HOME HELP FEEDBACK SUBSCRIPTIONS ARCHIVE SEARCH TABLE OF CONTENTS		QUICK SEARCH:   [advanced] Author: Keyword(s): Year:  Vol:  Page:

This Article Abstract Full Text (PDF) Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Citing Articles Citing Articles via HighWire Citing Articles via ISI Web of Science (10) Citing Articles via Google Scholar Google Scholar Articles by Kunz-Schughart, L. A. Articles by Wheatley, D. N. Search for Related Content PubMed PubMed Citation Articles by Kunz-Schughart, L. A. Articles by Wheatley, D. N.

VASCULAR BIOLOGY

Potential of fibroblasts to regulate the formation of three-dimensional vessel-like structures from endothelial cells in vitro

¹Institute of Pathology and ²Clinic of Cardiothoracic Surgery, University of Regensburg, Regensburg, Germany; and ³BioMedES, Aberdeen, United Kingdom

Submitted 25 May 2005 ; accepted in final form 30 November 2005

	ABSTRACT

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
The development of vessel-like structures in vitro to mimic as well as to realize the possibility of tissue-engineered small vascular networks presents a major challenge to cell biologists and biotechnologists. We aimed to establish a three-dimensional (3-D) culture system with an endothelial network that does not require artificial substrates or ECM compounds. By using human skin fibroblasts and endothelial cells (ECs) from the human umbilical vein (HUVECs) in diverse spheroid coculture strategies, we verified that fibroblast support and modulate EC migration, viability, and network formation in a 3-D tissue-like stromal environment. In mixed spheroid cultures consisting of human ECs and fibroblasts, a complex 3-D network with EC tubular structures, lumen formation, pinocytotic activity, and tight junction complexes has been identified on the basis of immunohistochemical and transmission electron microscopic imaging. Tubular networks with extensions up to 400 µm were achieved. When EC suspensions were used, EC migration and network formation were critically affected by the status of the fibroblast. However, the absence of EC migration into the center of 14-day, but not 3-day, precultured fibroblast spheroids could not be attributed to loss of F viability. In parallel, it was also confirmed that migrated ECs that entered cluster-like formations became apoptotic, whereas the majority of those forming vessel-like structures remained viable for >8 days. Our protocols allow us to study the nature of tubule formation in a manner more closely related to the in vivo situation as well as to understand the basis for the integration of capillary networks in tissue grafts and develop methods of quantifying the amount of angiogenesis in spheroids using fibroblast and other cells isolated from the same patient, along with ECs.

endothelium; angiogenesis; human umbilical vein endothelial cell; multicellular spheroid; coculture; tubular structures

With respect to the remodeling of vascular structures, much has already been learned from currently available model systems that mimic vessel formation under defined in vitro conditions and have been established to identify and assay molecules affecting angiogenesis. The most widely applied assay systems are combinations of Matrigel, collagen gels, or mixed ECM gels with ECs, which grow and differentiate to form loose network-like structures of cords in the presence of VEGF. However, gel-type EC culture systems have been developed on the basis of the assumption that the ECs form vessels in an appropriate matrix, grossly underestimating the intimate association and/or involvement of living fibroblasts. Because vessel sprouting and maturation are intricately tuned by diverse positive and negative regulators that are also expressed by adjacent cell types, coculture systems are clearly much more relevant, and although they are inevitably more complex, they nevertheless provide a better approximation of a three-dimensional (3-D) capillary-like network that would be found in vivo (46). This hypothesis is supported by the results of several in vitro studies, primarily in the field of cancer research. Janvier et al. (26), for example, showed that fibroblasts enhanced the formation of a capillary-like network in a sandwiched collagen gel culture system containing fibroblasts and PC-3 human prostate adenocarcinoma cells. In addition, Walter-Yohrling et al. (54) described that cocultures of an ovarian carcinoma cell line with myofibroblasts facilitated the invasion of ECs in a Matrigel-based 3-D culture system. The supporting effect of fibroblasts was also noted, however, in tumor cell-free applications. In 1994, Villaschi and Nicosia (53) documented microvessel stabilization and EC life span prolongation by fibroblasts in a collagen gel system and in the rat aorta assay. This was confirmed by Martin et al. (40), who showed that fibroblast in a bioengineered but otherwise naturally occurring ECM enhanced angiogenesis and EC motility, both in the rat aortic ring and in a Matrigel tube formation assay.

Considering that many normal physiological processes that rely on angiogenesis also require activation and migration of fibroblasts, cocultivation of fibroblast and ECs is a logical approach to enhance the quality of vessel-like structures formed in vitro. Indeed, one of the most authentic in vitro angiogenesis test systems was presented 5 years ago and showed extraordinary in vitro tubule formation in mixed cultures of human diploid fibroblast and HUVECs in a monolayer coculture (8, 14). Most notably, no artificial or additional matrix proteins were required, and vessel-like structures with true luminal development occurred during a 14-day period in multilayered areas formed by the two cell types. This occurred despite the fact that no extant vessels were present and that there was no involvement of blood or fluid producing backup pressure as would be found during angiogenesis in vivo. The crucial role of fibroblasts was recognized, and the conditions allowed experimental analysis of the cellular interactions with comparative ease. We set out to evaluate the power of human skin fibroblast in vessel formation in a true 3-D culture system that extends three cell layers but also avoids artificial matrices. By choosing the spheroid model system with cell cycle-arrested fibroblasts (36, 37) that are cocultured with HUVECs, we have established an easy-to-handle culture technique that allows one not only to investigate the impact of fibroblasts and other cells on vessel formation in a tissue-like environment but also to manipulate the local occurrence of tubule-like structures. Our system does not require any cell attachment to noncellular surfaces and thus could easily be applied in transplantation approaches.

	MATERIALS AND METHODS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Cell types and routine culturing. For most coculture experiments, two cell types were applied, human umbilical vein ECs (HUVECs) (23) and normal diploid human fibroblasts originating from foreskin material after routine circumcision. Primary fibroblast culturing and characterization were performed as detailed earlier (36, 39). Both EC and fibroblast stock cultures were stored in liquid N₂ in supplemented media or FCS, respectively, containing 10% DMSO and subsequently recultured as monolayers in complete endothelial growth medium (EGM)-2 medium (EC; Cambrex BioScience, Verviers, Belgium) or DMEM containing 0.2% phenol red (Biochrom, Berlin, Germany), 10% FCS, and 100 IU/ml penicillin + 100 µg/ml streptomycin (fibroblast; all from PAN-Biotech, Aidenbach, Germany). Cultures were incubated in a humidified atmosphere with 5% CO₂ in air at 37°C. Cell transfer and preparation of single-cell suspensions were performed using mild enzymatic dissociation with a 0.05% trypsin-0.02% EDTA solution in PBS (PAN-Biotech). Cell counts and cell volumes were routinely recorded using a CASY1 cell analyzer (Schärfe System, Reutlingen, Germany) for culture quality assessment. Cells used for cocultivation had cumulative population doubling of <25 (passages 4 and 5) and were free of mycoplasmal contamination as verified using the MycoAlert mycoplasma detection kit (Cambrex Bioscience).

