PLASMINOGEN ACTIVATOR INHIBITOR-1 (PAI-1): A KEY FACTOR LINKING FIBRINOLYSIS AND AGE-RELATED SUBCLINICAL AND CLINICAL CONDITIONS

Abstract

The close relationship existing between aging and thrombosis has growingly been studied in this last decade. The age-related development of a pro-thrombotic imbalance in the fibrinolysis homeostasis has been hypothesized at the basis of this increased cardiovascular and cerebrovascular risk. Fibrinolysis is the resulting of the interactions among multiple plasminogen activators and inhibitors constituing the enzymatic cascade, and ultimately leading to the degradation of fibrin. The plasminogen activator system plays a key role in a wide range of physiological and pathological processes. Plasminogen activator inhibitor-1 (PAI-1) is a member of the superfamily of serine-protease inhibitors (or serpins), and the principal inhibitor of both the tissue-type and the urinary-type plasminogen activator, the two plasminogen activators able to activate plasminogen. In this review, current evidence describing the central role played by PAI-1 in a number of age-related subclinical (i.e., inflammation, atherosclerosis, insulin resistance) and clinical (i.e., obesity, comorbidities, Werner syndrome) conditions is presented. Despite some controversial and unclear issues, PAI-1 represents an extremely promising marker which may become a biological parameter to be growingly considered in the prognostic evaluation, in the disease monitoring, and as treatment target of age-related conditions in the next future.

Introduction

One of the most relevant socio-economical phenomena occurred during the last decades is represented by the increasing age of Western Countries populations. It has been estimated that in 2050, 16.4% and 27.6% of the World and European populations will be represented by persons aged 65 years and older, respectively[1]. A third of the United States population in 2030 will be constituted by older persons[2]. About 10% of the Western Countries populations will be older than 80 years of age in the next 30 years[1]. The number of clinical conditions and need of social support increase with age. For example, in United States, subjects aged 65 years and older represent about 13% of the overall population, but are responsible for 35% of the health care costs[3]. Therefore, these estimates represent a major point of concern for public health, because of the possible consequences in the structure of our societies, and for the relevant effects on social, economical, and clinical burdens. Although age is not modifiable, the understanding of the aging process and the pathophysiological mechanisms underlying age-related conditions may still lead to the development of innovative preventive measures, and novel therapeutical targets.

Age is a well-established risk factor for cardiovascular disease, as also documented by the increasing incidence of thrombotic events with age[4]. The close relationship existing between aging and thrombosis has growingly been studied in this last decade[5]. The age-related development of a pro-thrombotic imbalance in the fibrinolysis homeostasis has been hypothesized at the basis of this increased cardiovascular and cerebrovascular risk[6–9].

Aging is responsible for a progressive impairment of the fibrinolytic system (as documented by an age-related prolongation of the euglobulin lysis time[9–11]), leading to pro-thrombotic modifications in circulating concentrations of hemostasis factors[9, 12–14]. It is noteworthy, before evaluating the potential consequences of such alterations, that the fibrinolytic system is composed by a wide spectrum of proteolytic enzymes which are involved in a multitude of processes (e.g, haemostatic balance, tissue remodeling, tumor invasion, angiogenesis, reproduction…)[15].

One of the most studied biomarkers of the fibrinolysis system is the plasminogen activator inhibitor-1 (PAI-1). Interestingly, its expression is not only elevated in the elderly, but also significantly enhanced in a variety of clinical conditions that are typical of the aging process (e.g., obesity, insulin resistance, psychosocial stress, decreased immune responses, increased inflammation, vascular sclerosis/remodelling)[16].

PAI-1: General Biology

Fibrinolysis is the resulting of interactions among multiple plasminogen activators and inhibitors consituting the enzymatic cascade ultimately leading to the degradation of fibrin. The plasminogen activator system plays a key role in a wide range of physiological and pathological processes, including coagulation, fibrinolysis, inflammation, wound healing, and malignancy[17]. A crucial reaction of the plasminogen activator system is the conversion of plasminogen to plasmin by plasminogen activators[18].

The main enzyme of the plasminogen activator system is plasmin. Plasmin is a serine protease playing a key role in the fibrinolysis cascade being responsible for the final degradation of fibrin and extracellular matrix proteins. Plasmin is similar to trypsin, and is generated from its precursor plasminogen by the plasminogen activators[17].

Plasminogen is primarily present in the plasma, and the liver represents its primary site of synthesis. However, plasminogen mRNA has been found in several mouse tissues, including adrenal, kidney, brain, testis, heart, lung, uterus, spleen, thymus, and gut, supporting the broad functional role played by the plasminogen activator system[19]. The activation of plasminogen into plasmin is mediated by two types of activators, urokinase-type plasminogen activator (u-PA) and tissue-type plasminogen activator (t-PA). The activity of both is regulated by specific plasminogen activator inhibitors (PAIs). The principal PAIs are PAI type 1 (PAI-1), initially called the endothelial cell PAI[20], PAI type 2 (PAI-2), also known as placental-type PAI[21, 22], and PAI type 3 (PAI-3), which is identical to protein C inhibitor[23, 24].

