Potential conflict of interest: Nothing to report.
Niemann-Pick type C (NPC) disease develops as a result of mutations in the NPC1 gene that encodes a protein involved in the net movement of unesterified cholesterol from the late endosomal/lysosomal compartment to the metabolically active pool of sterol in the cytosol of virtually every cell in the body. Although early publications emphasized the neurodegeneration occurring in children with this mutation, more recent clinical information suggests that serious liver disease also is an important part of this syndrome. These studies, therefore, were undertaken to characterize the liver dysfunction seen in mice with this same mutation. The NPC mouse develops significant hepatomegaly that reaches 8% of body weight at 5 to 6 weeks of age. This increase in liver size is associated with a linear increase in cholesterol content and with accumulation of amorphous cellular inclusions in both hepatocytes and macrophages. During the few weeks after birth, significant elevation of the plasma alkaline phosphatase level occurs, as also is seen in the human infant with this disease. At 4 to 5 weeks of age, plasma aminotransferase levels also rise abruptly. Histologically, at this time there is apoptosis, but no excess deposition of collagen or glycogen. mRNA expression is elevated for caspase 1, caspase 6, and several enzymes associated with sterol biosynthesis and bile acid formation. In conclusion, the NPC mouse has liver disease similar to that seen in the NPC infant and represents a relevant model for exploring the molecular events occurring in this form of liver disease. (HEPATOLOGY 2005;42:886–893.)
Niemann-Pick type C (NPC) disease is an autosomal recessive lipid storage disorder characterized by enlargement of the liver and spleen, progressive hepatic and neurological deterioration, and death in the second or third decade of life.1 The mutation responsible for approximately 95% of these cases has been mapped to a gene on chromosome 18q11 designated NPC1.2 This gene encodes a protein (NPC1) with 5 transmembrane domains that share homology with the sterol-sensing domain of 2 enzymes involved in cellular cholesterol homeostasis, 3-hydroxy-3-methylglutaryl coenzyme A reductase and sterol regulatory element binding protein cleavage-activating protein.2, 3
Most cells in the body take up cholesterol from the surrounding environment, using one or more members of the low-density lipoprotein receptor (LDLR) family, and then process this sterol through the clathrin-coated pit pathway to the metabolically accessible pool of sterol in the cytosol.4–6 Because the NPC1 protein apparently functions to move unesterified cholesterol from the late endosomal/lysosomal compartment of these cells to the metabolically active pool, the characteristic finding in the child (or mouse) with a mutation of this protein is the progressive accumulation of sterol in every tissue.7–10 However, the magnitude of this accumulation varies in different organs, depending on how much cholesterol is normally taken up through the clathrin-coated pit pathway in each of these particular tissues.11
Liver plays the key role in the clearance of circulating cholesterol carried in lipoproteins and so manifests the highest rate of sterol accumulation in both the human and mouse with a mutation in NPC1.5, 7, 9 The pathways responsible for this accumulation are outlined diagrammatically in Fig. 1. Cholesterol absorbed from the intestine and carried in the chylomicron remnant (CMr) is nearly completely cleared into the liver, using the LDLR and clathrin-coated pit pathway. Similarly, nearly all of the cholesterol carried in the remnant of very low density lipoprotein (VLDLr) and approximately 80% of the sterol in low-density lipoprotein (LDL) is cleared from the plasma by the liver.5, 12 Nearly all of this sterol becomes trapped in the late endosomal/lysosomal compartment when NPC1 is mutated and not functional. In contrast, as also shown in Fig. 1, cholesterol that is carried in high-density lipoprotein (HDL) and taken up into the liver by scavenger receptor BI (SR-BI) or that is synthesized in the hepatocyte is handled normally by the liver.13
Despite significant advances in understanding the pathogenesis of NPC disease, the liver dysfunction associated with this syndrome has been largely ignored and is poorly characterized. However, the literature suggests that between 45% and 65% of NPC patients present with neonatal cholestasis, and 10% of these will die of liver failure before 6 months of age.14–16 These patients typically do not have neurological manifestations. In another study, NPC disease was found to be the second most common genetic/metabolic cause of neonatal cholestasis after alpha-1-antitrypsin deficiency over a 20-year period at King's College Hospital in the United Kingdom.17 In a retrospective review over a 2-year period in the United States, NPC disease was the most common genetic/metabolic cause of neonatal cholestasis.15 Such data suggest that severe liver dysfunction in NPC disease is more common than previously believed.
