Revascularization of swine renal artery stenosis improves renal function but not the changes in vascular structure

Abstract

Renal revascularization with percutaneous transluminal renal angioplasty improves blood pressure and stenotic-kidney function in selected groups of patients, but the reversibility of intra-renal and microvascular remodeling remains unknown. This study tested the hypothesis that renal angioplasty improves the function and structure of the renal microcirculation in experimental chronic renal artery stenosis. Stenotic kidney function, hemodynamics and endothelial function were assessed in-vivo in pigs after a 10-weeks sham-treated unilateral renal artery stenosis, similar stenosis that underwent angioplasty and stenting 4 weeks earlier, or sham operated animals. Renal microvascular remodeling, angiogenic pathways, and fibrosis were investigated ex-vivo. Renal angioplasty decreased blood pressure, improved glomerular filtration rate and microvascular endothelial function, promoted the expression of angiogenic factors, and decreased renal apoptosis induced by renal artery stenosis. However, the spatial density of renal microvessels was partially improved after angioplasty, and renal blood flow was incompletely restored compared to sham-treated kidneys, as was interstitial fibrosis. Renal microvascular media-to-lumen ratio remained unchanged by angioplasty. The current study shows that revascularization of renal artery stenosis restores glomerular filtration rate and renal endothelial function four weeks later, while renal hemodynamics and structure are incompletely restored.

Introduction

Renal artery stenosis (RAS) induced mainly by atheromatous disease can accelerate hypertension and renal failure (1) and leads to adverse cardiovascular outcomes. The main treatment options for RAS include medical therapy and/or revascularization of the stenotic renal artery aimed to improve blood pressure control and preserve renal function.

The role of percutaneous transluminal renal angioplasty (PTRA), followed by stenting of the revascularized renal artery and optimal medical therapy, remains controversial. While restoration of kidney perfusion should restore renal function and improve blood pressure control, several prospective, randomized trials up to now fail to identify compelling benefits for cardiovascular or kidney outcomes (2). Notably, most studies in humans evaluating the response to PTRA are limited to changes in serum creatinine, glomerular filtration rate (GFR), blood pressure, or number of medications (3, 4).

Obstruction of flow in the renal artery initiates a complex cascade of events that remains incompletely understood. Using a swine model of progressive RAS, we have previously shown that the kidney exposed to chronic RAS shows functional deterioration accompanied by renal inflammation, fibrosis, and microvascular rarefaction and remodeling (5-8). Diminished intra-renal microvascular density may be the result of either altered or insufficient angiogenesis with or without acceleration of apoptosis. It is associated with impaired renal function and structure (8). However, little is known about the ability of revascularization to restore the renal microcirculation and attenuate microvascular injury.

The aims of this study were to explore the effect of PTRA and stenting on intrarenal microvascular remodeling, angiogenesis, and fibrosis. To test the hypothesis that PTRA can modify the function and structure of the renal microcirculation in the stenotic kidney, we utilized a large animal model of chronic experimental RAS. The pigs underwent revascularization with PTRA and stenting after 6 weeks of unilateral RAS, a duration associated with development of renovascular hypertension, kidney dysfunction, and renal fibrosis (9-11). They were studied in vivo 4 weeks later, to allow for detectable improvement of renal and endothelial functions (12, 13).

Results

Systemic and hemodynamic parameters

Six weeks after coil implantation, hemodynamically significant RAS (>60%) was present in all pigs (Figure 1A-B). PTRA was successfully performed in 6 pigs (degree of stenosis before PTRA 68.3±4.4%), and followed by a decrease of MAP to the baseline level (Figure 1C), whereas the remaining RAS pigs (degree of stenosis 71.5±6.3%, p=NS) had sham angiography. Four weeks later, the sham-treated RAS group still had a significant stenosis (75.3±5.9%) and elevated systolic blood pressure (p<0.05 vs. Sham), and MAP tended to increase as well (p=0.068), while in RAS+PTRA the renal artery was patent with a minimal residual stenosis (Table 1). Blood pressure after PTRA was lower than untreated RAS and not different from Sham (Table 1). Creatinine levels slightly increased in RAS and were lower after PTRA, while PRA was not different among the groups (Table 1).