Among the inhibitory factors, the rapid acting PAI-1 is one of the most important inhibitors of the plasma fibrinolytic activity. PAI-1 is a single chain glycoprotein member of the superfamily of serine-protease inhibitors (or serpins). It is composed by 379 aminoacids with an apparent molecular weight of 48 kDa. PAI-1 (or serpin E1) is the principal inhibitor of both the tissue-type plasminogen activator (t-PA) and the urinary-type plasminogen activator (u-PA), which are able to activate plasminogen by cleaving a specific Arg-Val peptide bond located within the protease domain. Differently from t-PA (which is mainly involved in intravascular fibrinolysis), u-PA exerts proteolytic effects as well as intracellular signaling functions by binding to its high affinity receptor on the cell surface[25].

The activation of plasminogen by t-PA is highly dependent on the presence of cofactors, such as fibrin, that bind and alter the conformation of plasminogen[26]. PAI-1 is produced by endothelial cells, megakaryocytes, smooth muscle cells, fibroblasts, monocytes/macrophages, adypocytes, endometrium, peritoneum, liver cells, mesothelial cells, and cardiac myocytes[15, 20]. Once synthetized, PAI-1 is mainly stored in platelets, though it can also be secreted to blood flow or deposited on the subendothelial matrix. After release into the bloodstream, PAI-1 is present either in an active form or, more frequently, complexed with either t-PA or vitronectin (a relatively thermostable glycoprotein which is able to stabilize and convert PAI-1 into an active form)[27, 28]. The increased expression of PAI-1 in vivo suppresses fibrinolysis, consequently leading to the pathological fibrin deposition and tissue damage[29, 30]. Moreover, PAI-1 directly interacts with vascular cells and is involved in the regulation of cell replication and angiogenesis[31–35]. PAI-1 is also considered an acute phase reactant, being closely influeced by inflammatory cytokines (e.g., interleukin-6 [IL-6], interleukin-1 [IL-1], tumor necrosis factor-α [TNF-α]), growth factors (e.g., tissue growth factor-β [TGF-β]), and hormones (e.g., insulin, glucocorticoids, adrenaline)[27, 36–38].

PAI-1 concentrations vary by race/ethnicity[39, 40] and gender[41], although differences in body composition and adipose tissue distribution may account for (large part of) this variability. Plasma PAI-1 concentrations show a well-documented circadian variability. Its concentrations reach a peak in the early morning (when acute coronary syndromes present a maximal circadian)[42–44].

Plasma PAI-1 concentrations have shown to be (at least partly) genetically determined. An association between one of the DNA sequence variations of the human PAI-1 gene, the promoter −675 4G/5G polymorphism, and plasma PAI-1 concentrations has been suggested[28]. It has been hypothesized that the additional guanine in the DNA strand of the promoter region (5G allele) may interfere with the transcription process by conferring the binding site to a transcriptor inhibitor or by altering it for a transcription factor[28]. Thus, the 4G homozygotes tend to have the highest and the 5G homozygotes having the lowest PAI-1 concentrations[45–47]. However, such relationship is still under study and for some aspects controversial. For example, Mannucci and colleagues[48] confirmed the association between the promoter −675 4G/5G polymorphism and plasma PAI-1 concentrations in centenarians. They found a significantly higher frequency of the 4G allele, and the homozygous 4G/4G genotype was found to be associated with high PAI-1 levels even in this very old age group. However, since high PAI-1 concentrations are commonly considered a risk factor for atherosclerotic diseases and major clinical conditions in adulthood, it is puzzling finding the corresponding genetic marker in a population at the extreme limits of human life[48]. Subsequent studies[49–51] failed to report significant differences in the distribution of the high risk allele between centenarians and younger individuals, although confirming the age-related increase of plasma PAI-1 concentrations. It is possible that variations in polymorphisms associated with atherosclerotic diseases and/or major clinical conditions may act independently of longevity. Probably, factors other than this polymorphism are more involved in the variability of plasma PAI-1 concentrations. For example, Verschuur and colleagues[52] reported that body mass index is a stronger determinant of PAI-1 concentrations compared to the PAI-1 promoter variation. At the same time, it has been suggested this genotype is associated with obesity and possibly modulates changes in adipose tissue distribution at menopause[53], with possible consequences on phenotypical and clinical outcomes.

Just recently, another polymorphism (i.e., the Pro/Pro genotype of the p53 codon 72 polymorphism) has been identified as potential determinant of PAI-1 concentrations older persons[54]. Additional studies are needed to confirm these preliminary findings and explore this novel field.

Besides of genetic determinants, several factors are able to affect PAI-1 synthesis and secretion. Metabolic determinants (e.g., insulin resistance[55, 56], body mass index[57], plasma lipids[58, 59][60]) surely play a major role in determining plasma PAI-1 circulating levels. It is noteworthy that advancing age (a major contributor to increased PAI-1 expression[61]) is also associated with higher insulin resistance[62] and atherothrombotic risk[63, 64]. PAI-1 expression is upregulated by cytokines (e.g., TNF-α[65], TGF-β[65], and interleukins[66]), and by hormones (e.g., glucocorticoid[67], insulin[68], adrenaline[69], and angiotensin II[70]). PAI-1 is also strongly induced by the endotoxin (lipopolysaccharide) of Gram-negative bacteria[65, 71], able to profoundly alter the fibrinolytic system determining a pro-thrombotic state[72].

