Hypertension, matrix metalloproteinases and target organ damage.

Abstract

The renal vasculature is highly susceptible to deleterious structural and functional changes induced by high blood pressure. In particular, hypertension is associated with an increase in extracellular matrix (ECM) deposition (especially collagen types I, III and IV) in renal resistance vessels, glomeruli and the interstitium, which results in the characteristic appearances of renal and glomerular fibrosis, and thickening of the glomerular basement membrane (GBM) [1]. Within the tubulointerstitial compartment of the kidney, transforming growth factor-β (TGF-β), connective tissue growth factor, endothelin-1 and angiotensin II play important roles in initiating fibrogenesis, in part by increasing rates of matrix protein synthesis and by altering processes that regulate ECM remodelling and turnover. In the 1980s, Gonzalez-Avila et al. [2] first highlighted the importance of impaired degradation of matrix proteins in the pathogenesis of interstitial fibrosis. They showed that collagen synthesis rates were unchanged but collagenolytic activity decreased 10- to 30-fold in a model of severe fibrosis induced by renal vein ligation [2]. The mechanisms underlying ECM accumulation in hypertensive renal disease are poorly understood, yet there is a need for targeted interventions to prevent, or reverse, the structural adaptations to high blood pressure in the kidney that lead to proteinuria, nephron loss and progressive functional impairment. Recent experimental studies with endothelin antagonists [3] and angiotensin receptor blockers [4] have provided useful information with regard to matrix protein synthesis, but the effects of hypertension on ECM degradation remain unclear. In this issue of the journal, Camp et al. [5] provide new data showing that hypertension has specific effects on pathways regulating the proteolysis of different ECM components in the kidney [5]. Nomenclature and regulation of ECM degradation There are four superfamilies of tissue proteinases that are variably expressed in different tissues: aspartic, cysteine, serine and metallo-type proteinases. Normal kidneys produce a number of proteases with specificity for matrix proteins, including some serine proteases and several enzymes that belong to the metalloproteinase family. Matrix metalloproteinases (MMPs) are zinc-dependent endoproteinases with the combined ability to degrade all the components of ECM at physiological pH [6]. More than 20 MMPs have been cloned and characterized, and classified into five main groups according to structural similarities and substrate affinities [7]: (i) the interstitial collagenases (e.g MMP-1 and MMP-8, which degrade collagen types I and III); (ii) two gelatinases, MMP-2 (also known as gelatinase A) and MMP-9 (also known as gelatinase B), which are particularly efficient in degrading collagen-IV, the major structural component of the mesangial matrix and GBM [8], and elastin; (iii) the stromelysins (e.g MMP-3 and MMP-10), which have broad substrate specificity for proteoglycans, fibronectin, laminin and GBM collagens [9]; (iv) the membrane-type (MT) subgroup of six MMPs (MT1-MMP to MT6-MMP), which play important roles in endothelial tubulogenesis and neovascularization [10]; and (v) a heterogeneous subgroup (e.g MMP-7, MMP-12, MMP-19 and MMP-20). MMPs are involved in a variety of physiological and pathological processes related to connective tissue turnover, remodelling, angiogenesis and atherosclerosis. Various cytokines, hormones and growth factors, as well as shear stress and oxidative stress, regulate MMPs at three levels (Fig. 1): (i) induction of gene expression; (ii) activation of the latent proenzymes; and (iii) inhibition by tissue inhibitors of metalloproteinases (TIMPs), of which TIMP-1, -2, -3 and -4 have been described [7]. A number of cytokines and growth factors stimulate transcription of MMPs (e.g interleukin-1, platelet- derived growth factor and tumour necrosis factor-α), whereas others (e.g TGF-β, heparin and corticosteroids), have an inhibitory effect on mRNA synthesis (Fig. 1). MMP-2 and MMP-9 are highly expressed in glomeruli [11], but they are differentially regulated at the level of gene transcription due to fundamental