Stimulation of Caveolin-1 Signaling Improves Arteriovenous Fistula Patency

Abstract

Objective:

Arteriovenous fistulae (AVF) are the most common access created for hemodialysis; however, many AVF fail to mature and require repeated intervention, suggesting a need to improve AVF maturation. Eph-B4 is the embryonic venous determinant that is functional in adult veins and can regulate AVF maturation. Caveolin-1 (Cav-1) is the major scaffolding protein of caveolae, a distinct microdomain that serves as a mechanosensor at the endothelial cell membrane. We hypothesized that Cav-1 function is critical for Eph-B4-mediated AVF maturation.

Approach and Results:

In a mouse aortocaval fistula model, both Cav-1 mRNA and protein were increased in the AVF compared to control veins. Cav-1 KO mice showed increased fistula wall thickening (p=0.0005) and outward remodeling (p<0.0001), with increased eNOS activity compared with WT mice. Ephrin-B2/Fc inhibited AVF outward remodeling in WT mice but not in Cav-1 KO mice, and was maintained in Cav-1 endothelial reconstituted (RC) mice (WT, p=0.0001; Cav-1 KO, p=0.7552; Cav-1 RC, p=0.0002). Cavtratin, a Cav-1 scaffolding domain peptide, decreased AVF wall thickness in WT mice as well as in Eph-B4 het mice compared to vehicle alone (WT, p=0.0235; Eph-B4 het, p=0.0431); cavtratin also increased AVF patency (day 42) in WT mice (p=0.0275).

Conclusions:

Endothelial Cav-1 mediates Eph-B4-mediated AVF maturation. The Eph-B4-Cav-1 axis regulates adaptive remodeling during venous adaptation to the fistula environment. Manipulation of Cav-1 function may be a translational strategy to enhance AVF patency.

Graphical Abstract

Introduction

End-stage renal disease (ESRD) is a serious public health issue with increasing mortality rates worldwide.¹ Patients with ESRD are largely treated with and dependent on hemodialysis as a renal replacement therapy. Arteriovenous fistulae (AVF) are the preferred vascular access for hemodialysis, with superior patency, reduced infection, and improved long-term outcomes including mortality compared to patients requiring an arteriovenous graft or a central venous catheter,² as well as lower total health care expenditures.³ Two-thirds of hemodialysis patients in the United States used an AVF as an access in 2013,⁴ in accordance with guidelines.⁵ AVF are not available immediately after creation since the outflow vein must mature, that is thicken and dilate, to support the higher magnitude of blood flow as well as cannulation with large bore needles that are required for hemodialysis.⁶ Unfortunately up to 60% of AVF fail to mature;^7–9 a recent systematic review reported primary patency rates at 1 year were only 60%.¹⁰ These imperfect AVF outcomes force patients with ESRD to undergo repeated vascular interventions or conversion to other suboptimal methods of access.^{7, 11, 12} As a consequence, to develop strategies to improve AVF maturation, better understanding of the basic mechanisms and molecular events leading to venous adaptation within a fistula environment is required.^{13, 14}

Caveolin-1 (Cav-1) is the major scaffolding protein of plasmalemmal caveolae, serving as a dynamic platform to concentrate signaling molecules critical for endothelial cell (EC) function.^{15, 16} In particular, Cav-1 inhibits nitric oxide (NO) signaling by binding and inhibiting endothelial NO synthase (eNOS),¹⁷ thus regulating vascular remodeling and angiogenesis. Cav-1 regulates arterial remodeling,^{18, 19} suggesting that Cav-1 functions as a mechanotransduction molecule in the endothelium and thus may also be capable of regulating venous adaptation.

Cav-1 interacts with several receptor tyrosine kinases (RTK)^20–23 and G-protein coupled receptors, and regulates their signal-mediated vascular remodeling.^{24, 25} Eph-B4, a member of the Eph receptor tyrosine kinase family, is the embryonic venous determinant. We previously showed that Eph-B4 mediates vein graft adaptation and AVF maturation with distinct patterns of altered vessel identity.^26–28 Successful AVF maturation also requires venous adaptation to a fistula environment that is similar to the arterial environment but is characterized by distinct biology and hemodynamics compared to vein graft adaptation.²⁹ Although Cav-1 is a determinant of Eph-B4-mediated vein graft adaptation,²⁷ the role of Cav-1 during AVF maturation has not been investigated. Since Eph-B4 expression increases during AVF maturation,²⁸ but not vein graft adaptation,²⁶ we hypothesized that Cav-1 mediates AVF maturation and that endothelial Cav-1 is a critical regulator of Eph-B4-mediated AVF maturation through inhibition of eNOS activity.

Methods

Human specimens

The principles outlined in the Declaration of Helsinki were followed, and approval of the Veterans Affairs institutional Human Investigation Committee was obtained; informed consent to use the samples was obtained. Deidentified specimens of AVF and control veins were obtained during revision operations of functional AVF.²⁸ Since most fistulae are currently revised by endovascular methods, open revisions are no longer performed regularly, and thus only n=3 human specimens were available for analysis.

