**3. Pathology and pathophysiologic mechanisms of hemodialysis vascular access dysfunction**

### **3.1 Pathology of Hemodialysis Vascular Access Stenosis in AVF and AVG**

Venous stenosis that occurs in both AVFs and AVGs is primarily due to neointimal hyperplasia. Venous stenosis in AVGs most frequently arises from the development of aggressive neointimal hyperplasia, characterized by (a) the presence of alpha smooth muscle actin positive cells myofibroblasts, and microvessels within the neointima, (b) an abundance of extracellular matrix components, (c) angiogenesis (neovascularization) within the neointima and adventitia, (d) a macrophage layer lining the perigraft region, and (e) an increased expression of mediators and inflammatory cytokines such as TGF-β, PDGF, and endothelin within the media, neointima and adventitia 59-64.

While the neointimal hyperplasia in AVFs is similar to AVGs in regards to pathogenesis, the venous stenosis that develops in AVFs is highly influenced by the capacity of the vein to vasodilitate and vascular injury from surgical technique 65. In AVFs the two main etiologies of failure are an initial failure to mature (nonmaturation) and a subsequent (late) venous stenosis 4. Similar to AVGs, venous neointimal hyperplasia in late AVF stenosis has been shown to be composed primarily of alpha smooth muscle actin positive cells, together with expression of mediators and cytokines such as TGF-β, PDGF, and endothelin within the media and intima of the vein 60,65. However, recently, the lesion of AVF nonmaturation at 6 weeks after AVF creation has also been described to have significant neointimal hyperplasia 66.

### **3.2 Pathophysiologic mechanisms of neointimal hyperplasia formation in hemodialysis access dysfunction**

The pathogenesis of venous neointimal hyperplasia in AVG stenosis and late AVF stenosis has been well described and is commonly divided into upstream and downstream events4. Upstream events are characterized as the initial events and insults that are responsible for endothelial and smooth muscle cell injury, which leads to a cascade of mediators (downstream events) that regulate oxidative stress, endothelial dysfunction, and inflammation (eventually resulting in venous neointimal hyperplasia). Upstream events that are believed to contribute to the pathogenesis of neointimal hyperplasia include 4,62,67-70: (1) surgical trauma at the time of AV surgery, (2) hemodynamic shear stress at the veinartery or vein-graft anastomosis, (3) bioincompatability of the AVG, (4) vessel injury due to dialysis needle punctures, (5) uremia resulting in endothelial dysfunction, and (6) repeated angioplasties causing further endothelial injury. Downstream events represent the response to endothelial (vascular) injury from the upstream events, resulting in the migration of smooth muscle cells from the media to the intima and eventually the development of neointimal hyperplasia 65.

The pathogenesis in AVFs that fail to mature (early failure) for dialysis, in contrast to AVG and late AVF failure, remains poorly understood. At a histological level early AVF failure is also characterized by aggressive neointimal hyperplasia in both animal and human models, seen as early as 1 month in animals 63,71 and 3 months in humans 64,66. The underlying factors

recent work in which uremic mice developed a 2-3 fold greater magnitude of neointimal hyperplasia at the arteriovenous anastomosis as compared to non-uremic animals in a mouse model of AVF stenosis 92, and a recent study which showed marked upregulation of monocyte chemoattractant protein-1 (MCP-1) in the venous segment of AVF compared to

In clinical studies, possible linkages have described the presence of inflammatory cells (macrophages and lymphocytes), cytokines such as TGF-ß and insulin-like growth factor-1 (IGF-1) and the magnitude of neointimal hyperplasia and venous stenosis within stenotic

Local bioincompatability to synthetic polytetrafluoroethylene (PTFE) material in AVGs could also result in local inflammation 95. In vitro studies have demonstrated that conditioned media obtained after the interaction of peripheral blood mononuclear cells (PBMCs) with PTFE graft material resulted in a significant upregulation of smooth muscle cell proliferation as compared to control media 96. This proliferative response has been shown to be attenuated by tumor necrosis-alpha (TNF-) inhibitors 96. Furthermore, the presence of macrophages that line PTFE graft material has been described in both experimental and clinical AVG stenosis with co-expression of inflammatory cytokines such

An intact and functional endothelium is essential for the vein to properly respond to acute changes in blood flow that occurs after creation of AVFs and AVGs 98. Nitric oxide (NO) is an important mediator responsible for these transformations 99,100. The presence of uremia in hemodialysis patients has been shown to exacerbate endothelial dysfunction, possibly through the pathways of inflammation and oxidative stress described above 101,102. In the specific context of vascular access stenosis, endothelial dysfunction is likely to be responsible for the development of pre-existing venous neointimal hyperplasia 77-81, medial hypertrophy 77,81 and radial artery intima-media thickening 103-105 that is present even before the creation of AVFs in uremic patients. Pre-existing arterial intima-media thickness has been correlated with future AVF dysfunction 103. Recently, pre-existing venous neointimal

