Potential conflict of interest: Nothing to report.
Several experimental models of cirrhosis have shown dysregulation of renal aquaporins in different phases of liver disease. We investigated the urinary excretion of both aquaporin-1 and aquaporin-2 in patients with cirrhosis at different stages of the disease. Twenty-four-hour urine was collected from 11 healthy volunteers, 13 patients with compensated cirrhosis (without ascites), and 20 patients with decompensated cirrhosis (11 with ascites without renal failure and 9 with hepatorenal syndrome). Aquaporin-1 and aquaporin-2 excretion was analyzed by immunoblotting. Urinary aquaporin-2 excretion was reduced in patients with cirrhosis compared to healthy subjects. A progressive decrease in urinary aquaporin-2 excretion was observed as the severity of cirrhosis increased, from compensated cirrhosis to cirrhosis with ascites and hepatorenal syndrome. Patients with hyponatremia had lower urinary aquaporin-2 excretion than patients without hyponatremia. Vasopressin plasma level did not correlate with aquaporin-2 excretion. There were no differences between healthy subjects and patients with cirrhosis with or without ascites in urinary excretion of aquaporin-1, but urinary aquaporin-1 excretion of those with hepatorenal syndrome was extremely low. In conclusion, patients with cirrhosis appear to exhibit a decreased abundance of renal aquaporin-2 and therefore lower water permeability in the collecting tubules. This may represent an adaptive renal response to sodium retention, with expansion of extracellular fluid volume and dilutional hyponatremia observed in those who have cirrhosis with ascites. Finally, aquaporin-1 does not appear to play a role in the progressive dysregulation of extracellular fluid volume in cirrhosis. (HEPATOLOGY 2006;44:1555–1563.)
Cirrhosis is associated with altered regulation of sodium and water metabolism. During the natural course of cirrhosis, progressive impairment of kidney function occurs, with the kidneys no longer able to maintain the volume of extracellular fluid within normal limits. This mainly occurs because of an abnormal increase in tubular sodium, which leads to an inability to adjust the amount of sodium excreted by the kidneys to the amount of sodium ingested with the diet. Sodium retention is often associated with an impaired ability to eliminate solute-free water, which may lead to dilutional hyponatremia from a disproportionate increase in total body water relative to total sodium content. As the disease progresses, vasoconstriction of renal circulation may develop, which causes renal hypoperfusion, a reduced glomerular filtration rate, and, eventually, renal failure (the so-called hepatorenal syndrome).1
The kidney plays a central role in regulating the balance of salt and water. Disordered regulation of sodium and water transport in the kidneys is responsible for altered salt and water balance in several pathophysiological states, including cirrhosis. Water reabsorption in the kidneys occurs mainly through renal water channels, or aquaporins, which are membrane proteins localized along the renal tubules. Several aquaporins are expressed in the kidney.2 Aquaporin-1 is expressed in the apical and basolateral membranes of the proximal tubule and the descending limb of Henle's loop cells, and aquaporin-2, recognized as the vasopressin-regulated water channel, is expressed in the apical membrane of the collecting-duct cells.3 Aquaporin-3 and aquaporin-4 are localized in the basolateral membrane of the collecting ducts. The water permeability of the renal tubules is mainly dependent on the amount of aquaporins at the membranes of epithelial tubular cells.
Several studies using experimental models of cirrhosis have reported contrasting results about aquaporin-2 expression in the kidney. Asahina et al. reported increases in aquaporin-2 following repetitive intraperitoneal injections of carbon tetrachloride.4 In contrast, in a previous study we were unable to demonstrate increased abundance of aquaporin-2 protein in rats with carbon tetrachloride-induced cirrhosis.5 However, the rats with cirrhosis with ascites showed increased trafficking of aquaporin-2 to the plasma membrane and, interestingly, increased aquaporin-1 in the kidneys. Studies of cirrhosis induced by common bile duct ligation have reported decreased expression of aquaporin-2 in the kidneys of rats with ascites.6, 7 These differences confirm that the physiological states reached in the 2 models are very different, even though both are associated with salt and water retention. The differences reemphasize the complexity of the mechanisms that regulate salt and water balance and point to the necessity of further research, including studies of the molecular pathophysiology of patients with cirrhosis in order to determine how the principles defined in animal studies apply to human disease.
