Leveraging of rifampicin-dosed cynomolgus monkeys to identify bile acid 3-O-sulfate conjugates as potential novel biomarkers for organic anion-transporting polypeptidess

Article abstract of DOI:10.1124/dmd.117.075275

In the search for novel bile acid (BA) biomarkers of liver organic anion-transporting polypeptides (OATPs), cynomolgus monkeys received oral rifampicin (RIF) at four dose levels (1, 3, 10, and 30 mg/kg) that generated plasma-free C_max values (0.06, 0.66, 2.57, and 7.79 μM, respectively) spanning the reported in vitro IC₅₀ values for OATP1B1 and OATP1B3 (≤1.7 μM). As expected, the area under the plasma concentration-time curve (AUC) of an OATP probe drug (i.v.²H₄-pitavastatin, 0.2 mg/kg) was increased 1.2-, 2.4-, 3.8-, and 4.5-fold, respectively. Plasma of RIF-dosed cynomolgus monkeys was subjected to a liquid chromatography-tandem mass spectrometry method that supported the analysis of 30 different BAs. Monkey urine was profiled, and we also determined that the impact of RIF on BA renal clearance was minimal. Although sulfated BAs comprised only 1% of the plasma BA pool, a robust RIF dose response (maximal ?50-fold increase in plasma AUC) was observed for the sulfates of five BAs [glycodeoxycholate (GDCA-S), glycochenodeoxycholate (GCDCA-S), taurochenodeoxycholate, deoxycholate (DCA-S), and taurodeoxycholate (TDCA-S)]. In vitro, RIF (≤100 μM) did not inhibit cynomolgus monkey liver cytosol-catalyzed BA sulfation and cynomolgus monkey hepatocyte-mediated uptake of representative sulfated BAs (GDCA-S, GCDCA-S, DCA-S, and TDCA-S) was sodium-independent and inhibited (≥70%) by RIF (5 μM); uptake of taurocholic acid was sensitive to sodium removal (74% decrease) and relatively refractory to RIF (≤21% inhibition). We concluded that sulfated BAs may serve as sensitive biomarkers of cynomolgus monkey OATPs and that exploration of their utility as circulating human OATP biomarkers is warranted.

Full text of DOI:10.1124/dmd.117.075275

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
DMD/2017/075275
Leveraging of Rifampicin-Dosed Cynomolgus Monkeys to Identify Bile Acid 3-O-Sulfate
Conjugates as Potential Novel Biomarkers for Organic Anion-Transporting Polypeptides
Rhishikesh Thakare, Hongying Gao, Rachel E. Kosa, Yi-An Bi, Manthena V. S. Varma,
Matthew A. Cerny, Raman Sharma, Max Kuhn, Bingshou Huang, Yiping Liu, Aijia Yu, Gregory
S. Walker, Mark Niosi, Larry Tremaine, Yazen Alnouti, and A. David Rodrigues
Pharmacokinetics, Pharmacodynamics & Metabolism, Medicine Design, Pfizer Inc., Groton,
Connecticut, USA (H. G., R. E. K., Y. B., M. V. S. V., M. A. C., R. S., G. S. W., M. N., L. T.
and A. D. R.); Nonclinical Statistics, Pfizer Inc., Groton, Connecticut, USA (M.K.); Department
of Pharmaceutical Sciences, College of Pharmacy, University of Nebraska Medical Center,
Omaha, Nebraska, USA (R. T. and Y. A.); WuXi AppTec, Waigaoqiao Free Trade Zone,
Shanghai 200131, China (B. H., Y. L. and A. Y.)
1

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
Running title: Bile Acid Sulfates as Potential Novel OATP Biomarkers
*Corresponding Author:
A. David Rodrigues, Ph.D.
Pharmacokinetics, Pharmacodynamics, & Metabolism, Medicine Design,
Eastern Point Road, Bldg 220/002/2565, Mail Stop 8220-2439,
Pfizer Inc., Groton, CT 06340 USA
Phone: +1-860-686-9311. Fax: +1-860-441-6402.
E-mail: a.david.rodrigues@pfizer.com
Number of text pages:
Number of tables:
44
3
Number of figures:
8
Number of references:
Number of words in Abstract:
47
250
Number of words in Introduction: 749
Number of words in Discussion: 1499
2

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
Abbreviations
BA, bile acid; DHEA, dehydroepiandrosterone; RIF, rifampicin; RIFsv, rifamycin SV; LC-
MS/MS, liquid chromatography-tandem mass spectrometry; PAPS, 3'-phosphoadenosine-5'-
phosphosulfate; SLC, solute carrier; DDI, drug-drug interaction; SULT2A1, sulfotransferase
2
A1; OATP, organic anion-transporting polypeptide; NTCP, sodium-taurocholate co-
transporting polypeptide; LCA, lithocholic acid; GLCA, glycolithocholic acid; TLCA,
taurolithocholic acid; UDCA, ursodeoxycholic acid; GUDCA, glycoursodeoxycholic acid;
TUDCA, tauroursodeoxycholic acid; DCA, deoxycholic acid; GDCA, glycodeoxycholic acid;
TDCA,
taurodeoxycholic
acid;
CDCA,
chenodeoxycholic
acid;
GCDCA,
glycochenodeoxycholic acid; TCDCA, taurochenodeoxycholic acid; CA, cholic acid; GCA,
glycocholic acid; TCA, taurocholic acid; LCA-S, lithocholic acid 3-O-sulfate; GLCA-S,
glycolithocholic acid 3-O-sulfate; TLCA-S, taurolithocholic acid 3-O-sulfate; UDCA-S,
ursodeoxycholic acid 3-O-sulfate; GUDCA-S, glycoursodeoxycholic acid 3-O-sulfate; TUDCA-
S, tauroursodeoxycholic acid 3-O-sulfate; CDCA-S, chenodeoxycholic acid 3-O-sulfate,
GCDCA-S, glycochenodeoxycholic acid 3-O-sulfate; TCDCA-S, taurochenodeoxycholic acid 3-
O-sulfate; DCA-S, deoxycholic acid 3-O-sulfate; GDCA-S, glycodeoxycholic acid 3-O-sulfate;
TDCA-S, taurodeoxycholic acid 3-O-sulfate; CA-S, cholic acid 3-O-sulfate; GCA-S, glycocholic
acid 3-O-sulfate; TCA-S, taurocholic acid 3-O-sulfate; OAT3, organic anion transporter 3;
AUCRplasma, AUC (area under the plasma concentration-time curve) ratio determined by dividing
the AUC0-24,plasma after RIF treatment by the AUC0-24,plasma after vehicle alone; FDR, false
discover rate; CLrenal ratio, renal clearance (CLrenal) after RIF treatment divided by the CLrenal
following vehicle alone.
3