Figure 1.

Renal angiography in a pig with renal artery stenosis (RAS) before (A) and 4 weeks after (B) revascularization with percutaneous transluminal renal angioplasty (PTRA) and stenting. (C) Representative evolution of mean arterial pressure measured using telemetry in a RAS pig after implantation of a local-irritant coil. RAS increased blood pressure, which rapidly returned to baseline levels after PTRA. D. Single-kidney renal blood flow (RBF) and Glomerular filtration rate (GFR) at baseline (Bsl) and in response to acetylcholine (Ach). * p<0.05 vs. Sham, ^Ψ p< 0.05 vs. Bsl.

Table 1.

Systemic characteristics and basal single-kidney hemodynamics (mean ± SEM) in sham, sham-treated renal artery stenosis (RAS), and RAS pigs 4 weeks after PTRA.

	Sham	RAS	RAS + PTRA
Number of kidneys	8	7	6
Body weight (Kg)	47.5 ± 1.5	56.0 ± 3.0	54.5 ± 3.4
Diastolic Blood Pressure (mmHg)	88.3 ± 4.0	107.6 ± 8.3	94.5 ± 8.1
Systolic Blood Pressure (mmHg)	111.5 ± 3.5	139.3 ± 9.6*	122.3 ± 10.3
Mean arterial pressure (mmHg)	96.0 ± 3.8	118.1 ± 8.6	103.8 ± 8.9
Degree of stenosis (%)	0	75.3 ± 5.9*	5.0 ± 3.4§
PRA, ng/ml/hrs	0.25 ± 0.04	0.22 ± 0.05	0.25 ± 0.14
Creatinine, mg/dl	1.15 ± 0.09	1.5 ± 0.08*	1.28 ± 0.12
Basal RBF, ml/min	584.8 ± 17.118	271.0 ± 84.5*	386.1 ± 39.5*
Basal GFR, ml/min	70.4 ± 2.4	32.7 ± 10.0*	58.2 ± 8.0
Ach-RBF, ml/min	663.8 ± 30.9 Ψ	334.0 ± 106.9*	539.7 ± 62.4 Ψ
Ach-GFR, ml/min	91.9 ± 2.8 Ψ	44.8 ± 14.5*	72.4 ± 10.6 Ψ
Renal volume, ml	123.9 ± 3.7	79.9 ± 21.0	114.1 ± 11.2
Cortical Perfusion (ml/min/cc tissue)	4.96 ± 0.11	3.85 ± 0.43	3.78 ± 0.23*

PRA: plasma renin activity; RBF: renal blood flow; GFR: glomerular filtration rate; Ach: acetylcholine.

^*p<0.05 vs. Sham,

^Ψp<0.05 vs. basal values,

^§p<0.05 vs. RAS.

Renal hemodynamic and function

Ten weeks after induction of RAS basal GFR and renal blood flow (RBF) were significantly lower in the RAS kidney compared to Sham, and the response to Acetylcholine (Ach) was attenuated (Figure 1D), indicating endothelial dysfunction. In the RAS+PTRA group, GFR was improved and not different from Sham, while RBF was not significantly higher than RAS, and remained lower compared to Sham (p<0.05). Nevertheless, both GFR and RBF responses to Ach were improved (p=0.016 and p=0.001 vs. baseline, respectively), suggesting an improvement of endothelial function after PTRA. The slight decrease in renal volume in RAS was partially reversed by PTRA but this has not reached statistical significance due to a high variability (Table 1). Cortical perfusion, which tended to decrease in RAS, reached statistical significance in RAS + PTRA group (p<0.01 vs. sham, Table 1) and was not different between the two RAS groups.

