Characterization of Arteriovenous Thrombus Formation and Propagation in a Mouse Arteriovenous Fistula Model

Abstract

Background

Compared with an arterial thrombus (AT) or a venous thrombus (VT), there is limited knowledge about arteriovenous thrombus (AVT). AVT develops in 69% of arteriovenous fistulae (AVF) and 50% of arteriovenous grafts (AVG) within 1 year. Thrombosis remains one of the major complications after creation of a vascular access often resulting in failure of the access.

Objective

To characterize and differentiate AVT from VT or AT.

Methods

An AVT model was established by a needle puncture through the aorta to the inferior vena cava (IVC) in wild type mice and different reporter mice and compared with a mouse venous thrombus (VT) model using IVC ligation. AVT was also examined under defined arteriovenous flow conditions. AVT was examined by gross view, histology, immunofluorescence, and scanning electron microscopy.

Results

AVT occurs immediately at the juxta-anastomotic area (JAA) after successful arteriovenous flow was established, with platelets being a major component of early AVT. Reduced injury to the endothelium resulted in smaller AVT, whereas local delivery of rapamycin to inhibit cell proliferation failed to decrease the volume of the AVT. Incomplete reendothelialization of the peri-fistula exit area correlated with growth of the AVT. AVT universally presents at the JAA in other arteriovenous models.

Conclusion

We provide the first detailed histopathological characterization of AVT induced by AVF. AVT originates from the injured vessel wall and is more similar to AT than VT. This model provides a valuable tool to characterize AVT. Both this AVT model and our data have potential for clinical translation.

Keywords

arteriovenous thrombus - arteriovenous flow - neointimal hyperplasia - juxta-anastomosis stenosis

Introduction

Symptomatic arterial thrombosis (AT) or venous thrombosis (VT) can cause ischemic stroke and myocardial infarction, or severe limb swelling and fatal pulmonary embolism respectively.[1] Compared with the large amount of research into AT and VT,[2] [3] there is still limited knowledge about arteriovenous thrombus (AVT) under arteriovenous flow. Arteriovenous flow occurs after arteriovenous fistula (AVF) or arteriovenous graft (AVG) creation and is characterized by acute increase in shear stress magnitudes and frequencies (disturbed flow) as well as oxygenation introduced into the low flow and low resistance venous environment. Symptomatic acute AVT accounts for 5 to 20% AVF failures during the first 30 days.[4] Chronic AVT occurs in 69% of AVF and 50% of AVG within a year,[5] [6] supporting a critical need to understand, characterize, and ultimately prevent the thrombus.

To recapitulate human AVF and AVG,[7] [8] different AVF and AVG animal models have been established. Such studies predominantly focus on the formation of late neointimal hyperplasia (NIH) but not early AVT.[9] [10] [11] Furthermore, unlike AT and VT, antiplatelet agents and anticoagulation are not recommended by the KDOQI Clinical Practice Guidelines to increase fistula maturation.[12] [13] [14] The unique nature of the AVT, compared with AT and VT, supports the need for a more comprehensive understanding of AVT formation.

Thrombosis always forms at the region of the vascular anastomosis secondary to the injury,[15] with disturbed flow being another strong risk factor for thrombus formation.[16] The juxta-anastomotic area (JAA) is suddenly exposed to increased disturbed arteriovenous flow after AVF and AVG creation in both human and animal models.[17] [18] [19] In this study, we explore AVT formation, organization, and growth in a mouse aortocaval model. We also examine AVT formation under different arteriovenous flow conditions complementing our studies with human AVF samples.

Materials and Methods

Human Samples

The principles outlined in the Declaration of Helsinki were followed. This study was approved by the Human Investigation Committee of Yale University. AVF samples were small circumferential segments of mature AVF that were harvested at the time of second-stage basilic vein transpositions that would have otherwise been discarded. Samples were fixed in 10% formalin prior to paraffin embedding and cut into 5-μm thick sections.

Animal Models

All animal experiments were performed in strict compliance with federal guidelines and with approval from the Institutional Animal Care and Use Committee of Yale University. Anesthesia and analgesia were given according to protocol; no heparin or antibiotics were used.

Cdh5-Cre/ERT2 (kindly provided by Prof. Ralf H. Adams, London, UK) mice were crossed with tdTomato reporter mice (B6.Cg-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J) to generate Cdh5-Cre/ERT2;ROSA26-tdTomato transgenic mice, where endothelial cells were labeled with tdTomato following tamoxifen (80 mg/kg/d i.p. for 5 days) injection.

Myh11-Cre/ERT2 mice were crossed with tdTomato reporter mice (B6.Cg-Gt(ROSA)26Sortm14(CAG-tdTomato)Hze/J) to generate Myh11-Cre/ERT2;ROSA26-tdTomato transgenic mice, where smooth muscle cells were labeled with tdTomato following tamoxifen (80 mg/kg/d i.p. for 5 days) injection.

PF4cre × Gt (ROSA) 26mT/mG (PF4-Cre × mT/mG) reporter mice were generated as previously reported; PF4cre drives membrane GFP expression in megakaryocytes and platelets, while all other cells are labeled with mT.[20]

At the time of surgery 9- to 11-week-old mice were used. Female and male C57Bl/6 mice, Cdh5-Cre/ERT2;ROSA26-tdTomato transgenic mice, PF4-Cre × mT/mG reporter mice were used; since the bacterial artificial chromosome containing Myh11-CreERT2 inserts on the Y chromosome, only male Myh11-Cre/ERT2;ROSA26-tdTomato transgenic mice were used. Microsurgical procedures were performed aseptically in a dedicated facility using a dissecting microscope (Leica MZ 95). Anesthesia was administered using 2 to 2.5% isoflurane, and extended-release buprenorphine (Ethiqa XR; North Brunswick, NJ) was administered subcutaneously (3.25 mg/kg body weight) for intraoperative and postoperative analgesia.

