Strain Background Modifies Phenotypes in the ATP8B1-
Sohela Shah1,2., Ukina R. Sanford1,2., Julie C. Vargas1,2¤a, Hongmei Xu1,2¤b, Annamiek Groen3, Coen C.
Paulusma3, James P. Grenert1,4, Ludmila Pawlikowska5,6, Saunak Sen1,7, Ronald P. J. Oude Elferink3,
Laura N. Bull1,2,6*
1UCSF Liver Center, University of California San Francisco, San Francisco, California, United States of America, 2Department of Medicine, University of California San
Francisco, San Francisco, California, United States of America, 3Academic Medical Center, Tytgat Institute for Liver and Intestinal Research, Amsterdam, The Netherlands,
4Department of Pathology, University of California San Francisco, San Francisco, California, United States of America, 5Department of Anesthesia and Perioperative Care,
University of California San Francisco, San Francisco, California, United States of America, 6Institute for Human Genetics, University of California San Francisco, San
Francisco, California, United States of America, 7Department of Epidemiology and Biostatistics, University of California San Francisco, San Francisco, California, United
States of America
Background: Mutations in ATP8B1 (FIC1) underlie cases of cholestatic disease, ranging from chronic and progressive
(progressive familial intrahepatic cholestasis) to intermittent (benign recurrent intrahepatic cholestasis). The ATP8B1-
deficient mouse serves as an animal model of human ATP8B1 deficiency.
Methodology/Principal Findings: We investigated the effect of genetic background on phenotypes of ATP8B1-deficient
and wild-type mice, using C57Bl/6 (B6), 129, and (B6-129) F1 strain backgrounds. B6 background resulted in greater
abnormalities in ATP8B1-deficient mice than did 129 and/or F1 background. ATP8B1-deficient pups of B6 background
gained less weight. In adult ATP8B1-deficient mice at baseline, those of B6 background had lower serum cholesterol levels,
higher serum alkaline phosphatase levels, and larger livers. After challenge with cholate-supplemented diet, these mice
exhibited higher serum alkaline phosphatase and bilirubin levels, greater weight loss and larger livers. ATP8B1-deficient
phenotypes in mice of F1 and 129 backgrounds are usually similar, suggesting that susceptibility to manifestations of
ATP8B1 deficiency may be recessive. We also detected differences in hepatobiliary phenotypes between wild-type mice of
Conclusions/Significance: Our results indicate that the ATP8B1-deficient mouse in a B6 background may be a better model
of human ATP8B1 deficiency and highlight the importance of informed background strain selection for mouse models of
Citation: Shah S, Sanford UR, Vargas JC, Xu H, Groen A, et al. (2010) Strain Background Modifies Phenotypes in the ATP8B1-Deficient Mouse. PLoS ONE 5(2):
Editor: Alfred Lewin, University of Florida, United States of America
Received September 4, 2009; Accepted January 11, 2010; Published February 1, 2010
Copyright: ? 2010 Shah et al. This is an open-access article distributed under the terms of the Creative Commons Attribution License, which permits
unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.
Funding: This work was supported by the following grants from the U.S. National Institutes of Health: R01 DK50697 to L. N. Bull, R01 GM078338 to S. Sen, and
R01 DK072187 to S. Erickson, and the UCSF Liver Center (NIH P30 DK026743). The funders had no role in study design, data collection and analysis, decision to
publish, or preparation of the manuscript.
Competing Interests: One co-author has what might be considered a possible competing interest, in that she now works in industry: Ms. Julie Vargas, who
worked on the studies reported here as a staff research associate (technician) in Dr. Bull’s laboratory, now works at F. Hoffmann-La Roche, ltd. In Basel, Switerland.
No other competing interest are present.
* E-mail: firstname.lastname@example.org
. These authors contributed equally to this work.
¤a Current address: F. Hoffmann-La Roche, Ltd., Basel, Switzerland
¤b Current address: Children’s Hospital of Chongqing Medical University, Chongqing, China
ATP8B1, also known as FIC1 (familial intrahepatic cholestasis
1), is an ATP-dependent membrane transport protein in the P-
type ATPase family . ATP8B1 belongs to the P4 subfamily of P-
type ATPases. Members of this subfamily appear to function in
phospholipid transport [2–4]; 14 P4 P-type ATPases are encoded
in the human genome. ATP8B1 is involved in transport of
phosphatidylserine from the outer to the inner leaflet of the plasma
membrane [5–7]. Mutations in ATP8B1 result in cholestatic
disease with an autosomal recessive mode of inheritance, and
ranging in severity from mild and episodic (benign recurrent
intrahepatic cholestasis, BRIC1) to chronic and progressive
(progressive familial intrahepatic cholestasis; PFIC1) [1,8,9].
Patients with severe ATP8B1 deficiency (i.e. PFIC1) typically
require liver transplantation prior to adulthood, due to liver failure
[10,11]. While severity and penetrance of ATP8B1 deficiency is
correlated with the predicted impact of the ATP8B1 mutation(s)
that a patient carries, additional as-yet-unidentified genetic and/or
environmental factors also have an influence .