Microvascular 3-D architecture and remodeling

Overall, transmural spatial density of renal microvessels across the cortical regions was lower in RAS (p<0.01 vs. Sham) and improved in RAS+PTRA (Table 2, Figure 2A-B, p>0.05 vs. Sham), while average vessel diameter was unchanged (p=0.7). Microvascular tortuousity tended to increase in both RAS and RAS+PTRA (p=0.12), suggesting angiogenic activity to restore microvessels (Table 2). In addition, small microvessels with a diameter under 40μm were reduced in RAS (p<0.001 vs. Sham) and improved by PTRA (p>0.05 vs. Sham). The density of 40-100μm microvessels tended to be reduced in RAS and RAS+PTRA (p=0.073 vs. Sham), while larger microvessels were unaffected in either group (Figure 2).

Table 2.

Renal cortical microvascular architecture assessed by micro-CT in sham, sham-treated renal artery stenosis (RAS), and RAS + PTRA pigs.

	Sham	RAS	RAS + PTRA
Spatial density (vessels/mm2)	2.70 ± 0.24	1.50 ± 0.15*	1.97 ± 0.06
Average vessel diameter (μm)	91 ± 4	102 ± 12	95 ± 9
Tortuousity (ratio)	1.30 ± 0.04	1.43 ± 0.05	1.51 ± 0.12

Values are means ± SEM

^*p<0.05 vs. Sham.

Figure 2.

Representative three-dimensional tomographic images of the cortical microcirculation in sham, RAS, and RAS + PTRA pigs (A), and spatial density (B, mean ± SEM) of different size microvessels. PTRA slightly increased intra-renal density of small microvessels. *p<0.05 vs. Sham.

The expression of VEGF (p<0.05 vs. Sham) and Flt-1 (p<0.05 vs. sham) increased in RAS, and Flk-1 tended to increase as well (p=0.1, Figure 3). In RAS + PTRA, VEGF increased further (p<0.01 vs. RAS) while Flt-1 and Flk-1 both decreased to normal levels (Figure 3). In addition, angiopoietin-1 expression was upregulated after PTRA (p<0.01 vs. RAS, Figure 3), while eNOS was unchanged. Moreover, MMP-2 was increased in RAS (p<0.05 vs. Sham, Figure 3), but further elevated in RAS +PTRA (p<0.001 vs. RAS, Figure 3) while the expression of its inhibitor TIMP-2 was similarly blunted in both groups (p<0.01 vs. Sham). The expression of tTG in RAS and RAS + PTRA was not significantly different from Sham (p=0.18, Figure 3). Increase in microvascular media to lumen ratio in RAS and RAS + PTRA groups compared to Sham (p<0.001 vs. Sham) indicated microvascular remodeling that was not reversed four weeks after PTRA (Figures 4).

Figure 3.

Immunoblots (upper panel) and densitometric quantification (bottom) demonstrating renal protein expression of VEGF, Flt1, Flk1, angiopoietin-1, eNOS, pro-MMP2, TIMP2 and tTG in Sham, RAS, and RAS + PTRA groups. PTRA increased VEGF and angiopoietin-1 expression, suggesting a proangiogenic milieu in these kidneys. Mean ± SEM, * p< 0.05 vs. Sham, § p<0.05 vs. RAS.

Figure 4.

A. Representative renal α-SMA staining in Sham, RAS, and RAS + PTRA kidneys. B. Quantification of media to lumen ratio (mean ± SEM) in the same groups. * p<0.001vs Sham.

Renal tissue remodeling

RAS showed increased renal apoptosis (fraction of TUNEL positive cells) compared with Sham (p<0.01, Figure 5), in association with increased expression of activated caspase-3. Both indices of apoptosis were significantly attenuated after PTRA, although activated caspase-3 expression remained higher than in Sham (Figure 5).

Figure 5.

Caspase 3 (A, brown) and TUNEL (B, green) staining and quantification (C), showing increased number of positive (apoptotic signal) cells in RAS, which was decreased after PTRA. *p<0.01 vs. Sham; § p<0.01 vs. RAS.