Mouse AVT model: After exposing the IVC and aorta, the aorta was clamped just below the renal artery, and an arteriovenous fistula was created by puncturing the distal aorta into the IVC using a 25-gauge or 30-gauge needle; the surrounding connective tissue was used to cover the aortic puncture site and gentle pressure was applied to achieve hemostasis, as we previously described.[18] Direct visualization of disturbed arteriovenous blood flow from the aorta to the IVC was used to assess initial technical success of AVF creation. No veins were ligated, neither proximally nor distally. In this model, the AVT forms on the exposed elastin, collagen, and extracellular matrix of the IVC and the aorta, along the needle track, and eventually protrudes into the IVC lumen mirroring the path of the high pressure disturbed arteriovenous flow through the fistula.[18]

Mouse inferior vena cava (IVC) ligation induced VT model: Briefly, after exposing the infra-renal IVC, all sides and posterior venous branches were ligated with 7–0 Prolene suture, and then the IVC was ligated below the renal vein using 6–0 suture.

Rat jugular-carotid AVF model (male Wistar rats, 6 to 8 weeks): After induction of anesthesia with inhaled isoflurane, a midline neck incision was made, and right carotid artery (CA) and jugular vein (JV) were exposed. The JV was then clamped proximally and ligated distally in the neck; the CA was clamped with atraumatic microvascular clamps and a 1-mm longitudinal arteriotomy was made. The JV was then anastomosed to the CA in an end-to-side configuration using a running 10–0 polypropylene suture to form a tension-free anastomosis. Hemostasis was achieved with gentle manual pressure. The skin incision was then closed in two layers using running 5–0 and 4–0 absorbable sutures. No heparin or antibiotics were used in this model.[19] [21] In this model, the AVT forms at the AVF anastomosis (the interface of exposed elastin, collagen, extracellular matrix of the jugular vein and carotid artery, and disturbed arteriovenous flow).

Pig CA-JV AVG model: Briefly, Yorkshire male pigs (mean weight, 48 kg; age, 3.4 months) were used. For anesthesia, pigs were induced with tiletamine-zolazepam (Telazol; 4.4 mg/kg), ketamine (2.2 mg/kg), and dexmedetomidine (0.02 mg/kg) and maintained on inhaled isoflurane (1–4%). A longitudinal midline incision was made in the neck, and the common carotid artery and internal jugular vein were exposed. The vessels were carefully controlled with silastic vessel loops, and a single heparin bolus (100 units/kg) was given and allowed to circulate for 3 minutes before vessel clamping; both the arteriotomy and the venotomy were 8 mm. Expanded polytetrafluoroethylene (ePTFE) (6-mm diameter, 6- to 7-cm length; W. L. Gore & Associates, Flagstaff, Ariz) was used as a graft. End-to-side anastomoses were created at 45 degrees between the graft and native vessel using continuous 6–0 polypropylene suture. In this model, the AVT forms at the graft–vein anastomosis (the interface of venous wall injury, the prosthetic material, and disturbed arteriovenous flow). The AVG and vessels were harvested at 1 week for histological analysis.[10]

Seprafilm (Baxter, United states; sodium hyaluronate and carboxymethylcellulose [HA/CMC]; biodegradable) was cut into 3 mm × 3 mm patches, and 10 µL saline, 10 µL 1% rhodamine solutions, or 10 µL rapamycin solution (10 mg/ml, pure ethanol) was added onto the Seprafilm patches and the liquid allowed to vaporize. The Seprafilm patch loaded with saline, rhodamine, or rapamycin was gently placed onto the JAA; samples were harvested at different time points and examined by histology.

Macroscopic Photographs

The mouse AVT samples and IVC ligation induced VT samples were harvested, and photographs were taken using the dissection microscope at different time points.

Histology

Animals were euthanized and perfused with normal saline followed by 10% formalin via the left ventricle under physiological pressure, and the AVF was extracted and then embedded in paraffin, cut in 5-μm cross sections, and stained with hematoxylin and eosin (H&E) and Elastin van Gieson (EVG).

Immunofluorescence

Tissue sections were deparaffined using xylene and rehydrated in a graded series of alcohols. Sections were heated in citric acid buffer (pH 6.0) at 100°C for 10 minutes for antigen retrieval. The sections were then blocked with 2% bovine serum albumin (BSA) for 1 hour at room temperature, before incubation overnight at 4 °C with the primary antibodies (CD31, R&D, AF3628, 1:400; α-SMA, Abcam, ab5694, 1:500; PCNA, Abcam, ab92552; SM22 α, Abcam, ab10135, 1:500; vWF, Abcam, ab6994, 1:200;). Sections were then treated with secondary antibodies at room temperature for 1 hour and stained with 4′,6-diamidino-2-phenylindole (DAPI, P36935; Invitrogen).

For Cdh5-Cre/ERT2;ROSA26-tdTomato transgenic mice, Myh11-Cre/ERT2;ROSA26-tdTomato transgenic mice, PF4-Cre × mT/mG reporter mice, mice were perfused with 4% paraformaldehyde, and the vessels were fixed overnight in 4% paraformaldehyde at 4°C, cryoprotected in 15% sucrose for 6 to 8 hours at 4°C, embedded in OCT, and 6-μm-thick sections were obtained. Length and thickness were measured using ImageJ (NIH Image, Bethesda, MD).[22]

Whole Mount Staining

The AVT (day 3) was harvested and gently dissected away from the aorta, and then the AVT was compressed between two slides. Samples were blocked in 1% BSA for 30 minute and then incubated overnight at 4°C with the primary antibodies. Sections were then treated with secondary antibodies at room temperature for 1 hour and stained with 4′,6-diamidino-2-phenylindole (DAPI, P36935; Invitrogen).