We previously generated mice homozygous for a mutation in
Atp8b1, the mouse ortholog of ATP8B1 . These are knock-in
PLoS ONE | www.plosone.org1 February 2010 | Volume 5 | Issue 2 | e8984
mice for the 923G.T point mutation identified in Amish PFIC
patients; this mutation results in an amino acid change in a highly
conserved residue, G308V . Mice homozygous for this missense
mutation are termed Atp8b1G308V/G308Vmice, or ‘ATP8B1 mutant’
mice. These mice exhibited defects in bile acid homeostasis, but
did not suffer from progressive cholestatic liver disease. When
challenged with a bile salt-supplemented diet, ATP8B1 mutant
mice displayed a more severe phenotype, including rapid weight
loss, greater liver enlargement, and biochemical evidence of
cholestasis, although the phenotype was still less severe than that
seen in human patients. Subsequent studies indicated that the
canalicular membrane in ATP8B1 mutant mice is susceptible to
damage by hydrophobic bile salts, and that hepatobiliary excretion
of hydrophobic bile salts is impaired [13,14].
Published studies have characterized phenotypes in male
ATP8B1 mutant mice, and with one exception , have
focused upon mice in a 129 strain background [12,13,15,16];
preliminary findings suggested that the ATP8B1 mutant
phenotype might differ between 129 and C57Bl/6J (B6).
Therefore, we compared effects of the Atp8b1 mutation in 129
and B6 strain backgrounds. Here, we present evaluation of
aspects of serum and bile biochemistry, as well as body- and
liver-weight related phenotypes, in male and female WT (wild-
type) and ATP8B1 mutant mice in B6, 129, and F1(B66129)
To investigate the simultaneous effects of mutation (WT and
ATP8B1 mutant), background strain (B6, 129, and F1), sex (male
and female), and diet (cholate-supplemented and control), we
conducted a factorial experiment. We studied $5 mice for each of
the (2636262=24) factorial combinations (‘‘full factorial exper-
iment’’) . This approach allowed us to study not only the
individual effects (‘‘main effects’’) of each factor (mutation, diet,
genetic background, and sex), but also whether the effect of a
factor depended on other factors (‘‘interactions’’) (Table 1). Unless
otherwise indicated, p-values reported in the text were derived
from ANOVA comparing the groups mentioned; when sex
differences were not apparent, data from males and females were
sometimes combined for these latter tests. For visual clarity, means
and standard errors of the mean (SEM) are used to summarize the
data in the figures, instead of showing the many p-values from
Pup Survival from the Mid-Nursing Period to Weaning Is
Lower in B6 Mice
Offspring of heterozygote couples were less likely to survive
from midway through the nursing period (,day 10) to weaning
(,day 21) if they were of B6, than of 129 or F1background
(Table 1; for offspring of all genotypes, p,0.0003 for B6 versus
129 and p,0.0001 for B6 versus F1, chi-squared test; this
difference, when analyzed separately for mutants and for pooled
WT and heterozygotes, remains significant between strains).
While 7% of B6 pups died during this period, well under 1% of
F1or 129 pups did. Amongst the B6 mice, 14% of mutant, and
5% of pooled WT and heterozygote, mice died during this period
ATP8B1 Mutant B6 Mice Exhibit Slower Weight Gain
during the Nursing Period
Previous study had suggested that mutant mice were slightly
smaller at weaning than WT and heterozygote littermates .
Table 1. Summary of the factorial experiment.
Main Effects Interactions
Pup survival-Y NANA
Mid-nursing weightY- NAY
Weaning weightYY NAY
% weight loss/dayYYYYYYY
Baseline serum cholesterolYY NAYY
Baseline serum ALPYY NAYY
Baseline bilirubin-- NA-
Baseline serum bile saltsYY NA-
Post-diet serum bile saltsY-
Post-cholate biliary cholesterolˆ
Post-cholate biliary phospholipidsˆ
Post-cholate biliary bile saltsˆ
% liver weight relative to final body weight #
The first column lists the phenotypes studied. Columns 2–5 list the main effects of 4 factors- genotype, strain, diet, and sex. Columns 6–11 list the interactions between
genotype, strain, diet, and sex. ‘Y’ indicates that a main effect or interaction influences the phenotype. ‘2’ indicates no main effect. ‘+’ indicates that the factor had no
main effect, but influences the phenotype when interacting with one or more of the other factors. ‘NA’ indicates that the factor was not included or assessed in the
experiment.ˆAs more data were available for F1 and B6 mice on cholate, than control, diet, only results of analysis of cholate diet are shown here. #For percent liver
weight relative to final body weight, a 3-way interaction was detected between strain, diet, and genotype.
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org2 February 2010 | Volume 5 | Issue 2 | e8984
We weighed pups born to heterozygote couples midway through
the nursing period and again at weaning. At both timepoints,
mutant mice trended smaller than their WT and heterozygote
littermates (Table 1, Figure 1 a and b). There was no effect of
strain on this difference midway through the nursing period. At
weaning, however, B6 mutant mice were 12% smaller than their
WT and heterozygous littermates; this mutation-dependent weight
difference was greater than that seen in 129 and F1mice (B6 versus
129 and F1males: p,0.01 for both comparisons; B6 versus F1
females: p,0.001; B6 versus 129 females: p,0.05). A mutation-
associated defect in weight gain during the nursing period thus
appears greater in mice of B6 background, as compared to 129
The Proportion of Mutant Pups Is Lower than Expected
Overall, among pups born to heterozygote couples, and
genotyped at weaning, the mutant allele had a frequency of
46%, which is slightly lower than the expected 50% frequency
(p,0.002). This indicates a mild survival benefit conferred by the
wild-type allele. The genotype frequencies were in Hardy-
Weinberg equilibrium (p=0.81), and were not observed to differ
with strain background (p=0.85). Consistent with the allelic
analysis, 21% of pups born to heterozygote couples, and
genotyped at weaning, were mutants. This is modestly lower than
the expectation of 25%, indicating a mild decrease in rate of
survival to weaning for mutant pups, compared to WT and
heterozygote littermates (p,0.01, chi-squared test, N=861).
