Early thrombus formation is required of eccentric and heterogenous neointimal hyperplasia under disturbed flow

Abstract

Background:

Anticoagulation and antiplatelet therapy effectively inhibit neointimal hyperplasia (NIH) in both arterial and venous systems but not in arteriovenous fistulae (AVF). The main site of AVF failure is the juxta-anastomotic area that is characterized by disturbed flow compared to laminar flow in the arterial inflow and the venous outflow. We hypothesize that early thrombus formation is required for eccentric and heterogenous NIH under disturbed flow.

Method:

Needle puncture and sutured AVF were created in C57BL/6 mice, in PF4-cre × mT/mG reporter mice, and in Wistar rats. Human AVF samples were harvested transpositions. The tissues were examined by histology, immunofluorescence, immunohistochemistry and en face staining.

Result:

Under disturbed flow, both mouse and human AVF showed eccentric and heterogenous NIH. Maladapted vein wall was characterized by eccentric and heterogenous neointima that was composed of a different abundance of thrombus and smooth muscle cells (SMC). PF4-cre × mT/mG reporter mice AVF showed that GFP-labeled platelets deposit on the wall directly facing the fistula exit with endothelial cells loss and continue to accumulate under disturbed flow. Neither disturbed flow with limited endothelial cell loss nor non-disturbed flow induced heterogenous neointima in different animal models.

Conclusion:

Early thrombus contributes to late heterogenous NIH in the presence of disturbed flow. Disturbed flow, large area of endothelial cell loss and thrombus formation are critical to form eccentric and heterogenous NIH. Categorization of adapted or maladapted walls may be helpful for therapy targeting heterogenous NIH.

Introduction

The nature of the relationship between thrombus formation and neointimal hyperplasia (NIH) has been poorly understood and controversial. Thrombus deposition has been observed on suture lines and venous graft luminal surfaces as early as three hours in animal models [1], and both heparin and aspirin decrease arterial and venous neointimal hyperplasia in preclinical models.[2],[3],[4],[5] The recently published VOYAGER PAD and COMPASS trials showed that anticoagulation is more effective compared to antiplatelet monotherapy [6], [7], suggesting involvement of coagulation in arterial NIH. In veins, anticoagulation prevents thrombosis and recurrence [8],[9] However, in AVF, neither antiplatelet therapy nor anticoagulation are effective to prevent AVF failure. In the Dialysis Access Consortium (DAC) trial, there was a 37% lower risk of thrombosis in the clopidogrel group compared to the placebo group; but this randomized, double-blinded clinical trial was stopped because of no significant benefit of clopidogrel.[10],[11] Since intraoperative anticoagulation using heparin during AVF creation also does not affect outcomes [12], the Kidney Disease Outcomes Quality Initiative (KDOQI) does not suggest the use of either heparin or clopidogrel to improve AVF maturation.[13] Moreover, pre-existing intimal hyperplasia in veins prior to AVF creation was not associated with AVF failure, suggesting that NIH is a pathological marker of injury and not a mechanism. [14], [15] Thus, the significance of early thrombus formation after vascular reconstruction remains unclear in AVF, and the clinical significance of inhibiting thrombus as a potential translational therapy is not understood.

There are different shapes of neointima after vascular surgeries in both human patients and animal models. Concentric and homogenous neointima, or eccentric and heterogenous neointima, forms under different flow patterns and types of procedures [16],[17],[18],[2]. Unfortunately, there is still no consensus on how to evaluate the severity of eccentric and heterogenous NIH [19],[20]. However, homogeneous and heterogeneous neointima can be categorized in patients using optical coherence tomography (OCT) [21]. After arterial stent implantation, heterogeneous neointima may have a high incidence of thrombus formation [22]. Furthermore, in-stent arterial neointimal area in lesions with homogeneous neointima decreased significantly, but neointimal area in lesions with heterogeneous neointima did not change significantly [23]. The importance of heterogenous neointima in the pathogenesis of restenosis has attracted attention. In a mouse vein graft model, vein graft mural thrombus may contribute to formation of heterogenous NIH [24]. These clinical and basic research data suggest a strong relationship between early thrombus and late heterogenous neointima.