Doppler Examination

Doppler ultrasound was used to detect the AVT and disturbed flow at the JAA in the mouse model on day 21. The Vevo 2 High Resolution Imaging System (VisualSonics, WA, USA) with probe UHF46X was used in all experiments. Color Doppler mode was used to confirm the presence of disturbed flow.

Sample Processing for Scanning Electron Microscopy

Samples of AVT acquired at different time points were fixed with 4% paraformaldehyde overnight at 40°C, followed by further fixation once the sample was pinned open and then fixed with 2.5% glutaraldehyde in 0.1 M sodium cacodylate buffer pH7.4 for 1 hour. Samples were rinsed in 0.1 M sodium cacodylate buffer and post fixed in 2% osmium tetroxide in 0.1 M sodium cacodylate buffer pH 7.4. These were rinsed in buffer and dehydrated through an ethanol series from 30 to 100%. The samples were dried using a Leica 300 critical point dryer with liquid carbon dioxide as transitional fluid. The samples were glued to aluminum stubs using a carbon graphite adhesive, and sputter coated with 6-nm platinum/palladium using a Cressington 208HR.

Digital images were acquired in Zeiss CrossBeam 550 between 1.5–2kV at a working distance of 8–12 mm.

Statistical Analysis

All data were analyzed using Prism 10 software (GraphPad Software, Inc, La Jolla, CA). n represents numbers of animals and error bars represent the SEM. The Shapiro-Wilk test was used to test for normality; the F test was performed to evaluate homogeneity of variances. For normally distributed data, two-group comparisons were performed with the unpaired Student t-test. P values of ≤0.05 were considered to indicate statistical significance.

Results

AVT Formation and Organization at the Juxta-Anastomotic Area (JAA) under Arteriovenous Flow in a Mouse Model

At the JAA in the mouse AVT model ([Fig. 1A]), the AVT can clearly been seen through the thin IVC wall on day 21 ([Fig. 1B]). Doppler examination confirmed that the AVT was at the JAA, in the presence of high arteriovenous flow in the fistula channel ([Fig. 1C]). Color Doppler also showed disturbed arteriovenous flow at the dilated JAA ([Fig. 1C]). The color of the acute AVT was white compared with the typical red early-stage venous thrombus (VT; [Fig. 1D]). Although VT, in a ligation induced VT model, extended significantly from day 1 to 7, the AVT did not grow significantly larger (gross pathology) from day 1 to 21 ([Fig. 1E]; [Supplementary Figure S1], [S2] (available in the online version only)).

Fig. 1 Gross view of the arteriovenous thrombus (AVT) at the juxta-anastomotic area (JAA) under arteriovenous flow in a mouse model. (A) Diagram showing the AVT model was established by puncturing from the aorta (Ao) to inferior vena cava (IVC) using a 25-gauge needle in mice. (B) Photograph showing the AVT through the thin IVC wall at day 21. (C) Ultrasound showing the AVT, arteriovenous flow at the JAA, and velocity in the fistula channel at day 21. (D) Venous thrombus (VT) induced by IVC ligation at 6 hours, aorta, AVT (dashed red line square) at 1 hour (h), 3 h and 6 h. Scale bar, 5 mm; n = 6. (E) Gross anterior and posterior views of AVT (dashed red line square) on days 1, 3, 7, and 21; scale bar, 5 mm; n = 6. (Illustration created by H.B using Microsoft PowerPoint.)

Scanning electron microscopy (SEM) was used to characterize the early and late AVT in this mouse model. Immediately after needle puncture, there was a small acute thrombus on the venous side ([Fig. 2A]). After 1 hour, the AVT became increasingly visible with more fibrin nets covering the AVT surface at 3 hours compared with 1 hour ([Fig. 2B]). Interestingly, the AVT became more compact at 6 hours compared with 1 and 3 hours ([Fig. 2B]). At day 21, the AVT had a smooth surface, and endothelial cells covered the body of the AVT ([Fig. 2C]). Higher magnification SEM images showed that fibrin and platelets were visible on the AVT surface ([Fig. 2D]), while EVG staining showed that red blood cells (RBC) were present inside the AVT ([Supplementary Fig. S3], available in the online version only).

Fig. 2 Scanning electron microscope (SEM) photos of acute, organized arteriovenous thrombus (AVT). (A) SEM photos showing the acute AVT immediately after needle puncture, n = 3. (B) SEM photos showing the acute AVT at 1, 3, and 6 hours; n = 3. (C) SEM photos showing the organized AVT at day 21; note the full coverage of endothelial cells on the AVT body; n = 3. (D) SEM photos showing the platelets (red arrowhead) and fibrin (yellow arrowhead) in the AVT immediately after puncture and at day 3; scale bar, 2 μm; n = 3. Ao, aorta; IVC, inferior vena cava.

To see the location of early AVT origination, AVT was harvested at 6 hours. Both EVG and immunofluorescence staining images showed AVT originated from the anastomosis surface without CD31 positive cell coverage ([Fig. 3A, B]). Not surprisingly, compared with the AVT area with a patent AVF at day 21, there was a significantly larger AVT area in occluded AVF ([Fig. 3C, D]). Longitudinal sections also showed that AVT localized at the JAA at day 21 ([Fig. 3E]). Temporal sections from day 1 to 21 showed F4/80 positive macrophage cells accumulated in the AVT at day 3 and decreased after day 7; α-actin positive cells infiltrated and increased with the same time course, and CD31 positive cells gradually covered the AVT body surface ([Fig. 3F]). Since 5-μm thick histological AVT sections show only a small sample of the AVT, day 3 AVT was whole mounted to establish a more complete characterization of AVT formation. There were few long (≈100 μm) α-actin positive cells in the AVT, as well as individual or cluster of fishnet pattern-like CD31-positive cells ([Fig. 3G]).