Findings were assessed at baseline, and after
short-term feeding of a diet supplemented with 0.5% cholate, or a
ATP8B1 Mutant B6 Mice Lose Substantial Weight upon
Rate of weight change per day was affected by genotype
(ATP8B1 mutant mice lost more weight than WT), strain (B6 mice
lost more weight than 129 and F1mice), diet (mice lost more
weight on cholate diet), and sex (females lost more weight, or
gained less, than males); there are also several interactions (Table 1;
Figure 2). On cholate diet, mutant mice of all strains lose weight
(p,0.001 for all comparisons of mice on control versus cholate
diet, except p,0.01 for F1females; Figure 2b). The weight loss is
greatest in mutant B6 mice, as compared to mutant 129 and F1
mice (p,0.05 to ,0.001).
Figure 1. ATP8B1 mutant B6 mice exhibit slower weight gain
during the nursing period. Pup weight in WT and ATP8B1 mutant
mice of B6, 129, and F1backgrounds at: A) mid-nursing period (,day
10) and B) weaning. Weights of mutant pups were normalized to those
of WT and heterozygote littermates; means and SEM are shown. Range
of N’s: a) B6 (n=14256), 129 (n=23257), and F1(n=702136) and b)
B6 (n=12253), 129 (n=23258), and F1(n=502138).
Figure 2. ATP8B1 mutant B6 mice lose substantial weight upon cholate feeding. Weight change per day in WT (A) and ATP8B1 mutant (B)
mice of B6, 129, and F1backgrounds after feeding of control (grey) or cholate (black) diet for 4–8 days; means and SEM are shown. N’s for cholate
diet: B6 (n=8213), 129 (n=18235), and F1(n=13218); and control diet: B6 (n=529), 129 (n=22229), and F1(n=11216).
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org3February 2010 | Volume 5 | Issue 2 | e8984
Among WT mice, only B6 females lost weight on cholate, as
compared to control, diet (p,0.001; Figure 2a). This finding
indicates that cholate diet may have a greater negative impact in B6
than intheotherstrains, eveninthe absenceoftheAtp8b1mutation.
Atp8b1 Mutation Results in Lowered Serum Cholesterol
in B6 Mice
At baseline, factor analysis showed overall effects of genotype
(mutants,WT), strain (B6,129 and F1), and sex (females,males)
on serum cholesterol levels (Table 1). In addition, there is a strain-
genotype interaction; in B6 mice only, serum cholesterol was lower
in mutants than in WT (males: p,0.001; females p,0.05) (Figure 3
a and b). Among sex-matched WT mice, serum cholesterol was
similar between strains, except for B6 females, which had lower
levels than 129 and F1females (p,0.001; Figure 3a). Among sex-
matched mutants, B6 mice had lower cholesterol levels than 129
mice (p,0.001, males and females); F1mice were intermediate
After Challenge with Cholate-Supplemented Diet,
ATP8B1 Mutant Mice of All Strains Have Low Serum
Genotype, strain, and sex have similar effects on post-diet serum
cholesterol levels, as at baseline. Overall, cholate feeding lowers
Figure 3. Atp8b1 mutation results in lowered serum cholesterol and increased serum alkaline phosphatase (sALP) levels in B6 mice.
Serum cholesterol and alkaline phosphatase levels in WT (A, C) and mutant (B, D) mice of B6, 129, and F1backgrounds at baseline (light grey) and
after feeding of cholate (black) or control (medium grey) diet for 4–8 days; means and SEM are shown. N’s at baseline: B6 (n=7215), 129 (n=42266),
and F1(n=21236); N’s for cholate diet: B6 (n=5212), 129 (n=15229), and F1(n=6219) and for control diet: B6 (n=5212), 129 (n=18229), and F1
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org4 February 2010 | Volume 5 | Issue 2 | e8984
serum cholesterol levels in ATP8B1 mutant mice compared to
control-fed mutants, and there are diet-genotype and diet-sex
interactions (Table 1). Cholesterol levels were significantly reduced
in male mutants of 129 and F1backgrounds after consumption of
cholate, as compared to control, diet (p,0.001 for both), but not
in the groups that had low cholesterol levels after consumption of
control diet as well: B6 males, or females of all backgrounds
In contrast, among WT mice, only B6 males had lower
cholesterol after consumption of cholate, than control, diet
(p,0.001). In WT mice, no strain differences were apparent fter
control diet, but after cholate diet, B6 mice had lower cholesterol
than 129 mice (males: p,0.05; females: p,0.01) (Figure 3a).
Atp8b1 Mutation Results in Increased Serum Alkaline
Phosphatase (sALP) Levels in B6 Mice
At baseline, genotype (mutant.WT), strain (B6.129 and F1),
and sex (females.males) affect sALP levels (Table 1; Figure 3c and
d). Serum ALP levels are higher in mutant B6 mice, as compared
to WT (p,0.001, males and females), but there is no difference
between mutant and WT mice of the other two strains (genotype-
strain interaction). Mutant B6 mice had notably higher sALP
levels than did sex-matched mutant mice of other backgrounds
(p,0.001, each comparison).
