzed using the Agilent Human Whole Genome (4 44k) Oligo Microarray (Agilent Technologies Inc.) as fi Materials and Methods previously described (15). Signi cance analysis of microarrays was performed in J-Express (www.molmine.com; Generation of single-cell clones ref. 16), by dividing samples into angiogenic and non- Angiogenic (clone-9) and nonangiogenic (clone-4) cell angiogenic (revertant) groups. Functional and pathway lines were established from the SW-872 cell line as previ- analysis of genes upregulated in revertant or angiogenic fi in vivo ously described (8, 14). For the rst passage, a new cell clones by 1.5-fold or more was generated using Ingenuity line (clone-9#2) was generated following subcutaneous Pathways Analysis (Ingenuity Systems; www.ingenuity. fi inoculation into a severe combined immunode cient com) as well as GOEAST (http://omicslab.genetics.ac. (SCID) mouse. Subclones with similar growth rates were cn/GOEAST; ref. 17). prepared for subsequent in vivo passage from monolayers of the clone-9#2 cell line following limiting dilution (average Real-time reverse transcription PCR 0.2 cells/well) in 96-well plates. Each clonal population was Real-time reverse transcription PCR (RT-PCR) analysis expanded to approximately 30 106 cells and inoculated was performed in triplicate on 20 ng cDNA prepared using into 5 SCID mice. The most angiogenic clone was selected to Thermoscript RT-PCR system using validated Taqman gene produce single-cell clones for the next in vivo selection cycle. expression assays (Applied Biosystems). Following outlier fi Speci cally, 31, 33, and 68 single-cell clones were tested in exclusion (samples with a Ct SD > 0.3), angiogenic and

www.aacrjournals.org Mol Cancer Res; 12(5) May 2014 755

Rogers et al.

nonangiogenic samples were compared by DDCt using glyc- second is more clock-like and is often governed by accumu- eraldehyde-3-phosphate dehydrogenase as endogenous lation of specific factors until a threshold is reached, thereby control. triggering the phenotypic change. For the angiogenic switch, the mutation model predicts stochastic conversion of cell Statistical analysis populations to angiogenic growth, versus maintenance of the Time to palpable tumor for the angiogenic and nonan- nonangiogenic phenotypes. In contrast, if the delay in giogenic cells was compared using the Kaplan–Meier meth- angiogenic growth is the result of a "clock," then one might od, with survival curves compared using log-rank test (16), expect drift in the timing of the clock to earlier or later time and confidence intervals determined using Greenwood for- points in various clones as a result of differences in angio- mula (15). For a subset of tumors, we compared the median genic/antiangiogenic factor accumulation, but relatively sta- time from initial palpation to a size of 500 mm3 using the ble dormant times within a given cell clone. Mann–Whitney U test. Two-tailed values of P < 0.05 were To better understand the nature of the process underlying considered statistically significant. the switch to the angiogenic phenotype (and exponential tumor growth), we examined the distribution of switch times Analysis of escape kinetics (Fig. 1C and D). If the process were clock-like (e.g., results Growth curves were fit to a model of exponential growth from local accumulation of a proangiogenic factor to a after a variable delayÀÁ using least squares fitting to the following predetermined threshold), escape times would be distributed mtðÞt equation: dtðÞ¼ d0 if t < t0; d010 0 if t > t0 with m in a Gaussian fashion. In contrast, if the process is governed d t n (growth rate), 0 (dormant tumor size), and 0 (length by events (e.g., mutations or other heritable regulatory of dormant period) as free variables. The hypothesis that events), then escape times will follow a multistage model ktðÞt n the data might be consistent with different growth rates curve of the form e 0 (18, 19). Comparing these (vs. tumor growth initiation times) was rejected by F test models using Akaike Information Criterion shows that in P 300 t fi d m ( < 10 ) after holding 0 to zero and again tting and our experimental model there is greater evidence for the for each animal, excluding the possibility that a small number mutation model (evidence ratio >109). Fitting the distribu- of exponentially growing tumor cells led to the delay in tumor tion of escape times to the multistage model equation yielded growth observed for the "non-angiogenic" tumors. a best-fit value of 1.8 for n (with n ¼ 2 fitting marginally worse; P ¼ 0.029). These findings indicate that, on average, Results approximately two changes were necessary to convert a In vivo growth of angiogenic and nonangiogenic human nonangiogenic tumor to an angiogenic tumor. – liposarcoma single-cell derived clones Not all clones from angiogenic tumors retain the growth We previously described an animal model of human phenotype of the original tumor liposarcoma, based on the phenomenon that all human Because the above analysis indicates that the angiogenic tumors are comprised of angiogenic and nonangiogenic fi switch requires two distinct mutations, it also predicts an tumor cell populations (8). Speci cally, two single-cell intermediate phenotype in which cells possess only one clones were derived from the original tumor population: mutation. The intermediate phenotype would manifest in one displayed angiogenic properties (clone-9) and one was in vivo a growth curve that displayed a delay in the acquisition of nonangiogenic (clone-4). Here, we characterize the exponential growth that was longer than the early-growing growth potential of these clonal populations in detail using a angiogenic phenotype and shorter than the dormant phe- xenograft model (Fig. 1). We observed that 100% of SCID n ¼ notype. Thus, to prove our model, we sought to identify such mice ( 13) inoculated with angiogenic clone-9 cells an intermediate. To that end, we passaged the original initiated exponential tumor growth early (within 1 week of 3 clonally derived, early growing, angiogenic tumor cells inoculation) and reached a mean size of 1,500 mm within (clone-9 cells) through 4 cycles in SCID mice (Fig. 2), with 1 month of inoculation. In contrast, mice inoculated with the idea that a fraction of such tumors would contain cells the nonangiogenic clone-4 cells developed only microscopic 3 fi that only carried a single mutation. tumors (less than 50 mm ) in the rst week after tumor cell After each in vivo cycle, the fastest-growing early tumor inoculation. The majority (97%, 78 of 80 mice) of clone-4 was used to establish a cell line and produce a new set of tumors, however, spontaneously initiated exponential tumor – in vivo single-cell derived clones used for testing of angio- growth at a median of 125 days ( 25 days, SD), a phe- genic activity. We selected 31, 33, and 68 single-cell clones nomenon previously described as the "angiogenic switch" in the second, third, and fourth in vivo passages, respectively. (9). Remarkably, 2.5% of clone-4 tumors (2 of 80 mice) These clones were chosen (from over 200 single-cell clones remained at this small size for more than 300 days. per in vivo passage) under in vitro conditions, such that they exhibited similar proliferation rates. Thus, any differences in Growth escape kinetics are inconsistent with single-event in vivo growth kinetics would not be due to intrinsic causation proliferation rates. Each clonally derived tumor cell popu- Biologic transitions are typically thought of in terms of two lation was inoculated into 5 SCID mice for determination of different processes. In the first, a transition is regulated by a in vivo growth phenotype. Observation of these tumors small number of discrete events such as mutations. The revealed that 40% to 70% of clones retained the growth

