Host Deficiency in Caveolin-2 Inhibits Lung Carcinoma Tumor Growth by Impairing Tumor Angiogenesis

Caveolins are key components of detergent-resistant cholesterol lipid-rich membranes, including lipid rafts and caveolae. Caveolin-1 (Cav-1) and -2 are ubiquitously expressed and interact with each other, whereas Cav-3 is muscle specific (1). Despite similar name, the amino acid sequence between Cav-1 and Cav-2 is only 38% identical (62% different; ref. 2), suggesting distinct functional roles for each of these proteins (reviewed in ref. 3). However, in contrast with extensively studied Cav-1, much less is known about Cav-2. Nevertheless, most recent studies suggest that Cav-2 could be involved in regulating various processes and functions, in particular in endothelial cells and other cell types (3–12).

To grow beyond approximately 2 mm³, tumors require increased supply of oxygen and nutrients, which is accomplished by angiogenesis, the formation of new blood vessels from preexisting vasculature, for example, from capillaries or venules (13, 14). This transition from the avascular to the angiogenic phase of tumor growth is often referred to as the “angiogenic switch” (13, 14). The angiogenic switch and the subsequent increase in tumor blood vessel density are the most critical mechanisms that allow tumors to overcome growth limitations due to insufficient blood supply. Despite extensive studies focused on tumor growth and tumor-induced angiogenesis (reviewed in refs. 15–17), the cellular and molecular mechanisms involved are far from understanding.

Availability of Cav-1 knockout (Cav-1 KO) mice generated by several independent research laboratories allowed for extensive characterization of the role of Cav-1 in tumor growth and tumor-induced angiogenesis (reviewed in ref. 24; refs. 18–23). However, to the best of our knowledge the role of Cav-2 expressed within the tumor microenvironment in tumor growth and tumor-induced angiogenesis remained unknown. In this study, using newly generated in our laboratory Cav-2 KO mice subjected to subcutaneous implantation with Lewis lung carcinoma (LLC) and B16-F10 melanoma cells, we have examined the role of host-expressed Cav-2 in regulating tumor growth and tumor growth–induced angiogenesis. Remarkably, the results of these studies determined that in contrast with wild-type (WT) mice, LLC tumors are unable to grow whereas B16-F10 tumors display retarded growth in the host microenvironment lacking Cav-2 expression. Further studies determined impaired pathologic angiogenesis in tumors implanted into Cav-2 KO mice.

LLC and B16-F10 cell lines (ATCC) were cultured in DMEM containing 10% FBS, 1% l-glutamine and 100 UI/mL of penicillin plus streptomycin in a humidified chamber at 37°C under 5% CO₂. Both cell lines were regularly authenticated according to the guidelines provided by the ATCC based on morphology (rounded–loosely attached or floating for LLCs and spindle-shaped plus epithelial-like for B16-F10), viability, recovery, and growth, most recently confirmed one month before use.

Six- to 8-week-old C57BL/6N Cav-2 KO and WT littermate mice originating from Charles River Laboratories were used for all experiments according to the protocol approved by the University of Missouri Animal Care and Use Committee.

Cav-2 KO mice were generated in C57BL/6N background with the assistance of Mouse Biology Program (UC Davis, CA) by deletion of entire exon 2 and a 5′ portion of exon 3 (See Supplementary Methods and Supplementary Fig. S1–S2 for details).

To examine the role of host-expressed Cav-2 in tumor growth, LLC or B16-F10 melanoma (10⁶ cells in 100 μL PBS) was injected s.c. in the lower back right and left flanks of 6- to 8-week-old Cav-2 KO and WT littermate mice. When tumors became palpable (typically on day 6 after implantation), LLC tumor growth was monitored every other day by measuring the length and width of the tumor using a caliper. Tumor volume was calculated using the following formula: Volume = 0.52 × (width)² × (length). In addition, at the end of the experiments (typically day 17 for LLC and day 14 for B16-F10), tumors were removed and tumor mass was determined by weighing.

LLC tumors were resected on day 6 (the earliest palpable tumors) or day 10, whereas B16-F10 on day 8 of the experiment and immersed in 10% neutral formalin for overnight fixation. Fixed tissues were processed for paraffin embedment and sectioned for histochemical and immunohistochemical assays. Sections were stained with hematoxylin and eosin (H&E) for evaluation of morphologic changes and with the picrosirius red (PSR) method to detect the presence of extracellular collagen.