Monolayer cocultures. To preevaluate the capacity of EC and F to form vessel-like structures in coculture, dissociated single-cell suspensions were seeded as mixtures in 24-well plates according to previous investigations and according to the protocol provided with a commercially available culture system for angiogenesis (TCS Cell Works, Botolph Claydon, UK) (8, 14). EGM-2 medium with and without VEGF supplement (1–2 ng/ml according to the manufacturer's instructions; Cambrex BioScience) was applied and refreshed every other day. The formation of tubule-like structures in these cocultures was evaluated microscopically using EC-specific immunohistochemical staining after 8, 15, and 21 days in culture.

Spheroid culture. Spheroid mono- and cocultures were grown in liquid overlay using agarose-coated 96-well plates (50 µl of 1.5% agarose in DMEM/well; Sigma-Aldrich) and EGM-2 medium with or without VEGF supplement (10, 38). For long-term cultures up to 22 days, medium was renewed every 48 h to provide sufficient nutrient supply after an initial steady-state phase of 3–4 days. Spheroid volumes were routinely determined using a calibrated image-processing system consisting of an inverted microscope (Axiovert 200) equipped with a digital camera (Axiocam MRc) and software (KS300) (all from Zeiss, Göttingen, Germany) with a protocol for reproducible morphometric analysis of spheroid data.

Different spheroid coculture strategies were used to demonstrate the impact of fibroblasts on the formation of a tubule-like network. For strategy 1, fibroblasts and EC were mixed in different concentrations (40:1, 4:1, or 2:1) and were seeded onto agarose-coated plates with a constant number of 1 x 10⁴ fibroblast/well. Fibroblast-EC mosaic spheroids were fixed and processed for immunohistochemical staining after defined times in coculture up to day 21. In the other approach (strategy 2), fibroblast spheroid monocultures were prepared with an initial seeding concentration of 1 x 10⁴ fibroblast/well. EC suspensions (1 x 10⁴) were added at days 3 and 14, respectively, after fibroblast inoculation. EC migration and tubule formation in these spheroid cocultures were monitored 2 and 8 days after addition of ECs.

Immunohistochemistry. Monolayer cultures were stained immunohistochemically according to an established protocol (8, 14). ECs were labeled using antibodies against human CD31 (platelet endothelial cell adhesion molecule 1) or CD146 (melanoma cell adhesion molecule) after fixation of the cultures with ice-cold 70% ethanol containing 15 mM glycine and blocked for nonspecific binding by incubation with 1% BSA in PBS. Anti-human CD31 antibodies were also applied to visualize ECs in spheroid cocultures using either immunohistochemical or immunofluorescence approaches. For both immunostaining techniques, spheroids and spheroid cocultures were fixed in 4% formalin, embedded in paraffin using a Hypercenter XP (Shandon, Frankfurt, Germany), and serially sectioned (5–6 µm). For immunohistochemistry, sections were processed according to routine protocols using the NexEs/HC module (Ventana Medical Systems, Tucson, AZ) to visualize the distribution of ECs and of the ECM compounds collagen types I and IV. Hematoxylin was used as a counterstain. All antibodies and secondary detection systems used for immunostaining are specified in Table 1. All unspecified chemicals were purchased from Sigma-Aldrich or Merck.

Because proteinase K treatment interfered with the labeling of CD31 antigen using commercially available anti-CD31 antibodies, we visualized ECs in parallel sections of cocultures using an established immunofluorescence protocol. Primary and secondary antibodies and incubation details are provided in Table 1. Chemicals from unspecified sources were purchased from Sigma-Aldrich or Merck.

Flow cytometry. Spheroid mono- and cocultures were also dissociated to determine the number of cells per spheroid using the CASY1 cell analyzer and to perform advanced flow cytometric analysis. In general, a defined number of spheroids (30–48) was collected for each measurement, and single-cell suspensions were prepared by applying enzymatic treatment for 5 min at 37°C using 0.05% trypsin-0.02% EDTA solution in PBS (PAN-Biotech) supplemented with 10 µg/ml collagenase type IA (Sigma-Aldrich). For analysis of F-to-EC distribution, cell suspensions were stained with FITC-conjugated anti-human CD31 (clone MEM-05; Immunotools, Friesoythe, Germany). Propidium iodide (PI, 2 µg/ml; Sigma-Aldrich) was applied routinely to exclude membrane-damaged cells. In some experimental series, a commercial allophycocyanin-conjugated AnnexinV kit (ApoScreen AnnexinV; Southern Biotech, Birmingham, AL) combined with a specific AnnexinV binding buffer (BD Biosciences, Heidelberg, Germany) was used according to the manufacturer's instructions to estimate the proportion of cells that underwent apoptosis. All experiments were performed at least in triplicate. Statistical significance was proved using an unpaired Student's t-test.

ELISA. VEGF concentration in supernatants of spheroid mono- and cocultures was measured after defined times using a commercially available ELISA kit (R&D Systems, Wiesbaden, Germany). Conditioned media of spheroid cultures grown in EGM-2 medium without VEGF supplement were collected at days 3, 5, 13, and 15 in culture. The 2-day time intervals were chosen to allow for calculation of 48-h cellular release rates, taking into account the respective cell numbers per spheroid at days 5 and 15, respectively. According to the feeding strategy with only 50% of the medium replaced at day 3 in culture and every 48 h thereafter, VEGF levels at day 3 or 13 had to be subtracted from the 5- and 15-day measurements. Unconditioned, fresh EGM-2 media with and without VEGF supplementation served as controls.

3-D reconstruction. To demonstrate the EC network in a representative mixed fibroblast-EC spheroid coculture, an entire series of 5- to 6-µm sections of one spheroid stained immunohistochemically for CD31 was imaged (45 individual sections). The resulting tagged image file format, or TIFF, files were processed and reassembled for 3-D reconstruction using Amira 3.1 software (TGS; Mercury Computer Systems, San Diego, CA). The procedure included image alignment, filtering, and conversion of color information using different algorithms (median and Gaussian) to extract CD31-stained areas (diaminobenzidine-positive, brown), as well as the production of gray-scale binary images. We then generated isosurfaces for 10–12 adjacent or for all sections using a voxel of 5 to dilate data sets into Z-direction and produced overlays.