Although increased laminar flow causes atrophy of an established neointima in models,[25],[26],[27] but the AVF juxta-anastomotic area (JAA) is characterized by highly disturbed flow [28]. We recently showed that focal endothelial cell (EC) loss at the JAA in the presence of disturbed flow is followed by thrombus deposition and NIH in both mice and rat AVF [28]. This research shows a complex relationship between thrombus and NIH under disturbed flow. We now hypothesize that early thrombus formation is required for eccentric and heterogenous neointimal hyperplasia under disturbed flow. We tested our hypothesis using animal models with disturbed or non-disturbed flow and using PF4-cre × mT/mG reporter mice.

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. One AVF sample with cannulation was obtained from an AVF revision secondary to severe arm swelling. 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.

PF4-cre×Gt (ROSA) 26mT/mG (PF4-Cre×mT/mG) reporter mice were generated as previously reported; PF4-cre drives membrane GFP expression in megakaryocytes and platelets, while all other cells are labeled with mT [29]. Mice were 9 to 11 weeks at time of surgery.

Mouse needle puncture AVF model (female and male C57Bl/6 mice without or with chronic kidney disease (CKD) induced by 5/6 nephrectomy; 9 to 11 weeks): The IVC and aorta were exposed and an AVF was created by puncturing the distal aorta into the IVC using a 25-gauge needle as previously described [28],[30]. Visualization of pulsatile arterial blood flow in the IVC was used to assess initial technical success of AVF creation.

Mouse carotid artery-jugular vein AVF model: After exposing the carotid artery and jugular vein, the jugular vein was clamped and a 0.5 mm venotomy was made using a 30G needle. The proximal carotid artery was clamped and the distal end was ligated using 8–0 suture. The carotid artery was anastomosed to the jugular vein in an end-to-side configuration. The AVF was harvested after 2 weeks.

Rat jugular-carotid AVF model (male Wistar rats, 6 to 8 weeks): The left carotid artery and jugular vein were exposed and a 1-mm longitudinal arteriotomy was made on the carotid artery; the jugular vein was anastomosed to the carotid artery in an end-to-side configuration [31].

Mouse aortic direct suture and patch angioplasty models: The mouse aorta was exposed and a 0.5 mm arteriotomy was made; the arteriotomy was closed a by running suture using 10–0 suture (direct suture group) [32]. In other mice, a pericardial patch was trimmed to approximately 1 mm × 1.5 mm, and the arteriotomy was closed with the pericardial patch using running 10–0 nylon suture (patch group) [33].

Mouse aortic clamp injury model: The mouse infrarenal aorta was dissected 5 mm in length. A microneedle holder was applied to the aorta for exactly 10 s and was then removed; the microneedle holder was reapplied slightly distally to the rest of the dissected aorta in sequential fashion [34].

Macroscopic photographs

The mouse AVF was harvested, and photographs were taken using the dissection microscope at day 7.

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 Elastin van Gieson (EVG).

Immunohistochemistry

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, sections were incubated overnight at 4°C with the primary antibody (α-SMA, Abcam, ab5694, 1:500). After overnight incubation, sections were incubated in EnVision reagents for 1 h at room temperature and treated with Liquid DAB Substrate Chromogen System (Dako). Finally, sections were counterstained with Dako Mayer’s Hematoxylin. Images were captured with light microscopy (Olympus Q-Color 5).

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; Collagen I, Abcam, ab34710, 1:50; PCNA, Abcam, ab92552; SM22 alpha, Abcam, ab10135, 1:500; Tissue factor, NOVUS, NBP2-67731, 1:100; vimentin, Abcam, ab45939, 1:200; vWF, Abcam, ab6994, 1:200; CD-68, Bio-Rad, MCA-1957, 1:200; CD4, abcam, Ab183685, 1:200; CD3, Abcam, ab11089, 1:100; CD45, Santa Cruz, sc1178, 1:100). 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 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 7-μm-thick sections were obtained. Length and thickness were measured using ImageJ (NIH Image, Bethesda, MD).[35]

En face staining

After animals were euthanized and perfused with normal saline followed by 10% formalin, the IVC from the bifurcation of iliac vein to renal vein was harvested, and its adventitia was gently removed; the IVC lumen was opened longitudinally. Samples were blocked in 1% BSA for 30 minutes and then incubated overnight at 4 °C with the primary antibodies. IVC samples were carefully placed on the slides with the luminal side facing upward and stained with DAPI.

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, 2-group comparisons were performed with the unpaired Student t test, and multiple group comparisons were performed with the 1-way ANOVA test. P values of ≤0.05 were considered to indicate statistical significance.

