-
The Author 2013. Published by Oxford University Press on behalf
of the Society of Toxicology. All rights reserved. For permissions,
please email: [email protected].
Evaluation of Serum Bile Acid Profiles as Biomarkers of Liver
Injury in Rodents
LinaLuo,* ShelliSchomaker,* ChristopherHoule, JiriAubrecht,* and
Jennifer L.Colangelo*,1
*Biomarkers of Drug Safety Research and Development and
Toxicologic Pathology of Drug Safety Research and Development,
Pfizer Inc., Groton, Connecticut 06340
1To whom correspondence should be addressed at Biomarkers of
Drug Safety Research and Development, Pfizer Inc., 8274-1429
Eastern Point Road, Groton, CT 06340. Fax: (860) 715-8045. E-mail:
[email protected].
Received June 19, 2013; accepted September 9, 2013
Bile acids (BAs) have been studied as potential biomarkers of
drug-induced liver injury. However, the relationship between
lev-els of individual BAs and specific forms of liver injury
remains to be fully understood. Thus, we set out to evaluate cholic
acid (CA), glycocholic acid (GCA), and taurocholic acid (TCA) as
potential biomarkers of liver injury in rodent toxicity studies. We
have developed a sensitive liquid chromatography-tandem mass
spec-trometry (LC/MS/MS) assay applicable to rat and mouse serum
and evaluated levels of the individual BAs in comparison with the
classical biomarkers of hepatotoxicity (alanine aminotransferase
[ALT], aspartate aminotransferase [AST], glutamate dehydro-genase
(GLDH), alkaline phosphatase, total bilirubin, gamma-glutamyl
transferase, and total BAs) and histopathology findings in animals
treated with model toxicants. The pattern of changes in the
individual BAs varied with different forms of liver injury. Animals
with histopathologic signs of hepatocellular necrosis showed
increases in all 3 BAs tested, as well as increases in ALT, AST,
GLDH, and total BAs. Animals with histopathologic signs of bile
duct hyperplasia (BDH) displayed increases in only conju-gated BAs
(GCA and TCA), a pattern not observed with the other toxicants.
Because BDH is detectable only via histopathology, our results
indicate the potential diagnostic value of examining indi-vidual
BAs levels in serum as biomarkers capable of differentiat-ing
specific forms of liver injury in rodent toxicity studies.
Key Words: bile acids; drug-induced liver necrosis; bile duct
hyperplasia; LC/MS/MS; biomarkers.
Adverse drug reactions, especially drug-induced liver injury
(DILI), represent a major challenge for drug develop-ment.
Hepatotoxicity has been considered the most frequent cause of
safety-related drug withdrawals for the past 50years (FDA, 2009;
Kola and Landis, 2004; Lazarou et al., 1998; Pirmohamed etal.,
2004). Serum enzymatic activity of alanine aminotransferase (ALT)
is considered the gold standard clini-cal chemistry biomarker of
liver injury in preclinical species and humans (Amacher, 2002;
Amacher etal., 1998; Ozer etal.,
2010). However, ALT assessments in preclinical studies may
present a challenge, especially when increases in ALT activity do
not correlate with histopathology findings (Ennulat etal., 2010).
In many cases, these increases can be attributed to induc-tion or
extrahepatic injury, such as muscle damage or metabolic state, and
might be addressed by additional biochemical param-eters. On the
other hand, histopathologic analysis may be the only indicator for
certain types of liver injury, such as bile duct hyperplasia (BDH).
BDH can occur secondary to other abnor-malities, such as portal
inflammation, cholestasis, and biliary and/or hepatocellular
injury, and can also occur as a primary lesion. Thus, the
development of additional biomarkers capable of facilitating the
interpretation of serum ALT increases and differentiating between
various histopathologic findings in pre-clinical toxicity studies
is important.