Because of the potential significance of this liver disorder in children with the NPC1 mutation, the current studies were undertaken to define the liver dysfunction present in a murine model of NPC disease, the BALB/c mouse with a mutation in the npc1 gene. This model has proved valuable in delineating many of the molecular and physiological defects found in the presence of the npc1 mutation and in testing potential therapies to reverse these defects. However, only one report has appeared of the hepatic lesions seen in this murine model, and this study was carried out in animals fed a very high-fat, high-cholesterol diet.18 For this reason, the current studies were undertaken to delineate in detail the biochemical, histological, and molecular abnormalities that occur in the liver of the NPC mouse maintained on low levels of dietary lipid intake.
NPC, Niemann-Pick type C; NPC1, NPC1 protein; LDLR, low density lipoprotein receptor; CMr, chylomicron remnant; VLDLr, very low density lipoprotein remnant; LDL, low-density lipoprotein; HDL, high-density lipoprotein; SR-BI, scavenger receptor type BI, AP, alkaline phosphatase; AST, aspartate aminotransferase; ALT, alanine aminotransferase.
Materials and Methods
Animals and Diets.
Wild type (npc1+/+) and homozygous mutant (npc1−/−) mice were generated from heterozygous npc1+/− animals with a BALB/c background. Breeding stock was originally obtained from Dr. Peter G. Pentchev at the National Institutes of Health.3 All animals were fed ad libitum a low-cholesterol rodent diet (no. 7001; Harlan Teklad, Madison, WI) after weaning at 19 days of age. This diet had a cholesterol content of 0.016% (wt/wt) and a total lipid content of 5% (wt/wt). Measurements were made in animals at different ages, varying from 1 to 75 days. Because no gender differences were observed in any experimental parameter, the experimental groups consisted of equal or nearly equal numbers of males and females. All experimental protocols were approved by the Institutional Animal Care and Use Committee.
Measurement of Tissue Cholesterol Concentration.
Animals were exsanguinated, and the brain and liver were removed and saponified. The sterol was extracted with petroleum ether, and the cholesterol concentration was measured by gas-liquid chromatography using stigmastanol (Sigma, St. Louis, MO) as an internal standard.19, 20 The values were expressed either as milligrams of cholesterol per gram wet weight of tissue (mg/g) or as milligrams of cholesterol per whole organ (mg/whole organ).
Measurement of Liver Triacylglycerol Content.
An aliquot of liver (300-400 mg) was extracted in chloroform:methanol (2:1, vol/vol) in the presence of [14C]triolein (American Radiolabeled Chemicals, Inc., St. Louis, MO). Duplicate 5-mL aliquots of this extract were dried under air, and the residue was redissolved in 1 mL hexane:methyl-t-butyl ether (100:1.5, vol/vol). This solution was run over a Sep-Pak Vac RC silica cartridge (500 mg) (Waters Corp., Milford, MA). After elution of the cholesteryl esters, the solvent was changed to hexane:methyl-t-butyl ether (96:4, vol/vol) for elution of the triacylglycerols. This fraction was dried under air and redissolved in chloroform–methanol (2:1, vol/vol). Aliquots of this solution were dried and used for the measurement of the recovery of the internal standard, and the mass of triacylglycerol was measured using Infinity Triglycerides Liquid Stable reagent (ThermoTrace, Noble Park, Australia). This value was used to calculate the milligrams triacylglycerol in the whole liver.
Measurement of Plasma Liver Enzyme Activities.
Plasma alkaline phosphatase, aspartate aminotransferase (AST), and alanine aminotransferase (ALT) activities (U/L) were measured by a commercial laboratory.
Livers were harvested from age-matched npc1+/+ and npc1−/− mice at 5, 12, 56, and 75 days of age. Tissues were fixed in either 4% paraformaldehyde or alcoholic formalin,21 and sections were then stained with hematoxylin-eosin, Masson Trichrome stain,22 periodic acid-Schiff (PAS),22 or for nuclear DNA fragmentation (terminal deoxynucleotidyl transferase-mediated nick-end labeling [TUNEL])23.