Mice

All animal experiments were performed in strict compliance with federal guidelines and with approval from the Yale University IACUC. Appropriate anesthesia and analgesia was given as previously described.^{30, 31} Mice used for this study included wild type C57BL6/J (WT), Cav-1 KO, Cav-1 endothelial reconstituted (Cav-1 RC), or Eph-B4 heterozygous (Eph-B4 het) mice. C57BL/6 WT, Cav-1 KO, and EphB4^+/− mice were purchased (The Jackson Laboratory). Cav-1 RC mice were generated on the Cav-1 KO genetic background and maintained as previously described.^{18, 32} All mice were male and 9–11 weeks of age; only male mice were studied since female sex is the only predictor of non-maturation of human AVF in some studies.³³ EphB4^+/− mice were B6.129S7 background and bred to C57BL/6 mice for at least ten additional generations so that littermate mice were used for reference groups.²⁷ Estimated sample sizes were calculated using a power analysis (Graphpad Statmate 2.00 software); 6 animals per sample would be needed for morphometry to allow detection of a difference in neointimal hyperplasia between groups assuming alpha=0.05, 80% power, SD=1.0 micron, and difference between groups of 2 microns. Fewer animals are needed for qPCR and Western blotting since these studies are more accurate, with smaller standard deviation, than morphometry of histology. However, when we found a significant difference with a smaller sample number, additional animal experiments were not performed to prevent unnecessary waste of precious resources and animals.

Infrarenal aorto-caval fistula model

Infrarenal aortocaval fistulae were created as previously described.³¹ Briefly, an aortocaval fistula was created by needle puncture from the aorta into the inferior vena cava (IVC). Sham procedures consisted of midline laparotomy without the AVF creation. The presence of the AVF was confirmed by direct visualization of pulsatile arterial blood flow in the IVC at creation and harvest, as well as by serial Doppler ultrasound (Vevo770 High Resolution Imaging System; VisualSonics Inc., Toronto, Ontario, Canada). In the latter, increased diastolic flow through the aorta and a high velocity pulsatile flow within the IVC were used to confirm the presence of an AVF during post-operative examinations.

In vivo measurement of fistula patency and dilation

Doppler ultrasound was also used to measure the diameter of the IVC and aorta during the study period as previously described.³⁰ Measurements were performed the day prior to operation (preoperative values) and serially post-operatively.

Recombinant peptide experiments

Peptides, corresponding to the scaffolding domain of Cav-1 (amino acids 82–101; DGIWKASFTTFTVTKYWFYR) were synthesized as a fusion peptide to the C terminus of the antennapedia internalization sequence (RQIKIWFQNRRMKWKK) by standard Fmoc chemistry (United BioSystems, Herndon, VA).³⁴ Mice were given control (truncated) antennapedia peptide (AP) or AP-Cav peptides, or cavtratin, via intraperitoneal injection (3.0 mg/kg) once daily beginning on day 0 and continuing throughout the experiments.

Eph-B4 stimulation in vivo

Eph-B4 was stimulated with Ephrin-B2/Fc (R&D, Minneapolis, MN). 24 hours prior to AVF creation, mice underwent ultrasound for measurement of preoperative vessel diameter as described above. While still under isoflurane anesthesia, Ephrin-B2/Fc (20 µg) diluted in 200 µL PBS was delivered by intraperitoneal injection; control mice received an equal volume injection of vehicle (PBS). The next day, an infrarenal aorto-caval AVF was created as described above. Additional intraperitoneal injections of Ephrin-B2/Fc or control vehicle were delivered every 48 hours beginning on postoperative day 1 and continued throughout the study period.

Histology

The venous limb of the AVF or control IVC from sham-operated mice were extracted en bloc after circulatory flushing with normal saline followed by 10% formalin. The tissue block was then embedded in paraffin and cut in 5-μm cross-sections. For wall thickness measurements, cross sections were obtained at 0.2 to 5.0 mm cranial to the AVF and stained with elastin van Gieson stain to highlight the elastic lamina. Digital images for the sections were captured with a microscopic system (BX40; Q Color 5, Olympus America, Center Valley, PA) and were analyzed using ImageJ 1.48v software (National Institute of Health, Bethesda, MD). Wall thickness, or intima + media thickness, was measured at 8 points (45°) around the vessel circumference and then averaged for each sample.²⁸ Additional unstained cross sections in this same region (0.2–5.0 mm cranial to the AVF) were used for immunohistochemistry or for immunofluorescence.

Immunohistochemistry

Sections were heated in citric acid buffer (pH 6.0) at 100°C for 10 min for antigen retrieval. Sections were then treated with 0.3% hydrogen for 30 min and blocked with 5% normal goat serum containing 0.05% Triton X‐100 (T‐PBS). Sections were then incubated overnight at 4°C with the following primary antibodies diluted in T-PBS: anti-Cav-1 (BD Biosciences, 610057, 610059), anti-Ki67 (Abcam, ab15580), or anti-cleaved caspase-3 (cell signaling, #9661). Isotype control that lacked specificity to the target but matched the class and type of the primary antibody was used as a negative control. After overnight incubation, the sections were incubated with Dako EnVision for 1 hour at room temperature and treated with Dako Liquid DAB+ Substrate Chromogen System (Dako; Carpinteria, CA) to detect the reaction products. Finally, the sections were counterstained with Dako Mayer’s Hematoxylin (Lillie’s Modification) Histological Staining Reagent (Dako). The integrated optical density of the immunoreactive signals in the vessel wall was analyzed using Image-J software. Relative density was graphed in arbitrary units. Cells staining positively for Ki67 and cleaved caspase-3 were directly counted in four high-power fields and the mean numbers of cells were then compared.