Transmission electron microscopy. Selected spheroid cocultures were fixed in 0.1 M cacodylate-buffered Karnovsky solution (2.5% glutaraldehyde and 1% paraformaldehyde overnight at room temperature) and postfixed in 1% osmium tetroxide (2 h), pH 7.3, dehydrated in graded ethanol, and embedded in EmBed-812 epoxy resin (all reagents were obtained from Science Services, Munich, Germany; automated LYNX tissue processor, Leica/Germany). After 48-h heat polymerization at 60°C, semithin (0.8 µm) sections were cut, stained with toluidine blue/fuchsin, and after light microscopic spheroid selection, the epon block was trimmed for ultrathin sectioning. Ultrathin (80 nm) sections were prepared with a diamond knife on a Reichert Ultracut-S ultramicrotome and double contrasted with aqueous 2% uranyl acetate and lead citrate solutions for 10 min each. The sections were examined using a LEO912AB electron microscope operating at 80 kV.

	RESULTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
EC network in 2-D fibroblast endothelial mixed cultures. Primary EC cultures differ individually in their capacity to form an EC network in in vitro angiogenesis systems. The capacity of HUVECs to be applied in 3-D coculture was therefore preassessed in a 2-D coculture angiogenesis system in parallel, which allowed us to assess selected human foreskin fibroblast cultures as a supportive stromal cell type. Figure 1 documents the development of a planar EC network in this two-cell-type system as a function of incubation time. Between days 15 and 21, clear EC sprouting activity was observed. According to previous studies in which this coculture model was used, a few EC sprouts were obtained even in medium without added VEGF, but sprouting became more pronounced with increasing supplementation up to 10 ng/ml VEGF. EC sprouting and network formation was also affected by the addition of other physiologically relevant paracrine modulators, including the pluripotent peptide growth factors TGF-₁ and PDGF (data not shown).

View larger version (84K): [in this window] [in a new window]   Fig. 1. Human skin fibroblasts (F) and human umbilical vein endothelial cells (HUVECs) in monolayer coculture form an endothelial cell (EC) network and show sprouting. Mixed cultures of F and EC were stained immunohistochemically with a mouse anti-human CD146 MAb to visualize the developing EC network as a function of time in monolayer cultures grown in supplemented endothelial growth medium (EGM)-2 with VEGF and bFGF. APAAP staining technique, Fast Red as chromogen. d7, day 7; d15, day 15; d21, day 21.

View larger version (53K): [in this window] [in a new window]   Fig. 2. Human skin F and HUVECs form a complex EC network in mixed three-dimensional (3-D) spheroid cocultures and show morphological differentiation phenomena. A: representative images of 5- to 6-µm median sections of paraffin-embedded spheroid cocultures established from F-EC mixtures seeded onto agarose-coated 96-well plates with F-to-EC ratios of 40 to 1 and 4 to 1, respectively. Spheroids were monitored after 3, 8, and 14 days in culture. EGM-2 supplemented with 1 ng/ml VEGF was renewed every 48 h, as detailed in MATERIALS AND METHODS. ECs were labeled with a mouse anti-human CD31 MAb using a streptavidin-peroxidase technique, diaminobenzidine for color development, and hematoxylin counterstaining. B: F + EC mixed spheroid cultures seeded at ratios of 40 to 1 and 4 to 1 were also dissociated, and cells were reanalyzed using an automated cell analyzer and flow cytometry to determine the number of membrane-intact F and ECs at days 3 and 15 in culture. Regions of interest to estimate the proportion of F to EC were gated in the forward vs. 90° scatter signal dot plot, the propidium iodide (PI) histograms (PI negative = membrane intact), and in the dot plot of CD31-specific staining vs. FL-2 autofluorescence. Cell counts per spheroid (means ± SD) of 3 independent experiments are shown at right. Number of ECs in mixed cultures differed by a factor of only 2 (but not 10) when the 40-to-1 and 4-to-1 experimental settings were compared.

View larger version (20K): [in this window] [in a new window]   Fig. 3. Addition of VEGF does not critically affect the formation of the EC network in mixed spheroid cultures. A: representative images of 5- to 6-µm median sections of paraffin-embedded spheroid cocultures established from F-to-EC mixtures (1 to 4) cultured over a period of 3, 14, or 21 days, respectively, in medium supplemented with or without VEGF (1 ng/ml). EC labeling was performed as described in Fig. 2A. B: analysis of cell suspensions derived from dissociated F + EC cocultures grown in medium supplemented with or without VEGF. Flow cytometry was performed as described in Fig. 2B. Mean cell counts per spheroid coculture (±SD) of 3 independent experiments are shown. Number of cells (F + ECs) decreased only at later growth phases (day 15). Number of membrane-intact ECs did not differ between the two culture conditions. C: VEGF content in supernatants of F monocultures and F + EC cocultures was determined by ELISA and is shown with respect to VEGF-supplemented and -nonsupplemented medium controls. According to these data, which show a rapid drop in VEGF content as a function of time, F released VEGF only at the onset of spheroid culturing. VEGF level is significantly lower in mixed spheroids compared with F monocultures up to day 13 in culture.

Impact of exogenous VEGF on EC network in 3-D mixed cultures. Superfluous VEGF medium supplementation was not required for the formation of the EC network in the fibroblast endothelial mixed spheroid culture system. This was indicated in immunohistochemically stained sections of spheroid cocultures during a period of 21 days (Fig. 3A) and confirmed using flow cytometric analysis (Fig. 3B). We therefore hypothesized that routine addition of 2% FCS plus fibroblast-derived paracrine factors provide a sufficient level of VEGF potentially required for EC network formation. Accordingly, VEGF concentrations were determined in supernatants of fibroblast mono- and fibroblast + EC cocultured spheroids at defined time intervals and cellular VEGF release rates (free VEGF/cell) were calculated on the basis of cell number per spheroid. The data in Fig. 3C show that the VEGF level in the supernatant of fibroblast spheroid cultures at day 3 was one-tenth that in VEGF-supplemented medium (100 pg/ml vs. 1,000 pg/ml). Unsupplemented medium controls contained only 10 pg/ml. Thus cellular VEGF release was 2.0 ± 0.6 fg/cell for the 72-h initiation interval. VEGF levels then continuously decreased throughout further cultivation as cellular release rates rapidly dropped to zero. VEGF levels in the fibroblast + EC cocultures already were significantly lower at day 3 in culture with a free VEGF concentration of 0.7 ± 0.1 fg/cell, indicating either reduced cellular release rates during the 72-h aggregation phase because of reciprocal interactions between fibroblast and EC or some VEGF consumption required for EC survival and network formation at early culture phases.

