Mrp4−/− mice have an impaired cytoprotective response in obstructive cholestasis


  • Potential conflict of interest: Nothing to report.


Mrp4 is a member of the multidrug resistance–associated gene family that is expressed on the basolateral membrane of hepatocytes and undergoes adaptive upregulation in response to cholestatic injury or bile acid feeding. However, the relative importance of Mrp4 in a protective adaptive response to cholestatic injury is not known. To address this issue, common bile duct ligation (CBDL) was performed in wild-type and Mrp4−/− mice and animals followed for 7 days. Histological analysis and serum aminotransferase levels revealed more severe liver injury in the absence of Mrp4 expression. Western analyses revealed that Mrp4, but not Mrp3, was significantly increased after CBDL in wild-type mice. Serum bile acid levels were significantly lower in Mrp4−/− mice than in wild-type CBDL mice, whereas serum bilirubin levels were the same, suggesting that Mrp4 was required to effectively extrude bile acids from the cholestatic liver. Mrp3 and Ostα-Ostβ were upregulated in Mrp4−/− mice but were unable to compensate for the loss of Mrp4. High-performance liquid chromatography analysis on liver extracts revealed that taurine tetrahydroxy bile acid/beta-muricholic acid ratios were increased twofold in Mrp4−/− mice. In conclusion, hepatic Mrp4 plays a unique and essential protective role in the adaptive response to obstructive cholestatic liver injury. (HEPATOLOGY 2006;43:1013–1021.)

Multidrug resistance–associated protein 4 (MRP4) (ABCC4) is an ATP-dependent organic anion transporter with broad substrate specificity.1–3 It is a member of the ABC transporter superfamily4 and is expressed in a variety of epithelia, including the basolateral and apical plasma membranes of the liver and kidneys, respectively.5–7 However, this array of substrates and specific tissue localization have not provided insight into Mrp4 function in vivo.

Mrp4 was first suggested to play a role in the adaptive response to the hepatic overload of bile acids following the genetic deletion of farnesoid X receptor (FXR), a bile acid sensor.8, 9 A subsequent study suggested that Mrp4 was not elevated upon feeding the hydrophobic bile acid, lithocholic acid.10 Despite this finding, other studies have demonstrated that hepatic Mrp4 is upregulated in both rats and mice after bile duct ligation6, 11 and in pediatric patients with progressive familial intrahepatic cholestasis.12 Recent studies demonstrate that MRP4 functions as an efflux pump for bile acids together with glutathione.13 Further support for a role for MRP4 in hepatic bile acid overload is the recent demonstration that Mrp4 is upregulated by the constitutive androstane receptor.14 Constitutive androstane receptor is a member of the nuclear hormone receptor superfamily that is required to elevate serum bile acids during cholestatic injury.15

Cholestatic injury in mice, rats, and humans can also result in adaptive responses in other basolateral transporters. Examples include Mrp3, which is primarily a bilirubin conjugate transporter and also a constitutive androstane receptor target,15 and the recently described heterodimeric bile acid and organic solute transporter, Ostα-Ostβ.16, 17 The relative contribution of Mrp4 in the protection against bile acid overload is unknown.

We used the recently developed Mrp4−/− mouse18 to determine whether the absence of Mrp4 alters the effect of common bile duct ligation (CBDL) on the level of serum and liver bile acids and the degree of liver injury. Our findings indicate that CBDL results in greater liver injury in the absence of Mrp4 than in wild-type mice, in parallel with a reduced ability to export bile acids into the plasma, where they can be cleared from the systemic circulation. Adaptive responses from other basolateral membrane bile acid transporters (Mrp3, Ostα-Ostβ) are unable to fully compensate for the absence of Mrp4 in this model.


Mrp, multidrug resistance–associated protein; CBDL, common bile duct ligation; FXR, farnesoid X receptor; SGPT, serum glutamic pyruvate aminotransferase; mRNA, messenger RNA.