Results

Early thrombus formation is required of eccentric and heterogenous neointima in both mouse and human AVF

Either concentric and homogenous, or eccentric and heterogenous, neointimas (Figure 1A) are formed after different human vascular surgeries and animal models [28],[17],[16],[18]. In humans, the JAA is defined as the first 2 – 5 cm in the outflow vein distal to the AVF, but there are few studies of the human JAA area due to the relative rarity of samples [19], [36],[37],[38]. We examined the pattern of neointima at the JAA in both female and male mouse and human AVF. In the mouse aortocaval needle puncture AVF model, there was disturbed flow at the JAA area; (Figure 1B, Supplementary Video). There was eccentric and heterogenous neointima at the JAA, both at days 7 and 21 (Figure 1C). We categorized the vein wall into ‘adapted’ (without eccentric and heterogenous neointima) or ‘maladapted’ (with eccentric and heterogenous neointima) wall. The adapted wall showed moderate native wall thickening in all layers, whereas the maladapted wall showed exuberant thickening of heterogenous neointima (Figure 1D). Eccentric and heterogenous neointima was composed of a different abundance of dual PCNA- and α-actin-positive cells, and organized thrombus, both at day 7 and day 21 (Figure 1E). In the human AVF, there was also eccentric and heterogenous neointima on the maladapted wall (Figure 1F, 1G). CD31 positive cells did not fully cover either the adapted or maladapted vein wall (Figure 1H). Human eccentric and heterogenous neointima was also composed of a different abundance of dual PCNA- and α-actin-positive cells, and organized thrombus (Figure 1H, white arrow). These data show that eccentric and heterogenous neointima forms at the JAA after AVF formation in both mouse and human AVF.

Figure 1: Eccentric and heterogenous neointima in the mouse and human arteriovenous fistula (AVF).

A) Illustration pictures showing the concentric and homogenous, eccentric and heterogenous neointimas under laminar flow and disturbed flow. B) Diagram showing the AVF created by puncturing from the aorta to inferior vena cava (IVC) using a 25-gauge needle, black dashed line square showing the juxta-anastomotic area (JAA). Ultrasound measurement of the diameters and waveforms of the aorta and IVC at day 21 after AVF creation at the JAA; A, aorta; IVC, inferior vena cava. C) Mouse AVF sections stained with Verhoeff Van Gieson (EVG) showing the eccentric and heterogenous neointima at day 7 and day 21, black “*” showing the eccentric and heterogenous neointima, black arrow area showing the eccentric and heterogenous neointima on the vein wall; scale bar, 1mm, n=3. D) Illustration picture showing the adapted and maladapted vein wall without or with eccentric and heterogeneous neointima in mouse AVF. E) Low power immunofluorescence photographs showing sections stained with PCNA (green), α-actin (red) and DAPI (blue) at day 7 and 21; white “*” showing the eccentric and heterogenous neointima, white arrow area showing the eccentric and heterogenous neointima on the vein wall; A, aorta; scale bar, 1mm; n=3. F) Light microscope showing the eccentric and heterogenous neointima in human AVF, read line area showing the eccentric and heterogenous neointima, blue line area showing the media, yellow line area showing the adventitia; the maladapted vein wall was categorized into zone 1 and 2 based on the irregular shape; scale bar, 2mm. G) Illustration picture showing the adapted and maladapted vein wall without or with eccentric and heterogeneous neointima in human AVF. H) Immunofluorescence photographs showing human AVF sections stained with CD31 (green), PCNA (green), α-actin (red) and DAPI (blue); white arrow area showing the thrombus residuals in the eccentric and heterogenous neointima on the maladapted vein wall; scale bar, 400μm.

We then examined the temporal changes of anastomotic neointima in the mouse AVF model since these specimens are not available from human AVF [28]. Temporal sections showed thrombus formed immediately after AVF creation and gradually organized and transformed to heterogenous neointima in both female and male mice, without or with CKD; the thrombus formed on the damaged wall of the IVC and the aorta at the site of the complete wall disruption forming the anastomosis (Figure 2A, 2B), with α-actin positive cell accumulation and CD31-positive cell coverage (Figure 2C). Sections stained with Von Willebrand Factor (vWF) also showed similar temporal changes (Supplementary Figure 1). Anastomotic neointima also showed thrombus residue with α-actin positive cells at day 42 (Figure 2D). We further categorized the anastomotic heterogenous neointima into smooth muscle-rich (SMC-rich) neointima or SMC-poor neointima based on the abundance of α-actin positive cells (Figure 2E). There were almost all α-actin positive cells in the SMC-rich neointima while fewer than 1/3 α-actin positive cells in the SMC-poor neointima (Figure 2F). There was a similar dual PCNA-and α-actin-positive cell percentage in the SMC-rich and SMC-poor neointima at day 7, but the percentage significantly decreased at day 21 (Figure 2E, 2F). There was a smaller area of α-actin positive cells in the SMC-poor neointima compared to the SMC-rich neointima both at day 7 and 21 (Figure 2E, 2F). This data shows that the anastomotic heterogenous neointima is gradually transformed from thrombus.

