Original Article

TISSUE ENGINEERING: Part A Volume 00, Number 00, 2014 ª Mary Ann Liebert, Inc. DOI: 10.1089/ten.tea.2014.0315

Tuning Graft- and Host-Derived Vascularization in Modular Tissue Constructs: A Potential Role of HIF1 Activation Gabrielle C. Lam, MASc,1 and Michael V. Sefton, ScD1,2

A better understanding of the factors governing the vascularization of engineered tissues is crucial for their advancement as therapeutic platforms. Here, we studied the effect of implant volume and cell densities on the in vivo vascularization of modular engineered tissue constructs. Sub-millimeter collagen modules containing adipose-derived mesenchymal stromal cells (adMSC) and enveloped by human umbilical vein endothelial cells (HUVEC) were subcutaneously implanted in severe-combined immunodeficient mice with a beige-mutation (SCID-bg) mice. Implant volume and cell density was varied relative to a base case, defined as a 0.01 mL implant containing 1.5 · 107 adMSC/mL and 3.9 · 106 HUVEC/mL. At 7 and 14 days post-transplantation, the constructs were harvested for immunohistochemical analysis of total (CD31 + ) and graft-derived (UEA1 + ) vessel formation, hypoxia-inducible factor 1-alpha (HIF1a) expression, infiltration of host-derived leukocytes (CD45), and macrophages (F4/80). Implant volume and cell density affected the relative contributions of hostversus graft-derived vascularization, highlighting that different mechanisms underlie the two processes. Graftderived vessel formation was most rapid and robust in implants with high HIF1a expression, namely large volume implants and implants with high adMSC and HUVEC density ( p < 0.01 compared to base case at day 7). Many HIF1a + cells were vessel-lining HUVEC, suggesting that HIF1 activation may be key to vessel assembly in the graft. Host vessel ingrowth, however, dominated the vascularization of small volume implants (of high and low adMSC density alike), which showed low HIF1a expression at day 7. Host vessels were sustained to day 14 when adMSC density alone was increased, presumably due to increased paracrine secretions. This study points to a potential role of HIF1 activation in the vascularization of tissue constructs, which may be harnessed to engineer robust vessels for therapeutic applications.

vascularization of modTE constructs. It is expected that these parameters could be harnessed to advance the use of modTE for therapeutic applications. Already, others have shown that cell proportions influence vascularization of tissue systems containing different mesenchymal stromal cell (MSC) and endothelial cell (EC) populations.8–10 Cell growth also varied spatially and with cell density in engineered tissues,11–13 presumably due to corresponding differences in local oxygen tensions.14,15 While prolonged hypoxia can have detrimental effects on cells, such as EC,16,17 it is also an important driver of angiogenesis. Hypoxia-inducible factor 1-alpha (HIF1a) is not degraded in the absence of oxygen, enabling translocation to the nucleus and exerting many of its effects as a transcription factor.18 As a master regulator of cellular response to stresses like hypoxia, HIF1a affects the transcription of numerous genes related to angiogenesis, survival, proliferation, and glucose

Introduction

V

ascularization of engineered tissue is an important challenge. Because oxygen diffusion is limited to *100 mm,1 an integral vasculature is crucial to fabricate tissues of clinically relevant size, cell loads, and therapeutic potential. A wide variety of angiogenic methods have been explored, including cell and growth factor delivery, with and without tissue scaffolds, as reviewed elsewhere.2 Modular tissue engineering (modTE) is a method of constructing tissue with integral vasculature,3 based on bottom-up assembly of cylindrical collagen pieces (modules), each enveloped in an endothelial layer. Due to its modular format, the technique is scalable, requires minimally invasive delivery, and can be used with various cell combinations.4–7 However, there currently lacks an understanding of the parameters, such as hypoxia, that affect

1 Institute of Biomaterials and Biomedical Engineering and 2Department of Chemical Engineering and Applied Chemistry, University of Toronto, Toronto, Ontario, Canada.

1

2

LAM AND SEFTON

metabolism.19–21 In hypoxia, HIF1 mediates EC survival and tube formation,22,23 augments MSC-secreted paracrine factors,24–26 and myeloid cell recruitment to ischemic sites.16,17,27–31 Together, these cells orchestrate an angiogenic response to reestablish oxygenation within the ischemic tissue. The objective of this study was to assess the effect of implant volume and cell density on the vascularization of modular tissues containing adipose-derived MSC (adMSC) and EC. It was hypothesized that manipulating these parameters affects the rate and degree of vascularization, via changing cellular response to hypoxic stress (as measured by HIF1a immunostaining) in vivo. Parameters were manipulated relative to a base case, which contained the same adMSC (1.5 · 107 cells/mL implant) and human umbilical vein endothelial cell (HUVEC) density (3.9 · 106 cells/mL implant) as the standard modular implant described before5, albeit with one-tenth of the standard implant volume (0.01 vs. 0.10 mL). Materials and Methods Cell culture

HUVECs and human adMSC were obtained commercially (Lonza). HUVECs were cultured in Endothelial Basal Medium (Lonza), supplemented with the Endothelial Growth Media Bullet-kit, while adMSC were grown in Dulbecco’s Modified Eagle’s Medium (Sigma-Aldrich), supplemented with 10% volume fetal bovine serum and 1% volume penicillin-streptomycin (Gibco). Medium was changed every 2–3 days, and cells passaged at 80% confluence. For all experiments detailed herein, HUVECs were used between passage 3–6, and adMSC between passage 4–5. Module fabrication

Modules embedded with adMSC (1.5 · 106 or 1.5 · 107 adMSC/mL of collagen solution) and enveloped by HUVEC were fabricated, as previously described.3 Briefly, a collagen solution comprised of purified bovine type 1 collagen (Advanced Biomatrix), 10-times concentrated alpha minimum essential media (Gibco; 128 mL/mL collagen), and 0.8 M sodium bicarbonate was prepared. adMSC were suspended in the collagen solution, filled within a gas-sterilized polyethylene tubing (Instech Laboratories; 760 mm or 380 mm diameter), and incubated at 37C for 1 h. After gelation, a mechanical cutter was used to section the PE tubing into pieces 2 mm long. Modules were released by gentle vortex, and cultured overnight in adMSC medium. The following day, HUVEC (105

cells/mL) and modules were suspended in equal volumes of HUVEC and adMSC media (10 mL total), and placed on a rocker at 37C for 1 h, after which a nonconfluent HUVEC layer was produced on the module surface. Modules were statically cultured overnight prior to implantation; during this time the HUVEC and adMSC shrink the modules to final diameters of 400 mm (from 760 mm) or 200 mm (from 380 mm). Module shrinkage, due to cell traction, resulted in an increase in adMSC density by a factor of *10 and formation of a confluent HUVEC layer on the module surface. Module implantation