differences in their gene promoter regions [12].Fig. 1: Schematic showing the regulation of matrix metalloproteinases (MMPs) at three levels: gene transcription, activation of latent proenzyme and via tissue inhibitors of metalloproteinases (TIMPs). TGF, transforming growth factor; IL, interleukin; TNF, tumour necrosis factor; ECM, extracellular matrix.There are also important interactions with the coagulation cascade. For example, thrombin is a multifunctional serine protease with biological effects that extend beyond haemostasis, including stimulation of type IV collagen synthesis and TIMP-1 expression [13]. In addition, plasmin (released by local tissue plasminogen activators) is a potent activator of most MMPs and promotes cleavage of the latent propeptides to the active molecule. Several studies performed in vivo and in vitro have demonstrated the effects of different cytokines and growth factors on the accumulation of ECM via alterations in the balance between synthesis and degradation [13,14]. For example, in renal interstitial fibrosis there is often no documented change in mRNA levels for the MMPs but, instead, significantly increased expression of the inducible inhibitor, TIMP-1 [15]. Tissue inhibitors of metalloproteinases TIMPS are a family of specific inhibitors of MMPs which are essential for the regulation of normal connective tissue metabolism [7]. TIMP-1 is synthesized by most connective tissue cell types, including mesangial cells and macrophages, and has broad spectrum inhibitory activity against most MMPs (except the MT-MMPs). In situ hybridization studies have shown that tubular epithelial cells and interstitial cells produce TIMP-1 in the renal tubulointerstitium [16]. In actively resorbing tissues, TIMP-1 is highly expressed and forms high-affinity, irreversible complexes with the active MMP enzymes. Thus, the net level of proteinase activity is dependent upon the relative concentrations of active MMPs and TIMPs. TIMP-1 mRNA levels are increased in association with renal interstitial fibrosis (e.g in diabetic nephropathy [15] and polycystic kidney disease [17]), but TIMP-1 deficient mice are not protected against interstitial fibrosis [18]. TIMP-2 has only 42% amino acid homology with TIMP-1 but a similar profile of MMP inhibitory activity. Whereas TIMP-1 is highly inducible by cytokines and growth factors, TIMP-2 expression is largely constitutive and closely matches the pattern of expression of MMP-2 [19]. Expression of TIMP-1 and TIMP-2 is relatively low in normal glomeruli and mesangial cells [11], but expression is increased significantly in patients with glomerulosclerosis [20]. TIMP-3 shares only 37% sequence homology with TIMP-1 and is localized mainly to the ECM, unlike other TIMPs [21]. TIMP-4 is the most abundant TIMP in the heart [22], and imbalances between MMP and TIMP in cardiac tissues have important consequences for long-term remodelling processes such as infarction, heart failure and cardiomyopathy [19]. Independent of their antiproteolytic properties, TIMPs exert a number of other biological effects in connective tissues, including growth factor activity, inhibition of apoptosis, inhibition of angiogenesis, and changes in cell morphology [23]. Disordered connective tissue turnover has many possible causes but mutations in the TIMP-3 gene illustrate how subtle imbalances in the MMP-mediated matrix remodelling process can have clinically significant effects on connective tissue structure and function [24]. Hypertension-induced disturbances of MMP and TIMP activities Several studies in rodent models have shown that hypertension is associated with ECM accumulation in the heart, major vessels and kidney [25], although the extent to which this involves increased collagen synthesis and/or decreased collagen degradation is not entirely clear. There is evidence that mechanical stretch increases matrix protein synthesis in the vasculature and kidney [26,27], but shear stress and mechanical strain also upregulate MMP synthesis, especially MMP-2 and MMP-9 in vascular smooth muscle cells and MMP-9 in endothelial cells [28–31]. This mechanical effect on MMP transcription and enzyme activation may at least partly depend upon TGF-β and c-Myc [28,30], and may be positively regulated by