In addition, trichrome staining (indicating renal fibrosis) increased in RAS (p<0.01 vs. Sham, Figure 6), and was unaffected by PTRA (p<0.05 vs. Sham, Figure 6). Moreover, cortical hydroxyproline content, reflecting collagen content, was elevated in the RAS group (16.4±3.6 vs. 5.3±0.8 ug/10 mg, p= 0.01 vs. Sham), and decreased but not was normalized after PTRA (8.5±0.4 μg/10 mg, p= 0.018 vs. RAS, p=0.002 vs. Sham, Figure 6). There were no detectable differences in the expressions of TGF-β and TIMP 1 among the groups (p=0.30 and p=0.39, respectively, Figure 6), but connective tissue growth factor (CTGF) expression tended to increase in RAS group and was statistically significantly elevated in RAS + PTRA group (p<0.05 vs. Sham, Figure 6). The expression of xanthine oxidase increased in the RAS kidney (p<0.05 vs. Sham) and decreased to sham levels after PTRA (p<0.05 vs. RAS, Figure 6), suggesting attenuation of oxidative stress induced by RAS.

Figure 6.

Representative images of trichrome (A) and quantification (B) of trichrome and hydroxyproline (mean ± SEM) in Sham, RAS, and RAS + PTRA groups. Immunoblots (C) and densitometric quantification (D) demonstrating renal protein expression of TGF beta, CTGF, TIMP1, and xanthine oxidase in Sham, RAS and RAS + PTRA, mean ± SEM, * p<0.05 vs. Sham, § p<0.05 vs. RAS.

Discussion

This study demonstrates that PTRA performed after 6 weeks of experimental renal artery stenosis partially restored GFR and renal endothelial function. These changes were associated with enhanced proliferation and maturation of new vessels, but renal perfusion, vascular wall remodeling, and interstitial fibrosis remained incompletely restored. Our study therefore demonstrates that PTRA in swine RAS can only partially restore the renal microcirculation. Importantly, these observations suggest that meaningful improvement in GFR could be achieved without complete reversal of renal tissue remodeling.

Our results extend clinical observations in human subjects with atherosclerotic renal artery disease. Hemodynamically significant RAS (>60% decrease in luminal diameter) is detected in almost 7% of individuals older than 65 years of age (14). RAS regularly accelerates renovascular hypertension, renal dysfunction, and increased cardiovascular mortality. Revascularization of the kidney using PTRA or other means can lead to a decrease in blood pressure, and in turn improvement in systemic endothelial function and oxidative stress (12). Remarkably, renal function improves only in a minority of patients with atherosclerotic RAS (1). However, the reasons for incomplete recovery of renal function in RAS remain obscure (2).

In our experimental swine model, arterial pressure increased within a few days after implantation of a local irritant coil as a result of a significant RAS. We have previously shown that this is associated with a decrease in GFR and RBF (9, 15) and with renal endothelial dysfunction (10, 16). Plasma renin activity is not necessarily elevated, as typical for a chronic phase of untreated renovascular hypertension (17), implicating recruitment of alternative pressor mechanisms that maintain hypertension (1). Our current studies with effective PTRA induced a rapid and stable decrease of arterial pressure, and by 4 weeks after PTRA, renal function (GFR and serum creatinine) improved as well. RBF improved only slightly, albeit not significantly, but its response to acetylcholine was restored, suggesting improvement of endothelial function that was impaired in RAS. Notably, the reduction in pressure distal to RAS leads to decreased vascular tone due to compensatory dilatation, which might attenuate the extent of microvascular responsiveness to Ach. Hence, improved RBF and GFR responses to Ach might be partly secondary to restoration of basal vascular tone rather than endothelial function. Nevertheless, the greater improvement in MAP than RBF suggests that vascular resistance distal to the stenosis remained elevated.