Special attention has been posed to mental and psychological stressors as possible determinants of the PAI-1 activity, and the PAI-1 has been indicated as a major stress-induced gene[73]. It is well-established that mental and physical stressors negatively influence the fibrinolytic activity[74, 75], potentially increasing the risk of thrombotic complications[73]. The stress response is a universally conserved cellular defense program, required for the homeostasis maintenance[76]. Stressful conditions (e.g., hyperthermia[77], acute inflammation[71], hypoxic stress[78], oxidative damage[79, 80], and fatigue/lack of energy[81]) produce rapid modifications in genes expression, including PAI-1[16]. Sympathetic nerve activity is associated with increased PAI-1 levels in older persons, suggesting this age-related modification as a potential contributor to the age-related increase of cardiovascular events[82]. In this context, it is noteworthy the activation of the hypothalamic-pituitary-adrenal axis and the sympathetic nervous system produced by stressors, leading to an increased secretion of glucocorticoids and adrenaline (two hormones able to induce PAI-1 expression in vivo[67, 69]).

Yamamoto and colleagues[73] have demonstrated that the magnitude of PAI-1 mRNA induction by stress is particularly enhanced in the adipose tissue, indicating adipocytes as primary actors. Interestingly and consistently with this observation, adipocyte-specific PAI-1 induction has been confirmed in obese mice[68] and in obese humans[83]. Therefore, if PAI-1 expression is dramatically induced by stress[84] with a particular contribution by the adipose tissue[43], obese individuals are particularly susceptible to stress-mediated pathological changes. In other words, excess of fat depots may enhance the stress-mediated induction of PAI-1, consequently causing thrombotic events in obese subjects[85].

Among exogenous stressors involved in the upregulation of PAI-1, smoking needs to be mentioned. In fact, it has been reported that smoking is associated with increased PAI-1 concentrations independently of age[86].

PAI-1 and obesity

Obesity, defined as an excessive amount of body fat in relation to lean body mass, is an increasingly prevalent condition in all age groups, including older persons[87]. The aging process is associated with relevant body composition modifications, namely a reduction of lean mass (beginning at about 20–30 years of age) and an increase of fat mass (peaking at 70 years of age). Moreover, a redistribution of fat depots (towards central and intramuscular adiposity) also occurs with aging, determining increased risk of insulin resistance and metabolic conditions[88, 89]. It has been estimated that the prevalence of obese older persons in the United States is currently well above the 30%[90].

During the last two decades, clinical studies have more and more clearly demonstrated that obesity is associated with impaired fibrinolysis[57, 91–95]. The concept that adipose tissue is not an inert energy storage site, but an active organ with crucial endocrine and metabolic functions has been growingly accepted in these last years, especially since it has been demonstrated its capacity to synthesize and secrete a variety of cytokines (the so-called adipokines)[96].

Among the adipokines (most of them characterized by pro-inflammatory properties), PAI-1 is one of the most relevant[97]. The excess of adipose tissue increases the production of PAI-1, leading to an impairment of the fibrinolytic system[83, 98]. Moreover, obesity is currently considered a low-grade inflammatory state[99], and is responsible for a series of inflammatory cytokines (in particular, IL-6 and TNF-α[100]) able to induce the PAI-1 overexpression. In this context, it is noteworthy that obesity is associated with increased adipose tissue macrophages infiltration (up to 40%)[101], and a change in adipose tissue macrophages polarization to a more pro-inflammatory state[102]. About one third of the total circulating IL-6 concentration is synthesized in the adipose tissue[103–105]. Obesity has also shown to be independently correlated with CRP levels[106, 107].

As mentioned, PAI-1 is partly produced and released by human adipose tissue[97, 108, 109]. Significant correlations of PAI-1 with a variety of adiposity measures (e.g., body mass index, waist circumference, waist-to-hip ratio, total fat mass, visceral adipose tissue, subcutaneous adipose tissue) and metabolic syndrome features (e.g., HOMA index, inflammatory biomarkers, insulin, glucose, triglycerides, HDL cholesterol [inverse]) features have been reported[110]. Sawdey and Loskutoff[65] originally reported that adipose tissue in mice expresses high PAI-1 concentrations. In a subsequent study it was shown that obese animals present several-fold increased plasma PAI-1 concentrations compared to lean mice, likely due to the higher amount of expressing adipose tissue[27]. Consistently, fat cells from obese subjects produce and secrete twice more PAI-1 than those from lean subjects[108, 111]. PAI-1 concentrations are also related to the lipid content and the volume of fat cells[108].

In a recent study, Sam and colleagues showed that visceral adipose tissue is positively associated with PAI-1, even after adjustment for body mass index[112]. Therefore, the amount of visceral adipose tissue provides more information on the fibrinolytic status compared to the measure of body mass index (which is still a mere measure of mass, and a gross correlate of adiposity). In this context, it is important to mention that the fat territory does not equally contribute to the secretion of this adipokine. In fact, visceral fat produces more PAI-1 than subcutaneous or femoral fat[97, 113]. Yudkin and colleagues[114] found no evidence that subcutaneous adipose tissue contributes significantly to circulating PAI-1 levels in lean subjects. On the other hand, Mavri and colleagues[98] demonstrated that only PAI-1 expression in the subcutaneous abdominal depot contributes to plasma PAI-1 concentrations, whereas no significant result was obtained for the relationship between PAI-1 mRNA in the subcutaneous femoral adipose tissue and circulating PAI-1 protein. Moreover, several studies have shown that increased levels of PAI-1 are particularly evident in subjects with abdominal obesity[57, 91, 92].