Similarly, we examined mixed spheroids cultured in EGM-2 medium without external addition of both VEGF and bFGF using flow cytometric approaches. Under these conditions, the total number of membrane intact cells was reduced by 10%. However, this was primarily due to a reduction of fibroblasts in these mixed cultures, at least during early culture phases (day 3). At days 8 and 15, the number of ECs was slightly reduced compared with bFGF-supplemented controls (data not shown). However, the difference in the number of ECs was not as pronounced as expected in a bFGF-deficient environment. As for VEGF, this phenomenon is under further investigation, including bFGF release in fibroblast spheroids.

Tubule-like structures in fibroblast endothelial mixed cultures. Serial sections through entire mixed spheroids seeded at a 4-to-1 ratio were stained for CD31 to confirm that the EC arrangement was not planar but was a truly 3-D interconnected cellular network (Fig. 4A). Reconstruction of the filtered immunohistochemical color information revealed a complex, capillary-like 3-D structure (Fig. 4B). To further show whether tubular structures with lumina formed by ECs were present in the in vitro system reflecting an in vivo-like capillary EC network, 14-day mixed spheroids were examined using transmission electron microscopy. Figure 5 highlights the tubular arrangements of ECs with lumen formation (cf. structures described in Ref. 8). Tubule walls were no thicker than a single EC layer. Points of tightly apposed cell membranes were frequent. EC tight junctions and junctional complexes (Fig. 5, A and B), as well as luminal pinocytotic activity, were clearly visible (Fig. 5B). In some luminal structures, fingerlike EC protrusions characteristic of EC overlaps and adherence could be distinguished. Cell debris and degradation material were found frequently in the lumina (Fig. 5C). The EC tubular structures were covered by a segmental basement membrane confining the ECs from surrounding fibroblasts (Fig. 5A). Rod-shaped microtubulated bodies (Weibel-Palade bodies), specific organelles present in ECs of arteries, veins, and capillaries in different quantities that contain von Willebrand factor, could not be identified clearly, although some cytoplasmic inclusions suggestive of such configurations were observed. Observations of myelin in EC cytoplasm were frequent. In some cases, zones of amorphous breakdown material within the cytoplasm of EC, as well as single ECs undergoing destruction, occurred, with both processes appearing to participate in lumen formation.

View larger version (52K): [in this window] [in a new window]   Fig. 4. EC in mixed spheroid cultures form a complex 3-D network. A: series of 45 consecutive, aligned, immunohistochemically stained 5- to 6-µm sections of one paraffin-embedded spheroid coculture seeded at a ratio of 40 to 1 and cultured over a 14-day period in VEGF-supplemented EGM-2. EC labeling was performed according to methods described in Fig. 2. Color information was filtered and converted to create binary images. B: isosurface calculation from original images (left lower panel) and overlay production from binary images shown in A of each of 11 adjacent sections (top left) or of all 45 sections (top right). Combination of isosurface and overlay productions (bottom right).

View larger version (154K): [in this window] [in a new window]   Fig. 5. EC network in mixed spheroid cultures shows tubulelike structures. A: transmission electron photomicrograph of a selected area of a 14-day-old spheroid coculture (4-to-1 F-to-EC ratio) displaying lumen formation by tubular arranged ECs. Note the numerous tight junctions laterally connecting adjacent cells (small open arrows) and segmental basement membrane formations (thick open arrows) at the adluminal cell pole, which delineate the tubular structure to the surrounding region (cf. Ref. 12). B: higher magnification of area depicted in A showing multiple pinocytotic vesicles (solid arrowheads) in the luminal cell surface as well as tight junctions (small open arrows) of the lateral junctional complex sealing the luminal space (asterisk). In one of the cells, cytoplasmic inclusion suggestive of a Weibel-Palade body is shown (long open arrow); however, due to the oblique plane of the section, it could not clearly be distinguished from myelin figures observed in many cells. C: luminal aspect of another tubular structure showing fingerlike cell protrusions (thick solid arrows) in the top part of the image and cell debris (solid cross) in the bottom part. Bar, 1 µm.

View larger version (47K): [in this window] [in a new window]   Fig. 6. ECs migrate into F spheroids. EC migration and network formation in 3-D coculture are critically affected by F-derived factors. Representative images of 5- to 6-µm sections of 3-day (top) and 14-day-old F spheroids (bottom) cocultured with EC suspensions (1 x 104 ECs/F spheroid) for an additional period of 2 to 3 or 8 days, respectively. Immunohistological processing of paraffin-embedded material and EC labeling (CD31) was performed as described in Fig. 2. EC migration is incomplete in long-term precultured F spheroids. Histomorphological observations imply that F differentiation or cell death phenomena throughout culturing confine EC migration.

View larger version (41K): [in this window] [in a new window]   Fig. 7. Lack of EC migration into F spheroids is not due to loss of F viability. EC network formation in F spheroids is causally related to EC survival. Representative images of 5- to 6-µm sections of F spheroid monocultures (A, top) at days 8, 15, and 21 stained with a green fluorescent DNA fragment end labeling to visualize apoptotic cells. F monocultures were also collected, dissociated, and analyzed using flow cytometry with fluorescently labeled AnnexinV and PI to discriminate membrane defect in parallel to early apoptotic cells. Same staining protocol was applied to visualize apoptotic cells in F spheroids precultured for 14 days (B, bottom left) or 3 days (B, bottom right) and cocultured with EC suspensions for 8 and 14 days, respectively. Parallel sections were labeled with an anti-CD31 antibody combined with an Alexa 546-conjugated secondary antibody to show EC distribution and networks. C: multicolor staining with a FITC-conjugated anti-human CD31, PI, and AnnexinV-allophycocyanin (APC) for flow cytometric analysis was performed with cell suspensions derived from 14-day precultured F spheroids that were cocultured with ECs for an additional 8 days. Membrane-intact F and ECs were defined as described in Fig. 2B. Cell loss, PI-positive, and AnnexinV-positive and -negative cell fractions from 3 individual experiments were averaged to show that the majority of cells were lost (left). Indeed, only a small (5%) proportion of the 10,000 ECs seeded per F spheroid attached and migrated. However, the majority of those that migrated and potentially differentiated into an EC network during a period of 8 days obviously was not apoptotic (center and right). Viable F-to-EC ratio in these cultures was between 10 to 1 and 20 to 1, in accordance with mixed culture systems.

