Autologous tissue patches acquire vascular identity depending on the environment

Abstract

Vascular identity is genetically determined, but can be altered during surgical procedures. We hypothesized that the environment of the procedure critically alters the identity of autologous tissue patches implanted into the arterial or venous environment. Autologous jugular vein or carotid artery was used as a patch to repair a rat aorta or inferior vena cava. In the aortic environment patches contained neointimal cells that were CD34/Ephrin-B2-dual positive but not CD34/Eph-B4-dual positive; patches expressed Ephrin-B2, notch-4 and dll-4 but not Eph-B4 and COUP-TFII. In the venous environment patches contained neointimal cells that were CD34/Eph-B4-dual positive but not CD34/Ephrin-B2-dual positive; patches expressed Eph-B4 and COUP-TFII but not Ephrin-B2, notch-4 and dll-4. These data show that autologous tissue patches heal by acquisition of the vascular identity determined by the environment into which they are implanted, suggesting some plasticity of adult vascular identity.

Introduction

Patch angioplasty is a common procedure to close medium-diameter vessels, such as the carotid or femoral arteries, after a longitudinal arteriotomy; commonly used patches include autologous vein or artery as well as synthetic patches.^1–3 future materials may include novel tissue engineered patches.⁴ Despite decades of use, the mechanisms of patch healing are not well established. Using a rat aorta and inferior vena cava (IVC) angioplasty model,⁵ we found that pericardial and polyester patches heal by infiltration of endothelial progenitor cells and neointimal endothelialization.^6–8

The role of vascular identity in healing after vascular procedures is becoming more well described; for example, in vein grafts, expression of Eph-B4 is lost and expression of Ephrin-B2 is not strongly induced, suggesting loss of venous identity.⁹ Interestingly, in arteriovenous fistulae, the outflow vein expresses both Eph-B4 as well as Ephrin-B2, that is it acquires both arterial and venous identities.¹⁰ After pericardial or polyester patch angioplasty, these acellular patches gain either arterial or venous identity based upon the location of implantation; patches implanted into the arterial environment acquire arterial identity whereas patches implanted into the venous environment acquire venous identity.^6–8 Interestingly, prosthetic patches in the venous environment exposed to an arteriovenous fistula acquire dual arterial-venous identity, suggest plasticity of implant identity in adults undergoing vascular surgery.⁸

Despite the common use of prosthetics in vascular surgery, including use of prosthetics for patch angioplasty, autologous tissue remains the gold standard for vessel closure and conduits. For example, saphenous vein grafts are commonly used for cardiac and peripheral bypass due to superior patency compared to prosthetic grafts.^11,12 Similarly, in cardiac surgery, internal mammary and radial arteries are used even in preference to vein bypass.^13–15 In addition, the use of autologous vein patches is associated with superior results compared to prosthetic patches in some studies.^2,16,17 We hypothesized that autologous tissue patches also acquire the identity of the environment into which they are implanted. We used the rat patch angioplasty model, implanting autologous jugular vein (JV) or carotid artery (CA) patches into the aorta or the inferior vena cava (IVC) to determine whether the implanted patches acquire arterial markers in the arterial environment or venous markers in the venous environment.

MATERIALS AND METHODS:

Animal Model

Male Wistar rats (6–8 weeks) were used for patch implantation (n=90). The aorta or infrarenal vena cava (IVC) was exposed, and arteriotomy or venotomy was made as previously described.⁶ The arteriotomy or venotomy was closed by running suture in the direct suture (DS) group using 10–0 suture. In the patch groups, the right jugular vein (JV) or the right carotid artery (CA) was harvested from the same rat, and trimmed to approximately 1.5 mm × 4 mm; the arteriotomy or venotomy was closed with a jugular vein (JV) or carotid artery (CA) patch using running 10–0 nylon suture (Supplementary figure 1). Rats were sacrificed after 5 minutes of surgery (day 0) or on postoperative day 14. No immunosuppressive agents, antibiotics, antiplatelet agents or heparin were given at any time. All experiments were approved by the Institutional Animal Care and Use Committee at the Yale University School of Medicine.