Efforts to identify and develop additional biomarkers for DILI
and/or to enhance the current biomarker panels have been initiated
(Amacher et al., 2005; Ozer et al., 2008). Standard clinical
chemistry panels currently include monitoring some combination of
ALT, aspartate aminotransferase (AST), alka-line phosphatase (ALP),
and gamma-glutamyl transferase (GGT) serum activity, as well as
serum concentration of total bilirubin (TBIL). However, these often
are of limited use for detecting BDH. Several research groups have
employed genomics, proteomics, and metabonomics platforms in hopes
of identifying potential biomarkers in this area. Recent atten-tion
has been given to bile acids (BAs), which have been identi-fied in
several metabonomic studies. These studies indicated that an
altered BA profile was a key characteristic of the toxic response
in the liver (Beckwith-Hall et al., 1998; Davis and Thompson, 1993;
Lin etal., 2009; Yamazaki etal., 2013).
BAs are a class of structurally similar compounds that play
essential roles in cholesterol homeostasis, lipid absorption, and
intestinal signaling (Xiang etal., 2010). Synthesized in the liver
from cholesterol, BAs are excreted into the small intestine via the
bile duct mainly as glycine or taurine conjugates and then
toxicological sciences 137(1), 1225
2014doi:10.1093/toxsci/kft221Advance Access publication October 1,
2013
by guest on April 6, 2015
http://toxsci.oxfordjournals.org/D
ownloaded from
-
Evaluation of SErum BilE acid ProfilES
undergo enterohepatic circulation with further metabolism by
bacterial and hepatic enzymes (Martin etal., 2007). Liver and
gastrointestinal diseases often disturb the hepatic synthesis and
clearance of individual BAs giving rise to quantitative changes in
the pattern of serum BAs (Burkard etal., 2005; Meng etal., 1997;
Thompson et al., 1987). In contrast, the measurement of total BAs
may provide a general overall assessment of liver function,
particularly in advanced stages of liver disease, but provide
little insight into specific liver pathology especially early on in
the disease process (Berg etal., 1986; Reyes and Sjvall, 2000).
Because the total BA test measures the sum of all serum BAs, which
is over 20 BAs in most species, and can be influenced by the
presence of other endogenous molecules, the analysis of individual
BAs has been proposed to provide valuable information regarding the
pathogenesis of toxic liver injury and disease (Alnouti etal.,
2008; Bentayeb etal., 2008; Ducroq etal., 2010; Turley and
Dietschy, 1978).
Numerous analytical methods have emerged to determine BA
concentrations in plasma and serum (Gatti etal., 1997; Lee etal.,
1997; Street etal., 1985; Thompson etal., 1987). The complexity of
metabolism, the typically low concentration of BAs in biological
fluids, and the existence of multiple isobaric structural isomers
make BA separation and quantitation chal-lenging (Janzen et al.,
2010; Ostrow, 1993). In recent years, liquid chromatography coupled
with mass spectrometry (LC/MS) has become an ideal option for the
analysis of individ-ual BAs due to the high sensitivity and
selectivity of the plat-form (Alnouti etal., 2008; Ando etal.,
2006; Bentayeb etal., 2008; Bobeldijk etal., 2008; Griffiths and
Sjvall, 2010; Hagio etal., 2009; Scherer etal., 2009). Other
benefits of utilizing the LC/MS platform for these analyses include
simple sample preparation, low sample volume requirements, and
relatively low cost for reagents. Of the 2 primary BAs, cholic acid
(CA) and chenodeoxycholic acid (CDCA), CA is more abundant in rats.
Because multiplexing individual BAs into a single LC/MS assay often
impairs accuracy and precision (Suzuki etal., 2013), we selected
the most relevant primary BA to rat, CA, and its direct conjugates,
glycocholic acid (GCA) and tauro-cholic acid (TCA), for our
studies.
Goals of this study were to develop a sensitive LC/MS method for
detection of CA, GCA, and TCA in rodent serum and to evaluate the
potential of BA serum profiles as a bio-marker of DILI with the
capability of differentiating biliary and hepatocellular damage in
rodents.
MATeriALS AnD MeTHoDS
Chemicals and reagents. CA, GCA, and TCA were purchased from
Sigma-Aldrich (St Louis, Missouri). d
4-CA and d
4-GCA were purchased
from C/D/N ISOTOPES Inc. (Quebec, Canada). d4-TCA was purchased
from
Toronto Research Chemicals Inc. (Ontario, Canada). HPLC grade
metha-nol, acetonitrile, water, and formic acid were purchased from
Honeywell Burdick & Jackson (Muskegon, Michigan).