756 Mol Cancer Res; 12(5) May 2014 Molecular Cancer Research

Reversion of Angiogenesis

Angiogenic Nonangiogenic A clone clone B 3,000 100 90 Nonangiogenic ) 2,500 n n 3 = 13 = 80 80 clone 2,000 70 60 Angiogenic clone 1,500 50 40 1,000 30 Tumor volume (mm Tumor volume 500 20 10

0 Freedom from palpable tumor (%) 0 0 100 200 300 0 50 100 150 200 250 300 Time (days) Time (days) C D

Figure 1. Growth and survival of angiogenic versus nonangiogenic liposarcoma xenografts and analysis of escape kinetics. A, size of subcutaneous xenografts of 5 106 liposarcoma cells from an angiogenic (gray) and nonangiogenic (black) clone originating from the same human tumor. B, freedom from palpable tumors of mice bearing these tumors. C, the time of escape for each of 80 mice is plotted (^) along with the fraction of mice that have not yet had their tumor escape from dormancy. Best-fit 1- and 2-event models are indicated by the dashed and solid curves, respectively. Residual fraction remaining for a two-event model is also indicated. D, tails and residuals of the curve fit highlighting differences in the model fits. properties of the parental line, whereas 20% to 50% dis- 76% of the clones grew early, 3% were revertant, and played an intermediate phenotype (Fig. 3). 21% were intermediate. As part of the fourth in vivo passage, Strikingly, after each in vivo passage, 3% to 6% of the 43% grew early, 6% were revertant, and 51% were inter- single-cell clones derived from the fastest-growing early- mediate. The overall distribution of early-growing, hybrid, arising clone of the previous passage remained at a micro- and revertant cell lines is consistent with that resulting scopic, dormant size for a median of 72 days. In other words, from a 26% chance of individual reversion events in a these cells reverted to a dormant phenotype (Fig. 3). Fol- two-event model (55%, 38%, 7%; P > 0.7 by c2 test). The lowing this dormancy period, most of the dormant tumors time it took for a tumor to reach 50 mm3 following spontaneously switched to the fast-growing phenotype, a inoculation has been summarized in Fig. 3 using Kaplan– process that was regulated by the acquisition of angiogenic Meier curves for each in vivo passage. Importantly, the potential in the original clone-9. In contrast, others reverted presence of a "revertant" group of clones, characterized with to the nonangiogenic clone-4 phenotype, remaining dor- a dormancy period and delay of exponential tumor growth mant for more than 200 days (Fig. 3B, E, and H). Specif- start was evident at each in vivo cycle. ically, during the second in vivo passage, 55% of the clones had an early-growth phenotype, 6% demonstrated a "rever- Early-growing and revertant clones did not differ in tant" phenotype (dormant period >1 month for at least 4/5 proliferation in vitro or in vivo, but did exhibit mice, i.e., in vivo growth similar to nonangiogenic clone-4 differences in expression of angiogenic regulators. cells), and 39% displayed the intermediate or "hybrid" We then sought to determine the underlying mechanism phenotype (Fig. 3). During the third in vivo passage, governing the in vivo properties of the early-growing and

www.aacrjournals.org Mol Cancer Res; 12(5) May 2014 757

Rogers et al.

Figure 2. Scheme for serial cloning, expansion, and phenotypic analysis of human liposarcoma cells.

revertant clones. We ﬁrst examined whether the two types clones and three early-growing clones from the second in vivo of clones possessed different in vitro growth kinetics. We passage and generated in vitro growth curves (Fig. 5). compared the in vitro proliferation potential of two revertant Importantly, we did not observe a signiﬁcant difference in

Figure 3. Growth pattern of clones derived from highly angiogenic tumor cell lines. Left, examples of growth patterns observed for the second (top), third (middle), and fourth (bottom) in vivo passages of the most angiogenic clone from the previous cycle. Middle, the fraction of clones exhibiting each of the phenotypes shown on the left. Right, tumor-free survival for clones exhibiting each of the three phenotypes shown on the left.