Modules were transferred to a 1 mL centrifuge tube and rinsed thrice with phosphate-buffered solution (PBS; Gibco). Subcutaneous injection of modules was performed using sterile techniques at the University of Toronto, with all procedures approved by and in accordance with ethical guidelines of the Faculty of Medicine Animal Care Committee as before.5 Severe-combined immunodeficient mice with a beige-mutation (SCID-bg) were obtained commercially (Charles River) and operated upon at 8–10 weeks of age. Mice were anesthetized by isofluorane inhalation, and hair on the dorsum was waxed. A subcutaneous pocket was created using an 18-gauge catheter, into which modules— preloaded within PE tubing in 200 mL PBS—were delivered. Animals were housed in sterile conditions and provided irradiated food and water. Subcutaneous implants were harvested 7 and 14 days after implantation (n = 5). Description of experiments and experimental parameters

The implant volume, adMSC density, and/or HUVEC density were varied within modular implants (summarized in Table 1) to explore their effects on module-induced vascularization. A small implant volume (0.01 mL) and standard adMSC density (1.5 · 107 cells/mL) with confluent HUVEC (3.9 · 106 cells/mL of implant modules; 400 mm diameter module) was chosen as the base case. adMSC and HUVEC density were the same as that used in an earlier report,5 albeit using one-tenth the implant volume. Because HUVEC adhere to the surface of modules, HUVEC density was varied by changing module diameter while maintaining implant volume constant. That is, the number of HUVEC scales with the total surface area of modules within the implant volume. Since there are more small-diameter modules in a given implant volume, the density and total number of HUVEC in an implant was

Table 1. Description of Implant Volume, adMSC and HUVEC Densities (Per mL of Modular Implant) and Number in Experimental Groups

Experimental group Base Group 1 (higher volume) Group 2 (higher adMSC density) Group 3 (higher adMSC/HUVEC density)

Implant volume (mL) 0.01 0.10 0.01 0.01

adMSC number (106 cells) 0.15 1.50 1.50 1.50

(79%) (79%) (97%) (94%)

adMSC density (106 cells/mL) 15.0 15.0 150 150

HUVEC number (106 cells) 0.039 0.39 0.039 0.098

(21%) (21%) (3%) (6%)

HUVEC density (106 cells/mL) 3.9 3.9 3.9 9.8

Percentage in cell number column is percentage of all cells that are HUVEC or adMSC, respectively. Cell density refers to the number of cells per mL of implant. adMSC, adipose-derived mesenchymal stromal cell; HUVEC, human umbilical vein endothelial cell.

MODULAR TISSUE VASCULARIZATION HIF1 ACTIVATION

3

changed by manipulating module diameter. Calculations of HUVEC density and total number are provided in Supplementary Data (Supplementary Data are available online at www.liebertpub.com/tea).

Scope XT (Aperio Technologies) in the Advanced Optical Microscopy Facility at Princess Margaret Hospital, Toronto. Digital images were processed and analyzed using ImageScope software.

Experimental series 1: varying implant volume. Implant volume was increased 10-fold compared to the base case, such that the implant volume of Group 1 was 0.10 mL. Because adMSC and HUVEC density were maintained the same as the base case, increasing implant volume also increased adMSC and HUVEC number. Module diameter was the same in both groups (400 mm after contraction), such that HUVEC density was the same in both the base case and in Group 1.

Vessel quantification. The implant area was defined as the collagen module bed and its surrounding fibrotic tissue (aligned layer of host-derived cells). The region of interest (ROI) was traced, and the area automatically computed by the software. CD31 + vessels within the ROI were counted; a vessel was defined as a lumen-containing structure enclosed entirely by EC. Vessel density was calculated as vessel number divided by implant area (vessels/mm2). This was repeated for UEA1 + vessels. The proportion of UEA1 + vessels (graft-derived) was determined as the ratio of UEA1 + to CD31 + vessel density.

Experimental series 2: varying adMSC density. adMSC density was increased within small volume implants, such that the total adMSC number was equivalent to that in Group 1 (1.5 · 106 total adMSC). The adMSC density within Group 2 implants was 1.5 · 108 adMSC/mL. Module diameter (determining HUVEC density) was the same as the base case.

Quantification of HIF1a + cells. The density of HIF1a +

cells within the implant was determined by defining three ROIs (200 · 200 mm) within each implant. ROIs were located in the implant interior, excluded the fibrotic capsule, and enclosed the maximum number of HIF1a + cells (hotspot). HIF1a + cells were identified as those with positively stained nuclei; those with cytoplasmic HIF1a staining only, and/or with nuclei stained solely for the hematoxylin (blue) were termed HIF1a - . An average density and proportion of HIF1a + nuclei was computed from the three ROIs.

Experimental series 3: varying adMSC and HUVEC density. Both adMSC and HUVEC density were increased in

Group 3 implants. adMSC density was increased to 1.5 · 108 adMSC/mL, as in experimental series 2. HUVEC density was increased about 2.5-fold by fabricating modules with a diameter of 200 mm (after contraction). Histology

Quantification of CD45 and F4/80-positive pixels. The number of positive CD45 or F4/80 pixels (intensity above 100) in the implant was quantified using the Positive Pixel Count algorithm of the ImageScope software. Pixel density was calculated based on the implant area.

Modular tissues were explanted at day 7 and 14, and fixed in 10% neutral buffered formalin for at least 24 h prior to immunohistochemical processing. Briefly, samples were embedded in paraffin wax, dewaxed with xylene, and rehydrated with graded alcohol. Sections 4 mm thick were stained for the following: Masson Trichome, anti-CD31 (Santa Cruz Biotechnology, Inc., SC-1506-R, 1:2000, rabbit polyclonal antibody), biotinylated Ulex Europaeus Agglutinin I (UEA1; Vector Laboratories B1065, 1:400; stains human EC only), anti-a smooth muscle actin (SMA; Sigma-Aldrich A5228, 1:200, mouse monoclonal antibody), anti-HIF1a (Novus Biologicals NB100-479; 1:2000, rabbit polyclonal antibody), antiCD45 (BD Biosciences 550539; rat monoclonal antibody), and anti-F4/80 (Serotec MCA497GA; monoclonal antibody).