Charcoal-stripped serum was from Bioreclamation (Westbury,
NewYork).
Galactosamine (GalN), microcystin-LR (MC),
-naphthylisothiocyanate (ANIT), acetaminophen (APAP), and
isoproterenol (ISO) were purchased from
Sigma. Drug candidates A, B, C, and D were obtained from Pfizer
(Groton, Connecticut).
Animals. All studies were approved and conducted under the
oversight of the Institutional Animal Care and Use Committee.
Studies were conducted on male Sprague Dawley rats (approximately
69 weeks old) or male CD-1 mice (approximately 68 weeks old), with
the exception of the ISO study that used Hanover Wistar rats
(approximately 69 weeks old). Hanover Wistar was selected for the
ISO study because we had conducted full characterization of the
cardiac toxicity in this strain, and no differences in serum
concentrations of individual BAs have been observed between the 2
strains of rats (data not pub-lished). Drinking water and a
standard commercial laboratory certified rodent diet were provided
ad libitum throughout the studies. All rats were in the fasted
state prior to necropsy.
General study design. Test articles (N=9) were selected to
induce liver injury or injury to other organs. These compounds
included classic hepato-toxicants, a cardiac toxicant, a testicular
toxicant, a compound known to elicit BDH, and a nontoxic comparator
compound. Within each study, a vehicle con-trol group was utilized
with the same number of animals as treatment groups. Treatments of
test article or vehicle were administrated to rats or mice for a
period of time by oral (PO) gavage or subcutaneous (SC) or
intraperitoneal (IP) injection (Table1). Blood samples were
collected prior to necropsy at the end of treatment.
Sample collection. Blood samples were collected via the vena
cava at necropsy, unless otherwise noted. Serum was collected in
tubes containing no anticoagulant and was obtained from all rats
dosed with vehicle or compound. Terminal serum samples were
collected at the time of necropsy. All serum sam-ples were stored
at 80C until analysis.
Clinical chemistry and histopathology. ALT, AST, ALP, GGT, GLDH,
TBIL, and total bile acids (TBAs) were analyzed by standard
clinical chemis-try techniques on the Siemens Advia 2400 platform.
Cerner HNA Millennium Laboratory Information System was used to
acquiredata.
Histopathologic examination of the liver was performed in all
cases with additional evaluation of other major organs when
appropriate. Organ samples were fixed in 10% buffered formalin,
embedded in paraffin, sectioned, and stained with hematoxylin and
eosin. Liver pathology was generally graded using a 4-point scheme
as follows: (1) minimal, (2) mild, (3) moderate, and (4)
marked.
Sample preparation for LC/MS/MS analysis. Fifty microliters of
each serum sample was placed into an individual well of a Sirocco
protein precipita-tion plate on the top of a 96-well collection
plate. Three hundred microliters of methanol was added to each well
for protein precipitation. Deuterated stand-ards of CA, GCA, and
TCA were used as internal standards (ISs). Ten microlit-ers of the
1.0g/ml working solution of IS mixture was added, followed by
vortex mixing. After 10 min of centrifugation, the Sirocco plate
was removed, the supernatant was evaporated to dryness under a
steady stream of N
2 at 37C.
The samples were reconstituted in 50l methanol/water (1/1,
vol/vol) and then 3l was injected onto the LC/MS system.
Calibration and quality control standard preparation. One
milligram per milliliter of BA stock solutions or IS stock
solutions were prepared by dissolv-ing each BA reference in
methanol. The individual BA stock solutions were combined and
diluted to achieve the concentration of 20g/ml for a working
standard solution or 1g/ml for a working IS solution. Nine
calibration stand-ard solutions ranging from 5 to 5000 ng/ml were
prepared by serially diluting the working standard solution into
charcoal-stripped serum. Quality control (QC) standards were
prepared in the same manner at 20, 500, and 4000 ng/ml in stripped
serum. The calibration standards and QC standards then went through
the sample preparation process described above.
