CsA is a cyclic endecapeptide which is derived from extracts of Tolypocladium inflatium Gams , a member of the Fungi imperfekti family (Figure 1) (Kahan 1999).
''The high treeless plain where this fungus was discovered, Hardarger Vidda, is a land populated by malevolent spirits. One man, Askelad, a Nordic folk hero, used his ingenuity and perseverance to subvert the spiritsí schemes and preserve the peace and harmony of the community'' (Kahan 1999).
Since the early 1980ís, CsA has been widely used to prevent rejection of organ transplants. It has increased graft survival and decreased patient mortality (Kahan et al. 1987; 1999). The use of CsA to treat autoimmune diseases has increased during recent years (Bach 1999). In Finland, indications of CsA include rheumatoid arthritis, psoriasis, atopic dermatitis, endogenous uveitis and nephrotic syndrome. The dose of CsA in autoimmune diseases is usually lower than that used in organ transplantations (Bach 1999). Unlike in organ transplantations, in autoimmune diseases the use of CsA is often temporary.
CsA was the first T-lymphocyte selective drug to inhibit lymphocyte function without concomitant myelosuppression. CsA binds to cyclophilin to inhibit phosphatase calcineurin, and thereby prevents the dephosphorylation of nuclear factor of activated T-cells that is essential for upregulation of messenger ribonucleic acid (mRNA) (Figure 2) (Liu et al. 1992). Thus, CsA interferes with the synthesis of many lymphokine mediators, particularly interleukin-2, which is critical for T-lymphocyte proliferation and maturition, and interferon- g , which is critical for macrophage activation. It has recently been proposed that increased production of TGF- b 1 is at least partly responsible for the immunosuppressive effects of CsA (Hojo et al. 1999; Nabel 1999). The mechanism for the induction of TGF- b 1 by CsA remains obscure (Nabel 1999).
The extent to which the adverse effects of CsA are related to the immunosuppressive mechanisms of the drug is controversial. Inhibition of calcineurin by CsA may alter the production of many biologically active agents, such as endothelin, NO and TGF- b 1 , which have been implicated in adverse renal and cardiovascular effects (Curtis 1998; Hemenway and Heitman 1999). On the other hand, calcineurin inhibition may protect against the development of left ventricular hypertrophy (Sussman et al. 1998).
Figure 2. Main mechanism for the immunosuppressive action of cyclosporine. CsA, cyclosporine; NF-AT, nuclear factor of activated T-cells; p, phosphate group. Adapted from Nabel (1999).
The most important adverse effects that restrict the use of CsA are nephrotoxicity and hypertension (Mason 1989; Taler et al. 1999). These effects have been detected in transplant recipients and in autoimmune patients. Other factors besides CsA may, however, also contribute to these abnormalities, and thus, determining the effects of CsA alone is difficult. In transplant patients, these confounding factors include rejection of the transplanted organ and concomitant glucocorticoid and other immunosuppressive treatments. In some autoimmune patients, for example those suffering from rheumatoid arthritis, the disease itself may cause renal and cardiovascular sequalae (Bach 1999).
A single dose of CsA causes a transient rise in blood pressure in healthy volunteers (Hansen et al. 1997). Long-term CsA treatment in organ transplant recipients increases the risk for hypertension compared with the therapy with azathioprine and glucocorticoids (Kahan et al. 1987; Schorn et al. 1988; Taler et al. 1999). CsA also induces hypertension in autoimmune patients (Bach 1999). While the incidence of hypertension is 29 - 54% in CsA-treated autoimmune patients, in transplant patients the corresponding figure is 71 - 100% (Taler et al. 1999). Hypertension is an independent risk factor for atherosclerosis in CsA-treated transplant patients (Kasiske 1988) . Furthermore, hypertension after transplantation is related with lower rates of graft survival (van Ypersele de Strihou et al. 1983). Hypertension is usually reversible after discontinuation of short-term CsA therapy (Taler et al. 1999), whereas continued treatment even at reduced doses frequently results in sustained hypertension (Schwartz et al. 1996).
Adverse renal effects of CsA can be divided into two categories: functional and histological abnormalities. CsA-induced renal dysfunction is related to reduced glomerular filtration rate and renal blood flow. This is generally thought to be secondary to vasoconstriction of the glomerular afferent arterioles, which causes a decrease in glomerular pressure (Murray et al. 1985; Myers et al. 1988; Hansen et al. 1997; Andoh and Bennett 1998). Renal dysfunction, in turn, causes an increase in serum creatinine concentration and a decrease in creatinine clearance. These functional changes are dose-dependent and are usually reversible after short-term CsA treatment (Andoh and Bennett 1998).