Statistical analyses

A one-way analysis of variance with the Tukey’s honest significant difference post hoc test was used to compare means between multiple experimental groups and time points. Differences were considered statistically significant at p < 0.05. Analyses were performed using Statistica software (StatSoft 1997). Results

While Figures 1–7 and the paragraphs below detail the results, a simple summary is provided in Table 2. For many cases, changing cell density or implant volume had no effect on measured outcomes; it was Group 3 (higher adMSC and

Quantitative analyses of histology

Histology slides were scanned at 20· magnification (40 · magnification for HIF1a-stained slides) using Scan-

Table 2. Summary of Experimental Effects on Graft-Derived and Total Vessel Formation, and HIF1a + Nuclei Relative to the Base Case Effect on graft-derived vessel Group Base Group 1 Group 2 Group 3

Change relative to base case

Proportion

Density

Effect on total vessel density

Effect on HIF1a + nuclei

— Larger volume Higher adMSC density Higher adMSC and HUVEC density

o + o +

o o o +

o o o +

o + o +

‘‘o’’ represents no statistically significant difference relative to the base case, and ‘‘ + ’’ indicates a statistically significant increase relative to the base case ( p < 0.05). HIF1a, hypoxia-inducible factor 1-alpha.

4

HUVEC density) that had a dramatic impact on all parameters. A larger volume increased the density of HIF1a + nuclei and the proportion of graft-derived vessels, but not total vessel density. Implant volume and cell density affect the mechanism of vascularization in modular implants

With the exception of Group 3 at day 7, total (host- and graft-derived, CD31 + ) vessel density was similar across all groups and time points (Fig. 1). Other apparent differences were not statistically significant. Not only did increasing both adMSC and HUVEC density (Group 3) markedly increase total vessel density compared

LAM AND SEFTON

to the base case ( p < 0.01), it was also increased relative to implants with high adMSC density alone (Group 2, p < 0.05). Also, the high total vessel density observed in Group 3 at day 7 was due to high graft-derived vessel contribution. By day 14, however, a significant decrease in total vessel density was observed for Group 3 ( p < 0.01), but not for any of the other groups. Representative histological images are presented in Figure 2. Although pronounced differences in total vessel density were not observed for Groups 1 and 2, manipulated parameters affected the relative contributions of host- versus graft-derived vessels (Fig. 1). When implant volume was increased 10-fold at constant cell density (Group 1, total cell

FIG. 1. Vessel density and proportions in modular implants of different volumes and cell densities. (A) Total (CD31 + ) vessel density; (B) Graft-derived (UEA1 + ) vessel density; (C) Proportion of graft-derived vessels to total (graft and hostderived) vessels within the implant. At day 7, total vessel density was significantly increased in implants of highest cell density (Group 3), and was attributed to increased graft-derived vessel density and proportion. Although total vessel density was similar across other groups, differences were observed in graft-derived vessel contribution. When compared to the base case at day 7, large implant volume (Group 1) increased graft-derived vessel proportion, an effect not observed by increasing adMSC density alone (Group 2). Interestingly, graft-derived vessel formation appears to be delayed in base case implants, with high UEA1 + vessel proportion at day 14. n = 5; – SEM; *p < 0.05 and **p < 0.01. adMSC, adipose-derived mesenchymal stromal cells. Color images available online at www.liebertpub.com/tea

MODULAR TISSUE VASCULARIZATION HIF1 ACTIVATION

5

FIG. 2. Histology images of vascularization in modular implants of different implant volumes and cell densities. Overall morphology of the implant is visualized by Masson’s Trichrome staining, where collagen is stained blue and cells are purple. Total vessel structures (host- and graft-derived ECs) are visualized by CD31 staining, while HUVEC (graft-derived) vessels are stained by UEA1 lectin. At day 7, total vessel density was greatest in implants of highest cell density (Group 3), and was attributed to high graft-derived vessel contribution. At this time, UEA1 + vessels also dominated the vascularization of Group 1 implants, with host-derived vessels localized to the implant periphery. On the other hand, host-derived vessels permeated base case and Group 2 implants, and were sustained to day 14 in implants with high adMSC density. Hostderived vessels infiltrated the interior of Group 1 implants by day 14. Scale bar is 300 mm. EC, endothelial cell; HUVEC, human umbilical vein endothelial cell. Color images available online at www.liebertpub.com/tea

numbers were also different), graft-derived vessel proportion was also increased at day 7 (Fig. 1C; p < 0.01). Unlike small volume implants, where host-derived (CD31 + UEA1 - ) vessels were well distributed throughout the implant, those of large volume implants were concentrated at the periphery (Fig. 2). Importantly, graft-derived vascularization was accelerated by increasing adMSC and HUVEC density (Group 3) or implant volume (Group 1) relative to the base case. That is, high graft vessel proportion was observed as early as

day 7 in Group 1 and 3 implants, compared with day 14 in the base case. This effect was not observed when adMSC density was increased alone. In the second experimental series, adMSC density was increased within small volume implants (Group 2), such that total adMSC number was equivalent to that in large volume implants (Group 1). There was no effect on graft-derived vessel density or proportion at day 7. However, unlike the base case, a high proportion of host-derived vessels was maintained to day 14 in Group 2 (Fig. 1C).

6 Implant volume and cell density affect HIF1a expression at day 7

Both implant volume and cell density affected HIF1a staining at day 7. However, by day 14, HIF1a staining was negative or, at most, weakly positive in the nucleus for all groups (Figs. 3 and 4). Increasing implant volume (and cell number) 10-fold resulted in strong HIF1a nuclear expression (Group 1). In fact, the proportion of HIF1a + nuclei was greater than the base case ( p < 0.01), although density (HIF1a + cells/mm2) was not significantly changed. Contrary to expectation, increasing adMSC density alone (Group 2) did not increase HIF1a staining relative to the base case. Perhaps the cell density chosen was not high enough to enhance HIF1a nuclear expression to a level comparable to a 10-fold increase in implant volume. To test this idea, both HUVEC and adMSC density were increased within small volume implants. Indeed, density and proportion of HIF1a + nuclei were significantly higher than other small volume implants (Base and Group 2, Fig. 3). Intense nuclear staining was also observed, as with large volume implants at day 7 (Fig. 4). In implants with strong HIF1a staining at day 7 (Groups 1 and 3), the identities of HIF1a + cells were examined using serially stained tissue sections (Fig. 5). Some HIF1a + cells were located within modules, and appeared to be differentiated adMSC based on their spindle-shaped morphology and aSMA expression. Other HIF1a + cells were located between modules—likely adMSC that had migrated from within the module or host-derived fibroblasts and/or inflammatory cells. Importantly, many intensely stained HIF1a + cells were EC lining blood vessels, most of which were graft-derived. On the contrary, vessels of the base case

LAM AND SEFTON

and Group 2 implants (graft- and host-derived alike) were mostly lined by EC that did not stain for HIF1a (Fig. 4, insets). Implant volume and cell density exert subtle effects on inflammatory cell recruitment