Immunohistochemical analysis was performed on paraffin sections using antibodies to Ki-67, cleaved caspase-3, CD31, and CD34. In addition, tumors were subjected to TUNEL assay according to kit directions (In situ Cell Death Detection Kit, POD; Roche Applied Science). Sections (5 μm) were mounted onto ProbeOn Plus microscope slides (Fisher Scientific Inc.). Before immunohistochemical analysis, sections were dewaxed in xylene, rehydrated through graded concentrations of ethanol, and rinsed in distilled water. Antigen retrieval was performed by heating sections in 10 mmol/L citrate buffer (pH 6.0) for 20 minutes. Slides were treated with 3% hydrogen peroxide (to inactivate endogenous peroxidase activity), and rinsed before incubation in blocking buffer of 5% BSA for 20 minutes. Sections were then incubated for 60 minutes at room temperature with each of the following primary antibodies: anti-CD31 antibody [1:50–1:100 dilution of a rabbit anti-CD31 polyclonal antibody (ab28364) Abcam, Inc.]; anti-CD34 antibody [1:100 dilution of a rat anti-CD34 monoclonal antibody (MEC-14.7, ab8158), Abcam]; anti–Ki-67 antibody [1:300 dilution of a rabbit anti–Ki-67 polyclonal (RB1510-P); Thermo Scientific); and anti–cleaved caspase-3 [1:100 dilution of a rabbit anti–cleaved caspase-3 polyclonal antibody (ASP 175, #9661), Cell Signaling Technology]. To detect the antibodies to CD34, sections were incubated for 30 minutes with a biotinylated secondary antibody (rabbit anti-rat IgG, DAKO) and then for 30 minutes with a streptavidin-linked horseradish peroxidase (HRP) product (DAKO). Sections probed with antibodies to CD31, Ki-67, and cleaved caspase-3, were incubated with EnVision, an HRP-labeled polymer conjugated to anti-rabbit antibodies (DAKO). Bound antibodies were visualized following incubation with 3,3′-diaminobenzidine solution (0.05% with 0.015% H₂O₂ in PBS; DAKO) for 3 to 5 minutes. Sections were counterstained with Mayer's hematoxylin, dehydrated, and cover-slipped for microscopic examination. Images were captured using Anxiovert IX70 microscope with 20X objective.

Lung and LLC tumor tissue was extracted from mice and snap-frozen with liquid nitrogen followed by extraction with tissue grinder and lysed in a RIPA lysis buffer containing: 50 mmol/L Tris HCl, 0.1 mmol/L EGTA, 0.1 mmol/L EDTA, 100 mmol/L leupeptin, 1 mmol/L phenylmethylsulfonyl fluoride, 1% (v/v) NP-40, 0.1% SDS, and 0.5% deoxycholic acid; pH 7.4, homogenized, and centrifuged for 10 minutes at 14,000 rpm and at 4°C. Insoluble material was removed and the supernatants were then mixed with Laemmli SDS loading buffer and boiled. Samples were subjected to SDSPAGE and immmunoblotting as described previously (7, 8). Briefly, an equal protein amount was loaded on SDSPAGE, and proteins were electrotransferred onto nitrocellulose membranes. The membranes were washed in Tris-buffered saline with 0.1% Tween, blocked in 5% milk, and incubated with primary antibodies against Cav-2, Hsp-90 (from BD Transduction Laboratories), total endothelial nitric oxide synthase eNOS (from Abcam), phospho-serine 1177–eNOS (P-S1177–eNOS), cleaved caspase-3 (from Cell Signaling Technology), thrombospondin-1 (TSP-1), or PARP-1 (from Santa Cruz Biotechnology) at 4°C overnight followed by incubation with HRP-labeled secondary goat anti-mouse antibody (Bio-Rad), and developed by enhanced chemiluminescence. Where appropriate, the densitometric values for specific immunoblots were determined using ImageJ followed by calculating and expressing densitometric ratios as mean ± SD from two replications based on one representative of two total experiments.

Quantitative real-time PCR (qRT-PCR) was performed similar to described previously (7, 8) with minor modifications (see Supplementary Methods for detailed description).