After Challenge with Cholate-Supplemented Diet, sALP
Increases in ATP8B1 Mutant Mice of All Strains
After dietary challenge, there are overall effects of genotype
(mutant.WT), strain (B6.129.F1), diet (cholate.control) and
sex (females.males). Cholate feeding increases sALP levels in
mutant mice (diet-genotype interaction; p,0.05 to ,0.001)
(Table 1; Figure 3d). Cholate-fed mutant mice of all strains had
higher sALP levels as compared to cholate-fed WT mice (p,0.001
for all comparisons; Figure 3c and d). In mutant mice after either
diet, B6 mice have notably higher sALP than do 129 or F1mice
(p,0.001 for all comparisons except ,0.01 for male B6 versus
129). In WT mice after cholate diet, B6 females have higher levels
than do females of 129 or F1 background (p,0.001). There are
also diet-sex and sex-genotype interactions.
Serum Bilirubin Concentration Increases in ATP8B1
Mutant Mice after Consumption of Cholate-
At baseline, no differences in bilirubin levels between groups of
mice were detected (Table 1; values not shown). Factor analysis of
post-dietary challenge data showed an overall effect of genotype,
strain and diet, and a diet-genotype interaction. For all strains,
mutant mice have higher serum bilirubin levels than WT mice
when fed cholate diet (p,0.001); after consumption of control diet,
the effect of genotype is significant only for the 129 strain (p,0.05;
Figure 4). After cholate diet, serum bilirubin was higher in B6,
than 129 or F1, mutant mice (p,0.001; Figure 4b). In contrast, for
mutant mice after control diet, and WT mice after both diets,
serum bilirubin levels in B6 and 129 strains were similar, and
higher than those seen in F1mice (p,0.001, each comparison).
Mutant, but not WT, mice of all strains have higher bilirubin
levels after cholate diet as compared to control diet (p,0.001; diet-
Serum Bile Salt Levels Are Higher in ATP8B1 Mutant Mice
At baseline, factor analysis showed overall effects of mutation
(mutants.WT; 129: p,0.001; F1: p,0.01; B6: p,0.05) and
strain (B6 and 129.F1) (Table 1, Figure 5). B6 mutant mice had
higher serum bile salt levels than did F1mutant mice (p,0.05;
Figure 5b). After dietary challenge, there is an overall effect of
genotype and a genotype-diet interaction (Table 1, Figure 5).
Mutant mice fed cholate diet have higher serum bile salt levels as
compared to those fed control diet (p,0.001, each comparison); in
WT mice, effect of diet is smaller (B6 and F1: p,0.05; 129: ns).
Factor analysis did not identify an overall effect of strain after
dietary challenge. Specifically after cholate feeding, however,
serum bile salts were modestly higher in mutant mice of B6 and
129 strain background, than in those of F1background (p,0.05,
Gallbladder Bile Composition after Consumption of
Cholate Diet Is Influenced by Atp8b1 Mutation and Strain
Bile was more amenable to collection by needle aspiration from
mice after cholate diet, as they tended to have well-filled
gallbladders, than from mice after control diet, whose gallbladders
often contained very little fluid. Therefore, we analyzed bile
composition after cholate feeding only.
Figure 4. Serum bilirubin concentration increases in ATP8B1
mutantmiceuponcholatefeeding. ProportionofWT(A) and ATP8B1
mutant (B) mice of B6, 129, and F1backgrounds with normal (light grey),
moderately elevated (medium grey), and highly elevated (black) serum
bilirubin levels in after feeding cholate or control diet for 4–8 days. N’s for
cholate diet: B6 (n=528), 129 (n=30231), and F1 (n=13225); and
control diet: B6 (n=10213), 129 (n=39248), and F1(n=18223).
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org5 February 2010 | Volume 5 | Issue 2 | e8984
were overall effects of genotype (mutant,WT) and strain (B6.129
and F1), and for cholesterol, a strain-genotype interaction (Table 1;
cholesterol and phospholipids concentrations than strain-matched
WT mice (p,0.01 to ,0.001; Figure 6a–d). Mutant mice of B6
background had higher cholesterol and phospholipid concentra-
tions than did mutant mice of F1or 129 backgrounds (p,0.01 to
cholesterol than did 129 mice (p,0.05: Figure 6a). Factor analysis
identified an effect of genotype (mutants,WT) on biliary bile salt
concentration, which attained significance in the 129 strain
(p,0.05; Figure 6e, f).
ATP8B1 Mutant B6 Mice Have Larger Livers than Mutant
There were overall effects of genotype (mutant.WT), strain
(B6.129, with F1intermediate), diet (cholate.control), and sex
(females.males) on liver weight as a proportion of final body
weight. We also detected genotype-strain, genotype-diet, and
strain-diet interactions, and a 3-way interaction between strain,
diet, and genotype (Table 1, Figure 7). Livers of mutant mice were
larger than those of sex-matched WT mice for B6 and F1mice fed
control (B6: p,0.001; F1: p,0.05) and cholate diet (p,0.001 for
all), but in 129 mice this difference was only seen after cholate diet
(p,0.001). After consumption of control diet, mutant mice of B6
background had larger livers than those of 129 or F1 backgrounds
(p,0.001, all comparisons; Figure 7b). After cholate diet, livers of
mutant 129 and F1, but not B6, mice are enlarged relative to after
control diet (p,0.001, except F1males: p,0.05). Nevertheless,
liver weights in male mutant mice after cholate diet consumption
were still higher in B6, than in 129 or F1, mice (p,0.001, both
comparisons); in female mutants after cholate feeding, B6 and F1
mice have larger livers than 129 mice (B6 versus F1: p,0.01; 129
versus F1: p,0.001; Figure 7b). In WT mice after either diet, livers
of B6 mice trended slightly larger than in 129 and F1mice (control
diet: p,0.001 for B6 versus 129 males, ns for others; cholate diet:
p,0.05 to p,0.001; Figure 7a).
