This paper was presented in part at the Digestive Disease Week, American Gastroenterological Association, New Orleans, LA, 1998, and at the Annual Meeting of the German Society for Internal Medicine, and published as abstracts in Gastroenterology 1998;114:A527 and Medizinische Klinik 2000;AI:69.
Previously, we identified needle-like and filamentous, putatively “anhydrous” cholesterol crystallization in vitro at very low phospholipid concentrations in model and native biles. Our aim now was to address whether spontaneous gallstone formation occurs in Mdr2 (Abcb4) knockout mice that are characterized by phospholipid-deficient bile. Biliary phenotypes and cholesterol crystallization sequences in fresh gallbladder biles and non-fixed liver sections were determined by direct and polarizing light microscopy. The physical chemical nature and composition of crystals and stones were determined by sucrose density centrifugation and before mass and infrared spectroscopy. Gallbladder biles of Mdr2−/− mice precipitate needle-like cholesterol crystals at 12 weeks of age on chow. After 15 weeks, more than 50% of Mdr2−/− mice develop gallbladder stones, with female mice displaying a markedly higher gallstone-susceptibility. Although gallbladder biles of Mdr2−/− mice contain only traces (≤ 1.1 mM) of phospholipid and cholesterol, they become supersaturated with cholesterol and plot in the left 2-phase zone of the ternary phase diagram, consistent with “anhydrous” cholesterol crystallization. Furthermore, more than 40% of adult female Mdr2−/− mice show intra- and extrahepatic bile duct stones. In conclusion, spontaneous gallstone formation is a new consistent feature of the Mdr2−/− phenotype. The Mdr2−/− mouse is therefore a model for low phospholipid-associated cholelithiasis recently described in humans with a dysfunctional mutation in the orthologous ABCB4 gene. The mouse model supports the concept that this gene is a monogenic risk factor for cholesterol gallstones and a target for novel therapeutic strategies. (HEPATOLOGY 2004;39:117–128.)
The multidrug resistance gene 2 (Mdr2) encodes the hepatic canalicular transporter for the major biliary phospholipid phosphatidylcholine (lecithin). This protein belongs to the ATP-binding cassette (ABC) transporter family, and accordingly the official gene symbol Abcb4 has been assigned. Mice with homozygous disruption of the Abcb4 gene (commonly denoted Mdr2−/−- mice) display almost complete absence of phospholipids from bile.1 The phospholipid-deficiency of Mdr2−/− mice2, 3 results in liver injury from chronic cholangitis, resembling human liver disease due to mutations of the orthologous gene ABCB4.4ABCB4 mutations result in a wide spectrum of phenotypes, ranging from progressive familial intrahepatic cholestasis (PFIC type 3) to adult cholestatic liver disorders characterized by elevated γ-GT levels.4 Under healthy conditions, mixed cholesterol-lecithin-bile salt micelles protect the biliary epithelium against the detergent properties of bile salts. The cholestatic liver injury in these patients is therefore commonly attributed to the “toxic” effects of bile salts on the apical membranes of hepato- and cholangiocytes.5 The liver injury in Mdr2−/− mice is characterized by segmental biliary strictures due to periductal fibrosis and fibro-obliteration of bile ducts.6
Biliary lecithin also plays a key role in solubilizing excess cholesterol in the form of unilamellar cholesterol-lecithin-vesicles. In agreement with this paradigm, Rosmorduc et al.7, 8 recently described a clinical entity characterized mainly by the occurrence of intrahepatic and gallbladder microlithiasis in young adults associated with ABCB4 mutations. This “peculiar” form of cholelithiasis was termed low phospholipid-associated cholelithiasis (LPAC). In earlier work, we had identified needle-like and filamentous cholesterol crystallization, apparently “anhydrous” in in vitro studies of model biles9, 10 and native biles with very low phospholipid/bile salt ratios.11, 12 Therefore, we hypothesized here that similar crystallization phenomena would be observed in phospholipid-deficient bile of Mdr2−/− mice in vivo and lead to gallstones. We compared the physical-chemistry of fresh hepatic and gallbladder biles between Mdr2−/− and wild-type mice. Because spontaneous gallstone formation, without a dietary challenge, had not been reported in previous studies of this animal model,1–3, 13–16 we carried out a systematic microscopic as well as physical-chemical analysis of the biliary phenotypes, identifying gallstone formation as an important feature of Mdr2−/− mice.