Individual bile acid analysis. Individual bile acid (IBA)
analysis was performed by LC/MS/MS. The mass spectrometer was an
ABSciex 5500 QTrap equipped with Turbo Spray ion source, operating
in negative mode. An Acquity UPLC system was interfaced to the
front end of the MS system. All chromatographic separations were
performed by gradient elution with an
13
by guest on April 6, 2015
http://toxsci.oxfordjournals.org/D
ownloaded from
-
Luo etaL.
TA
BL
e1
In V
ivo
Ani
mal
Stu
dies
Cat
egor
yC
ompo
und
Dos
e L
evel
sA
dmin
istr
atio
nV
ehic
leT
reat
men
tN
o. o
f Ani
mal
s/G
roup
Toxi
cant
s w
ith n
ecro
sis
Gal
N10
00 m
g/kg
IPPB
S24
h5
rats
MC
0.2
mg/
kgIP
PBS
24 h
8 ra
tsA
NIT
30 m
g/kg
PO0.
5% m
ethy
lcel
lulo
se24
h10
rat
s10
0 m
g/kg
PO0.
5% m
ethy
lcel
lulo
se24
h10
rat
sA
PAP
1000
mg/
kgPO
0.5%
met
hylc
ellu
lose
24 h
5 ra
tsB
DH
A15
0 m
g/kg
PO0.
5% m
ethy
lcel
lulo
se a
nd 5
% P
EG
400
5da
ys5
rats
300
mg/
kgPO
0.5%
met
hylc
ellu
lose
and
5%
PE
G40
05
days
5 ra
ts50
0 m
g/kg
PO0.
5% m
ethy
lcel
lulo
se a
nd 5
% P
EG
400
5da
ys5
rats
/6 m
ice
1000
mg/
kgPO
0.5%
met
hylc
ellu
lose
and
5%
PE
G40
05
days
5 ra
ts/6
mic
eB
DH
neg
ativ
e co
ntro
lB
150
mg/
kgPO
0.5%
met
hylc
ellu
lose
and
5%
PE
G40
05
days
5 ra
ts30
0 m
g/kg
PO0.
5% m
ethy
lcel
lulo
se a
nd 5
% P
EG
400
5da
ys5
rats
ISO
100g
/kg
SCPB
S6
h4
rats
500g
/kg
SCPB
S6
h4
rats
4000
g/
kgSC
PBS
6 h
4 ra
ts10
0g
/kg
SCPB
S24
h4
rats
500g
/kg
SCPB
S24
h4
rats
4000
g/
kgSC
PBS
24 h
4 ra
tsC
5 m
g/kg
PO20
% P
EG
400
and
20%
hyd
roxy
prop
yl b
etac
yclo
dext
rin
7da
ys
5 ra
ts50
mg/
kgPO
20%
PE
G40
0 an
d 20
% h
ydro
xypr
opyl
bet
acyc
lode
xtri
n7
days
5
rats
500
mg/
kgPO
20%
PE
G40
0 an
d 20
% h
ydro
xypr
opyl
bet
acyc
lode
xtri
n7
days
5
rats
75 m
g/kg
PO20
% h
yrox
ypro
pyl b
etac
yclo
dext
rin
21d
ays
5 ra
tsN
egat
ive
cont
rols
(ot
her
orga
n to
xica
nts)
D15
0 m
g/kg
PO20
% h
yrox
ypro
pyl b
etac
yclo
dext
rin
21d
ays
5 ra
ts
14
by guest on April 6, 2015
http://toxsci.oxfordjournals.org/D
ownloaded from
-
Evaluation of SErum BilE acid ProfilES
Acquity Shield RP18 column, 50 2.1 mm, 1.7m, maintained at 55C
at a flow rate of 400l/min. The gradient program started at 65%
mobile phase A(0.05% formic acid in water) and 35% mobile phase B
(5% acetonitrile in methanol), increased to 90% of B in 7 min,
decreased to 35% B in 0.5 min, and then held at 35% B for 2.5 min.
