Abstract
Pancreatic islets are a highly vascularized entity, and their transplantation into diabetic individuals requires optimal revascularization. In addition, β-cells in islets are extremely sensitive to inflammation. α-1-Antitrypsin (AAT), a circulating serine-protease inhibitor that is available for clinical use as an affinity-purified human product, has been shown to protect islets from graft failure in mouse transplantation models and to achieve readily vascularized islet grafts. AAT is known to induce vascular endothelial growth factor (VEGF) expression and release, as well as protect from proteolytic cleavage of VEGF by elastase, promote viability of endothelial cells, and enhance migration of myocytes. Our aim was to examine whether AAT enhances vasculogenesis toward islet grafts. We employed Matrigel-islet plugs as means to introduce islets in an explantable isolated compartment and examined vessel formation, vessel maturation, and inflammatory profile of explants 9 days after implantation. Also, we examined primary epithelial cell grafts that were prepared from lungs of mice that are transgenic for human AAT. In addition, aortic ring sprouting assay was performed, and HUVEC tube formation assays were studied in the presence of AAT. Our findings indicate that islet grafts exhibit mature vessels in the presence of AAT, as demonstrated by morphology, as well as expression of endothelial CD31, smooth muscle actin (SMA), and von Willebrand factor (vWF). Epithelial cells that express human AAT achieved a similar positive outcome. Aortic ring sprouting was enhanced in AAT-treated cultures and also in cultures that contained primary epithelial cells from human AAT transgenic animals in the absence of added AAT. According to the tube formation assay, HUVECs exhibited superior responses in the presence of AAT. We conclude that vasculogenesis toward islet grafts is enhanced in the presence of AAT. Together with the remarkable safety profile of AAT, the study supports its use in the relevant clinical setups.
Introduction
Pancreatic islets of Langerhans are a highly vascularized entity, in accordance with their nonredundant endocrine role (7,8). Indeed, islets exhibit marked sensitivity to hypoxia and display a strong angiogenic potential (25,43). Despite their ability to accept nutrients by diffusion, islets that are under 100 μm in size share capillaries that belong to the exocrine network, while larger islets directly accept one to three capillaries that branch out of local arterioles (77). The importance of intact blood supply to islets is exemplified by the fact that they are perfused by 5–25% of the pancreatic blood supply and exercise their own blood perfusion distribution regulation, while consisting only about 1% of the pancreatic mass (26,36). In an animal model for type 1 diabetes, a clear association was found between islet blood supply and islet dysfunction (7).
Isolated human islets are embolized intraportally into the oxygen-poor liver portal system during the process of islet transplantation for diabetic patients (53). The islets arrive at a state of advanced injury, after being severed, typically, from a brain-dead donor, removed from supporting extracellular matrix (58), and processed for engraftment in an injurious procedure that consists of enzymatic digestion and aggressive mechanical shear force (14). Lodged in the recipient's liver sinusoids, up to 60% of the damaged islets expire within the first 48 h regardless of antigen identity (2,53), while the surviving islets begin the process of revascularization (6,64). With the multiple challenges that the procedure of islet transplantation faces to date, inadequate angiogenesis is still considered to be a limiting factor toward optimal islet function (11,38,41). Furthermore, two of three immunosuppressive regimens that are included in the human islet transplantation program, namely, sirolimus and tacrolimus, have recently been shown to interfere with islet graft angiogenesis (29).
Islets are active participants in graft revascularization (6,64). Islet-derived vascular endothelial growth factor (VEGF) and its receptors are elevated in hypoxic islets (13,40) and participate in graft revascularization (32), along with mediators such as fibroblast growth factor (FGF) and endothelial nitric oxide synthetase (eNOS) (10,40). Markers that signify the elaboration of mature blood vessels in grafts include the endothelial cell marker cluster of differentiation 31 (CD31) (16) as well as smooth muscle actin (SMA) (45), von Willebrand factor (vWF) (51), and endothelial surface marker CD40 (52). Also, newly formed vessels are expected to be able to contain the flow and pressure of blood without leaking neither fluid nor erythrocytes (54). All the while, factors that oppose angiogenesis, such as thrombospondin (Thbs)-1 and plasminogen activator inhibitor (PAI)-1 (67), are required to be reduced in levels. Facilitation of an angiogenic environment, such that results in increased markers for neovascularization, would thus be of benefit during islet transplantation.
α-1-Antitrypsin (AAT) is a serine-protease inhibitor that is released into the circulation from the liver and, to a lesser extent, forms lung epithelial cells during hypoxic and inflammatory conditions [reviewed in Gooptu and Lomas (22)]. AAT is an anti-inflammatory agent and is elevated systemically during the acute phase response by up to sixfold, from an average of 1.3 mg/ml in healthy individuals (50). The primary target of AAT is neutrophil elastase. Patients with genetic deficiency in AAT exhibit 15–60% lower circulating AAT levels than normal and are prone to develop emphysema, a collapse of lung alveolar walls. AAT-deficient patients also exhibit an elevated rate of various types of vasculitis, mediated in part by increased elastase and proteinase-3 activities and higher levels of proinflammatory cytokines (18,20,37,39,55,56,60). To date, AAT-deficient patients are treated with life-long administrations of affinity-purified human AAT, once weekly, with a remarkable safety profile (42,49). With relevance to vascularization, AAT was recently shown to function as an antiapoptotic factor for endothelial cells and for vascular smooth muscle cells (47,48,62,63) and to facilitate the migration and proliferation of vascular smooth muscle cells during vessel formation (24). AAT increases VEGF expression in various cells (48,62,75) and prevents elastase from inactivating VEGF by enzymatic cleavage (31). Indeed, specific blockade of lung VEGF receptors in animals results in emphysema, despite the availability of local AAT (19,27,28), suggesting that AAT deficiency-related lung emphysema is related to reduced AAT-driven VEGF levels. Accordingly, the promoter sequence of AAT contains the hypoxia-responsive element that binds to hypoxia-inducing factor (HIF)-1α (70).