Measurement of Relative mRNA Expression in Liver Tissue.
Aliquots of liver tissue from animals at 56 days of age were frozen in liquid nitrogen. Total RNA was prepared from 4 liver samples of npc1+/+ and npc1−/− mice, and equal aliquots were pooled with an RNA STAT-60 kit (Tel-Test, Friendswood, TX). cDNA was synthesized from 5 μg DNase I-treated total RNA (Rnase free, FPLC pure, Amersham Pharmacia, Buckinghamshire, UK) using the SuperScript First-Strand Synthesis System (catalogue no. 11904-018, GIBCO/BRL) and random hexamer primers. Specific primers for each gene were designed by using PRIMER EXPRESS software (Perkin-Elmer, Boston, MA). The real-time reverse transcription polymerase chain reaction contained, in a final volume of 30 μL, 50 ng reverse-transcribed total RNA, 167 nmol forward and reverse primers, and 15 μL 2× SYBR green PCR Master Mix.24 PCRs were carried out by using the Applied Biosystems Prism 7700 Sequence Detection System (Foster City, CA). All reactions were done in triplicate. The relative amounts of all mRNAs were calculated using the comparative CT method (User Bulletin no. 2, Perkin-Elmer).
Measurement of Cholesterol and Fatty Acid Synthesis.
The rates of cholesterol and fatty acid synthesis were measured in vivo by using [3H]water as previously described.20
Whole Body and Organ Size During Development.
Initial experiments were undertaken to quantify the effect of the npc1 mutation on the growth of the whole animal, and on the absolute and relative weights of the liver and brain, the two organs primarily affected by this disease. As shown in Fig. 2A, the npc1+/+ and npc1−/− animals progressively gained weight over most of the 75 days of observation, although a pause occurred in the rate of gain in both groups just as the animals were weaned to solid food. However, the mean weights of the npc1−/− animals were lower than those of the npc1+/+ mice at nearly every age, and these differences became large between 49 and 75 days of age, when significant neurological abnormalities became manifest. In contrast, throughout this period of growth, the mean weight of the liver in the npc1−/− mice equaled or exceeded that seen in the control mice (Fig. 2B). These differences were even greater when relative liver weight was calculated (Fig. 2C). Relative liver size declined during the first 2 weeks of life as the mice went into a period of rapid body growth. Beyond this age, however, relative liver size increased, and this increase was greater in the npc1−/− animals than in the control mice. After approximately 42 days of age, the liver accounted for approximately 6% of body weight in the npc1+/+mice but 8% in the npc1−/− animals. The absolute weight of the brain was less in mutant mice (Fig. 2D). However, when corrected for the lower body weights seen in these npc1−/− animals, relative brain weight was the same in the mutant and control mice (Fig. 2E).
Liver and Brain Cholesterol Concentration and Content.
The striking effect of the npc1 mutation on sterol homeostasis is evident in Fig. 3. In the npc1+/+mice, the concentration of cholesterol was elevated at approximately 4 mg/g during the suckling period but promptly fell to approximately 2.2 mg/g after weaning when oral cholesterol intake decreased (A). In contrast, in the npc1−/− mice, this concentration was much higher at every age, although these values temporarily declined after weaning before again increasing in the older animals. This variation in concentration resulted from the variable rate of liver growth during this same period (Fig. 2B). A more accurate assessment of the absolute rate of cholesterol accumulation in the liver of the animals with the npc1 mutation is seen when whole organ sterol content is calculated. As is apparent, the absolute amount of cholesterol in the liver increased in an essentially linear manner from 0.8 mg at birth to 37 mg at 56 days of age (Fig. 3B). However, the content of cholesterol in the liver of the npc1+/+ animals increased from only 0.4 mg to 3.6 mg over this same interval. In contrast to these findings in the liver, the concentration of cholesterol in the brain of the npc1−/− animals was marginally less than in the control mice at all ages beyond approximately 28 days (Fig. 3C), and this difference was even greater when whole brain sterol content was calculated (Fig. 3D). This decrease represented the nerve cell death and demyelination known to be occurring at these ages in the npc1−/− animals.25–27
Assessment of Liver Cell Damage.