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
ABSTRACT
In the search for novel bile acid (BA) biomarkers of liver organic anion-transporting
polypeptides (OATPs), cynomolgus monkeys received oral rifampicin (RIF) at four dose levels
(1, 3, 10 and 30 mg/kg) that generated plasma free Cmax values (0.06, 0.66, 2.57 and 7.79 µM,
respectively) that covered the reported in vitro IC s for OATP1B1 and OATP1B3 (≤ 1.7 μM).
50
As expected, the area under the plasma concentration-time curve (AUC) of an OATP probe drug
2
(
intravenous H -pitavastatin, 0.2 mg/kg) was increased 1.2-, 2.4-, 3.8-, and 4.5-fold,
4
respectively. Plasma of RIF-dosed cynomolgus monkeys was subjected to a liquid
chromatography-tandem mass spectrometry method that supported the analysis of 30 different
BAs. Monkey urine was profiled also and it was determined that the impact of RIF on BA renal
clearance was minimal. Although sulfated BAs comprised only 1% of the plasma BA pool, a
robust RIF dose response (maximal ≥50-fold increase in plasma AUC) was observed for the
sulfates of five BAs (glycodeoxycholate [GDCA-S], glycochenodeoxycholate [GCDCA-S],
taurochenodeoxycholate, deoxycholate [DCA-S], and taurodeoxycholate [TDCA-S]). In vitro,
RIF (≤ 100 μM) did not inhibit cynomolgus monkey liver cytosol-catalyzed BA sulfation and
cynomolgus monkey hepatocyte-mediated uptake of representative sulfated BAs (GDCA-S,
GCDCA-S, DCA-S, and TDCA-S) was sodium-independent and inhibited (≥ 70%) by RIF (5
μM); uptake of taurocholic acid was sensitive to sodium removal (74% decrease) and relatively
refractory to RIF (≤21% inhibition). It is concluded that sulfated BAs may serve as sensitive
biomarkers of cynomolgus monkey OATPs. Exploration of their utility as circulating human
OATP biomarkers is warranted.
4

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
INTRODUCTION
It is now accepted that organic anion-transporting polypeptides (OATPs) mediate the
active uptake of numerous drugs into hepatocytes and hence govern their pharmacokinetic
profile and liver (free)-to-plasma (free) concentration ratio. OATPs can also serve as the loci of
important drug-drug interactions (DDIs) leading to changes in systemic and local drug
concentrations, possibly resulting in altered efficacy and toxicity profiles (Giacomini et al., 2010;
Yoshida et al., 2012). Consequently, tools have been developed to facilitate OATP inhibition
screening in vitro, drive model-based DDI in vitro-in vivo extrapolations, and support OATP
DDI risk assessment prior to human dosing (Poirier et al., 2007; Jamei et al., 2014; Vaidyanathan
et al., 2016; Yoshikado et al., 2016). The latter is particularly important because OATP activity
and expression is also known to be impacted by genotype and liver disease (Gong and Kim,
2
013; Clarke et al., 2014).
More recently, it has been envisioned that OATP biomarkers will greatly facilitate
clinical phenotyping and DDI studies, while possibly deferring more formal studies employing
drug probes (Lai et al., 2016; Yee et al., 2016). For example, Lai et al (2016) recently evaluated
plasma bilirubin, bilirubin glucuronide, and coproporphyrin isomers (I and III) as OATP
biomarkers in human subjects following a single dose of rifampicin (RIF). The authors noted a
4
.0- and 3.3-fold increase in the coproporphyrin I and III area under the plasma concentration-
time curve (AUC), respectively, consistent with in vitro data (Bednarczyk and Boiselle, 2016;
Shen et al., 2016). Likewise, Yee et al (2016) identified the 3-O-sulfate conjugates of
glycochenodeoxycholic acid (GCDCA-S), glycodeoxycholic acid (GDCA-S), and
taurolithocholic acid (TLCA-S) as candidate OATP biomarkers after dosing of SLCO1B1
genotyped subjects with cyclosporine.
5

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
Because of the high sequence identity with human OATPs, the cynomolgus monkey has been
used increasingly as a model to study OATP inhibition in vivo (Shen et al., 2013; Takahashi et
al., 2013; Shen et al., 2015). The utility of the cynomolgus monkey has extended also to the
search for OATP biomarkers, which has involved the administration of a single RIF dose and
reporting its impact on the plasma bilirubin, bilirubin glucuronide, coproporphyrins (I and III),
non-sulfated bile acids (BAs) and dehydroepiandrosterone (DHEA) 3-O-sulfate (Chu et al.,
2
015; Watanabe et al., 2015; Shen et al., 2016).
As described, an attempt was made to extend the work of Chu et al (2015) by profiling 30
different BAs in cynomolgus monkey plasma following single oral doses of RIF (1, 3, 10 and 30
mg/kg). It was possible to prepare synthetic standards of numerous BA 3-O-sulfates and apply a
liquid chromatography-tandem mass spectrometry (LC-MS/MS) method that has been used
successfully to profile non-sulfated and sulfated BAs in human serum and urine (Bathena et al.,
2
013). It should be noted that at the time of the study, BA sulfation in the cynomolgus monkey
was not well characterized, although the putative hydroxysteroid sulfotransferase (SULT2A1)
that catalyses BA sulfation in humans was known to be expressed in monkey liver (Nishimura et
al., 2008; Alnouti, 2009; Nishimura et al., 2009). It was also known that DHEA 3-O-sulfate was
detectable in cynomolgus monkeys following DHEA administration (Leblanc et al., 2003).
2
In the present study, animals also received a single i.v. dose of H -pitavastatin to ensure that
4
OATP was inhibited by RIF in a dose-dependent manner (Takahashi et al., 2013). In addition,
the plasma concentrations of RIF were determined at each of the four dose levels. Sulfated BAs
are known to be cleared renally in humans (Alnouti, 2009; Bathena et al., 2013; Tsuruya et al.,
2
016), therefore BA profiling was extended to include urine of control and RIF-dosed
cynomolgus monkeys. The different BAs were then assessed in terms of their utility as OATP
6

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
biomarkers: (a) detectability in control animals; (b) magnitude of the RIF dose response; and (c)
detectability in human serum. Based on these criteria, seven sulfated BAs (GCDCA-S, GDCA-
S, glycolithocholate 3-O-sulfate [GLCA-S], TLCA-S, taurochenodeoxycholate 3-O-sulfate
[
TCDCA-S], taurodeoxycholate 3-O-sulfate [TDCA-S], and deoxycholate 3-O-sulfate [DCA-S])
were identified as potential OATP biomarkers. For GCDCA-S, GDCA-S, and TLCA-S, the
results are consistent with human plasma metabolomic data from a recent OATP1B1 (SLCO1B1)
genome-wide association study and earlier reports describing TLCA-S as an OATP substrate in
vitro (Yee et al., 2016; Sasaki et al., 2002; Meng et al., 2002). The present study also included
an in vitro assessment of BA sulfation by cynomolgus monkey liver cytosol, as well as uptake
studies with GDCA-S, GCDCA-S, DCA-S, and TDCA-S (cynomolgus monkey plated primary
hepatocytes) in order to phenotype both in terms of OATP- and sodium-taurocholate co-
transporting polypeptide (NTCP)-mediated active uptake.
7