We also questioned whether ECM components accumulated in the fibroblast-spheroid center. We previously studied the distribution of collagen types I and IV and did not find systematic and massive alterations in these ECM components. Taking into account a collagen type I-positive tumor tissue control, it is noteworthy that this same collagen was below the level of detection in paraffin-embedded fibroblast spheroids. Collagen type IV in cocultures was found primarily in the EC areas.

	DISCUSSION

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
Spheroids as experimental tools. The idea of using the spheroid model as a basis for the study of angiogenesis, among other approaches, may seem obvious, considering that it has been used in biomedical research since the early 20th century to study the mechanisms of organogenesis and the expression of malignancy, as well as to mimic tissues in vivo (34, 37, 42, 43). However, attention was mostly focused on isolated cell types of epithelial and tumor origin. This (simple) spheroid system clearly is of intermediate complexity with regard to pathophysiological micromilieu and histomorphology, but it does not reflect organ- and/or tissue-specific cellular heterogeneity. Heterologous intercellular interactions began to be appreciated by cell culturists decades later, and investigators have made sporadic attempts to reestablish spheroid cultures during the past 15 years using diverse stromal cell populations relevant in tumor tissue, such as immune cells and fibroblasts (13, 33, 37, 38). The application of EC suspensions with tumor cell spheroids has not been successful, however, supposedly because ECs do not survive in this approach. This clearly was one of the rationales for Korff and Augustin (31) to establish a tumor cell-free EC spheroid model, even though such 3-D EC clusters were expected to be poor mimics of the in vivo situation. Indeed, interesting differentiation phenomena were observed at the periphery of these EC spheroids, and apoptosis in the spheroid center was manipulable. This model became more advanced with the use of two different approaches. In the first, Vernon and Sage (52) embedded EC spheroids initiated with a hanging drop technique in type I collagen to analyze radial EC migration and morphology. This model, radial invasion of matrix by aggregated cells, has been expanded using a confrontation culture approach with tumor cell spheroids (20, 21). In the second model, ECs and stromal cells, namely, smooth muscle cells, were co-inoculated in nonadherent, round-bottomed, 96-well plates to form EC-smooth muscle cell spheroid cocultures (32). To some extent, this system reflects an inside-out presentation of a vessel with a central region of smooth muscle cells surrounded by one layer of ECs oriented toward the nutrient supply and culture medium.

We used fibroblast but not smooth muscle cells as a stromal cell type interacting with ECs and found an entirely different pattern. An EC surface layer was rarely observed; instead, fibroblast and EC mixed to form a capillary-like EC network within the spheroid. By using EC suspensions, we confirmed that ECs not only attach to fibroblast and fibroblast-derived ECM but also migrate into fibroblast spheroids. This obviously is not the case with smooth muscle cells. Accordingly, it would be most relevant in advanced vessel engineering to create human three-cell-type (spheroid) culture systems consisting of ECs and both adjacent cell types relevant for vessel assembly and maturation.

Fibroblast-EC spheroid cocultures. We have successfully established the first cell system that does not require any artificial matrix material or animal cell systems and is capable of forming a 3-D network of EC tubular structures with extensions of up to 400 µm, applying human cell types on the basis of a similar strategy used by Bishop et al. (8). It has become evident that under these culture conditions, fibroblast could provide all factors, including VEGF, required for the survival, differentiation, and network formation of ECs in a 3-D environment but also that these processes critically depend on an optimal balance of heterologous interactions. For example, high proportions of ECs lead to cluster formation accompanied or followed by apoptotic EC death. This massive cell death, which is not found in areas of EC tubular structures in the spheroid cocultures, clearly does not reflect the process of EC apoptosis found in the middle of a cord interpreted to contribute to lumen formation and maturation of patent small blood vessels (15, 51). Both culture strategies applied in the present study imply an optimum fibroblast-to-EC ratio for EC attachment, survival, and network formation that is in the range of 10 to 1. Several observations support this interpretation. 1) In mixed cultures with high numbers of co-seeded ECs, the majority of ECs are lost during the 3-day initiation phase, with cultures at day 3 containing 10% EC. This ratio stayed constant throughout further cultivation; that is, the number of ECs decreased in parallel with a loss of membrane-intact fibroblast. 2) In cultures with a 20-to-1 fibroblast-to-EC ratio at day 3, the number of ECs was stable throughout cultivation (day 15 vs. day 3) despite the decrease in viable fibroblast. Accordingly, the F-to-EC ratio at day 15 in culture declined to 10 to 1. 3) In fibroblast spheroid-EC suspension cocultures (strategy 2), the number of ECs added was extremely high (10,000), but only 5–10% of these cells attached and migrated into 14-day-old fibroblast spheroids. Attached and migrated ECs accounted for a 10-to-1 fibroblast-to-EC ratio after 3 days in coculture.

EC proliferation in our spheroid coculture model was not analyzed in detail. However, the lack of spheroid growth in volume indicates cell quiescence of both EC and fibroblast populations, much as the situation in fibroblast and EC spheroid monocultures (32, 36). Indeed, spheroid volume decreased at later growth phases as indicated in the median sections shown in Fig. 2A. Flow cytometric analyses of dissociated spheroid cocultures revealed this impression, showing that the overall number of membrane-intact cells is reduced at later culture phases independently of VEGF supplementation (day 15 vs. days 3 and 8). However, it was also shown that the number of membrane-intact ECs only decreased along with fibroblast; that is, there was clearly no reduction (but rather an opposite tendency) in the number of membrane-intact ECs from days 3 through 8 in culture. In the present study, the proportion of apoptotic ECs according to AnnexinV labeling also was minimal. The phenomenon of F loss at later coculture stages is under further investigation, as is the observation that the addition of EC to fibroblast using both culture strategies in general is accompanied by a deprivation of membrane-intact fibroblast (see, e.g., Fig. 7A vs. Figs. 2B, 3B, and 7C).