Immunofluorescence

Antigen retrieval was done in the same manner as IHC described above. After pretreatment, sections were blocked with 5% normal goat serum in T‐PBS for 1hr at room temperature and then incubated overnight at 4°C with the following primary antibodies diluted in T-PBS: anti-Cav-1 (BD Biosciences, 610057, 610059), anti-phospho-eNOS (Abcam, ab195944), anti-eNOS (Santa Cruz, sc654), anti-phospho-Akt1 (Cell Signaling Technology, 9018), anti-alpha-actin (Abcam, ab5694), anti-vWF (Abcam, ab11713), anti-PECAM-1 (Santa Cruz, sc1506) or with isotype-matched primary antibodies serving as negative controls. Secondary reagents were Alexa Fluor conjugated antibodies (488 and 568) from Life technologies used at 1:250. Sections were stained with SlowFade® Gold Antifade Mountant with DAPI (Life Technologies). Digital fluorescence images were captured and intensity of immunoreactive signal was measured using Image J software (NIH, Bethesda, Maryland). Intensity of merge signal was determined by applying a color threshold selective for yellow signal. For each antibody, standard quality control procedures were undertaken to optimize antigen retrieval, primary antibody dilution, secondary antibody detection, and other factors for both signal and noise.

Western blotting

The venous limb of the AVF was harvested and treated with RIPA lysis buffer containing protease and phosphatase inhibitors (Roche). Equal amounts of protein were loaded and run in SDS-PAGE followed by Western blot analysis. Protein expression was probed with the following antibodies: Cav-1 (BD Biosciences, 610406), phospho-eNOS (Cell Signaling Technology, 9571), eNOS (Santa Cruz, sc-654), phospho-Akt1 (Cell Signaling Technology, 9018), Akt1 (Cell Signaling Technology, 2967), or GAPDH (Cell Signaling Technology, 2118). Membrane signals were detected using ECL detection reagent (GE Healthcare, Denville scientific).

RNA extraction and quantitative RT-PCR

The venous limb of the fistula or the IVC (sham) was harvested after isolation from surrounding arterial and connective tissues and total RNA was isolated using RNeasy Mini Kit with digested DNase I (Qiagen, Valencia, California, USA). RNA quality was confirmed using the 260/280 nm ratio (Nanodrop; Thermo Scientific, Wilmington, Delware, USA). Reverse transcription was performed using SuperScript III First-Strand Synthesis Supermix (Invitrogen; Life Technologies, Carlsbad, California, USA). Real-time PCR was performed using iQ SYBR Green Supermix (Bio-Rad, Hercules, California, USA). Samples were amplified for 35 cycles using the iQ5 Real-Time PCR Detection System (Bio-Rad). Correct target amplification and exclusion of nonspecific amplification was confirmed by 1.5% agarose gel electrophoresis; primer efficiencies were determined by melt curve analysis. Primers used are as listed in Table III. All samples were normalized by GAPDH RNA amplification.

Transmission electron microscopy

IVC tissue samples were fixed with 2.5% glutaraldehyde in 0.1 mol/L cacodylate buffer, postfixed with OsO4, and stained with uranyl acetate and lead citrate. Microtome sections were examined under a FEI Tecnai Biotwin TEM. Caveolae were identified by their characteristic flask shape, size (50 –100 nm), and location at the plasma membrane.³⁵

Cell culture

Mouse lung endothelial cells (MLEC) isolated from male mice were prepared as previously described,³⁶ and were cultured in EBM-2 supplemented with 20% FBS, L-glutamine, penicillin/streptomycin, and the EGM-2 Bulletkit (Lonza).

NO release analysis

NO production by MLEC was analyzed as previously described.³⁷ In brief, conditioned medium was examined 24 h after cavtratin treatment. The media was processed for the measurement of nitrite (NO_2⁻) by an NO-specific chemiluminescence analyzer (Sievers).

In vitro whole cell lysate isolation

Cells were treated with RIPA lysis buffer containing protease and phosphatase inhibitor cocktail. Cells were then scraped, sonicated and centrifuged.

Statistical analysis

All data were analyzed using Prism 7 software (GraphPad Software, Inc., La Jolla, CA). Error bars depict the SEM. The Shapiro-Wilk test was performed to analyze normality; F test were performed to evaluate homogeneity of variances. For two-group comparisons with normally distributed data, the unpaired Student’s t-test was used for data with equal variances among groups and the unpaired Student’s t-test (with Welch correction) was used for data with unequal variances. For multiple group comparisons with normally distributed data, the one-way ANOVA followed by the Tukey post-hoc test was used. For data that was not normally distributed, the nonparametric Mann-Whitney test was used for two-group comparisons and the Kruskal-Wallis was performed for multiple groups. Patency outcomes were analyzed with the use of a Cox proportional-hazards model. Kaplan–Meier curves were constructed to display the distribution of occlusion events detected over time. P values < 0.05 were considered significant.

Results

Increased Cav-1 expression and caveolae formation during AVF maturation

To determine if Cav-1 regulates AVF maturation, we evaluated Cav-1 expression in human and mouse AVF. In functional mature human AVF, expression of immunoreactive Cav-1 protein was increased compared to control vein (Figure 1A). Similarly, in a mouse AVF model there was increased expression of Cav-1 protein in the AVF compared to sham-operated veins (Figure 1B); transcripts of Cav-1 were increased in the AVF from day 3 to day 21 compared to control veins (Figure 1C). The increased Cav-1 in AVF colocalized predominantly with vWF-positive cells, with 3.8 times more immunoreactivity colocalized with vWF-positive cells compared with α-SMA positive cells (Figure 1D). Venous endothelium of control veins or AVF was examined with transmission electron microscopy; the number of caveolae per length of plasma membrane was increased 1.9-fold in the endothelium of AVF compared to control veins (Figure 1E). These results show that in AVF there is increased Cav-1 expression and caveolae formation in the venous endothelium.