PAI-1 and inflammation

Inflammation is a complex of host’s normal defense reactions to internal and external stressors underlying aging and age-related diseases[115]. Its crucial role in determining and modifying the aging process has even led to the creation of neologism "inflamm-aging"[116]. If the age-related inflammation (or inflamm-aging) trespasses an hypothesized individual threshold, the transition between successful and unsuccesful aging occurs.

As mentioned above, an important aspect of PAI-1 resides in its close relationship with inflammation[17]. The link between inflammation and fibrinolytic system is well-supported. Experimental in vivo studies performed in animal models as well as in humans have shown that TNF-α and IL-6 are major contributors to the increases of PAI-1[117, 118].

Among the most studied pro-inflammatory cytokines, TNF-α is one of the most relevant. It is an important stimulator of the PAI-1 expression in the adipose tissue[119, 120]. Sawdey and Loskutoff[65] were the first to report that TNF-α is involved in the regulation of PAI-1. Moreover, the expression of TNF-α is chronically elevated in adipose tissue both from obese rodents and humans[121]. Administration of TNF-α to lean mice significantly increased PAI-1 mRNA in adipose tissue, leading to a pattern similar to that found in obese mice[119]. Similar results were obtained in studies on human cells from adipose tissue[109, 111, 122] and endothelium[123]. Moreover, incubation of human fat biopsies with TNF-α inhibitors markedly reduced the production of PAI-1 mRNA[109]. Similarly, obese mice treated with antibodies against TNF-α showed increased insulin sensitivity,reduced plasma PAI-1 antigen concentrations, and lower levels of adipose-tissue PAI-1 and TGF-β mRNA[120].

IL-6 is an acute phase inflammatory reaction protein produced after TNF-α and IL-1 stimulation[124]. IL-6 induces the synthesis of CRP[125, 126], haptoglobin, and fibrinogen[127], and inhibit the synthesis of other reactants (e.g., antithrombin, albumin, and prealbumin[128]). IL-6 also functions as a traditional hormone by stimulating the hypothalamic-pituitary-adrenal axis to produce cortisol, the feedback of which inhibits IL-6 production[124]. IL-6 has relevant pro-thrombotic effects and promotes platelet aggregation[127, 129, 130]. It has been reported that the injection of IL-6 significantly increases PAI-1 antigen and tPA concentrations in animal models[127]. Treatment of human adipose tissue cells with IL-6 resulted in a ten-fold increase of PAI-1 mRNA and antigen release[131].

Interestingly, Woodhouse and colleagues[132] reported that body mass index (a gross correlate of adiposity), blood neutrophil count (an inflammatory parameter), and (inversely) serum ascorbate (an antioxidant marker) are significant predictors of PAI-1 concentrations, independently of potential confounders (including smoking and vitamin C supplementation use). This results may support the hypothesis of a vicious cycle existing among inflammation, oxidative (and/or antioxidant) status, and platelet activation as promoter of those pathophysiological modifications responsible for age-related conditions (including atherosclerotic diseases) and, ultimately, disability and mortality[133].

Several studies support the hypothesis that TNF-α[120, 122] and other cytokines (e.g., IL-1β)[122] may (at least partly) be responsible for the expression of TGF-β, another major stimulant of the PAI-1 biosynthesis in adipose tissue both from rodents[27, 65] and humans[111, 122, 134]. In humans with severe obesity, a direct correlaton between in PAI-1 and TGF-β mRNA expression has been reported in the visceral and subcutaneous adipose tissues[134]. Moreover, Birgel and colleagues showed that in vitro differentiated human adipocytes express TGF-β and two of three receptor subtypes[122]. Thus, it is plausible assuming that the upregulation of PAI-1 in the obese state involves a paracrine mechanism of TGF-β.

PAI-1 and sarcopenia

Adipose tissue is closely and inversely related to lean mass. This relationship becomes clearly evident with aging, whereas fat mass tends to increase while muscle mass undergoes a quantitative and qualitative reduction. This age-related modification of body composition, determining the age-related phenomena of "sarcopenia"[135] and "sarcopenic obesity"[136], has received particular attention in the last decade as potentially responsible of major health-related events in older persons (e.g., physical impairment, disability, mortality). It is interesting to note that while at the beginning of the exploration of this topic muscle mass was playing a primary role as determinant of negative health-related events, the latest contributions in the field have showed a major relevance of adipose tissue[137–139]. In particular, several reports have suggested that adipose tissue, and in particular intramuscular fat infiltrates, are able to significantly decrease the overall muscle quality, determining a reduction in its functionality[139, 140]. If adipose tissue is the endocrine and pro-inflammatory organ we previously described, the effect of fat depots location on skeletal muscle represents a novel field of research that needs to be pursued. In fact, the adipokines we previously presented with so many deleterious effects on homeostasis and cardiovascular health, also present detrimental effects on the skeletal muscle[141–145]. It is possible that the multiple stressors able to induce the PAI-1 expression in adipose tissue may similarly act in the intramuscular fat infiltrates typical of the aging muscle. The consequence of this induction might be the promotion of a pro-thrombotic state and proliferation of extracellular matrix within the muscle. This scenario may shift the equilibrium between proteolysis and protein synthesis (necessary for the mantainance of muscle mass) towards the former due to the possible catabolic and apoptotic stimuli of inflammation, oxidative damage, thrombosis, and hypoxia.