Fig. 3 Histological structure and organization of arteriovenous thrombus (AVT). (A) Verhoeff Van Gieson (EVG) staining showing native aorta and inferior vena cava (IVC), AVT at 6 hours; black arrow shows the native IVC wall; dashed black square shows AVT; scale bar, 1 mm or 200 μm; n = 6. (B) Immunofluorescence showing CD31 (green), α-actin (red), and 4′,6-diamidino-2-phenylindole (DAPI) (blue) of the native aorta and IVC, the anastomosis at 6 hours after arteriovenous fistulae (AVF) creation; white arrow showing loss of CD31 positive cells coverage and exposure of extracellular matrix; scale bar, 100 μm; n = 3. (C) Hematoxylin and eosin (H&E) staining showing AVT in patent and occluded AVF at day 21; scale bar, 1 mm. (D) Bar graph showing the AVT area/total luminal area in the patent and occluded AVF at day 21 (*), p = 0.0085, n = 6. (E) Longitudinal section stained with H&E at day 21 showing the outflow vein and stenotic JAA filled with AVT; scale bar, 1 mm; n = 3. Ao, aorta; black dashed line demarcates the outflow vein and JAA. (F) Immunofluorescence images showing the temporal changes of F4/80 positive cells, CD31 and α-actin positive cells in the AVT from day 1 to 21; scale bar, 100 μm; n = 6. White arrowheads show the IVC endothelial cells that are CD31 positive. (G) Whole mount staining images showing the α-actin and CD31 positive cells in the day 3 AVT, n = 3.

These data show that AVT forms immediately and organized gradually at the JAA under arteriovenous flow in the mouse model. Infiltrated macrophages, smooth muscle cell expansion, and endothelial cell coverage also contributed to AVT organization.

Platelet, Smooth Muscle Cell and Endothelial Cell Lineage Tracing in AVT

Since both endothelial cells (EC) and smooth muscle cells (SMC) can regulate AVT organization, we traced both cells and platelets using different reporter mice. PF4-Cre × mT/mG reporter mice were used to trace megakaryocyte-platelets as we previously described.[23] The GFP positive area occupied approximately 80% of the AVT area at day 7 and significantly decreased to approximately 20% at day 21 ([Fig. 4A, B]). Myh11-Cre/ERT2;ROSA26-tdTomato transgenic mice were used to trace SMC. There were few Myh11 positive cells in the AVT (day 7; [Fig. 4C]). The Myh11 positive cells occupied approximately 10% of the total cells and less than 5% of the AVT area ([Fig. 4C, D]). Most of the Myh11 positive cells were near the anastomosis (yellow arrow, [Fig. 4C]). Cdh5-Cre/ERT2;ROSA26-tdtomato transgenic mice were used to trace EC at day 21; Cdh5 positive cells were mainly on the AVT surface but not inside the AVT ([Fig. 4E]). The Cdh5 positive cells only occupied approximately 20% of the total cells and less than 5% of the AVT area ([Fig. 4E, F]); the middle portion of the AVT was not covered by Cdh5 positive cells ([Fig. 4E]). These data show that platelets gradually degraded, and adjacent EC and SMC migrated and infiltrated into the AVT to contribute to AVT organization.

Fig. 4 Platelet, smooth muscle cell and endothelial cell lineage tracing in arteriovenous thrombus (AVT). (A) Photographs showing platelets (GFP) in the AVT at day 7 and 21 in PF4cre × Gt (ROSA) 26mT/mG mouse; scale bar, 1 mm; n = 6. (B) Bar graph showing the GFP area in the AVT at day 7 and 21, *, p < 0.0001, t-test; n = 6. (C) AVT in Myh11-Cre/ERT2;ROSA26-tdTomato transgenic (Myh11) mice at day 7; α-actin (green), Myh11 (tomato), and 4′,6-diamidino-2-phenylindole (DAPI) (blue). High power photographs showing AVT stained with α-actin (green), SM22 α (green), and DAPI (blue); yellow arrows showing Myh11 positive cells in the AVT; scale bar, 1 mm or 100 μm; n = 6. (D) Bar graph showing the Myh11 positive cell percentage and area, n = 6. (E) AVT in Cdh5-Cre/ERT2;ROSA26-tdtomato transgenic (Cdh5) mice at day 21; CD31 (green), Cdh5 (tomato), and DAPI (blue). High power photographs showing AVT stained with CD31 (green), α-actin (green), and DAPI (blue); yellow arrow showing dual CD31 and Cdh5 positive cells in the AVT; scale bar, 1 mm or 100 μm; n = 3. Zoomed in images show the white dashed line rectangle area (not covered by Cdh5 positive cells). (F) Bar graph showing the Cdh5 positive cell percentage and area, n = 6.