For the experiments involving time-dependent growth of LLC tumor grafts, the data are expressed as mean ± SEM (n = 6–8). To determine the statistical significance, the two-way ANOVA was used followed by the Bonferroni post hoc test. For all other experiments the Student t test was performed. Differences were considered statistically significant *, P < 0.05.

To examine the role of host-expressed Cav-2 in tumor growth, LLC cells were implanted into WT and Cav-2 KO male mice and tumor growth was determined as described in experimental procedures. As expected, the volume of LLC tumors progressively increased in WT mice reaching approximately 17.1 ± 6.4, 51.4 ± 24.3, 69.6 ± 18.4, 94.8 ± 18.7, 201 ± 38, and 371.5 ± 136.9 mm³ (n = 8) on days 6, 8, 10, 12, 14, and 17, respectively (Fig. 1A, closed squares). Remarkably, in contrast with WT mice, LLC tumors implanted into Cav-2 KO mice did not display a significant growth gradually regressed, in particular between days 10 and 17 (Fig. 1A, open squares). Specifically, the volume of LLC tumors in Cav-2 KO mice reached approximately 15.5 ± 5.4, 30 ± 8.2, 23 ± 12.8, 13.2 ± 11.0, 8.8 ± 6.6, and 2.1 ± 2.1 mm³ on days 6, 8, 10, 12, 14, and 17, respectively (Fig. 1A, open squares). The dramatic reduction in the volume of Cav-2 KO LLC tumors on day 17 was independently confirmed upon surgical removal followed by photographing (Fig. 1B) and weighing. Specifically, the average tumor mass was approximately 285 ± 86 mg (n = 6) and only 0.9 ± 0.3 mg (n = 6) for WT and Cav-2 KO mice, respectively (Fig. 1C), indicating significant (P < 0.0001), nearly a 320-fold reduction of LLC tumor mass in Cav-2 KO compared with WT mice at day 17. To determine whether the defective LLC tumor growth in Cav-2 KO mice was permanent, four additional Cav-2 KO mice were maintained up to 2 months after LLC tumor implantation into two flanks per mouse (n = 8). All eight LLC tumors implanted into Cav-2 mice completely regressed and mice remained tumor free within this prolonged time frame (not shown), suggesting that host-expressed Cav-2 is required for LLC tumor growth. To test whether the observed defect in LLC tumor growth was gender dependent or independent, in addition to male mice (shown in Fig. 1) we have also tested female mice. Similar to Cav-2 KO male mice, Cav-2 KO female mice also displayed defective LLC tumor growth (Supplementary Fig. S3).

To determine whether the role of host-expressed Cav-2 is restricted to the LLC model or is more global, we used B16-F10 melanoma cells as an additional syngeneic model of subcutaneous tumor growth in C57BL6 mice. Our results showed statistically significant (P < 0.05) nearly 3-fold reduction in B16-F10 tumor growth in Cav-2 KO compared with WT mice at final day 14 of the experiment (Supplementary Fig. S4). Taken together, our data suggest that host Cav-2 deficiency negatively affects subcutaneous tumor growth in both LLC and B16-F10 models in vivo.

Because angiogenesis is critical for tumor growth, we hypothesized that the defective tumor growth in Cav-2 KO mice should be associated with reduced microvascular density (MVD) within tumor tissue. To test our hypothesis, LLC tumors were extracted on day 10 of the experiment, that is, before the largest tumors from WT mice reached maximal volume of approximately 100 mm³ and no statistically significant difference in tumor volume could be observed between WT and Cav-2 KO (See Fig. 1A). To determine MVD within LLC tumor tissue, 5-μm paraffin sections were immunohistochemically stained with anti-CD31 and anti-CD34 antibodies. As shown in Fig. 2A and B, Cav-2 KO tumors had considerably less CD31-positive structures (Fig. 2B) compared with WT counterparts (Fig. 2A). Similar to CD31, Cav-2 KO tumors also displayed reduced numbers of CD34-positive structures (Fig. 2F) compared with WT counterparts (Fig. 2E). Quantitative analysis revealed that Cav-2 KO tumors had significantly, approximately 13.2-fold reduced number of CD31-positive vessels (Fig. 2C) and approximately 8.5-fold reduced CD31-positive vessel area (Fig. 2D). Cav-2 KO tumors had also significantly, approximately 3.5-fold reduced number of CD34-positive vessels (Fig. 2G) and approximately 3.6-fold reduced CD34-positive vessel area (Fig. 2H).