758 Mol Cancer Res; 12(5) May 2014 Molecular Cancer Research

Reversion of Angiogenesis

Figure 4. Gross appearance of ABC tumors in mice. A, mice bearing an angiogenic clone. B, time-matched mice bearing a revertant clone. C, appearance of nonangiogenic clone in the subcutaneous space.

growth rates in vitro between the two types of clones. We also VEGF and the angiogenesis inhibitor thrombospondin-1 compared the in vivo growth rates via PCNA staining of (Fig. 6A–C). VEGF was chosen because manipulations that dormant and exponentially growing tumors. Although there induce VEGF expression have been shown to induce expo- was no significant difference in the fraction of nuclei that nential, angiogenic growth in otherwise dormant tumors stained positive for PCNA (Supplementary Fig. S1), we (11). Similarly, we have previously observed downregulation observed an increase in apoptotic cells in dormant tumors of thrombospondin-1 in dormant tumors that spontane- compared with exponentially expanding ones (8). Another ously initiate exponential growth (20). The revertant clones possible explanation for the failure of the microscopic lesions (for example, clone 10) demonstrated significant (4-fold formed by the revertant clones to grow macroscopically higher levels, P ¼ 0.01, Student t test) upregulation of would be a genetic or epigenetic event(s) that resulted in thrombospondin-1 when compared with the angiogenic the loss of the transformed phenotype. To test this possi- clones (for example, clone 29) that grow exponentially in bility, we assayed the ability of these angiogenic tumor– animals (Fig. 6A and B). This difference was evident in derived clonal cell lines to grow in an anchorage-indepen- both Western blot and ELISA assays. Interestingly, a dent manner. No differences were observed between early- similar thrombospondin-1 regulation pattern (3.5-fold growing and revertant cell lines. difference between the angiogenic and nonangiogenic cell Because we observed no differences in in vitro or in vivo lines; P ¼ 0.03) was observed in the original, patient- proliferation rates or the ability to form colonies in semisolid derived cell lines. In contrast, although VEGF was down- media between the early-growing and revertant clones, we regulated in the revertant clones as expected (6-fold in postulated that there may be a difference in net angiogenic revertant clone 10 vs. angiogenic clone 29, P < 0.0001), it output between the two clone types, as had occurred in the was slightly (20%) upregulated in the original nonan- parental lines. To begin to test this hypothesis, we examined giogenic clone-4 when compared with the angiogenic cells these clones for expression of the angiogenic stimulator fromthesametumor(Fig.6AandC).Thesedataindicate that thrombospondin-1 expression might play an important role in the interconversion of angiogenic and non- Revertant Cl9-2 angiogenic cells in this model. 1,200 Revertant Cl9-10 Angiogenic Cl9-28 The expression changes in VEGF and Tsp-1 between the Angiogenic Cl9-29 ) 1,000 3 Angiogenic Cl9-37 angiogenic cell lines and their revertants suggest that the cause of the revertant phenotype is a decrease in net angio- 800 genic activity. However, expression data alone cannot firmly 600 establish that the change in tumor growth kinetics results from lower angiogenesis-inducing activity in the revertant 400 clones than the early-growing "angiogenic" cell lines. There- Cell number (x 10 (x Cell number 200 fore, to determine whether the clones actually exhibited different angiogenic output, which could account for their 0 differential growth properties in vivo, we performed the 0 2 4 6 8 10 endothelial migration assay, a well-validated surrogate for Time (days) the angiogenic response (21). This allowed us to directly measure the net angiogenic output of the revertant and Figure 5. Cell division rate of angiogenic and nonangiogenic liposarcoma "angiogenic" clones. The ability of two early-growing (9-1 clones. In vitro proliferation of angiogenic (gray) and nonangiogenic and 9-29) and two revertant clones (9-2 and 9-10) to induce (black) clones in cell culture as measured by cell counting. the migration of human microvascular endothelial cells

www.aacrjournals.org Mol Cancer Res; 12(5) May 2014 759

Rogers et al.

AC5 Clone 4 Figure 6. Expression of known 28 29 10 2 *** cells angiogenesis modulators in 4 TSP-1 3 angiogenic and revertant cell lines. A, Western blot detection of VEGF 2 thrombospondin-1 (TSP-1) and VEGF expression in the indicated β-Actin pg/mL/1x10 1

Human VEGF levels angiogenic and revertant clones, Angiogenic Angiogenic with b-actin as loading control. 0 revertants B, ELISA measurement of (nonangiogenic) Original Clone 29 10 thrombospondin-1 production comparing the original angiogenic and nonangiogenic cell line pair versus representative angiogenic Nonangiogenic Angiogenic Angiogenic Revertant (clone 29) and revertant (clone 10) clones, derived in passage two. 50 150 B D ** C, ELISA measurement of VEGF * ** production in the indicated cell 40 * * * lines, as outlined in B. Bars, mean cells

6 100 t 30 values with SEM. The Student test was used to compare various 20 groups. D, HMVEC-d migration to 50 wells containing angiogenic cells ( ± SEM)

ng/mL/1x10 10 (clones 1 and 29) and revertant Human TSP-1 levels Migrated endothelial (clone 2 and 10) cell lines cultured 0 0 in serum-free media. Differences Original Clone Media Conditioned media from clone between each angiogenic line and 29 10 129102 each revertant line were signiﬁcant Angiogenic Angiogenic (P < 0.05 by ANOVA). revertants (nonangiogenic) Nonangiogenic Angiogenic Angiogenic Revertant