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
MATERIALS AND METHODS
Chemicals and Reagents.
3'-Phosphoadenosine-5'-phosphosulfate (PAPS), simvastatin,
rifamycin SV (RIFsv), and rifampicin (RIF) were purchased from Sigma-Aldrich (St Louis, MO,
2
USA). Deuterium labeled RIF ( H -RIF) were obtained from ALSACHIM (Illkirch,
8
Graffenstaden, France). Pitavastatin was purchased from Sequoia Research Products Ltd.
2
(
(
(
Oxford, UK). Deuterium labeled pitavastatin ( H -pitavastatin) was purchased from Clearsynth
4
Ontario, Canada). InVitroGro-HT and CP hepatocyte media were purchased from Celsis IVT
Baltmore, MD). Collagen I coated 24-well plates were obtained from BD Biosciences (Franklin
Lakes, NJ). Cryopreserved cynomolgus monkey hepatocytes were purchased from In vitro
ADMET Laboratories, LLC (Columbia, Maryland). BCR protein assay kit was purchased from
PIERCE (Rockford, IL). RIPA protein lysis buffer was purchased from TEKNOVA (Hollister,
CA). Methanol, acetonitrile, water, ammonium hydroxide were obtained from Fisher Scientific
(Fair Lawn, NJ, USA).
Cholic acid (CA), glycocholic acid (GCA), taurocholic acid (TCA), deuterium labeled
2
taurocholic acid ( H -TCA), chenodeoxycholic acid (CDCA), deuterium labeled
4
2
chenodeoxycholic acid ( H -CDCA), taurochenodeoxycholic acid (TCDCA), deoxycholic acid
4
(DCA), glycochenodeoxycholic acid (GCDCA), glycodeoxycholic acid (GDCA),
taurodeoxycholic acid (TDCA), lithocholic acid (LCA), glycolithocholic acid (GLCA),
taurolithocholic acid (TLCA), ursodeoxycholic acid (UDCA), tauroursodeoxycholic acid
(TUDCA), glycoursodeoxycholic acid (GUDCA), and deuterium labeled glycodeoxycholic acid
2
3
-O-sulfate ( H -GDCA-S) were purchased from IsoSciences (King of Prussia, PA, USA).
4
Lithocholic acid 3-O-sulfate (LCA-S), glycolithocholic acid 3-O-sulfate (GLCA-S),
chenodeoxycholic acid 3-O-sulfate (CDCA-S), glycochenodeoxycholic acid 3-O-sulfate
8

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
2
(
GCDCA-S), and deuterium labeled glycochenodeoxycholic acid 3-O-sulfate ( H -GCDCA-S)
5
were purchased from Toronto Research Chemicals (Toronto, Ontario, Canada).
Ursodeoxycholic acid 3-O-sulfate (UDCA-S) was obtained from ALSACHIM (Illkirch,
2
Graffenstaden, France). Deuterium labeled glycochenodeoxycholic acid ( H -GCDCA) was
4
purchased from C/D/N isotopes, Inc. (Pointe-Claire, Quebec, Canada). Taurolithocholic acid 3-
O-sulfate (TLCA-S) was purchased from Sigma-Aldrich (St Louis, MO, USA)
Refer to Supplemental Information for the chemical synthesis of taurochenodeoxycholic
acid 3-O-sulfate (TCDCA-S), deoxycholic acid 3-O-sulfate (DCA-S), glycodeoxycholic acid 3-O-
sulfate (GDCA-S), taurodeoxycholic acid 3-O-sulfate (TDCA-S), cholic acid 3-O-sulfate (CA-S),
glycocholic acid 3-O-sulfate (GCA-S), taurocholic acid 3-O-sulfate (TCA-S), and biosynthesis of
glycoursodeoxycholic acid 3-O-sulfate (GUDCA-S), and tauroursodeocycholic acid 3-O-sulfate
2
2
(
TUDCA-S). Biosynthesis of H -TDCA-S, and H -DCA-S is described also. An alternative
4 4
method for preparing some of the BA sulfates described herein has been described recently by
Donazzolo et al (2017).
Refer to Supplemental Information for all BA common names, chemical names,
structures, and CAS numbers.
Animal Handling, Dosing, Plasma Draws and Urine Collection. All experiments involving
animals were conducted at the Pfizer Groton (Connecticut) facilities (Association for Assessment
&
Accreditation of Laboratory Animal Care-accredited) and were reviewed and approved by the
Pfizer Institutional Animal Care and Use Committee. Male cynomolgus macaque Mauritian
monkeys (approximately 6 to 8.5 years of age) were used for these studies. A cross-over study
design was employed, where the same four animals were dosed over a series of five studies,
9

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
following a minimum one-week wash-out period between each study. One exception was the 3
mg/kg RIF dose group, where one of four monkeys was dosed only in that single study.
Animals were provided a normal food schedule the day before the study (meals at 8:00
am and 11:00 am, with one treat daily) and were allowed free access to water. On the day of the
study, monkeys were fed at approximately 1h and 3h post-dose and allowed water ad libitum.
RIF was administered via oral gavage at 0 (blank vehicle), 1, 3, 10, and 30 mg/kg. RIF was
given at a dose volume of 2 ml/kg in a 0.5% (w/v) methylcellulose (in water) suspension.
2
Approximately one hour and 15 minutes following the oral RIF administration, H -pitavastatin
4
was administered via an intravenous (i.v.) bolus (cephalic vein), at dose of 0.2 mg/kg, in a dosing
volume of 0.2 ml/kg; 2% DMSO (v/v) and 98% of TRIS-buffered saline (pH ~7.7). All i.v.
formulations were sterile filtered prior to administration. Serial blood samples were collected via
the femoral vein before dosing into K EDTA tubes and then at 0.083, 0.25, 0.5, 0.75, 1, 2, 3, 5,
2
6
, and 24 hours post i.v. dosing. Blood samples were stored on wet ice prior to being centrifuged
◦
to obtain plasma (3000 RPM, 10 minutes at 4 C; Jouan BR4i refrigerated centrifuge). Urine was
also collected (metabolism cages), on wet ice, pre-dose and at intervals of 0 to 6 hours and 6 to
2
4 hours post-dose. Due to instability and possible inter-conversion of lactone to pitavastatin,
each plasma and urine sample was equally divided into two aliquots prior to being stored frozen.
The first was untreated matrix, while the second was added to an equal volume of 0.1 M sodium
acetate buffer (pH 4) to stabilize pitavastatin. All urine and plasma samples, treated and
untreated, were kept cold during collection, after which they were stored frozen at -20°C. It is
known that BAs undergo enterohepatic recirculation, but no attempt was made to collect portal
vein blood from the different cynomolgus monkeys. Following LC-MS/MS analysis of femoral
vein-derived plasma, it was apparent that the concentration (plasma total) of the various BA
1
0

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
sulfates was low (≤ 30 nM). For some of the BA sulfates (TCDCA-S, GCDCA-S,TDCA-S, and
GDCA-S), unbound fraction in cynomolgus monkey plasma was determined (~0.016); maximal
free plasma concentration ~0.5 nM. It is assumed that even if free BA sulfate concentrations
were 100-fold higher in the portal vein, such concentrations would still likely be below the
apparent K for cynomolgus monkey OATPs. As described in Results, RIF dosing brought
m
about robust (≥ 10-fold) dose-dependent increase in the area under the plasma concentration-time
curve (AUC) for a number of BA sulfates (LCA-S, GLCA-S, TLCA-S, GCDCA-S, TCDCA-S,
DCA-S, GDCA-S, TDCA-S). Such a result is consistent with low substrate concentration-to-
OATP K ratios in vivo.
m
Liquid Chromatography-Tandem Mass Spectrometry (LC-MS/MS) Analysis. Analysis of
BAs and their 3-O-sulfate conjugates: Plasma and urine concentrations of BAs and their 3-O-
sulfate conjugates were measured in untreated matrices by LC-MS/MS, as described previously
with some modifications (Bathena et al., 2013). Briefly, a Waters ACQUITY ultra performance
®
liquid chromatography (UPLC) system (Waters, Milford, MA) coupled to an 5500 Q TRAP
quadrupole linear ion trap hybrid mass spectrometer (MS) with an electrospray ionization (ESI)
source (Applied Biosystems, MDS Sciex, Foster City, CA) was used. Chromatographic
®
separations were performed with an ACQUITY UPLC BEH C18 column (1.7 µm, 150 × 2.1
mm) maintained at 25°C and equipped with an in-line pre-column filter. The mobile phase
consisted of 7.5 mM ammonium bicarbonate, adjusted to pH 9.0 using ammonium hydroxide
(mobile phase A), and 30% acetonitrile in methanol (mobile phase B), at a total flow rate of 0.2
ml/min. The gradient profile was held at 52.5% mobile phase B for 12.75 min, increased linearly
to 68% in 0.25 min, held at 68% for 8.75 min, increased linearly to 90% in 0.25 min, held at 90%
1
1

