Rapid Perfusion and Network Remodeling in a Microvascular Construct After Implantation
http://www.100md.com
《动脉硬化血栓血管生物学》
From the Biomedical Engineering Program (H.Y.S.C., C.M.S., G.G., S.K.W., J.B.H.) and Physiological Sciences Program (B.R.S., S.K.W., J.B.H.), Vascular Research Group, University of Arizona, Tucson. B.R.S. is currently at the Boyer Center for Molecular Medicine, Yale School of Medicine, New Haven, Conn.
ABSTRACT
Objective— We have previously demonstrated the ability to construct 3-dimensional microvascular beds in vitro via angiogenesis from isolated, intact, microvessel fragments that retain endothelial cells and perivascular cells. Our objective was to develop and characterize an experimental model of tissue vascularization, based on the implantation of this microvascular construct, which recapitulated angiogenesis, vessel differentiation, and network maturation.
Methods and Results— On implantation in a severe combined-immunodeficient mouse model, vessels in the microvascular constructs rapidly inosculated with the recipient host circulation. Ink perfusion of implants via the left ventricle of the host demonstrated that vessel inosculation begins within the first day after implantation. Evaluation of explanted constructs over the course of 28 days revealed the presence of a mature functional microvascular bed. Using a probe specific for the original microvessel source, 91.7%±11% and 88.6%±19% of the vessels by day 5 and day 28 after implantation, respectively, were derived from the original microvessel isolate. Similar results were obtained when human-derived microvessels were used to build the microvascular construct.
Conclusions— With this model, we reproduce the important aspects of vascularization, angiogenesis, inosculation, and network remodeling. Furthermore, we demonstrate that the model accommodates human-derived vessel fragments, enabling the construction of human-mouse vascular chimeras.
Key Words: vascularization ? microcirculation ? angiogenesis ? human ? vascular remodeling
Introduction
Vascularization is the process by which perfusion pathway length and vessel segment number are increased and organized into a functional vascular bed. In normal situations, this effective increase in vessel density delivers more blood to the tissue, facilitating tissue growth and/or increased tissue activity.1,2 Consequently, vascularization is a primary component of tissue growth and repair, such as occurs during development,3 after an upstream occlusive event leading to tissue ischemia,4 or during proliferative events, as seen in tissue healing5,6 and tumor growth.7 Although we know of many factors and signals that initiate or terminate the vascularization process, little is known about the rules that govern vascularization as an integrated process that includes angiogenesis,3 arteriogenesis,8 vascular remodeling,9 vessel adaptation,10 and arterio-venous polarization.11
We have previously shown that isolated intact microvessel fragments retain angiogenic potential and are capable of forming a simple microvascular bed when cultured in a 3-dimensional collagen I gel.12 In this microvascular construct, the vessel fragments undergo stereotypical angiogenesis, forming neovessels that maintain patent lumen and perivascular cell associations. Furthermore, the vessel fragments within this culture system are responsive to proangiogenic conditions.12,13 All of this occurs in the absence of blood flow and relatively few nonvascular cells.
Here we report the development and characterization of an experimental model of tissue vascularization based on the implantation of this microvascular construct. Precultured or freshly formed microvascular constructs implanted subcutaneously inosculate with the host vasculature and expand into a microvascular bed containing perfused mature heterogeneous vessel elements. With this model, we reproduce the important aspects of vascularization, angiogenesis, inosculation, and network remodeling. Furthermore, we demonstrate that the model accommodates human-derived vessel fragments, enabling the construction of human-mouse vascular chimeras.
Methods
All animal procedures were performed according to University of Arizona IACUC-approved protocols. Mice were anesthetized with an intraperitoneal injection of 2.5% avertin (0.1 mL/10 g) or euthanized by CO2 asphyxiation. Rats were euthanized with an overdose of pentobarbital (150 mg/kg).
Microvascular Constructs
Rat fat microvessel fragments (RFMFs) were isolated from epididymal fat pads of retired breeder male Sprague-Dawley rats. Under aseptic conditions, harvested fat pads were finely minced with scissors, digested in 2 mg/mL collagenase plus 2 mg/mL EFAF-BSA in phosphate-buffered saline (PBS) for 8 minutes at 37°C with vigorous shaking and washed in 0.1% BSA-PBS. Tissue debris and large vessel pieces were removed by filtering the suspension through a sterile 500-μm-pore nylon screen. Microvessel fragments were captured by filtration of the remaining suspension on a 30-μm-pore nylon screen and recovered by vigorous flushing of the screen surface with 0.1% BSA-PBS. The type and lot number of collagenase used were predetermined to optimize fragment yield while maintaining microvessel integrity. Microvessel fragments (MF) were suspended (12 000 to 15 000 MF/mL) in ice-cold 3 mg/mL rat tail type I collagen (BD BioSciences, Bedford, Mass) prepared with Dulbecco minimum essential medium (DMEM) (1x final) and pH neutralized with 1 mol/L NaOH. MF/collagen suspensions were plated into individual wells (0.25 mL/well) of a 48-well plate and placed in a 37°C incubator for 20 minutes to polymerize the collagen. For culturing, an equal volume of 10% FBS/DMEM was added to each well and fed at day 4. No growth factors or supplements were added to the culture medium. Constructs were cultured for 5 to 7 days before implantation. For immediate implantation, newly formed constructs were transferred directly to the implant site without preculturing. Human fat microvessel fragments were isolated from discarded abdominoplasty adipose and cast into collagen gels using the same procedures as for rat fat microvessel fragments but were not cultured. In all cases, avascular control constructs were prepared from the same batches of collagen solution used to form microvascular constructs and at the same time.
Implantation
Microvascular constructs and avascular collagen gel controls were implanted in the subcutaneous position on the flanks of anesthetized female severe combined-immunodeficient (SCID) mice. For implantation, a subcutaneous pocket was formed between the skin and the underlying muscle anterior to the pelvis using blunt dissection through a small skin incision. Each host mouse received 2 implants: a microvascular construct on one side and an avascular control collagen gel on the contralateral side. The incision was then closed with 6-0 suture and the animal was allowed to recovered.
Histology and Histochemistry
Mice were anesthetized as mentioned and implants were removed, with a small portion of the underlying muscle fixed in 2% paraformaldehyde/PBS and processed into paraffin. General histology was performed on deparaffinized, 5- to 6-μm-thick sections (for details, see the data supplement available at http://atvb.ahajournals.org). Vessel density was determined by counting discreet GS-1–positive structures in the implanted construct from at least 5 different fields (of defined area) per section from 2 different implants. Individual counts were divided by the area of each field and averaged for each time point. For details of en bloc immunostaining, see the data supplement available at http://atvb.ahajournals.org.
Ink Perfusion
Mice containing implants were anesthetized with Avertin and placed supine on a dissecting stage. Through a left ventricular catheter (PE 60 tubing), blood was cleared with heparinized saline containing 10 μmol/L sodium nitroprusside followed by perfusion with India ink (Speedball 3398; Hunt Manufacturing, Statesville, NC), dialyzed against PBS, and filtered through #1 Whatman paper. All perfusions were performed at a maintained pressure of 90 to 100 mm Hg until all tissues in the mouse appeared dark (usually, this required 2 to 3 mL of ink solution). After ink perfusion, the implants were excised, fixed in 4% paraformaldehyde in PBS for 45 minutes at 4°C, and cleared in 100% glycerol for 20 minutes. Implants were bisected, and the 2 halves, cut-side up, were sandwiched between a microscope slide and coverslip for viewing with a standard light microscope.
In Situ Hybridization
A probe that detects a repeat element on the Y chromosome in rat14 was synthesized from rat-tail genomic DNA by PCR (forward: ggt tct aga ctg taa aac cca gac; reverse: act taa aac taa gct tat tgg cca), size-verified, and labeled with biotin using the Photoprobe labeling kit (Vector Laboratories, Burlingame, Calif). Labeled probe was hybridized to 8-μm sections using standard methods (for details, see http://.atvb.ahajournals.org). For determination of the percent Y chromosome-positive cells, serial sections (6-μm-thick) were stained by in situ hybridization for the Y chromosome or with hematoxylin to label all nuclei. The ratios of Y chromosome-positive counts to hematoxylin-positive counts (both derived from serial sections) from at least 5 areas of a section from 2 different implants were averaged.