(HMVEC) was compared in a modified Boyden chamber complement system, and relaxin signaling (Supplementary assay. We observed that the early-growing "angiogenic" Fig. S3A). Other pathways significantly overrepresented clones had induced 2- to 4-fold more migration than the included PPAR signaling, molecular mechanisms of cancer, revertant clones (P < 0.05 by ANOVA; Fig. 6D). Taken and sphingolipid metabolism. In revertant clones, the top together with the expression data and the previous findings canonical pathways represented were antigen presentation that clone-9 was more angiogenic than clone-4, the data pathway, dendritic cell maturation, autoimmune thyroid indicate that reversion from the early-growing, angiogenic disease signaling, and allograft rejection signaling (Supple- phenotype to dormancy is mediated by a decrease in net mentary Fig. S3B). Other pathways significantly overrepre- angiogenic output. sented included semaphorin signaling in neurons, NF-kB signaling, and death receptor signaling. Differences in in vitro gene expression were observed Gene Ontology analysis using GOEAST software identi- between angiogenic and nonangiogenic clones fied extracellular matrix structural constituent (P ¼ To obtain more comprehensive insight into the molecular 8.38106) as the top molecular function and proteinaceous differences between angiogenic and revertant cell lines, we extracellular matrix (P ¼ 2.461011) as the top cellular performed global gene expression profiling. Two indepen- constituent among genes upregulated in angiogenic clones in dent angiogenic and three independent nonangiogenic cell comparison with revertant clones. Receptor binding (P ¼ clones from the second cycle of cloning (clones 9-2, 9-10 vs. 8.01108) was a significantly overrepresented molecular 9-28, 9-29, 9-37, respectively) were selected and expression function in revertant clones and MHC class II protein of RNA transcripts was measured using Agilent microarrays. complex (P ¼ 8.71107) was one of the most significantly Differences in expression of the top 31 genes were confirmed overrepresented cellular constituents in revertant clones. by real-time RT-PCR (Fig. 7). Clustering of upregulated Immune response was a significantly overrepresented (P ¼ genes from either the angiogenic clone dataset or the rever- 6.3107) biologic process in the revertant liposarcoma tant clone dataset around distinct chromosomal locations clones. Finally, axon guidance was a significantly overrepre- was not observed (see Supplementary Fig. S2). sented biologic process in angiogenic clones (P ¼ 6.89105). Of the genes upregulated in angiogenic clones, the top Eph receptor A4 (1.76-fold) and Eph receptor B3 (1.52- canonical pathways represented as determined by Ingenuity fold) were expressed in angiogenic clones at higher levels Pathway Analysis were acute response phase signaling, than in revertant clones. Eph receptors and ephrins reg- ephrin receptor signaling, axonal guidance signaling, the ulate cell movements during neural crest cell migration,

760 Mol Cancer Res; 12(5) May 2014 Molecular Cancer Research

Reversion of Angiogenesis

CXCR4 MIA Pleiotrophin ADAM23 TNF ADAM20 ROBO1 CD33 MMP9 Figure 7. Real-time RT-PCR TSP-1 analysis of genes identiﬁed as Upregulated in Spondin-2 revertant clones TNFR-1 differentially regulated in ADAMTS9 angiogenic and revertant clones. NF-κB Expression of genes identiﬁed by BRCA-1 CYP26A microarray as differentially TP53 regulated was assessed by real- IL12A time PCR following reverse TRAF1 transcription using Taqman LCP1 HSPB2 probes. Bars, the ratio of gene CD95/Fas expression, nonangiogenic/ VEGF angiogenic (left of vertical axis), or DAXX Upregulated in ID2 angiogenic/nonangiogenic (right). PPARγ angiogenic clones RARRES2 STAT4 SEMA5B PDGFb IL12B

−130 −120 −110 −100 −90 −80 −70 −60 −50 −40 −30 −20 −10 0 10 20 Angiogenic/revertant fold difference in cDNA

gastrulation, neuronal axon guidance, and angiogenesis. 2.23-fold in angiogenic clones and this altered expression was Significantly, a number of downstream effectors of ephrin further confirmed by real-time PCR analysis. Genome-wide receptors, including syndecan 2 (SDC2), actin-related expression analysis of renal cell carcinoma has identified protein 2/3 complex (ARP2/3), and ephexin were also Semaphorin 5B as being upregulated in tumor tissue in elevated in expression in angiogenic clones compared with comparison with the normal surrounding tissue, and a pro- revertant clones. Previous studies have demonstrated that tumorigenic role for this semaphorin has been demonstrated blocking EphA class receptor activation inhibits angio- via promotion of renal carcinoma cell survival (32). genesisinmousetumormodels(22),andthelevelof Finally, ROBO1, which was upregulated by 1.33-fold in expression of EphA2 can be correlated with a more angiogenic clones on the microarray and 1.13-fold by real- aggressive phenotype in prostate cancer (23). Overexpres- time RT-PCR, is also involved in axonal guidance and is a sion of ephrin B3 in a mouse tumor xenograft model has positive regulator of angiogenesis when expressed on endo- been demonstrated to reduce microvessel density and thelial cells (33). PPARg was also confirmed as an upregu- tumor growth (24). In a corneal angiogenesis model, lated gene (1.32-fold) in angiogenic clones by real-time PCR stimulation of Eph B receptors enhanced fibroblast growth analysis, whereas other genes that influence PPARg signaling factor–induced blood vessel formation (25). such as retinoid X receptor a (RXRA; 1.59-fold), nuclear Semaphorin 7A and a number of its effectors, including receptor interacting protein 1 (NRIP1; 2.39-fold), and c-Fos feline sarcoma oncogene and LIM kinase, an inhibitor of (1.70-fold) were upregulated on the microarray. cofilin, were upregulated in revertant clones in comparison Taken together, these findings strongly support the con- with angiogenic clones. Semaphorins can have pro- or anti- clusion that the observed increase in a tumor's net angiogenic tumorigenic effects depending on the semaphorin expressed output is due to genetic mutations. Further, at different and the cell type where it is expressed within the tumor. times in tumor development there are transition states Antiangiogenic effects of semaphorin 3A and semaphorin 3F regulated by heterogeneous populations of tumor cells with have been described (26–29), whereas semaphorin 3C has varying degrees of angiogenic output. On the basis of these been shown to induce endothelial cell growth and migration findings, it is reasonable to assume that if tumors could be (30). The role of semaphorin 7A has not been extensively detected early enough, it may be possible to prevent the studied in carcinogenesis or angiogenesis, but there is some acquisition of the angiogenic phenotype and maintain dor- evidence that expression of its receptor plexin C1 is inversely mant lesions for extended periods of time. correlated with melanoma cell invasiveness (31). Expression of the plexin B3 receptor was also increased 1.6-fold in Discussion revertant clones in comparison with angiogenic clones in the Our studies provide clear evidence that the angiogenic current study. In contrast, semaphorin 5B was upregulated switch is reversible at the cellular level. Although the findings