Figure 1.

Caveolin-1 and caveolae formation are upregulated in arteriovenous fistula. A, Left, representative immunoblotting of Cav-1 in human fistula or normal vein. Right, densitometry of Cav-1 immunoblotting relative to GAPDH. *P=0.0036, student’s t-test, N=3 for each group. B, Left, representative immunoblotting for Cav-1 in mouse IVC harvested at day 21 after sham operation or AVF creation. Right, densitometry of Cav-1 immunoblotting relative to GAPDH. *P=0.005, student’s t-test, N=5 for each group. C, Time course of Cav-1 mRNA expression in the IVC after AVF creation. Normalized with the values from the IVC harvested at identical time points after sham operation. *P=0.019, ANOVA, N=4 for day 0, N=8 for day 1, 3, 7, 21. D, Left, immunofluorescence study of Cav-1, vWF, and α-SMA in fistula at day 7 after sham operation of AVF creation. White arrowheads denote double positive signal. Scale bar, 20μm. L, lumen of the IVC. Right, quantification of double positive signal. *P=0.0071, student’s t-test, N=4 for sham; N=6 for AVF. E, Left, TEM (transmission electron microscopy) images for IVC endothelial cells from C57/BL6 mice at day 7 after sham or AVF operation. Scale bar is 500nm. At least 20 individual images were analyzed per mouse for the quantification of caveolae/μm of plasma membrane. *P<0.0001, Mann-Whitney test, N=3 per group. Error bars represent the standard error of the mean (s.e.m.).

Absence of Cav-1 facilitates AVF remodeling

Venous remodeling that occurs during AVF maturation in this model is characterized by approximately 2.4-fold increased wall thickening and 1.5-fold increased dilation compared to mice with sham operations.³¹ Since our data shows increased Cav-1 expression during AVF maturation (Figure 1), we hypothesized that Cav-1 regulates flow-mediated venous remodeling such as occurs during AVF maturation. To determine whether reduced Cav-1 function plays a functional role to regulate AVF maturation, we examined AVF maturation in Cav-1 KO mice. Cav-1 KO mice have similar body weight and similar baseline wall thickness and diameter of their vena cavae and aortae compared to control mice (Supplemental Figure IA-F). As expected, the veins of Cav-1 KO mice did not have any detectable Cav-1 expression but had increased eNOS phosphorylation (Supplemental Figure IG).^{32, 34, 38, 39} Similarly, Cav-1 expression was only minimally detectable in EC derived from Cav-1 KO mice compared to that detectable in control mice, and there was increased eNOS phosphorylation (Figure 2A) and nitrite released from EC derived from Cav-1 KO mice (Figure 2B).

Figure 2.

Absence of Cav-1 facilitates remodeling of the fistula with eNOS activation. A, Left, representative immunoblotting for Cav-1, peNOS (S1177), eNOS in WT and Cav-1 KO mouse lung endothelial cells. Right, densitometry of immunoblotting (ratio of phospho- to total-eNOS). *P=0.0022, Mann-Whitney test, N=3 each group. B, Nitrite release in WT and Cav-1 KO endothelial cells. *P=0.0002, t-test, n=2 for WT; n=3 for Cav-1 KO. C, Left panels, IVC wall at 21 days after AVF creation in WT and Cav-1 KO mice. Sections were prepared with EVG staining. right, bar graph of IVC wall thickness. *P=0.0005, student’s t-test, n=7 for each group. Scale bar, 50μm. L, lumen of the fistula. D, Outward remodeling of fistula and aorta in WT and Cav-1 KO mice. Data are shown as relative value normalized by preoperative measurement. n=10 for WT, n=7 for Cav-1 KO mice. *P<0.0001 for IVC; P=0.1449 for aorta, 2-way ANOVA. E, Left, representative immunoblotting of Cav-1, peNOS, eNOS and GAPDH protein in WT and Cav-1 KO fistula at day 21. Right, densitometry of immunoblotting. *P=0.0497 for peNOS, **P=0.0135 for teNOS, student’s t-test, N=3 each group. F, Left, section of IVC at 21 days after AVF creation in WT and Cav-1 KO mice stained with anti-peNOS, anti-eNOS, and anti-vWF antibodies. The data represent 5 mice that showed similar results. Right, quantification of peNOS, eNOS and vWF positive signal. *P=0.0165, student’s t-test, N=5 for each group. G, Proliferation of IVC wall at day 21 after AVF creation in WT and Cav-1 KO mice. *P=0.0286, Mann-Whitney test. n=4 for each group. H, Apoptosis of IVC wall at day 21 after AVF creation in WT and Cav-1 KO mice. P=0.7941. n=4 for each group. I, Left, section of IVC at 21 days after AVF creation in WT and Cav-1 KO mice stained with anti-Cav-1, anti-α-SMA antibodies. Right, quantification of Cav-1- and α-SMA-positive signal. *P=0.0089, student’s t-test. N=5 for each group. Error bars represent the standard error of the mean (s.e.m.). J, Left panels, IVC wall at 21 days after AVF creation in WT and eNOS KO mice. Sections were prepared with EVG staining. right, bar graph of IVC wall thickness. *P=0.0019, student’s t-test, n=5 for each group. Scale bar, 50μm. L, lumen of the fistula vein. K, Left panels, IVC wall at 21 days after AVF creation in WT mice treated with L-NAME or control. Sections were prepared with EVG staining. right, bar graph of IVC wall thickness. *P=0.0211, student’s t-test, n=4 for each group. Scale bar, 50μm. L, lumen of the fistula vein.