The mass spectrometer was operated with the source and desolvation
temperatures set at 120C and 350C, respectively. The curtain gas
was 40 psi; the ion spray voltage was 4500 V; probe temperature was
600C; and ion source gas 1 and ion source gas 2 were 40 and 30 psi,
respectively. Deuterated standards d
4-TCA, d
4-GCA, and d
4-CA were used as
ISs for TCA, GCA, and CA, respectively. Analysis in the mass
spectrometer was performed in multiple reaction monitoring (MRM)
mode. The MRM tran-sitions (m/z) were 407.3>407.3 for CA,
464.3>74.3 for GCA, 514.3>80.3 for TCA, 411.3>411.3 for
d
4-CA, 468.2>73.8 for d
4-GCA, and 518.1>79.7 for d
4-
TCA. Peak integration and quantification were performed using
Analyst 1.5.1 software. Individual standard curves for each BA were
constructed by plotting the ratio of the BA peak area to its
deuterated standard peak area versus con-centration. Slope and
y-intercept were calculated using a linear curve fit with 1/x2
weighting. The concentrations of BAs in study samples were
calculated relative to the regressionline.
Freshly prepared standard curve and QC samples were included in
each analysis run. Arun was deemed acceptable if the QC samples
were 15% of the nominal concentrations and the coefficient of
variance (CV) did not exceed 10%.
Method validation. The method was validated using QC samples at
3 concentration levels (20, 200, and 2000 ng/ml) from the
calibration curve. Four replicates of each QC sample were analyzed
in a single run to determine the intraassay accuracy and precision.
This process was repeated 4 times over 4days in order to determine
the interassay accuracy and precision. Accuracy and precision were
calculated from the % relative error (RE) [%(meas-ured
theoretical)/theoretical concentration] and relative standard
deviation [%RSD=% standard deviation/mean], respectively. The assay
recovery was evaluated by comparing the mean detector response of
extracted QC samples at low, medium, and high concentrations (20,
200, and 1000 ng/ml) in 4 rep-licates to those of postextracted
serum blanks spiked at equal concentrations. The matrix effect was
estimated by comparing the extracted serum residue and the neat
solution.
Statistical analysis. Data are presented as individual animals
or group mean SD. Statistical analyses were conducted by 2-tailed
Students t test to compare drug treatment groups with vehicle
control groups. Values signifi-cantly different from control are
indicated as **p < .01 and *p < .05.
reSuLTS
Development of an LC/MS Assay for Detection of CA, GCA, and
TCA
We have developed a sensitive method for the quantifica-tion of
BAs using LC/MS. Because BAs are endogenous molecules and already
present in rodent serum, we used char-coal-stripped rat serum for
the standard curve and QC sample preparations. To increase the
method accuracy and precision, deuterium-labeled ISs for each
corresponding BA were used for the quantification. The accuracy and
precision of TCA have been reported to be acceptable only when
d
4-TCA was used
as the IS (Xiang et al., 2010). All BAs gave excellent linear
response over a 103 dynamic range of 55000 ng/ml, with
coef-ficients of determination (R2) above 0.999. The lower
detection limit for these BAs was 0.5 ng/ml. Mean intraassay
accuracy was 95%109% for the 3 IBAs, with a mean CV of 3%8%; mean
interassay accuracy was 97%106%, with a mean CV
of 4%10% (Table2). The mean recovery rates of the extrac-tion
procedures were between 103% and 105%. No significant matrix effect
was observed, and the signal difference was less than 5%. These
values were within the acceptable range, and the method was judged
to be suitably accurate and precise for the analytes.
Currently, no published data can be found on the reference
ranges of individual serum BA levels. The 3 IBA serum
con-centrations from various vehicle-treated rodents in toxicology
studies were generated here to provide baseline endogenous lev-els.
Endogenous levels of TCA, GCA, and CA in various vehi-cle-treated
rats (n=46) measured by this LC/MS assay were 0.184 0.153g/ml,
0.559 0.333g/ml, and 5.455 2.196g/ml, respectively; the serum
levels of TCA, GCA, and CA in various vehicle-treated mice (n = 18)
were 0.470 0.464g/ml, 0.004 0.002g/ml, and 0.091 0.11g/ml,
respectively.
Serum BA Profiles After Treatment With Model Liver Toxicants
As expected, the treatment of rats with model liver toxicants
GalN, MC, ANIT, and APAP caused a wide degree of
hepato-cellular/hepatobiliary effects detected via histopathology
and serum biochemical analyses (Table3). As expected, no
histo-pathologic or serum biomarker changes were observed in
vehi-cle-treated animals.