In a previous study[143], we reported no significant association between sarcopenia measures and circulating PAI-1 concentrations. However, these negative findings do not exclude that PAI-1 may still have a locally rather than systemic role in determining the loss of muscle mass. Interestingly, fibrosis (a common complication of prolonged muscle atrophy) is a well-described pathological hallmark of chronic myopathies[146]. Excessive fibrosis is responsible for lower muscle strength and elasticity, and inhibits the nutrients diffusion to myofibers[147]. Moreover, Zhu and colleagues[148] recently demonstrated that myostatin, an inhibitor of skeletal muscle growth, is characterized by fibrogenic properties as it stimulates proliferation of muscle fibroblasts and consequent production of extracellular matrix proteins. The excessive accumulation of extracellular matrix molecules can be counteracted by two major proteolytic systems represented by plasmin and the matrix metalloproteinases. For the purpose of the present review, the former is of particular interest. In fact, plasmin is able to limit extracellular matrix proteins accumulation, and its regulation (as previously mentioned) is mainly due to the PAI-1 activity. This concept is supported by studies in human and animal models showing that PAI-1 is upregulated (and, consequently, plasmin is downregulated) in glomerulosclerosis and liver, pulmonary, and cardiac fibrosis[149–152]. In a recent study, Naderi and colleagues[153] reported a seven-fold increased PAI-1 mRNA expression in muscle biposies from patients with neurogenic atrophy compared with control patients. Therefore, current evidence suggest that PAI-1 may constitute a novel target in diagnosing, monitoring, and hopefully treating muscular atrophy and atrophy-related fibrosis[153].

PAI-1 and metabolic syndrome

The metabolic syndrome, a risk factor for cardiovascular events[154, 155], is constituted by a combination of chronic inflammation, insulin resistance, and obesity. Recently, the well-established link existing between metabolic syndrome and cardiovascular disease has also been confirmed in older persons[156]. Since chronic inflammation and fat-driven body composition modifications represent two major features of the aging process, the high (and increasing) prevalence of metabolic syndrome in older persons is easily explicable. It has been estimated that nearly 45% of Americans aged 50 years ad older present this clinical condition[157].

Metabolic syndrome is characterized by a pro-thrombotic state due to the inhibition of the fibrinolytic pathway[158, 159]. In this scenario, PAI-1 may represent a very promising marker to study. In fact, it has been demonstrated that fibrinolytic dysfunction (defined by PAI-1 levels) mediates the increased risk of CVD in individuals with metabolic syndrome[160], and increased PAI-1 concentrations have been found both in blood and in coronary plaques of metabolic syndrome patients[15]. Moreover, plasma PAI-1 concentrations have been associated with metabolic syndrome features[161], especially with adipose tissue, insulin resistance, and lipidemic parameters[159, 162, 163].

A chronic low-grade inflammatory status has been indicated as a potential determinant or contributor to the metabolic syndrome. Among the wide range of inflammatory markers, the adipokine TNF-α has received special attention as a major stimulus upregulating the PAI-1 expression and secretion[111, 122, 164, 165]. Recently, You and colleagues[166] demonstrated a strong relationship between metabolic syndrome (and its severity) and adipokines (i.e., leptin, IL-6, PAI-1, and TNF-α) in older adults, independently of adiposity. Similar findings were also reported by Kressel and colleagues[167], documenting a positive correlation between the homeostasis model assessment (HOMA) index[168] and adipokines (including PAI-1).

It has been proposed that PAI-1 synthesis in metabolic syndrome might be promoted by the numerous homeostatic disturbances characterizing this condition. Experiments conducted in cellular coltures showing insulin, glucocorticoides, lipoproteins, triglycerides, and glucose promoting the PAI-1 production tend to confirm this hypothesis[58, 162, 169]. In particular, studies on cultured adipocytes showing the enhancement of PAI-1 production by insulin[68] and glucose[170] confirm its involvement in the insulin resistance condition[94].

PAI-1 and atherosclerosis

Atherosclerosis, an age-related vascular phenomenon, represents a major public health issue, being its clinical manifestations responsible for significant morbidity and mortality, especially in Western countries[171]. Although aging is clearly involved in the atherogenic process, a complete understanding of such relationship is not yet available. In fact, it is unclear whether the atherosclerotic lesions are caused by intrinsic and unalterable aging mechanisms in the arterial wall, or the age-dependent result of a chronic exposure to risk factors[172]. In other words, aging and atherosclerosis may share the same time-frame, or there may be a causal interaction between them[173]. Whatever the mechanism is, evidence clearly demonstrates a progressive accumulation of vascular modifications and dysfunctions with age[174].

Atherogenesis is a complex process involving vascular injury, lipid accumulation, and platelet and fibrin deposition. The vessel wall is a dynamic organ composed of distinct layers of cells (i.e., endothelial cells, vascular smooth muscle cells, fibroblasts) interacting in a complex autocrine-paracrine manner. It senses and actively responds to diverse vascular stressors, including mechanical forces, vasoactive and humoral factors, inflammatory agents, and thrombotic events[175].

Vascular tissue remodelling is the active process essential for the response to pathological stimuli, and includes a multitude of structural and functional modifications of the vessel wall via modulation of cellular adhesion, migration, proliferation, production and degradation of extracellular matrix, and cell death. Vascular injury frequently occurs at the earlier phases of the vascular remodelling, often leading to the activation of the coagulation cascade, and a consequent pro-thrombotic state with fibrin deposition. In addition to functioning within the vascular lumen to control fibrinolysis, the plasminogen activator system is active within the blood vessel wall, where it plays an important role in controlling vascular remodeling. Studies support a direct relationship between vascular remodelling and PAI-1 given the overexpression of PAI-1 (mRNA and protein) in the vascular wall adjacent to an arterial thrombus induced by mechanical injuries[175, 176].