CsA-induced histological renal damage is mainly characterised by tubulointerstitial fibrosis of striped pattern and arteriolopathy of the afferent arterioles (Antonovych et al. 1988; Mihatsch et al . 1988; Andoh and Bennett 1998). Interstitial fibrosis consists of streaks of fibrosis. Arteriolopathy is characterised by luminal narrowing and arteriolar wall thickening. These changes are composed of subendothelial edema, swelling of hypertrophied endothelial cells protruding into the lumen and eosinophilic granular transformation of vascular smooth muscle cells. However, other types of renal lesions, including glomerular and tubular damage, have also been linked with CsA treatment (Antonovych et al. 1988; Andoh and Bennett 1998). The histological changes develop mainly during long-term use and are usually irreversible once established even if CsA treatment is ceased, but the risk of histological changes is diminished with reducing the dose of CsA (Andoh and Bennett 1998).
As a consequence of the resistance of rats to adverse renal and cardiovascular effects of CsA, there was very little if any preclinical information on them in early clinical use (for review, see Curtis 1994). Therefore, the adverse renal and cardiovascular effects of CsA were somewhat unexpected. A great deal of effort was put into developing suitable animal models to study the mechanisms of these CsA-induced abnormalities.
The adverse renal effects of CsA were slight or absent in normotensive rats on normal-sodium diet (Elzinga et al. 1993; Porter et al. 1999). An exception to this was that CsA in a year-long study induced tubulointerstitial fibrosis (Lafayette et al. 1993). To study the nephrotoxicity of CsA, a rat model using sodium-depleted normotensive Sprague Dawley rats was developed (Burdmann et al. 1995; Pichler et al. 1995; Young et al. 1995b; Porter et al. 1999). In this rat model, a fairly high dose of CsA (15 mg/kg/day s.c.) for four weeks caused morphological and functional renal toxicity similar to that seen in CsA-treated patients.
According to several studies, CsA does not induce hypertension in normotensive rats even if high doses of up to 50 mg/kg/day are used (Nahman et al. 1988; Diederich et al. 1992; Lafayette et al. 1993; Basu et al. 1994; Stephan et al . 1995). A moderate rise in blood pressure by CsA at 25 mg/kg/day has, however, been reported in some studies of normotensive rats (Takeda et al. 1995; Oriji and Keiser 1998). The SHR model is a well-known genetic hypertension model that has many similarities to human essential hypertension (Okamoto et al. 1964; Phillips 1999). A few previous studies of CsA with SHR have been reported. Interestingly, CsA at a dose of 10 mg/kg/day delayed the onset of hypertension in two-week-old SHR (Sitsen and de Jong 1987). In SHR in the developmental phase of hypertension, CsA at doses of 20 - 100 mg/kg/day induced elevation of blood pressure and renal vascular lesions (Ryffel et al. 1986; Nahman et al. 1988). At 5 mg/kg/day, CsA caused a rise in blood pressure on normal-sodium diet (Basu et al. 1994; Mervaala et al. 1997). On high-sodium diet in SHR, CsA-induced hypertension manifests with concomitant renal dysfunction and glomerular damage (Mervaala et al. 1997; Pere et al. 1998).
The renin-angiotensin system (RAS) is an important regulator of blood pressure and renal function. The most important effector of RAS is octapeptide angiotensin II (Figure 3). It is formed in an enzyme cascade initiated by renin, which is secreted from juxtaglomerular cells of the kidney. While RAS is a circulatory hormone system, all of its components also exist locally in tissues, e.g. in the kidney, blood vessel wall, heart, brain and adrenal cortex (for review, see Ganten et al. 1983; Kim and Iwao 2000). Therefore, RAS seems to be a regional regulator as well.
The main stimulators of renin release are decreased sodium concentration in the distal tubule, decreased renal perfusion pressure and adrenergic b 1 -receptor stimulation. Angiotensin II, potassium, antidiuretic hormone and atrial natriuretic peptide reduce renin release. Renin cleaves angiotensinogen, which is derived from the liver, into inactive decapeptide angiotensin I (Kim and Iwao 2000).
There are different enzymatic routes of angiotensin II formation from angiotensin I (Figure 3). Angiotensin I is activated by angiotensin converting enzyme (ACE), which is mainly located on the surface of the vascular endothelium and the lung epithelium. Other enzymes, including chymase, which is present in the heart, can also convert angiotensin I to angiotensin II. The importance of these non-ACE-pathways is still controversial. In humans, the non-ACE pathways are reported to represent at least 40% of total angiotensin II formation (Hollenberg et al. 1998). Some evidence exists that ACE-independent pathways become quantitatively more important under pathological conditions such as diabetes (Hollenberg et al. 1998). In rats, however, ACE seems to be the most important enzyme for angiotensin II formation (Okunishi et al. 1993).
Figure 3. Renin-angiotensin system and kallikrein-kinin system. ACE, angiotensin converting enzyme; CAGE, chymostatin-sensitive angiotensin II-generating enzyme; NO, nitric oxide; t-PA, tissue plasminogen activator.