Aquaporin protein research in humans has focused on urinary aquaporin excretion because of the availability of urine and the conflict about whether it is ethical to obtain kidney tissue. To our knowledge, only 2 studies have so far been reported investigating urinary aquaporin excretion in patients with cirrhosis, again with contrasting results.8, 9 Our study had 2 objectives. The first objective was to investigate urinary aquaporin-2 excretion during the progressive impairment of sodium and water metabolism in cirrhosis. To address this, we studied patients in different phases of sodium and water balance disorder. The second objective was to extend the investigation to aquaporin-1 because some experimental data suggested it could contribute to the sodium and water balance disorder present in cirrhosis.5
Patients and Methods
The exclusion criteria for the healthy controls were: (1) a history of clinical signs of disease of the heart, lungs, liver, kidneys, or endocrine organs; (2) neoplastic disease; (3) arterial hypertension; (4) alcohol or drug abuse; (5) treatment with any type of drugs.
The diagnosis of cirrhosis was based on a liver biopsy or classical clinical, biochemical, and ultrasonographic findings. Exclusion criteria for patients with cirrhosis were: (1) history of clinical signs of heart or lung disease, arterial hypertension, or parenchymal renal failure; (2) gastrointestinal bleeding or active infection 7 days prior to the study; (3) treatment with cyclooxygenase inhibitors or other nephrotoxic drugs 30 days prior to the study; and (4) treatment with diuretics 5 days prior to the study. Patients were classified into 1 of 3 groups according to the severity of renal dysfunction: patients with compensated LC (patients without ascites), patients with ascites without renal failure, and patients with hepatorenal syndrome. Hepatorenal syndrome was diagnosed using widely accepted criteria.10
Informed consent was obtained from all participants. The study was approved by the ethics committee of the institution.
Water Deprivation and Water Loading Test in Control Subjects.
Two control subjects were submitted to an 8-hour period of water deprivation and 3 days later to an 8-hour period of water loading (400 mL/hour). Urine was collected during these periods. Then 10 mL of each urine sample was frozen at −80°C until assayed.
All subjects were allowed to drink freely during the experiment. A 24-hour urine was collected from each subject the day before the blood sample was drawn. Then 10 mL of each urine sample was frozen at −80°C until assayed.
Blood and Urine Analyses
Plasma and urine osmolarity were determined from the osmometric depression of the freezing point (Osmometer 3300; Advanced Instruments, Needham Heights, MA), and sodium, creatinine bilirubin, prothrombin time, and albumin were determined by standard analytical methods. Plasma renin activity and plasma concentration of aldosterone, vasopressin, and norepinephrine were determined by radioimmunoassay (Clinical Assays, Cambridge, MA; Diagnostic and Products Corporation, Los Angeles CA; Bühlman Laboratories AG, Basel Switzerland; and CAIBL Laboratories, Hamburg, Germany, respectively).
Measurement of Urinary Aquaporin-1 and Aquaporin-2
Urinary aquaporin-1 and aquaporin-2 were measured using a method described in detail elsewhere.11 Briefly, urine samples were thawed and centrifuged at 1000×g for 5 minutes at 4°C to remove cellular debris. A total of 150 μg of creatinine equivalent of each sample was then concentrated by ultrafiltration using Amicon® Ultra-4 Centrifugal Filter Devices (Millipore, Bedford, MA) with a 10,000-Da cutoff according to the protocol provided by the manufacturer. Concentrated urine samples were then prepared as previously described and subjected to immunoblot analysis in order to semiquantify the amount of aquaporin-1 and aquaporin-2 in the sample.
Electrophoresis and Immunoblotting of Proteins.
Electrophoresis was run for the entire set of samples on 12% polyacrylamide/SDS gels. Immunoblots were run as previously described.5 The densitometry values were normalized to the mean of the control group to facilitate comparisons.
Affinity-purified, peptide-derived polyclonal antibodies to aquaporin-1 and aquaporin-2 were used for immunoblotting.5 The specificity of the antibodies was tested by competition experiments.
Presentation of Data and Statistical Analyses
Quantitative data are presented as mean ± standard error of the mean (SEM). Statistical comparisons were done using ANOVA, the unpaired t test (when variances were the same), the Mann-Whitney rank-sum test (when variances were significantly different between groups), and the chi-square test. P values less than .05 were considered statistically significant.