Figure 2: Anastomotic thrombus transforms to heterogenous anastomotic neointima.

A) Illustration picture showing the anastomotic neointima. B) Temporal sections stained with EVG showing the anastomotic thrombus transformed to neointima in male and female mice without or with CKD, black arrow showing the transformation process from thrombus to neointima; n=3–6. C) Immunofluorescence photographs showing CD31 (red), α-actin (green) and DAPI (blue) at 6 hours, day 1, 7 and 21; yellow arrow showing the transformation process from thrombus to neointima; red arrow showing the blood flow from aorta to IVC; scale bar, 200μm; n=3. D) Anastomotic neointima stained with EVG; and CD31 (red), α-actin (green) and DAPI (blue); yellow arrow showing the thrombus residue. scale bar, 200μm; n=3. E) High power immunofluorescence photographs stained with PCNA (green), α-actin (red) and DAPI (blue) of the heterogenous anastomotic neointima at day 7 and day 21; white dashed line demarcated the SMC-rich and SMC-poor neointima; L, lumen; C, fistula channel; scale bar, 200μm; n=3. F) Bar graphs showing α-actin positive cells area (*, day 7, p<0.0001; day 21, p=0.0178; t-test), α-actin and PCNA dual positive cells (*, p= 0.5729; p=0.0153; t-test) in the SMC-rich and SMC-poor neointima; n=3.

We next examined the eccentric neointima on the vein wall at the JAA. Temporal sections showed heterogenous neointima on the maladapted wall but not on the adapted wall; there was a thicker heterogenous neointima in female AVF compared to male AVF (Figure 3A). There was a significantly smaller endothelial cell coverage on the maladapted wall compared to the adapted wall both at day 7 and 21 (Figure 3B, 3C). There were more PCNA- and α-actin-dual positive cells in the eccentric and heterogenous neointima on the maladapted wall compared to the adapted wall both at day 7 and 21 (Figure 3B, 3D), and there were decreased dual PCNA- and α-actin-positive cells in the adapted vein wall at day 21 compared to day 7 (Figure 3B, 3D). There was also a large amount of residual thrombus in the heterogenous neointima both at day 7 and 21 (Figure 3B). Tissue factor (TF)-positive cells were encapsulated by the partially covered CD31-positive cells (Figure 3E). There were also dual α-actin- and collagen-1-positive cells in the eccentric and heterogenous neointima at day 21 (Figure 3E). There were also CD45, CD8, CD68, CD3 and CD4 positive cells in the adapted and maladapted vein wall, consistent with inflammatory cells (Supplementary Figure 2). α-actin and SM22 α positive cells also colocalized with vimentin or collagen-1 in the adapted and maladapted wall (Supplementary Figure 3). This data shows that the vein wall adapts differently after AVF creation because of the eccentric and heterogenous neointima.

Figure 3: Eccentric and heterogenous neointima correlates with the maladapted vein wall.

A) Temporal sections stained with EVG showing the adapted and maladapted wall in female and male mice, without or with CKD; n=3–6. B) High power immunofluorescence photographs stained with CD31 (green), α-actin (red) and DAPI (blue); PCNA (green), α-actin (red) and DAPI (blue) of the heterogenous neointima at day 7 and 21; yellow arrows showing the thrombus; scale bar, 200μm, n=3. C) Bar graphs showing endothelial cell coverage; *, day 7, p<0.0001; day 21, p= 0.0010; t-test; n=3. D) Bar graphs showing PCNA and α-actin dual positive cells area; *, day 7, p=0.0191; day 21, p=0.0033; t-test; n=3. E) High power immunofluorescence photographs stained with CD31 (green), tissue factor (TF, red); α-actin (green), collagen-1 (red) and DAPI (blue) of the heterogenous neointima at day 21; yellow arrow showing the TF positive cells inside the heterogenous neointima; scale bar, 200μm, n=3.