At least two types of angiotensin receptors are present on the surface of the target cells: Angiotensin II type 1 (AT 1 ) and type 2 (AT 2 ) receptors (for review, see Stroth and Unger 1999). AT 1 receptors can be further sudivided into subtype A (AT 1A ) and subtype B (AT 1B ) receptors . AT 1A receptor is the major subtype mediating cardiovascular and renal effects of angiotensin II (Llorens-Cortes et al. 1994). AT 1 receptors are G-protein-coupled receptors with several second-messenger systems, including activation of phospholipase C. This results in activation of the phosphoinositol pathway, which leads to increased intracellular calcium levels and protein kinase C activation. These, in turn, activate growth-related immediate transcription factors, such as c-fos, c-myc and c-jun. Furthermore, inhibition of adenylate cyclase by angiotensin II reduces cyclic adenosine 3', 5'-monophosphate (cAMP) levels. AT 2 receptors are also coupled to G-proteins, but the signal transduction pathways are not fully known (Stroth and Unger 1999).
Most of the known cardiovascular effects of angiotensin II are mediated via AT 1 receptors. These include contraction of vascular smooth muscle, enhancement of calcium sensitivity of the contractile apparatus of the smooth muscle, increased endothelin production, facilitation of noradrenaline biosynthesis and release and inhibition of its reuptake in sympathetic nerve terminals, stimulation of catecholamine release from the adrenal medulla, positive inotropic and chronotropic action on the heart and facilitation of aldosterone biosynthesis and secretion in the adrenal cortex. Furthermore, AT 1 receptor stimulation augments tubular sodium reabsorption in the kidney, inhibits renin release, induces polydipsia and releases vasopressin from the pituitary gland (for review, see Stroth and Unger 1999; Kim and Iwao 2000).
Growing evidence shows that AT 2 receptors are also important in controlling the cardiovascular system (for review, see Stroth and Unger 1999). It seems that most of the effects mediated by AT 2 receptors are the opposite of those mediated by AT 1 . AT 2 receptor knock-out mice develop an increase in blood pressure and an increased sensitivity to the pressor actions of angiotensin II (Ichiki et al. 1995). The stimulation of AT 2 receptors is proposed to antagonise the AT 1 -mediated blood pressure increase (Scheuer and Perrone 1993), to cause vasodilation and an inhibition of angiogenesis (Munzenmaier and Greene 1996) and to increase endothelial NO production (Seyedi et al. 1995). In vitro evidence shows that stimulation of AT 2 receptors may cause apoptosis of cells (Yamada et al. 1996). However , it seems that both receptor subtypes can mediate vascular apoptosis in rats in vivo (Diep et al. 1999).
Even though RAS has an important role in regulating blood pressure, its role in hypertension is unclear. In some renovascular forms of hypertension, activation of RAS is the major reason for the development of hypertension. Otherwise, increased activity of RAS, measured as plasma renin activity (PRA) elevation or as increased plasma angiotensin II concentrations, is an irregular finding in SHR or in human essential hypertension (Paran et al. 1995). However, antagonism of RAS by ACE inhibitors and AT 1 receptor antagonists is effective in reducing blood pressure in SHR and in hypertensive patients (Timmermans and Smith 1996; Gavras 1997).
Moreover, angiotensin II may play a role in the development of renal and vascular damage independently of blood pressure. Via AT 1 receptors, angiotensin II can cause cell growth, differentiation and proliferation directly by affecting several kinase pathways and indirectly by inducing several growth factors, including TGF- b 1 and platelet-derived growth factor (for review, see Stroth and Unger 1999).
The kallikrein-kinin system plays a role in controlling blood pressure, renal haemodynamics and excretory functions (for review, see Majima and Katori 1995). Tissue and plasma kallikreins generate kinins from kininogens by hydrolysis (Figure 3). Kinins are oligopeptides that contain the bradykinin sequence in their structure. There are two types of bradykinin receptors, B 1 and B 2 receptors. G-protein-coupled B 2 receptors mediate most of the known effects of bradykinin. The activation of B 2 receptors causes phospholipase C activation and thus an increase in intracellular calcium concentration which stimulates the production of NO and vasoactive prostanoids such as prostaglandin E 2 (PGE 2 ) and PGI 2 . (Linz et al. 1995). The effects mediated via B 1 receptors are not well known, but recent evidence suggests that activation of these receptors results in vasoconstriction and decreased glomerular filtration rate (Schanstra et al . 2000). The expression of B 1 receptors is reported to be induced during inflammation in rats (Schanstra et al . 2000) and in renal disease in patients (Naicker et al. 1999). Kinins are destroyed by kininases. The two main kininases are kininase II, which has been found to be identical with ACE, and neutral endopeptidase.