AVF were created in control and Cav-1 KO mice. Postoperative survival was reduced in Cav-1 KO mice (46.7%, 7/15) compared to control mice (91.7%, 11/12; p=0.0187, log rank); however, the patency rate of AVF (day 21) was similar between Cav-1 KO mice (100%, 15/15) and WT mice (91.7%, 11/12; p=0.4346, log rank) (Supplemental Figure IH-I). Despite lack of any significant differences in venous diameter and wall thickness between control and Cav-1 KO mice at baseline (Supplemental Figure I), AVF in Cav-1 KO mice showed increased fistula wall thickness compared with AVF in control mice, both at day 7 (Supplemental Figure IIA) as well as day 21 (Figure 2C); both total wall thickness (30.9 ± 2.6 μm vs. 15.6 ± 2.0 μm; n=7; p=0.0005) as well as neointimal thickness (19.0 ± 3.1 μm vs. 7.0 ± 1.1 μm; n=7; p=0.002) were increased in AVF of Cav-1 KO mice (day 21). Serial Doppler ultrasound showed that AVF in Cav-1 KO mice had larger diameter, relative to baseline, compared to AVF in control mice (Figure 2D), despite lower blood pressure in Cav-1 KO mice,³² and similar blood flow velocity and shear stress in the AVF of control and Cav-1 KO mice (Supplemental Figure IIB). Interestingly, there was no difference in relative diameter of the artery leading into the AVF (Figure 2D).

In AVF created in Cav-1 KO mice (day 21) there was no detectable Cav-1 expression but increased phosphorylated and total eNOS immunoreactivity compared to AVF created in control mice (Figure 2E), despite similar total eNOS expression in the IVC at baseline (Supplemental Figure IG); increased total eNOS expression is consistent with increased endothelium and neointima in the fistula. Immunofluorescence showed increased intensity of both p-eNOS (2.4-fold) and eNOS (1.8-fold) in the endothelium of Cav-1 KO AVF compared to control AVF, with similar vWF intensity in the two groups (Figure 2F). Increased wall thickening in the AVF of Cav-1 KO mice was characterized by increased numbers of proliferating cells with no decrease in apoptotic cells (Figure 2G,2H); as expected, the increased numbers of α-SMA-positive cells in AVF of Cav-1 KO mice did not have any Cav-1 immunoreactivity (Figure 2I).

These data suggest that Cav-1 regulates eNOS activity in vivo during AVF remodeling; accordingly, we assessed eNOS function in wild type mice. In wild type mice, the ratio of phospho-eNOS to total-eNOS increased in the AVF compared to sham-operated veins (Supplemental Figure IIIA). eNOS knockout mice had AVF with thinner walls compared to control mice (Figure 2J), and with larger diameter consistent with elevated blood pressure (Supplemental Figure IIIB); treatment with L-NAME, a global nitric oxide synthase inhibitor, similarly resulted in thinner AVF walls with larger diameter compared to control mice (Figure 2K, Supplemental Figure IIIC). These data suggest that eNOS is functional during AVF remodeling and thus Cav-1 regulation of eNOS may be a mechanism of venous remodeling during AVF maturation.

In toto, these data suggest that global deletion of Cav-1 regulates venous remodeling in the fistula environment, characterized by increased venous wall thickness and diameter, as well as increased proliferation of smooth muscle cells (SMC), and is consistent with the inhibitory effect of Cav-1 on downstream function including eNOS.^{34, 40}

Endothelial Cav-1 regulates Eph-B4-mediated AVF remodeling

We have previous shown that the venous determinant Eph-B4 regulates AVF remodeling,²⁸ similar to the function of Cav-1 (Figure 2). Eph receptors translocate to or reside within caveolae,^{20, 21, 23, 41} and stimulated Eph-B4 associates with Cav-1 in vitro and in vivo,²⁷ suggesting involvement of Cav-1 in Eph-B4 signaling. To directly test whether endothelial Cav-1 acts as a mediator of Eph-B4 during venous remodeling such as occurs during AVF maturation, Eph-B4 function was stimulated with its ligand Ephrin-B2/Fc in control mice, Cav-1 KO mice, and Cav-1 KO mice with specific EC Cav-1 reconstitution (Cav-1 RC) mice.^{18, 32} As expected, control mice treated with Ephrin-B2/Fc showed less AVF wall thickness at day 21; however, AVF in Cav-1 KO mice had increased wall thickness that was not responsive to Ephrin-B2/Fc; AVF wall thickness in Cav-1 RC mice showed a trend towards response to Ephrin-B2/Fc (26% reduction, p=0.07; Figure 3A). After Ephrin-B2/Fc treatment there were reduced numbers of proliferating cells in AVF of control mice but not in AVF of Cav-1 KO mice; there were reduced numbers of proliferating cells in AVF of Cav-1 RC mice (Figure 3B). Ephrin-B2/Fc treatment did not change the numbers of apoptotic cells in the AVF of any of these groups (Figure 3C).

Figure 3.