Smooth Muscle Cell Coverage
All vascular cells and smooth muscle cells (SMCs) were colabeled in fixed intact microvascular constructs (cultured or implanted) by en bloc immunostaining (for details, see http://.atvb.ahajournals.org). The immunostained constructs were imaged using a 20x objective on a Biorad MRC-1024ES Confocal Microscope. Stacks of 512x512-μm images with 1.05x1.05-μm pixel resolution were generated with an image separation of 1.0 μm for each fluorescence channel. For determination of percent smooth muscle coverage, up to 4 stacks of images were obtained from constructs and the fractional area comprising total vascular cells (MHC-positive structures) and SMCs (alpha-actin-positive structures) were computed (for details, see http://atvb.ahajournals.org). MHC-positive structures with a total volume less than that of a 10-μm-long and 5 μm in diameter rod were removed from the image stacks, as were SMC components not associated with vascular components. Percent coverage of SMC for each stack was calculated as (SMC volume/vascular volume)x100.
Results
Isolated fat microvessel fragments undergo angiogenic sprouting within the first 5 days of culture in collagen I gels (see http://atvb.ahajournals.org). Sprouting occurs at the ends and from mid regions of individual fragments. Furthermore, the sprouting process is dynamic, with individual sprouts forming and regressing within a day. By 11 days in culture, fragments have grown to form a collection of elongated, simple neovessels (see http://atvb.ahajournals.org) with an average diameter of 24.8±6.8 μm (n=39). These neovessels contain patent lumen and a relatively low density of -actin-positive perivascular cells.12
Coordinated with angiogenesis, newly formed vessel elements must locate and inosculate with other elements of the vascular network before functioning as a blood perfusion circuit. We evaluated the ability of this microvascular construct to interact with an existing vasculature by implanting neovessel cultures into SCID mice. After 5 to 7 days of culture, constructs were placed under the skin, in direct contact with the dorsal musculature, and analyzed at various time points (days 1, 3, 5, 7, 10, 14, 21, and 28 after implantation). All microvascular constructs (4 or more per time point) contained vessels after implantation. Superficial, blood-filled vessels were associated with microvascular constructs but not with collagen gel controls (see http://atvb.ahajournals.org). Furthermore, the implanted microvascular constructs contained the full range of vessel types commonly seen in a mature functional vascular bed, including small arteries, arterioles, capillaries, venules, and veins (Figure 1a through 1d).
Figure 1. Vessels within microvascular implants are heterogeneous in structure and contain red blood cells. H&E stained sections of characteristic implanted microvascular constructs after 14 (a, b, d) or 10 (c) days of implantation in a subcutaneous pocket reveal the presence of arteries (A), arterioles (arrowhead in d), capillaries (arrows), venules (arrowhead in b), veins (V), and red blood cells within vessels.
We perfused ink into the left ventricle of the host mouse to determine when vessels within the construct become contiguous with the host mouse vasculature. By day 1, a limited number of microvessels in the implant contained ink indicating continuity with the mouse circulation (Figure 2a). By day 2, the construct vessels were beginning to assemble together, continuing to form interconnections by day 3 (Figure 2b and 2c). By 28 days after implantation, the vessels have refined into a mature-appearing vascular bed (Figure 2d). Orthogonal polarized spectral imaging, which selectively detects hemoglobin,15 of living implants reveals blood-containing vessel structures throughout the constructs by day 14 (Figure 2e). Only surface blood, because of the dissection to expose the construct was detected with orthogonal polarized spectral on control, vessel-free collagen gel implants (Figure 2f).
Figure 2. Vessels within microvascular constructs inosculate with the host vasculature and are perfusion-competent when implanted subcutaneously into SCID mice. a-d, Representative fields of microvascular constructs harvested on day 1 (a), day 2 (b), day 3 (c), or day 28 (d) after implantation after perfusion of the respective host mouse through the left ventricle with ink. e and f, Orthogonal polarized spectroscopic images of microvascular implants after 14 days of implantation (e) and at day 3 with avascular control gel (f). A 10-times magnification probe was used, which provided an imaging depth of 700 μm.
Staining of implant sections with the Griffonia simplicifolia-I lectin verified the presence of endothelial cell-comprised vessels (Figure 3a). Vessel densities (GS-1–positive structures) within the microvascular constructs increased >2-fold from day 5 to day 28 (Figure 3b), suggesting that angiogenesis continued within the implants during this time. Immunostaining of implants for -SMC actin, to corroborate the mature vessel phenotype observed in the histology, identified positive cells surrounding vessels of different caliber (see http://atvb.ahajournals.org). To determine whether the vessels within the implanted microvascular constructs were derived from the original cultured rat microvessels or, alternatively, were replaced by the ingrowth of host mouse vessels, we took advantage of the fact that the rat microvessels were isolated from males and the constructs were placed into female mice. Using a DNA probe specific to a repeat element of the rat Y chromosome,14 we analyzed sections from day 5 and day 28 implants by in situ hybridization. Vessel-like structures and single cells throughout the implant were detected with the Y chromosome-specific probe, whereas Y-positive cells were not observed within the adjacent muscle of the host mouse (Figure 3c). To determine the percentage of cells within the microvascular constructs that were derived from the original isolate, serial sections from day 5 and day 28 implants were stained with either hematoxylin, to identify all nuclei, or the Y-chromosome probe (please see http://atvb.ahajournals.org). Comparing the number of Y-positive cells to the total number of cells in serial sections indicates that nearly all of the cells (90%) within the implants, even after 4 weeks, comprised cells derived from the original rat microvessel isolate (Figure 3d).
Figure 3. Microvascular implants contain vascular endothelium and smooth muscle-positive perivascular cells that are nearly derived from the original cultured isolates. a, Day 5 implant stained with the rodent endothelial cell-specific lectin, GS-1, showing vessels of various caliber. b, Vessel density measurements determined from GS-1–stained sections of implanted microvascular constructs. c, Day 28 implant section labeled by in situ hybridization for a rat Y-chromosome-specific repeat. The "m" depicts the underlying host (mouse) musculature, while the "i" marks the implant. The arrow indicates the boundary between the 2 tissues. d, The fraction of Y-chromosome-positive cells in implanted constructs (see Methods) for day 5 and day 28 implants.
During angiogenesis and vessel maturation, perivascular cells dissociate and reassociate with vessel elements, respectively.16 Using confocal microscopy of en bloc immunostained constructs and volumetric reconstruction algorithms, we measured the percent coverage of vessel elements by perivascular cells (identified as -actin-positive cells) during the early and late stages of vascular activity in the constructs. Perivascular coverage during the culture period decreased from 83% in the freshly isolated fragments to nearly 50% coverage after 8 days in culture (Figure 4). After implantation and inosculation with the host vasculature, perivascular cell investment of construct vessels returned to normal levels (82%). Qualitatively, the vessel elements with the least coverage were predominately the new vessel sprouts and elongating neovessels. This reduction in coverage did not occur until after 3 days in culture. Typically, vessel sprouts from fragments do not form until 4 or 5 days in culture.12 During the active vessel growth phases, -actin-positive cells were often observed in extravascular spaces of the construct (data not shown).
Figure 4. Smooth muscle coverage of vessel elements changes during vascularization. Confocal image stacks of double-stained vessel elements (green=MHC and red=-actin) were used to generate vessel and perivascular volumes. Single-image planes of (a) a freshly isolated microvessel fragment (day 0), (b) a day 8 cultured fragment, and (c) day 28 postimplantation branched vessel element. d, Percentage of vessel elements covered by -actin-positive perivascular cells measured at different time points in the model. Coverage was assessed during culturing of microvascular constructs (day 0c-day 10c) and during implantation of constructs precultured for 8 days before implantation (day14i-day 28i).