For high performance liquid chromatography (HPLC), bile salt standards were purchased from Sigma Chemical (St. Louis, MO) and CalBiochem-Behring (San Diego, CA), with the exception of tauro-β-muricholate and tauro-ω-muricholate, which were generous gifts from Tokyo Tanabe (Tokyo, Japan). The purities of individual bile salts were at least 98% by thin layer chromatography (TLC) or HPLC. Dilauroyl-phosphatidylcholine [PC-(24:0)], dimyristoyl-PC [PC-(28:0)], diarachidoyl-PC [PC-(40:0)], and dibehenoyl-PC [PC-(44:0)] standards for mass spectrometry were obtained from Sigma Chemical with purities of at least 99% by TLC. N-oleoyl-D-sphingomyelin (18:1) was 95% pure by TLC (Sigma Chemical). HPLC reagents and all other chemicals used were of the highest purity commercially available and purchased from Sigma or Fisher Scientific (Pittsburgh, PA).
Gene-Targeted Mice and Diet
Breeding pairs of homozygous Mdr2−/− gene-targeted mice (official name FVB/N-Abcb4tm1Bor) and Mdr2+/+ wild-type mice on the FVB/NJ background were obtained from The Jackson Laboratory (Bar Harbor, ME). Mice were housed as groups of 5 per cage in a temperature-controlled room (22 ± 1°C) that was illuminated from 6.00 AM to 6.00 PM.
Genotypes of Mdr2−/− mice were confirmed by polymerase chain reaction (PCR) using neomycin specific primers (5′-CTT GGG TGG AGA GGC TAT TC-3′; 5′-AGG TGA GAT GAC AGG AGA TC-3′) and 10 to 20 ng genomic DNA isolated from individual tail snips by the rapid lysis technique (DNeasy Tissue Kit, Qiagen AG, Hilden, Germany). The PCR reactions contained 2 mM MgCl2, 10 mM dNTPs, 200 nM primers, and 1 U Taq DNA polymerase (Roche Diagnostics, Mannheim, Germany) in 50 μL. PCR cycling conditions were 95°C/30 s, 55°C/60 s, and 72°C/30 s for 35 cycles, and a final extension step of 10 min at 72°C.
The mice were weaned at 3 weeks of age and maintained on laboratory chow diet (Purina 5001, St. Louis, MO) that contains less than 0.02% (wt/wt) cholesterol, as determined by HPLC.17 After 6, 9, 12, and 15 weeks on chow diet (i.e., at 9, 12, 15, and 18 weeks of age, respectively), mice were anesthetized (10 mg Avertin IP, Sigma-Aldrich), and laparotomy and cholecystectomy were performed at 9.00 AM after overnight fasting.12 Gallbladder volumes were determined gravimetrically, assuming a density of 1 g/ml bile.12 In a second group of mice, the lower end of the common bile duct was ligated, the common bile duct was cannulated with a PE-10 polyethylene catheter (Becton Dickinson, Sparks, MD) below the entrance of the cystic duct, which was doubly ligated, and hepatic bile was collected by gravity.18
For routine biochemistry, serum samples were stored at −70°C until analysis of aspartate aminotransferase (AST), alanine aminotransferase (ALT), alkaline phosphatase (AP), and bilirubin by routine clinical chemistry testing on a Hitachi 717 analyzer (Boehringer Mannheim, Mannheim, Germany).
Mice received care according to the “Guide for the Care and Use of Laboratory Animals” prepared by the National Academy of Sciences (NIH publication 86-23). Protocols were approved by the Animal Care and Use Committees of Harvard University and Aachen University. Mice were monitored routinely for selected infectious agents, and gallbladder bile samples were cultured to exclude bacterial infection. As expected for rodents, all mice showed progressive weight gain from 20 to 26 gm at 9 weeks of age to 24 to 30 gm by the end of the experiments.1, 2
For detection of cholesterol crystals and quantification of gallstone number and size, the gallbladder of each mouse was opened at the fundus. The observers (D.Q.-H.W., F.L.) were unaware of the animals' genotypes. Fresh gallbladder biles were then observed using direct and polarizing light microscopy (Leitz microscope DM RB, Leica Microsystems, Wetzlar, Germany) without a cover slip.10, 12 After compression with a cover glass, phase contrast and Nomarski differential interference optics were also employed. Liquid crystals and solid cholesterol crystals were defined according to previous criteria.10–12 Also, 5 μL of fresh hepatic biles were examined microscopically (n = 5 per group). In addition, bile samples were incubated at 37°C under argon for up to 7 days and followed by sequential microscopic examination.