We recently reported that AAT monotherapy elevates VEGF expression levels in islet allografts and prolongs islet allograft survival in mice (34). The change in VEGF expression was demonstrated in Matrigel plugs containing islets. In these studies, it was noted that, during AAT therapy, islets exhibited an elaborate vascular bed, and intraislet VEGF expression levels appeared to respond to added AAT in a dose-dependent manner.
Since multiple favorable attributes of AAT may independently explain its benefits on islet survival, such as its tissue-protective activities (34), we hereby sought to directly examine whether AAT can promote mature blood vessel formation in the context of islet transplantation, both in animal models for cell transplantation as well as in angiogenesis models that are absent of grafted islets. Our hypothesis would be that AAT may facilitate neovascularization of islet grafts.
Additionally, by using genetically modified cells that express transgenic human AAT, we address the possibility that commercially available clinical grade AAT might exert relevant activities by virtue of carryover molecules from its plasma purification process (42).
Importantly, we address the anti-inflammatory environment afforded by AAT and confirm that it coincides with revascularization of isolation-damaged pancreatic islets.
Materials and Methods
Mice
Wild-type (WT) C57BL/6 were purchased from Harlan (Jerusalem, Israel). Lung-specific hAAT-transgenic mice with C57BL/6 background (kind gift from Churg A, University of British Columbia, Vancouver, Canada) were gene rated as described (17), bred in-house, and genotyped routinely, as described (34). Mice heterozygous for hAAT were generated by crossbreeding C57BL/6 with hAAT-homozygous parents. Litters were screened for the presence of the human AAT gene by tail DNA extraction (XNAT2 extraction kit; Sigma-Aldrich, St. Louis, MO, USA) followed by two-step nested PCR amplification using outer sequence (450 bp product), forward 5′-ACTCCTCCGTACCCTCA ACC-3′ and reverse 5′-GCATTGCCCAGGTATTTCAT-3′, and inner sequence (249 bp product) forward 5′-ACTGTC AACTTCGGGGACAC-3′ and reverse 5′-CATGCCTAAA CGCTTCATCA-3′. Circulating levels of hAAT in mice that were heterozygous for hAAT were below the limit of detection (10 ng/ml) as determined by specific ELISA (34) and were less than 0.7 μg/ml in hAAT homozygous mice, as previously reported (17). In all experimental groups, 6- to 8-week-old female mice were used. Animal experiments were approved by the Institutional Animal Care and Use Committee.
Matrigel-Islet Graft In Vivo Transplantation Model
Primary mouse islets were isolated as described (34), and Matrigel-islet grafts containing the following various supplements were prepared as described (34). Human AAT (Aralast™, Baxter, Westlake Village, CA, USA) or phosphate-buffered saline (PBS; Biological Industries, Beit Haemek, Israel) were added to 450 μl growth factor-reduced Matrigel (BD Biosciences, Bedford, MA, USA) at 4°C, together with 20 freshly isolated C57BL/6 islets. Matrigel-islet plugs were injected through a 21-gauge needle (Kalir Engineering, Ltd., HaMerkaz, Israel) into the interscapular subcutaneous dorsal space of 6-week-old transgenic mice heterozygote for hAAT as control recipients or transgenic mice homozygote for hAAT that were administered hAAT every 3 days starting from 1 day prior to transplantation [60 mg/kg IP, Aralast™, same as for islet transplantation in mice (30,34)]. Vascularization was observed 9 days after grafting, at which point animals were harvested and plugs were removed for photo documentation and histological analysis. Graft explantation sites were also photographed. Samples for histology were immediately immersed in 10% neutral-buffered formalin (Sigma-Aldrich). Blood vessel assessment was conducted by five individuals and consisted of blinded grading of the number of vessels per field, maximal and average vessel wall thickness, distance of vessels from adjacent host tissue, and frequency of fluid accumulation baths and hemorrhagic foci in 10 sequential slides, 50 μm apart.
Coagulation Assay
Plasma was collected from C57BL/6 mice into prothrombin time/partial thromboplastin time (PT/PTT) tubes (Estar Medical, Holon, Israel). Blood cells were removed by centrifugation, and 1 ml plasma was transferred to tubes containing thromboplastin (250 μl, 3 U/ml, Instrumentation Laboratory, Lexington, MA, USA) together with either 50 μl human AAT (1 mg/ml, Aralast™), heparin (250 U/ml, Rotexmedica, Trittau, Germany), or PBS. After 90 s, coagulation was determined by observing the shift of fluid in vertically reversed tubes.