Materials and Methods


Wild-type and Mrp4−/− mice were maintained in an equivalent 129/Sv and C57BL/6 mixed background and were housed in a temperature- and humidity-controlled room under a light cycle with free access to food and water. Animals were either subjected to CBDL or sham operation for a period of 7 days, a protocol approved by the Yale Animal Care and Use Committee. CBDL animals were injected subcutaneously with 40 μg vitamin K1 (Neogen Corp., Lexington, KY) on days 2, 4, and 6 following ligation. Animals were housed in metabolic cages in groups of 3-4 for the last 24 hours to collect urine. Upon sacrifice, serum and bile were obtained from the inferior vena cava and gall bladder, respectively, and the liver and kidneys were flushed with normal saline and quick-frozen in liquid nitrogen. Liver sections were fixed in 10% formalin for histology.

Quantitation of Necrosis.

A cross-section of the entire left lobe was fixed in 10% formalin in phosphate-buffered saline for 1 to 6 days, embedded in paraffin, sectioned, and stained with hematoxylin-eosin. The sections were coded and the extent of liver necrosis was graded blindly on a scale of 0-4.

Bile Acid Assay.

A kit (Trinity Biotech, Wicklow, Ireland) was used to quantify total 3α-hydroxy bile acids in serum, urine, bile, and extracted liver. Livers were homogenized in t-butanol/water at a ratio of 1:1 (1.0 mL/100 mg tissue) and extracted overnight at room temperature. After centrifugation, the supernatant was diluted at a ratio of 1:5 with normal saline and passed through a C18 column (Fisher Scientific, Pittsburgh, PA). The column was washed with water, and bile acids were eluted with methanol.

Bile acids retained in the livers of control and Mrp4−/− mouse livers after 7 days of CBDL were analyzed via high-performance liquid chromatography of conjugated bile acids using a modification of the method of Rossi et al.19 as previously described.20

Liver Enzymes.

Bilirubin and serum glutamic pyruvate aminotransferase (SGPT) were measured in serum using kits from Thermo Electron Corp. (Louisville, CO).

Quantitative Real-Time Polymerase Chain Reaction.

Tissue was homogenized in Trizol (Invitrogen, Carlsbad, CA), and total RNA was isolated following the manufacturer's instructions. Ten micrograms of RNA was used for reverse transcription using the ProStar First Strand RT-PCR kit (Stratagene, La Jolla, CA). An Applied Biosystems (Foster City, CA) 7500 Sequence Detection System was used for quantitation of specific messenger RNA (mRNA). Twenty-five–microliter reactions contained TaqMan Universal Master Mix (Applied Biosystems), specific primers at a concentration of 500 nmol/L, and probes at 250 nmol/L. Amplification of 1 to 2 μL of complementary DNA was performed in triplicate in 96-well plates via incubation at 50°C for 2 minutes and 95°C for 10 minutes, then cycled at 95°C for 15 seconds and 60°C for 1 minute for 40 cycles. Gapdh was run for each sample to normalize expression.


Protein Analysis (Western Blotting).

Polyclonal antibodies raised against Mrp2 (Abcc2), Ntcp (Slc10a1), and Oatp1 (Slco1a1) were kindly provided by Bruno Steiger; anti-Mrp4 was a gift from Dietrich Keppler5; Oct1 (Slc22a1) antibodies were provided by Hermann Koepsell.21 Antibodies to Oatp2 (Slco1a4) and Mdr were purchased from Santa Cruz Biotechnology (Santa Cruz, CA), and Mrp3 (Abcc3) antibodies were developed by our laboratory (Chen et al., unpublished observations). Antibodies to mouse Ostα and Ostβ were kindly provided by Ned Ballatori.16 All protein samples from livers and kidneys were analyzed as previously described.16 All samples were run together on the same gel/blot using a triple-wide system from C.B.S. Scientific (Del Mar, CA). Densitometric analysis was performed using Multi-Analyst software (Bio-Rad, Hercules, CA).

Statistical Analysis.