For liver histology, tissue specimens were fixed in 10% formalin, and paraffin sections were stained with hematoxylin and eosin (HE), or 0.1% Sirius red F3B in saturated picric acid (Chroma, Münster, Germany).19 For visualization of intrahepatic cholesterol crystals, tissue was cryostat-sectioned at 5-μm thickness, rinsed in water, and mounted. Crystals were visualized by polarizing light microscopy.
Gallstone composition was analyzed by infrared spectroscopy, as described previously.20, 21 Gallstones were finely ground with an agate mortar and pestle, and compressed with spectral grade KBr at 8 t/cm2 to form thin wafers. Infrared spectra between 4,000 and 400 cm-1 were recorded for 8 to 12 min and evaluated manually and by computer-assisted comparison with previously recorded spectra.20, 22
Sucrose Density Gradient Centrifugation
Cholesterol crystals were concentrated by ultracentrifugation using a tabletop ultracentrifuge at room temperature, layered on top of linear 2 to 20% sucrose density gradients, and centrifuged for 1 h at 201,000 g.23, 24 Fractions were aspirated from top to bottom and examined by polarizing light microscopy, as described above. Sucrose densities were verified using a temperature-compensated refractometer.
Cholesterol monohydrate and anhydrous cholesterol crystals were used as standards. To prepare cholesterol monohydrate crystals,25 cholesterol (Nu-Chek Prep Inc., Elysian, MN) was recrystallized 3 times from hot 95% (vol/vol) ethanol. The final solution was cooled to room temperature, and the cholesterol monohydrate crystals were harvested by filtration, washed repeatedly with water, and dried overnight in a dessicator thermostated at 40°C. For anhydrous cholesterol crystals,25 cholesterol was dissolved in hot (70°C) glacial acetic acid and the solution was allowed to cool. The crystals were harvested as described above and placed in a heated dessicator (80°C) overnight under vacuum to remove all remaining acetic acid.
Biliary Lipid Analysis
Lipids were extracted according to Bligh and Dyer.26 Total biliary phospholipid concentrations were measured as inorganic phosphorus by the method of Bartlett27 as well as enzymatically using phospholipase D and choline oxidase,28 employing an assay kit from Wako Chemicals GmbH (Neuss, Germany). Cholesterol and bile salt concentrations were determined using cholesterol oxidase and 3α-hydroxysteroid dehydrogenase, respectively, as described previously.12 Cholesterol saturation indexes (CSIs) of bile samples were calculated from the critical tables.29 Individual bile salts were determined by HPLC.12
Measurement of Biliary Phospholipid Molecular Species
Biliary lipid extracts from 10 μl of gallbladder bile were taken to dryness under N2, redissolved in 20 to 100 μl methanol/chloroform (2:1, vol/vol), and centrifuged in a tabletop centrifuge at 16,000 rpm for 5 min. Samples were processed exclusively in glassware to avoid contamination with plasticizers and related materials. Mass spectrometric analyses were performed with the triple quadrupole instrument TSQ 7000 (Finnigan MAT, Bremen, Germany) equipped with a nanoelectrospray source, as described previously.30, 31 After single-stage positive ion scans, PC and sphingomyelin molecular species were selectively analyzed in the tandem mode by (i) scanning for neutral loss of the choline phosphate head group (183 Dalton) from cationized PC or sphingomyelin molecules ([M+Na]+ adducts) and (ii) precursor scanning at m/z 184, which detects [M+H]+ ions of choline-containing phospholipids only.31 For each spectrum, 50 to 150 repetitive scans were averaged. To correct for the influence of the fatty acid chain length on the signal intensity of PC, a set of standards [PC-(24:0), PC-(28:0), PC-(40:0), PC-(44:0)] was used for calibration according to Brügger et al.31
Data are expressed as means ± SE. Four to 8 animals from each group were studied at each time point. For evaluation of statistical significance, differences were assessed by Student's t-test for independent samples. Statistical significance was defined as a 2-tailed probability P less than .05 unless otherwise stated. Analyses were performed with SPSS 10 software (SPSS, Chicago, IL).