Directed In Vivo Angiogenesis Assay (DIVAA)
Uniform angioreactors, which are semiclosed sterile silicon cylinders (angioreactors™, Cultrex, Trevigen Inc., Gaithersburg, MD, USA) were filled with 20 μl of Matrigel that was premixed at 4°C with 20 freshly isolated islets from C57BL/6 mice and 1 μl hAAT (final concentration 50 μg/ml), FGF (final concentration 0.5 μg/ml, Cultrex), or PBS. Two angioreactors were implanted at a distance of 1 cm from each side of the abdominal midline in the subcutaneous space of 6-week-old untreated C57BL/6 mice and transgenic mice homozygote for hAAT that were treated with hAAT (60 mg/kg) every 3 days starting 1 day prior to grafting. Nine days after transplantation, vascularization response was measured by intravenous injection of fluorescein isothiocyanate (FITC)-dextran (200 μl, 15 mg/ml, Sigma-Aldrich/Fluka) 20 min before animal harvest and graft explantation. Each angioreactor's content was removed from the cylinder, digested at 37°C by Dispase™ (Cell Spare™, Cultrex) and centrifuged to collect the cells. Fluid material was analyzed by spectrofluorimetry to determine FITC content, and cellular material was processed for total RNA extraction and RT-PCR.
Immunohistochemistry
Histological sections of paraffin-embedded samples were prepared by 5 μm sectioning and stained by hematoxylin and eosin (H&E; Jackson ImmunoResearch, West Grove, PA, USA) or by specific antibodies. Mouse insulin was stained with guinea pig anti-swine insulin (Dako, Glostrup, Denmark), followed by cyanine-3 (CY-3)-conjugated donkey anti-guinea pig IgG (Jackson ImmunoResearch). Mouse SMA was stained using monoclonal mouse anti-mouse SMA (Sigma-Aldrich), followed by CY-2-conjugated donkey anti-mouse IgG (Jackson ImmunoResearch). Nuclei were depicted by 4′,6-diamidino-2-phenylindole (DAPI) staining (1 μg/ml, Sigma-Aldrich). Immunofluorescence was detected by fluorescent microscope (Olympus BX60, Olympus UK Ltd., London, UK).
Gene Expression Analysis
RT-PCR was performed as described (34). Total RNA was extracted using PerfectPure RNA Tissue Kit (5 PRIME, Gaithersburg, MD, USA), and reverse transcription followed using Verso cDNA kit (Thermo Scientific, Waltham, MA, USA). PCR primer sequences are listed in Table 1. Each sample was amplified by at least two different numbers of cycles to ensure that amplification was in the exponential phase of the PCR. Gene expression profile was analyzed by densitometry using NIH ImageJ software (http://rsbweb.nih.gov/ij/) and normalized to β-actin.
Primer Sequences
SMA, smooth muscle actin; CD31, cluster of differentiation 31; VEGF, vascular endothelial growth factor; Thbs1, thrombospondin-1; IL-12p35, interleukin-12 p35; CD40L, CD40 ligand; FasL, Fas ligand; VEGFR, VEGF receptor; PAI-1, plasminogen activator inhibitor; vWF, von Willebrand factor.
Aortic Ring Assay
The aortic ring assay was performed as described (4). Briefly, thoracic aorta was removed from anesthetized wild-type mice, mechanically cleaned from surrounding fat tissue, and sliced evenly into 1-mm rings. In a 48-well plate (Greiner, Nürtingen, Germany), each well was added a rounded drop of growth factor-reduced Matrigel (170 μl, BD Biosciences) that was allowed to solidify at 37°C. A single aortic ring was placed in the center of each well and covered with 170 μl Matrigel. The wells were completed with human endothelial serum-free medium (GIBCO, Carlsbad, CA, USA) supplemented with 2% fetal calf serum (FCS), 50 U/ml penicillin, and 50 μg/ml streptomycin (all from Biological Industries, Kibbutz Beit Haemek, Israel), together with 0.5 mg/ml human AAT or endothelial cell growth supplement (ECGS, 200 μg/ml, BD Biosciences) in sixplicates. Aortic rings were also cocultured with 20 freshly isolated islets, or WT and hAAT-transgenic mouse sera at a final dilution of 1:50, or directly in contact with primary lung epithelial cells from WT and hAAT-transgenic mice. After the wells were incubated at 37°C for a week, the rings were photographed and supernatants were collected for determination of VEGF levels by ELISA (R&D Systems, Minneapolis, MN, USA). Interleukin (IL)-1β, IL-6, monocyte chemoattractant protein-1 (MCP-1), tumor necrosis factor-α (TNF-α), and keratinocyte chemoattractant [KC; also known as chemokine (C-X-C motif) ligand 1 or CXCL1] levels were determined using Q-Plex mouse cytokine chemiluminescence-based ELISA (Quansys Biosciences, Logan, UT, USA).
