Role of platelet factor 4 in arteriovenous fistula maturation failure: What do we know so far?

Abstract

The rate of arteriovenous fistula (AVF) maturation failure remains unacceptably high despite continuous efforts on technique improvement and careful pre-surgery planning. In fact, half of all newly created AVFs are unable to be used for hemodialysis (HD) without a salvage procedure. While vascular stenosis in the venous limb of the access is the culprit, the underlying factors leading to vascular narrowing and AVF maturation failure are yet to be determined. We have recently demonstrated that AVF non-maturation is associated with post-operative medial fibrosis and fibrotic stenosis, and post-operative intimal hyperplasia (IH) exacerbates the situation. Multiple pathological processes and signaling pathways are underlying the stenotic remodeling of the AVF. Our group has recently indicated that a pro-inflammatory cytokine platelet factor 4 (PF4/CXCL4) is upregulated in veins that fail to mature after AVF creation. Platelet factor 4 is a fibrosis marker and can be detected in vascular stenosis tissue, suggesting that it may contribute to AVF maturation failure through stimulation of fibrosis and development of fibrotic stenosis. Here, we present an overview of the how PF4-mediated fibrosis determines AVF maturation failure.

Introduction

Since Brescia and his colleagues¹ created the first arteriovenous fistula (AVF) in New York in 1965, this type of vascular access has been used to deliver hemodialysis (HD) in patients suffering from various types of kidney failure, including the most severe one: end stage renal disease (ESRD).² In fact, AVF is considered the best amongst all three vascular access methods to deliver HD.^3–5 Compared to the other two methods, which are arteriovenous grafts (AVGs) and central vein catheters (CVCs), AVFs have lower complication rates, thrombosis and infection rates, fewer access related hospitalization, better long-term patency, and lower health care related cost burden.^6–11 Patients with AVFs also have shown substantially lower mortality rates and increased life expectancy.⁶ Therefore, the AVF has become the preferred option in ESRD patients^12,13 and recommended by the Kidney Disease Outcome Quality Initiative (KDOQI) vascular access guidelines.¹⁴ However, the frequency of AVF primary failure has remained high for the last five decades.^15,16 Studies have shown that up to 23%–50% of newly created AVFs fail to mature and cannot be used for HD without undergoing a salvage procedure,^8–21 and a 60% failure to mature rate amongst total AVFs has been reported in the US.^22–24 This causes a significant negative impact on morbidity, mortality, and life quality of patients who require HD for survival, as well as increased medical costs.⁵

The role of outward remodeling in AVF non-maturation has been recently demonstrated,²⁵ Fibrosis has been previously detected in stenotic AVFs,²⁶ and our recent study demonstrated a close association between AVF maturation failure and post-operative fibrotic stenosis.²⁷ At the molecular level, the pathology behind AVF maturation failure has not been fully illustrated, but several proteins are thought to be critical during this process, such as heme oxygenase−1 and −2 (HO-1 and HO-2), monocyte chemoattractant protein-1 (MCP-1), Kruppel-like factor 2 (KLF2), transforming growth factor β1 (TGFβ1), and inflammatory markers.^28–30 The last one, such as matrix metalloproteinases (MMPs) and tissue inhibitor of metalloproteinases (TIMPs) have also been considered as potential predictors of AVF maturation.^28–30 Platelet factor 4 (PF4/CXCL4) is a pro-inflammatory cytokine that has previously been associated with fibrosis,^31–33 and we recently found it is upregulated in veins that fail to mature after AVF creation.³⁴ These discoveries suggested that PF4 plays in AVF maturation failure. Therefore, in this review, we summarize the recent findings that link fibrosis with AVF maturation failure, as well as the involvement of PF4 in fibrosis and explore the potential role of PF4 in AVF maturation failure.