Histology

Rats were anesthetized with isoflurane inhalation, and tissues were fixed by transcardial perfusion of phosphate buffered saline (PBS) followed by 10% formalin. Tissue was removed and fixed overnight in 10% formalin followed by a 24-hour immersion in 70 percent alcohol, then embedded in paraffin and sectioned (5 μm thickness). Slides were de-paraffinized and stained with EVG staining kit (Elastic stain kit, DAKO). Neointimal thickness was determined using the mean of measurements from the lumen surface edge to the edge of the patch in three independent areas, as previously described.⁴

Immunohistochemistry

Tissue sections were de-paraffinized and then incubated using primary antibodies overnight at 4°C. After overnight incubation, the sections were incubated with EnVision reagents or proper secondary antibody for 1 h at room temperature and treated with Dako Liquid DAB Substrate Chromogen System (Dako). The sections were then counterstained with Dako Mayer’s Hematoxylin.

Immunofluorescence

Tissue sections were de-paraffinized and then incubated with primary antibodies overnight at 4°C, Sections were then treated with secondary antibodies at room temperature for 1 hour. Sections were stained with SlowFade® Gold Antifade Mountant with DAPI (Life Technologies) and a coverslip applied. Digital fluorescence images were captured and intensity of immunoreactive signals was measured using Image J software (NIH, Bethesda, Maryland).

Western blot

Patches were carefully harvested and removed from surrounding tissue and snap frozen in liquid nitrogen. Samples were crushed and mixed with buffer including protease inhibitors (Roche, Complete Mini 12108700) prior to sonication (5 sec) and centrifugation (135000 rpm, 15 min). Equal amounts of protein from each experimental group were loaded for SDS-PAGE, followed by Western blot analysis with signals detected using the ECL detection reagent. Patches were analyzed individually, without combination of samples.

RNA extraction and quantitative PCR

Total RNA from the aorta or IVC or excised patch was isolated using the RNeasy Mini kit. Reverse transcription was performed using the SuperScript III First-Strand Synthesis Supermix (Invitrogen, Carlsbad, CA). Real-time quantitative PCR was performed using SYBR Green Supermix (Bio-Rad Laboratories, Hercules, CA) and amplified for 40 cycles using the iQ5 Real-Time PCR Detection System (Bio-Rad Laboratories). Primers are listed in Table 1 and their efficiencies were determined by melt curve analysis. All samples were normalized by GAPDH RNA amplification.

Table 1.

Primers.

Gene	Forward primer	Reverse primer
COUP-TFII	CGGAGGAACCTGAGCTACAC	CCACTTTGAGGCACTTTTTGA
delta like ligand-4	AAGGTGCCACTTCGGTTACAC	AATGACACATTCGTTCCTCTCTT
notch-4	CAGAACGCGGATCCCCTCAAGTT	TTCTGATTCCTTCCACCCGAGTTT
Ephrin-B2	ACCGCTAAGGACTGCAGACAG	GTCCAAGTGGGGATCTCCTAG
Eph-B4	CACCCAGCAGCTTGATCCTG	ACCAGGACCACACCCACAAC
GAPDH	AGACAGCCGCATCTTCTTGT	CCACAGTCTTCTGAGTGGCA

Primary and secondary antibodies

Primary antibodies included: anti-α-actin (Abcam, ab5694; IHC and IF, 1:100); anti-CD34 (R&D, AF4117; IF, 1:100); anti-CD68 (ED1; Abcam, ab31630; IHC, 1:100; IF, 1:50);anti- COUPTFII (Novus biologicals, NBP1–67885; IHC, 1:100), anti-dll-4 (santa cruz, sc-18640; IHC, 1:50), anti-Eph-B4 (Santa Cruz, sc-5536; IF, 1:50); anti-Ephrin-B2 (Santa Cruz, sc-1010; IF, 1:50); anti-GAPDH (Cell Signaling, 14C10; WB, 1:2000); anti-notch-4 antibody (Santa Cruz, SC5594; IHC, 1:50); Secondary antibodies used for IF were: donkey anti-goat Alexa-Fluor-488, donkey anti-rabbit Alexa-Fluor-488, donkey anti-rabbit Alexa-Fluor-568, donkey anti-mouse Alexa-Fluor-568 and chicken anti-mouse Alexa-Fluor-488 conjugated antibodies from Invitrogen (1:500). For IHC, sections were incubated with EnVision reagents for 1 h at room temperature and treated with Dako Liquid DAB+ Substrate Chromogen System (Dako). Finally, the sections were counterstained with Mayer’s hematoxylin.