A decreased activity of the renal kallikrein-kinin system has been found in essential hypertension and in rat models of genetic hypertension (for review, see Majima and Katori 1995). Thus, the renal kallikrein-kinin system may have a crucial role in the development of hypertension.
The cardiovascular effects of the kallikrein-kinin system have been studied, e.g., by using B 2 receptor antagonist icatibant (HOE 140). Inhibition of bradykinin breakdown by ACE inhibitors leads to increased B 2 receptor activation. Icatibant has been reported to partly antagonise the antihypertensive effects of ACE inhibitors in experimental animals (Bao et al. 1992; O'Sullivan and Harrap 1995). Therefore, increased availability of bradykinin seems to be responsible for some of the beneficial effects of ACE inhibitors.
Several lines of evidence suggest an involvement of RAS in CsA toxicity. In rats, activation of RAS by CsA is a consistent finding. An increase in PRA has been demonstrated in CsA-treated rats on sodium-depletion (Burdmann et al. 1995), normal-sodium (Abassi et al. 1996) or high-sodium diets (Mervaala et al. 1997). Increased PRA by CsA has been related to increased renin expression in the juxtaglomerular apparatus (Young et al. 1995a). In vitro , CsA stimulated renin production and release from juxtaglomerular cells (Kurtz et al. 1988). The mechanism for this direct effect is obscure, but it is proposed to be linked to cyclophilin binding (Kurtz et al. 1988; Lee 1997). Furthermore, CsA increased local expression of angiotensin II in the outer medulla and medullary rays (Ramirez et al. 2000).
CsA has been reported to elevate PRA in cardiac and liver transplant patients (Julien et al . 1993) and plasma angiotensin II levels in patients with psoriasis (Edwards et al. 1994). Data of the effects of CsA on RAS in man are, however, to some extent contradictory; normal or even low PRA have been reported (Bantle et al. 1987; Myers et al. 1988; Mason et al. 1991). There is evidence of CsA-induced juxtaglomerular hyperplasia related to increased total renin content and impaired intrarenal conversion of inactive prorenin to active renin (Bantle et al. 1987; Myers et al. 1988; Mason et al. 1991). CsA may raise angiotensin II levels by increasing the activity of ACE which has been detected in serum of patients (Letizia et al . 1995) as well as in serum and in lung tissue of rats (Erman et al . 1990). Therefore, the lack of increase in PRA in some clinical studies does not exclude activation of RAS. It has been suggested that CsA causes activation of tissue RAS, which is not always seen as a change in PRA (Mason et al. 1991; Young et al. 1995a; Ramirez et al. 2000).
Chronic administration of angiotensin II in rats produces renal injury resembling that observed in CsA nephropathy in man (Giachelli et al. 1994; Noble and Border 1997). Both angiotensin II and CsA cause an overexpression of TGF- b 1 , a growth factor which has been implicated in the pathophysiology of many fibrotic diseases of the kidney and other organs (Pichler et al. 1995; Noble and Border 1997). In addition, both angiotensin II (Giachelli et al. 1994) and CsA (Pichler et al. 1995) increase tubulointerstitial expression of osteopontin, an adhesion molecule associated with renal fibrosis. Thus, activation of RAS by CsA seems to be at least partly responsible for interstitial fibrosis through stimulation of TGF- b 1 and osteopontin expression.
Antagonism of RAS by ACE inhibitors and AT 1 receptor antagonists lowers blood pressure in CsA patients (Mourad et al . 1993; Sennesael et al . 1995; Hannedouche et al . 1996; Del Castillo et al . 1998; Taler et al . 1999). Clinical studies suggest that the CsA-induced hypertension can be treated with several classes of antihypertensive drugs including ACE inhibitors, AT 1 receptor antagonists, calcium antagonists and adrenergic b -receptor antagonists. However, some studies suggest that ACE inhibitors alone have limited antihypertensive effects directly after transplantation (Textor et al. 1988; Taler et al . 1999).
In normotensive rats during sodium deprivation, antagonism of RAS with enalapril and losartan prevented CsA-induced morphological nephrotoxicity without affecting blood pressure (Burdman et al. 1995; Pichler et al. 1995). In the same animal model, arteriolopathy and interstitial fibrosis were diminished by RAS inhibition with enalapril and losartan, but not by some other classes of antihypertensive drugs such as calcium antagonist nilvadipine, vasodilatator dihydralazine or diuretic furosemide (Shihab et al . 1997), while blood pressure remained unaffected by any of the treatments. AT 1 receptor antagonist losartan reduced CsA-induced interstitial fibrosis in sodium-deprived rats concomitantly with decreasing renal TGF- b 1 and osteopontin expression (Pichler et al. 1995). Additionally, in a one-year study of normotensive rats on normal-sodium diet, CsA caused tubulointerstitial fibrosis without affecting blood pressure (Lafayette et al. 1993). Both concomitant enalapril and a combination of potassium channel opener minoxidil, monoamine transporter inhibitor reserpine and diuretic hydrochlorothiazide lowered blood pressure, but only enalapril prevented morphological damage of the kidneys. These studies suggest that angiotensin II has a specific role in the development of CsA-induced morphological renal damage. The effect of angiotensin II seems to be, at least partly, independent of blood pressure changes.