Animal models have demonstrated a three-fold upregulation of plasma PAI-1 concentrations in apolipoprotein E (apoE) deficient mice (more prone to develop atherosclerosis) compared to wild-type mice[177]. Moreover, the overexpressed PAI-1 within vascular smooth cells reduced the cellularity of neointimal lesions in ApoE deficient mice[178]. This finding may support the hypothesis that increased expression of PAI-1 in atheroma may promote plaque rupture because of the lower cellular component of the fibrous cap[178]. In fact, locally elevated PAI-1 concentrations may predispose to acellular, thin-walled atherosclerotic plaques (consequently at risk of acute rupture). After the plaque rupture, a sudden release of PAI-1 from the injured plaque and platelets occurs, leading to a (more than 10-fold) increase of its concentrations, with consequent promotion of thrombogenesis[179, 180].

The presence of an increased PAI-1 expression in atherosclerotic lesions and atheromas is well-established[181–183], suggesting a primary role of this adipokine in atherogenic process[184]. Nevertheless, evidence on this topic is still controversial, since PAI-1 has shown both promoting and preventive properties in the vascular remodelling processes, a phenomenon sometimes referred as the "PAI-1 paradox"[175, 179]. It is likely that the complex vascular functions of PAI-1 may depend on the vascular bed, type of lesion, the experimental/clinical conditions, and the different molecules interacting with it[18, 175]. Fay and colleagues[18] recently explained that when vascular injury is associated with activation of the coagulation system and fibrin formation, PAI-1 may have atherogenic properties by stabilizing fibrin and, therefore, providing support to vascular smooth muscle cells migration. Alternatively, in the absence of fibrin formation, PAI-1 may inhibit cell migration within the vascular wall by inhibiting u-PA and/or blocking interactions between cells and vitronectin in the extracellular matrix. This anti-migratory effect of PAI-1 may inhibit formation of intimal hyperplasia, and simultaneously reduce the cellularity within the fibrous cap of atherosclerotic plaques (consequently promoting their rupture)[178, 179]. However, PAI-1 also exert direct anti-apoptotic and proliferative effects on proliferation of vascular smooth muscle cells[33, 185]. In this case, the cellular proliferation may be dominant and drive the vascular pathology, with PAI-1 responsible for the neointimal growth[18].

Studies performed on arteries of atherosclerotic subjects have documented significantly higher levels of PAI-1 mRNA in severely atherosclerotic vessels compared with subjects having normal or mildly affected arteries[181]. Moverover, it is well-known that a decreased fibrinolytic capacity (with increased plasma PAI-1 concentrations) is present in coronary artery disease[15, 186–188]. However, the exact role played by PAI-1 has been questioned because often determined by third factors potentially explaining the association fibrinolysis-atherosclerotic disease (e.g., diabetes mellitus, hypertension, obesity, dyslipidemia)[189–191]. For example, Folsom and colleagues reported that the association between PAI-1 and carotid intima-media thickness was largely resized after adjustment for lifestyle and medical covariates[192]. Interstingly, transgenic mice overexpressing PAI-1 developed coronary artery thrombosis in an age-dependent fashion (i.e., no thrombi in mice younger than 4 months; 90% of spontaneous thrombotic occlusions in mice older than 6 months), suggesting the possible role of additional age-related factors contributing to thrombogenesis[193].

The possible relevance of third factors in the explanation of the link between fibrinolysis and atherosclerotic conditions has indicated the metabolic syndrome as an interesting model to study. A bidirectional interaction between classical determinants of the metabolic syndrome and PAI-1 overexpression is a well-accepted concept, despite the difficulties in understanding the underlying cause-effect relationship. The adipose tissue accumulation typical of the metabolic syndrome produces a wide range of inducing factors for PAI-1 expression (e.g., TNF-α, TGF-β, cortisol, angiotensin II, oxidative stress, hypoxia) while, simultaneously, PAI-1 overexpression interferes with insulin signalling and adipocyte differentiation favouring the development of obesity and glucidolipidic disturbances characteristic of the metabolic syndrome. Therefore, PAI-1 has been indicated as a true component of the metabolic syndrome with a pathogenic role per se, with potentialities to be an attractive target for future interventions[194].

PAI-1 and diseases

The number of clinical conditions increases with age, independently of gender[195]. The presence of comorbidity (i.e., the concomitant presence of multiple diseases in the same individual) represents a major issue in older persons due to the relevant clinical[196, 197] and socio-economical consequences[198]. Franceschi and colleagues[116] suggested that exogenous (or environmental) stressors act on an endogenously (or biologically) predisposed individual to develop age-related diseases. Therefore, aging is a major determinant of this mechanism since a longer exposure leads to a higher risk of phenotypical alteration (i.e., onset of clinical disease).

The impairment of the fibrinolytic cascade (and the increase of PAI-1 concentrations) is easily associable with a wide range of thrombotic conditions[199], including myocardial infarction[187], stroke[200], peripheral artery disease[201], deep vein thrombosis[202–204], and disseminated intravascular coagulation[205]. Several histopathological conditions are also easy to relate to the increased PAI-1expression (e.g., lung fibrosis[151], glomerulonephritis[206, 207], and atherosclerosis[181, 208]). However, PAI-1 levels are increased also in a large number of other clinical conditions with no apparent/direct relationship with the fibrinolytic system. For example, PAI-1 concentrations have been found elevated in most of patients with (and at risk of) adult respiratory distress syndrome[209], and in early and late age-related maculopathy[210].