Implant volume and cell density had little effect on the density of CD45 + cells recruited to the implant. With the exception of day 7, at which time CD45 + pixel density was greater in the base case than in large volume implants (Group 1; p < 0.05), similar CD45 recruitment was observed in all groups (Fig. 6). Some effects were also observed in macrophage recruitment. Specifically, at day 14, F4/80 staining was significantly decreased in implants of high adMSC density (Group 2; p < 0.05) and high adMSC and HUVEC density (Group 3; p < 0.05) compared with the base case. Moreover, Group 2 and 3 implants saw a significant decrease in F4/80 + cell recruitment from day 7 to 14 ( p < 0.05), while a high density of host-derived macrophages was maintained in base case implants at day 14. CD45 + and F4/80 + cells were located at the periphery of large volume implants (Group 1), they were present as ‘‘concentrated’’ pockets within small volume implants (Base, Group 2 and 3) at day 7. By day 14, however, CD45 and F4/80 staining in the interior of large volume implants was significantly greater than at day 7 (Fig. 7). Discussion

Implant volume and cell density affected the degree and nature of the vascular and inflammatory responses, while

FIG. 3. Average density (A) and average proportion (B) of HIF1a + nuclei within modular implants of different volumes and cell densities. HIF1a + cells were defined by nuclear (not cytoplasmic) localization of HIF1a expression. At day 7, HIF1a + nuclei density and proportion was markedly higher in implants of highest cell density (Group 3). Increasing implant volume 10-fold (Group 1) resulted in a significant increase in HIF1a + nuclei proportion, compared to base case and Group 2 implants. By day 14, differences in HIF1a + nuclei density and proportion were resolved among all groups. n = 5; – SEM; **p < 0.01. HIF1a, hypoxia-inducible factor 1-alpha. Color images available online at www.liebertpub.com/tea

MODULAR TISSUE VASCULARIZATION HIF1 ACTIVATION

7

FIG. 4. Expression of HIF1a in modular implants of different volumes and cell densities. HIF1a + cells were characterized by nuclear staining (red arrows), while HIF1a - cells were identified by cytoplasmic staining or absence of nuclear staining (black arrows). HIF1a + and HIF1a + blood vessels are magnified in the inset. At day 7, strong nuclear HIF1a expression was observed in implants of large volume (Group 1) and of highest cell density (Group 3). Insets show that blood vessels in the interior of these implants were predominantly lined by HIF1a + ECs. This effect was resolved by day 14. On the contrary, low HIF1a expression was observed in small volume implants of low (base case) and high adMSC density (Group 2) at day 7 and 14 alike. By contrast, insets show that blood vessels are lined by the HIF1a - ECs at both day 7 and 14. Scale bar is 200 mm. EC, endothelial cells. Color images available online at www.liebertpub.com/tea altering the nature of HIF1a staining within modular implants. To some extent, these are tied together by hypoxia, but this may be somewhat oversimplified. HIF1a staining was nearly absent by day 14, suggesting that any effects were most pronounced in the early period following implantation. A key finding of this study was that implant volume and cell density affected the relative contributions of host- versus graft-derived vascularization in modular tissues, suggesting that growth of these two vessel sources are governed by different mechanisms. This is not surprising, given their intrinsic difference; graft-derived vessels assemble from

within the implant, while host vessels grow into the implant from surrounding tissue. Graft-derived vascularization may be governed by a HIF1-mediated EC response

HIF1a expression may be a driver of graft-derived vessel formation, but appears to have a less direct role in hostderived vessel ingrowth. While HIF1a staining was nearly absent in implants with low graft-derived vessel contribution at day 7 (Base and Group 2), the converse was true of implants with high graft-derived vascularization (Group 1

8

LAM AND SEFTON

FIG. 5. Identifying HIF1a-expressing cells in implants of large volume (Group 1) and of highest cell density (Group 3) at day 7, by serial staining with SMA, CD31, UEA1, and CD45. HIF1a + cells included graft-derived (CD31 + UEA1 + ) blood vessels in the implant interior (red circles), spindle-shaped SMA + CD45 - cells within and between modules (blue arrows), and CD45 + SMA - inflammatory cells between modules (black arrows). Unlike graft-derived vessels, vessels of host origin were predominantly lined by HIF1a - ECs (black circles) and located at the implant periphery. Scale bar is 200 mm. SMA, smooth muscle actin. Color images available online at www.liebertpub.com/tea and 3, Fig. 4). Importantly, intense HIF1a nuclear staining was observed in vessel-lining HUVEC of Group 1 and 3 implants at day 7, but generally absent in EC of host-derived vessels (Fig. 5). This difference may be associated with accelerated graftderived vessel assembly in Group 1 and 3 implants, com-

pared to the base case. Whereas graft-derived vessel proportion peaked at day 14 in base case implants, this peak occurred at day 7 in large volume implants (Group 1) or ones of high cell density (Group 3). Although these observations do not indicate a causal relationship, the postulate that EC-specific HIF1a expression

MODULAR TISSUE VASCULARIZATION HIF1 ACTIVATION

9

FIG. 6. Density of CD45 + (A) and F4/80 + pixels (B) within modular implants of different volumes and cell densities. At day 7, the density of CD45 + pixels was significantly higher in small volume implants (base case) than in large volume implants (Group 1) of the same cell density; this difference was resolved by day 14. In small volume implants with high adMSC density (Group 2 and 3), a significant decrease in F4/80 + pixel density was observed from day 7 to 14. n = 5; – SEM; *p < 0.05 and **p < 0.01. Color images available online at www.liebertpub.com/tea

drives graft-derived vessel formation is supported by the literature. First, HIF1 activation increases EC sensitivity to angiogenic signals, via upregulation of VEGFR-1,32 VEGFR-2,33 and Dll4.34 Short exposure or moderate levels of hypoxia (2.5% oxygen) enhances in vitro tube formation by EC22,35; this response is HIF1-mediated, since EC isolated from mice depleted of the HIFb subunit showed defective migration, proliferation, and tube formation.22 To confirm this role in the context of engineered tissue, one would need to knock-down HIF1a in transplanted EC. Host vessels rapidly permeate small volume implants and appear to be sustained by adMSC effects