Gallbladder Bile of Mdr2−/− Mice Precipitates Unusual Solid Crystals
Upon laparotomy, careful macroscopic and microscopic examination of the biliary system of female Mdr2−/− mice at 12 weeks of age and males at 15 weeks of age revealed the presence of needle-shaped crystals and stones in the gallbladder. No crystals or stones were observed before this age. Of note is that the bile of Mdr2−/− mice did not contain helical or tubular cholesterol crystals,10–12 and cholesterol monohydrate crystals, which are characterized by a plate-like shape and often display a notched corner, were not observed either.25 Figure 1 shows photomicrographs of typical gallstones. Individual needle-like crystals project from the edges of the stones, which also contain amorphous mucin gel (Fig. 1A). When the stones were broken by a slide cover, we found the core of stones to be composed of needle-like crystals and gelled mucin (Fig. 1B). The needle-like crystals are short, straight filamentous crystals (<20 μm). Figure 2 displays representative microscopic images of these crystals with Nomarski differential interference optics, indicating that their morphology resembles the filamentous and needle-like cholesterol crystals described previously by our group in model biles of similar low phospholipid composition.9, 10
Time Course and Gender Difference of Cholesterol Crystallization and Gallstone Formation
Figure 3 shows the prevalence of needle-like crystals and stones as functions of age for female (left panel) and male (right panel) Mdr2−/− mice. At 9 weeks of age, the macroscopic and microscopic examination of gallbladder biles shows no evidence of crystals or gallstones. However, mucin gel is detected in 20% of female Mdr2−/− mice. At 12 weeks, gallbladder biles of female Mdr2−/− mice contain needle-like crystals (33%) as well as gallstones (50%), and stone prevalence increases strongly with passage of time. At 15 weeks, 60% of female Mdr2−/− mice display needle-like crystals and 80% show gallstones. By 18 weeks, all female mice develop stones (Fig. 3, left panel). In male Mdr2−/− mice, needle-like crystals and gallstones form later (≥ 15 weeks) and at lower prevalence rates (Fig. 3, right panel) compared to female mice.
As expected, gallbladder biles of Mdr2+/+ (wild-type) mice are microscopically free of crystals or stones. In addition, we tested heterozygous Mdr2+/− mice at 12 and 15 weeks of age (n = 10–14), whose gallbladder biles do not precipitate crystals either. Furthermore, all gallbladder biles of both Mdr2+/+ and Mdr2+/− mice remain crystal-free when followed by sequential microscopic examination during incubation at 37°C for 7 days.
Number, Size, and Composition of Gallstones
Figure 4 shows the average gallstone numbers and diameters in gallbladders of female and male Mdr2−/− mice. The number of gallstones per mouse ranges from 1 to 8 with an average of 1.3 ± 0.3 stones per mouse. Most gallstones are smaller than 0.1 mm, with a mean stone size of 0.15 ± 0.04 mm. On average, both stone number and size are significantly higher in female compared to male mice.
Ultracentrifugation of gallbladder bile samples from Mdr2−/− mice on sucrose density gradients yields needle-like crystals at intermediate densities (d = 1.03–1.04 gm/ml) that maintain their crystal habits in water for up to 48 h. Standard needle-like anhydrous cholesterol and plate-like cholesterol monohydrate crystals (see Materials and Methods) result in 2 bands at d = 1.03 and d = 1.05 gm/ml, respectively. This indicates that the crystals harvested from Mdr2−/− mice may be anhydrous cholesterol, which undergoes subsequent hydration as water becomes incorporated within the crystal lattice.10, 23, 32 This phase transition has been confirmed recently by grazing incidence x-ray diffraction analysis of ultrathin cholesterol films on water.33
However, gallbladder stones of Mdr2−/− mice demonstrate densities greater than 1.05, consistent with stones containing additional compounds beside cholesterol, particularly bilirubin with d = 1.31 gm/ml.34 Indeed, trace quantities of unconjugated bilirubin are detected by mass spectroscopy (m/z 584) in gallstones of Mdr2−/− mice. However, infrared spectroscopy reveals pronounced bands at 1,032, 602, and 562 cm−1, demonstrating calcium phosphate to be the most likely major inorganic stone compound (80%). The predominant calcium salt was likely to be carbonate (apatite), as indicated by the typical absorption bands at 1,449, 1,415, and 873 cm−1.20
Biliary Lipid Compositions
Table 1 shows the biliary lipid compositions of gallbladder biles from Mdr2+/+ and Mdr2−/− mice of both genders on chow. As has been shown for other inbred mouse strains,12 the gallbladder biles of wild-type mice are not supersaturated with cholesterol with CSIs less than 1.00. The predominant phospholipid class in wild-type mice is phosphatidylcholine (PC), whereas sphingomyelin (SM) comprises only 2% of total phospholipids (Table 1). As expected for mice with homozygous disruption of the PC transporter Mdr2,3, 16 only minimal amounts of PC (<3% of controls) are detected in biles of Mdr2−/− mice. This decrease is not compensated for by SM, with SM levels being low and essentially identical to those in wild-type mice.