Smaller Injury to the Vessel but not Anti-proliferation Decreases the Volume of AVT

Because the AVT originates from the injured wall at the anastomosis site, we assessed whether a smaller injury would induce a smaller AVT with a different composition. We compared injury from a 30-G (0.31 mm) smaller diameter needle with a 25-G needle (0.51 mm). There was a smaller AVT in the 30-G needle group compared with 25-G needle group at day 7 ([Fig. 5A, B]). Immunofluorescence showed α-actin positive cells interspersed inside the CD41 positive AVT; there were also vWF and fibrinogen α positive areas inside the AVT in both groups ([Fig. 5C]). There were similar percentages of PCNA positive cells (≈80%, PCNA positive cells/total number of cells), mostly smooth muscle cells, in both groups ([Fig. 5C, D]). Since there was a high cell proliferation rate in the AVT at day 7, we also explored whether local delivery of rapamycin to inhibit cell proliferation could decrease the volume of the AVT. Seprafilm was used as a drug delivery system since it biodegrades within 3 weeks ([Fig. 5E]). As a positive control, local delivery of rhodamine (water soluble) using Seprafilm showed strong fluorescence on the vessel wall at day 3 ([Fig. 5F]). However, local delivery of rapamycin using Seprafilm failed to decrease the volume of AVT at day 21 ([Fig. 5G, H]). These data show that a smaller injury of the vessel wall, but not antiproliferative drugs, is more effective in decreasing the volume of an AVT.

Fig. 5 Smaller endothelium injury but not anti-proliferation decreases the volume of arteriovenous thrombus (AVT). (A) Gross view and Elastin van Gieson (EVG) staining showing the AVT in the 25-G and 30-G arteriovenous fistulae (AVF) at day 7; scale bar, 1 mm; n = 6. (B) Bar graph showing the AVT area in the 25-G and 30-G AVF at day 7, *, p < 0.0001, t-test; n = 6. (C) Merged immunofluorescence stained for α-actin (green), CD41 (red), and 4′,6-diamidino-2-phenylindole (DAPI) (blue); SM22α (green), PCNA (red), and DAPI (blue); in the AVT induced by 25-G and 30-G needles at day 7. Scale bar, 100 μm; n = 6. (D) Bar graphs showing the PCNA positive cells in the AVT induced by 25-G and 30-G needles; p = 0.8337, t-test; n = 3. (E) Biodegradable Seprafilm at day 7 and 21 after installation on the aorta and IVC wall in mouse, sections were stained with H&E; black arrow showing the Seprafilm degraded from day 7 to day 21; n = 3. (F) Fluorescence images showing Separfilm loaded with rhodamine at day 3 after installation, white arrows showing the location of Seprafilm, red arrows showing the rhodamine fluorescence; scale bar, 1 mm or 200 μm; n = 3. (G) EVG staining showing AVT at day 21 with Seprafilm alone (control) and with Seprafilm loaded with rapamycin; scale bar, 1 mm; n = 6. (H) Bar graph showing mean AVT area in the control and rapamycin groups at day 21, *, p = 0.7406, t-test; n = 3.

Incomplete Endothelialization at the Peri-fistula Exit Area

Since there is little growth of the AVT in patent AVF between days 3 and 21 when viewed directly ([Fig. 1E]), and cross sections show different sizes of AVT at different locations ([Figs. 1D] and [6H]),[18] we examined the AVT length with longitudinal sections. There was an increased length of AVT (from the anastomosis to the fistula exit) at day 21 compared with day 3 ([Fig. 6A]). Accordingly, we explored the endothelial cell coverage of the peri-fistula exit area. At days 7 and 21, there was incomplete CD31 positive EC coverage at the peri-fistula exit area (yellow arrow, [Fig. 6B]). SEM was used to examine the day 21 AVT to get a comprehensive understanding of the peri-fistula exit area. There was incomplete EC coverage in the peri-fistula exit area compared with the full coverage of the EC on the AVT body surface (black arrow, [Fig. 6C]). There were also a large number of platelets and fibrin fibers at the fistula exit ([Fig. 6C]). These data show that incomplete endothelialization at the peri-fistula area might be correlated with AVT growth under disturbed arteriovenous flow in this model.

Fig. 6 Incomplete endothelialization at the peri-fistula exit area is associated with the growth of arteriovenous thrombus (AVT). (A) Longitudinal sections stained with Elastin van Gieson (EVG) showing the AVT at day 3 and 21, black dashed line rectangle showing the AVT; bar graph showing the AVT length; *, p = 0.017; scale bar, 1 mm; n = 3. (B) Sections at the fistula exit stained with CD31 (green) and 4′,6-diamidino-2-phenylindole (DAPI) (blue) at day 7 and 21, yellow arrow showing the incomplete endothelialization at the peri-fistula exit area; scale bar, 1 mm or 200 μm; n = 3. (C) SEM images showing the incomplete endothelialization, platelets accumulation, and fibrin formation at the peri-fistula exit area compared with the AVT body at day 21; n = 3.

AVT is Universally Present at the JAA in Other Models of Vascular Access

Since AVT forms and grows in this mouse needle puncture model, we also examined whether AVT is present at the JAA in other vascular access models. In the rat jugular vein (JV) to carotid artery (CA) sutured AVF ([Fig. 7A]), an AVT immediately formed at the JAA ([Fig. 7B]), and there was decreased AVT area at 24 h compared with 3 h ([Fig. 7C]). The AVT was not clearly present after day 7, as we previously showed.[19] [21] Both at days 7 and 42, there was incomplete CD31 positive cell coverage at the JAA ([Fig. 7D, E]) as well as decreased vWF positive layer coverage at day 42 compared with day 7 ([Fig. 7D, E]). Immunofluorescence also showed fibrinogen α and CD41 positive staining on the surface at the JAA both at days 7 and 42 ([Fig. 7D]).