High circulating levels of PAI-1 have shown to significantly predict the onset of myocardial infarction[191]. Moreover, an acute increase in plasma PAI-1 concentrations in patients with acute ST-elevated myocardial infarction is a poor prognosis factor for short-term survival[211]. Although the 4G allele of the PAI-1 promoter 4G/5G polymorphism is associated with higher PAI-1 levels, definitive conclusions about the association between this polymorphism and the risk of cardio-and cerebrovascular events may still be too premature. A relationship between the 4G/4G genotype and myocardial infarction has been reported[47, 212]. Nevertheless, evidence is still too limited, and it is currently more reasonable to state that the risk due to the genetic component (i.e., PAI-1 promoter 4G/5G polymorphism) for developing major atherosclerotic conditions is neglegible compared to traditional risk factors and/or other novel biomarkers of risk[213].

An increased PAI-1 expression is a common feature of diabetic patients, potentially explaining their higher cardiovascular risk and unfavorable plaque evolution[214]. Festa and colleagues[215] demonstrated in a large cohort of healthy non diabetic subjects that increased levels of PAI-1 at baseline were significantly predictive of incident diabetes over the 5 years of follow-up. In a subsequent study[216], the same group of authors reported that the progression of PAI-1 concentrations over time was also associated with incident diabetes.

Elevated plasma PAI-1 levels have been indicated as a strong risk factor for stroke at old age[217]. The increased plasminogen activation may increase the laminin degradation in extracellular brain tissue, consequently reducing the brain resistance to ischemic damage[218]. At the same time, PAI-1 may also have detrimental effects as an acute phase molecule[219]. Interestingly, a recent paper proposed the adoption of three immuno-inflammatory and thrombotic/fibrinolytic markers (i.e., TNF-α, PAI-1, and t-PA) as screening measures to predict the stroke diagnosis in the acute care setting[200].

There is compelling clinical evidence considering PAI-1 as a key factor for tumor invasion and metastasis. High PAI-1 concentrations are considered a poor prognostic factor in a variety of cancers, including breast[220–224], lung[225], colorectal[226], and gastric cancer[227]. Several studies in animal models have shown that tumor growth is inhibited in PAI-1-deficient mice and stimulated in mice overexpressing this adipokine[228, 229]. PAI-1 may contribute to cancer dissemination through modulation of cell adhesion[230, 231], support of angiogenesis[232, 233], and stimulation of cell proliferation[234]. A genetic influence by PAI-1 promoter −675 4G/5G polymorphism has also been proposed in the modulation of the cancer invasion[235].

Hypofibrinolysis due to high PAI-1 concentrations is also commonly found in patients with high thrombotic risk, such as obesity and metabolic syndrome[159], pregnancy[236, 237], and in certain obstetric complications (e.g., preeclampsia)[238, 239]. Physical disability has been associated with higher PAI-1 levels, maybe because of the higher prevalence of vascular diseases and chronic inflammation disabled patients present[240]. In this field, a recent study adopting a principal component analysis approach by Hsu and colleagues[241] showed that PAI-1, IL-6, and CRP simultaneously contribute to determine physical function in older persons.

PAI-1 and Werner syndrome

Werner syndrome is an inherited progeroid syndrome firstly described by Werner in 1904[242]. It is characterized by premature aging, that is high incidence of degenerative disorders (including atrophy of skin, subcutaneous tissue, muscles and reproductive organs, impaired wound healing, lenticular cataracts, osteoporosis, severe atherosclerosis, hyperinsulinemia, hypercholesterolemia, type 2 diabetes mellitus, benign and malignant tumors)[243, 244]. Among these complications malignant tumors and atherosclerosis represent the major causes of death in patients with Werner syndrome. Autoptic studies performed in Werner syndrome patients have reported an unsual prevalence of severe atherosclerosis[243]. The cause of such severe atherosclerosis was initially attributed to several of the accompanying complications (e.g., diabetes mellitus, hyperinsulinemia, hyperlipidemia, hypertension)[245–247]. This explanation became insufficient when Murano and colleagues reported about a Werner syndrome patient with severe atherosclerosis despite of the absence or only moderate presence of other traditional risk factors[248]. The search for alternative/complementary causes of enhanced atherosclerotic disease in Werner syndrome patients led to suspect possible alterations of the fibrinolytic system. In fact, Murano and colleagues[248] showed that all the five Werner syndrome cases they described presented plasma PAI-1 concentrations above the upper limit (50 mg/mL) of the normal range. Therefore, the enhanced induction of those genes regulating the expression and release of PAI-1, as demonstrated in cultured skin fibroblasts from a subject with Werner syndrome[249], has been hypothesized. The importance of an impaired fibrinolysis at the basis of the pro-atherogenic status of the Werner syndrome is consistent with an earlier report describing high concentrations of blood fibronectin in this progeroid syndrome[250].

PAI-1 and potential therapies

Given the preminent role played by the fibrinolytic pathway in the development of several major health-related events, numerous studies have tested possible behavioral and pharmacological interventions aimed at limiting the detrimental pro-thrombotic state. Most of these studies have used PAI-1 as a marker of primary interest in the evaluation of the success or failure of the proposed intervention.

