Role of selectins in glomerulonephritis

INTRODUCTION

Selectins are an important family of adhesion molecules involved in capturing circulating leucocytes by inducing ‘rolling’ along the endothelium. Besides this, selectins have other important functions, such as a procoagulant effect, which is also important in inflammation. E-selectin (CD62E) can be synthesized and expressed by endothelial cells following exposure to proinflammatory stimuli (e.g. IL-1, TNF-α). P-selectin (CD62P) can respond quickly to proinflammatory stimuli because it is normally stored in α-granules of platelets and Weibel–Palade bodies of endothelial cells. Following stimulation with proinflammatory stimuli, P-selectin can translocate to the cell surface rapidly to interact with its ligands on leucocytes (Fig. 1). l-selectin (CD62L) is present on the surface of leucocytes and interacts with its ligands on endothelial cells. This early interaction between leucocytes and activated endothelium is important for subsequent migration of leucocytes to sites of inflammation. However, selectins may have more complex properties. Soluble selectins can be produced by shedding or alternatively splicing (Fig. 1) [1,2]. These soluble selectins can bind to their corresponding ligands and limit further interaction of cell associated selectins with their target cells [3,4]. These interesting, but complex, properties of selectins have stimulated a great deal of research into the controversial role of selectins in glomerulonephritis.

Fig. 1.

Complexity of P-selectin functions. Membrane associated form of P-selectin binds to its ligand when leucocytes roll along the endothelial surface. Soluble P-selectin is produced by alternatively splicing or shedding from cell surfaces. Soluble P-selectin may function as an antagonist by competing with membrane associated P-selectin.

SELECTINS IN HUMAN GLOMERULONEPHRITIS

Leucocyte infiltration is a prominent feature of many types of glomerulonephritis. It would be logical to expect that adhesion molecules and chemokines are important in the pathogenesis of glomerulonephritis. However, with the probable overlapping roles of different adhesion molecules, it is important to ask whether blockade of selectins is sufficient in abrogating the cascade of events resulting in leucocyte infiltration and thrombosis in glomerulonephritis.

Expression of E-selectin has been detected in the endothelial side of capillaries of glomeruli and peritubular capillaries of the interstitium in patients with IgA nephropathy and lupus nephritis [5]. However, in a study of non-crescentic acute post-streptococcal glomerulonephritis, there was no increase in glomerular expression of E-selectin [6]. In membranous glomerulonephritis, there was increased expression of E-selectin in peritubular capillaries [7]. Expression of P-selectin has been detected in the glomeruli and interstitium in proliferative glomerulonephritis (rapidly progressive glomerulonephritis, IgA nephropathy, lupus nephritis) but not in minimal change glomerulonephritis, focal glomerulosclerosis or membranous nephropathy [5,8]. Confocal microscopy demonstrated intense expression of P-selectin on the luminal side of glomerular capillaries, the endothelial cells of postcapillary venules in the interstitium and platelets. However, the role of P-selectin in pathogenesis is still not clear. Although the expression of P-selectin correlates with leucocyte infiltration and cellular proliferation, it does not correlate with the clinical severity of the disease (proteinuria, haematuria and serum creatinine). Soluble P-selectin was higher in the serum of patients with glomerular or interstitial expression of P-selectin, but could not be detected in the urine of any of these patients. In patients with IgA nephropathy, circulating T cells and B cells expressed a higher amount of l-selectin [9]. However, the relationship between this observation and pathogenesis of the disease remains speculative. Interestingly, expression of E-, P- and l-selectin in renal interstitium was found to correlate with the degree of chronic interstitial injury in glomerulonephritis [6].

Besides glomerulonephritis, it is interesting that higher levels of E- and P-selectin were detected in the glomeruli and interstitium of patients with diabetic nephropathy [5]. E-selectin expression on peritubular capillaries correlated with the macrophage infiltration of interstitium.

SELECTINS IN EXPERIMENTAL GLOMERULONEPHRITIS

Most of the work on the role of selections in experimental glomerulonephritis has been performed on well-characterized rodent models of nephrotoxic nephritis, which are induced by injection of anti-glomerular basement membrane antiserum. P-selectin, but not E- or l-selectin, was shown to be important in the pathogenesis of glomerular inflammation in some of these studies.

In the heterologous phase of nephrotoxic nephritis in mice, glomerular expression of E-selectin can be detected as early as 2h following injection of anti-GBM antibodies [10]. Pretreatment with anti-TNF-α antibodies prevented the expression of E-selectin. However, the importance of E-selectin remained doubtful, because intravenous injection of anti-E-selectin monoclonal antibody (MoAb) had no effect on the neutrophil infiltration or proteinuria.