Results also suggest that host-derived vascularization is unlike graft-derived vascularization in two other aspects. Foremost, while host vessels permeated small volume implants (Base, Group 2 and 3), they were localized to the periphery of large volume implants at day 7 (Group 1, Fig. 2). Since host vessel ingrowth involves simultaneous degradation and remodeling of the collagen matrix,36–39 it would follow that total infiltration of large volume implants requires longer times in large-volume implants. Second, the degree of host vessel ingrowth is dependent upon adMSC effects. Increasing adMSC density alone relative to the base case (Group 2, without corresponding changes in HUVEC number) did not enhance graft-derived vessel formation at day 7, but sustained high host-derived

vessel proportion to day 14 (Fig. 1; note the markedly low proportion of graft-derived UEA1 + vessels in Group 2 at day 14 compared to the base case, p < 0.01). Thus, graftderived vascularization is not further enhanced by adMSC effects, whereas host vessel ingrowth is. The adMSC density was increased within small volume implants (Group 2), such that total adMSC number was equivalent to that in large volume implants (Group 1) because the difference in total adMSC number in Base and Group 1 implants was a confounding variable. One could attribute observed differences in graft-derived vascularization to the amount of paracrine factors released by adMSC, independent of HIF1a expression. However, increasing adMSC density alone did not affect graft-derived vascularization. Only when both HUVEC and adMSC densities were increased (Group 3) was there a pronounced effect, for reasons proposed in the previous section. Paracrine secretions from MSCs can attract the ingrowth of host vessels.40,41 In fact, this effect is further amplified under hypoxic conditions, as HIF1 activation increases MSC secretion of VEGF-A, FGF-2, IL-6, and SDF-1.8,24,26,40 Some angiogenic factors even have synergistic effects on collateral growth.42,43 Not surprisingly, robust HIF1a staining was observed in cells within module implants at day 7, presumably differentiated adMSC, host-derived fibroblasts, and myeloid cells (Fig. 5). Apart from attracting hostderived vessels, these factors can rescue the survival and tubulogenesis of hypoxic ECs25,26,40 and recruit hostderived cells29,31,44,45 responsible for angiogenesis and

10

LAM AND SEFTON

FIG. 7. Histology images of inflammatory cell infiltration in modular implants of different implant volumes and cell densities. Overall morphology of the implant is visualized by Masson’s Trichrome staining (first row), in which collagen is stained blue and cell nuclei are purple. Leukocytes are stained by CD45 (second row), while host-derived macrophages are stained by F4/80 (third row). At day 7, leukocytes and macrophages infiltrated small volume implants (Control, Group 2 and 3) and were localized as concentrated pockets within the implant interior. On the contrary, CD45 + and F4/80 + cells surrounded the periphery of large volume implants (Group 1) at day 7, and didn’t infiltrate the implant interior until day 14. Macrophage presence was significantly reduced in small volume implants with high adMSC density (Group 2 and 3) from day 7 to 14. Scale bar is 500 mm. Color images available online at www.liebertpub.com/tea anastomosis of the new vasculature.46–49 Recruitment of angiogenic host cells is largely attributed to HIF1-mediated activation of the CXCR4/SDF-1 axis.45,50 HUVEC and adMSC have distinct but complementary roles in modular tissue vascularization

Although results indicate that increasing adMSC density alone had no effect on graft-derived vascularization, MSC are indispensable for robust vascularization of EC-containing tissue systems. In their absence, there is poor graftderived vessel formation in modular tissues5,51 and other

similar systems.10,52–54 Thus, adMSC are necessary for graft-derived vessel formation, although increasing their density alone (Group 2) did not enhance graft-derived vascularization in this study. Results further suggest that the number of surviving graft-derived EC dictates the level of graft-derived vascularization. Only when both HUVEC and adMSC density were increased (Group 3) was there significant enhancement of graft vessel formation. In the presence of adMSC and their secreted factors (presumably IL-6, VEGF-A,25 and FGF-240), HIF1a + EC organize into vessel structures, and they contribute to enhanced graft-derived vascularization.

MODULAR TISSUE VASCULARIZATION HIF1 ACTIVATION

EC and adMSC have distinct angiogenic roles, and their interactions are required for robust vascularization of engineered tissues. EC physically organize into vessel structures, although their survival in hypoxic stress is dependent upon paracrine factors from cells like adMSC,16,17,55 which act via activating the phosphatidylinositol 3-kinase (PI3K)Akt pathway,25 for instance. Few, if any, MSC can physically incorporate into the endothelium of new blood vessels,47,48 but play a supportive role via paracrine effects.53,54 HIF1 is a hypoxic sensor in engineered tissue

Results discussed thus far suggest that hypoxia may have an important role in vascularization of engineered tissue. Importantly, HIF1a is part of the cellular response to environmental stresses including, but not limited to, hypoxia. Although oxygen directly affects the post-translational stability of the HIF1a subunit, the HIF1 pathway is also activated by signaling molecules that bind to receptor tyrosine kinases, and subsequently activate the PI3K or mitogenactivated protein kinase pathways.19 The HIF1 pathway is therefore also activated in an oxygen-independent manner. That said, the two parameters investigated here—implant volume and cell density—are likely closely tied to hypoxia. Using oxygen microelectrodes, others demonstrated that oxygen tensions decreased spatially and with increasing cell density in engineered tissue comprised of fibroblasts in collagen.14 The same parameters also negatively impacted

11

cell survival within engineered cartilage11 and bone,13 presumably due to spatial variations in oxygen tension. It is thus reasonable to infer that implant volume and cell density similarly affected oxygen tension within modular engineered tissues, contributing to the HIF1a expression patterns observed. Modular tissue vascularization: a proposed model

A model of modular tissue vascularization is proposed (Fig. 8), based on results presented here. In this model, hypoxic stress is induced at implantation, activating HIF1 in graft-derived cells. Specific changes effected by HIF1 activation depend on the cell type affected. HIF1 activation in HUVEC increases their responsiveness not only to angiogenic signals, via upregulation of VEGFR1, VEGFR-2, and Dll4 (discussed earlier), but also to apoptotic signals, such as p53 and p38/MAPK. The decision to undergo apoptosis or angiogenesis is partially dependent upon the balance of HIF1a association with p53 versus HIFb.56,57 VEGF also promotes EC survival by binding to VEGFR-2, activating the PI3K/Akt pathway,58 and concomitantly decreasing p38/MAPK activation.16,59 VEGF secretion is an aspect of the beneficial effect of adMSC on HUVEC. Of note, the rescuing effect of hypoxic MSCconditioned medium on survival and tubulogenesis of hypoxic EC is partly attenuated by VEGF25 and FGF-2 neutralizing antibodies.40 Thus, EC apoptose in severe hypoxia

FIG. 8. Proposed model of modular tissue vascularization. Hypoxic stress is induced upon implantation of modules, and results in HIF1 activation in graft-derived cells. HIF1a nuclear translocation (and subsequent HIF1 activation) augments adMSC secretion of angiogenic factors, which (1) direct the ingrowth of host-derived vessels, (2) recruit host-derived angiogenic cells, and (3) tip the hypoxic HUVEC profile toward an angiogenic disposition. HIF1 activation in HUVEC can lead to apoptosis or tubulogenesis, a decision related to the balance of apoptotic (p53 and p38) versus prosurvival and angiogenic activation (PI3K/Akt and VEGF). Apoptotic HUVEC do not establish functional vessels. In the presence of adMSC-secreted paracrine factors, however, HUVEC organize into tubular structures, which can anastomose with ingrown host-derived vessels. Tubular structure formation is enhanced by paracrine factors released by host cells recruited to the implant site. Nonperfused vessel structures undergo apoptosis. PI3K, phosphatidylinositol 3-kinase. Color images available online at www.liebertpub.com/tea