To determine if preculturing is necessary for the microvascular constructs to be able to integrate with the host circulation, we implanted freshly formed microvascular constructs without preculturing. As before, microvessel fragments were isolated and embedded in collagen gels. However, these freshly formed constructs were immediately implanted without previous culturing. Similar to precultured implanted constructs, the noncultured constructs formed a recognizable vascular tree that was contiguous with the host circulation (Figure 5a and 5b). This prompted us to explore the possibility that human-derived vessel fragments could be used to build functional microvascular constructs. Unlike with rat-derived microvessels, human-derived microvessel fragments do not undergo angiogenesis in the collagen gels in vitro and instead either remain quiescent or dissociate into single cells (unpublished observation). As with noncultured rat-derived constructs, the freshly formed human microvascular constructs contained heterogenous vessel structures 15 days after implantation that were positive for the differentiation marker, von Willebrand factor (Figure 5c), and possessed -actin-positive cells (Figure 5d). Similar to what we observed with rat microvessels, constructs established from human sources contained primarily human (UEA-I-positive) vessels after 15 days of implantation (Figure 5e) with limited mouse (GS-1 positive) vessel invasion (Figure 5f), indicating that the original human fragments used to establish the construct were maintained in the implants.
Figure 5. Microvascular constructs assembled from rat-derived or human-derived freshly isolated microvessel fragments (no preculturing, see Methods) also form a vascular bed on implantation. a, En bloc fluorescence immunostaining for all cells (anti-MHC antibody) reveals that vessels within the construct form tree-like structures. b, Ink perfusion of the host and a superficial vessel connection (arrowheads) between host and construct (arrows indicate host muscle/construct boundary). Immunostaining of human-derived, day 15 implants for von Willebrand factor (c) or -actin (d) indicates that vessels (arrows) are differentiated and mature. Representative sections from the same human-derived implant stained with the human-specific lectin, UEA1 labels only vessels (arrows in e) within the construct, whereas the rodent-specific vascular marker GS-1 labels vessels within only the surrounding host mouse tissue (arrows in f). The dashed lines indicate the boundary between the implant and underlying host tissue.
Discussion
In the present study, we examined the capability of a cultured microvascular construct to interface with an existing vasculature and subsequently remodel into a functional perfusion circuit. Our goal was to develop an experimental model enabling the detailed study of the progression from angiogenesis through vessel differentiation to network maturation. The microvascular constructs consisted of a network of new microvessels in a collagen I gel that formed, via angiogenesis, from isolated microvessel fragments after in vitro culturing. In culture, the isolated microvessel fragments, which comprise the full spectrum of microvascular elements,12 form a collection of morphologically homogenous, simple vessels with low perivascular cell coverage. Once implanted, this simple vasculature inosculates with the host circulation and remodels to contain morphologically identifiable arteries, arterioles, capillaries, venules, and veins assembled into a perfused, vascular bed. Therefore, each vessel in this model (ie, a freshly isolated microvessel fragment) begins as a differentiated vessel element, dedifferentiates into a "generic-appearing" vessel during/after angiogenesis and subsequently redifferentiates into a specific element of a perfusion-competent vascular tree. In this regard, the entire vascularization process is recapitulated in this model including true angiogenesis (sprouting from preexisting parent vessels1), vessel differentiation, and network maturation.
A typical microvessel consists of endothelial cells arranged into a tube wrapped by one or more layers of perivascular cells (usually SMCs or pericytes). The interactions between the endothelial and perivascular cell layers are thought important in maintaining microvascular integrity and stability.17 During culturing of the microvessel fragments, -actin-positive perivascular cell coverage of vessels decreased to nearly half of that observed in isolated vessel fragments. This loss in coverage involved predominately vessel sprouts and growing neovessels. This is consistent with observations that perivascular cell withdrawal from vessel segments relaxes control of the endothelial cell tube and permits sprouting and vessel elongation during angiogenesis.18 After implantation, perivascular cell coverage increased to a level comparable to that observed in the microvessel fragments isolated from the mature adipose microvasculature indicating that vessel differentiation and maturation was being reestablished in the constructs. This is similar to a related study involving the implantation of 3-dimensional cultures of endothelial cells.19 In this study, endothelial cell tubes acquired perivascular cells from the host at a time coincident with vessel maturation (the presence of blood cells and loss of VCAM expression). In our study, perivascular cells were present in the microvascular constructs at the time of implantation (45% perivascular cell coverage). The continual presence of perivascular cells may make the microvessels within the microvascular construct capable of rapid maturation and could explain why they were able to rapidly establish an integrated perfusion-competent vasculature.
Although not specifically examined in this study, inosculation with the host circulation appears to be restricted to the construct boundaries. Few, if any, host vessels were observed in the construct, and construct vessels that were visibly contiguous with host vessels were superficially located on the construct. Whether inosculation occurred via end-end or end-side anastomoses between host and construct vessels is unclear. The formation of end-end connections would suggest an active involvement of the host circulation, perhaps via a coincidental angiogenesis response that causes vessel ends to grow toward each other. The presence of continuous vessels originating in the host tissue and spanning across the construct boundary to extend branches into the construct suggests that a host vascular response was involved. However, further studies with the constructs specifically directed at addressing these questions are required to determine the true mechanism by which vessel inosculation occurs.
It is important to point out that angiogenesis in the model occurs in the relative absence of nonvascular cells. Angiogenesis is often considered a vascular response to increased tissue needs and is associated with release of angiogenic factors by the surrounding tissue stroma.3 The occurrence of angiogenesis in our model with a minimal presence of stromal cells implies that vessels have an endogenous capability for new vessel formation. This is supported by the finding that the microvessel fragments in culture express Vegf A mRNA (data not shown). Interestingly, the neovessels resulting from angiogenesis in the constructs are homogeneous in character (24 μm average diameter and uniform low-level perivascular cell coverage). It is tempting to speculate that, perhaps, in the absence of stroma, this represents a default or baseline phenotype of the angiogenic neovessel. However, it is possible that these characteristics may be unique to the specific collagen gel environment used in the model or a consequence of culturing. However, culturing of the microvessel fragments was not required for the establishment of a functional vascular bed within the constructs after implantation. Freshly implanted constructs that had not been cultured still formed a perfusion-competent differentiated vascular bed. Therefore, no activities unique to culturing are essential for vascularization.
The appearance of flow throughout the entire construct precedes the presence of differentiated heterogenous vessel types. Hemodynamic stimuli to a vessel, namely shear stress and circumferential and axial wall stress, affects vessel diameter and wall thickness and possibly controls arteriole specification.20–22 Thus, it is likely that flow (shear stress) and pressure (wall stress) are contributing to vessel differentiation in the implanted constructs. Whether blood flow is the only stimulus establishing a completely mature network architecture is not clear. Observations in developing zebra fish embryos indicate that arteries and veins are specified before blood flow begins.23 Whether this is unique to the embryo or whether it may also occur in the adult remains to be determined. Our results do not exclude the intriguing possibility that formation of artery-side and vein-side vessel types may reflect some aspect of "memory" in which the original vessel types present within the microvessel fragment isolate retain some predetermination. This experimental model, with the particular inclusion of the in vitro culturing, should prove invaluable in addressing this question and those generally related to artery/vein specification and differentiation after angiogenesis in the adult.
Experiments with human-derived microvessel fragments demonstrate the potential usefulness of this model in constructing an experimental human microcirculation without the need to modify or engineer the human cells. Previous models examining the human microcirculation in vivo have involved implantation of cultured human microvessel endothelial cells24,25 or the transplantation of full-thickness skin grafts.26,27 However in implants using cultured endothelial cells, it was necessary to prolong endothelial cell survival by overexpressing either telomerase25 or bcl-228 through stable transfections of the endothelial cells. We present an alternate model of the human microcirculation involving the use of isolated human microvessel fragments and subsequent implantation into a SCID mouse. This human/mousechimera model, which is easier to perform than the full-thickness skin grafting and does not require additional manipulation to preserve vessel survival, produces an intact differentiated human microcirculation in vivo. The resulting human/mouse vascular chimeras will prove useful in further studies exploring the mechanisms and processes of angiogenesis and vascular integration in the human microcirculation, and will explore microvessel/tissue cell interactions such as capillary exchange, metastasis, and inflammation. Furthermore, the human-derived constructs represent an alternative strategy to vascularizing tissue-engineered constructs by producing a ready-made vasculature comprising endothelial cells and perivascular cells. The rapid initial perfusion and subsequent remodeling obtained with the microvascular construct, once implanted, would provide early support of stromal cells incorporated into an engineered tissue, earlier than observed with other strategies, and establish a stable perfusion bed.
