Pharmacokinetics of the cardioprotective drug dexrazoxane and its active metabolite adr-925 with focus on cardiomyocytes and the hearts

Article abstract of DOI:10.1124/jpet.117.244848

Dexrazoxane (DEX), the only cardioprotectant approved against anthracycline cardiotoxicity, has been traditionally deemed to be a prodrug of the iron-chelating metabolite ADR-925. However, pharmacokinetic profile of both agents, particularly with respect to the cells and tissues essential for its action (cardiomyocytes/myocardium), remains poorly understood. The aim of this study is to characterize the conversion and disposition of DEX to ADR-925 in vitro (primary cardiomyocytes) and in vivo (rabbits) under conditions where DEX is clearly cardioprotective against anthracycline cardiotoxicity. Our results show that DEX is hydrolyzed to ADR-925 in cell media independently of the presence of cardiomyocytes or their lysate. Furthermore, ADR-925 directly penetrates into the cells with contribution of active transport, and detectable concentrations occur earlier than after DEX incubation. In rabbits, ADR-925 was detected rapidly in plasma after DEX administration to form sustained concentrations thereafter. ADR-925 was not markedly retained in the myocardium, and its relative exposure was 5.7-fold lower than for DEX. Unlike liver tissue, myocardium homogenates did not accelerate the conversion of DEX to ADR-925 in vitro, suggesting that myocardial concentrations in vivomay originate fromits distribution from the central compartment. The pharmacokinetic parameters for both DEX and ADR-925 were determined by both noncompartmental analyses and population pharmacokinetics (including joint parent-metabolite model). Importantly, all determined parameters were closer to human than to rodent data. The present results open venues for the direct assessment of the cardioprotective effects of ADR-925 in vitro and in vivo to establish whether DEX is a drug or prodrug.

Full text of DOI:10.1124/jpet.117.244848

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Pharmacokinetics of the cardioprotective drug dexrazoxane and its
active metabolite ADR-925 with focus on cardiomyocytes and the heart
Eduard Jirkovský, Anna Jirkovská, Jan Bureš, Jaroslav Chládek, Olga Lenčová, Ján Stariat,
Zuzana Pokorná, Galina Karabanovich, Jaroslav Roh, Petra Brázdová, Tomáš Šimůnek, Petra
Kovaříková and Martin Štěrba
Department of Pharmacology, Faculty of Medicine in Hradec Králové, Charles University,
Šimkova 870, Hradec Králové, 500 03, Czech Republic.
(E.J., J.C. O.L-P., Z.P. P.B., M.Š.)
Department of Biochemical Sciences, Faculty of Pharmacy in Hradec Králové, Charles
University, Akademika Heyrovského 1203, 500 05 Hradec Králové, Czech Republic.
(A.J., T.Š.)
Department of Pharmaceutical Chemistry and Pharmaceutical Analysis, Faculty of Pharmacy
in Hradec Králové, Charles University, Akademika Heyrovského 1203, 500 05 Hradec
Králové, Czech Republic.
(J.B., J.S., P.K.)
Department of Inorganic and Organic Chemistry, Faculty of Pharmacy in Hradec Králové,
Charles University, Akademika Heyrovského 1203, 500 05 Hradec Králové, Czech Republic.
(G.K., J.R.)
1

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Running title page
a) Running Title:
DEX metabolism in vitro and in vivo
b) Corresponding authors$:
Dr. Martin Štěrba
Faculty of Medicine in Hradec Králové
Charles University
Šimkova 870
500 03 Hradec Králové
Czech Republic
Tel.: +420 495 816 312; Fax: +420 495 513 597
E-mail: sterbam@lfhk.cuni.cz
Dr. Petra Kovaříková
Faculty of Pharmacy in Hradec Králové
Charles University
Akademika Heyrovského 1203
500 05 Hradec Králové
Tel.: +420 495 067 236
2