Primary Mouse Lung Epithelial Cells
Lung epithelial cells were isolated as described (15), with minor modifications. Briefly, 4- to 6-week-old wild-type C57BL/6 and hAAT homozygote mice were anesthetized, and immediately after cervical dislocation, the lungs were inflated with collagenase through the trachea (1 mg/ml, type XI, Sigma-Aldrich). Whole lungs were then excised and incubated for 40 min at 37°C. Digested lungs were vortexed and filtered through a 100-μm nylon cell strainer (BD Falcon), and the pellet was washed in Hank's buffered salt solution (HBSS) containing 0.5% bovine serum albumin (BSA; both Sigma-Aldrich). The pellet was resuspended and seeded in M-199 medium (Biological Industries) supplemented with 10% FCS, 50 U/ml penicillin, and 50 μg/ml streptomycin. After 8 h, the medium was replaced in order to remove red blood cells and nonadherent cells. The epithelial nature of the resulting monolayer was verified by immunofluorescent staining for the broad epithelial marker pancytokeratin (sc-17843, Santa Cruz Biotechnology, Santa Cruz, CA, USA), and purity was determined as greater than 95%.
Tube Formation Assay
Human umbilical vein endothelial cells (HUVECs) were purchased from ScienCell (San Diego, CA, USA). Cells were seeded on T-75 flasks (Greiner) coated with poly-l-lysine (Sigma-Aldrich) at 5 × 103 cells/cm2. Endothelial cell medium (ECM, ScienCell) was replaced every 48 h until cells reached 90% confluence. Cells were collected by trypsinization for subsequent assaying (trypsin/EDTA, Sigma-Aldrich). For image analysis, cells were seeded in 96-well plates (Greiner) in 18-plicates coated with 50 μl Matrigel at 3 × 104 cells/well in Roswell Park Memorial Institute (RPMI 1640) medium (Sigma-Aldrich) supplemented with 1% FCS, 50 U/ml penicillin, and 50 μg/ml streptomycin. Endothelial cell growth supplement (ECGS, 200 μg/ml, Sigma-Aldrich) was added as positive control. For RT-PCR, cells were seeded in 24-well plates (Greiner) at 3 × 105 cells/well in 18-plicates and harvested 1.5, 3, 6, and 24 h later.
Statistics
Statistical analysis was performed using GraphPad Prism 5.0 software (GraphPad Prism, Pugh computers, Aberystwyth, UK). Results are presented as mean±SD. The significance of differences between groups was determined by Student's t test, at 95% confidence, for two groups/ conditions or by one-way ANOVA with a post-Bonferroni's multiple comparison test for three groups or conditions. Means were considered statistically different at p < 0.05.
Results
Islet-Driven Blood Vessel Formation Is Enhanced in the Presence of AAT
We first sought to determine whether AAT can facilitate mature blood vessel formation toward grafted islets in vivo.
In order to examine the formation of blood vessels that originate from the host and protrude toward engrafted pancreatic islets, 30 islets were embedded in 450 μl growth factor-reduced Matrigel that contained either no additive or biologically relevant concentrations of human AAT (0.5 mg/ml, n = 3 mice per group, repeated in three independent experiments) (Fig. 1). The grafts were introduced subcutaneously under the dorsal skin of recipient mice and allowed to solidify and to cultivate new blood vessels throughout the duration of 9 days. The grafts were then explanted, and the graft site and graft content inspected. As shown in Figure 1A, graft sites that belonged to Matrigelislet transplants that contained AAT exhibited a blood vessel network that appeared to have formed substantial connections between graft and host, as appreciated by the blood-red areas uncovered under the AAT-treated explants. On macroscopic inspection of the explants (Fig. 1B), diffused pattern was found in the control grafts, while a more distinct dark clump of blood vessels surrounded by a clear mass of Matrigel was observed in the AAT group. Matrigel plugs that contained no cells, with or without AAT, attracted no vessels upon explantation on day 9 and were completely clear (not shown). The appearance of hemorrhages in the implant sites of AAT-treated mice raised the concern that AAT might have promoted an anti-coagulatory state in the recipient, although it is described as a very weak anticoagulant (21). We confirmed that thrombin-induced mouse blood clotting is intact in the presence of AAT in a standard blood clot assay (not shown).
Pancreatic islet-driven blood vessel formation is enhanced in the presence of AAT. Subcutaneous Matrigel islet plugs containing 30 islets and supplemented with phosphate-buffered saline (PBS) (control, CT, n = 3) or human α-1-antitrypsin (AAT) (0.5 mg/ml, n = 3) were allowed 9 days to recruit blood vessels, while receiving IP saline (CT) or human AAT (AAT, 60 mg/kg) from day −1 and every 3 days after. (A) Day 9 transplantation sites. Representative photomicrographs and their anatomical localization imposed on a diagram and depicted by white dashed rectangle. (B) Day 9 Matrigel-islet plugs. Freshly retrieved representative explants. (C) Histology and immunohistochemistry. Control grafts (CT) exhibit fluid accumulation (top) and inflammatory cell infiltration (bottom) in two representative areas following hematoxylin and eosin staining; AAT-treated grafts exhibit intact islets (thick arrow), in proximity to cell-lined vessels (thin arrows) containing erythrocytes. A high-power image of a vessel and its positive staining for smooth muscle actin (SMA) is shown (bottom right). Representative experiment out of three independent experiments. 4′,6-Diamidino-2-phenylindole (DAPI) staining of the nucleus is shown.