Eleven healthy controls and 33 patients with cirrhosis were included in the study. Table 1 shows the demographic characteristics and liver and renal function test and vasoactive hormone data of the healthy subjects and patients with cirrhosis in the study divided into 3 groups: patients without ascites, patients with ascites without renal failure, and patients with hepatorenal syndrome. Patients with cirrhosis showed progressive impairment of liver and renal function according to test results as the disease progressed from compensated cirrhosis to cirrhosis with ascites without renal failure to hepatorenal syndrome. Likewise, progressive stimulation of the renin-aldosterone system, sympathetic nervous system, and vasopressin was observed. Thirteen of the 33 patients (39%; 8 with hepatorenal syndrome) had serum sodium concentrations below normal (135 mmol/L), and 8 of them (24%) met the criterion for hyponatremia (serum sodium ≤ 130 mmol/L). Interestingly, patients with hyponatremia had been so for a relatively long time (average 49 days, range 15-60 days).
Table 1. Demographic, Clinical, Biochemical, and Hormonal Data of Healthy Subjects and Patients With Cirrhosis
Patients Without Ascites
Patients With Ascites
Patients With Hepatorenal Syndrome
NOTE. Data are mean ± SEM.
Abbreviations: M, male; F, female; HCV, hepatitis C virus.
Characteristics of Immunoblot Technique With Urinary Proteins
Aquaporin-1 and Aquaporin-2 Proteins in Urine.
Figure 1A and Fig. 2A show immunoblots loaded with urine sample from a healthy subject in order to detect aquaporin-1 and aquaporin-2, respectively. To test for the specificity of the antibody labeling, competition experiments were carried out (Fig. 1B and Fig. 2B). Figure 1C and Fig. 2C show a linear relationship between urinary aquaporin-1 and aquaporin-2 band density and urinary creatinine loading in the gel.
Water Deprivation and Water Loading Test in Healthy Subject.
Figure 3 shows an immunoblot of urinary aquaporin-2 during an 8-hour period of water deprivation and an 8-hour period of water loading in a healthy subject. The abundance of the vasopressin-regulated water channel aquaporin-2 in the urine during the water deprivation test increased over that during the loading test. This suggests, as previously demonstrated,12, 13 that under physiological conditions urinary aquaporin-2 excretion can be used as an index of the action of vasopressin in the kidneys of healthy subjects.
Aquaporin-1 and Aquaporin-2 Excretion
We analyzed the urinary excretion of aquaporin-1 and aquaporin-2 of healthy subjects. Subject 1 had the highest aquaporin-1 and aquaporin-2 urinary excretion of those in the control group, whereas subject 2 had the lowest aquaporin-1 and aquaporin-2 urinary excretion (data not shown). Therefore, we loaded the urine from these 2 control subjects in all immunoblots, and we used the densitometric values of control subject 2 to normalize the aquaporin band density.
Urinary Aquaporin-1 Excretion.
We quantified the aquaporin-1 bands of patients with cirrhosis by densitometry and normalized the values by dividing them by the densitometric value of control subject 2 (the control subject with the lowest aquaporin-1 signal). As shown in Fig. 4, there were no differences between healthy subjects and patients with cirrhosis with or without ascites in urinary excretion of aquaporin-1, whereas it was markedly reduced in patients with hepatorenal syndrome. The mean urinary excretion of aquaporin-1 was: 0.91 ± 0.10 in healthy subjects, 1.18 ± 0.18 in patients without ascites (P = .36 compared to healthy subjects), 0.78 ± 0.20 in patients with ascites (P = .69 compared to healthy subjects), and 0.02 ± 0.10 in patients with hepatorenal syndrome (P = .0001 compared to healthy subjects).
Urinary Aquaporin-2 Excretion.