Table 1. Biliary Lipid Compositions of Gallbladder Biles
In Mdr2+/+ mice, PC and SM concentrations were determined by ESI-MS/MS. SM concentrations were calculated based on the 4 major species detected (16:0, 22:0, 24:0, 24:1).
For Mdr2+/+ mice, CSI values were calculated using the critical tables,29 and thus are estimates based on taurocholate; for Mdr2−/− mice, calculations were based on the bile salt pool compositions of Mdr2−/− mice (see Table 2) and the maximal cholesterol-solubilizing capacities of micellar solutions of the two principal taurine conjugated bile salts.35, 36
**, d, ‡
*P < 0.05; **P < 0.01, compared to Mdr2+/+ mice; ‡P < 0.05, compared to 9 weeks.
In 18-weeks-old male Mdr2+/+ mice, analysis of biliary lipids in the isotropic phase of gallbladder bile was not feasible due to the development of minuscule gallbladders (mean volume 2 ± 1 μl) filled with sticky mucin gel and needle-like crystals (see Results on gallbladder volumes).
Abbreviations: BS, bile salts; Ch, cholesterol; CSI, cholesterol saturation index; F, female; M, male; Mdr2, multidrug resistance gene 2; ND, not determined; PC, phosphatidylcholine; PL, phospholipids; SM, sphingomyelin; [TL], total lipid concentration.
Although biles of Mdr2−/− mice contain traces only of cholesterol (Table 1), they are supersaturated with cholesterol (CSI ≥1.05) in 12- to 15-week-old mice, when calculated for the maximal cholesterol-solubilizing capacities of micellar bile salt solutions in the virtual absence of PC.35, 36 Because different species and compositions of bile salts could influence cholesterol solubility, we studied model biles of cholesterol-lecithin-taurocholate-tauro-β-muricholate, which generated 2 new phase diagrams with a mixture of taurocholate-tauro-β-muricholate in a ratio of 1:1 (wt/wt) according to the analyzed bile salt species (Table 2) and the average total lipid concentrations (∼10 and 15 gm/dl) of the pooled gallbladder biles of Mdr2+/+ (Fig. 5, top panel) and Mdr2−/− mice (bottom panel).10, 12, 37 For purposes of illustration, we plot in Fig. 5 the relative biliary lipid compositions of Mdr2+/+ mice (top panel) and Mdr2−/− mice (bottom panel) on condensed ternary cholesterol-lecithin-taurocholate-tauro-β-muricholate phase diagrams. Also, we studied mixtures down to ∼0.4% phospholipids, and defined phase boundaries for biles with very low phospholipid/bile salt ratios (see the inset of Fig. 5). Gallbladder biles of Mdr2+/+ mice plot in the micellar zone, that is, that they are not saturated with cholesterol (CSI ≤0.53) and do not precipitate crystals. The cholesterol supersaturated biles of Mdr2−/− mice plot in region A of the left 2-phase zone. By phase analysis in model bile systems,10 biles in region A are predicted to contain needle-like, putatively “anhydrous” cholesterol crystals, exactly as observed in our experimental analysis (Figs. 2 and 3). These findings explain physical-chemically why Mdr2−/−- mice develop gallstones composed of an unusual crystalline cholesterol habit. Furthermore, during the experimental time course, biliary cholesterol concentrations and CSIs increase and PC concentrations decrease markedly (Table 1), consistent with onset of stone formation after 12 weeks of age.