12

and/or when there are insufficient levels of angiogenic factors from supportive cells. On the other hand, EC survive, migrate, proliferate, and contribute to tissue vascularization in the presence of adMSC-secreted VEGF.23 HIF1 activation in adMSC augments the secretion of known (VEGF-A, FGF-2, IL-6, and SDF-1) and unknown paracrine factors. Apart from their beneficial effect on HUVEC, these factors recruit host-derived angiogenic cells, like fibroblasts and myeloid cells, largely via HIF1dependent activation of the CXCR4/SDF-1 axis. Together, these cells contribute to tissue vascularization by (1) assuming a pericytic role,46,47 (2) augmenting the pool of angiogenic signals in the implant, and (3) directing the ingrowth of host-derived vessels. Tissue oxygenation is established as host-derived vessels anastomose with graft-derived vessel structures, become stabilized and perfused. Nonperfused vessels regress, due in part to reduced biomechanical stimulation of EC survival pathways.60,61 Ultimately, a vasculature suited to the metabolic demands of the modular tissue is established. Future directions

This study suggests a connection between HIF1 activation and vascularization of engineered tissue. To identify a causal relationship, however, requires specific knock-down and/or overexpression of HIF1 in the implant. The modular tissue platform could also be adapted for use with endothelial progenitor cells or blood outgrowth ECs to generate a patient-specific source of ECs, as part of a program to render modules more clinically useful. The scale-up of modules from the mouse to a human presents many problems (and opportunities) of which the choice of cell source is but one. Conclusions

Findings of this study indicate that implant volume and cell density affect the balance of host- versus graft-derived vessel contributions in the vascularization of modular implants. Graft-derived vessel formation was most rapid and robust in implants with pronounced HIF1a nuclear expression (large volume implants, and small volume implants of high adMSC and EC density), and weakest in implants with low HIF1a expression (small volume implants with low and high adMSC density). These observations led to the postulate that EC-specific upregulation of HIF1a at early stages is related to graft-derived vessel formation, but unlikely an important factor in the somewhat later host-derived vessel ingrowth. Host-derived vessels dominated the vascularization of small volume implants; their presence was sustained by increasing adMSC density, suggesting that paracrine effects underlie host-derived vessel ingrowth. Manipulating cellspecific expression of HIF1a may therefore be an effective method of tuning vascularization in engineered tissues. Acknowledgments

The authors acknowledge financial support from the Canadian Institutes of Health Research (CIHR). G.L. thanks CIHR and the CIHR Training Program in Regenerative Medicine for graduate scholarship support. We also thank Chuen Lo for his expertise in animal surgery, and the team

LAM AND SEFTON

at Toronto General Hospital’s Pathology Research group for their services in immunohistochemistry. Disclosure Statement

No competing financial interests exist. References

1. Carmeliet, P., and Jain, R.K. Angiogenesis in cancer and other diseases. Nature 407, 249, 2000. 2. Phelps, E.A., and Garcı´a, A.J. Engineering more than a cell: vascularization strategies in tissue engineering. Curr Opin Biotechnol 21, 704, 2010. 3. McGuigan, A.P., and Sefton, M.V. Vascularized organoid engineered by modular assembly enables blood perfusion. Proc Natl Acad Sci U S A 103, 11461, 2006. 4. Chamberlain, M.D., Gupta, R., and Sefton, M.V. Bone marrow-derived mesenchymal stromal cells enhance chimeric vessel development driven by endothelial cell-coated microtissues. Tissue Eng Part A 18, 285, 2012. 5. Butler, M.J., and Sefton, M.V. Cotransplantation of adipose-derived mesenchymal stromal cells and endothelial cells in a modular construct drives vascularization in SCID/ bg mice. Tissue Eng Part A 18, 1628, 2012. 6. Gupta, R., and Sefton, M.V. Application of an endothelialized modular construct for islet transplantation in syngeneic and allogeneic immunosuppressed rat models. Tissue Eng Part A 17, 2005, 2011. 7. Leung, B.M., and Sefton, M.V. A modular approach to cardiac tissue engineering. Tissue Eng Part A 16, 3207, 2010. 8. Melero-Martin, J.M., De Obaldia, M.E., Kang, S.Y., Khan, Z.A., Yuan, L., Oettgen, P., and Bischoff, J. Engineering robust and functional vascular networks in vivo with human adult and cord blood-derived progenitor cells. Circ Res 103, 194, 2008. 9. Verseijden, F., Posthumus-van Sluijs, S.J., Farrell, E., Van Neck, J.W., Hovius, S.E.R., Hofer, S.O.P., and Van Osch, G.J.V.M. Prevascular structures promote vascularization in engineered human adipose tissue constructs upon implantation. Cell Transplant 19, 1007, 2010. 10. Traktuev, D.O., Prater, D.N., Merfeld-Clauss, S., Sanjeevaiah, A.R., Saadatzadeh, M.R., Murphy, M., Johnstone, B.H., Ingram, D.A., and March, K.L. Robust functional vascular network formation in vivo by cooperation of adipose progenitor and endothelial cells. Circ Res 104, 1410, 2009. 11. Freed, L.E., Vunjak-Novakovic, G., Marquis, J.C., and Langer, R. Kinetics of chondrocyte growth in cell-polymer implants. Biotech Bioeng 43, 597, 1994. 12. Radisic, M., Euloth, M., Yang, L., Langer, R., Freed, L.E., and Vunjak-Novakovic, G. High-density seeding of myocyte cells for cardiac tissue engineering. Biotech Bioeng 82, 403, 2003. 13. Dunn, J.C.Y., Chan, W.Y., Cristini, V., Kim, J.S., Lowengrub, J., Singh, S., and Wu, B.M. Analysis of cell growth in three-dimensional scaffolds. Tissue Eng 12, 705, 2006. 14. Cheema, U., Brown, R.A., Alp, B., and MacRobert, A.J. Spatially defined oxygen gradients and vascular endothelial growth factor expression in an engineered 3D cell model. Cell Mol Life Sci 65, 177, 2008. 15. Malda, J., Rouwkema, J., Martens, D.E., Le Comte, E.P., Kooy, F.K., Tramper, J., Van Blitterswijk, C.A., and Riesle, J. Oxygen gradients in tissue-engineered PEGT/PBT cartilaginous constructs: measurement and modeling. Biotech Bioeng 86, 9, 2004.