The effects of drugs suppressing RAS on CsA-induced renal dysfunction are unclear. Theoretically, it is possible that suppression of RAS might be harmful. CsA reduces the glomerular filtration rate and renal blood flow by causing vasoconstriction of the glomerular afferent arterioles (Murray et al. 1985; Myers et al. 1988). When renal blood flow is substantially reduced, angiotensin II contributes to the maintenance of the glomerular filtration rate by constricting the efferent glomerular arterioles (Hall et al. 1999). Under such conditions, dilation of the efferent arterioles by drugs suppressing RAS may potentiate the reduction in the glomerular filtration rate.
In sodium-depleted normotensive rats, enalapril and losartan prevented CsA-induced renal morphological changes but not the decrease in glomerular filtration rate (Burdmann et al. 1995; Kon et al. 1995; Pichler et al. 1995). In this rat model, it seems that renal dysfunction is related to activation of the endothelin system, whereas morphological damage is related to activation of RAS (Burdmann et al. 1995; Kon et al. 1995; Pichler et al. 1995). There are case reports in which ACE inhibitors have worsened the renal function of CsA-treated kidney transplant patients (Garcia et al. 1994). In most human studies, however, ACE inhibitors have lowered blood pressure, but have not affected renal function in CsA-treated renal-allograft patients (Sennesael et al. 1995; Hausberg et al. 1999). Furthermore, enalapril prevented the CsA-induced decline in glomerular filtration rate in diabetic patients (Hannedouche et al. 1996). Likewise, in CsA-treated normotensive rats on normal-sodium diet, ACE inhibitor captopril and AT 1 receptor antagonist saralasin increased glomerular filtration rate and renal blood flow (Kaskel et al. 1987). Taken together, it seems that suppression of RAS is not harmful for renal function of all CsA-treated patients and may in fact improve it in some cases. However, a subgroup of patients may exist, for example, kidney transplant patients, who are more susceptible than others to renal dysfunction due to cotreatment of CsA and a drug suppressing the activity of RAS.
CsA may cause cardiovascular effects via RAS by mechanisms which are independent of the rate of angiotensin II formation. Long-term treatment with CsA upregulated AT 1 receptors in rat vascular and renal tissues (Iwai et al. 1993; Regitz-Zagrosek et al. 1995). Similarly, CsA-incubation upregulated AT 1 receptors in human vascular cells and potentiated the rise in intracellular calcium concentration in response to angiotensin II (Avdonin et al. 1999). According to culture studies of rat vascular smooth muscle cells, it seems that CsA increases the intracellular calcium concentration by several mechanisms: CsA stimulated transmembrane calcium influx (Lo Russo et al. 1996; Meyer-Lehnert et al. 1997), and augmented angiotensin II-stimulated calcium influx and efflux (Pfeilschifter and Ruegg 1987). There is evidence that CsA may also mobilise calcium from the intracellular stores by activating the phosphoinositol pathway (Lo Russo et al. 1997). In addition, calcium may be released from intracellular stores by CsA by a mechanism which does not involve activation of the phosphoinositol pathway, but is due to a direct effect on intracellular calcium stores (Gordjani et al . 2000). In accordance with the above-mentioned mechanisms, repeated administration of CsA has been reported to increase the vasoconstrictive effect of angiotensin II (Auch-Schwelk et al. 1993; Takeda et al. 1995).
Very recently, a different view about the interaction of CsA and RAS has been presented. CsA antagonised the harmful effects of RAS overexpression on end-organ damage in double transgenic rats harbouring human renin and angiotensinogen genes (Mervaala et al. 2000). The protection was due to anti-inflammatory properties of CsA by inhibition of interleukin-6 and inducible nitric oxide synthase (iNOS) expression. It seems that CsA may not only cause renal and cardiovascular toxicity via induction of RAS, but can also antagonise inflammatory responses to various stimuli, including angiotensin II.
Only a few reports exist about the role of the renal kallikrein-kinin system in CsA toxicity. In normotensive rats, CsA at 20 mg/kg/day for three days decreased renal kallikrein and bradykinin B 2 receptor levels (Bompart et al . 1996). In another study, CsA-administration at 25 mg/kg/day for 28 days increased renal kallikrein and bradykinin B 2 receptor levels (Wang et al. 1997). This was proposed to occur as compensation for drug-induced vasoconstriction and hypertension. In patients with psoriasis, a reduced urinary kallikrein excretion was detected after six-month treatment with CsA (Raman et al. 1998). No studies of CsA with icatibant have been reported. Thus, the renal kallikrein-kinin system may have a role in CsA toxicity, but requires further studies before any conclusions can be drawn.