Fibrosis and AVF maturation failure

Fibrosis

Fibrosis is an outcome of the dysregulated tissue repair process in response to various types of tissue injury.³⁵ It is characterized by excessive accumulation/deposition of extracellular matrix (ECM) components, such as collagens and fibronectins.^35–38 During tissue injuries endothelial cells are damaged, and the basement membrane is thus exposed, which causes a clotting process and platelets are recruited. This results in releasing of factors such as platelet-derived growth factors (PDGFs) and transforming growth factor β1 (TGFβ1) that lead to smooth muscle cell (SMC) and fibroblasts activation. Activated fibroblasts have increased contractility and able to secrete inflammatory mediators and synthesize ECM components.^35,39 Fibrosis can affect normal tissue and organ functions in a pathological state,^40,41 and may lead to kidney, cardiac, liver failure, vessel dysfunction, or even life threatening conditions, such as idiopathic pulmonary fibrosis in extreme cases.^42–45 In the vasculature, fibrosis is featured by reduced liminal diameter and arterial wall thickening caused by excessive deposition of ECM.^46,47 Vascular fibrosis involves proliferation of vascular smooth muscle cells (VSMCs), accumulation of ECM particularly collagen and fibronectin in the vascular wall, and inhibition of matrix degradation.^46,48,49

Fibrosis can result from many causes, inflammation is the most notably one in general.^35,50 Fibrosis can be triggered by both acute and chronic inflammation,³⁷ with inflammatory monocytes and tissue-resident macrophages acting as key regulators of tissue fibrosis.^51–53 Inflammation leads to injury of epithelial and often endothelial cells, which in turn enhances the release of inflammatory mediators and the recruitment of inflammatory cells, which activates effector cells and drive the fibrogenic process.⁵⁴ Fibroblasts, SMC, and myofibroblasts are all effector cells and key fibrosis mediators in many organs. They are responsible for the synthesis of ECM proteins,^37,55,56 mostly collagen I, III and fibronectin,^37,57 in all types of fibrosis.^58–60 Fibroblasts can also produce collagen IV in some instances.^61,62 Interestingly, SMC proliferation can be stimulated by monomeric/soluble collagen I, while soluble collagen IV stimulates SMC migration in vitro.⁶³ This suggests that there may be a feed forward signaling loop between SMC and collagens that contributes to fibrosis progression. Profibrotic macrophages interact with fibroblasts,⁶⁴ which also contributes to ECM deposition during fibrosis.^65–70 Fibroblasts can undergo phenotype switching into myofibroblasts during fibrosis,^71–74 and myofibroblasts that are contractile express α-smooth muscle actin (αSMA) as a biomarker.⁷⁵ Apart from fibroblasts and myofibroblasts, collagens can also be produced by hematopoietic, epithelial, and endothelial cells.⁷⁶

Hypertension is another important cause of fibrosis, particularly in cardiovascular fibrosis. Prohypertensive factors, including activation of the renin-angiotensinaldosterone system (RAAS), inflammation, oxidative stress, salt intake, etc. have all been shown to result in excessive arterial fibrosis and ECM deposition.⁷⁷ Vascular changes associated with hypertension are featured by decrease of nitric oxide (NO) level and increase of reactive oxygen species (ROS) production, proinflammatory and profibrotic pathway activation, suppression of ECM degradation including decrease of collagen turnover, and ECM remodeling especially the increase of fibrillar type I collagen.^77–79 All these effects combine to increase the possibility of fibrosis occurrence. In fact, ECM accumulation that resulted from increase of collagen synthesis and reduction of collagen degradation is the major reason of fibrosis development in response to chronic hypertension,⁸⁰ especially when considering that collagens I and III are the predominant components in vascular ECM.⁸¹ Apart from collagen, elastin also plays a role in vascular fibrosis. Deficiency of elastin may exacerbate vascular fibrosis and increase the stiffness of vessels that is caused by collagen excess.^81–83 Furthermore, imbalanced elastin/collagen ratio has been shown to lead to the deposition of excessive ECM proteins, particularly collagen and fibronectin.⁸¹

Other cardiovascular disease-associated risk factors, such as hyperglycemia, dyslipidemia (hypercholesterolemia), and hyperhomocysteinemia (HHcy) are also associated with vascular fibrosis initiation and progression.⁴⁶ Increase of collagens, fibronectin and altered matrix degradation have been shown involved in interstitial fibrogenesis in hypercholesterolemic rats.^84,85 Circulating oxidized low-density lipoproteins (ox-LDL) can also promote fibrosis through stimulating the synthesis and expression of TGFβ in hypercholesterolemia patients.^86–88 Several of the above features of vascular fibrosis, such as excessive ECM deposition and imbalanced elastin/collagen proportions have all been shown in HHcy.⁸⁹ The increase of ECM deposition in HHcy may be due to alteration of ECM metabolism that are caused by activation of MMPs,⁹⁰ while the elastin/collagen imbalance may be due to increase intracellular Ca²+ signaling.⁹¹