Statistical analysis

Data are expressed as the mean ± SEM. Statistical significance for these analyses was determined using ANOVA and Tukey’s multiple comparisons test for post-hoc testing, and t-tests as appropriate (Prism 6; GraphPad Software, La Jolla, CA). P values less than 0.05 were considered significant.

Results

Since prosthetic patches acquire vascular identity depending on their environment, we determined the vascular identity of autologous tissue patches used to close arteriotomies. Control aortae were directly closed with sutures, without any patch implantation; DS closure showed thin neointima covering the luminal side of the arteriotomy (day 14; Figure 1A, 1B); there was no aneurysm formation, and no significant difference of the luminal area compared to day 0 (Figure 1A). Closure of the aortic arteriotomy with an autologous JV patch resulted in patch dilation and thickening (day 14), with a thick neointima on the luminal side of the JV patch (Figure 1A, 1B); both the overall thickness of the jugular vein patch and the luminal area became significantly larger at day 14 compared to day 0 (Figure 1C, 1D). Closure of the arteriotomy with an autologous CA patch showed patch and neointimal thickening without any luminal dilation or degradation of elastin (Figure 1A–D). The thickness of neointima that formed on the jugular patch was significantly greater than that which formed on the closure with either direct suture or a carotid artery patch (Figure 1B), which corresponded to a similarly greater number of α-actin-positive cells in arteries closed with JV patches (Figure 1A). Similarly, there were greater numbers of CD68+ cells present in JV patches (Supplementary figure 2A, 2B). The larger number of CD68+ cells in JV patches were CD68/TGM2-dual-positive cells as well as CD68/IL10-dual-positive cells, consistent with M2 type macrophages; there were also a smaller number of CD68/iNOS-dual-positive cells, as well as CD68/TNF-α-dual-positive cells, consistent with M1 type macrophages (Supplementary figure 2D–H). CD34/VEGFR2-dual positive cells, e.g. endothelial progenitor cells, were present on the neointimal surface (Figure 1E); interestingly, all patches contained neointimal cells that were CD34/Ephrin-B2-dual positive but not CD34/Eph-B4-dual positive (Figure 1E–F), consistent with the presence of arterial progenitor cells in healing aortae regardless of the type of closure or patch.

Figure 1.

Closure of rat aorta. A) Representative photomicrographs of the rat aorta closed with direct suture (DS), jugular vein (JV) patch, or carotid artery (CA) patch, day 0 and day 14; first row (low power, EVG staining), scale bar, 1 mm; second row (high power, EVG staining); third row (high power, trichrome masson staining); fourth row (high power, α-actin); scale bar, 100 μm. L, lumen; N, neointima; J, jugular vein patch; C, carotid artery patch; day 0, n=3, day 14, n=5. B) Bar graph showing neointimal thickness, day 14; p=0.011 (ANOVA); n=3–5. C) Bar graph showing patch thickness, day 0 and day 14; *, p<0.05, vs. day 0 (t-test); n=3–5. D) Bar graph showing the luminal area of the aorta, day 0 and day 14; p=0.0258 (t-test); n=3–5. E) Immunofluorescence of the neointima after direct suture (DS), jugular vein (JV) patch, carotid artery (CA) patch, day 14. First row, merge of CD34 (green) and VEGFR2 (red); second row, merge of CD34 (green) and Ephrin-B2 (red); third row, merge of CD34 (green) and Eph-B4 (red); DAPI, blue; L, lumen; scale bar, 100 μm; n=3–5. F) Bar graph showing the percentage of CD34/Ephrin-B2-dual-positive or CD34/Eph-B4-dual positive cells among all CD34-positive neointimal cells, day 14; n=3–5.