According to histology performed on serial sample slides (Fig. 1C), 9-day-old control grafts contained signs of vessel rupture [i.e., fluid accumulation and non-cell-lined erythrocyte-filled pools (54)] as well as vast inflammatory infiltrates (CT, top and bottom, respectively). The appearance of mature blood vessels was rarely encountered in control samples. On the other hand, samples obtained from the AAT-treated grafts were predominated by mature blood vessels (top and bottom) and the association of less sites of fluid accumulation, non-cell-lined erythrocyte-filled pools, and inflammatory cell infiltrates. Vessel walls stained positive for SMA (bottom right) indicating mature vessels in the AAT-treated group.
The large diversity in graft appearance in the Matrigel grafts precluded quantification of vessel maturation. When assessing for number of vessels, wall thickness, distance from margin toward the center of grafts, content of red blood cell (RBC), density per field, existence of infiltration, edema, and hemorrhages, the groups provided a too large variation to allow quantification. In order to solve this issue, we modified the experimental setup to include “Directed In Vivo Angiogenesis Assay” (DIVAA), in which the grafts are enclosed in a fixed volume (Fig. 2). As shown, 20 μl Matrigel is contained within a uniform sterile silicon cylinder compartment that has a single opening (Fig. 2A, diagram in white). The cylinders were loaded with Matrigel containing 20 islets and either no additive, or 500 ng/ml FGF in the positive control group, or 0.5 mg/ml AAT in the experimental group (n = 10 per group).
AAT-treated islet grafts contain greater transcript numbers of endothelial cell markers and VEGF and reduced antiangiogenesis factors. Subcutaneous directed in vivo angiogenesis assay (DIVAA) cylinders containing 20 islets (illustration on left) were supplemented with PBS (CT, n = 6), fibroblast growth factor (FGF; 500 ng/ml, n = 9), or human AAT (0.5 mg/ml, n = 7) and allowed to recruit blood vessels for 9 days, while receiving IP saline (CT and FGF) or AAT (60 mg/kg) from day –1 and every 3 days after. (A) Day 9 grafts. Images of a cylinder before transplantation (day 0) and CT, FGF, and AAT cylinders after explantation on day 9. Arrows depict the opening of cylinders. Representative images. Gene expression profiles of grafts on day 9 that depict (B) endothelial markers and (C) angiogenic and antiangiogenic factors. CT set at 100%, mean±SD, *p < 0.05, and **p < 0.01. CD31, cluster of differentiation 31; Thbs1, thrombospondin 1; IL-12, interleukin-12; VEGF, vascular endothelial growth factor.
Upon explantation on day 9, grafts were photographed and cellular content was released for further examination. Consistent with the nonencapsulated Matrigel model (Fig. 1), cylinder content turned weak red after islets were allowed to generate vessels without supplemental therapy (Fig. 2A, CT). In the presence of FGF, a robust but limited hemorrhagic area was visible immediately adjacent to the opening of the cylinder (Fig. 2A, FGF). In the presence of AAT, however, there appears to have developed no hemorrhagic foci but rather a more uniform vasculature (Fig. 2A, AAT). Indeed, day 9 cylinders in the AAT group were more difficult to dissect away from the recipient animals as they were firmly connected to host tissue. Of note, plugs containing no cells and either with or without AAT were retrieved and displayed clear content with no invasive host elements (not shown).
In order to further quantify the differences between plugs in the presence of AAT, RT-PCR was performed on explanted silicone cylinder cell content with primers specific to relevant vessel neogenesis markers (Fig. 2B, C). Cells were released from Matrigel by digestion, and total RNA was extracted. As shown, in AAT-treated grafts, transcript levels of the endothelial cell marker CD31 were increased 1.26-fold from control, coinciding with elevated levels of SMA in the same group (2.10-fold from control). CD40 in the AAT-treated group, reported to be associated with blood vessel formation (73), was increased 1.29-fold from control.
As markers for vessel neogenesis predominated in AAT-treated grafts, we next examined factors specifically related to angiogenesis. The expression of two representative antiangiogenic factors that are transcriptionally controlled, thrombospondin 1 (Thbs-1) and IL-12, was significantly downregulated in AAT-treated grafts (Fig. 2C); transcript levels of VEGF, on the other hand, were elevated 1.47-fold from control. Other genes that were assessed did not yield significant changes between the conditions, including CD40 ligand (CD40L), CD14, Fas ligand (FasL), CD105, VEGF receptor (VEGFR), and insulin (not shown).
AAT Generated by Epithelial Cells Also Promotes Vessel Formation
It is possible that the beneficial effects of AAT are restricted to islet cells. Thus, to examine whether the outcome of AAT treatment is consistent across cell types, Matrigel implants were prepared with the cellular content replaced by primary mouse lung epithelial cells that were isolated from lungs of either wild-type mice or from mice transgenic for the human AAT (hAAT) construct (17) (Fig. 3). With keeping the experimental design consistent with the initial experiment performed in this study, grafts containing primary epithelial cells were placed under the skin of recipient mice, and vascularization was evaluated after 9 days.