Rats dosed with a single dose of 1000 mg/kg GalN for 24 h
revealed moderate to marked panlobular hepatocellular necro-sis
that was randomly distributed throughout the liver (Fig.1A).
Statistically significant, treatment-related increases in serum
ALT, AST, and GLDH activity levels, together with increased
concentrations of total BAs, were also noted. The LC/MS anal-ysis
of CA, GCA, and TCA levels from the GaIN-treated rats showed
statistically significant elevations of serum concentra-tions for
all measured BAs. CA, as one of the primary BAs in rats, remained
the most abundant BA in GalN-treated ani-mals (Fig.2).
Interestingly, the BA proportions changed after treatment with GalN
with CA accounting for 51% of the total 3 BAs measured as compared
with 93% in the control group. Conversely, TCA accounted for 41% of
the total 3 BAs in the profile after treatment compared with 2% in
controls.
Rats dosed with a single dose of 1000 mg/kg APAP for 24 h
produced characteristic hepatocellular necrosis that was
accom-panied by significant increases in ALT, AST, ALP, and GLDH
activity and elevation of total BA concentrations (Fig.1B). No
other clinical chemistry parameters, such as GGT and TBIL,
exhibited significant changes. LC/MS analysis revealed
sta-tistically significant changes in serum concentrations of the 3
IBAs when compared with their corresponding control groups
(Table3). APAP treatment resulted in an IBA profile similar to GalN
treatment with CA remaining the major BA (77% of the total 3 BAs)
as shown in Figure2.
Treatment of rats with 30 or 100 mg/kg of ANIT for 24 h caused
mild to moderate hepatobiliary portal inflamma-tion and biliary
degeneration and necrosis in most animals
15
by guest on April 6, 2015
http://toxsci.oxfordjournals.org/D
ownloaded from
-
Luo etaL.
TA
BL
e2
Per
form
ance
of
the
QC
Sta
ndar
ds in
the
LC
/MS
Qua
ntit
ativ
e A
ssay
for
3 B
As
IBA
s
0.02
g/
ml
0.2
g/m
l2
g/m
l
Ave
rage
(ng
/ml)
%R
SD%
RE
Ave
rage
(ng
/ml)
%R
SD%
RE
Ave
rage
(ng
/ml)
%R
SD%
RE
Intr
aday
val
idat
ion
(n=
4)
C
A0.
0198
399
0.19
37
972.
178
109
G
CA
0.02
003
100
0.20
85
104
2.06
510
3
TC
A0.
0190
395
0.20
15
100
2.17
510
8In
terd
ay v
alid
atio
n (n
=4
)
CA
0.01
966
980.
193
497
2.11
810
6
GC
A0.
0209
1010
40.
200
610
02.
046
102
T
CA
0.02
078
104
0.19
84
992.
046
102
Abb
revi
atio
ns: %
RSD
=%
sta
ndar
d de
viat
ion/
mea
n; %
RE
=%
(mea
sure
d
theo
retic
al)/
theo
retic
al c
once
ntra
tion.
(Fig.1C). These changes were slightly more extensive in rats
given 100 mg/kg of ANIT, as compared with rats given 30 mg/kg.
Focal or multifocal, minimal, randomly scattered areas of
hepatocellular necrosis were observed in some rats given each dose.
The biomarker analysis showed statistical increases in levels of
all tested biomarkers (ALT, AST, ALP, GLDH, GGT, TBIL, and total
BAs) at both dose levels. The LC/MS analy-sis of CA, GCA, and TCA
concentrations in ANIT-treated rats revealed statistically
significant elevations. Furthermore, the proportions of the 3 IBAs
were altered markedly. Conjugated BAs, especially TCA, were
substantially elevated in both the 30 and 100 mg/kg groups (682 and
856 fold increases over the control group) and became the
predominate BA in the treated groups. The unconjugated BA (CA)
exhibited a comparatively small, yet statistically significant
increase (Fig. 2). The ratio of the taurine conjugate over the
total 3 BAs was higher in the 100 mg/kg group as compared with the
30 mg/kg group (data not shown). Paradoxically, ALT, AST, and GLDH
had larger fold changes in the 30 mg/kg dose group (4.1, 5.7, and
64 control, respectively) than in the 100 mg/kg group (1.9, 3.4,
and 35 control, respectively). The fold increases of ALP, GGT, and
TBIL, which are generally considered indicators of hepatobiliary
toxicity, were slightly increased in the 100 mg/kg group as
compared with the 30 mg/kg group. This was consist-ent with TCA
levels, the predominant BA here, which showed dose-dependent
increases from 30 to 100 mg/kg (Table3).