Several physiological and pathological medical conditions, such as aging, diabetes, and obesity are known risk factors that would contribute to vascular fibrosis. Aging is one of the main causes of hypertension,⁷⁷ while diabetes would cause expansion of ECM with the occurrence of fibrosis in many tissues including the vasculature.⁹² High glucose has been shown to increase angiotensin II (Ang II) (will be discussed below) generation which will result in promoting fibrosis via stimulation of TGFβ secretion.⁹³ Vascular remodeling and stiffness, on the other hand, are shown to be associated with obesity⁹⁴ and insulin resistance.⁹⁵

Besides the vasculature, fibrosis also occurs in various anatomic locations due to different diseases and/or pathologies. Progressive forms of chronic kidney disease (CKD) are associated with excessive deposition of ECM, leading to ESRD,^96–98 and hypertension and diabetes are the two leading causes of renal fibrosis.³⁷ A lot of diseases can lead to pulmonary fibrosis, among which idiopathic pulmonary fibrosis is a progressive one without substantial inflammation.⁹⁹ Finally, some genetic mutations have been found to be associated with fibrosis. For example, mutations in telomerase reverse transcriptase (TERT) and mucin 5B (MUC5B) are associated with pulmonary fibrosis,^100–103 while mutations in the myh7 gene is associated with cardiac fibrosis,¹⁰⁴ and mutation in dystrophin (DMD) is associated with Duchenne muscular dystrophy-associated skeletal muscle fibrosis.¹⁰⁵

Molecules and molecular pathways associated with fibrosis

At the molecular level, fibrosis is the result of a complex crosstalk between various signaling pathways that are involved in injury response, such as TGFβ, Wingless and Int-1 (Wnt), Notch, Yes-associated protein/transcription co-activator with PDZ-binding motif (YAP/TAZ), AMP-activated protein kinase (AMPK), Ras, etc.,^106–109 and TGFβ pathway is considered to be the primary factor driving fibrosis.¹⁰⁸ TGFβ is expressed in many cell types, including endothelial cells, VSMCs, myofibroblasts, and adventitial macrophages,^77,110 while fibroblasts are the major targets of this factor,¹¹¹ although TGFβ also stimulates a fibrogenic phenotype in epithelial cells.¹¹² Intriguingly, some other cell types, such as endothelial cells and mural cells may convert to fibroblasts in response to TGFβ.¹¹¹ TGFβ reduces collagenase production while increases TIMPs and results in excessive collagen/ECM accumulation.^46,113,114

In the canonical TGFβ pathway, binding of ligand to TGFβ receptor II (TGFβRII) recruits and phosphorylates TGFβ receptor Is (TGFβRIs). These phosphorylated/activated TGFβRIs will then phosphorylate the effector protein Smads, which then migrate to the nucleus to function as transcription co-activators of many fibrosis associated genes.^37,115 The signaling of TGFβ can be enhanced by binding of TGFβ receptor III (TGFβRIII) or endoglin.¹¹⁶ TGFβ is expressed and stored in the ECM as a large latent complex (LLC), a molecule structure that comprises of an active TGFβ, a latency-associated protein (LAP), and a latent TGFβ binding protein (LTBP).¹¹⁷ Active TGFβ can be obtained through many mechanisms, including MMPs and other proteases,^{115,118–125} integrins such as αvβ6,^126–132 matrix proteins including the ED-A fibronectin isoform, fibulin-1c, and thrombospondin-1,^133–136 and physical and chemical conditions such as low pH, heat, radiation, shear stress, and ROS.¹³⁷ It is worth mentioning that besides activating TGFβ, some MMPs themselves are involved in fibrosis. For example, it has been shown that MMPs are necessary for SMC and fibroblasts migration that leads to fibrosis.^56,138 MMPs also degrades ECM proteins that result in ECM modification which leads to cardiovascular remodeling and fibrosis,^77,139,140 and MMP9 activation potentiates fibrosis.¹⁴¹