www.aacrjournals.org Mol Cancer Res; 12(5) May 2014 761

Rogers et al.

in this report are specific to the liposarcoma cell lines used, a secondary effect of the overall decrease in angiogenesis there is reason to believe that other tumor types may exhibit induced by these cells and that the primary determinants of the same plasticity in dormancy and angiogenesis, though the reduced angiogenic stimulus remain to be determined. not necessarily by the same mechanisms (34). Specifically, Pathway and gene ontological analysis of differentially we demonstrate the striking finding that a single-cell clone expressed genes in angiogenic and revertant clones highlight- that forms angiogenic tumors can also give rise to cells that ed a number of possible mechanisms for the differential are incapable of inducing angiogenesis. Importantly, this ability of these cells to stimulate tumor growth beyond a indicates that even in actively growing tumors, a significant particular size. A number of genes first identified for their fraction (perhaps a majority) of tumor cells are nonangio- roles in neuronal axon guidance signaling, including ephrins genic. Crucially, we observed no difference in growth rate and their receptors, semaphorins and plexin receptors were or anchorage-independent growth in angiogenic versus non- differentially regulated in the revertant and angiogenic angiogenic tumors. These findings demonstrate that proan- clones. These signals for axon guidance have since been giogenic events can exist independent of changes that affect demonstrated to have important roles in the regulation of cell growth rate, differentiation, and the transformed state. angiogenesis. In addition, extracellular matrix genes such as Further, the delay in growth for nonangiogenic tumors thrombospondin-2, fibronectin 1, laminin a IV, collagen demonstrates that proangiogenic events are required for type V a 1, and ADAMTS9 were differentially expressed in macroscopic tumor growth, and the kinetics of that delay angiogenic liposarcoma clones in comparison with revertant suggest that two such changes were necessary in this model. clones. Finally, revertant clones had an elevated expression of This notion is bolstered by the observation that a significant class II major histocompatibility molecules in comparison fraction (21%–52%) of clones from angiogenic tumors with angiogenic clones, resulting in an overrepresentation of exhibited an intermediate phenotype in which a fraction of many canonical pathways related to immune responses. tumors exhibited growth indistinguishable from angiogenic Increased expression of these genes in revertant clones may clones and a fraction exhibited growth indistinguishable indicate that a more efficacious presentation of tumor anti- from nonangiogenic clones. One possible explanation for gens to immune effector cells was in operation in this model. such behavior would be that the initiating cell clone had lost Subsequently, a more effective immune response in com- one, but not all of the changes necessary to become angio- parison with the angiogenic clones may have occurred when genic. As this clone was then expanded, additional changes implanted in vivo, resulting in retardation of revertant clone would result in some subpopulations that were angiogenic growth. However, as these experiments were conducted in and some that were nonangiogenic. This phenotype was SCID mice, which lack B and T lymphocytes, natural killer then manifest when these cell populations were injected cells would be the effectors of any immune response in this in vivo. model (44). Although there are inherent limitations to studying These studies raise an important question: If tumor expression changes in cell culture to explain complex in vivo growth initiates when a bare minimum of tumor cells exhibit phenomena, the fact that our analysis identified several genes a proangiogenic phenotype and tumor cells are constantly previously demonstrated to be critically involved in regu- reverting to the nonangiogenic phenotype, why do so few lating angiogenesis indicates that at least a portion of these tumors cease growth and/or spontaneously regress? Our changes also occur in vivo. In this study, we observed several results suggest that there must be some hysteresis in the differences in gene expression in cultured cells that later angiogenic switch. At least two processes may underlie this proved to have reverted to the nonangiogenic phenotype. property of the angiogenic switch. First, there are indications Marked among these was the upregulation of the antiangio- that the threshold for the initiation of angiogenesis is higher genic protein thrombospondin-1. Interestingly, we did not than that for its continuation. Processes such as pericyte observe a difference in VEGF expression that could account dissociation and degradation of basement membrane (with for the differences in growth that we noted. However, one its associated inhibitors) once accomplished need not be caveat to these studies is that VEGF expression that we repeated for vessels to be extended. Second, once a tumor has observed in vitro may not completely reflect VEGF expres- begun to expand, there is more (angiogenic factor produc- sion in the tumor implant. For example, VEGF might be ing) tumor mass available to supply angiogenic factors (such upregulated by the hypoxic environment in the nascent as VEGF, HGF, etc.) to a given region of the tumor tumor or stimulated in the tumor microenvironment. periphery. Thus, for a tumor whose radius has expanded Importantly, however, this effect is likely to be similar for 2-fold, the fraction of angiogenic cells can be reduced by a both the angiogenic and nonangiogenic tumors, and may similar amount without compromising the ability of the even be accentuated by the extended period of hypoxia tumor to induce blood vessels at its periphery. Indeed, once experienced by the nonangiogenic tumors. In our broader established, a tumor can expand even if a fraction of the examination of the transcriptome of angiogenic and non- periphery of the tumor has become net nonangiogenic. In angiogenic cell types, additional differences in gene expres- that case, the tumor will not expand in all directions, but will sion were observed. Many of the most highly upregulated become irregular in shape as only the angiogenic portion of genes in revertant clones have been previously implicated in the tumor expands. tumor progression and/or metastasis (35–43). This might The specific molecular basis for the change of a tumor cell suggest that upregulation of these genes in revertant clones is from the angiogenic to nonangiogenic phenotype and its