To determine MVD in B16-F10 melanoma tumors, staining with anti-CD31 antibody was performed on paraffin sections from smaller tumors extracted at day 8 of the experiment (Supplementary Fig. S5A and S5B). Quantitative analysis revealed significant, approximately 2.5-fold reduction in the number of CD31-positive vessels in B16-F10 tumors implanted into Cav-2 KO compared with WT mice (Supplementary Fig. S5C). Taken together, our data indicate impaired angiogenesis within LLC and B16-F10 tumors implanted into Cav-2 KO compared with WT mice.

To determine whether the defective LLC tumor growth in Cav-2 KO mice is associated with reduced cell proliferation within tumor tissue 5-μm paraffin sections from tumors extracted on day 10 were immunohistochemically stained with anti–Ki-67 antibody. As shown in Fig. 3, there was visibly reduced density of Ki-67–positive cells in LLC tumors implanted in Cav-2 KO (Fig. 3B) compared with WT mice (Fig. 3A). Quantitative analysis revealed significant (P < 0.0001), approximately 3.6-fold reduction of Ki-67–positive nuclei in LLC tumors implanted into Cav-2 KO (71.3 ± 7.41; n = 6) compared with WT mice (254 ± 22.0; n = 7; Fig. 3C). Our data with Ki-67 staining suggest reduced cell proliferation within LLC tumors 10 days after implantation into Cav-2 KO mice.

To determine the degree of fibrosis within tumors extracted on day 10 of the experiment, 5-μm paraffin sections were histochemically stained with PSR known to quantitatively detect extracellular collagen. In contrast with WT tumors in which very little PSR-positive signal could be observed (Fig. 3D), Cav-2 KO tumors displayed considerable staining with PSR (Fig. 3E). As shown in Fig. 3F, quantitative analysis revealed significant (P < 0.0001), approximately 13.3-fold increase of PSR-positive area in LLC tumors implanted into Cav-2 KO (20.8 ± 2.86; n = 8) compared with WT mice (1.56 ± 0.471; n = 8). Taken together, our data with PSR staining suggest increased deposition of extracellular collagen and fibrosis within LLC tumors implanted into Cav-2 KO mice.

To examine whether defective LLC tumor growth (Fig. 1) could also be associated with increased cellular apoptosis, 5-μm sections from LLC tumors extracted on day 10 were stained with anti-cleaved caspase-3 antibody (marker of early apoptosis) as well as subjected to TUNEL staining, which detects late apoptosis (intense staining of DNA fragmentation localized within cell nuclei) as well as nonapoptotic cellular death (intense whole-cell staining; ref. 25). Interestingly, although a limited number of cells was intensely stained with cleaved caspase-3 antibody in WT tumors (Fig. 4A), no clearly cleaved caspase-3–positive cells could be observed within Cav-2 KO tumors (Fig. 4B). Consistent with cleaved caspase-3, a limited number of cells with intense nuclear TUNEL staining could be observed in WT tumors (Fig. 4C). However, in contrast with cleaved caspase-3, nearly all cells within Cav-2 KO tumors displayed intense, whole-cell staining with TUNEL (Fig. 4D), suggesting massive nonapoptotic cell death within LLC tumors 10 days after implantation into Cav-2 KO mice. H&E staining of 5-μm paraffin sections from tumors extracted on day 10 revealed that in contrast with WT, in which hematoxylin-stained nuclei were intact and clearly distinguishable (Fig. 4E), the cellular/nuclear integrity in Cav-2 KO samples was compromised (Fig. 4F), suggesting massive cellular cell death within Cav-2 KO tumors. To shed more light on the mechanism of cell death in Cav-2 KO LLC tumors, we used immunoblotting with anti-cleaved caspase-3 antibody of tumor lysates generated between days 6 and 10 after implantation. Consistent with IHC data, the levels of cleaved caspase-3 were reduced in Cav-2 KO compared with WT LLC tumors in a time-dependent manner (Fig. 4G), suggesting diminished apoptotic signaling and caspase-3–independent death in Cav-2 KO LLC tumors. We then examined the expression levels of 50 kDa necrotic cleavage fragment of PARP-1 (26) and determined time-dependent increase in Cav-2 KO compared with WT LLC tumors (Fig. 4G, central immunoblot; Fig. 4H). Overall, our IHC and immunoblotting data suggest massive necrotic cell death within LLC tumors implanted into Cav-2 KO mice.