Epithelial cell-driven blood vessel formation is enhanced in the presence of endogenous transgenic human AAT. Primary mouse lung epithelial cells (2 × 105) that were isolated from either wild-type mice or from mice that express a surfactant promoter-controlled sequence of hAAT were mixed in Matrigel and grafted subcutaneously. (A) Day 9 graft histology. Fluid accumulation (top) and non-cell-lined erythrocyte accumulation (bottom) in wild-type epithelial cells and cell-lined vessels in hAAT-expressing epithelial cells. (B) Day 9 gene expression profile. Antiangiogenic gene products, CT set at 100%, mean±SD, **p < 0.01, and ***p < 0.001. PAI, plasminogen activator inhibitor.
A highly consistent outcome was achieved by hAAT-expressing epithelial cells, as depicted by histological samples; WT epithelial cells produced edematic pools and hemorrhages (Fig. 3A, left), while AAT-treated plugs resulted in enclosed vessels (Fig. 3A, right). Gene expression profile was consistent with an expression profile of proangiogenic character (Fig. 3B). Thus, the attraction of vessels toward grafts is not a unique attribute of the pancreatic islet.
Aortic Ring Vessel Sprouting Is Enhanced by AAT and by Epithelial Cells That Express hAAT
To examine whether the response to AAT is restricted to changes in the grafted cells (e.g., increased graft viability or alteration of molecular release profile), as opposed to isolated changes in blood vessel-related cells, we performed the aortic ring assay in the presence of AAT and in the absence of islets.
As shown in Figure 4A, aortic rings were embedded in Matrigel domes, and medium was added that was supplemented with either no additive (negative control, CT), ECGS (positive control), or AAT. Similarly, aortic rings were cultured in the presence of no additive, but rather with primary lung epithelial cells from either wild-type mice or from mice that express hAAT (Fig. 4B). Sprouting vessels were photographed after 7 days, and supernatants were assayed for inflammatory release factors and for VEGF.
Aortic ring vessel sprouting is increased in the presence of hAAT and in the presence of transgenic hAAT-expressing epithelial cells. Aortic rings were seeded in sixplicates in the presence of either (A) medium alone (CT), positive control [endothelial cell growth supplement (ECGS) 0.2 mg/ml] or hAAT (0.5 mg/ml). In addition, aortic rings were seeded in the presence of (B) 2.5 × 104 wild-type lung epithelial cells or hAAT-expressing lung epithelial cells. Images were acquired on day 12; three representative images are shown per condition. Arrows depict examples of sprouting vessels. (C) Supernatant analysis. Two chemokines and VEGF levels were determined after 6 days of culture. CT set at 100%, mean±SD, **p < 0.01, and ***p < 0.001. tg, transgenic; KC, keratinocyte chemoattractant [also known as chemokine (C-X-C motif) ligand 1; CXCL1]; MCP-1, monocyte chemoattractant protein-1.
As shown in Figure 4A and B, aortic rings displayed none or very low sprouting in the control groups (whether CT or cocultures of WT epithelial cells). However, prominent sprouting was observed in the presence of ECGS, as well as in the presence of AAT, whether its source was clinical grade AAT (Fig. 4A, AAT) or endogenous epithelial cell-derived transgenic AAT (Fig. 4B, hAAT-tg).
Importantly, the environment attained by AAT was low in inflammatory markers ({keratinocyte chemoattractant [KC; also known as chemokine (C-X-C motif) ligand 1 or CXCL1]} protein levels were reduced 5.63-fold, and MCP-1 protein levels were reduced 5.64-fold) and high in VEGF levels (1.77-fold more in the presence of AAT) (Fig. 4C). Other inflammatory markers were below detection [IL-1β, TNF-α, interferon-γ (IFN-γ), and IL-6, not shown].
Endothelial Cells Respond to hAAT by Tube Formation and a Proangiogenic Gene Expression Profile
Since vessels appeared to develop in a superior manner in the presence of AAT while driven by either islets, epithelia, or aortic rings, we turned to examine isolated endothelial cell behavior. In the endothelial cell tube formation assay, primary endothelial cells rapidly formed tubes upon culture without added stimulants (Fig. 5A, CT) and, to a larger extent, in the presence of the positive control, ECGS. When endothelial cells were added with AAT, however, the tubes that were formed appeared to be more pronounced than the control, in that the area that they covered was larger (quantification presented in Fig. 5B), the number of tubes was greater, and the number of enclosed areas within the tubes was higher than the control cells.
Tube formation assay enhanced in the presence of AAT. Tubes formed after seeding of human umbilical vein endothelial cells (HUVECs), in the presence of either ECGS or AAT, were analyzed. (A) Photomicrographs of endothelial cultures. Representative images of 18-plicates acquired after 24 h of culture. (B) Grading of tube formation. (C) Gene expression. RT-PCR performed on total RNA extracted from 3- and 24-h cultures. CT set at 100%, mean±SD, *p < 0.05, **p < 0.01, and ***p < 0.001. vWF, von Willebrand factor.