Disturbed flow, endothelial cell loss, and continuous platelet deposition and accumulation are indispensable for eccentric and heterogenous neointima

We then explored the reasons for the eccentric and heterogenous formation in AVF. We observed that the location of eccentric and heterogenous neointima, or thrombus before day 7 [28], directly faced the fistula exit and anastomosis at day 3 and 7 (Figure 4A). At day 21, the eccentric and heterogenous neointima also directly faced the fistula exit but not the fistula anastomosis (Figure 4B, 4C). All of the mice with AVF showed that eccentric and heterogenous neointima was not fully covered by CD31-positive cells at day 21 (Figure 4B, 4C). Both of the AVF lumen and anastomotic neointima were whole mounted and stained at day 7 (Figure 4D); there were no CD31-positive cells in the eccentric and heterogenous neointima on the maladapted wall (Figure 4D). Although CD31-positive cells showed different morphology throughout the outflow vein and in the anastomotic neointima, the eccentric and heterogenous neointima was only located at the site where endothelial cells were not present (Figure 4D). This data shows that disturbed flow together with endothelial cell loss contributes to the eccentric and heterogenous neointima formation.

Figure 4: Eccentric and heterogenous neointima locates at the area without endothelial cell coverage under disturbed flow.

A) Sections stained with EVG showing the eccentric and heterogenous neointima (or thrombus at day 3) directly facing the fistula flow exit at day 3 and 7, black arrow showing the fistula flow exit, scale bar, 1mm, n=3. B) Longitudinal section stained with EVG, and stained with vWF (red) and DAPI (blue) showing the fistula anastomosis and fistula exit at day 21, black or white “*” showing the anastomotic neointima; scale bar, 1mm; n=3. C) Sections stained with EVG, and stained with CD31 (green) and DAPI (blue) showing the eccentric and heterogenous neointima directly facing the fistula exit at day 21; black and white arrows show the fistula flow exit, yellow arrow shows the eccentric and heterogenous neointima, red arrow shows the anastomosis, black and white “*” show the anastomotic neointima; scale bar, 1mm; n=3. D) En face stained with CD31 (red) and DAPI (blue) showing different endothelial cell morphology on the vein wall and anastomotic heterogenous neointima under disturbed flow at day 7. Dashed line oval showing the eccentric and heterogenous neointima on the vein wall, note there was no CD31 positive cell coverage on the eccentric and heterogenous neointima surface at day 7; scale bar, 1mm or 100μm n=3.

Since platelets play an important role in thrombus formation and neointima formation,[29] we also traced platelets in the eccentric and heterogenous NIH using PF4-Cre×mT/mG reporter mice. We focused on the anastomosis and fistula exit levels (Figure 5A). At day 7, GFP-positive platelets occupied most of the area of the anastomotic neointima with few SM22α-positive cells. There was no obvious heterogenous neointima on the contralateral wall facing the fistula exit; at day 21, GFP-positive platelets were encapsulated by SM22α-positive cells, and there was a large size of eccentric and heterogenous neointima on the contralateral wall facing the fistula exit, with increased area of SM22α-positive cells at day 21 compared to day 7 (Figure 5B, 5C). At day 7, at the anastomosis level, there were GFP-positive platelets on the channel surface but not on the surface of the anastomotic neointima (Figure 5D); this data is consistent with our previous observation that the fistula channel showed delayed endothelial cell coverage and related to AVF failure [28]. There was an obvious layer of GFP-positive platelets on the maladapted wall but not on the adapted wall at day 7 (Figure 5D, yellow arrows). In addition, there was increased length of the adapted wall compared to the maladapted wall (Figure 5E). At day 21, the eccentric and heterogenous neointima with GFP-positive platelets was at the level of fistula exit but not the anastomosis (Figure 5F). There was increased thickness of the heterogenous neointima at the fistula exit level compared to the anastomosis level (Figure 5G). These data show that continuous platelet deposition and accumulation play an important role in the formation of eccentric and heterogenous NIH under disturbed flow.

Figure 5: Continuous deposition of platelets and thrombus formation are critical for formation of eccentric and heterogenous NIH under disturbed flow.