Expression of TGFβ has been found to correlate with progression of liver, lung, kidney, skin, and cardiac fibrosis, while reduction of fibrosis development upon inhibition of TGFβ pathway has been reported in experimental models.^89,142 TGFβ affects inflammation,¹⁴² enhances infiltration and/or proliferation of fibroblasts, induces proliferation and activation of intrinsic fibroblasts into collagen producing myofibroblasts, increases ECM synthesis, and inhibits degradation of collagen.^143–145 Studies have shown that activation of vascular TGFβ1 increases the synthesis of ECM proteins, such as fibronectin, collagens, and plasminogen activator inhibitor-1 (PAI-1),^146–151 and active TGFβ1 also promotes fibroblast differentiation into myofibroblasts,¹⁵² featured by initiation of αSMA expression in fibroblasts.^147,153 Upregulation of PAI-1, a downstream effector of TGFβ, would result in ECM accumulation via prevention of collagen degradation.⁷⁷ Concurrently, connective tissue growth factor (CTGF), another downstream effector of TGFβ, is implicated in fibroblast proliferation and a key mediator of ECM production during fibrosis.^133,154

As mentioned above, the TGFβ pathway induces fibrosis as a participant of a larger complex crosstalk, and TGFβR can mediate signals via mitogen-activated protein kinase-extracellular signal-regulated kinase (MAPKERK) and ERK pathway activation, which facilitates fibroblast survival and promotes fibrosis.¹⁵⁵ Meanwhile, TGFβR also signals through c-Jun N-terminal kinase (JNK), p38-MAPK, nuclear factor κ B (NF-κB), phosphoinositide 3-kinase-protein kinase B-mammalian target of rapamycin (PI3K-Akt-mTOR), Rho-associated coiled-coil forming protein kinase (ROCK),^{108,156–163} and cooperates with Wnt & Notch signaling cascade.¹⁶⁴ In addition, TGFβ1 is able to induce several miRNA species that have fibrotic effects.¹⁰⁸ In fact, a series of miRNAs, such as miR-192, miR-19b, miR-21, miR-29, miR-30, and miR-133 are all involved in fibrosis.^165–168

Interestingly, besides being the end product of fibrosis, collagens are associated with fibrosis onset. ECM fragments are important drivers of fibrosis by stimulating chemokine and proinflammatory cytokine production in inflammatory monocytes and macrophages, including the aforementioned activation of TGFβ.^134,136,169 The collagen fragment ac-PGP produced by MMPs indirectly induces inflammation by attracting leukocytes through chemotaxis.^170,171 This molecule also drives inflammation and ECM remodeling by enhancing MMP9 as well as neutrophil elastase secretion and activation in a pulmonary inflammation model.¹⁷²

Vascular fibrosis, different from fibrosis in other tissues, can be further promoted by vasoactive agents, such as Ang II, aldosterone, and endothelin-1 (ET-1). Ang II, other than being a blood pressure regulator, is also responsible for vascular cell growth/apoptosis, VSMC migration, inflammation responses, and ECM remodeling.¹⁷³ It promotes inflammation⁴⁶ and ECM synthesis which leads to vascular fibrosis through TGFβ, CTGF, interleukin-6 (IL6), tumor necrosis factor α (TNFα), and MCP-1.^174,175 Overexpression of miR-122 exacerbates Ang II-mediated ECM deposition.^176,177 TGFβ1 activation during aging and hypertension is mediated by Ang II^178,179 and ET-1,¹⁸⁰ as well as mechanical stress^149,181 and ROS.¹⁸² Ang II acts through its receptor angiotensin II receptor type 1 (AT-1R) and mineralocorticoid receptor (MR) to increase TGFβ signaling and promote ECM synthesis.^{140,183–186} Antagonizing or knocking out of AT-1R decrease fibrosis, reduce intima-medial thickness, and attenuate intima hyperplasia.^187–190 Similarly, conversion of Ang II into Ang-(1–7) by the angiotensin converting enzyme (ACE) homolog ACE2 has been shown to prevent cellular proliferation and vascular fibrosis.¹⁹¹ Loss of ACE2 will lead to increase of Ang II-mediated expression of TGFβ, CTGF, and collagens,^192,193 while overexpression of ACE2 prevents the Ang II-induced increase of proinflammatory reaction.^194–196 Inhibition of ACE with H₂S can control the vascular fibrosis progression characteristic of chronic hypertension.^197–199 In contrast, decrease of H₂S in HHcy ameliorates its ability to inhibit ACE, which results in increase of Ang II-dependent profibrotic stimulation.²⁰⁰ Ang II can also activate Smad in an TGFβ-independent manner,^180,201 and it can induce activation of p38 MAPK which is associated with the development and progression of fibrosis.^202–204 Other than that, Ang II has been reported to activate CTGF⁷³ and MMP2 and 9.⁷⁷