We have previously shown that deterioration of renal hemodynamic and function in our RAS model is accompanied by oxidative stress, apoptosis, fibrosis, and inflammation (6, 7). In the current study we observed that 4 weeks after PTRA, GFR and endothelial function improved, possibly secondary to the decrease in oxidative stress as reflected in downregulation of the expression of the radical producing enzyme xanthine oxidase, thought to play an important role in renal ischemic injury (18). Renal cell apoptosis was also reduced, suggesting reversal of renal “hibernation” (16) in a kidney with preserved regenerative capacity (19). Nevertheless, renal fibrosis induced by RAS was only partially improved after PTRA. Since TGF-beta and TIMP-1 were only slightly and not significantly upregulated by RAS, additional pro-fibrotic factors likely mediated renal fibrosis in this model. Indeed, trichrome staining and the hydroxyproline assay showed an increase in renal collagen content induced by RAS. Moreover, CTGF expression further increased after PTRA, trichrome staining was unaffected, and collagen content was not fully restored by PTRA, suggesting that the scarring process was incompletely arrested and that extracellular matrix deposited was not cleared from the kidney.

Furthermore, we have shown before that deterioration of renal function in RAS is associated with microvascular rarefaction and remodeling (5-8). We suggested that angiogenic competence might be linked to renal functional recovery (5). Neovascularization involves a sequence of events such as endothelial cell proliferation, migration, and differentiation, remodeling of extracellular matrix, and functional maturation of the newly assembled vessels, which are often mediated by VEGF (20). While the stenotic kidney showed elevated expression of VEGF, microvascular density was decreased. This might be partly due to the increased renal expression of the VEGF receptor Flt1, which can sequester VEGF to decrease its activity and interfere with neovascularization (21). Interestingly, we found that PTRA led to an increase of VEGF expression and microvascular density, especially of renal vessels <40μm in diameter, which was accompanied by a slight increase of tortuousity characteristic of angiogenic vessels, implying enhanced angiogenic efficacy.

The PTRA-induced neovascularization was likely mediated by interaction among several growth factors acting in concert. A decrease in Flt-1 might have contributed to an increase in effective VEGF availability, while the increase in renal expression of angiopoietin-1 in RAS + PTRA mediated maturation of the new vessels (22), and the interaction between VEGF and angiopoietin-1 fostered microvascular stabilization. MMP-2 was increased in RAS + PTRA and facilitated angiogenesis by degrading surrounding extracellular matrix and allowing endothelial cell invasion (23). Taken together, our results indicate an important increase in pro-angiogenic milieu in the RAS kidney after PTRA, albeit with only partial restoration of the microvessels lost in RAS. These observations suggest that augmentations of the pro-angiogenic milieu in the post-stenotic kidney may offer the potential for new vessel formation in patients undergoing endovascular procedures.

It is important to recognize that microvascular vessel wall remodeling was not changed after PTRA and stenting, as indicated by the elevated media/lumen ratio. Moreover, the increased fraction of TUNEL positive renal cells in RAS suggested increased apoptosis, as previously described (7), which was underscored by activation of caspase-3, an important effector of apoptosis. PTRA diminished both TUNEL staining and activated-caspase-3 expression, implying attenuation of the apoptotic process by PTRA. Nevertheless, caspase-3 expression did not return to Sham levels after PTRA, both renal interstitial fibrosis and microvascular remodeling were not restored, and neovascularization was incomplete. This study therefore demonstrates curtailed structural recovery 4 weeks after PTRA, which may in part account for the mere partial restoration of RBF in the stenotic kidney. In most patients with non-atherosclerotic RAS, technically successful PTRA is followed by a prompt decrease in blood pressure within 1-6h; in 93% it has stabilized within 24h, and remains stable at 6 months (24). Indeed, we observed that blood pressure in RAS started decreasing shortly after PTRA. We would expect to observe improvement in renal function by 4 weeks after PTRA, because an increase in GFR in human RAS has been documented within 5 days after angioplasty (13), but may occur even earlier in parallel to MAP. Furthermore, a recent study in patients with RAS showed significant improvement in endothelial function and a decrease in blood pressure and oxidative stress by one month after angioplasty (12). Therefore, our observation that 4 weeks after PTRA GFR increased while RBF and the renal microcirculation remained impaired, likely bears clinical relevance.