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
for 1 min, and finally brought back to 52.5% in 0.25 min followed by 4.75 min re-equilibration
(total run time of 28 min per sample). Ten µl of sample was injected for analysis. Quantitative
data were acquired in multiple reaction monitoring (MRM)-negative ESI mode. MRM
transitions and MS parameters for the different BAs and their respective 3-O-sulfate conjugates
are shown in Supplemental Table 1.
For preparation of calibration curves blank plasma and urine were obtained by charcoal
stripping as described previously (Bathena et al., 2013). Fourteen-point calibration curves were
prepared in stripped matrices by spiking ten µl of appropriate standard solution at final
concentrations ranging from 0.5 to 2500 ng/ml. For extraction of plasma samples, 1 ml of ice-
2
2
cold alkaline ACN (5% NH OH) containing H -GCDCA and H -CDCA as internal standards
4
4
4
was added to 100 µl of samples. Samples were then vortex mixed and centrifuged at 16,000 × g
for 10 min and the supernatants were aspirated, evaporated, and reconstituted in 100 µl of 50%
MeOH solution. Urine samples were extracted similar to plasma samples except that Tween 20
was added (final concentration of 0.2% v/v) to reduce non-specific binding.
2
2
Analysis of H -pitavastatin and RIF: The plasma concentrations of RIF and H -pitavastatin
4
4
were measured in plasma samples treated with 0.1 M sodium acetate buffer (pH 4) using the LC-
MS/MS system listed above. All standards and QCs were made in blank monkey plasma mixed
with an equal volume of 0.1 M sodium acetate buffer (pH 4). Standard and QC mixtures of the
2
analytes were made to encompass a range of concentrations (0.1 to 500 ng/ml, H -pitavastatin;
4
1
to 5,000 ng/ml, RIF). Samples were diluted to be measured in the linear range of the
instrument responses; with the high specificity of MRM (no interference in the blank matrixes)
and a wide dynamic range for each analyte, the dilution integrity was confirmed by independent
analysis of the drugs in the samples in separate assays. Aliquots of 50 µl of standards, QCs and
1
2

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
plasma samples were prepared by protein precipitation using 200 µl of acetonitrile containing an
2
internal standard mixture of simvastatin and H -RIF (100 ng/ml). The plates were vortexed for
8
2
minutes, centrifuged at 3000 rpm for 10 minutes, and 100 µl supernatants of the mixture were
transferred for LC-MS/MS analysis. Chromatographic separation was accomplished on a Waters
Acquity UPLC HSS T3 C18 column (1.8µm, 2.1 × 50 mm) maintained at 40 °C. The mobile
phase consisted of two solvents, solvent A (0.1% formic acid in water) and solvent B (0.1%
formic acid in acetonitrile). The total run time for each injection was 3 minutes. The flow rate
was 0.6 ml/min. The gradient was maintained at 5% B for 0.3 min, followed by a linear increase
to 95% B in 1.8 minutes, and kept at 95% B for 0.3 min then a linear decrease to 5% in 0.3
minutes. The column was equilibrated at 5% B for 0.3 min. A Valco VICI valve (Valco
Instruments Co., Houston, TX) was used to divert the first 0.3 min and the last 0.5 minutes of
HPLC effluent to waste. The injection volume was 2 µl. The analytes were monitored using
MRM with settings listed in Supplemental Table 2.
Determination of Uptake Clearance in the Presence of Cynomolgus Monkey Plated
Primary Hepatocytes. Thawing and seeding procedure for cynomolgus monkey hepatocytes
was the same as that described previously for human hepatocytes (Bi et al., 2006). In brief,
cryopreserved cynomolgus monkey primary hepatocytes (male animal; Lot number 10106012; In
vitro ADMET Laboratories, LLC. Columbia, Maryland) were thawed and seeded into 24-well
collagen coated plates using In Vitro-HT and In Vitro-CP hepatocyte media at a density of 0.35 x
6
1
0 cells/well (0.5 ml/well). After culturing for 6 hours, the uptake study was conducted. To
assess the rate of uptake and passive diffusion, the cells were pre-incubated with and without
0
(
DMSO only) RIFsv (1 mM) or RIF (5 µM) at 37 C for 10 minutes. The uptake was initiated by
1
3

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
2
2
2
the addition of 0.5 ml containing H -TCDCA-S (0.1 µM), H -TDCA-S (0.1 µM), H -DCA-S
4
4
4
2
2
2
(
0.1 µM), H -GCDCA-S (0.1 and 0.5 µM), H -GDCA-S (0.1 and 0.5 µM), H -TCA (0.1 and
5 4 4
0
.5 µM), and non-labeled pitavastatin (0.1 µM). To determine the effect of sodium on substrate
uptake, the cells were pre-incubated in Krebs-Henseleit buffer (KHB) without sodium (NaCl and
NaHCO replaced with choline chloride and choline bicarbonate, respectively) at 37 °C for 10
3
minutes (Ho et al., 2004). In all cases, incubations were terminated at 0.5, 1, 2 and 5 minutes by
washing the cells three times with ice-cold HBSS buffer. The cells were then lysed with 100%
methanol containing internal standard (diclofenac), centrifuged, and dried down under nitrogen
and reconstituted in 50/50 methanol/water. Chromatography was performed on a Waters Acquity
UPLC System (Milford, MA). The autosampler and column were kept at 10°C and 50°C,
respectively. Separation was achieved with a Waters BEH C18 column (2.1x50mm, 1.7 µm),
and a gradient of 7.5 mM ammonium bicarbonate (Mobile Phase A) and 70/30
methanol/acetonitrile (Mobile Phase B) at a flow rate of 0.2 ml/min. An initial mobile phase
composition of 50% B was held for 2 minutes, then ramped to 95% in 1.5 minutes, held at 95%
for 1 minute, and then returned to initial 50% B for 0.5 minute re-equilibration. The total
analysis time for each sample was 5 minutes. Data were collected on an AB Sciex API5500
(QTRAP) mass spectrometer (Foster City, CA, USA) using negative Turbo IonSpray™
electrospray ionization (ESI) and multiple reaction monitoring (MRM) mode with the following
2
2
2
transitions: 290.6/96.8 ( H -TDCA-S), 475.2/96.8 ( H -DCA-S), 533.3/453.4 ( H -GCDCA-S),
4
4
5
2
2
5
32.3/452.3 ( H -GDCA-S), and 514.2/79.8 ( H -TCA). Data acquisition and processing was
4 4
carried out with Analyst software version 1.6.2. (Applied Biosystems/MDS Sciex, Canada)
Analysis of pitavastatin was performed as described above. Stability of the various sulfated BA
◦
substrates was confirmed after their addition to assay buffer and incubation for 5 mins at 37 C.
1
4