To determine whether reduced MVD in LLC tumors implanted into Cav-2 KO is causative rather than a consequence of reduced proliferation and increased death observed on day 10 after tumor implantation (Figs 2, 3A–C and 4), we shifted our focus to the earliest palpable LLC tumors (6 days after implantation). Specifically, 5-μm paraffin sections of the earliest palpable LLC tumors were immunohistochemically stained with antibodies against CD31 and Ki-67 as well as with TUNEL (Fig. 5A–H) or histochemically stained with H&E to assess the general cell morphology. As shown in Fig. 5A–C, the earliest palpable Cav-2 KO tumors already displayed visibly reduced MVD determined on the basis of anti-CD31 staining (Fig. 5B) compared with WT counterparts (Fig. 5A). Quantitative analysis determined that the earliest palpable Cav-2 KO tumors had significantly (P < 0.0001), approximately 2.6-fold reduced number of CD31-positive vessels per field (Fig. 5C). Remarkably, in contrast with reduced MVD, no significant difference in proliferation determined on the basis of Ki-67–positive staining (Fig. 5D–F) nor significant difference in cell death assessed by TUNEL staining (Fig. 5G–I) could be observed between the earliest palpable WT and Cav-2 KO tumors (6 days after implantation). Consistent with similar cell proliferation and survival rate, H&E staining revealed relatively similar cell morphology in the earliest palpable LLC tumors implanted in WT (Fig. 5J) and in Cav-2 KO mice (Fig. 5K). Taken together, our data with reduced MVD but unaltered cell proliferation and death in the earliest palpable Cav-2 KO tumors suggest that host deficiency in Cav-2 results in reduced LLC tumor-induced angiogenesis before decreased LLC tumor cell proliferation/survival. Thus, impaired LLC tumor-induced angiogenesis is the causative mechanism responsible for reduced cell proliferation/survival and consequent defective LLC tumor growth in Cav-2 KO mice.

To determine molecular mechanisms responsible for reduced angiogenesis associated with defective LLC tumor growth, we performed qRT-PCR of mRNA levels for selected known factors involved in angiogenesis in tumors extracted 6 days after implantation. No appreciable differences in mRNA levels were observed for proangiogenic growth factors such as VEGF-A (Supplementary Fig. S6A) or PlGF (Supplementary Fig. S6B), whereas a nonsignificant trend for reduced mRNA levels of FGF-2 and TGFβ transcriptional targets of Notch pathway Hes-1 and Hey-1 in Cav-2 KO LLC tumors was observed (Supplementary Fig. S6C–S6F).

In contrast with the previously mentioned growth factors and transcriptional targets of Notch, there was significant, approximately 1.9-fold increase in mRNA level of antiangiogenic TSP-1 on day 6 (Fig. 6A). Analysis of tumors extracted at later time points, that is, days 8 and 10 determined highly significant approximately 2.4- and 4-fold increase in TSP-1 mRNA in Cav-2 KO tumors (Fig. 6B–C). The time-dependent increase in TSP-1 in Cav-2 KO compared with WT tumors was also observed at protein level using immunoblotting approach (Fig. 6D, top immunoblot; Fig. 6E). Next, we examined serine 1177 phosphorylation of endothelial nitric oxide synthase (P-S1177–eNOS) because it is an important endothelial-specific target for TSP-1, which is involved in tumor angiogenesis and permeability (reviewed in ref. 27). Remarkably, we observed profound time-dependent reduction in the levels of phospho-S1177–eNOS within Cav-2 KO tumors (Fig. 6D, second immunoblot from the top; Fig. 6F). Because VEGFR2 pathway activation in tumor endothelial cells is critical for tumor-induced angiogenesis, we performed immunoblotting with anti–phospho-Y1175–VEGFR2 antibody (Cell Signaling Technology) but were unable to detect a reliable signal within tumor lysates (not shown). However, consistent with reduced angiogenesis, we detected time-dependent reduction in expression levels of total VEGFR2 in Cav-2 KO tumors (Fig. 6D, fourth immunoblot from the top). Taken together, our data suggest that host deficiency in Cav-2 results in increased expression of antiangiogenic TSP-1 and inhibition of phospho-S1177–eNOS.

