Background
Tumor cells assign neighboring blood vessels to support their own blood supply for oxygen and nutrients and finally for intravasation (to enter into the blood vessels) and extravasation (metastatic spread) through promoting pathologic neovascularization/angiogenesis [
1‐
3]. This event is potentiated by tumor cells through the production of diffusible angiogenic factors [
4‐
6]. New blood vessel formation/angiogenesis and remodeling of the vessel is a complex event and is dependent on proliferation, differentiation, mobilization and attachment of endothelial cells (ECs) and mural cells (MCs) with different phenotypic variants such vascular smooth muscle cells (VSMCs) and pericytes (PCs) in an autocrine-paracrine manner [
7‐
11]. The literature on the molecular interactions of tumor cells with ECs for the angiogenic switch is appreciable, but less is known about mural cells.
VSMCs/PCs, which are located in different vascular systems according to their needs [
11], play critical roles in both normal and pathologic vascular development, integrity and its maintenance [
11‐
14]. Although VSMCs and PCs are morphologically similar, and express common molecular markers, they may function differently [
11]. The vascular SMCs provide structural support to the large vessels and are critical regulators of blood flow, while PCs appear to be involved in the early events of capillary sprouting. The PCs are regularly found lying at and in front of the advancing tips of endothelial sprouts and may serve as a guiding structure of endothelial outgrowth [
11,
12] and termination of the event [
13]. PCs are irregular in shape in tumors and loosely associated with ECs on tumor vessels [
15,
16], During new blood vessel formation and assembly, recruitment of PCs through the differentiation of precursor cells (mesenchymal), migration and attachment to the newly formed capillaries are vital events of this multistep process [
17,
18]. However, the role(s) of tumor cells in differentiation, recruitments and attachment of these cells are still under described. Therefore, we are interested to explore whether the tumor cells have the ability to differentiate, recruit and interact with PCs to establish new blood vessels for their maintenance.
Accumulated evidences have shown that both endothelial and non-endothelial cells recruit pericytes in tumor blood vessels through PDGF-B, its receptor (PDGF-Rβ) and VEGF signaling networks in a mouse fibrosarcoma model and in U87MG glioma model [
18,
19]. Recently, our studies have found that breast tumor cells are capable of modulating the migration of vascular SMCs
in vitro, and this event is mediated through vascular endothelial growth factor (VEGF)/B-form of platelet-derivative growth factor (PDGF-B) - neuropilin-1 (NRP-1) signaling pathways [
20,
21]. This study, for the first time to our knowledge, shed light on the molecular interactions of tumor cells with mesenchymal stem cells, and offers new opportunities to improve the understanding of the regulation of pathologic pericytes by cell-cell interactions through successive studies. The main objective of the present work is to extend our initial findings and test the hypothesis that the interaction of tumor cells with mural precursor cells may cause differentiation of precursor cells to PCs (mesenchymal to pericyte transition) and the recruitment/attachment for tumor angiogenesis.
To test this concept, we determined whether different tumor cell-derived conditioned media are able to differentiate and recruit the mesenchymal stem cells to pericytes. We demonstrate that the tumor cell-derived conditioned media are capable of differentiating the stem cells to PCs; ultimately recruiting them to bind with endothelial cells differentially. The studies also reveal that tumor cell-secreted PDGF is the responsible molecule for differentiation as well as for recruitment through a physical interaction with NRP-1.
Discussion
Differentiation of mural precursor cells to vascular SMCs/PCs and their recruitment are the fundamental events for the maturation of both normal and tumor blood vessels created from nascent vessels [
3,
20,
29]. One aspect that is clear from our previous studies is that tumor cells can cross-talk with vascular SMCs. The present study further described an important issue of whether tumor cells have any roles in PC biology, most specifically, enhancement of the differentiation of mesenchymal stem cells to pericytes and subsequent recruitment/attachment of PC for tumor angiogenesis through a specific molecular networking circuit.