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
BA Sulfation Catalyzed by PAPS-Fortified Cynomolgus Liver Cytosol. Refer to
Supplemental Information for details related to incubation conditions and LC-MS/MS analysis.
Pharmacokinetics Analysis. BAs: For each BA, the plasma AUC from 0 to 24 hour (AUC0-
24,plasma)
was derived from the concentration-time profile for each individual animal (trapezoidal
rule, Microsoft Office Excel). Renal clearance (CLrenal) was calculated by dividing the amount
excreted in urine from 0 to 24 hour (Ae0-24,urine) by the AUC0-24,plasma. AUC0-24,plasma ratios
(
AUCRplasma) were determined by dividing the AUC0-24,plasma after RIF treatment by the vehicle
alone AUC0-24,plasma. The CLrenal ratio was determined by dividing the CLrenal after RIF treatment
by the vehicle alone CLrenal
.
2
RIF and pitavastatin: The non-compartmental analyses of H -pitavastatin and RIF plasma
4
TM
concentration-time data were performed using Watson LIMS version 7.4.1 (Thermo Fisher
Scientific Inc, Waltham, MA ), which supported the generation of the various pharmacokinetic
parameters; AUC, T1/2 (half-life), CL (clearance), Vdss (volume of distribution), Cmax (peak
plasma concentration), and Tmax (time to peak plasma concentration). As for the different BAs,
2
H -pitavastatin AUCR
values were determined by dividing the AUC0-24,plasma after RIF
4
plasma
treatment by the vehicle alone AUC0-24,plasma
.
For RIF, it was possible to calculate RIF plasma free Cmax based on a plasma unbound fraction of
.265 (refer to Supplemental Information for RIF cynomolgus monkey plasma protein binding).
In turn, in vitro IC s for RIF with cynomolgus monkey OATP and NTCP (Shen et al., 2013;
0
5
0
Chu et al., 2015) were used to calculate % inhibition; % inhibition = 100 * {[I]/([I] + IC )}. It is
5
0
assumed that IC ~ K (when substrate concentration < K ). [I] represents plasma C of RIF
5
0
i
m
max
1
5

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
(
total or free). Consideration of free and total plasma Cmax is consistent with Vaidyanathan et al
2016); in the absence of i.v. RIF pharmacokinetic data, it was not possible to derive an
(
absorption rate constant for RIF and estimate its liver inlet (portal) concentration in the
cynomolgus monkey.
Statistical Analysis of RIF Dose Response. Plasma: Plasma profiles over time for each animal
were collapsed into an AUC using the trapezoidal rule and a Cmax score. The analyses were then
conducted using a linear mixed model, such that the within-animal correlations were accounted
for in the model. The R computing language was used for these calculations (R Core Team,
2
014). Outcomes were analyzed on the log (base 2) scale to make the residuals more normal.
Concentrations were entered into the model on the same scale and the p-value for the linear trend
is based on an F-statistic using Satterthwaite's approximation (Burghoff, 2016). False discovery
rate (FDR) estimates were computed using the Benjamini and Hochberg procedure (Benjamini
and Hochberg, 1995).
Urine: Renal clearance (CLrenal) values (ml/min/kg) of zero were treated as missing and analytes
with greater than five missing values were excluded. As such, 18 of the 30 BAs were analyzed.
As described for plasma, the analysis was conducted using a linear mixed model such that the
within-animal correlations were accounted for in the model. The BAs were analyzed on the
natural log scale to make the residuals more normal. Similar to plasma, the p-value for the linear
trend is based on an F-statistic using Satterthwaite's approximation and FDR estimates were
computed using the Benjamini and Hochberg procedure.
1
6

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
RESULTS
Pharmacokinetics of RIF in Cynomolgus Monkeys. As expected, a dose-dependent increase
in RIF AUC0-24,plasma, and plasma Cmax was observed over the dose range of 1 to 30 mg/kg (Table
1
and Supplemental Fig. 1). However, at the RIF doses of 3, 10 and 30 mg/kg, there was
evidence for a greater than proportional increase in AUC0-24,plasma (11-, 71- and 243-fold versus
AUC at 1 mg/kg) and plasma Cmax (12-, 45-, and 137-fold versus Cmax at 1 mg/kg). For RIF,
plasma unbound fraction was 0.265 and so the calculated free Cmax was 0.057, 0.659, 2.57 and
7
.79 µM at 1, 3, 10 and 30 mg/kg, respectively. Such concentrations exceed the reported in vitro
IC values (0.14 to 1.7 µM) for RIF with cynomolgus monkey OATP1B1 and OATP1B3 (Shen
5
0
et al., 2013; Chu et al., 2015).
2
Impact of RIF on H -Pitavastatin Pharmacokinetics. The pharmacokinetic parameters of
4
2
H -pitavastatin, an accepted cynomolgus monkey OATP probe drug (Takahashi et al., 2013),
4
were determined after dosing of an i.v. bolus (0.2 mg/kg) to vehicle control and RIF-dosed
2
cynomolgus monkeys (Table 2, Supplemental Fig. 1). Compared with the vehicle control, H -
4
pitavastatin CL was decreased 21%, 58%, 73%, and 76% with RIF treatment at 1, 3, 10, and 30
mg/kg dose levels, respectively. However, the change was only statistically significant at the 3
2
top RIF doses. Furthermore, H -pitavastatin AUC
was increased 1.2-, 2.4-, 3.8-, and
0-24,plasma
4
2
4
.5-fold at 1, 3, 10 and 30 mg/kg RIF, respectively. Both the Vdss and T1/2 of H -pitavastatin
4
were decreased (~70 %) at the three highest RIF doses. At all the RIF dose levels, recovery of
2
unchanged H -pitavastatin in urine was less than 5% of the dose. Overall, the impact of the
4
2
lowest RIF dose on H -pitavastatin pharmacokinetics was not statistically significant.
4
1
7