A) Sections stained with EVG showing the eccentric and heterogenous neointima directly facing the fistula exit at day 14; black arrow showing the fistula flow exit; scale bar, 1mm; n=3. B) Representative images of sections stained with SM22α (red) and DAPI (blue) at the fistula exit level from PF4-mT/mG mice at day 7 and 21; white arrow shows the fistula flow exit; white dashed line circle shows the heterogenous anastomotic neointima; note the SM22α-positive cells encapsulated the green platelets at day 21 but not at day 7; Ao, aorta; scale bars: 1mm. C) Bar graph showing the SM22 α positive cell area in the anastomotic neointima at day 7 and 21; p=0.0005, t-test; n=3. D) Representative low and high power images of sections at the anastomosis and fistula exit levels from PF4-mT/mG mice at day 7; C, fistula channel; yellow arrows show a layer of green platelets deposited on the wall directly facing the fistula exit; scale bars: 400μm or 100μm; n=4 E) Bar graph showing the maladapted and adapted wall length at day 7; p<0.0001, t-test; n=3.. F) Representative low and high power images of sections stained with SM22α (red) and DAPI (blue) at the anastomosis level and fistula exit level from PF4-mT/mG mice at day 21; C, fistula channel; yellow arrows showing a layer of green platelets deposited on the wall directly facing the fistula exit; scale bars: 400μm or 100μm; n=4. G) Bar graph showing the heterogenous neointimal thickness at the level of the anastomosis and the fistula exit at day 21; p=0.0010, t-test; n=3.

To directly test whether EC loss in the AVF contributes to maladapted neointima, we used AVF cannulation to induce focal EC loss (Figure 6A). There was a large eccentric and heterogenous neointima near the cannulation site in both human and mouse AVF (Figure 6B–D). This data shows that induction of focal EC loss with cannulation also induces thrombus formation and contributes to eccentric and heterogenous neointima formation in the AVF.

Figure 6: Cannulation of the outflow vein induces eccentric and heterogenous neointima in both human and mouse AVF.

A) Illustration photos showing the cannulation of human and mouse AVF. B) Section stained with EVG showing the eccentric and heterogenous neointima with or without cannulation, irregular dashed line area showing the eccentric and heterogenous neointima; black arrows show the needle cannulation sites. C) Sections stained with EVG showing the heterogenous neointima with or without cannulation, irregular dashed line area showing the cannulation route; scale bar, 1mm or 200μm, n=3. D) Immunofluorescence photographs showing sections stained with CD31 (red), α-actin (green) and DAPI (blue) at day 21 with or without cannulation; 1, 2, 3 showing the different areas; scale bar, 1mm or 200μm; n=3.

Eccentric and heterogenous neointima is observed in sutured AVF models

Since the needle puncture mouse AVF model is similar to the clinically used endoAVF, we verified these findings using the more commonly used sutured mouse (carotid artery to jugular vein) and rat (jugular vein to carotid artery) AVF models to examine the neointima pattern at the JAA (Figure 7A). There was also eccentric and heterogenous neointima at day 7 in the mouse sutured AVF (Figure 7B). Since the mouse sutured AVF has a small amount of tissue, we examined the anastomosis and outflow vein (the JAA) using the rat sutured AVF (days 7 and 21). There were unevenly distributed α-actin-positive cells in the eccentric and heterogenous neointima in the anastomosis and outflow vein (Figure 7C). Eccentric and heterogenous neointima was also observed in female rats with the sutured AVF [31]. These data show eccentric and heterogenous neointima formation is observed in sutured AVF models under disturbed flow.

Figure 7: JAA eccentric and heterogenous neointima is universal in sutured anastomosis AVF.

A) Illustration and surgical photos showing the male mouse carotid artery-jugular vein (CA-JV) fistula (end to side) model, ruler marks 1mm. B) Illustration photo and sections stained with EVG showing the JAA at day 7, note the eccentric and heterogenous neointima in the high-power photographs; scale bar, 1mm or 100μm, n=3. C) Immunohistochemistry photographs of the anastomosis and outflow vein (in the JAA) in a male rat JV-CA fistula (end to side) model at day 7, 21 and 42, low and high-power photographs showing the anastomosis and outflow vein stained with α-actin; note the eccentric and heterogenous neointima in the outflow vein; scale bar, 400μm or 100μm, n=3.

