Conflict of interest: Nothing to report.
Polymorphic Metabolism May Affect Rate of Disease Progression in PBC
It is a highly plausible hypothesis that primary biliary cirrhosis (PBC) may be caused by aberrant metabolism of xenobiotics, generating haptens which trigger an autoimmune response in genetically susceptible individuals. Kimura et al. postulated that xenobiotic structural analogues of lipoic acid may act as haptens in promoting the formation of antibodies which characterize PBC with apparent specificity for the lipoyl-rich domain of pyruvate dehydrogenase. Candidate genes include those responsible for drug and hormone metabolism and transporters involved in intestinal absorption and biliary secretion of potentially toxic molecules. Kimura and colleagues chose to study a series of these candidates to determine the prevalence of single nucleotide polymorphisms which are known to affect the biological functioning of their products and which play a role in detoxification of halogenated antibiotics. Cytochrome P450 2D6 (CYP2D6) is involved in metabolism of approximately 20% of drugs; the alcohol inducible CYP2E1 metabolizes estrogenic metabolites; MDR1 P-glycoprotein is active in bile secretion and intestinal absorption of various xenobiotics; pregnane X receptor (PXR) is a nuclear receptor for various molecules including steroid hormones which regulates the expression of CYP3A4 and MDR1 in the liver and intestine. In a study of 169 patients with PBC and matched controls no differences were found in the frequency of 4 polymorphic variants of CYP2D6, the c1 and c2 variants of CYP2E1or known polymorphic forms of MDR1, nor of PXR. These are important negative findings which exclude a number of candidate genetic traits from causality in conferring genetic susceptibility to development of PBC. The approach remains one which promises to bear fruit during screening for other likely contenders. An ancillary finding from this study was the observation that of the two CYP2E1 variants the c2 allele was associated with more advanced disease, as manifested by the depth of jaundice and presence of ascites, after correction for age and disease duration. Thus, genetically determined differences in xenobiotic metabolism that have no role in initiation of PBC may, nevertheless, influence the rate of disease progression. (See HEPATOLOGY 2005;41:55–63)
UDCA Fails to Induce CYP3A4
Many mechanisms have been postulated to account for the therapeutic benefit of ursodeoxycholic acid (UDCA) in primary biliary cirrhosis (PBC) including induction of cytochrome P450 3A4 (CYP3A4). The CYP3A subfamily represents the most abundant drug metabolizing enzyme in human liver and intestine and is responsible for metabolism of many endogenous compounds as well as about 50% of drugs in clinical use. Its ability to biotransform secondary bile acids to less toxic metabolites is protective during cholestasis. Previous studies in murine microsomes and human hepatocyte culture found CYP3A was induced by UDCA to a far greater extent than was seen with any other human bile acid. To test the hypothesis that therapeutic benefit from UDCA in PBC accrued via induction of CYP3A, Dilger et al. determined the metabolite profile of CYP3A4 substrates in PBC patients and controls while they were treated consecutively with UDCA and rifampicin, a well-known CYP3A4 inducer. Twelve patients with early PBC and controls took UDCA (15 mg/kg/day) for three weeks followed by rifampicin (600 mg/day) for one week. As anticipated, pharmacokinetic studies following exposure to rifampicin showed dramatic increases in conversion of budesonide to its phase I metabolites (6β-hydroxybudesonide and16α-hydroxyprednisolone) and conversion of endogenous cortisol to 6βhydroxycortisol. As an example of the magnitude of rifampicin induction the ratio of areas under plasma concentration-time curves for budesonide and its metabolite — (AUC16- α hydroxyprednisolone/AUC budesonide) rose from 8.6 ± 3.9 to 527 ± 248.7. The virtual elimination of detectable budesonide (reduction by 99%–100% of AUCbudesonide) from serum reflects the massive induction by rifampicin of CYP3A4 in intestine and liver as well as coordinate induction of the intestinal drug transporter protein P-glycoprotein for which budesonide is also a substrate (See figure). In stark contrast, following UDCA treatment, no significant difference was seen in CYP3A4-mediated metabolism of budesonide or cortisol in either controls or patients with PBC. (See HEPATOLOGY 2005;41:595–602)
The Promise of Proteomics
Proteomic array technology permits profiling of serum proteins according to molecular weight. Paradis et al. compared 44 patients with hepatocellular carcinoma (HCC) complicating cirrhosis with 38 patients who had cirrhosis but were unaffected by HCC. Their technique of surface-enhanced laser desorption ionization time-of-flight mass spectrometry (SELDI-TOF MS) demonstrated more than 250 protein peaks from each patient's serum. Comparing peak intensities revealed 30 proteins whose mean peak intensity differed significantly between the groups. Of these 13 were significantly higher in patients with HCC and 17 were higher in the non-HCC patients with cirrhosis. The single most discriminant peak (for a given peak intensity cutoff value) showed an AUC of the receiver operator characteristic curve of 0.85 with a sensitivity of 85% and specificity of 75% for diagnosis of HCC. In comparison, for the same cohort an α-fetoprotein >20 had an AUC of 0.72, sensitivity of 60%, and specificity of 73%. Following regression analysis an algorithm was developed embracing data on 6 peaks in the initial set of sera analyzed. Application of the algorithm to 42 independent samples correctly classified 92% of the samples as HCC or non-HCC cirrhosis. Analysis of the most discriminant 8,900-Da protein showed that it corresponded to the C-terminal fragment of the vitronectin precursor. Its peak intensity correlated with tumor size (See figure) Further analysis supported the view that its increased serum concentration results from enhanced digestion by upregulated metalloproteinase 2 in HCC rather than increased expression of vitronectin by HCC. (See HEPATOLOGY 2005;41:40–47)