MODULAR TISSUE VASCULARIZATION HIF1 ACTIVATION

16. Eguchi, R., Suzuki, A., Miyakaze, S., Kaji, K., and Ohta, T. Hypoxia induces apoptosis of HUVECs in an in vitro capillary model by activating proapoptotic signal p38 through suppression of ERK1/2. Cell Signal 19, 1121, 2007. 17. Luo, J., Martinez, J., Yin, X., Sanchez, A., Tripathy, D., and Grammas, P. Hypoxia induces angiogenic factors in brain microvascular endothelial cells. Microvasc Res 83, 138, 2012. 18. Ryan, H.E., Lo, J., and Johnson, R.S. HIF-1a is required for solid tumor formation and embryonic vascularization. EMBO J 17, 3005, 1998. 19. Semenza, G.L. Targeting HIF-1 for cancer therapy. Nat Rev Cancer 3, 721, 2003. 20. Pugh, C.W., and Ratcliffe, P.J. Regulation of angiogenesis by hypoxia: role of the HIF system. Nat Med 9, 677, 2003. 21. Wang, G.L., and Semenza, G.L. General involvement of hypoxia-inducible factor 1 in transcriptional response to hypoxia. Proc Natl Acad Sci U S A 90, 4304, 1993. 22. Han, Y., Yang, K., Proweller, A., Zhou, G., Jain, M.K., and Ramirez-Bergeron, D.L. Inhibition of ARNT severely compromises endothelial cell viability and function in response to moderate hypoxia. Angiogenesis 15, 409, 2012. 23. Tang, N., Wang, L., Esko, J., Giordano, F.J., Huang, Y., Gerber, H.P., Ferrara, N., and Johnson, R.S. Loss of HIF-1a in endothelial cells disrupts a hypoxia-driven VEGF autocrine loop necessary for tumorigenesis. Cancer Cell 6, 485, 2004. 24. Ohnishi, S., Yasuda, T., Kitamura, S., and Nagaya, N. Effect of hypoxia on gene expression of bone marrow-derived mesenchymal stem cells and mononuclear cells. Stem Cells 25, 1166, 2007. 25. Hung, S.C., Pochampally, R.R., Chen, S.C., Hsu, S.C., and Prockop, D.J. Angiogenic effects of human multipotent stromal cell conditioned medium activate the PI3K-Akt pathway in hypoxic endothelial cells to inhibit apoptosis, increase survival, and stimulate angiogenesis. Stem Cells 25, 2363, 2007. 26. Ben-Shoshan, J., Schwartz, S., Luboshits, G., MayselAuslender, S., Barzelay, A., Polak-Charcon, S., Tzahor, E., Barshack, I., Barak, A., Levkovitch-Verbin, H., Keren, G., and George, J. Constitutive expression of HIF-1a and HIF2a in bone marrow stromal cells differentially promotes their proangiogenic properties. Stem Cells 26, 2634, 2008. 27. Takahashi, T., Kalka, C., Masuda, H., Chen, D., Silver, M., Kearney, M., Magner, M., Isner, J.M., and Asahara, T. Ischemia- and cytokine-induced mobilization of bone marrow-derived endothelial progenitor cells for neovascularization. Nat Med 5, 434, 1999. 28. Liu, L., Yu, Q., Lin, J., Lai, X., Cao, W., Du, K., Wang, Y., Wu, K., Hu, Y., Zhang, L., Xiao, H., Duan, Y., and Huang, H. Hypoxia-inducible factor-1a is essential for hypoxiainduced mesenchymal stem cell mobilization into the peripheral blood. Stem Cells Dev 20, 1961, 2011. 29. Cramer, T., Yamanishi, Y., Clausen, B.E., Fo¨rster, I., Pawlinski, R., Mackman, N., Haase, V.H., Jaenisch, R., Corr, M., Nizet, V., Firestein, G.S., Gerber, H.P., Ferrara, N., and Johnson, R.S. HIF-1a is essential for myeloid cellmediated inflammation. Cell 112, 645, 2003. 30. Lord-Dufour, S., Copland, I.B., Levros, L.C., JR., Post, M., Das, A., Khosla, C., Galipeau, J., Rassart, E., and Annabi, B. Evidence for transcriptional regulation of the glucose-6phosphate transporter by HIF-1a: targeting G6PT with mumbaistatin analogs in hypoxic mesenchymal stromal cells. Stem Cells 27, 489, 2009.

13

31. Rosova´, I., Dao, M., Capoccia, B., Link, D., and Nolta, J.A. Hypoxic preconditioning results in increased motility and improved therapeutic potential of human mesenchymal stem cells. Stem Cells 26, 2173, 2008. 32. Gerber, H.P., Condorelli, F., Park, J., and Ferrara, N. Differential transcriptional regulation of the two vascular endothelial growth factor receptor genes. Flt-1, but not Flk-1/ KDR, is up-regulated by hypoxia. J Biol Chem 272, 23659, 1997. 33. Waltenberger, J., Mayr, U., Pentz, S., and Hombach, V. Functional upregulation of the vascular endothelial growth factor receptor KDR by hypoxia. Circulation 94, 1647, 1996. 34. Patel, N.S., Li, J.L., Generali, D., Poulsom, R., Cranston, D.W., and Harris, A.L. Up-regulation of delta-like 4 ligand in human tumor vasculature and the role of basal expression in endothelial cell function. Cancer Res 65, 8690, 2005. 35. Toffoli, S., Roegiers, A., Feron, O., Van Steenbrugge, M., Ninane, N., Raes, M., and Michiels, C. Intermittent hypoxia is an angiogenic inducer for endothelial cells: role of HIF1. Angiogenesis 12, 47, 2009. 36. Fisher, G., Gilbertson-Beadling, S., Powers, E.A., Petzold, G., Poorman, R., and Mitchell, M.A. Interstitial collagenase is required for angiogenesis in vitro. Dev Biol 162, 499, 1994. 37. Ingber, D.E. Mechanical signaling and the cellular response to extracellular matrix in angiogenesis and cardiovascular physiology. Circ Res 91, 877, 2002. 38. Korff, T., and Augustin, H.G. Tensional forces in fibrillar extracellular matrices control directional capillary sprouting. J Cell Sci 112, 3249, 1999. 39. Vernon, R.B., Angello, J.C., Iruela-Arispe, M.L., Lane, T.F., and Sage, E.H. Reorganization of basement membrane matrices by cellular traction promotes the formation of cellular networks in vitro. Lab Invest 66, 536, 1992. 40. Kinnaird, T., Stabile, E., Burnett, M.S., Lee, C.W., Barr, S., Fuchs, S., and Epstein, S.E. Marrow-derived stromal cells express genes encoding a broad spectrum of arteriogenic cytokines and promote in vitro and in vivo arteriogenesis through paracrine mechanisms. Circ Res 94, 678, 2004. 41. Kondoh, K., Koyama, H., Miyata, T., Takato, T., Hamada, H., and Shigematsu, H. Conduction performance of collateral vessels induced by vascular endothelial growth factor or basic fibroblast growth factor. Cardiovasc Res 61, 132, 2004. 42. Asahara, T., Bauters, C., Zheng, L.P., Takeshita, S., Bunting, S., Ferrara, N., Symes, J.F., and Isner, J.M. Synergistic effect of vascular endothelial growth factor and basic fibroblast growth factor on angiogenesis in vivo. Circulation 92, II365, 1995. 43. Carmeliet, P., Moons, L., Luttun, A., Vincenti, V., Compernolle, V., De Mol, M., Wu, Y., Bono, F., Devy, L., Beck, H., Scholz, D., Acker, T., Dipalma, T., Dewerchin, M., Noel, A., Stalmans, I., Barra, A., Blacher, S., Vandendriessche, T., Ponten, A., Eriksson, U., Plate, K.H., Foidart, J.M., Schaper, W., Charnock-Jones, D.S., Hicklin, D.J., Herbert, J.M., Collen, D., and Persico, M.G. Synergism between vascular endothelial growth factor and placental growth factor contributes to angiogenesis and plasma extravasation in pathological conditions. Nat Med 7, 575, 2001. 44. Grunewald, M., Avraham, I., Dor, Y., Bachar-Lustig, E., Itin, A., Yung, S., Chimenti, S., Landsman, L., Abramovitch,