Treatment of rats with a single dose of 0.2 mg/kg MC for 24 h
caused moderate to marked centrilobular necrosis with fre-quent
involvement of adjacent portal tracts, which include bil-iary
tracts. These findings were accompanied by significantly increased
activity levels for ALT, AST, GLDH, ALP, TBIL, and total BAs (Fig.
1D). The LC/MS analysis of CA, GCA, and TCA levels from the rats
treated with MC showed statistically significant elevations of
serum concentrations of all measured BAs. Again, the BA proportions
changed dramatically. TCA had the largest fold increases of 903
over the control (Table3) and became the major BA accounting for
58% of the total 3 BAs compared with 1.4% in control (Fig.2).
Serum BA Profiles and BDH
To evaluate the BA profile in response to BDH, we treated rats
with a well-characterized drug candidate (compound A) that was
discontinued from development due to prevalent BDH detected in rat
safety studies. To assess specificity of the bio-marker response,
we used a structurally similar discontinued drug candidate of the
same class (compound B) that did not cause BDH in rat safety
studies. As expected, histopathologic examination of rats dosed
with compound Arevealed mild to moderate BDH in the 3 highest dose
groups, 300, 500, and 1000 mg/kg (Fig. 1E). In agreement with
previous studies, the classical biomarkers of liver injury measured
in our stud-ies (ALT, AST, ALP, GLDH, GGT, TBIL, and TBAs) failed
to detect BDH. Although total BA levels detected by an enzyme
cyclingbased assay was unchanged, the LC/MS analysis of
16
by guest on April 6, 2015
http://toxsci.oxfordjournals.org/D
ownloaded from
-
Evaluation of SErum BilE acid ProfilES
TA
BL
e3
Sum
mar
y of
Stu
dy e
ndpo
ints
for
rat
s
Liv
er H
isto
path
olog
yaB
iom
arke
rsb
Cat
egor
yC
ompo
und
Dos
eN
ecro
sis
BD
HA
LT
AST
AL
PG
LD
HT
BIL
GG
TTo
tal B
As
CA
TC
AG
CA
Toxi
cant
s w
ith
necr
osis
Gal
N10
00 m
g/kg
(2
4 h)
HC
(3)
to (
4)
No
18**
33**
NP
121*
*N
PN
P3.
4**
4.4*
*82
**7.
1**
MC
0.2
mg/
kg
(24
h)H
C (
3) to
(4)
N
o18
1**
130*
*2.
2*66
3*66
**2.
09.
3**
5.4*
*90
3*39
**
AN
IT30
mg/
kg
(24
h)B
D, H
C (
2) to
(3)
No
4.1*
*6*
*1.
3*65
**23
**1.
7*22
**3.
7**
682*
*84
**
100
mg/
kg
(24
h)B
D, H
C (
2) to
(3)
No
1.9*
*3.
4**
1.4*
35**
24**
2.4*
*22
**2.
8**
856*
*52
**
APA
P10
00 m
g/kg
(2
4 h)
HC
(3)
to (
4)N
o12
*31
*1.
4*14
1**
1.0
1.0
4.5*
*5.
3**
26*
5.7*
BD
HA
150
mg/
kg
(5d
ays)
No
No
1.2
1.1
1.0
2.0
1.0
1.0
0.7
0.8
0.7
0.8
300
mg/
kg
(5d
ays)
No
(1)
1.4
0.8
0.9
1.4
1.0
1.0
1.0
1.2
13*
6.0*
500
mg/
kg
(5d
ays)
No
(1)
to (
2)1.
40.
90.
91.
41.
01.
01.
41.