Table 2 summarizes the relative bile salt compositions of pooled gallbladder biles (n = 10) of Mdr2−/− and wild-type mice. Tauro-β-muricholate and taurocholate are the predominant bile salts, comprising more than 97% of the bile salt pool. Of note, gallbladder bile of female Mdr2−/− mice contains higher taurocholate and lower tauro-β-muricholate levels than bile of male knockout mice, which is reflected by a markedly higher hydrophobicity index (HI) of the bile salt pool (−0.24 vs. −0.51). For both genders, bile salt compositions do not differ between wild-type and knockout mice.
Table 3 tabulates the molecular species of PC in gallbladder biles of Mdr2−/− and wild-type mice, as determined by electrospray ionization tandem mass spectrometry (ESI-MS/MS). For the major biliary PC species, no marked sex differences are observed. Three major PC species are present in wild-type mice: 16:0 to 18:2 (51.4%), 16:0 to 18:1 (19.7%), and 16:0 to 20:4 (∼10%). In addition, minor amounts of a variety of other PC species are detected. In contrast to Mdr2+/+ mice, bile of Mdr2−/− mice contains trace amounts of all PC species. Compared to wild-type mice, the relative compositions shift to greater proportions of the more hydrophobic PC-(18:0–18:2) and PC-(18:0–20:4) in knockout mice, which is reflected by a higher PC hydrophobicity index (0.36 vs. 0.23).
Table 3. Phosphatidylcholine Molecular Species in Gallbladder Biles
The gallbladder volumes of Mdr2+/+ mice range from 2 μl to 10 μl and are similar to values obtained for other inbred strains of mice.12, 17 At 9 weeks of age, gallbladders of male and female Mdr2−/− mice are 3 times larger than gallbladders of wild-type mice (9 ± 1 μl vs. 3 ± 1 μl, P <.01). With passage of time, gallbladder sizes decrease significantly (P <.01) in male Mdr2−/− mice from 11 ± 1 μl at 9 weeks to 2 ± 1 μl at 18 weeks of age, with several old male mice showing very small gallbladders filled with sticky mucin gel and needle-like crystals. Gallbladder sizes of female Mdr2−/− mice (12 ± 1 μl) are significantly (P <.01) larger compared to male mice and do not decrease with age.
Mature Mdr2−/− Mice Develop Bile Duct and Intrahepatic Stones
Figure 6 illustrates that Mdr2−/− mice of both sexes display enlarged, but otherwise normal livers. However, as illustrated in Fig. 7, histologic assessment of liver injury reveals characteristic changes. Whereas parenchymal damage is mild, Mdr2−/− mice display marked proliferation of bile ducts, which is accompanied by periportal and periductal fibrosis, resulting in fibro-obliteration and segmental strictures of bile ducts (Fig. 7D and E). Table 4 shows that the cholestatic liver damage in Mdr2−/− mice is reflected by significantly elevated serum transaminase and AP activities, as well as slightly increased bilirubin levels, similar to previous reports.1, 2, 6 These findings demonstrate that the liver injury of Mdr2−/− mice resembles human diseases characterized by progressive sclerosing cholangitis,6 for example, primary sclerosing cholangitis and PFIC (type 3), the latter being caused by ABCB4 deficiency and the resulting absence of biliary phospholipids.4
Table 4. Liver Chemistry Tests
Values were determined in 3–6 animals per group and represent means ± SE.
‡‡, b, **, ##, ‡, *
*P < 0.05; **P < 0.01, compared to Mdr2+/+ mice; ‡P < 0.05; ‡‡P < 0.01, compared to male mice; #P < 0.05, ##P < 0.01, compared to 9 weeks.
Upon macroscopic examination of the biliary system of 7-month-old female Mdr2−/− mice we noted dilated extrahepatic duct, which contain multiple concrements impacted in the ducts. By polarizing light microscopy, these show stones of similar morphology compared to the gallbladder stones (Fig. 1) and are composed of needle-like crystals. Whereas only very few mice display bile duct stones by 18 weeks of age, 42% of 7- to 10-month-old female but no male mice develop choledocholithiasis (n = 26). Panels H and I of Fig. 7 display photomicrographs of frozen liver sections of these mature mice using polarizing light microscopy and reveal occluded intrahepatic bile ducts, which are filled with crystalline occlusions composed of needle-like crystals plus mucin gel. In contrast, fresh hepatic biles collected via an acute biliary fistula do not contain any solid crystals when analyzed by light or electron microscopy (data not shown).