In summary, we present a model of vascularization based on the formation of a microvascular construct grown from intact microvessel elements. Two key features of this experimental model of vascularization are: (1) the vessels within the construct progress through stereotypical stages of vascularization and (2) it enables examination of vascularization in the absence of confounding tissues. In vascularizing tissues, in can be difficult to capture only those vessels within the tissue that are undergoing angiogenesis or network remodeling for subsequent analysis. This is caused, in part, by the regional heterogeneity in vascularization stimuli present throughout many actively remodeling or repairing tissues, resulting in vessels of the tissue being at different stages of vascularization at the same time.4,29 In contrast to repairing tissues, our vascularization model is fairly well synchronized. Microvessel fragments are harvested from relatively stationary microvascular beds and cast into the collagen gels at the same time. Consequently, they begin the vascularization process at the same time and synchronously progress through stereotypical phases of vascularization. Culturing of the microvessels recapitulates only the early angiogenesis phase of vascularization, enabling a detailed study of angiogenic sprouting and other angiogenesis-specific activities. Furthermore, the ability to culture the microvessel fragments without affecting their ability to inosculate and carry blood extends the usefulness of the model to include gene delivery and drug treatment experimental approaches. An important aspect of the model is the relative absence of host cells in the construct and the persistence of the original vessels when implanted. This aspect of the microvascular construct simplifies interpretation of experimental results by limiting relevance to vascular cells. However, nonvascular tissue cells could be included with the microvessels in the construct, which would permit study of tissue/vessel interactions. The microvascular construct implanted into the SCID mouse serves as a complete model of vascularization and enables evaluation of the molecular, cellular, and physiological mechanisms mediating the active remodeling of a vasculature.
Acknowledgments
This research was supported in part by an Arizona Elks Major Projects Award (J.B.H.), The Mary Lou Hemler Arnold Award for Heart Disease, Stroke, and Vascular Research at the University of Arizona Sarver Heart Center (J.B.H.), an AHA predoctoral fellowship 0010162Z (B.R.S.), and NIH/HLBI awards HL63873 (S.K.W.) and HL67067 (J.B.H.).
References
Risau W. Mechanisms of angiogenesis. Nature. 1997; 386: 671–674.
Buschmann I, Schaper W. Arteriogenesis Versus Angiogenesis: Two Mechanisms of Vessel Growth. News Physiol Sci. 1999; 14: 121–125.
Breier G, Damert A, Plate KH, Risau W. Angiogenesis in embryos and ischemic diseases. Thromb Haemost. 1997; 78: 678–683.
Couffinhal T, Silver M, Zheng LP, Kearney M, Witzenbichler B, Isner JM. Mouse model of angiogenesis. Am J Pathol. 1998; 152: 1667–1679.
Thakral KK, Goodson WH, III, Hunt TK. Stimulation of wound blood vessel growth by wound macrophages. J Surg Res. 1979; 26: 430–436.
Carmeliet P, Collen D. Molecular analysis of blood vessel formation and disease. Am J Physiol. 1997; 273: H2091–H2104.
Folkman J, Cotran R. Relation of vascular proliferation to tumor growth. Int Rev Exp Pathol. 1976; 16: 207–248.
Schaper W, Buschmann I. Arteriogenesis, the good and bad of it. Cardiovasc Res. 1999; 43: 835–837.
Gibbons GH, Dzau VJ. The emerging concept of vascular remodeling. N Engl J Med. 1994; 330: 1431–1438.
Pries AR, Reglin B, Secomb TW. Structural adaptation of microvascular networks: functional roles of adaptive responses. Am J Physiol Heart Circ Physiol. 2001; 281: H1015–H1025.
Holder N, Klein R. Eph receptors and ephrins: effectors of morphogenesis. Development. 1999; 126: 2033–2044.
Hoying JB, Boswell CA, Williams SK. Angiogenic potential of microvessel fragments established in three- dimensional collagen gels. In Vitro Cell Dev Biol Anim. 1996; 32: 409–419.
Carter WB, Uy K, Ward MD, Hoying JB. Parathyroid-induced angiogenesis is VEGF-dependent. Surgery. 2000; 128: 458–464.
Essers J, de Stoppelaar JM, Hoebee B. A new rat repetitive DNA family shows preferential localization on chromosome 3, 12 and Y after fluorescence in situ hybridization and contains a subfamily which is Y chromosome specific. Cytogenet Cell Genet. 1995; 69: 246–252.
Groner W, Winkelman JW, Harris AG, Ince C, Bouma GJ, Messmer K, Nadeau RG. Orthogonal polarization spectral imaging: a new method for study of the microcirculation. Nat Med. 1999; 5: 1209–1212.
Hirschi KK, D’Amore PA. Pericytes in the microvasculature. Cardiovasc Res. 1996; 32: 687–698.
Lindahl P, Johansson BR, Leveen P, Betsholtz C. Pericyte loss and microaneurysm formation in PDGF-B-deficient mice. Science. 1997; 277: 242–245.
Orlidge A, D’Amore PA. Inhibition of capillary endothelial cell growth by pericytes and smooth muscle cells. J Cell Biol. 1987; 105: 1455–1462.
Schechner JS, Nath AK, Zheng L, Kluger MS, Hughes CC, Sierra-Honigmann MR, Lorber MI, Tellides G, Kashgarian M, Bothwell AL, Pober JS. In vivo formation of complex microvessels lined by human endothelial cells in an immunodeficient mouse. Proc Natl Acad Sci U S A. 2000; 97: 9191–9196.
Pries AR, Secomb TW. Microcirculatory network structures and models. Ann Biomed Engl. 2000; 28: 916–921.
Jackson ZS, Gotlieb AI, Langille BL. Wall tissue remodeling regulates longitudinal tension in arteries. Circ Res. 2002; 90: 918–925.
Sullivan CJ, Hoying JB. Flow-dependent remodeling in the carotid artery of fibroblast growth factor-2 knockout mice. Arterioscler Thromb Vasc Biol. 2002; 22: 1100–1105.
Lawson ND, Scheer N, Pham VN, Kim CH, Chitnis AB, Campos-Ortega JA, Weinstein BM. Notch signaling is required for arterial-venous differentiation during embryonic vascular development. Development. 2001; 128: 3675–3683.
Nor JE, Peters MC, Christensen JB, Sutorik MM, Linn S, Khan MK, Addison CL, Mooney DJ, Polverini PJ. Engineering and characterization of functional human microvessels in immunodeficient mice. Lab Invest. 2001; 81: 453–463.
Yang J, Nagavarapu U, Relloma K, Sjaastad MD, Moss WC, Passaniti A, Herron GS. Telomerized human microvasculature is functional in vivo. Nat Biotechnol. 2001; 19: 219–224.
Christofidou-Solomidou M, Bridges M, Murphy GF, Albelda SM, DeLisser HM. Expression and function of endothelial cell alpha v integrin receptors in wound-induced human angiogenesis in human skin/SCID mice chimeras. Am J Pathol. 1997; 151: 975–983.
Pilewski JM, Yan HC, Juhasz I, Christofidou-Solomidou M, Williams J, Murphy GF, Albelda SM. Modulation of adhesion molecules by cytokines in vivo using human/severe combined immunodeficient (SCID) mouse chimeras. J Clin Immunol. 1995; 15: 122S–129S.
Nor JE, Christensen J, Mooney DJ, Polverini PJ. Vascular endothelial growth factor (VEGF)-mediated angiogenesis is associated with enhanced endothelial cell survival and induction of Bcl-2 expression. Am J Pathol. 1999; 154: 375–384.