Eph-B4 regulation of fistula remodeling partially depends on endothelial Cav-1. A, Left, wall thickness of IVC at 21 days after AVF creation in WT, Cav-1 KO, Cav-1 RC mice treated with ephrin-B2/Fc or PBS. Sections were prepared with EVG staining, right, bar graph of IVC wall thickness. *P=0.0192 for WT, P=0.8533 for Cav-1 KO, P=0.0761 for Cav-1 RC mice. student’s t-test, N=3-8. Scale bar, 50μm. L, lumen of the fistula. B, Proliferation of IVC wall at day 21 after AVF creation in WT, Cav-1 KO, Cav-1 RC mice with or without ephrin-B2/Fc treatment. *P=0.0445 for WT, **P=0.0216 for Cav-1 RC. student’s t-test, n=3 for each group. C, Apoptosis of IVC wall at day 21 after AVF creation in WT, Cav-1 KO, Cav-1 RC mice with or without ephrin-B2/Fc treatment. D, Relative fistula diameter with ephrin-B2/Fc or control treatment. *P=0.0001 for WT, P=0.7552 for Cav-1 KO, *P=0.0002 for Cav-1 RC, 2-way ANOVA.E, Left, section of IVC at 21 days after AVF creation in WT, Cav-1 KO, Cav-1 RC mice treated with ephrin-B2/Fc or control stained with anti-peNOS. The data represent 3 mice per group that showed similar results. Scale bar, 50μm. L, lumen of the fistula vein. Right, quantification of peNOS positive signal. *P=0.0212 for WT, student’s t-test, **P=0.0402 for Cav-1 RC, student’s t-test with Welch correction. N=3-4 for each group.

Similarly, there was less outward remodeling during AVF maturation in control mice treated with Ephrin-B2/Fc; Cav-1 KO mice treated with Ephrin-B2/Fc showed no difference in outward remodeling, whereas Cav-1 RC mice showed less dilation in response to Ephrin-B2/Fc, similar to the responsiveness of control mice (Figure 3D). Ephrin-B2/Fc treatment also decreased the intensity of p-eNOS immunoreactivity in the AVF endothelium of control mice but not in Cav-1 KO mice; this reduction was partially restored in Cav-1 RC mice (p=0.0325; Figure 3E). eNOS KO mice treated with Ephrin-B2/Fc did not show reduced wall thickness, but this was not surprising since eNOS KO mice did not show any wall thickening at baseline (Supplemental Figure IIID, Figure 2J). These results suggest that Eph-B4-mediated AVF remodeling is abolished in Cav-1 KO mice and partially restored in endothelial Cav-1 RC mice; endothelial Cav-1 regulation of Eph-B4-mediated AVF remodeling is associated with diminished eNOS phosphorylation and cell proliferation.

Augmentation of Cav-1 signaling regulates flow-mediated venous remodeling and improves AVF patency

Since endothelial Cav-1 inhibits wall thickening and diameter expansion during AVF maturation (Figure 2), and reduced wall thickness is associated with improved AVF patency,²⁸ we hypothesized that simulation of Cav-1 function would enhance AVF patency. Therefore, we used cavtratin, a chimeric peptide with a cellular internalization sequence fused to the Cav-1 scaffolding domain that enhances Cav-1 function and inhibits eNOS activity,³⁴ to alter flow-mediated venous remodeling during AVF maturation. As expected, cavtratin inhibited nitrite release from EC in vitro (Figure 4A). Wild type mice treated systemically with cavtratin showed less thickened AVF walls (day 21) compared to control mice (Figure 4B). In addition, cavtratin treatment was associated with less diameter expansion of the AVF but not the artery compared to control mice (Figure 4C). AVF in mice treated with cavtratin showed reduced numbers of proliferating cells with no increase in apoptotic cells in the AVF (Figure 4D,4E).

Figure 4.

Stimulation of Cav-1 signaling inhibits AVF remodeling and improves patency. A, Effect of cavtratin on nitrite release in WT endothelial cells. P=0.0197, student’s t-test, N=3. B, Left panels, IVC wall at day 21 after AVF creation with cavtratin or control treatment, right, bar graph of IVC wall thickness. *P=0.0408, student’s t-test with Welch correction. n=4 for each group. Scale bar, 50μm. L, lumen of the fistula vein. C, Outward remodeling of fistula and aorta in WT mice with cavtratin or control treatment. Data are shown as relative value normalized by preoperative measurement. *P=0.0085 for IVC; P=0.0810 for aorta, two-way ANOVA. N=7 for cavtratin, N=4 for control. D, Proliferation of IVC wall at day 21 after AVF creation in WT mice with control or cavtratin treatment. *P=0.0260, student’s t-test. N=4 for each group. E, Apoptosis of IVC wall at day 21 after AVF creation in WT mice with control or cavtratin treatment. P=0.9862. N=4 for each group. F, Line graph showing AVF patency in mice treated with control or cavtratin injections; failures < 7 days not included. Hazard ratio 0.10; *P=0.0275 by log-rank test. n=13 for control; n=12 for cavtratin. G, Bar graph of IVC wall thickness at 42 days after AVF creation in mice with control or cavtratin treatment. *P=0.0261, student’s t-test. N=7 for cavtratin, N=4 for control. H, Outward remodeling of fistula and aorta in WT mice with cavtratin or control treatment. Data are shown as relative value normalized by preoperative measurement. *P= 0.0474 for IVC, P= 0.1627 for aorta, two-way ANOVA. N=12 for cavtratin; N=5 for control. I, Left, representative immunoblotting for peNOS (S1177), teNOS, pAkt1 (S473), and Akt1 in mouse IVC harvested at day 42 after AVF creation. Right, densitometry of immunoblotting (ratio of phospho- to total-protein). *P=0.0391 for peNOS/teNOS; P=0.5790 for pAkt1/tAkt1 , student’s t-test, N=3 for each group. J, Left, section of IVC at 42 days after AVF creation in WT mice with cavtratin or control treatment stained with anti-peNOS, anti-pAkt1, anti-PECAM-1 and anti-α-SMA antibodies. The data represent 3 mice per group that showed similar results. Scale bar, 50μm. L, lumen of the fistula vein. Right, quantification of double positive signal. *P=0.0260, student’s t-test, N=3 for each group. K, Left panels, wall thickness of IVC at 21 days after AVF creation in Eph-B4+/+ and Eph-B4+/− mice treated with AP-Cav or AP (vehicle). Sections were prepared with EVG staining. Right, bar graph of IVC wall thickness. *P= 0.0205 for Eph-B4+/+, Mann-Whitney test. **P= 0.0431 for Eph-B4+/−mice, student’s t-test. N=3-8 for each group. Scale bar, 50μm. L, lumen of the fistula..L, Proposed schema of signaling that regulate fistula remodeling. Error bars represent the standard error of the mean (s.e.m.).