Although immature neovasculature is known to be leaky and fragile and depends on a continuous supply of VEGF for survival (6, 45), vascular maturation is associated with arrest of angiogenesis through contact inhibition of EC proliferation in coculture systems with pericytes or smooth muscle cells (5, 7, 29, 47). Accordingly, we hypothesize that our system mimics a relatively mature state of the tubular network that might also be found in vivo, although this hypothesis must rely on further detailed work to fully characterize the system. Our data clearly show that the formation of the EC network does not require the constant addition of exogenous VEGF. Fibroblast were shown to produce some VEGF at least at the onset of coculturing. However, the level of VEGF was quite small and also rapidly decreased as a function of time in culture. Thus we conclude that either VEGF released at the onset of coculturing is operating and sufficient for EC survival or ECs become autonomous and thus require less VEGF as they become organized.

In contrast to tumor spheroid-based experiments (data not shown), ECs in suspension were capable of adhering to and migrating into the ECM-rich fibroblast spheroids. In the present study, interestingly, EC migration was restricted to peripheral areas when the period of fibroblast spheroid pre-culturing was increased. This lack of EC migration cannot be due to loss of fibroblast viability in the spheroid center. An impact of the early drop in VEGF production by fibroblast in 3-D culture is under discussion. However, we emphasize that EC migration into old fibroblast spheroids was limited in the presence of VEGF-supplemented media. Our morphological observations imply that the underlying mechanism may be causally related to fibroblast differentiation phenomena that could be identified in future studies, because they may become particularly useful for the selective manipulation of EC migration and network formation. In the present study, nutrient and oxygen gradients from the spheroid periphery to the center may play a role, although previously published data imply that central hypoxia with PO₂ values <2% (14 mmHg) is not expected in fibroblast spheroids with an 400-µm diameter (35).

Reconstruction of tubular structures. Many tests for angiogenesis are based on the measurement of the ability of flat ECs to turn into tubular structures. Assays mimicking vasculogenesis and/or angiogenesis (or simply tubulogenesis) are gel based, often fibrin or collagen clots, but most commonly Matrigel, a matrix-rich product prepared from Engelbreht-Holm-Swarm tumor cells. More recently, an advanced biomatrix gel-type system consisting of complex mixtures of ECM components, e.g., collagen types I and IV, laminin, fibronectin, heparan sulfate proteoglycans, and other commercially available compounds have been developed that supposedly better represent the matrix architecture of stromal and basement membrane compartments to be applied for tumorigenesis and angiogenesis studies. Also, more complex systems, such as the organotypic (aortic ring), the chick aortic arch, and Matrigel sponge assays, were developed to analyze vascular responses to paracrine factors and antiangiogenic or proangiogenic substances in vitro. Several reviews covering different aspects of the multifaceted angiogenic process reflected in these in vitro assay systems can be found in the literature (3, 20, 48). Despite some general criticisms of the relevance of these in vitro vs. in vivo test systems, most investigators recognize that currently available in vitro assays, although useful in complex mechanistic studies, usually cannot be used in tissue engineering of blood vessels. One limitation is that a truly 3-D EC network and vessel maturation are seldom achieved, although the formation in a quasi-3-D monolayer system has its merits (8). One also must constantly keep in mind that in vitro systems do not have back-pressure of blood within the opening vessels. Another problem is the reliance of several models on nonhuman material or cells, which generally bars their reimplantation into humans unless immunosuppressed.

A startlingly ambitious study to create long-lasting blood vessels was presented recently by Koike et al. (28), who showed that 10T1/2 mesenchymal precursor cells seeded together with HUVECs in a 3-D fibronectin type I collagen gel critically supported the formation and stabilization of perfused vessels after implantation of defined pieces of the solidified gel construct into the cranial window in severe combined immunodeficiency mice. Although this study, again, is one that combines human ECs with a mouse (mesenchymal) cell type and matrix, it is in strong accord with our hypotheses and findings. Exactly how integration of existing vessels with the endothelial network in spheroids (or tissue grafts) occurs remains an intriguing yet highly complex behavioral process of ECs. However, our model is a nonadherent system that can be manipulated easily to be transferred into animal models to study this process.

Conclusions and perspectives. Today's cell biologists and tissue engineers face two particular components of complex systems, namely, three-dimensionality and heterogeneity (1). Both of these issues were integral to our study in verifying that human fibroblast are essential supporters of human EC survival and migration, as well as important modulators (contributors) of expanded capillary-like networks in vitro. This extends the key role played by fibroblast in other aspects of wound healing (44). The particular focus of our study was to develop a system for future morphogenetic studies that would be useful in gaining deeper insight into the complex process of microvascular reconstruction both for studying tumor angiogenesis and for future-oriented, patient-specific, tissue-engineering purposes in regenerative medicine.

To test different culture protocols, we applied HUVECs as the most widely used EC population of human origin in vitro. However, HUVECs are embryonic and may differ from organ-specific adult ECs (20). It is our firm intent in future experiments to apply fibroblast and ECs from the same person when possible for tissue-engineering purposes. It will also be interesting to evaluate the capacity of adult macro- and microvascular ECs to form such a 3-D network in the spheroid system. On the basis of previous observations using the planar coculture model with human dermal microvascular ECs (8, 14), we hypothesize that different EC populations can be applied successfully under the culture conditions described. The same is true for fibroblast; adult fibroblast can be a highly heterogeneous multifunctional cell population despite their similar morphology (22, 49). However, fibroblast of different origins (skin, breast, and from normal and tumor stroma) were used in previous monolayer coculture systems, and the formation of tubular structures was in principle induced by all fibroblast types. Also, much of our work relies on the addition of FCS, which needs to be substituted so that homologous, and consequently autologous, sera can be used as medium supplements. A prospective experimental setup shall not only obviate the need for any animal-derived material but also improve the quality of the formed vessel-like structures by a defined increase in complexity, e.g., by addition of smooth muscle cells. Finally, it is most relevant to apply fibroblast + EC coculture strategies in nonspheroid alternative culture systems such as microstructures or biodegradable polymers that are designed to produce particular 3-D shapes for tissue-engineering purposes.

Two additional, specific developments are proposed, the first being to study the interaction of tumor cells with ECs, knowing that fibroblast must also play a part in this process. More than this, however, the second phase follows the abilities of different tumor types to instruct the fibroblasts and thereby control the levels of tumor angiogenesis that they indirectly induce. The spheroid coculture model that we have devised may become useful for assaying pro- and antiangiogenic substances in a 3-D tissue context, but it has not been designed specifically for this purpose. Indeed, the system is related more to the study of the angiogenic process itself as a developmental process throughout and after tubule network development as a result of heterologous intercellular interactions. Thus it may be particularly useful to study and identify (cellular) modulators systematically in the process of vascular maturation rather than during the early onset of angiogenesis.

	GRANTS

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
This work was supported by the Bavarian State Ministry of Sciences, Research and the Arts.