Eberhard A, Kahlert S, Goede V, Hemmerlein B, Plate KH, Augustin HG. Heterogeneity of angiogenesis and blood vessel maturation in human tumors: implications for antiangiogenic tumor therapies. Cancer Res. 2000; 60: 1388–1393.(Benjamin R. Shepherd; Hel)
ABSTRACT
Objective— We have previously demonstrated the ability to construct 3-dimensional microvascular beds in vitro via angiogenesis from isolated, intact, microvessel fragments that retain endothelial cells and perivascular cells. Our objective was to develop and characterize an experimental model of tissue vascularization, based on the implantation of this microvascular construct, which recapitulated angiogenesis, vessel differentiation, and network maturation.
Methods and Results— On implantation in a severe combined-immunodeficient mouse model, vessels in the microvascular constructs rapidly inosculated with the recipient host circulation. Ink perfusion of implants via the left ventricle of the host demonstrated that vessel inosculation begins within the first day after implantation. Evaluation of explanted constructs over the course of 28 days revealed the presence of a mature functional microvascular bed. Using a probe specific for the original microvessel source, 91.7%±11% and 88.6%±19% of the vessels by day 5 and day 28 after implantation, respectively, were derived from the original microvessel isolate. Similar results were obtained when human-derived microvessels were used to build the microvascular construct.
Conclusions— With this model, we reproduce the important aspects of vascularization, angiogenesis, inosculation, and network remodeling. Furthermore, we demonstrate that the model accommodates human-derived vessel fragments, enabling the construction of human-mouse vascular chimeras.
Key Words: vascularization ? microcirculation ? angiogenesis ? human ? vascular remodeling
Introduction
Vascularization is the process by which perfusion pathway length and vessel segment number are increased and organized into a functional vascular bed. In normal situations, this effective increase in vessel density delivers more blood to the tissue, facilitating tissue growth and/or increased tissue activity.1,2 Consequently, vascularization is a primary component of tissue growth and repair, such as occurs during development,3 after an upstream occlusive event leading to tissue ischemia,4 or during proliferative events, as seen in tissue healing5,6 and tumor growth.7 Although we know of many factors and signals that initiate or terminate the vascularization process, little is known about the rules that govern vascularization as an integrated process that includes angiogenesis,3 arteriogenesis,8 vascular remodeling,9 vessel adaptation,10 and arterio-venous polarization.11
We have previously shown that isolated intact microvessel fragments retain angiogenic potential and are capable of forming a simple microvascular bed when cultured in a 3-dimensional collagen I gel.12 In this microvascular construct, the vessel fragments undergo stereotypical angiogenesis, forming neovessels that maintain patent lumen and perivascular cell associations. Furthermore, the vessel fragments within this culture system are responsive to proangiogenic conditions.12,13 All of this occurs in the absence of blood flow and relatively few nonvascular cells.
Here we report the development and characterization of an experimental model of tissue vascularization based on the implantation of this microvascular construct. Precultured or freshly formed microvascular constructs implanted subcutaneously inosculate with the host vasculature and expand into a microvascular bed containing perfused mature heterogeneous vessel elements. With this model, we reproduce the important aspects of vascularization, angiogenesis, inosculation, and network remodeling. Furthermore, we demonstrate that the model accommodates human-derived vessel fragments, enabling the construction of human-mouse vascular chimeras.
Methods
All animal procedures were performed according to University of Arizona IACUC-approved protocols. Mice were anesthetized with an intraperitoneal injection of 2.5% avertin (0.1 mL/10 g) or euthanized by CO2 asphyxiation. Rats were euthanized with an overdose of pentobarbital (150 mg/kg).
Microvascular Constructs
Rat fat microvessel fragments (RFMFs) were isolated from epididymal fat pads of retired breeder male Sprague-Dawley rats. Under aseptic conditions, harvested fat pads were finely minced with scissors, digested in 2 mg/mL collagenase plus 2 mg/mL EFAF-BSA in phosphate-buffered saline (PBS) for 8 minutes at 37°C with vigorous shaking and washed in 0.1% BSA-PBS. Tissue debris and large vessel pieces were removed by filtering the suspension through a sterile 500-μm-pore nylon screen. Microvessel fragments were captured by filtration of the remaining suspension on a 30-μm-pore nylon screen and recovered by vigorous flushing of the screen surface with 0.1% BSA-PBS. The type and lot number of collagenase used were predetermined to optimize fragment yield while maintaining microvessel integrity. Microvessel fragments (MF) were suspended (12 000 to 15 000 MF/mL) in ice-cold 3 mg/mL rat tail type I collagen (BD BioSciences, Bedford, Mass) prepared with Dulbecco minimum essential medium (DMEM) (1x final) and pH neutralized with 1 mol/L NaOH. MF/collagen suspensions were plated into individual wells (0.25 mL/well) of a 48-well plate and placed in a 37°C incubator for 20 minutes to polymerize the collagen. For culturing, an equal volume of 10% FBS/DMEM was added to each well and fed at day 4. No growth factors or supplements were added to the culture medium. Constructs were cultured for 5 to 7 days before implantation. For immediate implantation, newly formed constructs were transferred directly to the implant site without preculturing. Human fat microvessel fragments were isolated from discarded abdominoplasty adipose and cast into collagen gels using the same procedures as for rat fat microvessel fragments but were not cultured. In all cases, avascular control constructs were prepared from the same batches of collagen solution used to form microvascular constructs and at the same time.
Implantation
Microvascular constructs and avascular collagen gel controls were implanted in the subcutaneous position on the flanks of anesthetized female severe combined-immunodeficient (SCID) mice. For implantation, a subcutaneous pocket was formed between the skin and the underlying muscle anterior to the pelvis using blunt dissection through a small skin incision. Each host mouse received 2 implants: a microvascular construct on one side and an avascular control collagen gel on the contralateral side. The incision was then closed with 6-0 suture and the animal was allowed to recovered.
Histology and Histochemistry
Mice were anesthetized as mentioned and implants were removed, with a small portion of the underlying muscle fixed in 2% paraformaldehyde/PBS and processed into paraffin. General histology was performed on deparaffinized, 5- to 6-μm-thick sections (for details, see the data supplement available at http://atvb.ahajournals.org). Vessel density was determined by counting discreet GS-1–positive structures in the implanted construct from at least 5 different fields (of defined area) per section from 2 different implants. Individual counts were divided by the area of each field and averaged for each time point. For details of en bloc immunostaining, see the data supplement available at http://atvb.ahajournals.org.
Ink Perfusion
Mice containing implants were anesthetized with Avertin and placed supine on a dissecting stage. Through a left ventricular catheter (PE 60 tubing), blood was cleared with heparinized saline containing 10 μmol/L sodium nitroprusside followed by perfusion with India ink (Speedball 3398; Hunt Manufacturing, Statesville, NC), dialyzed against PBS, and filtered through #1 Whatman paper. All perfusions were performed at a maintained pressure of 90 to 100 mm Hg until all tissues in the mouse appeared dark (usually, this required 2 to 3 mL of ink solution). After ink perfusion, the implants were excised, fixed in 4% paraformaldehyde in PBS for 45 minutes at 4°C, and cleared in 100% glycerol for 20 minutes. Implants were bisected, and the 2 halves, cut-side up, were sandwiched between a microscope slide and coverslip for viewing with a standard light microscope.
In Situ Hybridization
A probe that detects a repeat element on the Y chromosome in rat14 was synthesized from rat-tail genomic DNA by PCR (forward: ggt tct aga ctg taa aac cca gac; reverse: act taa aac taa gct tat tgg cca), size-verified, and labeled with biotin using the Photoprobe labeling kit (Vector Laboratories, Burlingame, Calif). Labeled probe was hybridized to 8-μm sections using standard methods (for details, see http://.atvb.ahajournals.org). For determination of the percent Y chromosome-positive cells, serial sections (6-μm-thick) were stained by in situ hybridization for the Y chromosome or with hematoxylin to label all nuclei. The ratios of Y chromosome-positive counts to hematoxylin-positive counts (both derived from serial sections) from at least 5 areas of a section from 2 different implants were averaged.