The endothelial cell layer participates in the control of vascular tone and blood pressure by producing relaxing and contracting factors, thus modulating the tone of the underlying smooth muscle. Dysfunction of vascular endothelium, characterised by a decrease in endothelium-dependent relaxation, is shown in human essential hypertension and in various models of experimental hypertension (for review, see McIntyre et al. 1999). In SHR, the dysfunction occurs already before the onset of hypertension (Jameson et al. 1993). Impaired endothelial dysfunction is reported in the normotensive offspring of hypertensive patients (Taddei et al. 1996). Endothelial dysfunction can be improved by antihypertensive therapy in rats and in humans (for review, see Vapaatalo et al. 2000).
Nitric oxide (NO) is a lipophilic gas which is formed enzymatically from its precursor L-arginine by three isoforms of NO synthase (NOS), endothelial eNOS, neuronal nNOS and inducible iNOS (for review, see Stuehr 1999). NO diffuses to the vascular smooth muscle cells and activates guanylate cyclase, which leads to the increased production of cyclic guanosine 3',5'-monophosphate (cGMP) (Figure 4). cGMP-dependent kinase regulates several pathways involved in calcium homeostasis. NO also hyperpolarises vascular smooth muscle cells via cGMP-dependent pathway as well as direct effects on calcium-dependent potassium channels and Na + -K + ATPase (for review, see Busse and Fleming 1995). In vivo , NO is destroyed mainly by superoxide anion (for review, see Stuehr 1999). Because of its short half-life (0.1 seconds in vivo ), NO acts primarily as a paracrine hormone (Loscalzo and Welch 1995). Activity of NOS isoforms is regulated in different ways. Constitutive NO production by eNOS is controlled by intracellular calcium concentration following physical (shear stress) or chemical stimuli of various receptors for endogenous substances. iNOS is controlled mainly by inflammatory cytokines such as interleukin-1 and tumor necrosis factor- a (Rao 2000). High amounts of NO are released by iNOS in response to inflammatory stimuli from a variety of cell types (Rao 2000).
Figure 4. L-arginine - nitric oxide system. B 2 , bradykinin B 2 receptor; cGMP, cyclic guanosine 3',5'-monophosphate; eNOS, endothelial nitric oxide synthase; GC, guanylate cyclase; GTP, guanosine 5'-triphosphate; M, muscarinic receptor; NO, nitric oxide; .OH, hydroxyl radical; ONOO - , peroxynitrite.
In addition to NO, prostanoids have a major role in the control of vascular function. Prostanoids, including PGI 2 and PGE 2 , are produced from arachidonic acid, which is liberated from the cell membrane phospholipids by phospholipase A 2 . The effects of PGI 2 , the most important vasodilating prostaglandin, are mediated by membrane receptors on the smooth muscle cells. Stimulation of these receptors activates adenylate cyclase, which leads to increased intracellular concentrations of cAMP, and further to hyperpolarisation of the cell membrane, reduction of intracellular calcium concentration and decreased sensitivity of contractile proteins to calcium (for review, see Cohen and Vanhoutte 1995).
Recent evidence suggests that a factor called endothelium-derived hyperpolarising factor contributes to endothelium-mediated vasodilation by increasing the conductance of smooth muscle cells to potassium ions (for review, see Cohen and Vanhoutte 1995). The release of endothelium-derived hyperpolarising factor is believed to be initiated by an increase in free calcium concentration in the endothelial cell. The chemical structure of the substance or possibly of several different substances is not known. Endothelium-derived hyperpolarisation may, at least partly, account for nonprostanoid products of arachidonic acid metabolism (Cohen and Vanhoutte 1995; Fisslthaler et al. 2000).
The endothelium is also a source of contracting factors. A gene codes for three isoforms of endothelins (1, 2 and 3), which are generated in two steps from preproendothelins. First, neutral endopeptidase forms the intermediate product proendothelins, which are further activated by endothelin converting enzyme (Benigni and Remuzzi 1999). They act via two receptor subtypes (A and B). Other important endothelium-derived contractile factors include cyclooxygenase product thromboxane A 2 and prostaglandin endoperoxide intermediates prostaglandin G 2 and H 2 (Cohen 1995). All of these substances contract smooth muscle cells via thromboxane A 2 receptors (Cohen 1995).
Endothelial cells are both significant sources and targets of reactive oxygen products, including superoxide anion, hydroxyl radical, hydrogen peroxide and peroxynitrite (for review, see McIntyre et al. 1999). These products can inhibit vascular relaxation by neutralising the effects of NO, but they may also have direct vasoconstrictive properties (Cosentino et al. 1994). The main sources of reactive oxygen products are respiratory chain enzymes in the mitochondria, such as aldehyde oxidase, dihydroorotic dehydrogenases, flavin dehydrogenases, peroxidases, cyclooxygenase, NOS, and auto-oxidation of a large group of compounds including catecholamines, flavins and ferredoxin. Many enzymes, including superoxide dismutases, can inactivate reactive oxygen products (McIntyre et al. 1999).