Neither disturbed flow with limited endothelial loss nor non-disturbed flow induces heterogenous neointima

Since there were both disturbed flow and endothelial cell loss at the needle entry site in the inflow artery in this mouse AVF model [28], we examined temporal changes at the aortic entry site (0.515 mm in diameter; Figure 8). Thrombus formed immediately and became smaller after 6 hours (Figure 8A, 8B). At the needle entry site, there was an increased luminal area after day 7 and thin elastin fibers at day 42 (Figure 8A), consistent with this site being an arteriovenous fistula inflow artery aneurysm (AVFIA) (Figure 8A, 8D, 8E). Female mice also showed homogenous neointima in the AVFIA (Supplementary Figure 4). Spatial changes of the AVFIA at day 21 showed similar homogenous AVFIA neointima (Figure 8F). The aorta wall separates the original aortic lumen and AVFIA lumen because of the larger size of the AVFIA (Figure 8G, 8H). mGFP-positive platelets were found only at the aortic entry site at day 7 (Figure 8I). Cross sections also showed homogenously distributed α-actin-positive cells in the AVFIA neointima (Figure 8J), and some SM22α-positive cells expressed collagen-1 (Figure 8J). Longitudinal sections showed that the AVFIA neointima is homogenous, covered by CD31-positive cells and α-actin-positive cells (Figure 8K, 8L). The AVFIA wall achieved complete re-endothelization at day 21 similar to native aorta (Supplementary Figure 5). We also examined the neointima pattern in other mouse models without disturbed flow. There was only a small neointima in the mouse aortic direct suture model, and a homogenous neointima in the aortic clamp injury model and the aortic patch angioplasty model (Supplementary Figure 6). These data show that neither disturbed flow with limited endothelial cell loss nor non-disturbed flow induces heterogenous neointima.

Figure 8: Disturbed flow with limited endothelial cell loss induces homogenous neointima in the AVF inflow artery.

A) Temporal low and high-power photographs stained with EVG showing the gradual changes from thrombus to AVFIA wall from day 0 to day 42 in male mouse; blue arrow showing the thrombus, yellow arrow showing the AVFIA wall, black arrowhead showing the new formed elastin fibers at day 42; IVC, inferior vena cava; scale bar, 400 μm and 200 μm; n=6–10. B) Bar graph showing the thrombus area at day 0, 10 seconds, 6 and 24 hours; p=0.0002, one-way ANOVA; n=3. C) Bar graph showing the luminal area at day 0, 7, 21 and 42; p<0.0001, one-way ANOVA; n=3. D) Illustration photos showing the formation of homogenous neointima of the AVFIA. E) In situ photo showing the AVFIA (yellow dashed line square) at day 21, ruler marks 1mm. F) Series cross-sectional photographs stained with EVG showing homogenous neointima of the AVFIA from cranial to caudal at day 21, black dashed line rectangle showing the AVFIA; black arrow showing the aortic wall, yellow arrow showing the homogenous neointima of the AVFIA; IVC, inferior vena cava; scale bar, 400 μm or 200 μm; n=3. G) Infrequent sections showing AVFIA at day 21, black dashed line square showing the AVFIA, black arrow showing the AVFIA lumen, IVC, inferior vena cava; scale bar, 400 μm or 200 μm. H) Series sections of merged immunofluorescence stained for CD31 (green) and DAPI (blue) of the AVFIA at day 21; from caudal to cephalic (1–4), note the coverage of CD31 positive cells on the AVFIA neointima, white arrow showing the process of diminishing of the native aortic wall dividing the original aorta and AVFIA; scale bar, 200 μm; n=3. I) Representative images of AVFIA from PF4-mT/mG mice at day 7, scale bar, 100 μm; n=5. J) Merged immunofluorescence stained for CD31 (green), α-actin (red), and DAPI (blue); SM22α (green), collagen-1 (red) and DAPI (blue) of the AVFIA at day 21; white dashed line square showing the AVFIA; note the homogenous AVFIA neointima; scale bar, 200 μm; n=3. K) Longitudinal section stained with EVG showing the AVFIA at day 21, black arrow showing the aortic wall, yellow arrow showing the AVFIA; IVC, inferior vena cava; scale bar, 1mm and 400 μm; n=3. L) Longitudinal sections stained for CD31 (green), α-actin (red), and DAPI (blue) of the AVFIA at day 21; white dashed line square showing the AVFIA, white arrow showing the aortic wall, yellow arrow showing the AVFIA; IVC, inferior vena cava; scale bar, 1mm or 200 μm; n=3.

Discussion

We present studies that support early thrombus formation is required for heterogenous and eccentric neointimal hyperplasia under disturbed flow. Neither disturbed flow with limited endothelial loss nor non-disturbed flow induces heterogenous neointima. This data suggests targeted new directions for precision therapy, e.g. early antithrombotic and late anti-SMC proliferation to prevent NIH after vascular surgery.