We performed RT-PCR in order to closely observe changes in the transcription profile under the different culture conditions. Expression of β-actin was determined to represent positive changes in endothelial cell population size; as shown in Figure 5C, 3 and 24 h after seeding (top and bottom panels, accordingly), the population of endothelial cells had grown significantly in the presence of AAT, exceeding that which was achieved by ECGS alone. VEGF expression exhibited an early positive response to AAT, as soon as 3 h after seeding (2.80-fold, AAT vs. CT), and also after 24 h (3.89-fold, AAT vs. CT), as shown after normalized to β-actin readouts. The maturation marker VWF was similarly elevated within the first 3 h and also after 24 h (1.56-fold from control by ECGS and 8.82-fold from control by AAT). Importantly, changes in VEGF expression appeared to precede changes in population size.
Discussion
In type 1 diabetes, pancreatic islet β-cells undergo an immune attack and suffer significant damage, culminating in insulin deficiency [reviewed in van Belle et al. (65)]. The current therapeutic approach consists of exogenous insulin replacement, affording glycemic control, albeit suboptimal. Islet transplantation is currently evaluated in multiple centers as a therapeutic option for individuals with significant islet loss. Successful human islet grafts render the recipients normoglycemic and insulin independent while under an immunosuppressive protocol. Aside from rapid islet loss within the first days after engraftment, islet grafts appear to expire and loose function between the first and fifth year in the striking majority of diabetic recipients.
While potentially beneficial for angiogenesis, inflammation is critically injurious to islets. The damage to islet β-cells is, in fact, prevalent as early as pancreas procurement, whereby donor islets endure isolation-related stress that includes hypoxia, enzymatic digestion, mechanical stress, and endogenous inflammation. The mere severing from donor vasculature serves as a stress for islet β-cells, and processes such as apoptosis and necrosis rapidly ensue, facilitating graft immunogenicity. Rapid revascularization is, therefore, beneficial to human islet grafts, provided that it is afforded during conditions that are noninflammatory and are islet protective.
α-1-Antitrypsin is the most common serine-protease inhibitor in the circulation. Patients with AAT deficiency receive plasma-derived affinity-purified hAAT once a week for life, a course established as safe and, indeed, protective of pulmonary tissue (42,49). The rationale for examining the potential benefit for AAT toward better islet vascularization stems from a collection of recent studies. AAT has been shown to induce VEGF expression (48,62,75) and to prevent elastase from executing an enzymatic cleavage of mature VEGF (31), as well as facilitate smooth muscle myocyte migration and proliferation (24), while reducing apoptosis in endothelial cells and smooth muscle cells (47,48,62,63). In AAT deficiency, individuals are at high risk to develop emphysema, which can be mimicked in animal models by blockade of the VEGF pathway in the presence of normal local levels of AAT (19,27,28), suggesting a tight connection between AAT and VEGF, and between VEGF in AAT deficiency-related emphysema. Indeed, AAT promoter contains a hypoxic response element that binds HIF-1α, and the expression of AAT is upregulated during hypoxia (70). Interestingly, AAT has been shown to be produced by human islet cells in the presence of inflammatory mediators (5). Taken together, we hypothesized that AAT may play a role in the positive angiogenic outcomes of inflammation by its own local upregulation as a gene that bares responsive elements for IL-1β (5), its promotion of VEGF production, and its blockade of antiangiogenic factors.
Previous findings in our lab have established that islet allografts draw mature blood vessels in mice during long-term AAT therapy, as achieved by the use of mice transgenic for human AAT (34). Thus, we sought to specifically examine the effect of AAT on islet graft-derived angiogenesis and include the hAAT transgenic mice in the experimental systems herein. For example, we used primary epithelial cell grafts derived from lungs of mice transgenic for human AAT. In this system, AAT appears to have enhanced the production of VEGF and inhibited the expression of antiangiogenic factors, resulting in improved maturity of blood vessels toward islet grafts and epithelial cell implants, as well as increased vessel protrusion from aortic rings in the absence of islets. Angiogenesis in islet grafts and in epithelial cell implants is evident by an increase in the transcript levels of CD31, CD40, and SMA; although the presence of CD40 alone is not indicative of the cell type in these conditions, the low inflammatory infiltrate that accompanied AAT treatments here and in other reports (30,35) suggests that the source of CD40 transcripts is primarily nonimmune. Macroscopically, grafts that were treated with AAT appeared to contain less edematic and hemorrhagic pools, suggestive of more mature and non-leaking vessels (54). As grafts were extracted and examined for gene expression profile, we found higher expression of VEGF, eNOS, and VWF and reduced expression of IL-12 and Thbs1, altogether a proangiogenic panel of changes (1).
We chose to perform two distinct vascularization models, namely, free Matrigel plugs and fixed-silicon Matrigelcontaining plugs. The latter method provided smaller intraassay variance, particularly with regard to gene expression changes, yet was less favorable in as far as the processing for histology. These outcomes corroborated the findings obtained in the free Matrigel plug method, in which variability compromised statistical significance. The use of RT-PCR to examine changes in these systems granted extensive use of small amounts of material yet limited the study to factors that are governed by mRNA changes. Thus, it would be interesting to expand this study to include the examination of activated factors that might exhibit changes in the presence of AAT.