Our study was limited by the use of pigs without co-morbid conditions, and by a short duration of RAS relative to the human disease. Human RAS is multi-factorial and particularly dependent on the severity, evolution, and duration of the stenosis, as well as on other concurrent or pre-existing pathophysiological conditions like essential hypertension, diabetes, or hypercholesterolemia, which impair the renal microvasculature and likely modulate its response to revascularization. Nevertheless, renal structure and function in the swine model are similar to human kidneys, and our results bear relevance and may shed light on the reversibility of renal injury following revascularization.

In conclusion, the present study shows that PTRA in chronic porcine RAS partially restored the renal microvascular network and allowed GFR recovery. However, our study suggests that the improvement in GFR and decrease in blood pressure 4 weeks after PTRA may conceal residual structural damage even in the post-stenotic kidney uncomplicated by atherosclerosis. These observations suggest that the improvement in renal function can precede or be achieved without complete eradication of renal scarring. Future studies will need to determine whether such residual damage might resolve, interfere with renal functional reserve, or predispose the kidney for future injury.

Methods

Experimental protocol and MDCT analysis

The institutional Animal Care and Use Committee approved all the procedures. Twenty one female domestic pigs (45-68 kg) were studied after 10 weeks of observation. At baseline, all the pigs were anesthetized with an intra-muscular injection of telazol (5 mg/kg) and xylazine (2 mg/kg), intubated, and mechanically ventilated with room air. Anesthesia was maintained with a mixture of ketamine (0.2 mg/kg/min) and xylazine (0.03 mg/kg/min) in normal saline (0.05 μl/kg/min). In thirteen pigs, RAS was induced at baseline by a local-irritant coil placed in the main renal artery, inducing gradual development of unilateral RAS by 5-7 days, as described previously (9, 25, 26). A PhysioTel® telemetry system (Data Sciences International, Arden Hills, MN) was implanted on the same day in a femoral artery to continuously monitor mean arterial pressure (MAP) in all animals (25, 26). MAP was recorded at 5-min intervals and averaged for each 24-hour period. Levels reported in the table 1 were those obtained for 2 days preceding the in vivo studies.

Six weeks after induction of RAS, the pigs were anesthetized again and randomized into two groups that underwent either sham-treatment (RAS, n=7) or PTRA (RAS + PTRA, n=6). The degree of stenosis was evaluated by selective renal angiography, followed by PTRA or sham procedure. A 7F balloon catheter was engaged in the proximal-middle section of the renal artery under fluoroscopic guidance and the balloon was inflated, resulting in expansion of a tantalum stent to full balloon diameter, to restore luminal opening. Then, the balloon was deflated and removed, leaving the stent embedded in the vascular wall. A group of normal kidneys that did not undergo coil implantation served as controls (Sham, n=8).

Four weeks later the pigs were again anesthetized similarly. Blood samples were collected from the inferior vena cava for measurement of plasma renin activity (PRA, radio-immunoassay) and creatinine (colorimetric assay) (8). Single kidney renal function was then evaluated in vivo using helical multi-detector computer tomography (MDCT) for assessment of basal regional-renal perfusion, RBF, and GFR, as previously described in detail (5-7, 27, 28). Briefly, this method was performed with sequential acquisition of 160 consecutive scans over a period of almost 3 minutes after a central venous injection of iopamidol (0.5 mL/kg per 2 seconds). MDCT images were reconstructed and displayed with the Analyze™ software package (Biomedical Imaging Resource, Mayo Clinic). Regions of interest were selected from cross-sectional images from the aorta, renal cortex, and medulla. Average tissue attenuation in each region was plotted over time and fitted by curve-fitting algorithms to obtain measures of renal function. Cortical and medullary volumes were calculated by Analyze™, and RBF as the sum of the products of cortical and medullary perfusions and corresponding volumes. GFR was calculated from the cortical curve utilizing the slope of the proximal tubular curve. The same procedure was repeated after 15 minutes towards the end of a 10-minute supra-renal infusion of Ach (5 μg/kg/min) to test endothelium-dependent microvascular reactivity. Briefly, a tracker catheter (Prowler Microcatheter, Cordis), introduced from the carotid artery was placed above the renal arteries for infusion of Ach. Hemodynamics and function were therefore measured over a stable 3-minute observation period at baseline and during Ach infusion.