Hepatic Levels of Triglycerides and Cholesterol Are
Influenced by Diet and Atp8b1 Mutation
Triglycerides (TG), total cholesterol (TC), and free cholesterol
(FC) were quantified in snap-frozen liver tissue from a represen-
tative subset (143 mice) of the study sample; amount of esterified
cholesterol (EC) was calculated by subtracting FC from TC.
No overall effects of genotype, diet, strain, or sex were detected
for TG and FC; however, for both, diet-genotype and diet-
genotype-sex interactions were detected (Figure 8A–D). TG were
lower in B6 mutant females after cholate diet than in both B6
mutant females after control diet (p,0.01) and B6 WT females
after cholate diet (p,0.05).
An overall effect of diet (cholate.control) was detected for TC
and EC. After consumption of cholate diet, TC was higher in 129
WT mice (p,0.05), and EC was higher in 129 mutant mice
(p,0.01), than respective groups after control diet (Figure 8E–H).
Although factorial analysis did not detect an overall strain effect,
after cholate diet, 129 WT mice had higher EC than B6 WT mice
We have previously shown that mice lacking Atp8b1 exhibit a
mild form of human ATP8B1 deficiency, but do not suffer from
progressive cholestatic liver disease [12–14,16]; most of this work
was performed in 129 mice. Atp8b1 mice were previously found to
exhibit some of the characteristic phenotypic features of human
ATP8B1 deficiency, such as elevation of bile salts, bilirubin and
liver enzyme activities in serum, when fed a diet supplemented
with 0.5% cholate. Results from the analyses presented here
indicate strain-genotype interaction; the manifestation of many
phenotypic features in ATP8B1 mutant mice depends on strain
ATP8B1 mutant B6 mice manifest a number of phenotypes
that have correlates in human ATP8B1 deficiency, and are not
detected, or less readily apparent, in the 129 strain background.
In the B6 strain, Atp8b1 mutation is associated with lower serum
cholesterol, higher sALP, and higher serum bilirubin, mirroring
Figure 5. Serum Bile salts levels are higher in ATP8B1 mutant mice. Serum bile salt levels in WT (A) and Atp8b1 mutant (B) mice of B6, 129,
and F1backgrounds at baseline (light grey) and after feeding of cholate (black) or control (medium grey) diet for 4–8 days; means and SEM are
shown. N’s at baseline: B6 (n=9215), 129 (n=48243), and F1(n=43245). N’s for cholate diet: B6 (n=7213), 129 (n=18235), and F1(n=13218);
and control diet: B6 (n=529), 129 (n=22229), and F1(n=11216).
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org6February 2010 | Volume 5 | Issue 2 | e8984
findings in human ATP8B1-deficient patients [18–20]. Similarly,
slow growth during the nursing period, and substantial weight
loss on cholate diet, phenotypes reminiscent of the failure-to-
thrive seen in human patients, are most notable in mutant mice of
B6 background. Mutant mice of B6 background also exhibit
greater hepatomegaly. Results of the hepatic lipid assays
performed, as well as histological evaluation of a subset of study
samples (data not shown), indicate that this enlargement is not
due to lipid accumulation, in agreement with findings in human
ATP8B1-deficient patients. Instead, histologic assessment of
mitotic activity (data not shown) suggests that B6 mutant mice
may have increased cell proliferation, relative to 129 and F1
mutant mice. With respect to these phenotypes, the ATP8B1
mutant mouse in the B6 background strain may be a better
model of human ATP8B1 deficiency than is the mutant mouse in
the 129 strain.
A consistent finding in our study was that serum cholesterol was
decreased in mutant, as compared to WT, mice in the B6
background; after cholate feeding, this mutation effect was present
in all strains. It has been well-established both in mice and humans
that cholestasis leads to a decrease in HDL and its main
apolipoprotein apoA1 [21,22]. On the other hand, some forms
Figure 6. Composition of gallbladder bile after feeding of cholate-supplemented diet is influenced by ATP8B1 mutation and strain.
Bile cholesterol, phospholipid, and bile salt levels in WT (A, C, & E) and ATP8B1 mutant (B, D, & F) mice of B6, 129, and F1backgrounds after feeding of
cholate diet for 4–8 days; means and SEM are shown. N’s: B6 (n=11219), 129 (n=39252), and F1(n=19222).
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org7 February 2010 | Volume 5 | Issue 2 | e8984
of chronic cholestasis are associated with increased cholesterol in
the VLDL/LDL fraction, at least partly due to the appearance of
Lipoprotein X (LpX) in the serum . Formation of LpX
critically depends on the canalicular transporters responsible for
biliary lipid secretion and LpX is not found in states of cholestasis
caused by transport defects, including PFIC [24,25]. In addition,
in mice the majority of serum cholesterol is in HDL, and therefore
decreased HDL formation in cholestasis will have a lowering effect
on total serum cholesterol. Feeding of a cholate-supplemented diet
aggravates the intrahepatic cholestasis, which will lead to a further
reduction of serum cholesterol.