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Czech Republic
E-mail: petra.kovarikova@faf.cuni.cz
c) Counts:
Number of text pages: 58
Number of tables: 3
Number of figures: 9
Number of references: 53
Number of words in the abstract: 246
Number of words in the introduction: 720
Number of words in the discussion: 1685
Supplementary material: Supplementary Figure 1, Supplementary Figure 2, Supplemental
Method.
d) Abbreviations:
ANT – anthracycline antibiotics; CV – coefficient of variation; DEX – dexrazoxane; ISV –
inter-subject variability; NCA – non-compartmental analyses; NVCM – neonatal rat ventricular
cardiomyocytes; % RSE – percent of relative standard error; RVPC – residual variability in
plasma concentrations.
e) Section:
Metabolism, Transport, and Pharmacogenomics
3

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Abstract
Dexrazoxane (DEX), the only cardioprotectant approved against anthracycline cardiotoxicity,
has been traditionally deemed to be a prodrug of the iron-chelating metabolite ADR-925.
However, pharmacokinetic profile of both agents, particularly with respect to the cells and
tissues essential for its action (cardiomyocytes/myocardium), remains poorly understood. The
aim of this study is to characterize the conversion and disposition of DEX to ADR-925 in vitro
(primary cardiomyocytes) and in vivo (rabbits) under conditions where DEX is clearly
cardioprotective against anthracycline cardiotoxicity. Our results show that DEX is hydrolyzed
to ADR-925 in cell media independently of the presence of cardiomyocytes or their lysate.
Furthermore, ADR-925 directly penetrates into the cells with contribution of active transport,
and detectable concentrations occur earlier than after DEX incubation. In rabbits, ADR-925
was detected rapidly in plasma after DEX administration to form sustained concentrations
thereafter. ADR-925 was not markedly retained in the myocardium, and its relative exposure
was 5.7-fold lower than for DEX. Unlike liver tissue, myocardium homogenates did not
accelerate the conversion of DEX to ADR-925 in vitro suggesting that myocardial
concentrations in vivo may originate from its distribution from the central compartment. The
pharmacokinetic parameters for both DEX and ADR-925 were determined by both NCA and
population pharmacokinetics (including joint parent-metabolite model). Importantly, all
determined parameters were closer to human than to rodent data. The present results open
venues for the direct assessment of the cardioprotective effects of ADR-925 in vitro and in vivo
to establish whether DEX is a drug or prodrug.
4

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Introduction
Dexrazoxane (DEX, Fig. 1) is the only cardioprotective agent approved for clinical use against
cardiotoxicity induced by anthracycline (ANT) anticancer agents (Sterba et al., 2013;
Vejpongsa and Yeh, 2014a), and besides that it is also approved for treatment of accidental
anthracycline extravasation (Hasinoff, 2006; Langer, 2010). Over the decades of its preclinical
development, it has been documented that the cardioprotective effects of DEX are robust in all
species tested irrespectively of the ANT derivative (Sterba et al., 2013). Furthermore, clinical
efficacy has been firmly established in numerous trials and by their meta-analyses (van Dalen
et al., 2011; Vejpongsa and Yeh, 2014a; Hutchins et al., 2017).
DEX is traditionally believed to be a prodrug that is metabolized to an active metabolite - ADR-
925 via two intermediate hydrolytic products (termed B and C, Fig. 1). The first step is catalyzed
by dihydropyrimidase (initial single ring-opening step), while second step to ADR-925 is
catalyzed by dihydroorotase. ADR-925 is deemed to protect cardiomyocytes from ANT-
induced oxidative damage by chelating intracellular iron (Hasinoff et al., 1998; Cvetkovic and
Scott, 2005). Although this “ROS and iron” hypothesis is very popular, numerous concerns
have been raised over the last decade (Sterba et al., 2013). Some experts have questioned even
the fundamental concept of DEX action as a prodrug (Hasinoff and Herman, 2007), and further
evidence suggests that DEX may protect the heart by catalytic inhibition of topoisomerase IIβ
(Lyu et al., 2007; Vejpongsa and Yeh, 2014a; Lencova-Popelova et al., 2016).
It is noteworthy that despite several important contributions (Hasinoff et al., 1998; Hasinoff
and Aoyama, 1999a; Hasinoff and Aoyama, 1999b; Schroeder et al., 2002; Schroeder et al.,
2003; Schroeder and Hasinoff, 2005; Schroeder et al., 2005), understanding the metabolism of
DEX to its putative active metabolite ADR-925 is still considerably limited. This becomes
particularly apparent when attempting to put the pharmacodynamics of DEX on a solid
5