In this study, using newly generated mice with global loss of Cav-2, we show for the first time that LLC cells subcutaneously implanted into Cav-2 KO mice display defective tumor growth and angiogenesis. In addition, our studies reveal reduced tumor growth and angiogenesis using an independent syngeneic model involving B16-F10 melanoma cells. Moreover, we demonstrate that impaired tumor-induced angiogenesis is responsible for defective LLC tumor growth. Finally, we identify increased levels of antiangiogenic TSP-1 as well as impaired S1177 phosphorylation of eNOS as likely molecular mechanisms associated with defective LLC tumor growth and angiogenesis in Cav-2 KO mice.

The failure of LLC tumors to grow in Cav-2 KO mice is remarkable. In particular, the fact that LLC tumors implanted in Cav-2 KO mice not only grew much slower than their WT counterparts within the earlier phase, that is, up to approximately 10 days after implantation, but that they showed a clear regression pattern within a typical experimental time frame of up to 17 days, and that all Cav-2 KO mice were tumor free when left for a period of up to 2 months after implantation, suggests that host-expressed Cav-2 is absolutely necessary for LLC tumor growth within subcutaneous environment. Interestingly, although growth of B16 was significantly slower in Cav-2 KO mice, the phenotype was less robust, suggesting a greater ability of host-expressed Cav-2 to promote growth of LLC than B16 melanoma skin grafts.

The data showing tumor-promoting role of host-expressed Cav-2 is in agreement with several clinical observations involving various human cancers. For instance, Cav-2 expression was upregulated in esophageal (28) and urothelial carcinoma of the urinary bladder (29, 30). Moreover, Cav-2 expression was closely associated with basal-like immunophenotype and proved to be a prognostic factor of breast cancer (31, 32). The expression levels of Cav-2 were also upregulated during prostate cancer progression (33). Another study reported a strong correlation between Cav-2 negativity within tumors and a 5-year survival of patients with stage I lung adenocarcinoma and multivariate analyses taking into account the age and asbestos fiber content revealed that Cav-2 positivity might be an independent unfavorable prognostic factor (34). In addition to the previously discussed published clinical studies, several datasets available through searchable database cBioPortal (35, 36) indicate Cav-2 gene amplification in other tumor types such as glioblastoma, ovarian serous cystadenocarcinoma, stomach adenocarcinoma, or melanoma. Thus, in light of the previously discussed clinical data and our new data with defective or retarded subcutaneous tumor growth in Cav-2 KO mice using LLC and B16-F10 syngeneic tumor models, future mechanistic studies examining the role of Cav-2 expressed in tumor cells versus tumor microenvironment in regulating tumor progression are clearly warranted.

The data showing markedly reduced MVD within tumors implanted in Cav-2 KO mice and when WT tumors were still very small, strongly suggest that Cav-2 expressed in tumor microenvironment promotes pathologic angiogenesis induced in LLC and B16-F10 models. In addition, an obviously reduced cell proliferation and enhanced cell death as well as fibrosis within LLC tumors observed 10 days after implantation in Cav-2 KO mice are consistent with inability of LLC skin grafts to grow in Cav-2–deficient microenvironment. However, given that LLC tumors extracted from Cav-2 KO mice 10 days after implantation, in addition to robustly diminished MVD appear to display a massive cell death and fibrosis, the question remained to whether a defective angiogenesis is the primary mechanism and a cause, simply coincides, or is a consequence of reduced proliferation and enhanced cell death as well as fibrosis within LLC tumors. To answer the latter question, we performed additional experiments involving the earliest palpable tumors extracted from WT and Cav-2 KO mice only 6 days after LLC implantation. These experiments determined significantly reduced MVD but comparable cell proliferation and survival within the earliest palpable LLC tumors implanted into Cav-2 KO mice, indicating that defective angiogenesis occurs before reduced cell proliferation and survival. Thus, the ability to promote pathologic angiogenesis and the consequent enhancement of tumor cell proliferation and survival are the critical mechanisms via which host-expressed Cav-2 drives tumor growth. Overall, our data with defective angiogenesis being responsible for the inability of LLC tumors to grow within Cav-2 KO microenvironment is consistent with the notion that angiogenesis plays an important role in the tumor growth and its blood supply (13, 14, 16, 17).