A major finding of this study is that spontaneous cholesterol stone formation in gallbladder bile of Mdr2−/− mice in vivo confirms predictions from in vitro studies that studied model biles with very low phospholipid/bile salt ratios.9, 10, 40 Moreover, the findings in the Mdr2−/− mouse model are in agreement with recent observations of ABCB4 mutations in adult patients with cholesterol gallstones.7 Rosmorduc et al.7, 8 showed that ABCB4 loss-of-function mutations in either the homozygous or heterozygous state are linked with LPAC. LPAC is characterized by (i) clustering within families; (ii) onset of symptoms before the age of 40 years; (iii) intrahepatic hyperechoic foci, intrahepatic sludge, or microlithiasis; and (iv) recurrence of symptoms after cholecystectomy.7, 8
The link between Mdr2 deficiency and cholelithiasis is consistent with the increased prevalence of gallstones in PFIC type 34 and intrahepatic cholestasis of pregnancy,41 which are associated with ABCB4 mutations in individual patients.4 Shoda et al.42 demonstrated recently that the formation of cholesterol and cholesterol-rich brown pigment stones in intrahepatic bile ducts may be caused by a decreased hepatic efflux of phospholipids because these patients showed significantly decreased levels of ABCB4 mRNA and protein levels as well as reduced phospholipid concentrations in gallbladder bile.42 Fracchia et al.43 further supported this concept by demonstrating that patients with cholesterol microlithiasis display significantly decreased mean percent molar concentrations of phospholipids in duodenal bile. Interestingly, using polarizing light microscopy of hepatic cryostat sections, we observed intrahepatic crystal agglomerates in 7-month-old Mdr2−/− mice (Fig. 7). It is unlikely that the full picture of fibro-obliterative sclerosing cholangitis observed in 8-week-old Mdr2−/− mice6 is the consequence of stone formation, because no stones are present in male animals around that time and fresh hepatic bile samples do not contain any crystals. Moreover, periductal fibrosis and other features of bile duct damage are seen as soon as 2 weeks after birth.6 We therefore speculate that stagnation of bile as a result of strictures and segmental biliary obstruction may promote intrahepatic stone formation in the presence of an abnormal bile composition. Notwithstanding, this study establishes the Mdr2−/− mouse as an animal model for low phospholipid-associated microlithiasis and primary hepatolithiasis. With respect to low phospholipid secretion rates in human disease, it was reported that some non-obese human gallstone patients display significantly reduced phospholipid secretion rates that were out of proportion to low bile salt secretion rates.44 This suggests that the regulation of the hepatic PC transporter might be defective in some non-obese patients with cholesterol gallstone disease. Recently, needle-like cholesterol crystals have been documented in patients with acalculous gallbladder disease,45 albeit the possible pathophysiologic link of this disease entity to low biliary phospholipid concentrations remains to be investigated.
Whereas recent studies in humans7, 8 identified ABCB4 as monogenetic risk factor for cholelithiasis, the Mdr2−/− mouse provides a physical-chemical explanation for the formation of gallstones. Cholesterol and lecithin molecules are secreted by individual transporters from the canalicular membrane and form vesicles within the canalicular space.46, 47 The compositions of mixed vesicles can be inferred from the cholesterol/lecithin molar ratio in bile. Cholesterol-supersaturated vesicles with a cholesterol/lecithin ratio of ∼0.50 precipitate classic plate-like cholesterol crystals.18, 46 In contrast, biles of Mdr2−/− mice (Table 1) can display a cholesterol/lecithin ratio greater than 1.00 (0.65–2.24). Similar to Mdr2−/− mice, bile compositions of LPAC patients show cholesterol supersaturation in conjunction with low phospholipid concentrations, resulting in cholesterol/lecithin ratios between 0.44 and 0.93.7, 43 The bile compositions of Mdr2−/− mice plot in the lower left region A of the biliary phase diagram, which nucleates filamentous and needle-like cholesterol crystals.10 However, the present study cannot discriminate whether the primary crystalline form is truly “anhydrous” cholesterol that undergoes a polymorphic transition or represents a novel monohydrate polymorph, as indicated by a distinct diffraction pattern upon electron diffraction of single crystals in nucleating model bile of similar lipid composition to that of Mdr2−/− mice (Dr. F.M. Konikoff, Tel Aviv University, Dr. Y. Talmon, Technion University of Haifa, personal communication, June 2003). Cholesterol crystals in Mdr2−/− mice could originate from bile salt-dissolution of residual cholesterol-rich vesicles15; alternatively, as evidenced by their increased hydrophobicity indexes (Table 3), direct bile salt elution of structural phospholipids and cholesterol from the canalicular membrane might occur.