Physical exercise has shown to be one of the most successful intervention on a wide range of subclinical and clinical conditions. Encouraging and beneficial effects of it have also been consistently reported on PAI-1 circulating levels (and, therefore, on the fibrinolysis). Physical exercise enhances fibrinolytic activity by increasing t-PA activity and lowering PAI-1 concentrations[251]. It has been reported that regular physical exercise lowers markers of platelet activation[9, 252] and prevents age-related endothelial dysfunction[253]. In a clinical trial on patients with intermittent claudication due to peripheral artery disease, Killewich and colleagues[254] showed that patients enrolled in the physical exercise group experienced significant improvements of the fibrinolytic activity (i.e., 23% decrease in PAI-1 activity, 28% increase in t-PA activity). In particular, participants with baseline higher levels of PAI-1 were also those obtaining the major benefits from the intervention. More recently, a large study of older persons has demonstrated that current physical activity level is inversely associated (in a dose-dependent manner) with a wide range of hemostatic and inflammatory biomarkers (including CRP, fibrinogen, platelet and white cell count, and t-PA antigen [which reflects PAI-1 levels]), independently of prevalent cardiovascular disease, smoking habit, and obesity[255]. Consistent results were also obtained by DeSouza and colleagues[256] showing significant PAI-1 reductions soon after physical exercise in both hypertensive and normotensive subjects. This beneficial effect was maintained after 30 and 60 minutes post-exercise in both groups[256].

One of the best strategies to improve the obesity-related fibrinolytic impairment is also represented by the body weight loss[257]. Weight loss has been associated with a marked reduction in PAI-1 expressions[111, 258], with studies more generally reporting significant improvements of the overall disturbed fibrinolytic system[259]. It is possible that elderly and obese subjects (presenting both increased inflammation and PAI-1 levels) may particularly benefit from this intervention not only by relieving inhibition of the fibrinolytic system[260], but also from its anti-inflammatory properties[261, 262]. It has been suggested that the weight loss-induced reductions of PAI-1 mRNA levels measured in one fat depot do not necessarily reflect changes occurring in other fat depots nor are representative of plasma PAI-1 concentrations[98, 263]. This variability is consistent with the different production of PAI-1 reported for different adipose tissue depots[83, 91, 257].

A relationship between alcohol consumption and inflammation has been suggested by several studies, possibly indicating a beneficial action in the prevention of cardiovascular disease. However, no significant association of alcohol intake with TNF-α and PAI-1 concentrations has been reported[264].

Several pharmacological interventions have been proposed as potential modifiers of the fibrinolytic process. In this context, particular interest has been raised by hypoglycemic agents. Thiazolidinediones, such as troglitazone or pioglitazone, have demonstrated beneficial effects on the fibrinolytic system, especially on PAI-1 levels[265–267]. At the same time, these molecules have shown to positively influence other components of the metabolic syndrome, including inflammation[268].

The activation of the renin-angiotensin system is closely related to PAI-1[269]. Angiotensin II (which is also produced by the adipose tissue) is the biologically active product of angiotensinogen processing, and has shown to stimulate PAI-1 expression in human adipocytes at the transcriptional level[270]. Beneficial effects of ACE-inhibitors on cardioprotection seem to be mediated by mechanisms that are partly independent of blood pressure lowering. Unfortunately, evidence is still insufficient and controversial for what concerns their effect on the fibrinolytic cascade. Vaughan and colleagues[271] reported positive effects of ramipril on PAI-1 concentrations in patients with acute myocardial infarction. At the same time, no significant effect of fosinopril versus placebo was reported on PAI-1 levels in a recent trial we performed in subjects with a high cardiovascular risk profile[272]. Lack of data to draw definitive conclusions also exists for the role played by angiotensin receptor antagonists in the regulation of fibrinolysis[273–275].

Another class of medications showing promising results about their capacity to limit the development of a prothrombotic state are hypolipidemizing drugs[17, 276], in particular statins. Statins have relevant effects on the plasminogen activator system[17]. For example, it has been shown that a 4-week treatment with simvastatin in rabbits fed with an atherogenic diet was able to significantly reduce plasma PAI-1 and PAI-1 mRNA levels[277]. In human endothelial cells and vascular smooth muscle cells, statins downregulate the activation of inflammatory transcription factors (including nuclear factor kappa-light-chain-enhancer of activated B cells, or NF-kB), potentially explaining their potential anti-inflammatory role[278]. In line with the hypothesis of an inhibitory effect of statins on inflammation, statins have been found to counteract the enhancing effect of TNF-α on PAI-1[279]. Consistently, Qu and colleagues recently showed that, besides of improving lipid profiles, atorvastatin and rosuvastatin treatments decrease PAI-1 and CRP concentrations[280]. Beneficial effects of fibrates on PAI-1 antigen and activity have been reported in clinical studies, although results are still inconsistent and insufficient[281, 282].

Conclusions

There is still a lot to understand in the complex network of interactions existing between the fibrinolytic system, inflammation, oxidative (and antioxidant) status, adipose tissue (and skeletal muscle), metabolic syndrome, and atherosclerotic diseases. Current evidence, despite some contradictions and controversies, tend to indicate PAI-1 as an extremely promising marker potentially linking several of these pathways, organs, and conditions. This adipokine may become a biological parameter to be considered in the prognostic evaluation, in the disease monitoring, and as treatment target in the next future.