The initial objective of the present work was to establish the working hypothesis that the interaction between tumor cells and mural precursor cells causes the differentiation of precursor cells to PCs, which ultimately proliferate and recruit to establish a mature and durable vessel. We found that when 10T1/2 cells, C3 H mouse embryonic mesenchymal stem cells (C3H/10T1/2) as characterized by us and others [
22] (Fig.
1), were grown in different breast and pancreatic cancer cell-conditioned media for 2-4 days, the stem cells differentiate into pericytes with abnormally high expression of α-SMA along with Desmin (Fig.
2). These features are identical to the pericytes on tumor vessels [
15,
16]. Thus, we can anticipate that tumor cells have the capability to generate specific signals for mesenchymal to pericyte transitions.
Our next goal was to identified the signaling networks involved in tumor cell-induced MPT. The pericytes can originate from various cell lineages [
14,
30] and commonly come from mesenchymal stem cells [
14,
31]. There are two important molecular signaling pathways that are involved in the development of pericytes [
32], TGF-β and PDGF-B signaling pathways [
14,
33,
34]. Previous studies have shown that embryonic stem cells are able to differentiate into endothelial cells if they are exposed to VEGF, while they can be differentiated into pericytes in presence of PDGF-B [
35‐
37]. In our
in vitro model, we found that tumor cell-secreted PDGF-B plays a critical role in MPT (Figs.3 and 4). Most importantly, we found that the tumor cell-secreted PDGF-B-induced MPT event is mediated through NRP-1 (neuroplin-1), which is a co-receptor for semaphorins with key roles in axon guidance, a docking receptor for VEGF
165[
38,
39] and exhibits a physical interaction with PDGF-B in vascular SMC to enhance their migration [
20,
38].
One of the key steps of the termination of angiogenesis is the incorporation of pericytes/vascular SMCs into the newly formed vessels [
13,
40,
41]. We explored if tumor cells can enhance the recruitment and incorporation of pericytes into the newly formed endothelial tubes under
in vitro and
in vivo setups. The studies showed that in addition to induction of MPT, the tumor cell-generated signals including PDGF-B are able to increase the recruitment and attachment of newly differentiated pericytes into the blood vessels (Figs.
5 and
6). However, the impact of different tumor cell lines that cause pericytes to attach to blood vessels is different. Less aggressive cell lines, such as MCF-7 non-invasive breast tumor cell-generated CM caused more abundant recruitment/attachment of pericytes to the blood vessels as compared to aggressive cell lines (i.e., MDA-MB-231 and MIA-PaCa-2)-generated CM (Fig.
5). Recent studies reveal that pericytes limit metastatic spread in pancreatic β-islet cell tumorigenesis in PDGFRb
ret/ret mice [
42] and in patients with colorectal cancer [
43]. In agreement with these findings, less abundant recruitment of pericytes by aggressive tumor cell-CM may correlate with metastatic spread of these cells. Because MCF-7, MDA-MB-231 and MIA-PaCa-2 cells highly expressed PDGF-B but the CM of later two tumor cell lines were unable to facilitate pericytes to attach more abundantly to the capillary-like structures as compared to MCF-7-CM, these studies indicate that some additional factor(s) is essential for proper adhesion of pericytes to the vessels, which is either missing or inhibited by aggressive cancer cells for metastatic spread. Further studies are warranted. Finally, the studies also show that PDGF-B signaling not only influences MPT and recruitment of pericytes but also affects
in vitro capillary stability by unknown mechanisms (Fig.
5D, last panel).
Conclusions
In conclusion, this work, as depicted in Fig.
6C, provides direct evidence that tumor cells enhance the mesenchymal to pericyte transition event for the recruitment and adhesion of pericytes on newly formed blood vessels to terminate the tumor angiogenesis process. These multistep events are apparently mediated by tumor cell-secreted PDGF-B signaling molecule. NRP-1 may play a critical role in this event. We anticipate a direct link between the disparity in recruitment of pericytes by various tumor cells and their metastatic potency. In support of this study, targeting tumor cell-secreted PDGF-B or NRP-1 in pericytes by inhibitors may efficiently diminish the tumor angiogenesis, tumor growth and metastatic spread.