Smooth Muscle Cell Coverage
All vascular cells and smooth muscle cells (SMCs) were colabeled in fixed intact microvascular constructs (cultured or implanted) by en bloc immunostaining (for details, see http://.atvb.ahajournals.org). The immunostained constructs were imaged using a 20x objective on a Biorad MRC-1024ES Confocal Microscope. Stacks of 512x512-μm images with 1.05x1.05-μm pixel resolution were generated with an image separation of 1.0 μm for each fluorescence channel. For determination of percent smooth muscle coverage, up to 4 stacks of images were obtained from constructs and the fractional area comprising total vascular cells (MHC-positive structures) and SMCs (alpha-actin-positive structures) were computed (for details, see http://atvb.ahajournals.org). MHC-positive structures with a total volume less than that of a 10-μm-long and 5 μm in diameter rod were removed from the image stacks, as were SMC components not associated with vascular components. Percent coverage of SMC for each stack was calculated as (SMC volume/vascular volume)x100.
Results
Isolated fat microvessel fragments undergo angiogenic sprouting within the first 5 days of culture in collagen I gels (see http://atvb.ahajournals.org). Sprouting occurs at the ends and from mid regions of individual fragments. Furthermore, the sprouting process is dynamic, with individual sprouts forming and regressing within a day. By 11 days in culture, fragments have grown to form a collection of elongated, simple neovessels (see http://atvb.ahajournals.org) with an average diameter of 24.8±6.8 μm (n=39). These neovessels contain patent lumen and a relatively low density of -actin-positive perivascular cells.12
Coordinated with angiogenesis, newly formed vessel elements must locate and inosculate with other elements of the vascular network before functioning as a blood perfusion circuit. We evaluated the ability of this microvascular construct to interact with an existing vasculature by implanting neovessel cultures into SCID mice. After 5 to 7 days of culture, constructs were placed under the skin, in direct contact with the dorsal musculature, and analyzed at various time points (days 1, 3, 5, 7, 10, 14, 21, and 28 after implantation). All microvascular constructs (4 or more per time point) contained vessels after implantation. Superficial, blood-filled vessels were associated with microvascular constructs but not with collagen gel controls (see http://atvb.ahajournals.org). Furthermore, the implanted microvascular constructs contained the full range of vessel types commonly seen in a mature functional vascular bed, including small arteries, arterioles, capillaries, venules, and veins (Figure 1a through 1d).
Figure 1. Vessels within microvascular implants are heterogeneous in structure and contain red blood cells. H&E stained sections of characteristic implanted microvascular constructs after 14 (a, b, d) or 10 (c) days of implantation in a subcutaneous pocket reveal the presence of arteries (A), arterioles (arrowhead in d), capillaries (arrows), venules (arrowhead in b), veins (V), and red blood cells within vessels.
We perfused ink into the left ventricle of the host mouse to determine when vessels within the construct become contiguous with the host mouse vasculature. By day 1, a limited number of microvessels in the implant contained ink indicating continuity with the mouse circulation (Figure 2a). By day 2, the construct vessels were beginning to assemble together, continuing to form interconnections by day 3 (Figure 2b and 2c). By 28 days after implantation, the vessels have refined into a mature-appearing vascular bed (Figure 2d). Orthogonal polarized spectral imaging, which selectively detects hemoglobin,15 of living implants reveals blood-containing vessel structures throughout the constructs by day 14 (Figure 2e). Only surface blood, because of the dissection to expose the construct was detected with orthogonal polarized spectral on control, vessel-free collagen gel implants (Figure 2f).
Figure 2. Vessels within microvascular constructs inosculate with the host vasculature and are perfusion-competent when implanted subcutaneously into SCID mice. a-d, Representative fields of microvascular constructs harvested on day 1 (a), day 2 (b), day 3 (c), or day 28 (d) after implantation after perfusion of the respective host mouse through the left ventricle with ink. e and f, Orthogonal polarized spectroscopic images of microvascular implants after 14 days of implantation (e) and at day 3 with avascular control gel (f). A 10-times magnification probe was used, which provided an imaging depth of 700 μm.
Staining of implant sections with the Griffonia simplicifolia-I lectin verified the presence of endothelial cell-comprised vessels (Figure 3a). Vessel densities (GS-1–positive structures) within the microvascular constructs increased >2-fold from day 5 to day 28 (Figure 3b), suggesting that angiogenesis continued within the implants during this time. Immunostaining of implants for -SMC actin, to corroborate the mature vessel phenotype observed in the histology, identified positive cells surrounding vessels of different caliber (see http://atvb.ahajournals.org). To determine whether the vessels within the implanted microvascular constructs were derived from the original cultured rat microvessels or, alternatively, were replaced by the ingrowth of host mouse vessels, we took advantage of the fact that the rat microvessels were isolated from males and the constructs were placed into female mice. Using a DNA probe specific to a repeat element of the rat Y chromosome,14 we analyzed sections from day 5 and day 28 implants by in situ hybridization. Vessel-like structures and single cells throughout the implant were detected with the Y chromosome-specific probe, whereas Y-positive cells were not observed within the adjacent muscle of the host mouse (Figure 3c). To determine the percentage of cells within the microvascular constructs that were derived from the original isolate, serial sections from day 5 and day 28 implants were stained with either hematoxylin, to identify all nuclei, or the Y-chromosome probe (please see http://atvb.ahajournals.org). Comparing the number of Y-positive cells to the total number of cells in serial sections indicates that nearly all of the cells (90%) within the implants, even after 4 weeks, comprised cells derived from the original rat microvessel isolate (Figure 3d).
Figure 3. Microvascular implants contain vascular endothelium and smooth muscle-positive perivascular cells that are nearly derived from the original cultured isolates. a, Day 5 implant stained with the rodent endothelial cell-specific lectin, GS-1, showing vessels of various caliber. b, Vessel density measurements determined from GS-1–stained sections of implanted microvascular constructs. c, Day 28 implant section labeled by in situ hybridization for a rat Y-chromosome-specific repeat. The "m" depicts the underlying host (mouse) musculature, while the "i" marks the implant. The arrow indicates the boundary between the 2 tissues. d, The fraction of Y-chromosome-positive cells in implanted constructs (see Methods) for day 5 and day 28 implants.
During angiogenesis and vessel maturation, perivascular cells dissociate and reassociate with vessel elements, respectively.16 Using confocal microscopy of en bloc immunostained constructs and volumetric reconstruction algorithms, we measured the percent coverage of vessel elements by perivascular cells (identified as -actin-positive cells) during the early and late stages of vascular activity in the constructs. Perivascular coverage during the culture period decreased from 83% in the freshly isolated fragments to nearly 50% coverage after 8 days in culture (Figure 4). After implantation and inosculation with the host vasculature, perivascular cell investment of construct vessels returned to normal levels (82%). Qualitatively, the vessel elements with the least coverage were predominately the new vessel sprouts and elongating neovessels. This reduction in coverage did not occur until after 3 days in culture. Typically, vessel sprouts from fragments do not form until 4 or 5 days in culture.12 During the active vessel growth phases, -actin-positive cells were often observed in extravascular spaces of the construct (data not shown).
Figure 4. Smooth muscle coverage of vessel elements changes during vascularization. Confocal image stacks of double-stained vessel elements (green=MHC and red=-actin) were used to generate vessel and perivascular volumes. Single-image planes of (a) a freshly isolated microvessel fragment (day 0), (b) a day 8 cultured fragment, and (c) day 28 postimplantation branched vessel element. d, Percentage of vessel elements covered by -actin-positive perivascular cells measured at different time points in the model. Coverage was assessed during culturing of microvascular constructs (day 0c-day 10c) and during implantation of constructs precultured for 8 days before implantation (day14i-day 28i).