Asymmetrical dimethylarginine (ADMA) is an endogenous inhibitor of NO synthase and has been implicated as an important contributor to endothelial dysfunction in ESRD patients 106. ADMA is not excreted in ESRD patients and its levels have been reported to be two to six times higher in this patient population as compared to non-uremic individuals 107. In a recent clinical study in AVFs, patients with elevated ADMA levels at the time of percutaneous transluminal angioplasty of an initial AVF stenosis had a significantly

Although the traditional paradigm for the pathogenesis of neointimal hyperplasia has emphasized the migration of smooth muscle cells from the media to the intima, where they proliferate and contribute to the final volume of neointimal hyperplasia, a number of studies have reported that following coronary angioplasty or saphenous vein bypass grafting there is also a migration of cells (fibroblasts) from the adventitia, through the media, and into the intima, where these cells transform into "myofibroblasts" 109-111. In dialysis access, a number

hyperplasia has been linked to poor AVF maturation in a small clinical study 77.

rats deficient in the MCP-1 gene 93.

as basic fibroblast growth factor (bFGF) 61,97.

increased risk of a recurrent AVF stenosis 108.

**3.2.4 Alternative origins of neointimal cells** 

**3.2.3 Endothelial dysfunction** 

AVFs 94.

(upstream events) which may contribute to early AVF failure, include 4,72-81: (1) small diameter sizes in the vein and artery, (2) surgical injury at the time AV fistula placement, (3) previous venipunctures, (4) development of accessory veins after surgery, (5) hemodynamic shear stress at the AV anastomosis, (6) a genetic predisposition to vascular constriction and neointimal hyperplasia, and (7) pre-existing venous neointimal hyperplasia.

The subsequent sections will focus on the downstream events and three main mechanisms responsible for neointimal hyperplasia such as oxidative stress, inflammation, endothelial dysfunction, and alternative origins of neointimal-derived cells.

### **3.2.1 Oxidative stress**

Many of the upstream mechanisms above (particularly hemodynamic shear stress and angioplasty injury) have been documented to result in an increase in the production of free radicals and its downstream products nitrotyrosine and latter (peroxynitrate). The latter is a potent upregulator of the matrix metalloproteinases (MMPs) 82,83. MMPs are key enzymes that cause breakdown of extracellular matrix proteins such as collagen and elastin which facilitate the migration of vascular smooth muscle cells (VSMCs) in neointimal hyperplasia formation 84. MMPs, paradoxically, have also been shown to facilitate a beneficial dilatation of the feeding artery (through degradation of the internal elastic laminae) in both rabbit and mouse AVF models 82,85. Experimental studies of AVGs have demonstrated a differential upregulation of MMP-2 at the graft-vein anastomosis, with early expression (9 days) in the adventitia and a later expression (19 days) within the intima, supporting the concept of cellular migration from the adventitia to the intima 86. Furthermore, linkages between hemodynamic shear stress and the expression of oxidative stress markers and cytokines have also been described in a porcine model of AVG stenosis 87. Clinical studies of stenotic and thrombotic AVGs and AVFs have also demonstrated an upregulation of MMPs 88, and have documented the co-localization of oxidative stress markers with inflammatory cytokines such as transforming growth factor-beta (TGF-), and platelet-derived growth factor (PDGF), within the neointima of both stenotic AVGs and AVFs 60.

Heme-oxygenase-1 (HO-1) is an important enzyme pathway which has been shown to confer protective effects in the vascular endothelium and other organ systems through its anti-inflammatory, antioxidant, or antiproliferative actions and properties 89. Experimental studies in AVFs have described an increase in both the magnitude of arteriovenous stenosis and the frequency of thrombosis following the creation of AVFs in HO-1 knock out mice (increased baseline oxidative stress) as compared to wild type animals 90. Furthermore, in the HO-1 knockout mice, there was significant induction of MMP-9 expression in the vein at 1 week compared to wild type mice, suggesting that MMP expression in vascular tissue and its deleterious effects with regard to promoting cellular migration may in part be inhibited by HO-1. Clinical studies have demonstrated a higher frequency of AVF failure in patients with heme-oxygenase-1 (HO-1) gene polymorphisms with long GT repeats (resulting in increased oxidative stress) 73.