Materials and methods
Cell culture
The mouse embryonic mesenchymal stem cells, C3H/10T1/2, non-invasive breast tumor cells MCF-7, T-47 D and ZR-75-1, invasive breast cancer cells MDA-MB-231and invasive pancreatic carcinoma cells MIA-PaCa-2, were obtained from American Type Culture Collection (ATCC, Manassas, VA) and grown in Dulbecco's modified Eagle's medium (DMEM, Sigma Chemical Co, St. Louis, MO) supplemented with 10% fetal calf serum and antibiotics, in a humidified incubator at 37°C in an atmosphere containing 5% CO2 and 95% O2. Human aortic smooth muscle cells (AOSMC) and HUVEC (human umbilical vein endothelial cells) were purchased from Cambrex and grown in smooth muscle cells basal media (SmBM) with various growth factors (SmGM-2, i.e., insulin, FGF, EGF and 2% serum) and EGM-2 bullet kit (EBM-2, the basal medium supplemented with growth factors and 5% serum) respectively.
Animals
FVB/N mice (6-8 weeks of age), purchased from Taconic (Hudson, NY) were housed in autoclaved cages fitted with high efficiency filter-tops and with autoclaved bedding. The animals were fed irradiated Purina chow. The room was kept at 25°C with a 12-h light-dark cycle. The animal studies were carried out as per the guidelines established in the Guide for the Care and Use of Laboratory Animals, US Department of Health and Human Resources (NIH 1985) and VA Animal Care facilities.
Reagents
B-forms of PDGF protein and Giemsa were purchased from Sigma Chemical Co. Polyclonal rabbit anti-PDGF-BB, mouse monoclonal alpha smooth muscle-actin and Desmin antibodies were purchased from abCam (Cambridge, MA). Goat-anti mouse E-cadherin and β-Catenin and Goat-anti mouse Cytokeratin-19 and Vimentin were obtained from BD Transduction and Labvision (Neomarker) respectively. A Qtracker cell labeling kit was purchased from Invitrogen (Molecular Probes, Eugene, Oregon). Matrigel was purchased from BD Biosciences.
Preparation of tumor cell-Conditioned Media (TCM)
The procedure of preparation was the same as previously described [
20]. Briefly, MCF-7, MDA-MB-231 and MIA-PaCa-2 cells were grown in DMEM media for 24 hours and collected. They were centrifuged for 10 min at 2000 rpm at 4°C to remove the cells or cell debris. After, centrifugation, media were collected and used for the experiments.
Cell Count
~70% confluent 10T1/2 cells were incubated with different tumor cell-derived condition media for 24 hours. Cells were trypsinized after 24 hours and stained with trypan blue for the count in Cellometer Auto T4 (Nexcelom Bioscience, Lawrence, MA 01843).
Immunofluorescence
Briefly, ~60% 10T1/2 confluent cells that were grown in a one-well slide chamber were incubated with or without PDGF-BB (Platelet-derived growth factor-BB) or different tumor conditioned media (MCF-7, MDA-MB-231 and MIA-PaCa-2) for 24 and 72 hours at 37°C and fixed with methanol. The slides were then permeabilized with Triton-X-100 and incubated with blocking solution (Histostain kit, Zymed Laboratories, CA) for 10 mins at room temperature, and cells were reacted with mouse monoclonal alpha smooth muscle actin antibody overnight at 4°C. After being washed with PBS, the cells were incubated with goat anti-mouse IgG fluorescent conjugated secondary antibody for an hour at room temperature (1:1000 dilutions, Alexa Fluor 594, Molecular Probes). Finally, the washed cells were mounted in PBS-glycerin and examined under a fluorescent microscope.