Aldosterone has been shown to promote collagen I synthesis in rat.^205,206 It also stimulates PAI-1 expression and induces CTGF expression via p38 MAPK cascade and MR.^207,208 Inhibiting MR with galectin-3 attenuates hypertension, inflammation, and vascular fibrosis in an aldosterone-salt challenge rat model.²⁰⁹ ET-1, in turn, can stimulate fibroblast-induced collagen synthesis.^210,211

In patients suffering from chronic hypertension, the production of advanced glycation end products (AGEs) has been accelerated.^92,212 This molecule stimulates the release of TGFβ.^213–215 It is also notoriously known to halt matrix degradation by both irreversibly crosslinking with ECM macromolecules including collagen backbones²¹⁶ and reducing the activity of MMPs.^217,218

Apart from AGEs, tissue transglutaminase (TG2) is also a crosslinker, which promotes vascular stiffness by increasing cell surface-ECM crosslink.²¹⁹ ECM proteins such as fibronectin, collagen, and laminin have all been reported as substrates of TG2.⁷⁷ Under fibrotic condition, TG2/ECM protein crosslinks are increased,⁷⁷ while inhibition of TG2 has been shown to attenuate fibrosis in a mouse model.²²⁰ Interestingly, TG2 itself also activates TGFβ/Smad signaling and promotes collagen synthesis and fibrosis.²²¹

In lymphatic collecting vessels, the C-C chemokine receptor 7 (CCR7) has been implicated in preventing fibrosis. Increased collagen deposition has been shown in the adventitia of lymphatic collecting vessels of CCR7^−/− mice and fibroblasts can be observed within the excess adventitial collagen. This fibrosis is reversed by re-expression of CCR7 in dendritic cells (DCs).²²²

There are also many other signal pathways and molecules important to fibrosis. The Wnt pathway, besides cross-talking with TGFβ, is also important to fibrosis by itself. Activation of canonical Wnt pathway has been reported in kidney, liver, skin, and cardiac fibrosis,^223,224 while bleomycin-induced pulmonary fibrosis has been shown to be highly dependent on the Wnt-1 pathway.²²⁵ Wang et al.²²⁶ have shown that Wnt-1-inducible signaling protein 1 (WISP-1) can mediate TGFβ1-driven renal fibrosis in a mouse model, and induction of TGFβ and collagen expression in fibroblasts by Wnt3a has been shown by Poon et al.²²⁷ Interleukins (ILs), such as IL-17A, type 2-associated IL-4, and IL-13 have been shown to induce fibrosis.^35,228–233 The transcription factor SPI1 (PU.1) polarizes resting fibroblasts and repolarizes ECM degrading inflammatory fibroblasts to an ECM-producing fibrotic phenotype,^35,234 and inactivation of SPI1 leads to regression of fibrosis.²³⁴ The glycoprotein OX40-OX40 ligand (OX40/OX40L) axis promotes collagen synthesis/fibrosis and regulates cytokine balance toward a proinflammatory and profibrotic profile.²³⁵ Blockade of this protein prevents inflammation driven fibrosis in skin, lung, vessels, etc., and anti-OX40L antibody can reduce the severity of bleomycin-induced fibrosis.²³⁵ Finally, fibrosis can also be induced by PDGF, integrins, and the microbiome, among other factors.^236–239

Fibrosis and AVF

Fibrosis had not been considered a major factor in AVF maturation failure until recently.²⁷ It has been suggested that fibrosis has a causative association with irregular shapes, disrupted membranes, pseudopodia-like projections, and intracellular collagen-containing vacuoles of VSMCs populating the media of pre-access veins.²⁴⁰ This indicates that vein fibrosis may result from and/or contribute to the transformation of differentiated VSMCs into the myofibroblastic and migratory phenotype that contribute to intimal hyperplasia (IH) and AVF failure.²⁴¹ This is supported by our recent findings using lineage tracing techniques that demonstrate differentiated venous VSMCs ultimately contribute to both medial wall thickening and IH in a mouse jugular branch end to carotid side anastomosis model.²⁴² In another study, we have found that one downstream target of the TGFβ/Smad pathway, tetraspanin 2 (TSPAN2), is downregulated in both an in vitro VSMC de-differentiation model and in failed human AVF samples.²⁴³ This further links vascular fibrosis with AVF maturation failure. In a different mouse model, Mishra et al.²⁴⁴ have found that an aorta-to-vena cava AVF would result in fibrosis around the capillaries as well as interstitial regions at week 6 post-surgery, and treatment with H₂S at this point will result in attenuation of pathological condition by week 8–10. Medial fibrosis is common in CKD patients,²⁴⁵ and increased vascular stiffness caused by medial fibrosis has been proposed as a limiting factor for arterial dilation and AVF maturation in patients with CKD.²⁴¹