Increased superoxide production has been shown in endothelial cells from stroke-prone SHR (Grunfeld et al . 1995). In addition, aortic rings from SHR are more sensitive to contractions induced by oxygen-derived free radicals than those from Wistar-Kyoto rats (Auch-Schwelk et al . 1989). Increased vascular oxidative stress has also been demonstrated in SHR in vivo (Suzuki et al . 1995; 1998). These studies suggest that genetic hypertension is associated with increased oxidative stress.
Endothelial dysfunction was earlier thought to be attributed mostly to decreased production of NO either due to decreased activity of eNOS or to a deficiency in the availability of L-arginine (Palmer et al. 1988). L-arginine causes vasodilation and lowering of blood pressure in experimental animals and in man (for review, see Li and Forstermann 2000). However, recent evidence suggests that impaired endothelium-dependent relaxation is associated with increased rather than decreased expression of eNOS in vasculature (Boulomie et al. 1997). eNOS can also catalyse superoxide formation in a reaction which is primarily regulated by cofactor tetrahydrobiopterin rather than L-arginine (Vasquez-Vivar et al. 1998). Thus, it seems that for normal endothelial function the balance between NO and superoxide anion is more important than the absolute levels of either alone (McIntyre et al. 1999).
Repeated in vivo administration of fairly high doses of CsA has been found to impair endothelium-dependent relaxation to acetylcholine in the rat aorta (Mikkelsen et al. 1992; Auch-Schwelk et al. 1994; Verbeke et al. 1995; Kim et al. 1996), in femoral arteries (Gallego et al. 1994) and in mesenteric arteries (Rego et al. 1990). In CsA-treated transplant recipients, renal vasodilatory response to L-arginine infusion was reduced with a concomitant lowering of urinary nitrate excretion (Gaston et al. 1995). Likewise, CsA treatment in rheumatoid arthritis patients attenuated the vasodilatory responses to isoprenalin and PGE 2 (Stein et al. 1995). CsA inhibited endothelium-dependent relaxation in vivo in liver transplant recipients (Albillos et al. 1998) and in vitro in subcutaneous resistance vessels obtained from normal subjects undergoing routine surgery (Richards et al. 1990).
Some evidence suggests that endothelial dysfunction is due to direct toxicity of CsA on endothelial cells. CsA-induced dysfunction was associated with concomitant endothelial damage in rats (Kim et al . 1996). In addition, incubation of bovine aortic endothelial cells with CsA at a therapeutically relevant concentration of 1 m M resulted in cell lysis (Zoja et al. 1986).
CsA has been reported to impair endothelium-independent relaxations in rats (Gallego et al . 1993; Kim et al . 1996), but endothelium-independent relaxations have also been reported to remain unaltered (Oriji and Keiser 1998). Response to nitroglycerin was unchanged by CsA in patients (Stein et al. 1995).
CsA contracted renal arteries and arterioles of rats in vitro (Lanese et al. 1994; Epstein et al. 1998). Repeated administration of CsA has been reported to increase (Rego et al. 1990), decrease (Roullet et al. 1995) or have no effect (Ressureicao et al. 1995) on contractile responses in mesenteric arteries of rats. Likewise, repeated administration of CsA has been reported to sensitise renal arteries of rats to noradrenaline (Mikkelsen et al. 1992), but this has not been confirmed by all studies (Diederich et al. 1992). In patients, contractile responses to noradrenaline remained unaltered by CsA (Stein et al. 1995).
Taken together, CsA causes vascular dysfunction in vivo which is characterised by decreased endothelial relaxations in man and in rats. In addition, CsA increases vascular contractility in vitro .
According to previous studies, CsA seems to have different effects on iNOS and eNOS expression (Table 2 and 3). Increased NO synthesis and eNOS mRNA and protein levels have been reported in endothelial cells exposed to CsA in vitro (Lopez-Ongil et al . 1996). The induction of eNOS seems to be due to transcriptional induction of the eNOS gene (Navarro-Antolin et al. 2000). CsA also caused monocyte eNOS upregulation in renal transplant patients (Calo et al. 1998) and eNOS overexpression in the rat kidneys (Amore et al . 1995) and in the renal cortex of uninephroctomised rats (Bobadilla et al. 1998). In addition, acute infusion of CsA into human forearm vessels increased NO activity in vivo (Stroes et al . 1997). However, long-term treatment with CsA did not affect renal or vascular eNOS expression in rats (Vaziri et al. 1998), suggesting that induction of eNOS by CsA is unclear in vivo .