762 Mol Cancer Res; 12(5) May 2014 Molecular Cancer Research

Reversion of Angiogenesis

subsequent reversal is currently unclear. There is evidence in further growth and harmless to patients with cancer, signif- the literature that the angiogenic phenotype can be regulated icantly extending their lives. by mutations to specific oncogenes and tumor suppressors (45, 46); however, these mutations are typically associated Disclosure of Potential Conflicts of Interest with substantial changes in cell growth and/or survival that No potential conflicts of interest were disclosed. we do not observe in our clones. There are multiple possible mechanistic explanations for the observed reversion in phe- Authors' Contributions notype. For example, an epigenetic event, such as methyl- Conception and design: M.S. Rogers, A.M. Oyan, M. Lampa, K.-H. Kalland, ation of the tsp-1 promoter, which gave rise to a proangio- R.S. Watnick, J. Folkman, G.N. Naumov genic phenotype could be lost, restoring expression. Also, an Development of methodology: M.S. Rogers, D. Zurakowski, A.M. Oyan, M. Lampa, K.-H. Kalland, O. Straume, R.S. Watnick, J. Folkman increase in copy number of a gene encoding a proangiogenic Acquisition of data (provided animals, acquired and managed patients, provided gene could be lost or an increase in copy number of gene facilities, etc.): K. Novak, E. Lifshits, A.M. Oyan, M. Lampa, S.-Y. Kang, encoding an antiangiogenic gene could be acquired in the K.-H. Kalland, O. Straume, L.A. Akslen, R.S. Watnick, J. Folkman, G.N. Naumov Analysis and interpretation of data (e.g., statistical analysis, biostatistics, compu- revertant cell line. Finally, a mutation could arise down- tational analysis): M.S. Rogers, D. Zurakowski, L. Cryan, A. Blois, T.H. Bø, stream of a gene in a pathway that drives angiogenesis and A.M. Oyan, M. Lampa, K. Naxerova, K.-H. Kalland, O. Straume, L.A. Akslen, R.S. Watnick, J. Folkman, G.N. Naumov results in the overexpression of an angiogenic factor or loss of Writing, review, and/or revision of the manuscript: M.S. Rogers, D. Zurakowski, an antiangiogenic factor (47). The identification of the L. Cryan, A. Blois, A.M. Oyan, M. Lampa, K.-H. Kalland, O. Straume, L.A. Akslen, specific events that drive reversion to the dormant phenotype R.S. Watnick, G.N. Naumov Administrative, technical, or material support (i.e., reporting or organizing data, is an active area of investigation for us, as the precise constructing databases): M.S. Rogers, A.M. Oyan, M. Lampa, L.A. Akslen understanding of this process may have clinical implications. Study supervision: M. Lampa, J. Folkman, G.N. Naumov If angiogenic tumor cells can be induced to exhibit a nonangiogenic phenotype, or if nonangiogenic clones can Grant Support be given a selective advantage, then it may be possible to This study was funded by a Department of Defense Synergistic Idea Award reverse the angiogenic switch in a given tumor. Alternatively, #BC074070 W81XWH-08-1-0710 (to M.S. Rogers), a Department of Defense Breast Cancer Innovator Award #W81XWH-04-1-0316, The Breast Cancer Research if the reversion to the dormant phenotype is governed by a Foundation, and private philanthropic funds (to J. Folkman). genetically driven increase in the expression of a secreted The costs of publication of this article were defrayed in part by the payment of page antiangiogenic protein, such as Tsp-1, then tumors could be charges. This article must therefore be hereby marked advertisement in accordance with reverted to dormancy by stimulating that pathway. The 18 U.S.C. Section 1734 solely to indicate this fact. conversion of an actively growing tumor to the dormant Received October 4, 2013; revised December 31, 2013; accepted January 24, 2014; phenotype would thus render that tumor incapable of published OnlineFirst February 26, 2014.