All data were analyzed using the Student t test and are expressed as the mean ± SD. A P value of less than .05 was considered significant.


CBDL Results in More Liver Cell Injury in Mrp4-Null Mice.

Liver histology was obtained in control and Mrp4-null male mice 7 days after CBDL (Fig. 1A-B). Visual inspection revealed punched-out lesions of variable size in both groups of animals after CBDL as previously described.11, 22 A blind assessment of the liver tissue revealed a trend toward more liver cell injury from CBDL Mrp4-null mice compared with CBDL wild-type animals. Cell injury, graded on a qualitative 0-4 scale, was 2.14 ± 1.21 for CBDL controls vs 3.0 ± 0.79 for CBDL Mrp4−/− mice (P = .20). No cell injury was noted in sham-operated control livers from either group.

Figure 1.

Liver necrosis and serum liver enzyme levels are increased in Mrp4−/− CBDL mice. Liver histology from (A) Mrp4+/+ CBDL mice (n = 7) and (B) Mrp4−/− CBDL mice (n = 5). Necrotic areas (arrows) were easily distinguished from surrounding parenchyma and were more extensive in the Mrp4−/− mice. (C) Serum aminotranferase (alanine aminotransferase/glutamic pyruvate aminotransferase) levels from Mrp4+/+ and Mrp4−/− sham-operated and CBDL mice (n = 6 for each group). Enzyme levels were significantly higher in serum from the CBDL Mrp4−/− mice. *P < .01. SGPT, serum glutamic pyruvate aminotransferase; Mrp4, multidrug resistance–associated protein 4; BDL, bile duct ligation.

SGPT Is Increased in Mrp4−/− CBDL Mice.

SGPT measurements revealed significant increases in serum from Mrp4−/− mice 7 days after CBDL compared with wild-type mice (Fig. 1C). SGPT levels in wild-type CBDL mice were 41.98 ± 17.8 U/L compared with 187.38 ± 131.99 U/L (P < .01) in Mrp4−/− mice, consistent with histological analyses. Sham concentrations of serum SGPT were 29.1 ± 13.5 and 34.7 ± 14.4 U/L in wild-type and Mrp4−/− mice, respectively. Taken together, these observations indicate that the absence of Mrp4 increases susceptibility to liver cell injury after CBDL.

Serum Bile Acid Levels Are Lower in Mrp4−/− CBDL Mice, but Bilirubin Is Unchanged.

As previously described, serum bile acids were substantially increased in CBDL wild-type animals compared with their sham-operated controls.11, 22 However, they were fourfold lower in CBDL Mrp4−/− mice compared with CBDL wild-type mice, despite the same duration of cholestasis (Fig. 2A). This finding suggests that Mrp4 may have an important functional role in the extrusion of bile acids from the cholestatic liver. In contrast, there were no differences in bile acid concentrations in liver tissue, bile, or urine between the two groups of animals after CBDL for 7 days. Liver concentrations in wild-type CBDL animals were 881.4 ± 80.3 nmol/g liver compared with 907.2 ± 84.9 in −/− CBDL mice. Concentrations in bile were 48.8 ± 33.0 and 49.2 ± 30.4 mmol/L in wild-type and Mrp4−/− mice after CBDL, respectively (P > .05). As expected, urine concentrations of bile acids were negligible in the two sham-operated control groups and increased significantly without differences to 87.3 ± 39.0 and 89.6 ± 126.7 μmol/L in CBDL wild-type and Mrp4−/− mice, respectively.

Figure 2.

Following CBDL, serum bile acids but not bilirubin levels are significantly less in the Mrp4−/− mouse. (A) The increases in serum bile acids after CBDL for 7 days were significantly smaller in Mrp4−/− mice than in wild-type mice. *P < .005. (B) There were no differences in serum bilirubin levels between the two groups (n = 6 for +/+ and −/− sham-operated mice and −/− CBDL mice; n = 7 for +/+ CBDL mice). BDL, bile duct ligation.