reactive oxygen species [31]. By contrast, expression of TIMP-1 and TIMP-2 mRNA in vascular smooth muscle cells is not affected by mechanical stretch [31]. Clinical studies have shown that serum concentrations of the interstitial collagenase MMP-1 are decreased whereas levels of TIMP-1 are increased in patients with hypertension [32] (Fig. 2), and it has been suggested that elevated TIMP-1 might be a useful non-invasive marker of left ventricular diastolic dysfunction and fibrosis [33]. Serum TIMP-1 concentrations > 500 ng/ml showed good specificity and positive predictive value for detecting diastolic dysfunction among untreated patients with hypertension [33]. By contrast, serum concentrations of the gelatinases, MMP-2 and MMP-9, have been reported as decreased or unchanged in patients with hypertension compared to normotensive controls [34,35]. Clinical studies have also related changes in circulating MMP and TIMP levels to complications associated with hypertension, including microalbuminuria [36], vascular dementia [37] and a prothrombotic state [38].Fig. 2: (a) Serum concentrations of free matrix metalloproteinases (MMP)-1 (and mean value) in 23 normotensive (NT) and 37 essential hypertensive (HT) subjects (difference P < 0.001) and (b) corresponding measurements of serum free tissue inhibitor of metalloproteinase (TIMP)-1 levels in the same subjects (difference P < 0.001). Reproduced with permission [32].Hypertensive renal disease Even though hypertension is associated with renal fibrogenesis, MMP activity is increased in the kidneys of spontaneously hypertensive rats (SHRs) compared to normotensive Wistar controls [39], but very little information is available about differences in the balance of MMP and TIMP activities in the kidney during the development of primary hypertension. In this issue of the journal, Camp et al. [5] compared MMP and TIMP activities in the cortex and medulla of SHRs at 2 weeks and 6 weeks of age, relative to normotensive Wistar rats. Specifically, the investigators measured activity of the two gelatinases, MMP-2 and MMP-9, and MMP-7, and quantified TIMP-4 protein by immunoblotting. The development of hypertension in 6-week-old SHRs was associated with a fibrogenic response and collagen deposition in glomeruli and the tubular interstitium. In the renal cortex of SHRs, gelatinase activity increased after 6 weeks with no significant change in TIMP-4, and there was evidence of significant elastinolytic activity. By contrast, in the renal medulla, activity of MMP-9 and MMP-7 increased in association with increased TIMP-4. In situ labelling indicated very little TIMP-4 in glomeruli but abundant TIMP-4 in the epithelial layer of tubules. The interpretation of these results is not entirely straight-forward, but the authors speculate that, in the glomeruli of SHRs, activation of MMP-2 and MMP-9 results in relatively greater elastin degradation (unopposed by TIMP-4) compared to collagen degradation. In turn, this change in the elastin–collagen ratio might increase glomerular stiffness and create a functional disturbance that resembles increased cardiac stiffness and diastolic dysfunction. Unfortunately, the experiments are compromised by not measuring the activity of other TIMPs and MMPs in the cortex and medulla, and by not examining individual cell types and tissue compartments. The effects of other MMPs and TIMPs, especially TIMP-1, may be underestimated. Nevertheless, these data provide several important conclusions: (i) imbalances in the intra-renal MMP–TIMP axis occur early in the course of primary hypertension and are therefore likely to be important in the pathogenesis of structural and functional changes induced by high blood pressure; (ii) TIMP-4 is expressed mainly in the tubular epithelium, rather than the glomerulus, which resembles the pattern of distribution reported for TIMP-1 [16]; and (iii) differential collagenolytic and elastinolytic activities within the glomerulus may affect capillary structure and function, giving rise to the renal equivalent of diastolic dysfunction (relative stiffness with delayed relaxation). Much has been reported about the effects of MMPs in the heart, but further work is required to explore this complex enzyme system in hypertensive renal injury.