Female, Fat, and Fertile—Breeding Gallstones
The old adage about predisposition to gallstone formation has been well attested by a recent study. Ultrasonography was performed on 3,254 women during pregnancy and postpartum. Those with gallstones present on their initial ultrasound or having a history of cholecystectomy were excluded. De novo gallbladder sludge or stones or progression of baseline sludge to stones developed progressively throughout pregnancy, reaching an incidence of 10.2% by 4–6 weeks postpartum. This lithogenic influence of pregnancy was reversed subsequently with regression of sludge and stone. Even so, cholecystectomy was undertaken in 0.8% of the women during the first year after delivery. To determine which factors carried the greatest risk for gallstone formation during pregnancy serum glucose, lipids, leptin, estradiol, and progesterone were measured at 26–28 weeks' gestation. Prepregnancy BMI proved to be a strong predictor of incident gallbladder disease with an incidence of 2.7% for BMI<25 and 11.7% for BMI >30 kg/m2 (P < .001). The tendency to form sludge or stones correlated with serum triglyceride levels and inversely with HDL cholesterol. After adjusting for BMI, of the various metabolic influences studied serum leptin was shown to be most strongly associated, increased risk of gallbladder disease being conferred with an odds ratio of 1.05 (CI, 1.01–1.11) for each 1-ng/dL increase in serum leptin concentration. Conversely, having adjusted for serum leptin there was no additional risk of gallbladder disease from being obese. Further studies to elucidate this association will be eagerly awaited. (See HEPATOLOGY 2005;41:359–365)
COX-2 Inhibition and Renal Function in Cirrhosis
Exposure to nonsteroidal antiinflammatory drugs (NSAIDs) which act via inhibition of cyclooxygenase (COX) carries a risk of precipitating renal failure in patients with cirrhosis. It has been claimed that this risk may not pertain to the more recently developed selective COX-2 inhibitors which specifically antagonize the inducible isoform of the enzyme which is involved in the inflammatory response. Clària et al. have investigated these claims in patients with cirrhosis and ascites in whom serum creatinine was less than 1.5 mg/dL, plasma renin activity greater than 4 ng/mL/h and glomerular filtration rate (GFR) more than 40 mL/min at commencement of the study. Patients were allocated by double-blind randomization to treatment with the COX-2 inhibitor celecoxib (200 mg every 12 hours for a total of 5 doses), placebo or the nonselective NSAID naproxen (500 mg every 12 hours for 5 doses). As previously described, naproxen (n = 6) had measurable effects on various aspects of renal function with significant reduction (P < .05) in GFR, renal plasma flow, urinary prostaglandin E2 excretion, diuretic, and natriuretic response to furosemide. None of these aspects of renal function changed significantly in groups on placebo (n = 6) or celecoxib (n = 7). Similarly, the known effect of NSAIDs in inhibition of platelet aggregation and thromboxane B2 production were documented during naproxen treatment but were not detectable during exposure to celecoxib. The pattern of results supports the contention that COX-1 but not COX-2 is necessary to maintain renal homeostasis, responsiveness to loop diuretics and prevention of the hepatorenal syndrome in patients with cirrhosis, ascites and high renin activity. Longer term studies may be anticipated in which these advantages are confirmed during exposure of patients with cirrhosis to selective COX-2 inhibitors which have been prescribed for their antiinflammatory effect.(See HEPATOLOGY 2005;41:579–587)
Site-Specific Promotion of Portal Vein and Hepatic Vein Thrombosis
Discovery of genetic abnormalities has facilitated study of the contribution of hereditary factors to thrombosis. Sixty-five patients with thrombosis of the extrahepatic portal vein (EHPVO) in isolation (19) or in combination with superior mesenteric and/or splenic vein thrombosis (46) were screened for hematological disorders. A myeloproliferative disorder, found in 35%, was the most common risk factor. The frequency of “gain of function” DNA mutations in Factor V (factor V Leiden) and in the promoter of the prothrombin gene (G20210A) were determined. EHPVO was significantly associated with the prothrombin mutation (OR, 8.1 [95% CI, 3.8–17.5]) approximately double the odds ratio for lower limb deep vein thrombosis (DVT). In contrast, there was no association of EHPVO with Factor V Leiden, despite the latter's strong association with DVT and, from previous publications, with the Budd Chiari syndrome (BCS). Similarly, there was no association of EHPVO with oral contraception, strengthening the notion that its etiology in terms of predisposing factors differs significantly from BCS and DVT. Primignani et al. also assayed the functional activity of antithrombin III, protein C, and protein S and measured plasma homocysteine. Taken together, deficiencies of antithrombin, protein C, and protein S were significantly associated with EHPVO (OR, 4.5; 95% CI, 1.1–18.0). Given the caveat that diminished activity of these factors occurs secondary to EHPVO the authors insisted on a corresponding finding in at least one blood relative before ascribing deficiency to heredity. (See HEPATOLOGY 2005;41:603–608).