Bile acid (BA)-dependent bile flow occurs from the concerted action of (1) absorption of BAs at the basolateral membrane, (2) transhepatic transport, and (3) secretion at the canalicular membrane. The molecular nature of the proteins, including membrane transporters, that are involved in this process have been defined over the last decade, and the mechanisms by which they are regulated are beginning to be uncovered. Although NTCP (sodium taurocholate co-transporting polypeptide; SLC10A1) is the major transporter involved in Na-dependent uptake of conjugated BA absorption from portal blood, multiple members of the organic anion transporting (OATP) family have been implicated as having some contribution in the uptake of BA as well as other organic anions such as bilirubin and many drugs.1 The holy grail in the canalicular secretion of BAs was resolved when bile salt export pump (BSEP) (ABCB11) was cloned and shown to be a member of the multidrug-resistant, ATP-binding cassette (ABC) family of transporters.2 The discovery of the molecular identity of the transport proteins and their cognate genes led to the definition of a genetic basis of inherited cholestatic syndromes collectively called progressive familial inherited cholestasis (PFIC). We now know that whereas PFIC1 is due to mutations in FIC1 (a phosphatidyl serine [PS] flippase) and BSEP in some patients with benign recurrent intrahepatic cholestasis (BRIC), PFIC2 is due to mutations in BSEP, and PFIC3 is due to mutations in multidrug-resistant P-glycoprotein 3 (MDR3).
Multiple (approximately 30) missense, premature termination, and frameshift mutations in BSEP have been identified in patients with PFIC2. In addition, BSEP mutations have also been found in a subset of BRIC patients with no changes in FIC1.3 Recent studies have reported BSEP mutations in acquired cholestasis (drug-induced) and in some cases identified single-nucleotide polymorphisms (SNPs) in the gene.4 Because premature termination and frameshift mutations typically result in proteins that are truncated, abnormal, and hence nonfunctional, most of the studies have concentrated on missense mutations and their effect on BSEP function using heterologous systems employing Sf9 insect cells (baculovirus-encoded complementary DNA [cDNA]) or Madin-Darby canine kidney (MDCK) cell line (using plasmid-encoded or adenovirus-encoded cDNA). These studies have been carried out using mouse, rat, and human BSEP cDNAs. These studies examined the following BSEP mutations and their effects on function and targeting to the apical membrane: (1) G238V, (2) E297G, (3) C336S, (4) D482G, (5) G928R, (6) R1153C, and (7) R1268Q. Wang et al.5 introduced these missense mutations into rat Bsep and examined Bsep expression in MDCK and Sf9 cells transfected with the mutated DNA constructs. Their data revealed that although 5 mutations (G238V, E297G, G982R, R1153C, and R1268Q) prevented apical targeting of Bsep, 4 mutations (E297G, G982R, R1153C, and R1268Q) in addition also eliminated transport activity. Studies by Plaas et al.6 analyzed the effect of D482G in mouse Bsep in HepG2 cells and found that this mutation did not affect transport whereas membrane and intracellular protein amount was significantly affected combined with a decrease in glycosylation. Similar to the observations with ΔF508 mutation in cystic fibrosis transmembrane conductance regulator (ΔF508-CFTR),7 growth of cells at a lower temperature led to increased messenger RNA and protein along with increased membrane protein and glycosylation. Thus, these data were consistent with the data obtained by Wang et al. in that the D482G mutation in rat and mouse Bsep altered the trafficking and processing of the Bsep protein. In their earlier studies, Hayashi, Sugiyama, and colleagues8 introduced 2 common mutations found among European PFIC2 patients (E297G and D482G) in human BSEP and studied them by adenovirus-mediated infection into MDCK and human embryonic kidney 293 cells. Their results showed that whereas most of the D482G and some of the E297G mutants underwent core glycosylation, there was a significant reduction in the intracellular protein content together with decreased membrane trafficking. Treatment with the proteasome inhibitor MG132 (carbobenzoxy-L-leucyle-L-leucyle-L-norralinol) increased the BSEP protein content in endoplasmic reticulum which was core glycosylated, suggesting increased proteasomal degradation of the mutated protein. Thus, studies from 3 different laboratories made the common observation that the D482G mutation in BSEP led to decreased protein whereas decreased membrane trafficking was seen in the case of human and mouse proteins. Hayashi and Sugiyama now show, as reported in this issue of HEPATOLOGY, that this membrane trafficking defect could be corrected with the use of the agent sodium 4-phenyl butyrate (4-PBA).