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
2
Profiling of Plasma BAs at Different Doses of RIF. As with H -pitavastatin, the plasma
4
levels of various BAs, as well as their corresponding 3-O-sulfate conjugates, were measured in
cynomolgus monkeys following increasing doses of RIF. In total, 30 different BAs were
monitored. Sulfated BAs, in particular GDCA-S, TDCA-S, GCDCA-S, TCDCA-S, GLCA-S,
and TLCA-S, presented a marked dose-dependent increase in their plasma concentration-time
profile (Fig.1, Supplemental Table 3).
In contrast, the plasma concentration-time profile of
non-sulfated BAs, particularly DCA, GDCA, TDCA, GCDCA, TCDCA, UDCA, GUDCA,
TUDCA, CA, GCA, and TCA, showed a relatively weak increase at the RIF doses of 10 and 30
mg/kg (Supplemental Fig. 2). Chu et al. have also reported a weak increase in non-sulfated BAs
after an oral RIF dose of 18 mg/kg (Chu et al., 2015). RIF treatment did not cause significant
changes in LCA, GLCA and TLCA. Unfortunately, GUDCA-S, TUDCA-S, CA-S, TCA-S, and
GCA-S plasma levels were low and remained undetectable even after RIF treatment. Overall,
the presence of various sulfo-conjugates in plasma was indicative of BA sulfation in the
cynomolgus monkey. However, the pool of cynomolgus monkey BAs in circulation was distinct
from that reported for human subjects; the percentages of sulfated (1.3% versus 28.4%) and
amidated (21.5% versus 75.4%) BAs were low for monkey versus human, respectively
(Supplemental Table 4).
To complement the in vivo studies, the sulfation of six representative BAs (GCDCA,
GUDCA, LCA, GLCA, GCA, and TLCA) was investigated after incubation with cynomolgus
monkey liver cytosol (Supplemental Fig. 3); the availability of authentic standards supported the
identification of the 3-O-sulfate as the major product of PAPS-fortified monkey liver cytosol.
2
Although a good correlation was obtained between human and cynomolgus monkey (R =
0
.905), a low activity ratio (cynomolgus monkey-to-human) was obtained for the formation of
1
8

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
LCA-S, GLCA-S, TLCA-S, and GUDCA-S (0.20, 0.25, 0.36, and 0.37, respectively). In
comparison, GCDCA-S (activity ratio = 0.89) and GCA-S (activity ratio = 3.0) rendered higher
activity ratios comparable to those of DHEA 3-O-sulfate (activity ratio of 1.3). Importantly, RIF
(up to 100 µM) was shown not to inhibit cynomolgus monkey liver cytosol-catalyzed BA
sulfation (Supplemental Fig. 4).
Assessment of RIF Dose Response. AUCRplasma values for the various BAs after RIF treatment
are shown in Figure 2. The highest AUCRplasma (≥ 78) was observed for GDCA-S and GCDCA-
S, followed by TDCA-S, DCA-S and TCDCA-S (~50) and GLCA-S and TLCA-S (~30). A
relatively modest AUCRplasma (5 to 10) was observed for CA, GCA and TCA. The remaining
non-sulfated BAs presented low AUCRplasma values (2 to 5). Fold changes in plasma Cmax after
RIF treatment are shown in Figure 3. Similar to the AUC changes, the highest (50- to 80-) fold
increase was observed for DCA-S, GDCA-S, TCDCA-S, TDCA-S, and GCDCA-S followed by
a 20-fold increase for GLCA-S and TLCA-S at RIF dose of 30 mg/kg. AUC0
,
and plasma
-24 plasma
Cmax values at each RIF dose and vehicle control are shown in Supplemental Table 3.
Statistical analysis for linear trend in AUC0-24 plasma, and Cmax and FDR for each BA are
,
shown in Table 3. Overall, with the exception of CA-S, all BA 3-O-sulfates were characterized
by a significant linear trend (p-value < 0.01) with a slope of more than 0.6700 and a FDR less
than 10% for AUC0
,
. GDCA-S, GCDCA-S, TDCA-S, TCDCA-S, GLCA-S, and TLCA-S
-24 plasma
were the most significant sulfate conjugates with a p-value for the linear trend of less than 0.001,
a slope of more than 0.7500 and FDR less than 1% for AUC0 as well as Cmax. In contrast,
,
-24 plasma
most of the non-sulfated bile acids did not show a statistically significant linear trend and FDR
1
9

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
was more than 10% for AUC0-24,plasma and Cmax. Only TCA and TDCA showed a p-value less
than 0.01 and ~ 10% FDR.
2
Importantly, the plasma AUCRplasma values showed a good linear correlation (R >
0
.7000) between pitavastatin and GDCA-S, TDCA-S, GCDCA-S, TCDCA-S, GLCA-S, and
2
TLCA-S (Fig. 4). Similarly, a good linear correlation (R > 0.7500) between RIF plasma free
2
Cmax and H -pitavastatin, GDCA-S, TDCA-S, GCDCA-S, TCDCA-S, GLCA-S, and
4
TLCA-S AUCRplasma was observed (Fig. 5).
As described previously, an effort was made to administer RIF over a dose range that
generated a wide range of plasma total (0.2 to 29 μM) and free (0.06 to 7.8 μM) Cmax values
(Table 1). In so doing, it was possible to investigate the dose-dependent inhibition of OATPs
and NTCP. Based on in vitro IC data, dose-dependent inhibition of OATP1B1 (16% to ≥ 96%)
5
0
and OATP1B3 (<10% to ≥ 85%) was expected, with less inhibition of OATP2B1 (<10% to
1%) and NTCP (<10% to 29%) anticipated (Supplemental Table 5 and 6). Despite the effort to
3
ensure dose-dependent inhibition, however, the AUCRplasma values for the various BA sulfates
differed markedly. The highest maximal AUCRplasma values (≥ 78) were obtained with GDCA-S
and GCDCA-S, followed by TDCA-S, DCA-S and TCDCA-S (AUCRplasma ~ 50) and GLCA-S
and TLCA-S (AUCRplasma ~ 30).
Impact of RIF on BA Renal Clearance. Because the 3-O-sulfate conjugates of the various
BAs are recovered in human urine (Bathena et al., 2013; Bathena et al., 2015; Tsuruya et al.,
2
016), the present study was extended to include the profiling of urine of control and RIF-dosed
cynomolgus monkeys. In this regard, a dose-dependent increase in the amounts of GDCA-S,
TDCA-S, GCDCA-S, TCDCA-S, GLCA-S, and TLCA-S excreted in urine was observed
2
0

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
(Supplemental Table 3). CLrenal ratios and statistical analysis for each BA are shown in Figure 6
and Table 3, respectively. A weak dose-dependent increase in CLrenal was observed for UDCA,
GDCA, and DCA. However, there was no statistically significant effect of RIF treatment on the
CLrenal of sulfated and non-sulfated bile acids. For the former, this is in marked contrast to the
changes in plasma AUC and Cmax following RIF treatment.
Incubation of GCDCA-S, GDCA-S, DCA-S, TDCA-S, TCA and Pitavastatin with
Cynomolgus Monkey Plated Hepatocytes. Based on the availability of deuterium-labeled
material and RIF-dependent AUCRplasma values in vivo, four sulfated BAs (GCDCA-S, GDCA-
S, DCA-S, and TDCA-S) were chosen for study as solute carrier (SLC) substrates in vitro after
incubation with plated cynomolgus monkey primary hepatocytes (Figures 7 and 8). TCA and
pitavastatin were also incubated as representative cynomolgus monkey NTCP (> OATP) and
OATP (>NTCP) substrates, respectively (Chu et al., 2015; Takahashi et al., 2013). To discern
the role of NTCP versus OATPs, incubations were performed in the presence and absence of
sodium (NTCP is sodium-dependent). In addition, RIF (5 μM) and RIFsv (1 mM) were
deployed as cynomolgus monkey OATP-selective and pan-SLC inhibitors, respectively
(
Supplemental Table 5; Shen et al., 2013; Chu et al., 2015; Hong Shen, personal
communication).
When incubated with RIFsv, the uptake of TCA, GDCA-S, GCDCA-S, and pitavastatin
was markedly inhibited (≥ 92%). For the four substrates, such a result is consistent with
relatively high rates of active (versus passive; ≤ 8%) uptake (Figure 7A). As shown in Figure 7B,
the uptake of GCDCA-S and GDCA-S was minimally impacted by the removal of sodium. By
contrast, uptake of TCA (74%) and pitavastatin (~44%) was decreased in the presence of
2
1