Tissue patches gain arterial identity in the arterial environment

Since these data suggest that closure of an arteriotomy with an autologous tissue patch may acquire arterial identity, we next determined whether the type of patch closure, e.g. autologous arterial or venous patch, played a role in the type of identity that was formed. In control aortae directly sutured closed, there was no difference in Ephrin-B2 or Eph-B4 transcript numbers at day 0 and 14 (Figure 2A). In aortae closed with JV patches, there were decreased numbers of Eph-B4 transcripts and increased Ephrin-B2 transcripts in the patch at day 14 compared to day 0 (Figure 2B); however, there were no changes in in transcript numbers in CA patches (Figure 2C). Similarly, there was increased Ephrin-B2 and decreased Eph-B4 protein expression in JV patches but not in CA patches or in aortae directly sutured (Figure 2D–F). Immunofluorescence was used to determine the expression of Ephrin-B2 and Eph-B4 in the neointima; in control vessels, Ephrin-B2 was predominantly expressed in native arterial endothelial cells whereas Eph-B4 was predominantly expressed in native venous endothelial cells (Figure 2G, first and second columns). However, by day 14, neointimal endothelial cells expressed Ephrin-B2 with only weak expression of Eph-B4 regardless of the type of patch; there were a small number of α-actin/Ephrin-B2-dual positive neointimal cells in all patches (Figure 2G). Quantification of Ephrin-B2 immunoreactivity in the neointimal endothelial cells showed no significant differences between patches, with similar Ephrin-B2 immunoreactivity in the patches compared to the native aorta and the directly sutured aorta, but significantly higher than in the native jugular vein (Figure 2H).

Figure 2.

Venous patches lose Eph-B4 and gain Ephrin-B2 expression in the rat aorta. A-C) Bar graphs show relative mRNA transcript expression of Eph-B4 and Ephrin-B2 after DS (A), JV patch (B), or CA patch (C), day 0 and day 14; *, p< 0.05 (t-test); n=3. D) Representative Western blot showing Ephrin-B2, Eph-B4 and GAPDH expression after DS, JV patch, or CA patch, day 0 and day 14; n=3. E-F) Bar graphs show densitometry of Ephrin-B2 (E) and Eph-B4 (F) protein expression after DS, JV patch, or CA patch, day 0 and day 14; *, p< 0.05 (t-test); n=3. G) Immunofluorescence of the neointima in native aorta, jugular vein (JV), day 0; and after direct suture (DS), jugular vein (JV) patch, carotid artery (CA) patch, day 14. First row, merge of Eph-B4 (red) and Ephrin-B2 (green); second row, merge of Ephrin-B2 (green) and vWF (red); third row, merge of Ephrin-B2 (green) and α-actin (red); DAPI, blue; L, lumen; scale bar, 100 μm; n=5. H) Bar graph showing Ephrin-B2 density in the endothelial cells, day 14; *, p<0.005 (ANOVA); n=5.

We confirmed these data using the arterial markers notch-4 and dll-4 and the venous marker COUP-TFII; there were increased numbers of dll-4 and notch-4 transcripts and decreased numbers of COUP-TFII transcripts in JV patches without any changes in CA patches or in aortae directly sutured (day 14, Figure 3A–C). Control arteries showed notch-4 and dll-4 immunoreactivity in endothelial cells whereas COUP-TFII immunoreactivity was present in venous endothelial cells (Figure 3D, first and second columns). At day 14, neointimal endothelial cells showed notch-4 and dll-4 immunoreactivity but without detectable COUP-TFII immunoreactivity (Figure 3D); quantification showed similar notch-4, dll-4 and COUP-TFII immunoreactivity in JV and CA patches compared to the native aorta and directly sutured aortae (Figure 3E–3G), that is strong Ephrin-B2, notch-4 and dll-4 and weak Eph-B4 and COUP-TFII immunoreactivity in patches used to close an artery, regardless of the source of the patch. This data suggests that arterial patches placed into the arterial environment retain their arterial identity, but venous patches lose venous identity and gain arterial identity in the arterial environment.