There were also no differences in total serum bilirubin levels between wild-type and Mrp4−/− mice 7 days after CBDL (Fig 2B), suggesting that Mrp4 is not a major export pump for bilirubin conjugates in cholestatic liver. Bilirubin levels were undetectable in both groups of sham-operated animals.

Liver Bile Acid Analyses.

High-performance liquid chromatography analysis of bile acids from livers of wild-type (n = 7) and Mrp4−/− (n = 5) mice following 7 days of CBDL revealed significant differences in the composition of the retained species (Table 1). Livers from Mrp4−/− bile duct obstructed animals had significantly less of the major bile acid component, tauro-beta muricholic acid, and significantly more taurine tetrahydroxy bile acids. The relative percentage of other less prominent components, glycine beta muricholic acid, taurocholate and glycocholate, were not significantly different between the two groups of mice.

Table 1. Hepatic Bile Acid Composition
 Wild-Type (n = 7)Mrp4-Null (n = 5)P Value
  • NOTE. Hepatic bile acid composition was analyzed via high-performance liquid chromatography. Significantly less tauro beta-muricholic acid was seen in −/− CBDL livers compared with +/+ livers. Conversely, there were significantly more taurine tetrahydroxy bile acids in the −/− livers.

  • *

    Data are expressed as the percentage of total bile acids extracted.

  • Abbreviation: Mrp4, multidrug resistance–associated protein 4.

Taurine (OH)44.53 ± 2.169.14 ± 3.51.02*
Tauro β muricholic90.51 ± 3.4779.24 ± 7.45.03*
Glycine β muricholic1.3 ± 1.535.56 ± 7.50.51
Taurocholate3.47 ± 2.364.94 ± 4.21.46
Glycocholate0.16 ± 0.421.08 ± 1.23.19

Mrp2, Mrp3, Mrp4, Ntcp, and Bsep mRNA Expression.

Significant increases in Mrp3 mRNA expression were observed in livers 7 days after CBDL in wild-type mice as previously observed11, 22 (Fig. 3). As reported, Mrp4 mRNA was also significantly upregulated in wild-type mice after CBDL.11 Mrp3 mRNA expression was also increased in Mrp4−/− mice, but it was not statistically significant. Ntcp mRNA was significantly downregulated after CBDL (Fig. 3) as described in rats23 and mice.24 However, there were no significant changes in Mrp2 mRNA in any of the experimental groups as described previously in mice,11 in contrast to rat liver following CBDL25, 26 (Fig. 3). Bsep mRNA also remained unchanged in all experimental groups as previously reported after CBDL for 14 days in mice.22 A small increase in Bsep mRNA was reported 7 days after CBDL in mice by others.27

Figure 3.

Quantitative polymerase chain reaction mRNA levels for canalicular and basolateral liver transporters in wild-type and Mrp4−/− mice after 7 days of CBDL (n = 6 for +/+ sham-operated mice and −/− sham-operated and CBDL mice; n = 7 for +/+ CBDL mice). There were no significant differences in Mrp2 between the groups. Mrp3 mRNA was increased compared with sham-operated controls but reached significance only in the +/+ group (*P = .001). Mrp4 message levels were significantly increased in the +/+ CBDL group compared with the sham-operated control group (*P < .01). mRNA levels of Ntcp were significantly decreased in both Mrp4+/+ and Mrp4−/− livers (*P < .002). Levels of Bsep mRNA were unchanged between all groups. CBDL, common bile duct ligation; Mrp, multidrug resistance–associated protein.

Mrp4 Protein Expression Is Increased in the Liver After CBDL in Wild-Type Mice.