Fig. 7 Arteriovenous thrombus (AVT) universally presents at the juxta-anastomotic area (JAA) under other arteriovenous flow conditions. (A) Illustration picture of the rat JV-CA arteriovenous fistulae (AVF) model. (B) Elastin van Gieson (EVG) staining showing the AVT at 0, ½, 1, 3, 6, and 24 h at the JAA; black dashed eclipse and black arrow showing AVT, scale bar 1 mm or 400 μm, n = 3. (C) Bar graph showing AVT area at 0, 3, and 24 h; p = 0.002, one-way ANOVA; *, p = 0.002, versus 0 h; **, p = 0.0024, versus 3 h. n = 3. (D) Merged immunofluorescence stained for CD31 (red), α-actin (green), and 4′,6-diamidino-2-phenylindole (DAPI) (blue); vWF (green), fibrinogen α (red), and DAPI (blue); α-actin (green), CD41 (red), and DAPI (blue); at the JAA. Dotted red line arrow showing incomplete CD31 positive cell coverage at the JAA, green arrow showing vWF positive staining, yellow arrow showing the CD41 positive platelets; white dashed line area showing the AVT; CA, carotid artery; JV, jugular vein; scale bar, 1 mm and 200 μm; n = 3. (E) Bar graph showing CD31 positive coverage (p = 0.9043, t-test) and vWF positive cell coverage (*, p < 0.0001, t-test); n = 3. (F) Sections stained with α-actin (green), CD41 (red), and DAPI (blue); vWF (green), fibrinogen α (red), and DAPI (blue); in the human native vein and AVT. Scale bar, 200 μm; n = 3. (G) Bar graph showing CD41 positive area (*, p = 0.0007), α-actin positive area (*, p = 0.0002), vWF positive area (*, p < 0.0001, t-test), and fibrinogen α (*, p = 0.0003); t-test, n = 3. (H) Illustration of the pig carotid artery (CA)–jugular vein (JV) AVG model using expanded polytetrafluoroethylene (ePTFE). (I) EVG staining showing the AVT at the graft–vein anastomosis at day 7; red arrow showing the blood flow direction, scale bar, 1 mm and 200 μm; n = 3. (J) Sections stained with α-actin (green), CD41 (red), and DAPI (blue); vWF (green), fibrinogen α (red), and DAPI (blue); of the AVT; scale bar, 200 μm; n = 3. (Illustration created by H.B. using Microsoft PowerPoint.)

In human sutured basilic vein to brachial artery AVF, there was an increased area of CD41 positive platelets, decreased area of α-actin positive cells, increased area of vWF positive cells, and increased area of fibrinogen α in the AVT compared with the native vein ([Fig. 7F, G]). In the pig arteriovenous (jugular vein to carotid artery) graft (ePTFE) model ([Fig. 7H]), an AVT was present at the graft–vein anastomosis at day 7, mainly on the vein wall but not on the luminal side of the graft ([Fig. 7I]); vWF, fibrinogen α, and CD41 were detected in the AVT ([Fig. 7J]). These data show that formation of AVT may be a universal phenomenon at the JAA under disturbed arteriovenous flow.

Discussion

We show formation, organization, and growth of AVT at the JAA in a mouse model. Platelets are a major component of the early AVT. Smaller injury to the endothelium, but not antiproliferative drugs, effectively decreased AVT volume. Incomplete EC coverage at the peri-fistula exit area correlated with slow growth of the AVT. Our data advance the understanding of AVT under arteriovenous fistula flow, and may have a translational application in the prevention of AVT growth in human AVF and AVG.

Compared with the significant understanding of VT and AT,[24] [25] AVT has not been explored in pathophysiological detail in both human subjects and in animal models.[1] There is still limited knowledge on AVT, with one potential reason being the lack of an animal model.[9] Another reason might be the frequent focus on late-stage neointimal hyperplasia (after day 7) in many studies, and thus temporal observation at early time points (before day 3) has not been commonly reported. Moreover, it may be difficult to discriminate the small organized AVT from neointima.[23] [26] [27] We show that AVT has a different morphology compared with the mouse VT induced by IVC total ligation ([Fig. 1], [Supplementary Fig. S1], available in the online version only),[28] suggesting different mechanisms of formation and progression between AVT and VT. AVT originates from the injured needle puncture route (exposed elastin, collagen, and extracellular matrix; [Fig. 3A, B]), whereas VT is largely induced by static or decreased blood flow.[28] AVT forms rapidly and then grows slowly ([Figs. 1E] and [6]). The fistula exit can face different directions inside the IVC lumen ([Fig. 2C]), resulting in variable distance from the fistula exit to the IVC wall. When the fistula exit directly contacts the contralateral wall, fistula occlusion can occur; therefore, we categorized the AVT into asymptomatic (silent) or symptomatic AVT in this mouse model. Asymptomatic AVT results in varying degrees of luminal stenosis but remains patent after day 21 ([Fig. 3C], patent). Symptomatic AVT results in occlusion after day 21 in this model ([Fig. 3C], occluded). Notably, the sharp difference between VT and AVT suggests, at least in part, differential responsiveness to antiplatelet and anticoagulation therapies.

Decreasing the volume and growth of AVT may decrease the severity of JAA stenosis and the potential of vascular access failure. We previously showed that smaller injury to the vessels induces less neointimal hyperplasia compared with increased injuries in several models.[29] Similarly, we showed that a smaller diameter needle induces smaller injury and a smaller AVT, supporting the “no touch” technique that is associated with reduced AVT compared with extensive dissection.[19] Although rapamycin is an effective drug to inhibit SMC proliferation and neointimal hyperplasia in both arterial and venous systems,[29] [30] rapamycin failed to decrease the size of the AVT in this mouse AVF model ([Fig. 5]). The failure of rapamycin to decrease AVT may be because cells do not predominate in the early AVT in this mouse model, analogous to the uncertain role of rapamycin to inhibit cell proliferation in the treatment of human stenotic AVF.[31] [32] [33] Although AVT is exposed to the disturbed arteriovenous flow (both arterial and venous flows), AVT appears more like the AT than VT.[1] [34] Current KDOQI guidelines do not suggest antiplatelet and anticoagulation therapies to increase AVF maturation[14]; as such, we did not include both therapies in this research since these additional experiments need comprehensive testing of mechanisms and temporal patterns to be rigorous.