We quantified the sum of both bands (glycosylated and nonglycosylated) shown in the immunoblot by densitometry and normalized the values by dividing them by the densitometric value of control subject 2 (the control subject with the lowest aquaporin-2 signal). The urinary excretion of aquaporin-2 in patients with cirrhosis was lower than that in healthy subjects. As shown in Fig. 5, a progressive decrease in urinary aquaporin-2 excretion was observed among patients with cirrhosis as the severity of the disease increased from cirrhosis without ascites to cirrhosis with ascites without renal failure to hepatorenal syndrome. The mean urinary aquaporin-2 excretion was: 2.72 ± 0.61 in healthy subjects, 0.63 ± 0.18 in patients without ascites (P = .0007 compared to healthy subjects), 0.45 ± 0.28 in patients with ascites without renal failure (P = .0004 compared to healthy subjects), and 0.22 ± 0.11 in patients with hepatorenal syndrome (P < .0001 compared to healthy subjects). When patients with cirrhosis were categorized according to serum sodium concentration, it was observed that patients with hyponatremia (serum sodium ≤ 130 mmol/L) had significantly lower urinary aquaporin-2 excretion than patients without hyponatremia (Fig. 6). A significant difference was also observed when a serum sodium cutoff of 135 mmol/L instead of 130 mmol/L was used (data not shown). Overall, there was no significant relationship between plasma vasopressin level and urinary aquaporin-2 excretion among the patients in the study (Fig. 7), as previously found in rats and healthy humans.13–16
The main findings of our study were: (1) excretion of urinary aquaporin-2 of patients with cirrhosis was lower than that of healthy subjects; and (2) urinary excretion of aquaporin-1 of patients with cirrhosis was similar to that of healthy subjects but was markedly reduced in patients with hepatorenal syndrome.
Aquaporin-2 Protein in Urine.
Aquaporin-2 is recognized as the vasopressin-regulated water channel. Vasopressin regulates aquaporin-2 to increase principal cell water permeability in 2 ways. Its short-term effect is to trigger translocation of aquaporin-2-containing intracellular vesicles to the apical plasma membrane, and its long-term effect is to increase synthesis of aquaporin-2 in the principal cells of the collecting duct.3, 17 Previous studies have demonstrated that the daily excretion of aquaporin in the urine is approximately 3%-4% of the total amount of aquaporin-2 in the kidneys,14, 18 indicating that the main excretion of aquaporin-2 is not a result of shedding whole cells and that the aquaporin-2 in the urine is associated with membrane structures: vesicles with defined membranes, also called exosomes; multivesicular bodies; large membrane fractions that represent luminal plasma membrane fragments; and vesicle debris. These previous studies demonstrated that urinary excretion of aquaporin-2 is mediated by an apical pathway tightly regulated by vasopressin. After long-term water deprivation or the administration of 1-desamin-8-D-arginine vasopressin, expression of aquaporin-2 is increased in intracellular vesicles and the apical plasma membrane (small subapical vesicles and multivesicular bodies) of principal cells, resulting in increased water permeability of these cells and in increased aquaporin-2 urinary excretion. Under conditions known to suppress the endogenous release of vasopressin, aquaporin-2 levels in the plasma membrane are reduced by translocation into an intracellular reservoir. In experimental models this has been shown to be associated with decreased urinary excretion of aquaporin-2, indicating a correlation between vasopressin plasma level, level of aquaporin-2 from the apical plasma membrane, and aquaporin-2 urinary excretion. Similar results have been observed in humans.13, 15, 16 In the current study, we also detected in healthy subjects an increase in urinary aquaporin-2 excretion under water deprivation conditions and a marked decrease in urinary aquaporin-2 after a water-loading period, further supporting this correlation.
One of the main findings of our study was that patients with cirrhosis had reduced aquaporin-2 in the urine. A progressive decrease in urinary aquaporin-2 excretion was observed as the severity of the abnormal regulation of fluid balance and renal dysfunction increased. In fact, the reduction in urinary aquaporin-2 excretion was greater in patients with ascites without renal failure than in patients without ascites, and patients with hepatorenal syndrome excreted even less aquaporin-2 than did patients with ascites. Therefore, the results show that with marked expansion of the volume of extracellular fluid, a reduced level of serum sodium, and increased plasma levels of aldosterone, norepinephrine, and vasopressin, all characteristic of advanced cirrhosis, there is reduced urinary excretion of aquaporin-2. These results contrast with those obtained in patients with congestive heart failure, another state of expanded extracellular fluid volume. Studies of humans with decompensated congestive heart failure have shown increased aquaporin-2 excretion9, 19 and a close correlation between urinary aquaporin-2 and vasopressin levels. Thus, the findings show that urinary aquaporin-2 excretion is not uniform across the pathophysiological states of extracellular volume expansion. At this point, it is unclear why the 2 states differ in the effect on urinary aquaporin-2 excretion. In general, patients with advanced cirrhosis more frequently have chronic hyponatremia and their plasma vasopressin levels are higher than those with congestive heart failure. These differences may explain, at least in part, the different findings for urinary aquaporin-2 excretion.