In the mouse aortocaval AVF model there is increased neointimal hyperplasia with wall thickening and reduced patency between days 21 and 42, similar to the aggressive juxta-anastomotic neointimal hyperplasia that is associated with high rates of early failure of human AVF.³¹ In wild type mice treated with cavtratin, AVF had significantly increased patency by day 42 (p=0.0275, log-rank); at day 42 AVF were occluded in 0 of 12 mice (0%) receiving cavtratin, but were occluded in 4 of 13 mice (30.8%) receiving control peptide (Figure 4F). Consistent with this data, cavtratin-treated mice showed less thickened AVF walls (p=0.0261) and less outward venous remodeling (p=0.0474) at day 42 (Figures 4G, 4H). At day 42, the AVF of mice treated with cavtratin showed less p-eNOS immunoreactivity in the AVF (Western blot of whole vessel lysate), without any change in p-Akt1 immunoreactivity, compared to AVF of control mice (Figure 4I); immunofluorescence showed reduced p-eNOS immunoreactivity in the endothelium of cavtratin-treated mice (Figure 4J). These data suggest that stimulation of Cav-1 signaling with cavtratin improves AVF patency and is associated with less wall thickness, less outward remodeling, and less eNOS phosphorylation during AVF maturation.

Since Cav-1 activity is associated with reduced wall thickness (Figures 2C, 2I, 4B, 4G), and Cav-1 mediates Eph-B4-mediated AVF remodeling (Figure 3), we next determined if stimulation of Cav-1 activity could rescue diminished Eph-B4 function during AVF remodeling, that is Cav-1 is downstream of Eph-B4. Eph-B4 heterozygous mice have similar body weight, baseline IVC and aorta diameter, and similar postoperative survival and patency after AVF creation compared to control mice (Supplemental Figure IV); however, AVF in Eph-B4 heterozygous mice develop thicker walls (day 21) compared to control mice.²⁸ Eph-B4 heterozygous mice treated with cavtratin showed reduced AVF wall thickness similar to control mice similarly treated (day 21; Figure 4K), suggesting that stimulation of Cav-1 with cavtratin can rescue diminished Eph-B4 function, that is Cav-1 is downstream of Eph-B4 (Figure 4L). In toto, these data suggest that Cav-1 signaling promotes AVF patency and may be a mechanism for Eph-B4-mediated AVF maturation.

Discussion

This study identifies Cav-1 as an essential regulator of flow mediated venous adaptation in mouse AVF model, suggesting that Cav-1 may be a potential pharmacological target for intervention to prevent failure of the venous conduits in patients with arterial occlusive disease or with ESRD. Using a mouse aortocaval fistula model, we show that stimulation of Cav-1 signaling with cavtratin suppressed AVF remodeling, resulting in improved AVF patency (Figure 4). In addition, we show that endothelial Cav-1 is a critical mediator of Eph-B4 function during AVF remodeling (Figure 3).

This is the first report to show that Cav-1 regulates flow-mediated venous adaptation in the fistula environment, which is distinctly different than the arterial environment.²⁹ Understanding venous remodeling remains a significant clinical challenge, since venous conduits that are used for hemodialysis access or vein grafts frequently fail due to insufficient venous remodeling.^{10, 42, 43} Although some venous wall thickening and outward remodeling are consistent components of successful venous adaptation to the fistula and arterial environments,^{6, 42} excessive remodeling may result in neointimal hyperplasia and clinical failure. Our data shows that Cav-1 limits wall thickening and exuberant neointimal hyperplasia (Figure 2C) and that stimulation of Cav-1 signaling regulated AVF remodeling and reduced the risk of AVF failure (hazard ratio 0.1; Figure 4F), suggesting that Cav-1 may be another potential translational target.

Cav-1, the essential protein for the formation of plasmalemmal caveolae, plays a major role in mechanotransduction of dynamic shear stress changes in EC by interacting with several signaling protein families,^{18, 44} and has diverse functions including regulation of inflammation, tumor progression, and atherosclerosis.^{34, 45, 46} In the vascular system, Cav-1 inhibits arterial remodeling,^{18, 46} consistent with our data demonstrating an inhibitory function of Cav-1 during venous remodeling. Caveolae density in endothelial plasmalemma is increased in fistulae in vivo (Figure 1), consistent with previous in vitro studies showing that endothelial caveolae increase in response to shear stress.⁴⁷ However, the intracellular role of Cav-1 is separable from its role in caveolae biogenesis; caveolae formation and downstream signaling events occur through independent mechanisms.⁴⁸ Therefore, the direct implication of increased caveolae density in AVF or flow-mediated venous adaptation is currently unknown.