Our team has been using this mouse AVF model for more than 15 years. Although we briefly described AVT formation,[35] we have focused on the outflow vein neointima hyperplasia in most of our reports. More recently, we briefly showed some AVT data in our description of the complex temporal and spatial changes of the JAA in this mouse model.[18] [23] Here, we used three reporter mice, temporal analyses of gross and SEM images, different methods to decrease AVT, comparison of AVT to VT, as well as showing data from different models of vascular access to illustrate the details of AVT complexity. We also found this model has clinical translational applications for the approximately half a million patients in the United States who are dependent on hemodialysis; the incidence of thrombus under arteriovenous flow has been reported to be between 5 and 15%, and accounts for 65 to 85% of cases of permanent access loss; AVT occurs 0.1 to 0.5 times in AVF and approximately 0.5 to 2.0 times in AVG per year.[4] [36] We show that different sizes of AVT were found at the JAA under disturbed arteriovenous flows ([Fig. 5]). In humans, symptomatic AVT cases that induce occlusion are treated promptly, whereas how asymptomatic (silent) AVT transform at the JAA has not been investigated, despite many studies showing a potential relation of early asymptomatic AVT and late neointima hyperplasia in humans.[36] [37] [38] Thus, examination both of early AVT and late neointimal hyperplasia may be a promising method to understand the mechanism of AVF failure, and combined therapy addressing both mechanisms might be more effective than single antiplatelet or antiproliferation strategy.

In conclusion, AVT is universally present at the JAA under disturbed arteriovenous flow in different models of vascular access. Continuous AVT growth contributes to JAA stenosis in this mouse model. Our data advance the understanding of the unique nature of AVT and may have translational implications in preventing AVT formation and growth in human patients.

Contributors' Statement

H.B., Z.L., and A.D. conceived the ideas, designed experiments, performed data analysis, wrote and revised the manuscript; B.H, Y.C., and J.H. helped with the manuscript; A.D. obtained funding. All authors approved the final version of the manuscript.

Conflict of Interest

The authors declare that they have no conflict of interest.

Acknowledgment

We thank Dr. Shirley Liu et al for the pig AVG samples, Dr. Yasin Oduk for the Seprafilm, and Morven Graham for taking the SEM images. The authors would like to thank the Center for Cellular and Molecular Imaging, and the Electron Microscopy Facility at the Yale School of Medicine for assistance with the work presented here.

Data Availability Statement

The authors confirm that the data supporting the results of this study are available in the article. If required, further information can be obtained on request from the corresponding author Dr. Alan Dardik ([e-mail]). The authors will share more detailed information about any of the animal models or analytical procedures with anyone who wishes to use these models or procedures in their research.

^‡ These authors contributed equally to this article.