What are the molecular mechanisms responsible for impaired tumor-induced angiogenesis and inhibition of tumor growth in Cav-2 KO mice? Because TSP-1 is the major antiangiogenic factor most extensively characterized in the tumor microenvironment (reviewed in ref. 37), it is plausible that time-dependent upregulation of TSP-1 at mRNA and protein levels observed within LLC tumors implanted into Cav-2 KO mice is an important molecular mechanism via which host deficiency in Cav-2 inhibits tumor angiogenesis and growth. Although TSP-1 was reported to inhibit angiogenesis via multiple mechanisms, inhibition of endothelial cell–specific VEGR2- and eNOS-dependent pathway(s) appear to be prominent (27, 37). The endothelial cell–expressed VEGFR2-dependent pathway is particularly important in tumor angiogenesis (16, 17, 38). Notably, TSP-1 was shown to regulate VEGFR2 phosphorylation at Y1175 and overall VEGFR2 signaling in endothelial cells (39, 40). Thus, using IHC and immunoblotting approaches, we attempted to compare phospoho-Y1175–VEGFR2 levels but were unable to detect specific signal in tumor samples (not shown). Hence, to further examine activation of the VEGR2-dependent pathway within LLC tumors, we next focused on phospho-S1177–eNOS, which is an important endothelial-specific downstream target of phospoho-Y1175–VEGFR2 and is key for VEGFR2-dependent angiogenesis and permeability (41–44). Remarkably, we observed robust suppression of phospho-S1177–eNOS, which closely correlated with upregulation of TSP-1 within LLC tumors implanted into Cav-2 KO mice, suggesting impaired S1177–eNOS phosphorylation as a likely consequence of inhibition of the VEGFR2-dependent pathway by TSP-1. Future mechanistic studies examining how host Cav-2 deficiency upregulates TSP-1within tumor microenvironment are clearly warranted.

Although global Cav-2 KO mice are ideal model to determine the role of host-expressed Cav-2 in tumor growth and tumor growth–induced pathologic angiogenesis, future studies involving various cell type/tissue–specific rescue of Cav-2 in global Cav-2 KO background will be helpful to unequivocally determine cell-type(s)–specific and more detailed molecular mechanisms involved in tumor angiogenesis and tumor growth-promoting role of Cav-2. In addition, it will also be interesting to determine the significance of Cav-2 in additional mouse models of cancer, in particular in orthotopic models of tumor growth and metastasis.

Antiangiogenic therapy is a very important strategy in cancer treatment; however, due to frequently developed resistance, for instance, to anti-VEGF therapy, identifying additional targets for antiangiogenic therapy is crucial (15). Therefore, our new data showing that Cav-2 expressed in host microenvironment promotes tumor-induced angiogenesis suggest that Cav-2 could possibly be an important target for antiangiogenic therapy and cancer treatment and thus require further investigation. Moreover, except hyperproliferation involving Flk-1 (VEGFR2)-positive cells in the lung, Cav-2 KO mice generated by Razani and colleagues (45) and our group are healthy, reproduce normally, and do not display major defects in physiologic angiogenesis Thus, targeting Cav-2 could possibly selectively disrupt tumor microvessels without affecting normal vasculature.

No potential conflicts of interest were disclosed.

Conception and design: Y. Liu, G. Sowa

Development of methodology: Y. Liu, L. Xie, G. Sowa

Acquisition of data (provided animals, acquired and managed patients, provided facilities, etc.): Y. Liu, L. Xie, G. Sowa

Analysis and interpretation of data (e.g., statistical analysis, biostatistics, computational analysis): Y. Liu, L. Xie, G. Sowa

Writing, review, and/or revision of the manuscript: Y. Liu, G. Sowa

Administrative, technical, or material support (i.e., reporting or organizing data, constructing databases): Y. Liu

Study supervision: G. Sowa

Other (assisted in the experiments involving LLC and B16-F10 tumor cell implantation): S. Jang

This work was supported by the grant from the NIH (1R01HL081860; G. Sowa).

The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked advertisement in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.