Recently, we performed a study using venous samples pairs collected from patients who underwent two-stage upper-arm AVF surgeries at a single center, including pre-access veins from the first-stage access creation and AVF cross-sections from the second-stage transposition. Analysis of these samples showed that post-operative medial fibrosis and circumferential alignment of fibers around the lumen are closely associated with AVF maturation failure.²⁷ Meanwhile, we also noticed that non-matured AVFs presented higher level of collagen deposition in the media layer.²⁷ Interestingly, although not associated with stenosis development and AVF maturation failure by itself,²⁴⁶ post-operative IH showed an association with AVF maturation failure in patients with high post-operative medial fibrosis.²⁷ Consistently, we have also observed that AVFs with co-existing IH and medial fibrosis had a lower increase in venous diameter during remodeling.²⁷ A subsequent study further demonstrated increased collagen crosslinking in AVFs that failed and higher vascular expression of the collagen crosslinking enzyme lysyl oxidase (LOX).²⁴⁷

Our discovery of the association between fibrosis and AVF maturation failure is consistent with recent findings by Cai et al. Using RNA-sequence profiling of murine VSMCs, this group has shown that lower AVF patency rate is associated with higher expression of fibrotic genes.²⁴⁸ At 28 days after AVF creation, the patency rate, venous diameter, average neointima/media + adventitia area ratio, and areas and average cell density in neointima are all lower in female mice compared to male. In contrast, increased TGFβ1, TGFβRI, and collagen I are shown in VSMCs of the AVF from female mice, together with higher levels of α-SMA and pSmad3 and reduced bone morphogenetic protein 7 (BMP7) and interleukin 17 receptor b (IL-17Rb) expression. Most pSmad3 have been found to locate in the media layer and neointima, which promotes constrictive remodeling and venous fibrosis in the AVF from female mice.²⁴⁸ Furthermore, MR, being one of the receptors of Ang II which is critical to vascular fibrosis and IH,¹⁸⁷ has been found upregulated in dysfunctional AVF.¹⁸⁷

The knowledge regarding molecular cues and mechanisms responsible for the development of fibrosis in AVF is still limited. However, several studies have indicated that TGFβ may play an important role in AVF maturation failure,^249,250 while others have shown that TGFβ is heavily involved in vascular remodeling.^251,252 By performing bulk RNA-seq analysis of human pre-access veins, our group recently identified nine transcripts (false discovery rate [FDR] ≤ 0.05) encoding for fibrosis related proteins upregulated in veins that failed versus veins that had successful AVF maturation.³⁴ Among these nine genes, PF4/CXCL4 has been previously linked to the development of fibrosis.32,33

PF4, fibrosis, and AVF maturation failure

PF4 and fibrosis

PF4 is a very abundant chemokine that is released during platelet activation.²⁵³ It is mainly expressed by megakaryocytes, but can also be secreted from monocytes, macrophages, and DCs.^254–257 PF4 is involved in thrombosis, hemostasis, and innate immunity, including effects on monocyte and neutrophil chemotaxis,^258,259 as well as promoting host inflammatory response in many ways. PF4 activates granulocytes by causing release of lysosomal enzymes from neutrophils,²⁶⁰ stimulates CD4+CD25+ T cells proliferation,²⁶¹ prevents the apoptosis of monocytes and promotes their differentiation into macrophages,²⁶² and stimulates natural killer (NK) cells to produce IL8 and migrate in inflamed environments.^263,264