The results of our study have to be discussed in light of the current concepts about the pathogenesis of hyponatremia in cirrhosis.20–22 Hyponatremia results from impairment of the kidneys' capacity to excrete solute-free water. Impairment of solute-free water excretion may already be occurring in patients without ascites, but is very common in patients with ascites and is almost universal in patients with hepatorenal syndrome.20–22 The main pathogenic mechanisms leading to this disorder are impaired delivery of filtrate to the distal nephron because of increased proximal sodium reabsorption and increased plasma levels of vasopressin as a result of nonosmotic hypersecretion of vasopressin from the neurohypophysis. Increased plasma vasopressin would stimulate the V2 receptors on the capillary side of the principal cells of the collecting duct, which would result in increased permeability to water through an aquaporin-2-mediated process. Water would then be reabsorbed in an amount disproportionate to the amount of sodium retained by the kidneys, and dilutional hyponatremia would develop. According to this pathogenic concept of hyponatremia in cirrhosis, urinary aquaporin-2 would be expected to be increased, not reduced, as was found in our study. Nevertheless, if the AVP-aquaporin-2 system were persistently activated in cirrhosis, solute-free water reabsorption would be permanently increased relative to sodium retention, and serum sodium concentration would decline steadily to levels incompatible with life. However, this is usually not the case in clinical practice. In fact, in most patients with cirrhosis, hyponatremia develops in relation to some complication (infection, bleeding, etc.) but then is usually stable for weeks or a few months, as was true in patients in our study. Therefore, when hyponatremia develops in those with cirrhosis, it is likely that a protective mechanism is set in motion to prevent constant solute-free reabsorption of water in the kidneys and unrelenting hyponatremia. The reduction in urinary excretion of aquaporin-2 observed in our study may be such a mechanism.
Along these lines, it is important to point out that downregulation of aquaporin-2 is a crucial mechanism of the phenomenon of escaping the effects of vasopressin observed in healthy subjects and experimental animals. It has long been known23 that healthy individuals who have been given an infusion of vasopressin will initially retain water for several days and then show an escape phenomenon characterized by an increased volume of urine and a reduction in urine osmolality and body weight, despite continued administration of vasopressin. Recent studies have demonstrated that this phenomenon at least in part is a result of decreased expression of aquaporin-2 water channels in the renal collecting duct, which occurs independently of vasopressin and osmolality.24 The exact mechanism or mechanisms underlying this effect are not known but may be related to increased nitric oxide and/or prostaglandin synthesis in the kidney.24 As already noted, the reduced urinary excretion of aquaporin-2 found in patients with advanced cirrhosis may represent an escape mechanism similar to that found in healthy subjects receiving vasopressin to prevent progressive reduction in serum sodium levels. The finding of reduced urinary excretion of aquaporin-2 in the early stages of cirrhosis, prior to the development of overt hyponatremia, is intriguing and deserves further confirmation and investigation.
Comparison of the Results on Urinary Aquaporin-2 Excretion in Those With Cirrhosis With Those of Previous Studies.