Figure 3.

Venous patches lose COUP-TFII and gain notch-4 and dll-4 expression the rat aorta. A-C) Bar graphs show relative mRNA transcript expression of dll-4, notch-4 and COUP-TF II after DS (A), JV patch (B), or CA patch (C), day 0 and day 14; *, p<0.05; (t-test); n=3. D) Representative photomicrographs show native aorta, native jugular vein, day 0, and neointima of DS, JV patch, or CA patch, day 14. Upper row, anti-dll-4; second row, anti-notch-4; third row, anti- COUP-TF II; scale bar, 100 μm; L, lumen; yellow arrows show the positive cells; n=3–5. E-G) Bar graphs show percentages of notch-4 positive (E), dll-4 positive (F), or COUP-TF II-positive (G) cells among endothelial cells, day 14; *, p<0.005 (ANOVA); n=3–5.

Tissue patches gain venous identity in the venous environment

Since these data show that tissue patches acquire arterial identity when placed into the arterial environment, regardless of whether the patch was originally arterial or venous, we examined whether tissue patches gain venous identity when placed into the venous environment. Control vena cavae were directly closed with sutures, without any patch implantation; DS closure showed thin neointima covering the luminal side of the venotomy (day 14; Figure 4A, 4B); there was no significant difference of the luminal area compared to day 0 (Figure 4D). Closure of the venotomy with a JV patch resulted in a small amount of patch thickening (day 14), with some neointima on the luminal side of the JV patch (Figure 4A–C); however, there was no significant dilation or increase in luminal area (Figure 4A, 4D). Closure of the venotomy with a CA patch showed a small amount of patch and neointimal thickening without any luminal dilation (Figure 4A–D). The thickness of neointima that formed on the patches was not significantly different than that which formed on the closure with DS (Figure 4B), which corresponded to the lack of increased number of α-actin-positive cells in the patches (Figure 4A). There were similar numbers of CD68+ cells present in JV and CA patches as in the DS veins (Supplementary figure 2A, 2C), although the few number of CD68-positive cells were not identifiable as either M1 or M2 type (Supplementary figure 2D–H). CD34/VEGFR2-dual positive cells, e.g. endothelial progenitor cells, were present on the neointimal surface (Figure 4E), and all patches contained neointimal cells that were CD34/Eph-B4-dual positive but not CD34/Ephrin-B2-dual positive (Figure 4E–F), consistent with the presence of venous progenitor cells in healing vena cavae regardless of the type of closure or patch.

Figure 4.

Closure of rat inferior vena cava. A) Representative photomicrographs of the rat inferior vena cava closed with direct suture (DS), jugular vein (JV) patch, or carotid artery (CA) patch, day 0 and day 14; first row (low power, EVG staining), scale bar, 1 mm; second row (high power, EVG staining); third row (high power, trichrome masson staining); fourth row (high power, a-actin); scale bar, 100 μm. L, lumen; N, neointima; J, jugular vein patch; C, carotid artery patch; day 0, n=3, day 14, n=5. B) Bar graph showing neointimal thickness, day 14; *, p> 0.05 (ANOVA); n=3–5. C) Bar graph showing patch thickness, day 0 and day 14; p< 0.05 (t-test); n=3–5. D) Bar graph showing the luminal area of the vena cava, day 0 and day 14; n=3–5. E) Immunofluorescence of the neointima after direct suture (DS), jugular vein (JV) patch, carotid artery (CA) patch, day 14. First row, merge of CD34 (green) and VEGFR2 (red); second row, merge of CD34 (green) and Eph-B4 (red); third row, merge of CD34 (green) and Ephrin-B2 (red); DAPI, blue; L, lumen; scale bar, 100 μm; n=3–5. F) Bar graph showing the percentage of CD34/Eph-B4-dual-positive or CD34/Ephrin-B2-dual-positive cells among all CD34-positive neointimal cells, day 14; n=3–5.