The plasma level of 3 enzymes was next measured to assess liver cell damage in these same groups of animals. As seen in Fig. 4A, the plasma alkaline phosphatase (AP) level was elevated in both groups of newborn pups, although the value was higher in the npc1−/− mice (815 U/L) than in the control animals (578 U/L). However, this value progressively declined in the npc1+/+ mice over the next 28 days, but increased abruptly (to 965 U/L) in the mutant animals before returning to much lower levels in the older mice. In contrast, the plasma levels of both ALT and AST were normal or only minimally elevated during this initial 28-day period. However, both enzymes began to increase dramatically at approximately 35 days of age, reaching values in the mutant mice that were approximately 15-fold higher than those seen in the control animals at 75 days of age (Fig. 4B-C).
To correlate these changes in hepatic cholesterol accumulation and plasma enzyme levels with cellular damage, histological examination of the liver was next carried out at several critical times during the evolution of this disease. As seen in Fig. 5A-B, at 5 days of age there was no observable difference in liver histology in the npc1−/− mice even though the cholesterol content of these cells was already significantly elevated (2.7 vs. 0.7 mg/whole organ) compared with the control mice. Foci of extramedullary hematopoiesis were seen. No evidence was seen of inflammation or cholestasis in either group of animals. At 12 days of age, when the cellular cholesterol content was further elevated (6.0 vs. 1.2 mg/whole organ) the hepatocytes were swollen and showed finely vacuolated cytoplasm (Fig. 5C-D). Within the sinusoids, small numbers of large, pale-stained macrophages were identified. In addition, increased numbers of hepatocytes in mitosis were seen in the sections from the npc1−/− mice, but not in those from the npc1+/+ animals. No evidence was seen of periportal inflammation. At 56 days of age (36.8 vs. 3.3 mg/whole organ), differences were more dramatic (Fig. 5E-F). Large numbers of foamy macrophages were seen within the sinusoids of the livers from the npc1−/− animals, and, commonly, these macrophages were in clusters of 3 to 4 cells. Apoptosis of hepatocytes was common, but mitotic activity was rare. Again, there was no evidence of periportal inflammation. At 75 days of age (34.4 vs. 2.9 mg/whole organ), the number of sinusoidal macrophages increased further to represent approximately one third of the cross-sectional area of the sections examined. Apoptosis was common in macrophages and hepatocytes (Fig. 5G-H), and fragmented nuclear DNA was identified on TUNEL-stained sections (Fig. 5I-J).
PAS staining showed minimal accumulation of glycogen in the hepatocytes of the npc1−/− animals at 12 days of age (Fig. 6A-B). Furthermore, the amount of glycogen accumulation seen in these 12-day-old animals was not different from that observed in the npc1−/− mice at 75 days of age (Fig. 6B,D). Liver samples obtained from 12-day-old animals and stained with trichrome stain showed only minimal collagen deposition, mainly around the central veins, and there were no apparent differences in the livers from the mutant and control mice (Fig. 6E-F). Furthermore, no significant increase was seen in collagen deposition even in the npc1−/− animals at 75 days of age (Fig. 6G-H), as has been previously reported.28
Hepatic Lipid Content and Synthesis, and Expression of mRNA Levels.
Finally, several metabolic parameters and the relative mRNA expression for a number of proteins were measured in 56-day-old npc1+/+ and npc1−/− mice, as shown in Fig. 7. The content of cholesterol (panel A) was very high, as was the rate of sterol synthesis (panel B) in the npc1−/− animals. However, the content of triacylglycerol (panel C) and the rate of fatty acid synthesis (panel D) were lower in the mutant than in the control mice. This difference in the rate of cholesterol synthesis was reflected in the relative mRNA expression in the npc1−/− animals for HMGCoA reductase (HMGCoAR) (panel E), the rate-limiting enzyme in the biosynthetic pathway. Importantly, the relative mRNA levels for tumor necrosis factor alpha (panel F), caspase 1 (panel G), and caspase 6 (panel H) were also elevated in the livers of the mutant mice.