Mrp4 protein was significantly increased in the liver following CBDL as previously described in 7-day CBDL mice11 and 14-day CBDL rats6 (Fig. 4A-B). In contrast, Mrp3 liver protein was significantly increased in Mrp4−/− mice despite unchanged Mrp3 mRNA levels, but was unchanged in wild-type CBDL mice despite increases in Mrp3 mRNA. This is also in contrast to our previous findings, in which Mrp3 protein was increased 14 days after CBDL in wild-type (C57BL/6) mice.22 These discrepancies may be related to strain differences and/or the severity of the cholestatic injury after 7 versus 14 days of CBDL. Ntcp, Oatp1, and Oct1 protein expression decreased in both wild-type and Mrp4−/− CBDL mice as previously reported in CBDL rats21, 23, 26 and mice,28, 29 although Ntcp reached statistical significance only in the Mrp4−/− group (Fig. 4A-B). Mdr1 and Oatp2 protein expression increased in both wild-type and Mrp4−/− CBDL mice as expected,29 while Bsep protein expression remained unchanged. Mrp2 showed no change in protein expression in wild-type CBDL mice, but increased significantly in Mrp4−/− bile duct–ligated mice. However, the Mrp4−/− sham levels of Mrp2 mice were significantly lower than those of wild-type sham mice; thus, CBDL restored Mrp2 levels only to “normal” basal levels in the Mrp4−/− mice.

Figure 4.

(A) Representative Western blots of liver transporter protein expression in wild-type and Mrp4−/− sham-operated and CBDL mice. (B) Densitometry analysis. Data are expressed in arbitrary density units ± SD (n = 6 for +/+ and −/− sham-operated mice; n = 7 for +/+ CBDL mice; n = 5 for −/− CBDL mice). Mrp2: *P < .0001 versus −/− sham-operated mice. Mrp3: *P < .0005 versus −/− sham-operated mice. Mrp4: *P = .0005. Mdr: *P < .005 versus respective sham-operated animals. Ntcp: *P < .0001 versus −/− sham-operated mice. Oct1 and Oatp1: *P < .0001 versus respective sham-operated controls. Oatp2: *P < .001 versus respective shams. BDL, bile duct ligation; Mrp, multidrug resistance–associated protein; CBDL, common bile duct ligation.

Renal Mrp mRNAs Increased in Mrp4 CBDL Mice.

Seven days after CBDL, renal Mrp3 mRNA was increased in Mrp4−/− and wild-type mice, although the difference was only significant in the Mrp4−/− group (Fig. 5A). In contrast, no change in Mrp2 mRNA was observed in the kidneys of wild-type or Mrp4−/− mice. However, renal Mrp4 mRNA increased significantly after CBDL in wild-type mice (Fig. 5A).

Figure 5.

(A) Quantitative polymerase chain reaction for Mrp2, Mrp3, Mrp4, Ostα, and Ostβ in mouse kidney (n = 6 for +/+ and −/− sham-operated mice and −/− CBDL mice; n = 7 for +/+ CBDL mice). Mrp3−/− CBDL: *P = .004 versus −/− sham-operated mice. Mrp4: *P = .01. Ostα and Ostβ mRNAs were significantly increased in both wild-type and Mrp4−/− mice (CBDL: *P < .01 vs. respective sham-operated mice; #P = .01 vs. +/+ sham-operated mice; ##P < .05 vs. +/+ CBDL mice). (B) Representative Western blots for Mrp2, Mrp3, Mrp4, Ostα, and Ostβ and (C) corresponding densitometry. Mrp2: *P = .04 versus −/− sham-operated mice. Mrp3: *P = .03 versus −/− sham-operated mice. Ostα protein was increased only in Mrp4−/− mice (*P < .02), whereas Ostβ protein was more than doubled in both groups (*P < .04). CBDL, common bile duct ligation; Mrp, multidrug resistance–associated protein; BDL, bile duct ligation.

CBDL Increases Mrp2 Protein in the Kidneys of Mrp4−/− Mice.

