Exercising the nuclear option to treat cholestasis: CAR and PXR ligands


  • Saul J. Karpen M.D., Ph.D.

    Corresponding author
    1. Texas Children's Liver Center, Department of Pediatrics, Baylor College of Medicine, Houston, TX
    • Texas Children's Liver Center, Department of Pediatrics/GI, Hepatology & Nutrition, Baylor College of Medicine, One Baylor Plaza, Houston, TX 77030
    Search for more papers by this author
    • fax: 832-825-4893

  • See Article on Page 420

  • Potential conflict of interest: Nothing to report.

“No good deed goes unpunished.” This catchy phrase, attributed to the 20th century American diplomat and playwright Clare Boothe Luce,* may well apply to the multitasking hepatocyte and its broad responsibility to handle toxins, both from within (endobiotics) and without (xenobiotics). Some cell has to do it, and the majority of this responsibility lies within the hepatocyte, perfectly poised to be a first-line defense from potential xenobiotics present in both diet and drug therapies. In brief, the hepatocyte is a sensor and responder to toxins—able to disarm and place these toxins into blood or bile, destined for delivery outside the body in urine or feces, respectively. In normal circumstances, the hepatocyte is able to perform this function by engaging a marvelously rich and diverse means of self-protection and detoxification that involves a host of gene products. Although the mechanisms are broad, there are three generalizable components, usually divided into phases. First, hydroxylation (phase 1), then conjugation (phase 2) to increase solubility, and finally export (phase 3) out of the hepatocyte across either the sinusoidal or canalicular membranes. In cholestasis, the route of biliary excretion of toxins is markedly diminished or even blocked, leading to the accumulation of intracellular biliary constituents including bile acids, a prototypic endobiotic. The enterohepatic circulation of these detergent molecules is usually not a burden to the hepatocyte, but in cholestasis this biliary constituent is retained within hepatocytes and this “good deed” of handling bile acids soon becomes a form of cellular “punishment” in the form of markedly disordered structure and function, as well as engaging a procession to cell death. One would hope that if bile acid efflux across the canalicular membrane is blocked, then the hepatocyte can figure out ways to boost its ability to unload them through export across the sinusoidal membrane.


NR, nuclear receptor.

From work over the past few years, it is clear that the hepatocyte does possess the molecular means to handle intracellular accumulation of bile acids, ultimately resulting in sinusoidal export. Interestingly, these phase 1–3 processes are not static, but highly regulatable.1 One of the more exciting findings has been that bile acids are themselves cholestatic sensors, and can engage processes of self-protection, mainly through altering the transcriptional program in the hepatocyte that centers on the activities of the nuclear receptor (NR) for bile acids, FXR.2, 3 Although FXR-regulated genes lead to reduced sinusoidal bile acid import, reduced bile acid synthesis, and increased bile acid canalicular export, it is not clear that such processes are critical for the hepatocyte to respond to increased bile acid retention and increase sinusoidal export. In fact, it may be harmful in some cholestatic models like bile duct ligation, although this is controversial.4, 5 Moreover, if the processes of handling retained bile acids were perfectly effective and self-protective, then liver disease would not develop in cholestasis. This wish is unfulfilled, as we note daily in our cholestatic patients, perhaps most readily evident in children who have a specific genetic defect in the canalicular bile acid exporter, BSEP (Abcb11; disease known as PFIC2), as well as those with cholangiopathies.6 For essentially all of these diseases, liver transplantation is the only durable route to a cure. Thus, the natural mechanisms for handling bile acid retention are present, but inadequate, providing a fertile field for consideration of engaging other non-FXR-dependent components of phase 1–3 detoxification of bile acids as potentially amenable to enhancement.

In this issue of HEPATOLOGY, Wagner et al. from Michael Trauner's group in Graz, Austria, provide a series of experiments that support the concept there may very well be a means to pharmacologically augment bile acid handling in cholestasis.7 Their work is based on recent findings from several labs that detoxification and elimination pathways are under the control of genes regulated by members of the NR superfamily.3, 8, 9 This family of multipartite gene regulators now totals about 48 members in humans, and involves most notably steroid hormone receptors, as well as a host of other receptors, many of which are highly expressed in liver. The significant aspects of nuclear receptors as gene regulators is that they bind to DNA to regulate target genes, but are also themselves regulated by ligands, as is the case of FXR by bile acids. There is a small select group of NR superfamily members intimately involved in regulating critical detoxifying pathways in hepatocytes and other cells, and we are learning more about the actual effector mechanisms and target genes. Like FXR, these are members of the type 2 NR subfamily, identified as those that function as heterodimer partners with the nuclear receptor RXR, itself activated by retinoids. When these RXR-containing NR pairs bind to their sequences in the regulatory regions of target genes, they recruit cofactors in the transcription apparatus to activate the expression of many phase 1, 2, and 3 genes. Among the critical partners of RXR that mediate this regulation are CAR, PXR, RAR, LXR, PPARα, PPARγ and of course, FXR. Work in the NR field has clearly identified two of these nuclear receptor family members, CAR and PXR, as critically necessary for activating many genes involved in detoxification pathways in liver and other tissues, with the greatest amount of support for regulating phase 1 and 2 genes.10–12 In mice that are deficient in either of these genes, they have clear, select and impressively debilitated capabilities of detoxifying xenobiotics and endobiotics.13–15 This is most evident in the regulation of the central P450 detoxifying enzyme Cyp3a4 (Cyp3a11 in mice), which is strongly activated by both CAR and PXR ligands, and reduced in activity in CAR−/− and PXR−/− mice. Moreover, these two NRs are critical to the handling of bile acids, notably lithocholic acid.16 The central question is whether or not it is possible to increase the capacity of the hepatocyte to not only detoxify, but also rid itself of, excess bile acids by activating these same processes with CAR and PXR ligands. Wagner and colleagues have set out to do just that.