The functions of PF4 are mediated by its interaction with glycosaminoglycans, growth factors, lipoprotein receptor-related proteins, chemokine and growth factor receptors, integrins, or with several of those ligands at the same time.³¹ PF4 binds to C-X-C chemokine receptor 3B (CXCR3B) in microvascular endothelial cells to reduce DNA synthesis and promote apoptosis.²⁶⁵ It also induces intracellular Ca²+ release and the migration of activated human T lymphocytes via CXCR3A.²⁶⁶ PF4 drives the chemotaxis of monocytic THP-1 cells by binding to CCR1,²⁶⁷ and the interaction between PF4 and LDL-R leads to retention of LDL on fibroblasts surface, which affects LDL metabolism.²⁶⁸ Signaling through LDL receptor-related protein 1 (LRP1), PF4 promotes VSMCs to transform into a synthetic inflammatory phenotype characterized by a decline in differentiation markers, increased cytokine production, and cell proliferation.²⁶⁹ LRP1 acts as a repressor of the canonical Wnt pathway,²⁷⁰ and PF4 signaling through this receptor would contribute to Wnt pathway activation and β-catenin overexpression. It is believed that β-catenin could enhance TGFβ/Smads activity and increase cell differentiation as well as inflammatory and pro-fibrotic gene transcription.^{108,227,271,272} By interacting with integrins,^31,273,274 PF4 has been shown to interfere with the cell cycle in Chinese hamster ovary (CHO) cells^31,273 and cell migration in leukocytes.²⁷⁴

Interestingly, PF4 has also been reported to bind to one type of TGFβRI, activin receptor-like kinase 5 (ALK5), on NRK 49F fibroblasts and thus blocks it from binding to TGFβ1.²⁷⁵ The blockade of TGFβ signaling through ALK5 may have important implications for vascular remodeling upon tissue injury such as anastomosis creation in AVFs, because TGFβ signaling could be diverted through the ALK1 receptor instead. TGFβ signaling through ALK5 has been implicated with normal outward remodeling in AVF under laminar unidirectional blood flow, while signaling through ALK1 has been related to inward remodeling and stenosis when oscillating/disturbed blood flow is presented.²⁹ It is known that both pathways, ALK1-Smad1/5 and ALK5-Smad2/3, are activated and upregulated during vascular remodeling and fibrosis.²⁷⁶ Moreover, the interaction and balance between these two branches, and the presence of other factors and signals are all capable of modulating TGFβ pathway.^277–282 The role of PF4 on the regulation of this balance between branches has yet to be studied and would be an interesting topic to be investigated further.

PF4 has been previously shown as a fibrosis marker. Higher mRNA levels of PF4 can be seen in mice with liver fibrosis, while PF4 knockout mice have significantly reduce liver damage.³³ Activated platelets contribute to fibrosis through the release of PDGF, PF4, and TGFβ.^283–285 Increase of PF4 has been reported upon treatment of bleomycin, a drug that causes fibrosis, in lung tissue.²⁸⁶ In recent years, more evidence in line with this initial study have been reported. PF4 induces a pro-inflammatory and pro-fibrotic phenotype in monocyte-derived DCs, increasing ECM production, and inducing myofibroblast differentiation.²⁸⁷ Differential expression of hematopoietic PF4 marks the progression of bone marrow fibrosis, as the absence of PF4 decreases the activation of fibrotic pathways in megakaryocytes and inflammation in fibrosisdriving cells in three primary myelofibrosis models.²⁸⁸ PF4 is the marker of systemic sclerosis (SSc), a disease characterized by skin and internal organ fibrosis.^289,290 In patients who are suffering from this disease, higher levels of PF4 are produced in plasmacytoid dendritic cells (pDCs) in response to toll-like receptor 8 (TLR8),^291–293 and PF4 complexes with DNA/RNA to induce interferon α (IFN-α) activity in pDCs and stimulates the production of anti-PF4 antibodies from B-cells.²⁹⁰ Both increased levels of PF4 in circulation and anti-PF4 activity correlate with lung fibrosis.^290,293,294 At the same time, the presence of pDCs in the skin correlates with PF4 levels, supporting a mechanistic link for skin fibrosis in SSc patients.²⁹¹

On the other hand, Zhang et al.²⁹⁵ have shown that PF4 treatment decreases interstitial inflammation and fibrosis after renal transplantation at day 56. To our knowledge, this is the only report that indicates an inhibitory effect of PF4 on fibrosis.