14

45.

46.

47.

48.

49.

50.

51. 52. 53.

54.

LAM AND SEFTON

R., and Keshet, E. VEGF-induced adult neovascularization: recruitment, retention, and role of accessory cells. Cell 124, 175, 2006. Ceradini, D.J., Kulkarni, A.R., Callaghan, M.J., Tepper, O.M., Bastidas, N., Kleinman, M.E., Capla, J.M., Galiano, R.D., Levine, J.P., and Gurtner, G.C. Progenitor cell trafficking is regulated by hypoxic gradients through HIF-1 induction of SDF-1. Nat Med 10, 858, 2004. Rajantie, I., Ilmonen, M., Alminaite, A., Ozerdem, U., Alitalo, K., and Salven, P. Adult bone marrow-derived cells recruited during angiogenesis comprise precursors for periendothelial vascular mural cells. Blood 104, 2084, 2004. Ziegelhoeffer, T., Fernandez, B., Kostin, S., Heil, M., Voswinckel, R., Helisch, A., and Schaper, W. Bone marrowderived cells do not incorporate into the adult growing vasculature. Circ Res 94, 230, 2004. Peters, B.A., Diaz, L.A., Jr., Polyak, K., Meszler, L., Romans, K., Guinan, E.C., Antin, J.H., Myerson, D., Hamilton, S.R., Vogelstein, B., Kinzler, K.W., and Lengauer, C. Contribution of bone marrow-derived endothelial cells to human tumor vasculature. Nat Med 11, 261, 2005. Fantin, A., Vieira, J.M., Gestri, G., Denti, L., Schwarz, Q., Prykhozhij, S., Peri, F., Wilson, S.W., and Ruhrberg, C. Tissue macrophages act as cellular chaperones for vascular anastomosis downstream of VEGF-mediated endothelial tip cell induction. Blood 116, 829, 2010. Staller, P., Sulitkova, J., Lisztwan, J., Moch, H., Oakeley, E.J., and Krek, W. Chemokine receptor CXCR4 downregulated by von Hippel-Lindau tumour suppressor pVHL. Nature 425, 307, 2003. Gupta, R., Van Rooijen, N., and Sefton, M.V. Fate of endothelialized modular constructs implanted in an omental pouch in nude rats. Tissue Eng Part A 15, 2875, 2009. Koike, N., Fukumura, D., Gralla, O., Au, P., Schechner, J.S., and Jain, R.K. Creation of long-lasting blood vessels. Nature 428, 138, 2004. Lee, W.Y., Tsai, H.W., Chiang, J.H., Hwang, S.M., Chen, D.Y., Hsu, L.W., Hung, Y.W., Chang, Y., and Sung, H.W. Core-shell cell bodies composed of human cbMSCs and HUVECs for functional vasculogenesis. Biomaterials 32, 8446, 2011. Au, P., Tam, J., Fukumura, D., and Jain, R.K. Bone marrow derived mesenchymal stem cells facilitate engineering of long-lasting functional vasculature. Blood 111, 4551, 2008.

55. Stubbs, S.L., Hsiao, S.T., Peshavariya, H.M., Lim, S.Y., Dusting, G.J., and Dilley, R.J. Hypoxic preconditioning enhances survival of human adipose-derived stem cells and conditions endothelial cells in vitro. Stem Cells Dev 21, 1887, 2012. 56. Suzuki, H., Tomida, A., and Tsuruo, T. Dephosphorylated hypoxia-inducible factor 1a as a mediator of p53-dependent apoptosis during hypoxia. Oncogene 20, 5779, 2001. 57. Ravi, R., Mookerjee, B., Bhujwalla, Z.M., Sutter, C.H., Artemov, D., Zeng, Q., Dillehay, L.E., Madan, A., Semenza, G.L., and Bedi, A. Regulation of tumor angiogenesis by p53-induced degradation of hypoxia- inducible factor 1a. Genes Dev 14, 34, 2000. 58. Gerber, H.P., McMurtrey, A., Kowalski, J., Yan, M., Keyt, B.A., Dixit, V., and Ferrara, N. Vascular endothelial growth factor regulates endothelial cell survival through the phosphatidylinositol 3’-kinase/Akt signal transduction pathway: requirement for Flk-1/KDR activation. J Biol Chem 273, 30336, 1998. 59. Gratton, J.P., Morales-Ruiz, M., Kureishi, Y., Fulton, D., Walsh, K., and Sessa, W.C. Akt down-regulation of p38 signaling provides a novel mechanism of vascular endothelial growth factor-mediated cytoprotection in endothelial cells. J Biol Chem 276, 30359, 2001. 60. Ando, J., and Yamamoto, K. Vascular mechanobiology: endothelial cell responses to fluid shear stress. Circ J 73, 1983, 2009. 61. Califano, J.P., and Reinhart-King, C.A. Exogenous and endogenous force regulation of endothelial cell behavior. J Biomech 43, 79, 2010.

Address correspondence to: Michael V. Sefton, ScD Institute of Biomaterials and Biomedical Engineering University of Toronto 164 College St., Suite 407 Toronto M5S 3G9 Ontario Canada E-mail: [email protected] Received: May 29, 2014 Accepted: September 15, 2014 Online Publication Date: October 30, 2014