Results for most of the evaluated phenotypes indicate greater
abnormalities in mutant mice of B6, as compared to 129, strain
background; however, regarding bile composition, we detect
differences between WT and mutant mice of 129 background
that are not apparent in the B6 strain. We have previously
reported that ATP8B1 mutant mice characteristically have
increased biliary secretion of cholesterol compared with WT of
the same genetic background [13,14]. This enhanced secretion
occurs independent of the function of Abcg5/8 and therefore
most likely represents direct extraction of cholesterol due to a
reduced resistance of the canalicular membrane to the detergent
action of bile salts. In the present study, we find reduced
concentration of cholesterol (and phospholipids) in bile of mutant
mice of 129 and F1background. These contrasting results are
likely due to important differences between these studies. In
previous studies, we evaluated hepatic bile secretion after acute
infusion of taurocholate, while the current study is focused on
composition of gallbladder bile after chronic challenge by feeding
of a cholate-supplemented diet. Based on these observations, we
hypothesize that the increased cholesterol secretion occurs only
during acute bile salt challenge, in mutant mice of 129 and F1
background. Cholesterol extraction from the canalicular mem-
brane subsequently leads to decreased membrane cholesterol
content and cholestasis, also reducing the normal Abcg5/8
mediated cholesterol secretion into bile . Hence, in a chronic
situation reduced cholesterol secretion may be observed, as
opposed to increased cholesterol secretion in an acute situation of
bile salt infusion.
Our findings imply the existence of modifier loci regulating the
ATP8B1 mutant phenotype; the presence of similar loci in people
may underlie, at least in part, the varying severity and nature of
disease manifestations that can be seen, even between patients
carrying the same, or similar, ATP8B1 mutations . For many of
these phenotypes, mutant F1and 129 mice are similar to each
other, while mutant B6 mice differ, and are more abnormal. These
findings suggest that susceptibility to cholestasis-related pheno-
types in these mice is recessive; however such determination can
only be made after studying the phenotypes in an experimental
cross. Genetic mapping studies of Atp8b1 mutant mice, employing
an intercross, or a backcross onto B6 background, will allow us to
identify modifier loci of ATP8B1 mutant phenotypes. Mapping of
such loci in mice, and correlating of results with human genotype-
phenotype studies, may provide novel insight into the function of
ATP8B1 and the biological mechanisms of ATP8B1 deficiency in
We have also identified strain-dependent differences in WT
mice apparent at baseline and/or after dietary challenge. B6 WT
mice have lower serum cholesterol (both sexes), and higher sALP
(females), as well as greater weight loss (both sexes), and more
enlarged livers (both sexes) than do WT 129, and sometimes F1,
mice. Some of these findings are magnified upon cholate feeding,
suggesting a greater innate sensitivity to cholate feeding in the B6,
than 129, strain. Our findings may have general implications for
choice of strain when studying hepatobiliary phenotypes, either in
WT mice, or in mice with targeted mutations in other genes
influencing hepatobiliary phenotypes.
Materials and Methods
All mice were maintained in a specific-pathogen-free animal
facility in San Francisco; studies were conducted under a protocol
approved by the UCSF IACUC. The embryonic stem cell line
used in generation of the mice was derived from the 129S4 strain,
and mice of the genetically most closely related commercially
available 129 substrain, 129S1, were subsequently used for
breeding . We therefore backcrossed the Atp8b1 mutation
separately onto the B6, 129S1, and 129S4 strains. The studies
Figure 7. ATP8B1 mutant B6 mice have larger livers than mutant 129 mice. Liver weight as a proportion of final body weight in WT (A) and
ATP8B1 mutant (B) mice of B6, 129, and F1 backgrounds after feeding of cholate (black) or control (grey) diet for 4–8 days; means and SEM are shown.
N’s for cholate diet: B6 (n=7213), 129 (n=18235), and F1 (n=13218); and control diet: B6 (n=529), 129 (n=22229), and F1 (n=11216).
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org8 February 2010 | Volume 5 | Issue 2 | e8984
Figure 8. Hepatic levels of triglycerides and cholesterol are influenced by diet and Atp8b1 mutation. Hepatic triglycerides and
cholesterol in WT (A, C, E, and G) and mutant (B, D, F, and H) mice of B6, 129, and F1 background after cholate (black) and control (grey) diet for 4–8
days; mean and SEM are shown. N’s for cholate diet: B6 (n=4210), 129 (n=8220), and F1 (n=429); and control diet: B6 (n=4211), 129 (8219),
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org9 February 2010 | Volume 5 | Issue 2 | e8984
reported here were performed during the course of this
backcrossing. For B6, the first experiments were performed on
mice after 5 generations of backcrossing, with a median of 12
backcross generations for phenotyped mice. For phenotypes
presented here, little or no effect of 129 substrain (129S1 versus
129S4) was detected, so data for these 129 substrains were pooled.
A minority of WT mice was of pure strain stock, rather than
derived from backcrossing. For F1mice, similar numbers were
generated with B6 as the paternal, and as the maternal, strain.
At approximately 21 days after birth, pups were weaned,
weighed and tagged. For many litters born to heterozygote
couples, pups had also been weighed at approximately day 10, and
uniquely identified at that time, so that weight gain during the 2nd
half of the nursing period could be evaluated; we did not want to
disturb the litters earlier than that, to avoid increasing the risk of
mothers abandoning their litters. To account for effects of factors
such as litter size and precise age in days on pup weight, weights of
mutant pups were normalized to those of WT and heterozygous
Mice were aged a minimum of 3 months prior to study diet
administration (Dyets, Inc. catalog number 10191460.5% cholic
acid, sodium salt [Calbiochem]; a small, initial pilot study was
performed with a highly similar diet [K4068.02, Arie Blok
Diervoeders, Woerden, The Netherlands]). Mice were anesthe-
tized, and a ‘baseline’ blood collection was performed. Then,
standard mouse chow was replaced with control or cholate-
supplemented diet. To optimize the number of days mice would
be on study diet, we assessed impact of diet administration for
varying lengths of time, monitoring mouse body weight and
condition. Mutant B6 mice could not consistently remain on the
cholate diet for .6 days without demonstrating excessive weight
loss. Therefore, .95% of the mice in this study underwent dietary
challenge for 6–7 days. To make fullest use of data from animals
studied while optimizing diet length, we assessed impact of
number of days on diet on phenotypes. Regression analysis did not
identify differences attributable to number of days on diet over 4–8
days, so data from all mice on the diet for 4–8 days were pooled for
analysis. In total, results from 396 mice are included in this study,
tallied by strain as follows: 72 B6, 108 129S1, 100 129S4, and 116
F1(the latter including mice derived from crossing B6 with 129S1,
and B6 with 129S4).