In a series of studies using antibodies to P-selectin, the results supported the idea that glomerular endothelial expression of P-selectin is important in infiltration of neutrophils and monocytes/macrophages during the development of glomerulonephritis. In the heterologous phase (complement independent) of nephrotoxic nephritis in mice, P-selectin expression was up-regulated in glomerular endothelium. Platelet depletion had no effect on glomerular P-selectin expression. Pretreatment with anti-P-selectin antibodies reduced neutrophil infiltration and injury [11]. Using a rat model of accelerated nephrotoxic nephritis, it was shown that P-selectin was important in macrophage infiltration of the glomeruli and subsequent injury [12]. However, work with P-selectin knockout mice has given very surprising results. P-selectin knockout mice had higher neutrophil infiltration and more severe glomerulonephritis than the wild-type mice following induction of nephritis [13]. This raised the possibility that P-selectin may have anti-inflammatory functions in some situations. This may be achieved by (1) production of anti-inflammatory molecules, e.g. lipoxin A₄ (LXA₄) and (2) production of soluble selectin molecules.

For example, LXA₄ inhibits neutrophil functions, such as chemotaxis, adhesion to endothelial cells and transmigration across endothelium. P-selectin-dependent interactions between neutrophils and platelets are an important source of LXA₄ (reviewed in [14]). The renal LXA₄ level was lower in the P-selectin knockout mice than the wild-type mice following induction of the nephritis [13]. The exact importance of LXA₄ remains inconclusive, because platelet infusion into P-selectin knockout mice restored the amount of LXA₄, but only partly corrected the differences in neutrophil infiltration and had no effect on proteinuria [13].

In an accelerated model of nephrotoxic nephritis, P-selectin knockout mice had an increased amount of C3 deposition, platelet accumulation and fibrin deposition in the glomeruli, and more albuminuria than the wild-type mice [15]. Soluble P-selectin was detected in the serum of wild-type mice, but not in the P-selectin knockout mice, following induction of nephritis. This raised the possibility that soluble P-selectin may have anti-inflammatory effects. The relative contribution of endothelial P-selectin and platelet derived P-selectin was examined in bone marrow reconstituted mice. The Pt+/EC− chimeras showed a similar severity of nephrotoxic nephritis as the knockout mice (Pt-/EC-). Chimeras with endothelial cells expressing P-selectin (Pt−/EC+) had a similar concentration of soluble P-selectin in plasma as wild-type mice (Pt+/EC+), and had less glomerular injury than the knockout mice (Pt−/EC−). Further work suggested that the soluble P-selectin was produced by shedding from extra-renal endothelium. In fact, earlier work on characterization of P-selectin showed that soluble forms of P-selectin could be detected in the plasma. This soluble P-selectin may be produced by alternatively splicing or shedding. Soluble P-selectin has been shown to bind to cell surface ligands and have anti-inflammatory properties in vitro[4].

However, it is not clear what regulates the production of soluble P-selectin (alternatively splicing and shedding) [2], and the contribution of soluble P-selectin in vivo remains to be investigated.

In a murine model of lupus nephritis, injection of radiolabelled MoAb to E-selectin was used to detect low level of expression of E-selectin (but not P-selectin) in the kidneys from older MRL/lpr mice, although the level was below the detection of conventional immunohistology [16]. In the ConA model of immune complex nephritis, injection of MoAb to P-selectin was shown to reduce neutrophil and platelet accumulation in the glomeruli, but there was only a mild reduction in proteinuria [17].

NON-ANTIBODY BLOCKADE OF SELECTINS

Specific antibodies have been useful in defining the role of individual selectins. With better characterization of ligands for selectins, soluble selectin ligands have been used in blockade of selectins. Some of these soluble ligands have the advantage of blockade of multiple selectins with a single inhibitor. They also have less antigenicity and lower cost of production.

Attempts to use soluble inhibitors have had variable success. Fucoidean F7 is an oligosaccaride that blocks both P- and l-selectin. In the heterologous phase of nephrotoxic nephritis in Wistar rat, pretreatment with Fucioidean F7 reduced leucocyte rolling in venules, but had no effect on glomerular injury and proteinuria [18]. In an LPS-enhanced model of heterologous phase of nephrotoxic nephritis in rats, injection of a MoAb to P-selectin or a synthetic selectin inhibitor, a sulphatide derivative of sLex which blocks P- and l-selectin, reduced accumulation of neutrophils and platelets and reduced thrombosis in the glomeruli [19]. In this issue of the journal, Ogawa and colleagues used a synthetic sulphated polysaccharide to reduce macrophage infiltration and crescentic glomerulonephritis in nephrotoxic nephritis in WKY rats [20]. Their inhibitor can bind to both l-selectin and P-selectin. They demonstrated that P-selectin, but not l-selectin, was up-regulated in glomerular endothelium following induction of nephritis. Using specific MoAb to P- and l-selection, they found convincing evidence that their synthetic inhibitor abrogates crescentic glomerulonephritis through blockade of P-selectin. Although this well-conducted study seems to challenge the results of knockout mice studies, it is explained better by the fact that P-selectin has both proinflammatory and anti-inflammatory properties. The result of individual studies depends probably on the balance of the pro- and anti-inflammatory roles of P-selectin in each situation. Nevertheless, there is no need to be despondent about this complexity. With careful characterization of the role of P-selectin in vivo, it should still be a useful target for clinical intervention.

Future work should focus on understanding the wide range of functions of P-selectin in vivo. Understanding of the factors that regulate production of soluble selectins will provide another approach by which to regulate the balance of the inflammatory process. This knowledge may provide important therapeutic advances to the management of glomerulonephritis and other renal diseases, such as diabetic nephropathy.