727
*18
**
1000
mg/
kg
(5d
ays)
No
(2)
1.4
0.8
0.9
1.3
1.0
1.0
1.3
1.5
69**
10**
BD
H n
egat
ive
cont
rol
B15
0 m
g/kg
(5
day
s)N
oN
o1.
01.
20.
90.
81.
01.
00.
831.
10.
60.
6
300
mg/
kg
(5d
ays)
No
No
0.9
0.8
1.0
0.8
1.0
1.0
0.84
0.7
1.6
1.6
17
by guest on April 6, 2015
http://toxsci.oxfordjournals.org/D
ownloaded from
-
Luo etaL.
Neg
ativ
e co
ntro
ls
(oth
er o
rgan
to
xica
nts)
ISO
c10
0g
/kg
(6 h
)N
oN
o1.
21.
00.
82.
31.
01.
01.
21.
21.
30.
7
500g
/kg
(6 h
)N
oN
o1.
21.
9*0.
81.
11.
01.
02.
8*2.
01.
70.
5
4000
g/
kg
(6 h
)N
oN
o1.
23.
2**
0.9
1.2
1.0
1.0
1.0
1.3
0.7
0.1
100g
/kg
(24
h)N
oN
o1.
01.
01.
01.
11.
01.
02.
11.
71.
21.
3
500g
/kg
(24
h)N
oN
o1.
01.
30.
91.
21.
01.
01.
41.
21.
11.
8
4000
g/
kg
(24
h)N
oN
o1.
01.
51.
01.
41.
01.
01.
61.
30.
80.
6
Cd
5 m
g/kg
(7
day
s)N
oN
o1.
3*1.
10.
9N
P1.
01.
2N
P0.
60.
51.
1
50 m
g/kg
(7
day
s)N
oN
o1.
8**
0.9
1.1
NP
1.0
1.2
NP
0.7
0.8
0.4
500
mg/
kg
(7d
ays)
No
No
3.2*
*3.
5**
1.1
NP
1.0
1.1
NP
1.1
0.8
0.6
De
75 m
g/kg
(2
1da
ys)
No
No
1.7*
*1.
11.
3*N
P1.
01.
0N
P1.
21.
21.
2
150
mg/
kg
(21
days
)N
oN
o2.
1**
1.2
1.3*
NP
1.0
1.0
NP
0.8
0.8
0.2*
*
a HC
, hep
atoc
yte;
BD
, bile
duc
t or
hepa
tobi
llary
; the
num
ber
in th
e pa
rent
hese
s st
ands
for
the
seve
rity
of
liver
his
topa
th fi
ndin
gs: (
1) m
inim
al; (
2) m
ild; (
3) m
oder
ate;
(4)
mar
ked.
b Val
ues
repr
esen
t fol
d ch
ange
s co
mpa
red
with
thei
r co
rres
pond
ing
cont
rols
; NP,
not
per
form
ed.
c At
the
6-h
time
poin
t ch
ange
s in
the
hea
rt c
onsi
sted
pre
dom
inan
tly o
f m
inim
al t
o m
ild s
ubep
icar
dial
hem
orrh
age/
edem
a at
all
dose
lev
els;
by
24 h
, thi
s pr
ogre
ssed
to
also
inc
lude
min
imal
to
mild
m
yoca
rdia
l deg
ener
atio
n/ne
cros
is a
nd in
flam
mat
ion
at a
ll do
se le
vels
.d M
inim
al m
esen
teri
c lip
id d
eple
tion
in th
e 50
0 m
g/kg
gro
up; m
oder
ate
to m
arke
d zy
mog
en d
ecre
ase
in p
ancr
eas
and
mes
ente
ric
lipid
dep
letio
n in
the
500
mg/
kg g
roup
.e M
inim
al e
pidy
dim
al s
perm
atic
gra
nulo
mas
in th
e 75
mg/
kg g
roup
with
mild
epi
dydi
mal
inte
rstit
ial i
nflam
mat
ion
and
skel
etal
mus
cle
infla
mm
atio
n in
the
150
mg/
kg g
roup
.St
atis
tical
ly s
igni
fican
t cha
nges
are
indi
cate
d by
*p
< .0
5, *
*p