We established a standard protocol in which mice were fasted
for $4 hours, then anesthetized, and blood, gallbladder bile,
liver, and spleen (for isolation of DNA to confirm genotype)
were collected at sacrifice. Serum ALP, cholesterol, and
bilirubin were assayed in a clinical laboratory. Serum bile salts
were assayed using the Total Bile Acid Assay kit (DZ042A-K,
Diazyme Labs, USA). Cholesterol, phospholipids, and bile salts
in gallbladder bile were assayed as previously described .
Free cholesterol, total cholesterol, and total triglycerides were
measured in lipid extracts from snap-frozen liver tissue using
commercial assay kits (Wako Free Cholesterol E [435-35801,
Wako Diagnostics]; Infinity cholesterol reagent [TR13521,
Fisher Diagnostics]; Infinity triglyceride reagent [TR22321,
Since 24 factorial combinations are present, there were 276
possible ways of grouping the mice into two groups based on their
mutation status, genetic background, diet and sex. To reduce the
number of comparisons examined, and simplify the process of
determining which factors affect a trait of interest, we adopted the
following procedure: For each trait of interest, we fit a full factorial
model with all main effects, two-factor, three-factor and the four-
factor interaction . Then we performed backward selection
using the Bayesian Information Criterion . This approach
balanced the explanatory power of the model against model
complexity, yielding a parsimonious list of factors that influenced
the trait of interest. Lower order terms were always included if a
higher-order interaction was present. For most phenotypes we used
linear regression to fit the models. Serum bilirubin was analyzed as
an ordinal variable. Baseline bilirubin data were binned into 2
levels, normal (,0.1 mg/dl) and elevated ($0.1 mg/dl), and
analyzed using a binomial linear model. Post-diet data were binned
into 3 levels, normal (,0.1 mg/dl), moderatelyelevated (0.1–2 mg/
dl), or highly elevated (.2 mg/dl), and analyzed using proportional
we used ridge regression, using the equivalent of one mouse with no
association with any of the factors. These analyses were designed
using the R programming language ; for the annotated code,
see Supplementary Files S1, S2, S3.
P-values reported in the text are derived from analysis of
variance (ANOVA) with Tukey’s post-test (performed on the
relevant sub-groups), or the chi-square test. These analyses were
performed using PRISM 5.0 (Graphpad Software, Inc.) or
programmed in Excel (Microsoft Corp.). Except for serum
bilirubin and rate of weight loss, data were log-transformed for
all analyses. (As serum bile salt levels were undetectable in a small
number of mice, and therefore recorded as zero, we added 0.1
prior to log transformation.) For subtle effects, significance is
occasionally obtained in the factorial analysis, but not in ANOVA,
due to differences in power.
Found at: doi:10.1371/journal.pone.0008984.s001 (0.00 MB
Found at: doi:10.1371/journal.pone.0008984.s002 (0.01 MB
Found at: doi:10.1371/journal.pone.0008984.s003 (0.01 MB
We would like to thank R. Jaenisch for providing the 129S4 WT mouse
line, M.D. Kendrick for technical assistance, A.S. Knisely and S.K.
Erickson for helpful discussion, and the Pathology & Imaging Core of the
UCSF Liver Center (P30 DK026743) for assistance with histopathology.
Conceived and designed the experiments: AG CP LP SS ROE LNB.
Performed the experiments: URS JCV HX AG CP JPG ROE LNB.
Analyzed the data: SS JPG SS LNB. Wrote the paper: SS URS AG CP
JPG LP SS ROE LNB. Proofread the manuscript: JCV. Reviewed the
1. Bull LN, van Eijk MJ, Pawlikowska L, DeYoung JA, Juijn JA, et al. (1998) A
gene encoding a P-type ATPase mutated in two forms of hereditary cholestasis.
Nat Genet 18: 219–224.
2. Tang X, Halleck MS, Schlegel RA, Williamson P (1996) A subfamily of P-type
ATPases with aminophospholipid transporting activity. Science 272:
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org10 February 2010 | Volume 5 | Issue 2 | e8984
3. Folmer DE, Elferink RP, Paulusma CC (2009) P4 ATPases - Lipid flippases and
their role in disease. Biochim Biophys Acta 1791: 628–635.
4. Muthusamy BP, Natarajan P, Zhou X, Graham TR (2009) Linking
phospholipid flippases to vesicle-mediated protein transport. Biochim Biophys
Acta 1791: 612–619.
5. Ujhazy P, Ortiz D, Misra S, Li S, Moseley J, et al. (2001) Familial intrahepatic
cholestasis 1: Studies of localization and function. Hepatology 34: 768–775.
6. Paulusma CC, Folmer DE, Ho-Mok KS, de Waart DR, Hilarius PM, et al.
(2008) ATP8B1 requires an accessory protein for endoplasmic reticulum exit and
plasma membrane lipid flippase activity. Hepatology 47: 268–278.