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
sodium-free buffer. As expected, RIF elicited relatively weak inhibition of TCA uptake when
compared to pitavastatin (14% versus 58%). Uptake of both GCDCA-S (69% inhibition) and
GDCA-S (82% inhibition) was sensitive to RIF. Although the exact contribution of OATP1B1
and OATP1B3 was not determine, in the absence of selective inhibitors and established relative
activity factors for cynomolgus monkey transporters, it is concluded that uptake of both
GCDCA-S and GDCA-S (0.5 µM) in the presence of cynomolgus monkey hepatocytes is
dominated by OATPs and that their profile is distinct from that of TCA. Based on the results
presented in Figure 8, the same can be said for two additional BA sulfates (TDCA-S and DCA-
S). In this instance, RIF (5 µM) was shown to inhibit the uptake of TCA, pitavastatin, GCDCA-
S, GDCA-S, TDCA-S and DCA-S (0.1 µM) by 21%, 73%, 92%, 83%, 95% and 80%,
respectively.
2
2

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
DISCUSSION
Metabolism of various BAs is complex and involves oxidation, amidation,
glucuronidation, and sulfation. Once conjugated, BAs are also subjected to transporter-mediated
uptake (e.g., OATP and NTCP) and efflux (Akita et al., 2001; Sasaki et al., 2002; Zelcer et al.,
2
003a, 2003b; Tsuruya et al., 2016; Rodrigues et al., 2014). Therefore, the BA pool of most
species is complex, and subject to enterohepatic recirculation and renal clearance. This means
that the BA profiling requires robust LC-MS/MS methods with access to a large number of
authentic standards (Bathena et al., 2013). To support the present study, therefore, authentic
standards of a number of non-commercially available BA 3-O-sulfates were prepared and 30
different BAs were profiled in cynomolgus monkey plasma and urine.
Although in vivo sulfation of DHEA has been reported in cynomolgus monkeys (Leblanc
et al., 2003), and SULT2A1 (human sulfotransferase involved in BA 3-O-sulfation) is known to
be expressed in cynomolgus monkey liver (Nishimura et al., 2008; Alnouti, 2009; Nishimura et
al., 2009), there have been no reports describing the BA sulfation in the cynomolgus monkey.
For the first time, it was possible to report that DHEA and various BAs undergo sulfation in vitro
(Supplemental Fig. 3). Importantly, the presence of BA sulfates in cynomolgus monkey urine is
consistent with human data (Bathena et al., 2013; Tsuruya et al., 2016). However, the fraction of
the BA pool in circulation as the sulfated species was low in cynomolgus monkey versus human
(28.4% versus 1.3%; Supplemental Table 4), and likely reflects species differences in CLrenal and
formation clearance. In agreement, the CLrenal of GCDCA-S is lower in humans (31 versus 0.05
ml/min per kg) (Supplemental Table 3; Tsuruya et al., 2016) and the rate of BA sulfation in vitro
was lower in the presence of cynomolgus monkey cytosol (Supplemental Fig. 3). A species
difference in OATP-mediated hepatic uptake clearance is also a possibility.
2
3

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
GCDCA-S has been shown to undergo active renal secretion and has been proposed as a
biomarker for human organic anion transporter 3 (OAT3) (Tsuruya et al., 2016). Although the
cynomolgus monkey OAT3 has been expressed and characterized, nothing is known of its ability
to transport BA sulfates and its inhibition by RIF (Tahara et al., 2005). Therefore, as part of the
present study, it was important to assess the impact of RIF on BA sulfation and renal clearance.
This ensured that any observed changes in BA AUC0-24,plasma were reflective of hepatic
transporter inhibition. In vitro data indicated that RIF (up to 100 μM) did not inhibit liver
cytosol-catalyzed BA sulfation (Supplemental Fig. 4). Likewise, profiling of cynomolgus
monkey urine supported calculation of CLrenal for the different BAs and it was determined that
the impact of RIF was minimal (Table 3; Supplemental Table 3). Importantly, there is evidence
indicating that BA sulfates also serve as substrates of human canalicular multidrug resistance-
associated protein (MRP) MRP2 and basolateral MRP3 and MRP4 (Akita et al., 2001; Sasaki et
al., 2002; Zelcer et al., 2003a, 2003b). The possibility that single dose RIF can inhibit
cynomolgus monkey MRP2 has already been considered by Chu et al (2015). In this instance,
the authors showed that RIF is a relatively weak inhibitor of cynomolgus monkey MRP2 (IC =
5
0
1
18 µM) and argued for relatively minimal inhibition based on estimated free RIF liver levels
(~10 µM). Although no RIF inhibition data are available for monkey MRP3 and MRP4, given
that robust AUCRplasma values for the various BA sulfates it is assumed that neither transporter is
inhibited. Alternatively, RIF could induce MRP3 and MRP4 and compromise the interpretation
of the data (Marshall et al., 2005; Badolo et al., 2015). However, it can be argued that significant
and sustained induction of both MRP proteins is unlikely following a single RIF dose.
Importantly, BA sulfate plasma concentration-time profiles were consistent with relatively rapid
2
4

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
(within 6 hr) dose-dependent inhibition of hepatic uptake. Moreover, plasma concentrations for
five of the six BA sulfates were trending towards pre-RIF dose levels by 24 hr (Fig. 1).
To date, efforts to inhibit cynomolgus monkey OATP in vivo have involved oral
administration of a single RIF dose that generates a plasma total Cmax of ~10 μM (Chu et al 2015;
Shen et al., 2013; and Takahashi et al., 2013). As described, it was possible to administer RIF at
four dose levels of 1, 3, 10, and 30 mg/kg and obtain mean plasma total Cmax values of 0.2, 2.5,
9
.7, and 29 µM, respectively; corresponding to a plasma free Cmax of 0.06, 0.67, 2.57 and 7.79
μM, respectively (Table 1). Evidently, the pharmacokinetic profile of RIF in the cynomolgus
monkey was non-linear, likely reflecting saturation of first pass, and consistent with human data
2
(
Acocella, 1978). Despite the non-linearity, there was a good linear correlation (R > 0.9332)
between RIF plasma Cmax and AUCRplasma for a number of BA sulfates (Fig. 5) and, because
expression of hepatic OATP2B1 is relatively low in cynomolgus monkey, such a dose response
is more likely reflective of OATP1B1 and OATP1B3 inhibition (Wang et al., 2015). Based on
the in vitro IC s reported for both OATPs (≤ 1.7 µM; Supplemental Table 5), and assuming that
5
0
RIF plasma concentrations support estimates of inhibition (Vaidyanathan et al., 2016), up to
>85% (RIF free Cmax) and >96% (RIF total Cmax) inhibition is anticipated (Supplemental Table
2
6
3
). In agreement, a clear RIF dose-dependent increase in H -pitavastatin AUCR
(1.2, 2.4,
plasma
4
.8 and 4.5) was evident (Fig. 5). Because of weaker inhibition by RIF (IC50 ≥ 35.1 µM,
Supplemental Table 5), less NTCP inhibition (10% to 29%) is expected (Supplemental Table 6).
This is an important consideration, because certain BAs are known to favor NTCP over OATPs
(Meier et al., 1997; Maeda et al., 2006; Dong et al., 2015: Suga et al, 2017). For example, the
uptake of TCA by primary cynomolgus monkey hepatocytes was shown to be highly sodium
dependent and relatively refractory to RIF (Figures 7 and 8). Although no formal in vitro-in vivo
2
5