Previous studies of urinary aquaporin-2 excretion in patients with cirrhosis reported contrasting results. Ivarsen et al. detected an increase in urinary aquaporin-2 excretion in patients with cirrhosis, whereas Pedersen et al. observed no changes in aquaporin-2 excretion.8, 9 Several aspects of these studies could explain the differences between their findings about aquaporin-2 urine excretion with those of our study. First, in the present work we studied patients with cirrhosis in different phases of sodium and water retention, that is, patients without ascites, patients with ascites, and patients with hepatorenal syndrome. Alteration in sodium and water balance metabolism is a progressive disorder that has been clearly documented to occur in patients with cirrhosis, and the steady states achieved in these different phases cannot be compared.1 Pedersen et al. studied 14 patients with cirrhosis, 6 of whom had ascites, analyzing aquaporin-2 urinary excretion in the patients as a single group. In contrast, Ivarsen et al. used the Child-Pugh score to stratify patients according to the clinical severity of their liver disease. However, it is known this score may not correspond to alterations in sodium and water balance metabolism. Second, all patients in this study were free of treatment (specifically diuretic drugs) known to alter the abundance of aquaporin protein in the kidneys and therefore in urine.25–27 Indeed, increased excretion of urinary aquaporin-2 has been demonstrated during furosemide treatment.27 In the Pedersen study, 9 of 14 patients with cirrhosis received diuretics up to 24 hours prior to urine collection, and in the Ivarsen study all patients were receiving diuretic treatment at the time of urine collection. Finally, although the technique used to detect urine aquaporin-2 differed among the 3 studies—dot-blot analysis in the Ivarsen study, radioimmunoassay in the Pedersen study, and immunoblot in our study—we do not think this can explain the differences found in aquaporin-2 urinary excretion. The 3 studies had similar procedures for urine collection and similar processing of the urine samples prior to analysis of the aquaporin-2. This is an important point—altering protein detection in urine—which has recently been documented.28 Therefore, the differences in the population studied and in the selection of study participants could explain the different results for aquaporin-2 urine excretion.
Aquaporin-1 Protein in Urine.
Results from our study detected aquaporin-1 protein in urine, demonstrating that human urine contains this water channel. Little is known about regulation of aquaporin-1. Knocking out aquaporin-1 demonstrated that this proximal water channel is necessary for the urinary concentration process.29 Aquaporin-1 has been shown to increase in the kidneys of rats with cirrhosis induced by carbon tetrachloride, pointing to potentially important effects of having cirrhosis on the proximal tubules, which may account for the increased fluid absorption in this segment. Micropuncture and lithium clearance studies have also found evidence of increased sodium and water reabsorption in the proximal tubules of animals with cirrhosis and of patients.30, 31 We observed, however, that the urinary excretion of aquaporin-1 was not altered in patients with cirrhosis with or without ascites. By contrast, urinary excretion of aquaporin-1 was extremely reduced in patients with hepatorenal syndrome. The decreased aquaporin-1 in the urine could theoretically be associated with decreased aquaporin-1 at the plasma membrane level. At this point, it is not known how the possible decrease in aquaporin-1 in the kidney could affect fluid reabsorption of the proximal tubule segment because it is known that aquaporin-1 is normally expressed at high levels in the proximal tubule. In addition, patients with hepatorenal syndrome had a decreased glomerular filtration rate. It is not known if renal insufficiency could alter the urinary excretion of aquaporins. Further studies are needed in humans to specifically investigate the origin of aquaporin-1 in urine.
Limitations of the Study.
This study had several limitations. First, the number of patients in the study was relatively small. Therefore, further studies with a larger number of patients, particularly those with ascites and hyponatremia without hepatorenal syndrome, would be important. Second, the patients with hyponatremia had had this condition for an average of 7 weeks. Therefore, it would be worthwhile to study patients with cirrhosis who are just developing hyponatremia or have developed it very recently in order to assess whether urinary aquaporin-2 excretion in this group differs from that of patients with more chronic hyponatremia. Third, patients were studied at a single point in time; it would be interesting to study patients at several time points in order to find out whether urinary aquaporin-1 or aquaporin-2 excretion change over time. Finally, it would also be important to study aquaporin-2 excretion in patients with cirrhosis whose serum sodium concentration has been normalized by the administration of V2 receptor antagonists, drugs that increase solute-free water excretion and improve serum sodium levels.32, 33
In conclusion, patients with cirrhosis, particularly those with hyponatremia and hepatorenal syndrome, have decreased urinary aquaporin-2 excretion relative to healthy subjects, which suggests there is down-regulation of aquaporin-2 expression in the kidneys of those with cirrhosis. This may be a compensatory response to the hyponatremia and extracellular fluid volume expansion that patients with cirrhosis have. In addition, unlike with healthy subjects and patients with decompensated congestive heart failure, the vasopressin plasma levels of patients with cirrhosis do not correlate with urinary aquaporin-2 excretion. Finally, aquaporin-1 does not appear to play a role in the progressive abnormalities of fluid balance and renal function in patients with cirrhosis.