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
pharmacokinetic ground (Sterba et al., 2013). Clinical data on the metabolism of DEX to ADR-
925 are extremely sparse. There is only a single small clinical trial (Schroeder et al., 2003)
where ADR-925 was systematically determined in plasma. Unfortunately, DEX was part of an
atypical chemotherapeutic combination with high-dose etoposide, the number of patients was
limited and the calculated half-life of DEX was longer compared to values commonly reported
elsewhere (Cvetkovic and Scott, 2005). Most data from animal experiments were obtained by
the Hasinoff group, and all in vivo experiments published in extenso used only rats (Schroeder
and Hasinoff, 2002; Schroeder and Hasinoff, 2005). Moreover, ADR-925 concentrations in the
myocardium, where its cardioprotective action should occur, are not completely described. The
only available data come from the latter study where concentrations of ADR-925 were
determined in a single time point (different for DEX and ADR-925 administration). In addition,
the translatability of these in vivo findings may not be straightforward due to the strikingly
higher elimination rate of DEX in rodents (rats and mice) compared to humans (Hasinoff et al.,
1998).
DEX metabolism has also been studied in vitro with isolated cardiomyocytes and hepatocytes
(Schroeder et al., 2005). However, its concentrations were determined in cell culture media but
not within the intracellular compartment where the cardioprotective effects of ADR-925 should
occur. Indeed, the relatively polar ADR-925 need not penetrate easily out of the cells, which
may make the precise evaluation of DEX metabolism difficult. Furthermore, experiments
performed with DEX in cardiomyocytes were short (2 hours), and concentrations of ADR-925
have not been determined. Hence, it remains unknown whether the putative active metabolite
achieves markedly prolonged intracellular concentrations in cardiomyocytes, which can be
essential for the cardioprotective efficacy of DEX.
Thus, a plasma concentration-time profile of the putative active metabolite (ADR-925) under
standard conditions when DEX is administered as a single agent in a clearly cardioprotective
6

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
schedule remains unknown. In addition, information on concentration-time profiles of both
DEX and ADR-925 in the heart and cardiomyocytes (i.e., at the site of drug action) is missing.
Therefore, it is uncertain whether and to what extent DEX is metabolized in the heart and
cardiomyocytes, and whether ADR-925 can effectively penetrate this compartment from cell
media and circulation.
The aim of the present study is to thoroughly investigate the metabolism of DEX to ADR-925
in vitro and in vivo on the schedule in which the drug is clearly cardioprotective, with a
particular focus on cardiomyocytes and the heart.
7

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Materials and Methods
Drugs and chemicals
For HPLC analysis, gradient grade methanol (MeOH) was purchased from Merck (Darmstadt,
Germany). Ammonium formate, 1,2-bis(3,5-dioxopiperazin-1-yl)ethane (I.S.DEX), 1,3-
diaminopropane-N,N,N,N-tetraacetic acid (I.S.ADR-925) were obtained from Sigma–Aldrich
(Schnelldorf, Germany). Milli-Q water was produced by a Millipore purification system
(Schwalbach, Germany).
For cell culture experiments, Dulbecco’s modified Eagle’s medium containing the nutrient
mixture F-12 (DMEM/F12), penicillin/streptomycin solution (5000 U/mL; P/S) and sodium
pyruvate solution (100 mM; PYR) were purchased from Lonza (Verviers, Belgium). Horse
serum (HS) and fetal bovine serum (FBS) were purchased from Sigma-Aldrich or Biosera
(Nuaille, France) and were heat-inactivated prior to use (56 °C, 30 minutes). For isolation of
cardiomyocytes, collagenase type II (Thermo Fisher Scientific, Waltham, USA) and pancreatin
(Sigma-Aldrich) were used. Other chemicals used for in vitro and in vivo experiments (e.g.,
constituents of various buffers) were purchased from Sigma–Aldrich or Penta (Prague, Czech
Republic). Solutions of DEX and ADR-925 for all in vitro experiments were prepared as 1000x
concentrated stock solutions in DMSO.
DEX (MW = 268 g/mol) was obtained from Huaren Chemicals (Changzhou, China).
Cardioxane (dexrazoxane, Clinigen Healthcare, Burton‐on‐Trent, UK) was used for in vivo
pharmacokinetic studies. The racemic form of ADR-925 (ICRF-198) (MW = 304 g/mol) was
prepared in house; for a full description of the procedure, see the Supplementary Methods.
Ketamine (Narketan, Vétoquinol, Lure, France), midazolam (Midazolam Torrex, Chiesi
Pharmaceuticals, Wien, Austria) and pentobarbital (Sigma-Aldrich) were used for anesthesia.
8