Besides the Mdr2−/− mouse, the deer mouse (Peromyscus maniculatus) is the only animal known to form gallstones spontaneously in the wild,48 whereas all other animal models require dietary challenges to increase cholesterol concentrations in bile. Following this paradigm, Mdr2−/− mice were reported to develop gallbladder stones upon dietary challenge with cholic acid (0.1%) for more than 3 weeks,13, 16 but spontaneous gallstone formation has not been recorded in this knockout mouse model. However, development of gallstones in Mdr2−/− mice resembles stone formation in guinea pigs, which are also characterized by very low biliary phospholipid/bile salt ratios.49 Under specific conditions such as vitamin C deficiency or lincomycin-treatment, gallstones of guinea pigs contain randomly aggregated needle-like cholesterol crystals, as well as high amounts of calcium carbonate and calcium phosphate,50, 51 all of which are reminiscent of stones of Mdr2−/− mice. The carbonate content is likely to result from increased bicarbonate secretion in obstructed bile ducts.52
From our model bile systems, we predicted that with a dearth of phospholipids, a cholesterol/phospholipid ratio of ∼1.0 (Table 1) results in precipitation of short needle-like crystals that remain stable and do not evolve rapidly into typical plate-like cholesterol monohydrate crystals.53 In fact, this crystal type predominates in bile of Mdr2−/− mice. The bile of Mdr2−/− mice contains trace amounts of phospholipids, whose molecular species may influence cholesterol crystal habits and transition sequences of crystal intermediates during cholesterol crystallization from bile salt-rich biles.53 Whereas common biliary PC species (16:0–18:2, 16:0–18:1) are detected in bile of Mdr2+/+ mice, bile of Mdr2−/− mice contains rarer structural membrane PCs that presumably are not pumped out by MDR2 (Table 3). The presence of these and other membrane lipids (data not shown) can be attributed to bile-salt induced biliary mucosal injury in Mdr2−/− mice.5 Compared to controls, the bile is relatively depleted in PC species with saturated acyl chains and contains higher amounts of long unsaturated PCs. In vitro, these more hydrophobic PC species are associated with retarded precipitation of short filaments and no formation of metastable helical or tubular intermediates32, 54; and in fact, no intermediate helix formation9, 10 was observed in gallbladder biles of Mdr2−/− mice in vivo.
In addition to phospholipids, bile salt molecular species can influence cholesterol crystallization. In this study, female Mdr2−/− mice display a more hydrophobic bile salt pool due to lower tauro-β-muricholate levels in bile (Table 2). Increasing bile salt hydrophobicity has been demonstrated in vitro to enhance cholesterol crystallization by reducing the liquid crystal region E of the phase diagram, where solid crystal precipitation is a forbidden phase transition.10, 40, 55 This likely explains the earlier onset and higher prevalence of gallstones in female Mdr2−/− mice. These findings provide a rationale for substitution of more hydrophobic bile salts with the more hydrophilic ursodeoxycholic acid in LPAC patients. Moreover, the Mdr2−/− model predicts that a more hydrophilic bile salt pool might not be sufficient to fully prevent cholecysto- and hepatolithiasis associated with phospholipid deficiency and might even display deleterious effects in the presence of fully developed strictures and mechanical obstruction.6 In the future, the Mdr2−/− mouse is expected to serve as an experimental model for the design and testing of new therapeutic interventions for human low phospholipid-associated cholelithiasis.
Dr. Wang is a recipient of a New Scholar Award from the Ellison Medical Foundation (1999–2003). The authors are indebted to Prof. Dr. Wolf D. Lehmann (Central Spectroscopy Department, German Cancer Research Center, Heidelberg, Germany) who generously provided tandem mass spectrometry for analysis of biliary phospholipids and to Dr. Reinhart Kluge (Aachen University) for veterinarian support.