More than a decade ago, Li et al.9 showed, using baculovirus-encoded ΔF508-CFTR, that the phosphorylation-regulated chloride channel property of CFTR was not affected by the mutation thus narrowing the effect to an altered trafficking to the membrane. Subsequent to these studies, Rubenstein et al.7 used 4-PBA and elegantly showed that membrane trafficking defects in ΔF508-CFTR can be rescued by this compound. The agent 4-PBA is a less toxic form of butyrate which has been employed in many studies to increase gene expression by rendering chromatin more permissive to transcription because of its ability to inhibit histone deacetylases. However, its use in reversing defective targeting of a membrane transporter had not been shown. However, Rubenstein et al. demonstrated that nasal epithelia obtained from patients with ΔF508-CFTR exposed to 0.1-2 mM 4-PBA for 7 days restored forskolin-activated chloride secretion. In addition, cultures of primary nasal epithelia and a bronchial epithelial cell line (IB3-1) when treated with 4-PBA led to the appearance of CFTR protein with higher molecular mass suggestive of addition or modification of carbohydrates and processing in the Golgi. Encouraged by this finding and later clinical trials of 4-PBA,10 Hayashi and Sugiyama proposed that mutations that introduced trafficking defects in BSEP could be rescued by treatment with 4-PBA. Proof of this principle is illustrated in their article appearing in this issue.
Hayashi et al. show that when 4-PBA was added to MDCK II cells infected with wild-type or mutant (E297G and D482G) BSEP cDNA-containing adenoviruses, it led to increased appearance of mature functional protein in the membrane (as assessed by biotinylation and transcellular transport) compared to untreated cells. Kinetic analysis showed that this increase was due to an increase in Vmax(reflecting transporter density on the membrane) rather than Km(affinity of the transporter for substrate). What is different about these studies from those on CFTR is the fact that trafficking of the wild-type protein can also be increased by 4-PBA in the BSEP-infected cells. Although 4-PBA was used in earlier work because of its effect on transcription enhancement, mechanisms of how it improves maturation of membrane proteins (CFTR and now BSEP) and targeting to the correct membrane domain is not clear. In this context, some headway has been achieved through a few recent studies. Using microarray and proteomic profiling approaches, these investigators identified as many as 85 proteins that were modulated by 4-PBA treatment in the bronchial epithelial cell line IB3-1.11, 12 Notable among these proteins were heat shock protein chaperones. However, other protein classes that were altered by 4-PBA addition included members of structural elements, cellular defense, protein biosynthesis, and ion transport families. Such a microarray/proteomic approach in cells of hepatocellular origin might potentially be useful in development of newer compounds with protein maturation capacity. Multiple pathways and interactions with proteins such as HS1-associated protein X-1 (HAX-1) and myosin II regulatory light chain (MLC2) are involved in BSEP trafficking to the membrane (Fig. 1), which seemed to regulate 2 separate intracellular pools of the protein in the hepatocyte.13 It remains to be seen whether 4-PBA affects any of these pathways in addition to facilitating Golgi processing, although the data from this study show that it increases the cell surface–resident BSEP. Also not known is what effect 4-PBA will have on other missense mutations (G238V, G982R, R1153C, and R1268Q) studied by Wang et al. If it turns out that 4-PBA improves the targeting of the other BSEP missense mutations its use will have much broader implications and may even be considered for use in MRP2 targeting mutations (Delta R, M [del R1392-M1393], and R768W)14, 15 observed in Dubin-Johnson syndrome. It is also unclear why the D482G mutation in rat Bsep did not affect apical targeting but did so when mouse or human BSEP was expressed possibly arguing for a dominant effect in a species-specific fashion. Hayashi et al. also show that these effects of 4-PBA are tenable in vivo because Sprague-Dawley rats injected with this compound showed enhanced Bsep expression in the canalicular membrane which resulted in increased biliary secretion of [3H]taurocholate. Thus, their case for potential use of 4-PBA in cholestatic patients with a decreased expression of BSEP might be well worth considering going forward.
In summary, the studies published in this issue by Hayashi and Sugiyama show great promise in considering therapeutic approaches to ameliorating diseases involving mutations in liver transporters, at least from the point of targeting-defective mutations. Combined with approaches involving gene therapy schemes and the use of nuclear receptor ligands, the future looks promising. Additional studies are needed to further refine technologies and develop additional small-molecule compounds that are more effective in the hepatocyte. Once experimental studies using animal models and cell lines have demonstrated the efficacy and safety of these therapeutics, rigorous clinical trials followed by FDA approval may provide hope for this class of liver disease in humans.