The findings in vivo were supported by in vitro experiments. The rationale for in vitro extension of the work was to allow better control over the cell types examined and also to evade some variability inherent to whole animal models. In addition, the study of endothelial cells or aortic rings excluded direct protective effects that AAT has on islets, which may have afforded islet-containing systems relative benefit. Knowingly, these in vitro assays do not represent vasculogenesis but rather focus on more specific aspects of this complex process. We recently published a protocol for the aortic ring assay (4). Here, adding human AAT to aortic rings resulted in vessel sprouting and in elevated levels of VEGF in the culture media; a similar outcome was demonstrated in cultured HUVECs, which were further analyzed for gene expression at various time points. Consistent with the findings, thus far, HUVECs displayed increased expression of VEGF and VWF. As expected, β-actin was uniform between treatment conditions at 3 h of culture; however, after 24 h, AAT-treated cells exhibited twofold greater total β-actin transcript numbers, superseding the effect obtained by the positive control, that is, ECGS. Interestingly, the elevation in VEGF expression preceded the elevation in VWF expression, and both genes appeared to be more pronounced in AAT-treated cells over ECGS-treated cells. The overall result consisted of tube formation by HUVECs that covered a greater area, enclosed more spaces, and were larger in number in AAT-treated cultures.
Importantly, it was essential to address the inflammatory environment under study conditions. It is rather widely accepted that inflammation is required for angiogenesis. Here, we present angiogenic advances in the presence of an anti-inflammatory profile. Our results indicate that at least two inflammatory markers, the chemokines KC and MCP-1, are reduced in aortic ring assay media in the presence of AAT, at the time that VEGF is elevated in the very same samples. hAAT is also known to reduce IL-1β, IL-6, TNF-α, and MCP-1 in multiple reports and across cell types and experimental systems (34,35,61,69) while elevating the production of IL-1 receptor antagonist (34,57,61) and IL-10 (34,46). This important outcome suggests that it is possible to generate an increase in VEGF and to facilitate blood vessel maturation, while various aspects of inflammation are suppressed. Hypoxic cells are known to attract blood vessels without inflammation (71), and macrophages are known to facilitate angiogenesis while differentiated into the noninflammatory M2 subtype (72). The anti-inflammatory attributes of AAT are extensively described in literature and include studies that particularly regard islet transplantation; being an acute phase reactant, AAT functions in situations where tissue preservation is required, excessive inflammation necessitates interventionary anti-inflammatory molecules, and tissue recovery, including vascular advancements, are critical. For islet graft survival, this is a beneficial environment, as is the blockade of local inflammation.
Tumors require angiogenesis. The trigger in this case involves most probably the lack of local oxygen as the size of the tumor mass exceeds 1–2 mm3, driving the production of VEGF and its receptors to allow blood vessel formation toward the tumor cells. The tumors thus promote local injury and a rise in inflammatory mediators that further advance blood vessel formation (44,68). Tumors also require matrix metallopeptidase 12 (MMP-12) in order to spread; the activity of this metaloprotease was recently shown to be inhibited by AAT (12). Thus, it is not unexpected that AAT inhibits angiogenesis in the vicinity of tumors, as well as inhibits tumor growth (23). Indeed, individuals that receive AAT over extended periods of time at doses that exceed normal plasma values for AAT exhibit reduced frequency of lung cancer (74).
Several studies indicate that AAT exhibits an antiapoptotic activity. For example, AAT was shown to inhibit apoptosis in endothelial cells (47), smooth muscle cells (48), and islet β-cells (35,76), as well as hepatocytes (66) and lung alveolar epithelial cells (48), perhaps by direct intracellular uptake and subsequent inhibition of caspase-3 (47). It is, therefore, possible that the benefit of AAT during angiogenesis may be attributed to protection from cell death afforded to particular cell types such as, in the case of vascularization, endothelial cells and smooth muscle cells. Added to the protection that AAT provides islet β-cells, it is possible that it is not angiogenic but rather allows the systems studied here benefits that result in improved cell survival and mature protrusion of new vessels. It would be interesting to assess the impact of AAT in culture on islet cell and intraislet endothelial cell recovery after isolation, which in turn may contribute to faster/improved neovascularization upon transplantation.
The impressive safety record of long-term administration of human AAT to individuals with varying degrees of deficiency (42,49) supports its use in indications such as cell transplantation (33,59,69). In particular, this approach is important when considering cells as sensitive to inflammation as pancreatic islet β-cells, grafted in a procedure, which currently combats poor multicenter transplantation outcomes and is rendered inherently limited by the fact that sirolimus and tacrolimus are inhibitory for angiogenesis (9,13,29) and are islet β-cell toxic (3,9). The effects of AAT under conditions that include mammalian target of rapamycin (mTOR) inhibitors should be assessed in extended experiments, addressing both β-cell toxicity and angiogenesis. Nonetheless, with the addition of enhanced maturation of blood vessel toward grafted islets, one may conclude that the array of activities exerted by AAT call for its extended use in relevant clinical setups.
Footnotes
Acknowledgments
we thank Mrs. Valeria Frishman for her excellent technical assistance. This work was supported by Juvenile Diabetes Research Foundation (2-2007-103) and Israel Science Foundation (1027/07). The authors declare no conflict of interest.