PF4 in AVF maturation failure

As mentioned above, we have shown that PF4 is upregulated in veins that fail maturation after AVF creation,³⁴ and recently our group found localization of PF4 predominantly in the adventitia (Figure 1). The adventitial layer of veins consists of fibroblasts and may be important for the generation of fibrotic stenosis in failed AVF. Previous studies by our group and others have indicated that the vast majority of cells that comprise the hyperplastic intima are myofibroblasts.^296,297 A pig AVG model demonstrated that fibroblasts in the adventitia are able to differentiate into myofibroblasts by expressing αSMA,²⁹⁸ and these myofibroblasts are capable of migrating inward to form IH.^298,299 Meanwhile, PF4 can be detected from vascular stenosis tissue derived from chronic thromboembolic pulmonary hypertension (CTEPH) patients together with extracellular collagen I.³⁰⁰ All these results suggest that PF4 may play an important role in adventitial fibroblasts differentiation and stenosis generation. This is in line with our most recent discovery that PF4 treatment of isolated human primary vein fibroblasts could stimulate upregulation of αSMA and ECM components at both RNA and protein levels (Figure 2).

Figure 1.

Immunofluorescent staining of veins with PF4 (red). Stronger PF4 staining are shown in the adventitia (a, d and g). DAPI staining and the merged images are shown in (b), (e), (h), and (c), (f), (i), respectively. Each scale bar represents 10 μm.

Figure 2.

PF4 induces αSMA (ACTA2) and collagen I (Col1A1). The representative western blot is shown in (c). The statistics of the qRT-PCR of ACTA2 and Col1A1, and that of the western blot of ACTA2 and Col1A1 are shown in (a, b, d and e) respectively (n = 4).

Note: Part (c) is a representative figure, the quantification and significance for “*” of part (c) is showing in parts (d) and (e).

Very recently, we have found that the intracellular expression of PF4 and itgb6, the gene encoding for integrin β6 subunit, are associated to each other, and overexpressing PF4 will lead to upregulation of itgb6 at the protein level, while knocking down of PF4 results in downregulation of itgb6 (unpublished data). As mentioned above, integrin αvβ6 is required for the release of the active TGFβ from the LAP-TGFβ complex,¹²⁹ and TGFβ can directly stimulate the expression of ECM proteins^301,302 and is considered the most important factor for fibrosis.^107,108

Conclusion and future direction

The molecular cause of AVF maturation failure may be multifactorial but a common manifestation is overt vascular fibrosis combined with IH in the AVF.²⁷ Therefore, finding the mechanisms that promote fibrosis and myofibroblasts activation during post-operative AVF remodeling is crucial to the design of anti-stenotic therapies. There is emerging evidence that reveals a potential involvement of PF4 in fibrosis, including the finding of upregulation of PF4 in veins that fail.³⁴ Hypothetical mechanisms that link upregulation of PF4 with fibrosis onset and AVF maturation failure are summarized in Figure 3. Briefly, increase of PF4 in extracellular space may directly induce upregulation of αSMA and ECM proteins. Molecular features of fibrosis, such as fibroblasts differentiation into myofibroblasts, that is, expression of αSMA and ECM proteins, may be stimulated by PF4 via interaction with its various receptors. In contrast, increase of intracellular PF4 may lead to upregulation of itgb6 and thus elevates the level of αvβ6, which in turn promotes the activation of the TGFβ pathway and subsequently induces fibrosis.

Figure 3.

Potential mechanisms of PF4 involvement in fibrosis. Extracellular PF4 may directly induce upregulation of αSMA and ECM proteins (left-hand-side). In contrast, upregulation of intracellular PF4 may lead to increase of ITGB6 and membrane αvβ6, which may promote TGFβ pathway activation, and subsequently stimulate fibrosis (right-hand-side). All these mechanisms may contribute to fibrosis remodeling associated with AVF maturation failure.

Considering the multiple biological processes in which PF4 is involved including hemostasis, PF4 itself may not be a good therapeutic target to improve AVF maturation.^258,259 However, the downstream receptors of this molecule, especially those responsible for fibrosis onset, αSMA, and ECM protein expression, may serve as good candidates for developing therapies to reduce the rate of AVF maturation failure. Therefore, further studies that focus on PF4 and its downstream receptors should be performed to illustrate the molecular mechanisms behind the involvement of PF4 in fibrosis. This may shed light on our goal of improving AVF patency, HD delivery, and the quality of life of ESRD patients.