In control vena cavae directly sutured closed, there was no difference in Eph-B4 or Ephrin-B2 transcription numbers at day 0 and 14 (Figure 5A). In vena cavae closed with JV patches, there were also no changes in Eph-B4 or Ephrin-B2 transcript numbers (Figure 5B); however, in CA patches there were increased numbers of Eph-B4 transcripts and decreased Ephrin-B2 transcripts at day 14 compared to day 0 (Figure 5C). Similarly, there was increased Eph-B4 and decreased Ephrin-B2 protein expression in CA patches but not in JV patches or in vena cavae directly sutured (Figure 5D–F). Immunofluorescence was used to determine the expression of Eph-B4 and Ephrin-B2 in the neointima; in control vessels, Eph-B4 was predominantly expressed in native venous endothelial cells whereas Ephrin-B2 was predominantly expressed in native arterial endothelial cells (Figure 5G, first and second columns). However, by day 14, neointimal endothelial cells expressed Eph-B4 with only weak expression of Ephrin-B2 regardless of the type of patch; there were a small number of α-actin/Eph-B4-dual positive neointimal cells in all patches (Figure 5G). Quantification of Eph-B4 immunoreactivity in the neointimal endothelial cells showed no significant differences between patches, with similar Eph-B4 immunoreactivity in the patches compared to the native vena cava and the directly sutured vena cava, but significantly higher than in native artery (Figure 5H).

Figure 5.

Arterial patches lose Ephrin-B2 and gain Eph-B4 expression in the rat inferior vena cava. A-C) Bar graphs show relative mRNA transcript expression of Eph-B4 and Ephrin-B2 after DS (A), JV patch (B), or CA patch (C), day 0 and day 14; *, p< 0.05 (t-test); n=3. D) Representative Western blot showing Ephrin-B2, Eph-B4 and GAPDH expression after DS, JV patch, or CA patch, day 0 and day 14; n=3. E-F) Bar graphs show densitometry of Eph-B4 (E) and Ephrin-B2 (F) protein expression after DS, JV patch, or CA patch, day 0 and day 14; *, p< 0.005 vs.day 0 (t-test); n=3. G) Immunofluorescence of the neointima in native inferior vena cava (IVC), carotid artery (CA), day 0; and after direct suture (DS), jugular vein (JV) patch, carotid artery (CA) patch, day 14. First row, merge of Eph-B4 (red) and Ephrin-B2 (green); second row, merge of Eph-B4 (green) and vWF (red); third row, merge of Eph-B4 (green) and α-actin (red); DAPI, blue; L, lumen; scale bar, 100 μm; n=5. H) Bar graph showing Eph-B4 density in the endothelial cells, day 14; *, p<0.005 (ANOVA); n=5.

We confirmed these data using the venous marker COUP-TFII and the arterial markers notch-4 and dll-4; there was decreased numbers of dll-4 and notch-4 transcripts and increased numbers of COUP-TFII transcripts in CA patches without any changes in JV patches or in vena cavae directly sutured (day 14, Figure 6A–C). Control vena cavae showed COUP-TFII immunoreactivity present, without notch-4 and dll-4 immunoreactivity, in endothelial cells whereas arteries showed notch-4 and dll-4 immunoreactivity in endothelial cells (Figure 6D, first and second columns). At day 14, neointimal endothelial cells showed COUP-TFII immunoreactivity with minimal notch-4 and dll-4 immunoreactivity (Figure 6D); quantification showed similar notch-4, dll-4 and COUP-TFII immunoreactivity in JV and CA patches compared to the native vena cava and directly sutured vena cava (Figure 6E–G), that is strong Eph-B4 and COUP-TFII and weak Ephrin-B2, notch-4 and dll-4 immunoreactivity in patches used to close a vein, regardless of the source of the patch. This data suggests that venous patches placed into the venous environment retain their venous identity, but arterial patches lose arterial identity and gain venous identity in the venous environment.

Figure 6.