	DISCLOSURES

 
We claim no conflict of interest (e.g., consultancies, stock ownership, equity interests, patent-licensing arrangements, lack of access to data, lack of control of the decision to publish).

	ACKNOWLEDGMENTS

 
We gratefully acknowledge the effort of Thomas Wondrak in 3-D reconstruction analyses and the technical assistance of Beate Voll and Heiko I. Siegmund in performing transmission electron microscopy.

	FOOTNOTES

 

Address for reprint requests and other correspondence: L. A. Kunz-Schughart, Institute of Pathology, Univ. of Regensburg, Franz-Josef-Strauss-Allee 11, D-93053 Regensburg, Germany (e-mail: leoni.kunz-schughart{at}klinik.uni-regensburg.de)

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

	REFERENCES

TOP ABSTRACT MATERIALS AND METHODS RESULTS DISCUSSION GRANTS REFERENCES

 
1. Abbott A. Cell culture: biology's new dimension. Nature 424: 870–872, 2003.[CrossRef][Medline]

2. Albo D, Wang TN, and Tuszynski GP. Antiangiogenic therapy. Curr Pharm Des 10: 27–37, 2004.[CrossRef][ISI][Medline]

3. Auerbach R, Lewis R, Shinners B, Kubai L, and Akhtar N. Angiogenesis assays: a critical overview. Clin Chem 49: 32–40, 2003.[Abstract/Free Full Text]

4. Augustin HG. Translating angiogenesis research into the clinic: the challenges ahead. Br J Radiol 76: S3–S10, 2003.[Abstract/Free Full Text]

5. Banin VV. Role of pericytes in mechanism of vessel neovascularisation in the regenerating connective tissue (Article in Russian). Morfologiia 125: 45–50, 2004.[Medline]

6. Benjamin LE, Golijanin D, Itin A, Pode D, and Keshet E. Selective ablation of immature blood vessels in established human tumors follows vascular endothelial growth factor withdrawal. J Clin Invest 103: 159–165, 1999.[ISI][Medline]

7. Benjamin LE, Hemo I, and Keshet E. A plasticity window for blood vessel remodelling is defined by pericyte coverage of the preformed endothelial network and is regulated by PDGF-B and VEGF. Development 125: 1591–1598, 1998.[Abstract]

8. Bishop ET, Bell GT, Bloor S, Broom IJ, Hendry NF, and Wheatley DN. An in vitro model of angiogenesis: basic features. Angiogenesis 3: 335–344, 1999.[CrossRef][Medline]

9. Bodolay E, Koch AE, Kim J, Szegedi G, and Szekanecz Z. Angiogenesis and chemokines in rheumatoid arthritis and other systemic inflammatory rheumatic diseases. J Cell Mol Med 6: 357–376, 2002.[ISI][Medline]

10. Carlsson J and Yuhas JM. Liquid overlay culture of cellular spheroids. In: Spheroids in Cancer Research, edited by Acker H, Carlsson J, Durand R, and Sutherland RM. Berlin, Germany: Springer-Verlag, 1984, p. 1–23.

11. Carmeliet P and Jain RK. Angiogenesis in cancer and other diseases. Nature 407: 249–257, 2000.[CrossRef][Medline]

12. Darland DC, Massingham LJ, Smith SR, Piek E, Saint-Geniez M, and D'Amore PA. Pericyte production of cell-associated VEGF is differentiation-dependent and is associated with endothelial survival. Dev Biol 264: 275–288, 2003.[CrossRef][ISI][Medline]

13. De Ridder L, Cornelissen M, and de Ridder D. Autologous spheroid culture: a screening tool for human brain tumour invasion. Crit Rev Oncol Hematol 36: 107–122, 2000.[ISI][Medline]

14. Donovan D, Brown NJ, Bishop ET, and Lewis CE. Comparison of three in vitro human "angiogenesis" assays with capillaries formed in vivo. Angiogenesis 4: 113–121, 2001.[CrossRef][Medline]

15. Fierlbeck W, Liu A, Coyle R, and Ballermann BJ. Endothelial cell apoptosis during glomerular capillary lumen formation in vivo. J Am Soc Nephrol 14: 1349–1354, 2003.[Abstract/Free Full Text]

16. Folkman J. Angiogenesis inhibitors: a new class of drugs. Cancer Biol Ther 2: S127–S133, 2003.[ISI][Medline]

17. Folkman J. Role of angiogenesis in tumor growth and metastasis. Semin Oncol 29: 15–18, 2002.[ISI][Medline]

18. Folkman J. Fundamental concepts of the angiogenic process. Curr Mol Med 3: 643–651, 2003.[CrossRef][ISI][Medline]

19. Gagne P, Akalu A, and Brooks PC. Challenges facing antiangiogenic therapy for cancer: impact of the tumor extracellular environment. Expert Rev Anticancer Ther 4: 129–140, 2004.[CrossRef][Medline]

20. Goldbrunner RH, Bendszus M, and Tonn JC. Models for angiogenesis in gliomas. Cancer Treat Res 117: 115–135, 2004.[Medline]

21. Goldbrunner RH, Wagner S, Roosen K, and Tonn JC. Models for assessment of angiogenesis in gliomas. J Neurooncol 50: 53–62, 2000.[CrossRef][Medline]

22. Grégoire M and Lieubeau B. The role of fibroblasts in tumor behavior. Cancer Metastasis Rev 14: 339–350, 1995.[CrossRef][ISI][Medline]

23. Jaffé EA, Nachman RL, Becker CG, and Minick CR. Culture of human endothelial cells derived from umbilical veins: identification by morphologic and immunologic criteria. J Clin Invest 52: 2745–2756, 1973.[ISI][Medline]

24. Jain RK. Molecular regulation of vessel maturation. Nat Med 9: 685–693, 2003.[CrossRef][ISI][Medline]

25. Jain RK. Normalization of tumor vasculature: an emerging concept in antiangiogenic therapy. Science 307: 58–62, 2005.[Abstract/Free Full Text]

26. Janvier R, Sourla A, Koutsilieris M, and Doillon CJ. Stromal fibroblasts are required for PC-3 human prostate cancer cells to produce capillary-like formation of endothelial cells in a three-dimensional co-culture system. Anticancer Res 17: 1551–1557, 1997.[ISI][Medline]

27. Kerbel R and Folkman J. Clinical translation of angiogenesis inhibitors. Nat Rev Cancer 2: 727–739, 2002.[CrossRef][ISI][Medline]