Although Mrp2 protein was increased significantly in the kidneys of Mrp4−/− mice, it did not reach statistical significance in wild-type mice after CBDL, unlike in the rat30, 31 (Fig. 5B-C). This suggests that the cholestatic insult at 7 days was more severe in the Mrp4−/− mice, resulting in a greater stimulus for adaptive responses in the kidney than in the wild-type cholestatic animals. The increases in Mrp3 mRNA in the kidneys of both wild-type and Mrp4−/− mice after CBDL was not reflected in increases in protein expression. In contrast, there was a small but significant decrease in Mrp3 protein in the kidneys of Mrp4−/− CBDL mice. Despite an increase in mRNA, protein levels for Mrp4 were unchanged after 7 days of CBDL.

Adaptive Responses in Ostα and Ostβ Expression.

In the liver, Ostα mRNA expression was significantly decreased in wild-type mice and was unchanged in Mrp4−/− mice after CBDL (Fig. 6A). Ostα protein significantly increased in both groups after CBDL (P < .01) (Fig. 6B). In contrast to Ostα, Ostβ mRNA expression increased in the livers of wild-type mice after CBDL (P < .001). Interestingly, the sham-operated Mrp4−/− animals had increased levels of Ostβ mRNA compared with the wild-type shams (P < .0001) (Fig. 6A). However, no further increase in Ostβ expression occurred after CBDL in Mrp4−/− mice. Ostβ protein was undetectable in mouse liver in all groups.

Figure 6.

(A) Ostα and Ostβ expression in mouse liver. Ostα: P < .02 versus +/+ CBDL mice, P < .03 versus −/− sham-operated mice. Ostβ is significantly increased after CBDL in wild-type mice (P = .0005), sham-operated Mrp4−/− mice (P < .0001), and CBDL Mrp4−/− mice (*P = .003) (n = 6 for +/+ sham-operated mice and −/− sham-operated and CBDL mice; n = 7 for +/+ CBDL mice) compared with sham-operated wild-type mice. (B) Ostα protein expression is significantly increased in both wild-type (n = 6; *P = .0006) and Mrp4-null mice (n = 6 for sham-operated mice; n = 5 for CBDL mice; *P < .01) compared with their respective sham-operated control mice. CBDL, common bile duct ligation; BDL, bile duct ligation.

In the kidneys, where Ostα and Ostβ are more highly expressed than in the liver, Ostα and Ostβ mRNA levels increased after CBDL in both wild-type and Mrp4−/− mice (Fig. 5A). Ostα protein increased only in the Mrp4−/− mice after CBDL, whereas the beta subunit protein was significantly increased in both the wild-type and Mrp4−/− mice. (Fig. 5B-C).


There is accumulating evidence that several bile acid and other organic solute transporters undergo adaptive regulation in the liver, kidneys, and intestine in response to cholestatic liver injury.32, 33 These adaptive responses, which have been identified in cholestatic animal models and in humans with cholestasis, occur at both transcriptional and posttranscriptional levels and are presumed to minimize the hepatic and systemic accumulation of bile acids and other potentially toxic substances, thus reducing tissue damage. These adaptations include the downregulation of the major liver conjugated bile acid uptake transporter and certain sinusoidal/basolateral organic anion and cation uptake systems (e.g., Ntcp, Oatp1, Oct1). These adaptive responses in transporter expression reduce the uptake of bile acids, small organic cations, and other organic anionic substances from portal blood. At the same time, there is an adaptive expression of several alternative ABC transporters on the basolateral membrane of hepatocytes that undergo significant upregulation, including MRP3/Mrp3 and MRP4/Mrp4. These proteins are believed, largely on the basis of in vitro studies, to participate in the extrusion of conjugated bile acids, bilirubin, and other organic solutes. Ostα-Ostβ may also contribute to bile acid extrusion across the basolateral membrane of hepatocytes and cholangiocytes, particularly in human hepatocytes, where this heterodimeric transporter is more highly expressed than in murine liver.16, 17 However, the relative importance of these basolateral transporters and their own particular role in responding to cholestatic stress in the liver are not well understood.

To address this question, we examined the effect of CBDL on the expression of a variety of these liver and renal transporters in the Mrp4−/− mouse where upregulation of Mrp4 cannot occur. Our findings confirm several effects on transporter expression and provide the following insights into the role of Mrp4 in adapting to the cholestatic response.