References 1. Folkman J. Tumor angiogenesis: therapeutic implications. N Engl angiogenic balance accompanies loss of tumor dormancy. Faseb J J Med 1971;285:1182–6. 2002;16:1361–70. 2. Naumov GN, Folkman J, Straume O. Tumor dormancy due to failure of 12. Naumov GN, Akslen LA, Folkman J. Role of angiogenesis in human angiogenesis: role of the microenvironment. Clin Exp Metastasis tumor dormancy: animal models of the angiogenic switch. Cell Cycle 2009;26:51–60. 2006;5:1779–87. 3. Kumar T, Patel N, Talwar A. Spontaneous regression of thoracic 13. Orimo A, Gupta PB, Sgroi DC, Arenzana-Seisdedos F, Delaunay T, malignancies. Respir Med 2010;104:1543–50. Naeem R, et al. Stromal fibroblasts present in invasive human breast 4. Papac RJ. Spontaneous regression of cancer: possible mechanisms. carcinomas promote tumor growth and angiogenesis through elevated In Vivo 1998;12:571–8. SDF-1/CXCL12 secretion. Cell 2005;121:335–48. 5. Weidner N, Folkman J, Pozza F, Bevilacqua P, Allred EN, Moore DH, 14. Almog N, Henke V, Flores L, Hlatky L, Kung AL, Wright RD, et al. et al. Tumor angiogenesis: a new significant and independent prog- Prolonged dormancy of human liposarcoma is associated with nostic indicator in early-stage breast carcinoma. J Natl Cancer Inst impaired tumor angiogenesis. Faseb J 2006;20:947–9. 1992;84:1875–87. 15. Ke XS, Qu Y, Rostad K, Li WC, Lin B, Halvorsen OJ, et al. Genome-wide 6. Weidner N, Semple JP, Welch WR, Folkman J. Tumor angiogenesis profiling of histone h3 lysine 4 and lysine 27 trimethylation reveals an and metastasis–correlation in invasive breast carcinoma. N Engl J Med epigenetic signature in prostate carcinogenesis. PLoS One 2009;4: 1991;324:1–8. e4687. 7. Belien JA, van Diest PJ, Baak JP. Relationships between vasculari- 16. Dysvik B, Jonassen I. J-Express: exploring gene expression data using zation and proliferation in invasive breast cancer. J Pathol 1999; Java. Bioinformatics 2001;17:369–70. 189:309–18. 17. Zheng Q, Wang XJ. GOEAST: a web-based software toolkit for 8. Achilles EG, Fernandez A, Allred EN, Kisker O, Udagawa T, Beecken Gene Ontology enrichment analysis. Nucleic Acids Res 2008;36: WD, et al. Heterogeneity of angiogenic activity in a human liposar- W358–63. coma: a proposed mechanism for "no take" of human tumors in mice. 18. Ashley DJ. The two "hit" and multiple "hit" theories of carcinogenesis. J Natl Cancer Inst 2001;93:1075–81. Br J Cancer 1969;23:313–28. 9. Naumov GN, Folkman J, Straume O, Akslen LA. Tumor-vascular 19. Knudson AG Jr. Mutation and cancer: statistical study of retinoblas- interactions and tumor dormancy. Apmis 2008;116:569–85. toma. Proc Natl Acad Sci U S A 1971;68:820–3. 10. Udagawa T. Tumor dormancy of primary and secondary cancers. 20. Naumov GN, Bender E, Zurakowski D, Kang SY, Sampson D, Flynn Apmis 2008;116:615–28. E, et al. A model of human tumor dormancy: an angiogenic switch 11. Udagawa T, Fernandez A, Achilles EG, Folkman J, D'Amato RJ. from the nonangiogenic phenotype. J Natl Cancer Inst 2006;98: Persistence of microscopic human cancers in mice: alterations in the 316–25.

www.aacrjournals.org Mol Cancer Res; 12(5) May 2014 763

Rogers et al.