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Isolation of rat cardiomyocytes and in vitro pharmacokinetic experiments
Neonatal rat ventricular cardiomyocytes (NVCM) have been used for in vitro pharmacokinetic
studies because this is one of the most established and extensively used models for the in vitro
study of cardioprotective effects of DEX (Hasinoff, 2002; Hasinoff et al., 2003a; Simunek et
al., 2008; Jirkovska-Vavrova et al., 2015). In addition, this system was previously employed in
the initial evaluation of DEX metabolism (Schroeder et al., 2005). NVCM were isolated from
1–3-day-old Wistar rats as described previously (Vavrova et al., 2013). Briefly, neonatal hearts
were minced in ADS buffer (1.2 mM MgSO4.7H2O; 116 mM NaCl; 5.3 mM KCl; 1.13 mM
NaH2PO4.H2O; 20 mM HEPES, pH 7.4) on ice and serially digested at 37 °C with collagenase
type II (65 U/mL) and pancreatin (0.4 mg/mL). To minimize non-myocyte cell contamination,
a 2-hour pre-plating period (approx. 20 hearts per 150 mm Petri dish) was used before plating
the cells on gelatine-coated 60-mm Petri dishes at a density of 2 x 105 cells per cm2. NVCM
were cultured at 37 °C in a 5 % CO2 atmosphere in DMEM/F12 medium supplemented with
10 % HS, 5 % FBS, 4 mM PYR and 1 % P/S (pH 7.4). After 40 hours, the medium was changed
to DMEM/F12 supplemented with 5 % FBS, 4 mM PYR and 1 % P/S. For all experiments, the
medium was changed to serum- and pyruvate-free DMEM/F12 with 1 % P/S. The use of
experimental animals for primary cell culture preparation has been approved by the Animal
Welfare Committee of the Faculty of Pharmacy, Charles University.
For investigation of DEX metabolism to ADR-925, NVCM (4.8 × 106 cells) were incubated
with 100 µM DEX. This DEX concentration has been repeatedly shown to be protective against
ANT toxicity in this cell model (Hasinoff, 2002; Adamcova et al., 2007; Jirkovska-Vavrova et
al., 2015). Incubation in the same setting was performed with ADR-925 (100 µM) to determine
the ability of the metabolite to penetrate into NVCM. After 3, 6, 9, 12 and 24 hours of
incubation in separate dishes with the compounds, samples of the cell culture media were taken,
and NVCM were briefly washed twice with PBS at 4 °C, harvested by cell scraping and
9

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
centrifuged (700 x g; 10 minutes; 4 °C). Incubations of both compounds with cell-free culture
medium were performed to determine their chemical degradation under these conditions. To
investigate the possible involvement of active transport of ADR-925 to NVCM, its
concentration inside the cells was determined after a 24-hour incubation with the compound at
concentrations of 100, 300 and 1000 µM ADR-925 at both 4 and 37 °C. All samples of cell
culture media and dry cell pellets were shock-frozen in liquid nitrogen immediately after
collection and kept at -80 °C until their analysis. All experiments were performed in
quadruplicate.
To determine the contribution of NVCM to conversion of DEX to ADR-925, the lysates of
freshly harvested NVCM (4.8 × 106 cells) were used. Pellets of the cells were resuspended in
400 µL of ADS buffer with DEX (100 µM) and sonicated (5 second stroke). The cell-free ADS
buffer supplemented with the drug served as a control. The mixture was incubated for 24 hours
in dark with slow shaking (37 °C). DEX and ADR-925 concentrations were determined at
selected time points (0, 1, 3, 6 and 24 hours) in 50 µL samples of supernatant (3 minutes, RT,
5 000 x g) obtained from the incubation mixture. Internal standards were added, and the
samples were treated with 300 µL of ice-cold MeOH, vortexed and centrifuged (2 minutes,
4 °C, 10 000 x g). Final supernatants were filtered through a 0.22 µm porosity filter and
immediately analyzed by HPLC-MS.
In vivo pharmacokinetic experiments in rabbits
All in vivo experiments and procedures were approved and supervised by Animal Welfare
Committee of the Faculty of Medicine in Hradec Kralove, Charles University. Chinchilla male
rabbits (n = 30, 3.8 kg) were randomized into 6 groups (n = 5 each) and fasted overnight
before the experiment. Four groups received a single dose of DEX (60 mg/kg, i.p.), and the
10