To determine if preculturing is necessary for the microvascular constructs to be able to integrate with the host circulation, we implanted freshly formed microvascular constructs without preculturing. As before, microvessel fragments were isolated and embedded in collagen gels. However, these freshly formed constructs were immediately implanted without previous culturing. Similar to precultured implanted constructs, the noncultured constructs formed a recognizable vascular tree that was contiguous with the host circulation (Figure 5a and 5b). This prompted us to explore the possibility that human-derived vessel fragments could be used to build functional microvascular constructs. Unlike with rat-derived microvessels, human-derived microvessel fragments do not undergo angiogenesis in the collagen gels in vitro and instead either remain quiescent or dissociate into single cells (unpublished observation). As with noncultured rat-derived constructs, the freshly formed human microvascular constructs contained heterogenous vessel structures 15 days after implantation that were positive for the differentiation marker, von Willebrand factor (Figure 5c), and possessed -actin-positive cells (Figure 5d). Similar to what we observed with rat microvessels, constructs established from human sources contained primarily human (UEA-I-positive) vessels after 15 days of implantation (Figure 5e) with limited mouse (GS-1 positive) vessel invasion (Figure 5f), indicating that the original human fragments used to establish the construct were maintained in the implants.
Figure 5. Microvascular constructs assembled from rat-derived or human-derived freshly isolated microvessel fragments (no preculturing, see Methods) also form a vascular bed on implantation. a, En bloc fluorescence immunostaining for all cells (anti-MHC antibody) reveals that vessels within the construct form tree-like structures. b, Ink perfusion of the host and a superficial vessel connection (arrowheads) between host and construct (arrows indicate host muscle/construct boundary). Immunostaining of human-derived, day 15 implants for von Willebrand factor (c) or -actin (d) indicates that vessels (arrows) are differentiated and mature. Representative sections from the same human-derived implant stained with the human-specific lectin, UEA1 labels only vessels (arrows in e) within the construct, whereas the rodent-specific vascular marker GS-1 labels vessels within only the surrounding host mouse tissue (arrows in f). The dashed lines indicate the boundary between the implant and underlying host tissue.
Discussion
In the present study, we examined the capability of a cultured microvascular construct to interface with an existing vasculature and subsequently remodel into a functional perfusion circuit. Our goal was to develop an experimental model enabling the detailed study of the progression from angiogenesis through vessel differentiation to network maturation. The microvascular constructs consisted of a network of new microvessels in a collagen I gel that formed, via angiogenesis, from isolated microvessel fragments after in vitro culturing. In culture, the isolated microvessel fragments, which comprise the full spectrum of microvascular elements,12 form a collection of morphologically homogenous, simple vessels with low perivascular cell coverage. Once implanted, this simple vasculature inosculates with the host circulation and remodels to contain morphologically identifiable arteries, arterioles, capillaries, venules, and veins assembled into a perfused, vascular bed. Therefore, each vessel in this model (ie, a freshly isolated microvessel fragment) begins as a differentiated vessel element, dedifferentiates into a "generic-appearing" vessel during/after angiogenesis and subsequently redifferentiates into a specific element of a perfusion-competent vascular tree. In this regard, the entire vascularization process is recapitulated in this model including true angiogenesis (sprouting from preexisting parent vessels1), vessel differentiation, and network maturation.
A typical microvessel consists of endothelial cells arranged into a tube wrapped by one or more layers of perivascular cells (usually SMCs or pericytes). The interactions between the endothelial and perivascular cell layers are thought important in maintaining microvascular integrity and stability.17 During culturing of the microvessel fragments, -actin-positive perivascular cell coverage of vessels decreased to nearly half of that observed in isolated vessel fragments. This loss in coverage involved predominately vessel sprouts and growing neovessels. This is consistent with observations that perivascular cell withdrawal from vessel segments relaxes control of the endothelial cell tube and permits sprouting and vessel elongation during angiogenesis.18 After implantation, perivascular cell coverage increased to a level comparable to that observed in the microvessel fragments isolated from the mature adipose microvasculature indicating that vessel differentiation and maturation was being reestablished in the constructs. This is similar to a related study involving the implantation of 3-dimensional cultures of endothelial cells.19 In this study, endothelial cell tubes acquired perivascular cells from the host at a time coincident with vessel maturation (the presence of blood cells and loss of VCAM expression). In our study, perivascular cells were present in the microvascular constructs at the time of implantation (45% perivascular cell coverage). The continual presence of perivascular cells may make the microvessels within the microvascular construct capable of rapid maturation and could explain why they were able to rapidly establish an integrated perfusion-competent vasculature.
Although not specifically examined in this study, inosculation with the host circulation appears to be restricted to the construct boundaries. Few, if any, host vessels were observed in the construct, and construct vessels that were visibly contiguous with host vessels were superficially located on the construct. Whether inosculation occurred via end-end or end-side anastomoses between host and construct vessels is unclear. The formation of end-end connections would suggest an active involvement of the host circulation, perhaps via a coincidental angiogenesis response that causes vessel ends to grow toward each other. The presence of continuous vessels originating in the host tissue and spanning across the construct boundary to extend branches into the construct suggests that a host vascular response was involved. However, further studies with the constructs specifically directed at addressing these questions are required to determine the true mechanism by which vessel inosculation occurs.
It is important to point out that angiogenesis in the model occurs in the relative absence of nonvascular cells. Angiogenesis is often considered a vascular response to increased tissue needs and is associated with release of angiogenic factors by the surrounding tissue stroma.3 The occurrence of angiogenesis in our model with a minimal presence of stromal cells implies that vessels have an endogenous capability for new vessel formation. This is supported by the finding that the microvessel fragments in culture express Vegf A mRNA (data not shown). Interestingly, the neovessels resulting from angiogenesis in the constructs are homogeneous in character (24 μm average diameter and uniform low-level perivascular cell coverage). It is tempting to speculate that, perhaps, in the absence of stroma, this represents a default or baseline phenotype of the angiogenic neovessel. However, it is possible that these characteristics may be unique to the specific collagen gel environment used in the model or a consequence of culturing. However, culturing of the microvessel fragments was not required for the establishment of a functional vascular bed within the constructs after implantation. Freshly implanted constructs that had not been cultured still formed a perfusion-competent differentiated vascular bed. Therefore, no activities unique to culturing are essential for vascularization.
The appearance of flow throughout the entire construct precedes the presence of differentiated heterogenous vessel types. Hemodynamic stimuli to a vessel, namely shear stress and circumferential and axial wall stress, affects vessel diameter and wall thickness and possibly controls arteriole specification.20–22 Thus, it is likely that flow (shear stress) and pressure (wall stress) are contributing to vessel differentiation in the implanted constructs. Whether blood flow is the only stimulus establishing a completely mature network architecture is not clear. Observations in developing zebra fish embryos indicate that arteries and veins are specified before blood flow begins.23 Whether this is unique to the embryo or whether it may also occur in the adult remains to be determined. Our results do not exclude the intriguing possibility that formation of artery-side and vein-side vessel types may reflect some aspect of "memory" in which the original vessel types present within the microvessel fragment isolate retain some predetermination. This experimental model, with the particular inclusion of the in vitro culturing, should prove invaluable in addressing this question and those generally related to artery/vein specification and differentiation after angiogenesis in the adult.
Experiments with human-derived microvessel fragments demonstrate the potential usefulness of this model in constructing an experimental human microcirculation without the need to modify or engineer the human cells. Previous models examining the human microcirculation in vivo have involved implantation of cultured human microvessel endothelial cells24,25 or the transplantation of full-thickness skin grafts.26,27 However in implants using cultured endothelial cells, it was necessary to prolong endothelial cell survival by overexpressing either telomerase25 or bcl-228 through stable transfections of the endothelial cells. We present an alternate model of the human microcirculation involving the use of isolated human microvessel fragments and subsequent implantation into a SCID mouse. This human/mousechimera model, which is easier to perform than the full-thickness skin grafting and does not require additional manipulation to preserve vessel survival, produces an intact differentiated human microcirculation in vivo. The resulting human/mouse vascular chimeras will prove useful in further studies exploring the mechanisms and processes of angiogenesis and vascular integration in the human microcirculation, and will explore microvessel/tissue cell interactions such as capillary exchange, metastasis, and inflammation. Furthermore, the human-derived constructs represent an alternative strategy to vascularizing tissue-engineered constructs by producing a ready-made vasculature comprising endothelial cells and perivascular cells. The rapid initial perfusion and subsequent remodeling obtained with the microvascular construct, once implanted, would provide early support of stromal cells incorporated into an engineered tissue, earlier than observed with other strategies, and establish a stable perfusion bed.