. Effects of cyclosporine A on iNOS activity and/or expression.
iNOS, inducible nitric oxide synthase; IL-1
; LPS, lipopolysaccaride;
NOS, nitric oxide synthase; TNF-
. Effects of cyclosporine A (CsA) on eNOS activity and/or
expression. cGMP, cyclic guanosine 3',5'-monophosphate; eNOS, endothelial
nitric oxide synthase; NOS, nitric oxide synthase.
CsA inhibited transcription of iNOS in murine macrophage and vascular smooth muscle cells (Dusting et al. 1999). In addition, depressed renal iNOS expression has been reported in vivo in rats after CsA treatment (Bobadilla et al. 1998; Vaziri et al. 1998; Mervaala et al . 2000). CsA also impaired iNOS in rat aortic smooth muscle cells (Marumo et al. 1995) and inhibited NO production in cultured macrophages (Conde et al. 1995).
Little is known about the effects of CsA on nNOS. Long-term CsA administration has been reported to decrease nNOS expression in uninephroctomised rats (Bobadilla et al. 1998).
L-arginine has been shown to have a beneficial effect on CsA-induced endothelial dysfunction (Gallego et al . 1993; Prieto et al . 1997; Oriji and Keiser 1998) and on renal haemodynamics in CsA-treated rats (De Nicola et al . 1993). Short-term treatment with L-arginine also reversed CsA-induced hypertension (Oriji and Keiser 1998).
Several lines of evidence suggest that CsA-induced renal and cardiovascular toxicity may be due to oxidative stress. CsA increases hypoxia and free radical production in the rat kidney (Zhong et al . 1998). Antioxidants, such as a -tocopherol, ascorbate, lazaroids and superoxide dismutase/catalase, have been shown to diminish CsA-induced renal toxicity (Wang and Salahudeen 1994; Wolf et al . 1994).
The mechanisms and the role of oxidative stress in CsA toxicity are unknown. Free radicals may be derived directly from CsA or its metabolites (for review, see Buetler et al. 2000). Reperfusion subsequent to CsA-induced renal vasoconstriction and hypoxia may also increase formation of free radicals. One possible mechanism of CsA-evoked oxidative stress is eNOS hyperactivity. Chronically administered CsA increased eNOS activity in the rat kidney parallel with decreased renal function (Amore et al . 1995). The functional nephrotoxicity by CsA was prevented by L-arginine (Amore et al . 1995). The protective effect of L-arginine was reversed by NOS inhibitor, suggesting that NO was crucial in the effect of L-arginine (Amore et al . 1995). While the intracellular L-arginine is sufficient for basal NO production in endothelial cells (Arnal et al . 1995), the substrate supply may be inadequate for the escalated demands of activated NOS. During insufficient L-arginine supply, the increase in NOS activity would not correspond to increased production of NO. Instead, NOS may generate reactive superoxide anion and other oxygen radicals (Stamler et al . 1992). L-arginine treatment might prevent this by shifting NOS activity toward NO generation and reduced oxidant injury.
Some mechanisms of CsA-induced radical oxygen species generation besides eNOS hyperactivity have been suggested. These include alterations in calcium homeostasis, leading to vasoconstriction and CsA metabolism by cytochrome P450 system (CYP450), more specifically, the CYP450 3A isoenzyme (Ahmed et al. 1995; Buetler et al. 2000).
Kidneys play a pivotal role in the regulation of sodium balance. Several modulatory factors can control sodium excretion to a certain degree. These include RAS, aldosterone, the renal renal kallikrein-kinin system, dietary intake of potassium, calcium and magnesium, natriuretic peptides and endothelium-derived factors. According to the pressure-natriuresis theory presented by Guyton et al. (1972), an increase in dietary sodium beyond the control of these regulatory factors causes a rise in blood pressure to get rid of excess sodium. High sodium intake is one of the risk factors for development of hypertension in rats and in humans (Mervaala et al. 1992; Dyer et al. 1994).
In rats, short-term CsA administration reduced urinary excretion of sodium (Ryffel et al . 1986; Kaskel et al . 1987). Vasoconstriction of renal arterioles causes a reduction in perfusion pressure that may cause sodium retention and, therefore, a rise in blood pressure. In SHR, CsA caused a pronounced rise in blood pressure and renal dysfunction during high intake of sodium (Mervaala et al. 1997). In contrast, during normal-sodium diet, the blood pressure increasing effect of CsA was only moderate and no effect on renal function was observed. Evidence also exists for disturbances in renal sodium handling playing a role in CsA-induced hypertension in patients. Sodium restriction has been shown to lower blood pressure in hypertensive transplant recipients treated with CsA, but not in hypertensive transplant recipients treated with azathioprine and glucocorticoids (Curtis et al. 1988; Curtis 1994). These findings together suggest that high intake of sodium is a risk factor for CsA-induced hypertension and renal toxicity.