28. Koike N, Fukumura D, Gralla O, Au P, Schechner JS, and Jain RK. Tissue engineering: creation of long-lasting blood vessels. Nature 428: 138–139, 2004.[CrossRef][Medline]

29. Kondo T, Hosoya K, Hori S, Tomi M, Ohtsuki S, and Terasaki T. PKC/MAPK signaling suppression by retinal pericyte conditioned medium prevents retinal endothelial cell proliferation. J Cell Physiol 203: 378–386, 2005.[CrossRef][ISI][Medline]

30. Konttinen YT, Mackiewicz Z, Ruuttila P, eponis A, Sukura A, Povilenaite D, Hukkanen M, and Virtanen I. Vascular damage and lack of angiogenesis in systemic sclerosis skin. Clin Rheumatol 22: 196–202, 2003.[CrossRef][ISI][Medline]

31. Korff T and Augustin HG. Integration of endothelial cells in multicellular spheroids prevents apoptosis and induces differentiation. J Cell Biol 143: 1341–1352, 1998.[Abstract/Free Full Text]

32. Korff T, Kimmina S, Martiny-Baron G, and Augustin HG. Blood vessel maturation in a 3-dimensional spheroidal coculture model: direct contact with smooth muscle cells regulates endothelial cell quiescence and abrogates VEGF responsiveness. FASEB J 15: 447–457, 2001.[Abstract/Free Full Text]

33. Kunz-Schughart LA. Multicellular tumor spheroids: intermediates between monolayer culture and in vivo tumor. Cell Biol Int 23: 157–161, 1999.[CrossRef][ISI][Medline]

34. Kunz-Schughart LA, Freyer JP, Hofstaedter F, and Ebner R. The use of 3-D cultures for high-throughput screening: the multicellular spheroid model. J Biomol Screen 9: 273–285, 2004.[Abstract]

35. Kunz-Schughart LA, Groebe K, and Mueller-Klieser W. Three-dimensional cell culture induces novel proliferative and metabolic alterations associated with oncogenic transformation. Int J Cancer 66: 578–586, 1996.[CrossRef][ISI][Medline]

36. Kunz-Schughart LA, Heyder P, Schroeder J, and Knuechel R. A heterologous 3-D coculture model of breast tumor cells and fibroblasts to study tumor-associated fibroblast differentiation. Exp Cell Res 266: 74–86, 2001.[CrossRef][ISI][Medline]

37. Kunz-Schughart LA, Kreutz M, and Knuechel R. Multicellular spheroids: a three-dimensional in vitro culture system to study tumour biology. Int J Exp Pathol 79: 1–23, 1998.[CrossRef][ISI][Medline]

38. Kunz-Schughart LA and Mueller-Klieser W. Three-dimensional culture. In: Animal Cell Culture, edited by Masters JRW. Oxford, UK: Oxford University, 2000, p. 123–148.

39. Kunz-Schughart LA, Wenninger S, Neumeier T, Seidl P, and Knuechel R. Three-dimensional tissue structure affects sensitivity of fibroblasts to TGF-₁. Am J Physiol Cell Physiol 284: C209–C219, 2003.[Abstract/Free Full Text]

40. Martin TA, Harding KG, and Jiang WG. Regulation of angiogenesis and endothelial cell motility by matrix-bound fibroblasts. Angiogenesis 3: 69–76, 1999.[CrossRef][Medline]

41. Middleton J, Americh L, Gayon R, Julien D, Aguilar L, Amalric F, and Girard JP. Endothelial cell phenotypes in the rheumatoid synovium: activated, angiogenic, apoptotic and leaky. Arthritis Res Ther 6: 60–72, 2004.[ISI][Medline]

42. Mueller-Klieser W. Three-dimensional cell cultures: from molecular mechanisms to clinical applications. Am J Physiol Cell Physiol 273: C1109–C1123, 1997.[Abstract/Free Full Text]

43. Mueller-Klieser W. Tumor biology and experimental therapeutics. Crit Rev Oncol Hematol 36: 123–139, 2000.[ISI][Medline]

44. Muir IF, Padilla-Lamb A, Stewart JE, and Wheatley DN. Growth inhibition of cultured fibroblasts by extracts from human dermis. Br J Plast Surg 50: 186–193, 1997.[CrossRef][ISI][Medline]

45. Nehls V, Schuchardt E, and Drenckhahn D. The effect of fibroblasts, vascular smooth muscle cells, and pericytes on sprout formation of endothelial cells in a fibrin gel angiogenesis system. Microvasc Res 48: 349–363, 1994.[CrossRef][ISI][Medline]

46. Nicosia RF and Villaschi S. Autoregulation of angiogenesis by cells of the vessel wall. Int Rev Cytol 185: 1–43, 1999.[ISI][Medline]

47. Orlidge A and D'Amore PA. Inhibition of capillary endothelial cell growth by pericytes and smooth muscle cells. J Cell Biol 105: 1455–1462, 1987.[Abstract/Free Full Text]

48. Oudar O. Spheroids: relation between tumour and endothelial cells. Crit Rev Oncol Hematol 36: 99–106, 2000.[ISI][Medline]

49. Schor AM, Rushton G, Ferguson JE, Howell A, Redford J, and Schor SL. Phenotypic heterogeneity in breast fibroblasts: functional anomaly in fibroblasts from histologically normal tissue adjacent to carcinoma. Int J Cancer 59: 25–32, 1994.[ISI][Medline]

50. Szekanecz Z, Gaspar L, and Koch AE. Angiogenesis in rheumatoid arthritis. Front Biosci 10: 1739–1753, 2005.[ISI][Medline]

51. Tertemiz F, Kayisli UA, Arici A, and Demir R. Apoptosis contributes to vascular lumen formation and vascular branching in human placental vasculogenesis. Biol Reprod 72: 727–735, 2005.[Abstract/Free Full Text]

52. Vernon RB and Sage EH. A novel, quantitative model for study of endothelial cell migration and sprout formation within three-dimensional collagen matrices. Microvasc Res 57: 118–133, 1999.[CrossRef][ISI][Medline]

53. Villaschi S and Nicosia RF. Paracrine interactions between fibroblasts and endothelial cells in a serum-free coculture model: modulation of angiogenesis and collagen gel contraction. Lab Invest 71: 291–299, 1994.[ISI]

54. Walter-Yohrling J, Pratt BM, Ledbetter S, and Teicher BA. Myofibroblasts enable invasion of endothelial cells into three-dimensional tumor cell clusters: a novel in vitro tumor model. Cancer Chemother Pharmacol 52: 263–269, 2003.[CrossRef][ISI][Medline]