In summary, we present a model of vascularization based on the formation of a microvascular construct grown from intact microvessel elements. Two key features of this experimental model of vascularization are: (1) the vessels within the construct progress through stereotypical stages of vascularization and (2) it enables examination of vascularization in the absence of confounding tissues. In vascularizing tissues, in can be difficult to capture only those vessels within the tissue that are undergoing angiogenesis or network remodeling for subsequent analysis. This is caused, in part, by the regional heterogeneity in vascularization stimuli present throughout many actively remodeling or repairing tissues, resulting in vessels of the tissue being at different stages of vascularization at the same time.4,29 In contrast to repairing tissues, our vascularization model is fairly well synchronized. Microvessel fragments are harvested from relatively stationary microvascular beds and cast into the collagen gels at the same time. Consequently, they begin the vascularization process at the same time and synchronously progress through stereotypical phases of vascularization. Culturing of the microvessels recapitulates only the early angiogenesis phase of vascularization, enabling a detailed study of angiogenic sprouting and other angiogenesis-specific activities. Furthermore, the ability to culture the microvessel fragments without affecting their ability to inosculate and carry blood extends the usefulness of the model to include gene delivery and drug treatment experimental approaches. An important aspect of the model is the relative absence of host cells in the construct and the persistence of the original vessels when implanted. This aspect of the microvascular construct simplifies interpretation of experimental results by limiting relevance to vascular cells. However, nonvascular tissue cells could be included with the microvessels in the construct, which would permit study of tissue/vessel interactions. The microvascular construct implanted into the SCID mouse serves as a complete model of vascularization and enables evaluation of the molecular, cellular, and physiological mechanisms mediating the active remodeling of a vasculature.
Acknowledgments
This research was supported in part by an Arizona Elks Major Projects Award (J.B.H.), The Mary Lou Hemler Arnold Award for Heart Disease, Stroke, and Vascular Research at the University of Arizona Sarver Heart Center (J.B.H.), an AHA predoctoral fellowship 0010162Z (B.R.S.), and NIH/HLBI awards HL63873 (S.K.W.) and HL67067 (J.B.H.).
References
Risau W. Mechanisms of angiogenesis. Nature. 1997; 386: 671–674.
Buschmann I, Schaper W. Arteriogenesis Versus Angiogenesis: Two Mechanisms of Vessel Growth. News Physiol Sci. 1999; 14: 121–125.
Breier G, Damert A, Plate KH, Risau W. Angiogenesis in embryos and ischemic diseases. Thromb Haemost. 1997; 78: 678–683.
Couffinhal T, Silver M, Zheng LP, Kearney M, Witzenbichler B, Isner JM. Mouse model of angiogenesis. Am J Pathol. 1998; 152: 1667–1679.
Thakral KK, Goodson WH, III, Hunt TK. Stimulation of wound blood vessel growth by wound macrophages. J Surg Res. 1979; 26: 430–436.
Carmeliet P, Collen D. Molecular analysis of blood vessel formation and disease. Am J Physiol. 1997; 273: H2091–H2104.
Folkman J, Cotran R. Relation of vascular proliferation to tumor growth. Int Rev Exp Pathol. 1976; 16: 207–248.
Schaper W, Buschmann I. Arteriogenesis, the good and bad of it. Cardiovasc Res. 1999; 43: 835–837.
Gibbons GH, Dzau VJ. The emerging concept of vascular remodeling. N Engl J Med. 1994; 330: 1431–1438.
Pries AR, Reglin B, Secomb TW. Structural adaptation of microvascular networks: functional roles of adaptive responses. Am J Physiol Heart Circ Physiol. 2001; 281: H1015–H1025.
Holder N, Klein R. Eph receptors and ephrins: effectors of morphogenesis. Development. 1999; 126: 2033–2044.
Hoying JB, Boswell CA, Williams SK. Angiogenic potential of microvessel fragments established in three- dimensional collagen gels. In Vitro Cell Dev Biol Anim. 1996; 32: 409–419.
Carter WB, Uy K, Ward MD, Hoying JB. Parathyroid-induced angiogenesis is VEGF-dependent. Surgery. 2000; 128: 458–464.
Essers J, de Stoppelaar JM, Hoebee B. A new rat repetitive DNA family shows preferential localization on chromosome 3, 12 and Y after fluorescence in situ hybridization and contains a subfamily which is Y chromosome specific. Cytogenet Cell Genet. 1995; 69: 246–252.
Groner W, Winkelman JW, Harris AG, Ince C, Bouma GJ, Messmer K, Nadeau RG. Orthogonal polarization spectral imaging: a new method for study of the microcirculation. Nat Med. 1999; 5: 1209–1212.
Hirschi KK, D’Amore PA. Pericytes in the microvasculature. Cardiovasc Res. 1996; 32: 687–698.
Lindahl P, Johansson BR, Leveen P, Betsholtz C. Pericyte loss and microaneurysm formation in PDGF-B-deficient mice. Science. 1997; 277: 242–245.
Orlidge A, D’Amore PA. Inhibition of capillary endothelial cell growth by pericytes and smooth muscle cells. J Cell Biol. 1987; 105: 1455–1462.
Schechner JS, Nath AK, Zheng L, Kluger MS, Hughes CC, Sierra-Honigmann MR, Lorber MI, Tellides G, Kashgarian M, Bothwell AL, Pober JS. In vivo formation of complex microvessels lined by human endothelial cells in an immunodeficient mouse. Proc Natl Acad Sci U S A. 2000; 97: 9191–9196.
Pries AR, Secomb TW. Microcirculatory network structures and models. Ann Biomed Engl. 2000; 28: 916–921.
Jackson ZS, Gotlieb AI, Langille BL. Wall tissue remodeling regulates longitudinal tension in arteries. Circ Res. 2002; 90: 918–925.
Sullivan CJ, Hoying JB. Flow-dependent remodeling in the carotid artery of fibroblast growth factor-2 knockout mice. Arterioscler Thromb Vasc Biol. 2002; 22: 1100–1105.
Lawson ND, Scheer N, Pham VN, Kim CH, Chitnis AB, Campos-Ortega JA, Weinstein BM. Notch signaling is required for arterial-venous differentiation during embryonic vascular development. Development. 2001; 128: 3675–3683.
Nor JE, Peters MC, Christensen JB, Sutorik MM, Linn S, Khan MK, Addison CL, Mooney DJ, Polverini PJ. Engineering and characterization of functional human microvessels in immunodeficient mice. Lab Invest. 2001; 81: 453–463.
Yang J, Nagavarapu U, Relloma K, Sjaastad MD, Moss WC, Passaniti A, Herron GS. Telomerized human microvasculature is functional in vivo. Nat Biotechnol. 2001; 19: 219–224.
Christofidou-Solomidou M, Bridges M, Murphy GF, Albelda SM, DeLisser HM. Expression and function of endothelial cell alpha v integrin receptors in wound-induced human angiogenesis in human skin/SCID mice chimeras. Am J Pathol. 1997; 151: 975–983.
Pilewski JM, Yan HC, Juhasz I, Christofidou-Solomidou M, Williams J, Murphy GF, Albelda SM. Modulation of adhesion molecules by cytokines in vivo using human/severe combined immunodeficient (SCID) mouse chimeras. J Clin Immunol. 1995; 15: 122S–129S.
Nor JE, Christensen J, Mooney DJ, Polverini PJ. Vascular endothelial growth factor (VEGF)-mediated angiogenesis is associated with enhanced endothelial cell survival and induction of Bcl-2 expression. Am J Pathol. 1999; 154: 375–384.
Eberhard A, Kahlert S, Goede V, Hemmerlein B, Plate KH, Augustin HG. Heterogeneity of angiogenesis and blood vessel maturation in human tumors: implications for antiangiogenic tumor therapies. Cancer Res. 2000; 60: 1388–1393.(Benjamin R. Shepherd; Hel)