21. Staton CA, Reed MW, Brown NJ. A critical analysis of current in vitro 35. Uchida D, Kuribayashi N, Kinouchi M, Ohe G, Tamatani T, Nagai H, and in vivo angiogenesis assays. Int J Exp Pathol 2009;90:195–221. et al. Expression and function of CXCR4 in human salivary gland 22. Brantley DM, Cheng N, Thompson EJ, Lin Q, Brekken RA, Thorpe PE, cancers. Clin Exp Metastasis 2013;30:133–42. et al. Soluble Eph A receptors inhibit tumor angiogenesis and pro- 36. Li KC, Huang YH, Ho CY, Chu CY, Cha ST, Tsai HH, et al. The role of IL- gression in vivo. Oncogene 2002;21:7011–26. 8 in the SDF-1alpha/CXCR4-induced angiogenesis of laryngeal and 23. Zeng G, Hu Z, Kinch MS, Pan CX, Flockhart DA, Kao C, et al. High-level hypopharyngeal squamous cell carcinoma. Oral Oncol 2012;48: expression of EphA2 receptor tyrosine kinase in prostatic intraepithe- 507–15. lial neoplasia. Am J Pathol 2003;163:2271–6. 37. Strasser GA, Kaminker JS, Tessier-Lavigne M. Microarray analysis of 24. Liu W, Jung YD, Ahmad SA, McCarty MF, Stoeltzing O, Reinmuth N, retinal endothelial tip cells identifies CXCR4 as a mediator of tip cell et al. Effects of overexpression of ephrin-B2 on tumour growth in morphology and branching. Blood 2010;115:5102–10. human colorectal cancer. Br J Cancer 2004;90:1620–6. 38. de Nigris F, Crudele V, Giovane A, Casamassimi A, Giordano A, Garban 25. Kojima T, Chang JH, Azar DT. Proangiogenic role of ephrinB1/EphB1 HJ, et al. CXCR4/YY1 inhibition impairs VEGF network and angiogen- in basic fibroblast growth factor-induced corneal angiogenesis. Am esis during malignancy. Proc Natl Acad Sci U S A 2010;107:14484–9. J Pathol 2007;170:764–73. 39. Sasahira T, Kirita T, Kurihara M, Yamamoto K, Bhawal UK, Bosserhoff 26. Miao HQ, Soker S, Feiner L, Alonso JL, Raper JA, Klagsbrun M. AK, et al. MIA-dependent angiogenesis and lymphangiogenesis are Neuropilin-1 mediates collapsin-1/semaphorin III inhibition of endo- closely associated with progression, nodal metastasis and poor prog- thelial cell motility: functional competition of collapsin-1 and vascular nosis in tongue squamous cell carcinoma. Eur J Cancer 2010;46: endothelial growth factor-165. J Cell Biol 1999;146:233–42. 2285–94. 27. Acevedo LM, Barillas S, Weis SM, Gothert JR, Cheresh DA. Sema- 40. Schaller UC, Bosserhoff AK, Neubauer AS, Buettner R, Kampik A, phorin 3A suppresses VEGF-mediated angiogenesis yet acts as a Mueller AJ. Melanoma inhibitory activity: a novel serum marker for vascular permeability factor. Blood 2008;111:2674–80. uveal melanoma. Melanoma Res 2002;12:593–9. 28. Bielenberg DR, Hida Y, Shimizu A, Kaipainen A, Kreuter M, Kim CC, 41. Chang Y, Zuka M, Perez-Pinera P, Astudillo A, Mortimer J, Berenson et al. Semaphorin 3F, a chemorepulsant for endothelial cells, induces a JR, et al. Secretion of pleiotrophin stimulates breast cancer progres- poorly vascularized, encapsulated, nonmetastatic tumor phenotype. sion through remodeling of the tumor microenvironment. Proc Natl J Clin Invest 2004;114:1260–71. Acad Sci U S A 2007;104:10888–93. 29. Guttmann-Raviv N, Shraga-Heled N, Varshavsky A, Guimaraes- 42. Zhang N, Zhong R, Perez-Pinera P, Herradon G, Ezquerra L, Wang ZY, Sternberg C, Kessler O, Neufeld G. Semaphorin-3A and sema- et al. Identification of the angiogenesis signaling domain in pleiotrophin phorin-3F work together to repel endothelial cells and to inhibit defines a mechanism of the angiogenic switch. Biochem Biophys Res their survival by induction of apoptosis. J Biol Chem 2007;282: Commun 2006;343:653–8. 26294–305. 43. Verbisck NV, Costa ET, Costa FF, Cavalher FP, Costa MD, Muras A, 30. Banu N, Teichman J, Dunlap-Brown M, Villegas G, Tufro A. Sema- et al. ADAM23 negatively modulates alpha(v)beta(3) integrin activation phorin 3C regulates endothelial cell function by increasing integrin during metastasis. Cancer Res 2009;69:5546–52. activity. Faseb J 2006;20:2150–2. 44. Shi B, Vinyals A, Alia P, Broceno C, Chen F, Adrover M, et al. Differential 31. Scott GA, McClelland LA, Fricke AF, Fender A. Plexin C1, a receptor expression of MHC class II molecules in highly metastatic breast for semaphorin 7a, inactivates cofilin and is a potential tumor cancer cells is mediated by the regulation of the CIITA transcription suppressor for melanoma progression. J Invest Dermatol 2009;129: Implication of CIITA in tumor and metastasis development. Int J 954–63. Biochem Cell Biol 2006;38:544–62. 32. Hirota E, Yan L, Tsunoda T, Ashida S, Fujime M, Shuin T, et al. 45. Dameron KM, Volpert OV, Tainsky MA, Bouck N. Control of angio- Genome-wide gene expression profiles of clear cell renal cell carci- genesis in fibroblasts by p53 regulation of thrombospondin-1. Science noma: identification of molecular targets for treatment of renal cell 1994;265:1582–4. carcinoma. Int J Oncol 2006;29:799–827. 46. Rak J, Mitsuhashi Y, Bayko L, Filmus J, Shirasawa S, Sasazuki T, et al. 33. Wang B, Xiao Y, Ding BB, Zhang N, Yuan X, Gui L, et al. Induction of Mutant ras oncogenes upregulate VEGF/VPF expression: implications tumor angiogenesis by Slit-Robo signaling and inhibition of cancer for induction and inhibition of tumor angiogenesis. Cancer Res growth by blocking Robo activity. Cancer Cell 2003;4:19–29. 1995;55:4575–80. 34. Straume O, Shimamura T, Lampa MJ, Carretero J, Oyan AM, Jia D, 47. Bayko L, Rak J, Man S, Bicknell R, Ferrara N, Kerbel RS. The dormant in et al. Suppression of heat shock protein 27 induces long-term dor- vivo phenotype of early stage primary human melanoma: termination mancy in human breast cancer. Proc Natl Acad Sci U S A by overexpression of vascular endothelial growth factor. Angiogenesis 2012;109:8699–704. 1998;2:203–17.

764 Mol Cancer Res; 12(5) May 2014 Molecular Cancer Research

Spontaneous Reversion of the Angiogenic Phenotype to a Nonangiogenic and Dormant State in Human Tumors

Michael S. Rogers, Katherine Novak, David Zurakowski, et al.

Mol Cancer Res 2014;12:754-764. Published OnlineFirst February 26, 2014.

Updated version Access the most recent version of this article at: doi:10.1158/1541-7786.MCR-13-0532-T

Supplementary Access the most recent supplemental material at: Material http://mcr.aacrjournals.org/content/suppl/2014/02/26/1541-7786.MCR-13-0532-T.DC1

Cited articles This article cites 47 articles, 11 of which you can access for free at: http://mcr.aacrjournals.org/content/12/5/754.full#ref-list-1

E-mail alerts Sign up to receive free email-alerts related to this article or journal.

Reprints and To order reprints of this article or to subscribe to the journal, contact the AACR Publications Department at Subscriptions [email protected].

Permissions To request permission to re-use all or part of this article, use this link http://mcr.aacrjournals.org/content/12/5/754. Click on "Request Permissions" which will take you to the Copyright Clearance Center's (CCC) Rightslink site.