• References

• 1 Alkarithi G, Duval C, Shi Y, Macrae FL, Ariëns RAS. Thrombus structural composition in cardiovascular disease. Arterioscler Thromb Vasc Biol 2021; 41 (09) 2370-2383
• 2 Undas A, Ariëns RA. Fibrin clot structure and function: a role in the pathophysiology of arterial and venous thromboembolic diseases. Arterioscler Thromb Vasc Biol 2011; 31 (12) e88-e99
• 3 Kattula S, Byrnes JR, Wolberg AS. Fibrinogen and fibrin in hemostasis and thrombosis. Arterioscler Thromb Vasc Biol 2017; 37 (03) e13-e21
• 4 Korn A, Alipour H, Zane J. et al. Factors associated with early thrombosis after arteriovenous fistula creation. Ann Vasc Surg 2018; 49: 281-284
• 5 Gibson KD, Gillen DL, Caps MT, Kohler TR, Sherrard DJ, Stehman-Breen CO. Vascular access survival and incidence of revisions: a comparison of prosthetic grafts, simple autogenous fistulas, and venous transposition fistulas from the United States Renal Data System Dialysis Morbidity and Mortality Study. J Vasc Surg 2001; 34 (04) 694-700
• 6 Kanterman RY, Vesely TM, Pilgram TK, Guy BW, Windus DW, Picus D. Dialysis access grafts: anatomic location of venous stenosis and results of angioplasty. Radiology 1995; 195 (01) 135-139
• 7 Singh D, Harita C. A tribute to the inventor of arteriovenous fistula for hemodialysis: James E Cimino. Indian J Vasc Endovasc Surg 2019; 6(2): 127-128
• 8 Lok CE, Huber TS, Orchanian-Cheff A, Rajan DK. Arteriovenous access for hemodialysis: a review. JAMA 2024; 331 (15) 1307-1317
• 9 Li Y, Hu K, Li Y, Lu C, Guo Y, Wang W. The rodent models of arteriovenous fistula. Front Cardiovasc Med 2024; 11: 1293568
• 10 Liu S, Wang T, Wang J. et al. Reduced patency in left-sided arteriovenous grafts in a porcine model. J Vasc Surg 2020; 72 (01) 305-317.e6
• 11 Bai H, Dardik A, Xing Y. Decellularized carotid artery functions as an arteriovenous graft. J Surg Res 2019; 234: 33-39
• 12 Eikelboom JW, Connolly SJ, Bosch J. et al; COMPASS Investigators. Rivaroxaban with or without aspirin in stable cardiovascular disease. N Engl J Med 2017; 377 (14) 1319-1330
• 13 Stevens SM, Woller SC, Kreuziger LB. et al. Antithrombotic therapy for VTE disease: second update of the CHEST Guideline and expert panel report. Chest 2021; 160 (06) e545-e608
• 14 Lok CE, Huber TS, Lee T. et al; National Kidney Foundation. KDOQI clinical practice guideline for vascular access: 2019 update. Am J Kidney Dis 2020; 75 (4, Suppl 2): S1-S164
• 15 Hanasono MM, Butler CE. Prevention and treatment of thrombosis in microvascular surgery. J Reconstr Microsurg 2008; 24 (05) 305-314
• 16 Azuma T, Fukushima T. Disturbance of blood flow as a factor of thrombus formation. Thromb Res 1976; 8 (2, suppl): 375-380
• 17 Ene-Iordache B, Semperboni C, Dubini G, Remuzzi A. Disturbed flow in a patient-specific arteriovenous fistula for hemodialysis: multidirectional and reciprocating near-wall flow patterns. J Biomech 2015; 48 (10) 2195-2200
• 18 Bai H, Varsanik MA, Thaxton C. et al. Disturbed flow in the juxta-anastomotic area of an arteriovenous fistula correlates with endothelial loss, acute thrombus formation, and neointimal hyperplasia. Am J Physiol Heart Circ Physiol 2024; 326 (06) H1446-H1461
• 19 Bai H, Sadaghianloo N, Gorecka J. et al. Artery to vein configuration of arteriovenous fistula improves hemodynamics to increase maturation and patency. Sci Transl Med 2020; 12 (557) eaax7613
• 20 Zeng Z, Xia L, Fan X. et al. Platelet-derived miR-223 promotes a phenotypic switch in arterial injury repair. J Clin Invest 2019; 129 (03) 1372-1386
• 21 Bai H, Wei S, Xie B. et al. Endothelial nitric oxide synthase (eNOS) mediates neointimal thickness in arteriovenous fistulae with different anastomotic angles in rats. J Vasc Access 2022; 23 (03) 403-411
• 22 Schneider CA, Rasband WS, Eliceiri KW. NIH Image to ImageJ: 25 years of image analysis. Nat Methods 2012; 9 (07) 671-675
• 23 Bai H, Li Z, Zhang W. et al. Early thrombus formation is required for eccentric and heterogeneous neointimal hyperplasia under disturbed flow. J Thromb Haemost 2024; 22 (12) 3614-3628
• 24 Diaz JA, Obi AT, Myers Jr DD. et al. Critical review of mouse models of venous thrombosis. Arterioscler Thromb Vasc Biol 2012; 32 (03) 556-562
• 25 Khan F, Tritschler T, Kahn SR, Rodger MA. Venous thromboembolism. Lancet 2021; 398 (10294): 64-77
• 26 Torsney E, Mayr U, Zou Y, Thompson WD, Hu Y, Xu Q. Thrombosis and neointima formation in vein grafts are inhibited by locally applied aspirin through endothelial protection. Circ Res 2004; 94 (11) 1466-1473
• 27 Richter GM, Palmaz JC, Noeldge G, Tio F. Relationship between blood flow, thrombus, and neointima in stents. J Vasc Interv Radiol 1999; 10 (05) 598-604
• 28 Diaz JA, Saha P, Cooley B. et al. Choosing a mouse model of venous thrombosis: a consensus assessment of utility and application. J Thromb Haemost 2019; 17 (04) 699-707
• 29 Bai H, Sun P, Wu H. et al. The application of tissue-engineered fish swim bladder vascular graft. Commun Biol 2021; 4 (01) 1153
• 30 Bai H, Wu H, Zhang L. et al. Adventitial injection of HA/SA hydrogel loaded with PLGA rapamycin nanoparticle inhibits neointimal hyperplasia in a rat aortic wire injury model. Drug Deliv Transl Res 2022; 12 (12) 2950-2959
• 31 Barcena AJR, Perez JVD, Bernardino MR. et al. Controlled delivery of rosuvastatin or rapamycin through electrospun bismuth nanoparticle-infused perivascular wraps promotes arteriovenous fistula maturation. ACS Appl Mater Interfaces 2024; 16 (26) 33159-33168
• 32 Lookstein RA, Haruguchi H, Ouriel K. et al; IN.PACT AV Access Investigators. Drug-coated balloons for dysfunctional dialysis arteriovenous fistulas. N Engl J Med 2020; 383 (08) 733-742
• 33 Karunanithy N, Robinson EJ, Ahmad F. et al. A multicenter randomized controlled trial indicates that paclitaxel-coated balloons provide no benefit for arteriovenous fistulas. Kidney Int 2021; 100 (02) 447-456
• 34 Chernysh IN, Nagaswami C, Kosolapova S. et al. The distinctive structure and composition of arterial and venous thrombi and pulmonary emboli. Sci Rep 2020; 10 (01) 5112
• 35 Yamamoto K, Protack CD, Kuwahara G. et al. Disturbed shear stress reduces Klf2 expression in arterial-venous fistulae in vivo. Physiol Rep 2015; 3 (03) e12348
• 36 Quencer KB, Oklu R. Hemodialysis access thrombosis. Cardiovasc Diagn Ther 2017; 7 (Suppl. 03) S299-S308
• 37 Chang CJ, Ko YS, Ko PJ. et al. Thrombosed arteriovenous fistula for hemodialysis access is characterized by a marked inflammatory activity. Kidney Int 2005; 68 (03) 1312-1319
• 38 Roy-Chaudhury P, Arend L, Zhang J. et al. Neointimal hyperplasia in early arteriovenous fistula failure. Am J Kidney Dis 2007; 50 (05) 782-790

Publication History

Received: 25 September 2025

Accepted after revision: 14 December 2025

Accepted Manuscript online:
16 December 2025

Article published online:
31 December 2025