Western blot analysis
The Western blot analysis was performed in cell lysates treated with different cells derived conditioned medium according to the method described previously [
20]. Cell lysates were obtained by adding lysis buffer containing 50 mM Tris-Cl, pH-8.0, 0.1% SDS, 150 mM NaCl, 1% Nonidet P-40 and protease inhibitor cocktail including 1 μg/ml of Aprotinin, 1 μg/ml leupepsin and 1 mM PMSF and centrifuged at 4°C. The supernatants were collected and protein concentrations were measured with coomassie blue reagent assay (Bio-Rad, Richmond, CA). Equal amounts of protein (10 μg) were subjected to 7.5 - 10% SDS-PAGE and blotted onto a nitrocellulose membrane. Membranes were incubated with specific antibodies anti-mouse monoclonal smooth muscle actin overnight and HRP-conjugated secondary antibodies were incubated for 30 min at room temperature. Signals were detected with Super Signal Ultra Chemiluminescent substrate (Pierce, Rockford, IL) by using ID Image Analysis software Version 3.6 (Eastman Kodak Company, Rochester, NY).
Neutralization assay
To determine the identity of the growth factor in the tumor cell-derived condition media that were mediating the conversion of stem cells into PCs, the media, conditioned by semi-confluent MCF-7, MDA-MB-231 and MIA-PaCa-2 cells, were preincubated overnight at 4°C with different concentrations (i.e., 200 and 500 ng/ml) of a neutralizing polyclonal antibody against PDGF-BB. The semi confluent mesenchymal stem cells, 10T1/2, were then incubated with regular medium (as a control), tumor cell-derived conditioned media, neutralizing media [tumor cell-derived condition media (CM) neutralized with anti-PDGF-BB antibody and rabbit polyclonal IgG (as a negative control)] for 24 hours at 37°C. After that, cell lysates were collected to perform the Western blot analysis for the detection of the expression level of anti-smooth muscle actin.
In vitro angiogenesis and Binding assay
Approximately 80% confluent 10T1/2 cells were labeled with Qtracker cell labeling kit (highly fluorescent Q-dot nanocrystals) obtained from Invitrogen (Molecular Probes). For 3 D cultures, which has been described earlier [
44], Matrigel (150 μl) was polymerized in an 8-well chambered slide. After polymerization, endothelial cell-specific media or tumor cell-generated conditioned media were added. HUVEC and labeled 10T1/2 cells (10,000 cells/well) were seeded into each well and incubated for approximately 20 h which determined the binding efficiency of 10T1/2 cells with capillary-like structure generated by HUVEC in regular media or different tumor cell-derived condition media. Quantification of the number of capillary-like structures and attached Q-dots were carried out using the NIS Elements software program attached with the Nikon photographic fluorescence microscope.
In vivo Angiogenesis Assay
The Gelfoam-implantation angiogenesis assay was performed according to the previous method [
45,
46]. Briefly, three sets of FVB/N mice (6-8 weeks old; four in each set) were anesthetized immediately before implantation of Gelfoam (Gelfoam
®, Pharmacia & Upjohn Company, NY, USA). Gelfoam (8 × 8 mm), presoaked with autoclaved water, regular media or TCM, was transplanted subcutaneously in mice. The transplanted mice were maintained for 5-6 days. In the mean time, Q-dot labeled 10T1/2 cells were cultured in regular media or TCM. After a maintenance period, labeled 10T1/2 cells (2 × 10
5) were injected subcutaneously near implanted Gelfoam according to the experimental set-up. The implanted Gelfoam was taken out carefully after 2 days, and the angiogenesis was detected and quantified using an inverted fluorescence microscope. The protocol has been depicted in Additional file
1: Fig. S1.
Statistical Analysis
All experiments were performed in triplicate for each of the observations. Each of the data represents the mean ± SE from the three separate experiments. Statistical analysis was performed between the two groups of data by an unpaired Student's t-test. A P-value less than 0.05 were considered as statistically significant.
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
KD, SKB and SB initiated the project and designed the study. KD and GD performed all the treatments and the immunofluorescence. KD, MM, IH and SM mainly were involved in western blot analysis. KD has assisted with the in vitro angiogenesis assay. IH with the help of SM performed the in vivo angiogenesis assay. All authors helped in discussing, reading and proofreading the final manuscript.