The physiological characteristics of the phenotype of the npc1−/− animal (and human) are very unusual, as illustrated by these studies. The fundamental problem caused by mutation of the NPC1 protein is that unesterified cholesterol, derived from apoB-containing lipoproteins processed through the clathrin-coated pit pathway, becomes trapped in late endosomes/lysosomes (Fig. 1). In contrast, cholesterol newly synthesized in the extrahepatic organs and delivered to the liver through the SR-BI pathway is processed normally to biliary sterols and bile acids (Fig. 1).13 Because the sterol trapped in the late endosomal/lysosomal compartment cannot become part of the metabolically active pool in the cell, paradoxically, the hepatocyte senses a shortage of cholesterol and, to compensate, increases the rate of synthesis. Thus, in the face of a massively expanded hepatic cholesterol pool (Figs. 3B, 7A), an increase occurs in the relative mRNA expression of HMGCoAR (Fig. 7E), the rate-limiting enzyme in the biosynthetic pathway, and a corresponding increase in overall sterol synthesis (Fig. 7B). In this manner, sterol balance across the liver and whole animal is maintained, but the mass of unesterified cholesterol in the late endosomal/lysosomal compartment progressively expands with age (Fig. 3B).
This expansion clearly leads to significant abnormalities of liver function. During the suckling period, when both absolute (Fig. 2B) and relative (Fig. 2C) hepatic growth is slow and the intake of dietary sterol from milk is high, the concentration of cholesterol in the liver of the pups increases abruptly (Fig. 3A). This increase is associated with a similar marked rise in the plasma alkaline phosphatase level that then returns toward normal in the older animals (Fig. 4A). Notably, most (65% in one series14) newborn infants with NPC disease also manifest prolonged cholestasis that usually disappears with age.14, 15, 29 Only later in the young adult mice do the aminotransferase levels increase significantly (Fig. 4B-C) at a time when there is also an increase in the relative mRNA expression of tumor necrosis factor alpha and caspases 1 and 6 (Fig. 7F-H) and the histological appearance of apoptotic cells (Fig. 5H,J). This finding again is similar to reports in humans with NPC disease that approximately half of those children who recover from the period of prolonged neonatal cholestasis manifest persistent liver disease with elevated aminotransferase levels.14
Similarities also were seen in the histology of the diseased liver in the mice and humans. Both species show vesicular cytosolic inclusions, activation of sinusoidal macrophages, and apoptotic cells (Fig. 5).14, 15 In some young children, however, increased fibrotic tissue has been described in liver biopsies with interlobular and intralobular bridging and, in some cases, full-blown cirrhosis.14, 15 In contrast, there was no increase in hepatic fibrosis in the npc1−/− mice detected by trichrome staining (Fig. 6H). However, a two-fold increase in hepatic hydroxyproline content has been reported in such animals even though excess fibrosis could not be detected histologically.28 Because these animals developed obvious neurological symptoms beginning at 5 to 6 weeks and died at 11 to 12 weeks of age, possibly there was insufficient time for the development of detectable fibrosis even though there was significant liver cell damage in this murine model.
These findings in the liver differ in several respects from those in the brain. As in the liver, there is accumulation of unesterified cholesterol in the neurons of the central nervous system with selective cell death.10, 25, 27, 30 Unlike the liver, however, this neurodegeneration is associated with demyelination so that the mass of unesterified cholesterol in the central nervous system decreases slightly (Fig. 3D) rather than increasing, as it does in the liver (Fig. 3B).30 However, as these studies show, the mice and humans with NPC disease both develop hepatic and neurological dysfunction. However, which factors determine the relative severity of each of these lesions in a given individual is not clear. Children who initially manifest significant liver abnormalities apparently die at a median age of 8.3 years, whereas those who have predominantly neurological symptoms die at a median age of 10.5 years.14
In conclusion, these and other studies indicate that this murine model represents a very useful experimental animal in which to further explore the pathogenesis of both the liver and central nervous system disease. Furthermore, because the mechanisms of cholesterol delivery to the liver are now well understood and because new pharmaceutical agents are available to interrupt this delivery, it may be possible in the near future to design therapy that will considerably ameliorate the severity of the liver disease and so improve the quality of life for children with NPC disease.
The authors thank Claire Havel, Jennifer Weissblum and Norma Anderson for excellent technical assistance and Merikay Presley for expert preparation of the manuscript.