Cav-1 is predominantly expressed in adipocytes, vascular SMC, EC, and fibroblasts, with endothelial Cav-1 playing a central role in blood vessel, pulmonary, and cardiac functions.³² In our AVF model, genetic global deletion of Cav-1 enhanced venous remodeling (Figure 2), whereas reconstitution of Cav-1 specifically in the endothelium restored a similar remodeling pattern as seen in WT mice (Figure 3, Supplemental Figure II); these data suggest an important role of endothelial Cav-1 in venous remodeling.

Cav-1 interacts with several RTK^20–23 and G-protein coupled receptors, and regulates their signal-mediated vascular remodeling.^{24, 25} For example, Cav-1 regulates AT1R signal-mediated vascular remodeling in vascular smooth muscle cells and endothelial cells.^{24, 25} Cav-1 also regulates VEGFR-2 activity;⁴⁹ the expression of VEGFR-2 is upregulated in the mouse AVF model.²⁸ Of translational interest, a RTK specifically associated with venous adaptation could be a target to improve AVF maturation; we focused on Eph-B4, an embryonic venous determinant active in adult endothelium, since we previously showed that Eph-B4 mediates vein graft adaptation and AVF maturation, and that Cav-1 is a determinant of Eph-B4-mediated vein graft adaptation. Cav-1 appears to be a downstream effector of Eph-B4 function since Cav-1 is not critical for Eph-B4 tyrosine phosphorylation, but interaction with Cav-1 are critical aspects of Eph-B4 signaling and function.²⁷

In addition to these multiple pathways, Cav-1 mediation of venous remodeling may involve eNOS. eNOS is a resident protein of caveolae, and Cav-1 functions as a tonic inhibitor of eNOS-mediated signal transduction in vivo;^{34, 40} Cav-1 KO mice show 5-fold increased systemic NO levels compared to wild type littermates.⁵⁰ Our data that Cav-1 KO mice have increased eNOS phosphorylation (Figure 2) whereas stimulation of Cav-1 activity with cavtratin inhibits eNOS phosphorylation (Figure 4) in AVF endothelium suggests that Cav-1 regulates AVF remodeling via eNOS, and this is corroborated by the data showing reduced thickening with reduced eNOS activity (Figure 2J-K). As such, Cav-1 regulation of AVF remodeling is likely to involve eNOS as a mechanism; although cavtratin inhibits the functional activity of broad classes of proteins in vitro, cavtratin selectively targets eNOS and reduces its activity in endothelial cells in vivo,³⁴ suggesting its potential for translational therapy.

Since veins ordinarily produce less NO compared with arteries,⁵¹ the abnormal venous remodeling and increased eNOS phosphorylation in Cav-1 KO mice (Figure 2) suggests a potential mechanism for excessive neointimal hyperplasia that can lead to AVF failure.⁵⁰ Although NO is generally considered to play a protective role in the maintenance of vascular homeostasis, several studies that reported excessive NO as a contributor of vessel pathologies,⁵² and our data may be consistent with this finding. Although short-term and occasional treatment with NO donors, such as those used to treat angina, is unlikely to cause vascular damage, our findings also indicate the need for caution in implementing long-term treatment with NO donors or with gene-therapy-augmented NOS expression.⁵² Augmentation of Cav-1 activity with cavtratin may provide a new therapeutic approach for regulation of NO production, perhaps mitigating the proinflammatory effects of excessive NO in EC;^{15, 34, 53} however, it is important to determine whether abnormal regulation of the NO pathway contributes to abnormal venous remodeling that is associated with AVF failure. While NO is essential for venous remodeling, uncoupled eNOS generating superoxide anion and peroxynitrite may impair AVF function.⁵⁴ We have previously shown that nitrotyrosine, a surrogate marker for peroxynitrite, is elevated during AVF remodeling.⁵⁵

In summary, our data suggests that Cav-1 acts as an inhibitory regulator of flow-mediated venous adaptation; stimulation of Cav-1-specific signaling inhibited eNOS activity and improved AVF patency. Endothelial Cav-1 is also a critical regulator of Eph-B4-mediated AVF maturation, possibly reflecting the dual function of Cav-1 as a scaffold of the caveolar signaling platform as well as a direct repressor of intracellular signaling events. Manipulation of an endothelial mechanosensor such as Cav-1 may be a new strategy to improve AVF maturation and optimize fistula use for patients with ESRD.

Table 1.

Primer sequences

Gene	Forward	Reverse
GAPDH	AATGTGTCCGTCGTGGATCTGA	AGTGTAGCCCAAGATGCCCTTC
Cav-1	ACGACGTGGTCAAGATTGACTTT	CCCAGATGTGCAGGAAGGAG

Highlights.

• Cav-1 expression and caveolae formation increases in the maturing AVF.

• Global deletion of Cav-1 regulates fistula remodeling with increased eNOS activity.

• Endothelial Cav-1 regulates Eph-B4-mediated fistula remodeling.

• Augmentation of Cav-1 signaling with cavtratin may be a strategy to improve AVF patency.

Nonstandard Abbreviations and Acronyms

AP

antennapedia

AVF

arteriovenous fistula

Cav-1

caveolin-1

EC

endothelial cell

eNOS

endothelial nitric oxide synthase

ESRD

end-stage renal disease

IVC

inferior vena cava

NO

nitric oxide

RTK

receptor tyrosine kinase

SMC

smooth muscle cell