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
exercise was attempted, it assumed that the RIF dose-dependent AUCRplasma values for TCA
(1.2, 2.0, 1.8 and 5.4) are largely reflective of NTCP inhibition (Supplemental Tables 3 and 6).
On the other hand, GCDCA-S, GDCA-S, TDCA-S and DCA-S behaved as OATP substrates in
the presence of cynomolgus monkey hepatocytes (Figures 7 and 8), consistent with reports of a
fifth sulphated BA (TLCA-S) presenting as an OATP substrate in vitro (Meng et al., 2002;
Sasaki et al., 2002).
As described, five sulfated BAs in plasma responded robustly to RIF in a dose-dependent
manner (maximal AUCRplasma ≥ 50). Two additional BA sulfates, GLCA-S and TLCA-S, were
also found to respond to RIF (maximal AUCRplasma ~ 30) (Fig. 2). From the standpoint of
detectability in control animals, magnitude of the RIF dose response, and detectability in human
plasma, these seven BA sulfates (GCDCA-S, TCDCA-S, DCA-S, GDCA-S, TDCA-S, GLCA-S
and TLCA-S) all present as potential sensitive OATP biomarkers. Importantly, the RIF dose-
response obtained was far greater than the increases (<10-fold) reported for cynomolgus monkey
plasma bilirubin, bilirubin glucuronide, coproporphyrin (I and III), non-sulfated BAs, and DHEA
sulfate (Chu et al., 2015; Watanabe et al., 2015; Shen et al., 2016).
Although the results presented herein showcase various BA sulfates as sensitive
cynomolgus monkey OATP biomarkers, it cannot be assumed that the results translate directly to
human subjects. For example, the balance of OATP- versus NTCP-mediated liver uptake, and
the contributions of individual OATPs, may not be the same across species. Despite the caveats,
the results of the present work are consistent with plasma metabolomic data from a recent
OATP1B1 (SLCO1B1) genome-wide association study that identified GCDCA-S, GDCA-S and
TLCA-S as potential OATP1B1 biomarkers (Yee et al., 2016). Such BA sulfates could
potentially serve as sensitive human OATP biomarkers and provide the necessary dynamic range
2
6

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
to support perpetrator differentiation (weak, moderate, versus potent inhibition), enable the study
of OATP genotype-phenotype associations, and facilitate phenotyping of hepatobiliary diseased
subjects (Gong and Kim, 2013; Clarke et al., 2014). In this regard, sulfated BAs could be
superior human OATP biomarkers when compared to plasma coproporphyrin I and III
(AUCRplasma ~4.0) (Lai et al., 2016). As discussed previously, because circulating sulfated BAs
are substrates of renal transporters also, they may serve as dual liver OATP and renal OAT3
biomarkers. In agreement, an increase in “urinary sulfated bile acids” has been reported in
patients with various hepatobiliary diseases, which may in part reflect altered liver OATP
function (Kobayashi et al., 2002; Nanashima et al., 2009; Bathena et al., 2015). Consistent with
human data, the amount of sulfated BAs in cynomolgus monkey urine increased with RIF dose
(Supplemental Table 3).
Based on the results of the present study, it is concluded that the cynomolgus monkey can
form BA sulfates that comprise only ~1% of the plasma BA pool and present as sensitive hepatic
OATP biomarkers. Given the number of similarly sulfated BAs that are detectable in human
serum (Bathena et al., 2013), it is envisioned that they will be increasingly studied as human
OATP substrates and compared to other biomarkers such as coproporphyrin I and III.
2
7

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
ACKNOWLEDGMENTS
The authors would like to thank Dr. Hong Shen (Bristol-Myers Squibb Co., Princeton, USA) for
providing rifamycin SV in vitro inhibition data for cynomolgus monkey OATPs and NTCP
expressed in HEK293 cells (Supplemental Table 5). The authors also thank Dana Gates and
2
Sweta Modi (Pfizer Inc.) for their help preparing the H -pitavastatin and rifampicin dose
4
formulations, and John Deschenes (Pfizer Inc.) for his help conducting the cynomolgus monkey
pharmacokinetics study. Sangwoo Ryu and Keith Riccardi (Pfizer Inc.) are acknowledged for
conducting equilibrium dialysis of RIF with cynomolgus monkey plasma (Supplemental
Information). Brian Rago (Pfizer Inc.) is acknowledged for conducting bioanalysis of
cynomolgus monkey hepatocyte incubates.
2
8

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
Authorship contributions
Participated in research design: A.D.R., H.G., M.V.S.V., L.T., R.T., R.E.K
Conducted experiments: R.T., H.G., R.E.K., Y.B., M.A.C., R.S., G.S.W, M.N.
Contributed new reagents and analytical procedures: R.T., Y.A., B.H., Y.L., A.Y., G.S.W.
Performed data analysis: R.T., M.V.S.V., Y.B., M.K., A.D.R., M.N.
Wrote or contributed to the writing of the manuscript: R.T., H.G., R.E.K., Y.B., M.A.C., R.S.,
M.K., Y.A., M.V.S.V, A.D.R.
2
9

DMD Fast Forward. Published on April 10, 2017 as DOI: 10.1124/dmd.117.075275
This article has not been copyedited and formatted. The final version may differ from this version.
REFERENCES
Acocella G (1978) Clinical pharmacokinetics of rifampicin. Clin Pharmacokinet 3:108-127.
Akita H, Suzuki H, Ito K, Kinoshita S, Sato N, Takikawa H, and Sugiyama Y (2001)
Characterization of bile acid transport mediated by multidrug resistance associated
protein 2 and bile salt export pump. Biochim Biophys Acta 1511:7-16.
Alnouti Y (2009) Bile acid sulfation: a pathway of bile acid elimination and detoxification.
Toxicol Sci 108:225-246.
Badolo L, Jensen B, Säll C, Norinder U, Kallunki P, and Montanari D (2015) Evaluation of 309
molecules as inducers of CYP3A4, CYP2B6, CYP1A2, OATP1B1, OCT1, MDR1,
MRP2, MRP3 and BCRP in cryopreserved human hepatocytes in sandwich culture.
Xenobiotica 45:177-1787.
Bathena SP, Mukherjee S, Olivera M, and Alnouti Y (2013) The profile of bile acids and their
sulfate metabolites in human urine and serum. J Chromatogr B Analyt Technol Biomed
Life Sci 942-943:53-62.
Bathena SP, Thakare R, Gautam N, Mukherjee S, Olivera M, Meza J, and Alnouti Y (2015)
Urinary bile acids as biomarkers for liver diseases II. Signature profiles in patients.
Toxicol Sci 143:308-318.
Bednarczyk D and Boiselle C (2016) Organic anion transporting polypeptide (OATP)-mediated
transport of coproporphyrins I and III. Xenobiotica 46:457-466.
Benjamini Y and Hochberg Y (1995) Controlling the false discovery rate: a practical and
powerful approach to multiple testing. Journal of the Royal Statistical Society Series B
(Methodological) 57:289-300.
3
0