7. Cai SY, Gautam S, Nguyen T, Soroka CJ, Rahner C, et al. (2008) ATP8B1
Deficiency Disrupts the Bile Canalicular Membrane Bilayer Structure in
Hepatocytes, But FXR Expression and Activity Are Maintained. Gastroenter-
ology 136: 1060–9.
8. Klomp LW, Bull LN, Knisely AS, van Der Doelen MA, Juijn JA, et al. (2000) A
missense mutation in FIC1 is associated with greenland familial cholestasis.
Hepatology 32: 1337–1341.
9. Klomp LW, Vargas JC, van Mil SW, Pawlikowska L, Strautnieks SS, et al.
(2004) Characterization of mutations in ATP8B1 associated with hereditary
cholestasis. Hepatology 40: 27–38.
10. Clayton RJ, Iber FL, Ruebner BH, McKusick VA (1969) Byler disease. Fatal
familial intrahepatic cholestasis in an Amish kindred. Am J Dis Child 117:
11. Linarelli LG, Williams CN, Phillips MJ (1972) Byler’s disease: fatal intrahepatic
cholestasis. J Pediatr 81: 484–492.
12. Pawlikowska L, Groen AK, Eppens EF, Kunne C, Ottenhoff R, et al. (2004) A
mouse genetic model for familial cholestasis caused by ATP8B1 mutatons reveals
perturbed bile salt homeostasis but no impairment in bile secretion. Hum Mol
Genet 15: 881–892.
13. Paulusma CC, Groen A, Kunne C, Ho-Mok KS, Spijkerboer AL, et al. (2006)
Atp8b1 deficiency in mice reduces resistance of the canalicular membrane to
hydrophobic bile salts and impairs bile salt transport. Hepatology 44: 195–204.
14. Groen A, Kunne C, Jongsma G, van den Oever K, Mok KS, et al. (2008)
Abcg5/8 independent biliary cholesterol excretion in Atp8b1-deficient mice.
Gastroenterology 134: 2091–2100.
15. Groen A, Kunne C, Oude Elferink RP (2006) Increased serum concentrations of
secondary bile salts during cholate feeding are due to coprophagy. A study with
wild-type and Atp8b1-deficient mice. Mol Pharm 3: 756–761.
16. Groen A, Kunne C, Paulusma CC, Kramer W, Agellon LB, et al. (2007)
Intestinal bile salt absorption in Atp8b1 deficient mice. J Hepatol 47: 114–122.
17. Cox DR, Reid N (2000) The Theory of the Design of Experiments: CRC Press.
18. Nagasaka H, Chiba H, Hui SP, Takikawa H, Miida T, et al. (2007) Depletion of
high-density lipoprotein and appearance of triglyceride-rich low-density
lipoprotein in a Japanese patient with FIC1 deficiency manifesting benign
recurrent intrahepatic cholestasis. J Pediatr Gastroenterol Nutr 45: 96–105.
19. Nagasaka H, Yorifuji T, Kosugiyama K, Egawa H, Kawai M, et al. (2004)
Resistance to parathyroid hormone in two patients with familial intrahepatic
cholestasis: possible involvement of the ATP8B1 gene in calcium regulation via
parathyroid hormone. J Pediatr Gastroenterol Nutr 39: 404–409.
20. Pawlikowska L, Strautnieks S, Jankowska I, Czubkowski P, Emerick K, et al.
Differences in presentation and progression between severe FIC1 and BSEP
deficiencies. J Hepat in press.
21. Claudel T, Sturm E, Duez H, Torra IP, Sirvent A, et al. (2002) Bile acid-
activated nuclear receptor FXR suppresses apolipoprotein A-I transcription via a
negative FXR response element. J Clin Invest 109: 961–971.
22. Kimmings N, Sewnath ME, Mairuhu WM, Van Zanten AP, Rauws EA, et al.
(2001) The abnormal lipid spectrum in malignant obstructive jaundice in
relation to endotoxin sensitivity and the result of preoperative biliary drainage.
Surgery 129: 282–291.
23. Hamilton RL, Havel RJ, Kane JP, Blaurock AE, Sata T (1971) Cholestasis:
lamellar structure of the abnormal human serum lipoprotein. Science 172:
24. Elferink RP, Ottenhoff R, van Marle J, Frijters CM, Smith AJ, et al. (1998) Class
III P-glycoproteins mediate the formation of lipoprotein X in the mouse. J Clin
Invest 102: 1749–1757.
25. Nagasaka H, Yorifuji T, Egawa H, Yanai H, Fujisawa T, et al. (2005) Evaluation
of risk for atherosclerosis in Alagille syndrome and progressive familial
intrahepatic cholestasis: two congenital cholestatic diseases with different
lipoprotein metabolisms. J Pediatr 146: 329–335.
26. Paulusma CC, de Waart DR, Kunne C, Mok KS, Elferink RP (2009) Activity of
the bile salt export pump (ABCB11) is critically dependent on canalicular
membrane cholesterol content. J Biol Chem 284: 9947–9954.
27. Groen AK, Van Wijland MJ, Frederiks WM, Smit JJ, Schinkel AH, et al. (1995)
Regulation of protein secretion into bile: studies in mice with a disrupted mdr2
p-glycoprotein gene. Gastroenterology 109: 1997–2006.
28. Schwarz G (1978) Estimating the Dimension of a Model. Ann Stat 6: 461–464.
29. R Development Core Team. (2005) R: A language and environment for
statistical computing. R Foundation for Statistical Computing. Vienna: Austria.
ISBN 3-900051-07-0, URL http://www.R-project.org.
ATP8B1 Mutant Mouse and Strain
PLoS ONE | www.plosone.org11 February 2010 | Volume 5 | Issue 2 | e8984