Arterial patches gain COUP-TFII and lose notch-4 and dll-4 expression the rat inferior vena cava. A-C) Bar graphs show relative mRNA transcript expression of dll-4, notch-4 and COUP-TF II after DS (A), JV patch (B), or CA patch (C), day 0 and day 14; *, p< 0.05 (t-test); n=3. D) Representative photomicrographs show native IVC, native carotid artery, day 0, and neointima of DS, JV patch, or CA patch, day 14. Upper row, anti-dll-4; second row, anti-notch-4; third row, anti-COUP-TF II; scale bar, 100 μm; L, lumen; yellow arrows show the positive cells; n=3–5. E-G) Bar graphs show percentages of notch-4 positive (E), dll-4 positive (F), or COUP-TF II-positive (G) cells among endothelial cells, day 14; *, p<0.005 (ANOVA); n=3–5.

Discussion

Although prosthetic patches heal by acquisition of environmentally-determined vascular identity, the mechanism by which autologous tissue patches heal was not previously described. We show that autologous tissue patches placed into the arterial environment have identifiable arterial endothelial progenitor cells (Figure 1); arterial-derived patches retain Ephrin-B2, notch-4 and dll-4 expression and venous-derived patches acquire this expression pattern (Figures 2 and 3). Conversely, autologous tissue patches placed into the venous environment have identifiable venous endothelial progenitor cells (Figure 4); venous-derived patches retain Eph-B4 and COUP-TFII expression and arterial-derived patches acquire this expression pattern (Figures 5 and 6). These data show that autologous tissue patches heal by acquisition of the vascular identity determined by the environment into which they are implanted, suggesting some plasticity of adult vascular identity.

We show that venous patches lose venous identity and gain arterial identity in the arterial environment (Figures 1–3). Various experiments Similar changes have been shown in vitro as well as in embryos,^18,19 and we have previously shown loss of venous identity without gain of identity both in vitro ²⁰ and in vivo.⁹ Using a mouse cuffed vein graft model, Koga et al showed that arterial markers were upregulated.²¹ We have shown that both venous and arterial markers are expressed in the arteriovenous fistula environment that has some features of the arterial environment.^10,22 Our finding that autologous venous patches express arterial identity markers in the arterial environment is consistent with these data as well as with our previous data in prosthetic patches.^6,8 It is possible that regulation of vascular identity depends on environmental features such shear stress. Endothelial cell Ephrin-B2 expression is controlled by microenvironmental determinants,²³ and shear stress stimulates expression of Ephrin-B2 in endothelial cells.²⁴ We have shown increased Ephrin-B2 expression in prosthetic patches exposed to increased shear stress.^8,10 However, our finding that arterial progenitor cells are present (Figure 1) suggests that progenitor cells may play a role in acquisition of identity;²⁵ it is unknown whether these progenitor cells are infiltrating from the environment or whether they are activated from endogenous vessel cells.

We also show that arterial patches lose arterial identity and gain venous identity in the venous environment (Figures 4–6). Patches in the venous environment are uncommon but occasionally reported to be used in some oncology resections ^26,27 and living donor liver transplantation;²⁸ however, arterial patches are generally not performed for venous reconstruction. Nevertheless, our findings complement the findings in the arterial environment and strengthen the possibility that environmental cues such as pressure, shear stress and oxygen tension are critical for determination of vascular identity.

We also find that macrophages may play a role in determination of vessel identity, with more M2 type macrophages present than M1 type macrophages at day 14 (Supplemental Figure 2). We have previously shown that macrophages, and especially M2 type macrophages, are critical for vascular remodeling such as occurs during vein graft adaptation and arteriovenous fistula remodeling.^29,30 Macrophages play a complicated role in arterial remodeling such as occurs during aneurysm formation.^31,32 Interestingly, there were more M2 type macrophages in the JV patches compared to the CA patches (Supplemental Figure 2), suggesting a role for macrophages in the developing neointima. However, the relationship between macrophages and development of vascular identity is not clear.

We find that acquisition of vascular identity is dependent upon the environment and not dependent on the particular biomaterial. Regulation of vascular identity may be a strategy to improve the outcomes of vascular interventions.