Coagulation, thrombus formation and NIH occurs after vascular surgery [39],[5]. Thrombus formation and organization contribute to NIH in animal studies [40] [24], but there are different mechanisms of thrombus organization and (neo)intimal hyperplasia[41],[42]. Luminal thrombus organization occurs on the luminal side of the endothelium and involves circulating fibrocytes that penetrate the thrombus before becoming myofibroblasts to support the structure. The mechanisms of thrombus resolution mechanism are complex, and include platelet-driven contraction of the thrombus, followed by a redistribution of different components and compressive deformation of red blood cells; contraction also affects obstructiveness, permeability, and fibrinolysis of the thrombus. [43] Thrombosis occurs in human AVF; 5%–20% of fistulae fail due to early thrombosis, with early thrombosis occurring before 30 days.[44] When AVF thrombose, thrombectomy is an established surgical technique to remove the thrombus with clinical success in nearly two-thirds of cases.[45] Venotomy and manual removal of chronic and organized thrombi from occluded native AVF is also safe and effective.[46] These clinical observations confirm the importance of avoiding thrombosis in human AVF to maintain access patency. (Neo)intimal hyperplasia occurs secondary to the accumulation, migration, and proliferation of wall myofibrocytes in response to injury. Platelet-derived growth factor and transforming growth factor β released from platelets in the thrombus may play an important role in the formation of heterogenous NIH.[47],[48] Our data suggests that thrombus formation may be the critical component that contributes to heterogenous NIH. Interestingly, focus on the eccentric and heterogenous neointima at the level of fistula exit might be more scientifically rigorous and clinically relevant compared to the neointima in the outflow vein, with both the length and the thickness of the heterogenous neointima being important for patency in this model. Further categorization of heterogenous NIH into SMC-rich and SMC-poor NIH may contribute to a more effective anti-NIH therapy, and thus temporal release of targeted therapeutic drugs - such as early antithrombotic and late anti-SMC - may be more effective to inhibit NIH [49].

Although an effective therapy to inhibit NIH has not emerged, anti-proliferation therapy is commonly used but the overall clinical outcome remains inadequate [50], [51],[52],[53]. Anticoagulation using heparin and antiplatelet therapy using aspirin can inhibit NIH both in basic and clinical studies [54],[2],[55], [5], [56]. We showed that there is a large amount of early thrombus replaced by anastomotic heterogenous neointima at the JAA in this AVF model, supporting a role for disturbed flow in formation of an early thrombus and late heterogenous NIH; further examination of how both anticoagulation and antiplatelet therapy affect NIH in this model is needed. The area of EC loss is also critical to the formation of heterogenous NIH in addition to the contribution of disturbed flow, as we showed that exposure of the inflow entry site with only limited EC loss (0.515 mm in diameter) to disturbed flow yielded a homogenous SMC distribution (Figure 7). On the contrary, we previously showed heterogenous NIH in a large pseudoaneurysm that has a large area without endothelial coverage and in the presence of disturbed flow [57]. These data show that both disturbed flow and a large area of EC loss are critical for formation of heterogenous NIH. The location of EC loss appears to determine the location of the eccentric neointima.

There are several limitations in this research. First, although our data suggests that organized thrombus is part of the source of luminal thickening during venous remodeling in the mouse AVF model, this may not be the cause of human AVF failure, since there are no temporal human AVF samples available. Similarly, although we used shear stress, vein diameter, and wall thickness in the outflow vein to determine the maturation in this mouse model, maturation is not completely the same between mouse and human AVF [58],[59]. Second, although juxta-anastomotic NIH is one of the common reasons that lead AVF failure in humans, [36] decreased NIH does not necessarily correlate with increased maturation in this needle puncture mouse AVF model; similarly, preexisting NIH in human veins used for AVF creation may modestly associate with maturation failure.[60], [61] Third, like other rodent AVF models, there is a large difference in hemodynamics between mice and humans, both before and after AVF creation; differences in multiple hemodynamic parameters influence the development and progression of NIH.[62] Last, current theories of the mechanisms of venous remodeling rely on the mobilization and transdifferentiation of smooth muscle cells (SMC) to the intima, but our data suggest that the SMC might invade into the thrombus, which will require validation.

In conclusion, we show that early thrombus contributes to late heterogenous NIH formation under disturbed flow. Disturbed flow, large area of endothelial cell loss, continuous platelets deposition and accumulation are indispensable for eccentric and heterogenous neointima. Categorization of adapted or maladapted walls may be helpful for therapy targeting heterogenous NIH.