Few days following completion of all functional studies, the pigs were euthanized (100mg/kg sodium pentobarbital). Kidneys were removed, immersed in 4°C saline solution containing heparin (10 units/mL) and divided into different lobes, stored at -80°C; immersed in 10% buffered formalin, or cannulated and prepared for micro-CT.

Micro CT analysis

All kidneys were perfused with 0.9% saline (containing heparin) at 10 ml/min under physiologic perfusion pressure (100 mmHg), using a saline-filled cannula ligated in a segmental artery. After 5-10 min, this was replaced with an intravascular radio-opaque silicone polymer (Microfil MV122; Flow Tech, Carver, MA, 0.8 ml/min) until the polymer drained freely from the segmental vein. A lobe of the polymer-filled tissue was subsequently trimmed, prepared and encased in paraffin, and scanned at 0.5-degree increments using a micro-CT scanner, as described previously (8). Three-dimensional volume images were then reconstructed at cubic voxels of 20ˆ3 μm, and displayed at 40-μm cubic voxels for analysis. The spatial density, average diameter and tortuousity of cortical microvessels were calculated as we have previously described (8, 29), and classified according to diameter as small (<40 μm), medium (40-100 μm), or large (>100 μm) vessels.

Immunohistochemistry and Western blotting

In vitro studies were performed to assess mechanisms responsible for formation and maintenance of the renal microvasculature, as well as fibrogenic factors and oxidative stress in the kidney. Standard Western blotting protocols were performed (7) with specific antibodies previously used or that cross-react with swine tissue. We used antibodies against vascular endothelial growth factor (VEGF), its receptors Flt-1 and Flk-1 (1:200, Santa Cruz Biotechnology), angiopoietin-1 (1:750, Novus Biologicals), endothelial nitric oxide synthase (eNOS), matrix metalloproteinase (MMP)-2, transforming growth factor (TGF)-beta, tissue inhibitors of metalloproteinase (TIMP)-1 and -2 (1:200, Santa Cruz Biotechnology), and tissue transglutaminase (tTG, 1:500, Novus Biologicals). To assess representative mediators of oxidative stress, protein expression of xanthine oxidase (1:10,000, Chemicon International) was also evaluated. β-actin (1:3000, Sigma) was used as loading control.

Staining was performed in 5μm paraffin slides for Masson trichrome and α-smooth muscle actin (SMA, 1:50, Sigma), or in frozen slides for activated caspase-3 (1:200, Santa Cruz), following standard procedures (5, 27). Each representative staining was quantified semi-automatically in 13 fields, expressed as percentage of staining of total surface area or number of positive cells/slide. The microvascular media/lumen ratio was measured in α-SMA-positive microvessels under 500μm in diameter. Apoptotic signals were characterized in renal cells with the TUNEL method using the deadEnd Fluorometric TUNEL system (Promega, Madison, WI), as we have previously shown (8). Apoptotic cells were quantified in 15 fields as the fraction of TUNEL-positive nuclei in each field.

Collagen content

The kidney cortex of each animal was used for a hydroxyproline assay as described previously (30). Briefly, the cortex was weighed, minced, homogenized, and diluted in PBS to 100mg cortex weight/ml. One hundred-microliter samples were then hydrolyzed in 12 mol/l HCl and duplicate samples analyzed by hydroxyproline assay.

Statistical Analysis

Quantitative data are expressed as mean ± SEM. We tested Gaussian distribution using Kolmogorov- Smirnov test and equality of variances with Bartlett test. Statistical comparisons among different experimental groups or studies were subsequently performed using one-way ANOVA with Newman-Keuls post-hoc test or non-parametric Kruskal-Wallis one-way ANOVA on ranks test with Dunns post-hoc test. Paired t-test was used for comparison within groups. Statistical significance was accepted for p≤0.05.