Mrp4−/− Mice Develop More Extensive Liver Parenchyma Cell Injury Than Wild-Type Mice When Subjected to 7 Days of Common Bile Duct Obstruction.

This finding, based on both histological and liver serum enzyme (SGPT) analyses, indicates that the cholestatic phenotype is more severe in the absence of this ABC efflux pump; therefore, the induction of Mrp4 is an important adaptive protective mechanism in cholestatic liver injury. Furthermore, by inference, other adaptive changes in bile acid transporter expression—including upregulation of Mrp3 and Ostα-Ostβ—are not able to fully compensate for the loss of Mrp4 after CBDL in the mouse.

Serum Bile Acid Levels Are Significantly Lower in 7-Day CBDL Mrp4−/− Mice Than in Wild-Type Mice.

Steady-state serum bile acid levels result from a combination of mechanical regurgitation of conjugated bile acids from the obstructed biliary system, transporter mediated efflux from hepatocytes, and clearance from the circulation by extrahepatic mechanisms, particularly glomerular filtration and renal tubular excretion.30, 34, 35 Assuming that CBDL results in equivalent degrees of bile acid regurgitation and renal clearance in both wild-type and Mrp4−/− mice, the lower levels of serum bile acids seem most likely to be due to a reduced ability of Mrp4−/− mice to export bile acids from hepatocytes back into the systemic circulation. (This assumption discounts any role that renal Mrp4 might play in bile acid excretion, which, in its absence, would tend to minimize the serum differences.) This conclusion seems even more justified given that serum bilirubin levels were similar in both groups of animals after 7 days of CBDL. Thus, both animal groups were equally able to efflux serum bilirubin into the circulation. This is exactly opposite to two recent reports where similar studies were performed in CBDL Mrp3−/− mice.36, 37 These studies resulted in lower serum levels of bilirubin in the Mrp3−/− mice, whereas serum bile acid levels and liver histology were not significantly different from the wild-type CBDL control group. Together with our findings, these results strongly suggest that bilirubin conjugates, particularly glucuronide conjugates, are the preferred substrate for Mrp3, whereas bile acid conjugates are the preferred substrate for Mrp4 in the cholestatic liver and that Mrp4, rather than Mrp3 or Ostα-Ostβ, is the predominant alternative cytoprotective basolateral transporter in this murine model. This latter conclusion is also consistent with findings in patients with progressive familial intrahepatic cholestasis, where Mrp4 but not Mrp3 is significantly upregulated.12 Nevertheless, it is likely that Ostα-Ostβ plays a more prominent role in human liver rather than mouse liver, because this heterodimeric transport protein is more highly upregulated at the basolateral membrane of cholestatic human hepatocytes relative to changes in expression of Mrp3 and Mrp4.17

Taurine Tetrahydroxy Bile Acid/Beta-Muricholic Acid Ratios Are Increased Twofold in Mrp4−/− Mice Compared With Wild-Type Mice After CBDL.

These differences in liver bile acid composition following CBDL are also consistent with greater cholestatic injury in Mrp4−/− mice, because the formation of tetrahydroxy bile acids represents an alternative pathway for detoxifying bile acids by phase 1 metabolism in the endoplasmic reticulum. Formation of tetrahydroxy bile acids has also been described in the Bsep−/− mouse38 and in children with neonatal cholestasis.39 Fxr−/− mice also demonstrate enhanced hydroxylation reactions following CBDL that are believed to be catalyzed by Cyp3a11.9, 40, 41

Hepatic Mrp3 Protein Expression Increased Significantly Only in the Mrp4−/− Mice, but Not in CBDL Wild-Type Mice, While Ostα Protein Increased in Both.