What Wagner et al. first determined was whether or not specific CAR and PXR ligands can activate genes involved in bile acid detoxification/export in native, non-pathological conditions in mice. In the cytochrome P450 family, CYP3a11 and its important polyspecific cousin, CYP2b10, were upregulated by both CAR and PXR ligands while an increase in phase 2 conjugating enzyme gene expression was activated primarily by CAR ligands. But what these authors have added, and what was not know before, was whether phase 3 (export) was also engaged by these ligands. This is critical, because without an increase in export, phase 1 and 2 modified bile acids may still cause damage through their persistent intracellular presence. Wagner et al. found that the hepatocyte does respond accordingly, by increasing the expression of sinusoidal transporter genes that can unload the cell of these toxins, place them in circulation, ultimately headed for renal excretion. More so with CAR than PXR ligands, RNA and protein levels of two broadly specific sinusoidal exporters in the MRP family, MRP3 and MRP4, are markedly upregulated by appropriate ligands. Thus, in normal mice, proof of principle for the entire process to detoxify and unload the hepatocyte of bile acids via phases 1–3 is evident.

While it is all well and good in the normal hepatocyte to increase bile acid detoxifying pathways with drugs, it would be more relevant if one can do so in the face of cholestasis, as it appears that such autoregulatory processes are adequate in the normal state. Determining if such agents can enhance the insufficient response to retained bile acids in cholestasis is really the goal. Using the most profound form of acute cholestasis, common bile duct ligation, Wagner et al. sought to determine if CAR and PXR ligands can even work in this state, and help the bile acid-laden hepatocyte. This, in fact seems so, but the results of treatment show a mixed final picture. On the positive side, treatments, especially with CAR agonists, lead to more hydroxylation of bile acids, with findings of relatively high amounts of the usually undetectable tetrahydroxylated and pentahydroxylated bile acids in urine. Serum bile acid levels rose to values over 1 mmol/L after bile duct ligation, whereas treatments with the CAR and PXR agonists, reduced this by half or more. Notably, urinary excretion of bile acids increased nearly threefold in response to CAR ligands. Finally, serum bilirubin levels dropped in response to CAR ligands, although this may reflect a combination of effects, including activation of bilirubin detoxifying pathways.17 But there is significant concern, because treatment with CAR and PXR ligands nearly doubled the increase in serum ALT levels above that seen with bile duct ligation alone. Whether or not there is increased and permanent histological damage, or if there is a dose effect, is extremely important and remains to be determined.

What is the relevance of these studies to possible new treatments for cholestatic liver diseases? The fact that the hepatocyte possesses a regulatable means of reducing intracellular concentrations of bile acids (and other toxins) provides an exciting and rational approach to directing research for new and effective anti-cholestatic agents, of which we have none to choose from at present. This overall concept makes sense, but requires validation in a variety of models, including chronic cholestatic and biliary tract-centered ones. Moreover, one cannot base the appropriateness of exploring the use of such agents until there are multiple ones to test and explore, akin to the history and use of selective estrogen receptor modulators, or SERMs.8 We would need a panel of potential SCARMs to find the right mix of regulating detoxification and export, while minimizing any cellular damage. The concept of increasing both detoxification and sinusoidal export seems quite reasonable, because trying to increase canalicular bile acid export in liver and biliary tract diseases where cholestasis and slow bile acid output are prominent features may not be effective at all (Fig. 1). I cannot think of a single significant liver disease where unloading more bile acids across the sinusoidal membrane would be inherently undesirable, nor would it be significantly damaging to the whole body, unless, perhaps, there was significant renal impairment. This is all speculative, but with studies like these from Wagner et al., one can envision a future where selective ligands for CAR, PXR, FXR, and perhaps other NRs, can be used in cholestasis and liver disease. This may not be as far away as one may think, as antidiabetic PPARγ ligands show promise as antifibrotic agents in both animals and humans, and two agents used clinically for decades in cholestasis have marked affinities for human NRs—phenobarbital is a CAR activator, while rifampicin is a potent PXR agonist.18–22

Figure 1.

Nuclear receptors CAR and PXR regulate bile acid detoxification and sinusoidal export. Outlined in green are functions primarily regulated by CAR; in purple, by PXR. The relative amounts of green and purple denote the relative contribution to these functions by the respective CAR or PXR ligands. The concepts of regulation of phase 1, 2, and 3 genes by CAR and PXR ligands represented in this highly simplified schematic are best understood by referring to Wagner et al.7

Studies like these can help determine if the dire need for safe and effective anti-cholestatic drugs can be filled by NR ligands, especially those that may selectively activate CAR or PXR. I am optimistic that with further studies like these, there will be a time in the near future when hepatologists will be confronted by a choice of anti-cholestatic agents, rather than the current state of frank paucity. In 1981 in HEPATOLOGY, Popper provided a wonderful review where he noted the need to learn more about cholestasis, and hoped for a time where therapeutic agents would specifically address the problems associated with retained bile.23 He finished this prescient review with the following challenge to future researchers:

Reflections of the crystal ball are dimmer for the more distant future, but they encourage dreams, for instance, that recombinant DNA … may substitute for defective enzymes … and replace liver function more effectively than do transplantation or artificial livers. Rational therapy of intrahepatic cholestasis remains elusive. H. Popper

With studies like those from Wagner et al., we are closer in time to answering this call to explore rationally-designed therapeutic agents that just may allow the hepatocyte to continue to do its good deeds.