Previous CBDL studies after 14 days, rather than 7 days, in both mice and rats found significant increases in Mrp3 mRNA and protein.22, 42, 43 Similar findings have been reported even after 7 days of bile duct ligation in both control and Fxr−/− mice.11 Differences in mouse strains in these studies might contribute to these discrepancies in Mrp3 expression following CBDL. OST/Ostα-OST/Ostβ expression also increases in response to CBDL (as confirmed in the present study) in the mouse and in patients with primary biliary cirrhosis.17 Thus, induction of Mrp3 and OST/Ostα-OST/Ostβ is a reproducible finding in the cholestatic liver. However, our finding that there is more advanced cholestatic injury in Mrp4−/− mice after CBDL emphasizes that neither Mrp3 nor Ostα-Ostβ can fully compensate for the absence of Mrp4, even though both are fully capable of transporting bile acid conjugates.3, 36, 37, 44, 45

Mrp2 mRNA levels in wild-type mice were not significantly different from sham-operated controls, which is consistent with earlier reports at 5 and 7—but not 14—days after CBDL in the mouse.11, 27 However, Mrp2 protein was significantly greater in the Mrp4−/− CBDL mice compared with the Mrp4−/− sham-operated mice, possibly reflecting an attempt to extrude bile acid conjugates into bile.46 This finding might account for the lack of difference in hepatic bile acid levels in Mrp4−/− and wild-type CBDL mice. As in previous reports, there was no change in Bsep mRNA or protein in either group after CBDL,11, 27 confirming that this important bile acid transporter continues to be expressed and extrudes bile acids into the obstructed biliary tree as evidenced by unchanged levels of bile acids in the bile of both groups.

Renal Ostα and Ostβ mRNA and Protein Expression Are Upregulated in Wild-Type and Mrp4−/− Mice After Bile Duct Ligation.

These findings are in contrast to the Mrp responses in which only Mrp4 mRNA is upregulated in wild-type mice and Mrp3 mRNA is upregulated in Mrp4−/− mice. However, Mrp2 is the only Mrp protein that is increased in Mrp4−/− mice. These findings differ from those in rats where renal Mrp2 protein is significantly increased after CBDL, presumably by posttranscriptional mechanisms.30, 31 However, cholic acid feeding in Fxr+/+ and Fxr−/− mice increases Mrp2 mRNA in the kidneys.11 Again, the lack of significant increases in Mrp2 protein in the kidneys of wild-type CBDL mice in the present study may reflect a less severe degree of cholestasis. As described elsewhere, bile acids are normally absent from the urine as a result of efficient hepatic clearance and reabsorption of bile acids in the proximal tubule by the apical sodium-dependent bile acid transporter Asbt (Slc10a2)30 on the luminal membrane and the heterodimeric transporter Ostα-Ostβ47–49 on the basolateral membrane.16 The present study does not clarify the role of Mrp4 in the renal excretion of bile acids, because similar concentrations of bile acids were measured in urine in both groups after CBDL. Mrp4 is located on the apical membrane of the rat proximal tubule together with Mrp2.6, 7 The increase in Mrp2 protein in Mrp4−/− mice after CBDL suggests that Mrp2 may assume some of Mrp4's function in this situation and increase renal bile acid excretion. The role of Mrp3 in the cholestatic kidney remains unexplained.

In conclusion, the findings in this study demonstrate that: (1) the induction of Mrp4 in mouse liver during obstructive cholestasis plays a vital role in protecting the tissue from damage presumably induced by the accumulation of bile acids, and (2) in the absence of Mrp4, the adaptive responses of other basolateral membrane transporters in the liver, including downregulation of Ntcp, Oatp1, and Oct1 and upregulation of Mrp3 and Ostα-Ostβ, are unable to fully compensate, resulting in a higher degree of tissue damage. This suggests that Mrp4 has preferred substrate specificity, most likely for conjugated bile acids. Although human disease specifically related to Mrp4 mutations or impairment has not yet been described, the present study of Mrp4−/− mice provides valuable insights into its function in response to cholestatic liver injury.


The authors thank Alan Hofmann and Ping Lam for helpful discussions regarding bile acid composition and analysis.