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
experiment was terminated at different time points after drug administration: 0.5, 3, 6 and
12 hours. The other two groups of rabbits received ADR-925 at the same dose (60 mg/kg) in
30-minutes i.v. infusion into the marginal ear vein to describe the pharmacokinetics of ADR-
925 alone, and the experiment was terminated 3 and 12 hours after the start of the infusion. The
particular DEX dose and route of administration have been repeatedly shown to provide
significant cardioprotection against chronic ANT cardiotoxicity in rabbits when administered
0.5 hours before each ANT dose (Simunek et al., 2004; Popelova et al., 2009; Jirkovsky et al.,
2013), and is based on the upper limit of DEX dose previously used in humans cardioprotective
study (Swain and Vici, 2004) and also for treatment of ANT extravasation(Kane et al., 2008),
i.e. the ratio 1:20 to ANT dose (3 mg/kg). The i.p. route of administration has been typically
used in most cardioprotective experiments in small laboratory animals including rabbits to
avoid possible difficulties with repeated administration of both studied agents i.v.. The rabbit
model was among the first employed to study the cardioprotective effects of DEX in vivo and
until today, it is one of the best established models providing constantly reliable data (Herman
and Ferrans, 1998; Sterba et al., 2013).
Blood samples (≈ 1 mL) were collected from the contralateral ear vein at selected time points
(5 minutes–12 hours) after drug administration into BD Vacutainers (Becton Dickenson and
Company, Plymouth, UK) containing lithium heparin. Plasma samples were harvested
immediately, shock-frozen in liquid nitrogen and stored at -80 °C until analysis.
For drug administration and blood sampling, light anesthesia was induced in rabbits with a
mixture of ketamine (25 mg/kg, i.m.) and midazolam (2.5 mg/kg, i.m.); sedation was then
maintained with midazolam when needed. The blood withdrawn from the animals was
compensated with an appropriate volume of sterile saline. Urine was collected throughout the
experiment, volumes were determined and samples were frozen in liquid nitrogen and stored in
-80 °C until analyzed.
11

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
At the defined time points, rabbits were overdosed with pentobarbital. The heart was rapidly
excised, briefly retrogradely perfused with ice-cold saline and dried with gauze. Left ventricles
were dissected and shock frozen in liquid nitrogen. Similarly, the liver and soleus muscle were
also collected and frozen in liquid nitrogen. In addition, the volume of urine in the urinary
bladder was determined, and urine samples were collected and frozen for analysis.
To determine the effects of tissue homogenates on DEX conversion to ADR-925, freshly
harvested rabbit LV myocardium and liver obtained from 2 rabbits were used. Samples were
homogenized in 5 volumes of ADS buffer on ice in a glass/PTFE Potter-Elvehjem tissue
grinder. Freshly prepared homogenates were diluted with ADS buffer to a final volume of 1 mL
to prepare homogenate samples containing two different amounts of the tissue (10 and 100 mg
per sample). DEX in a final concentration of 100 µM was added, and this mixture was incubated
for 24 hours in dark with slow shaking (37 °C, n = 4 in each tissue amount). The tissue-free
ADS buffer supplemented with the drug served as a control. DEX and ADR-925 concentrations
were determined at selected time points (0, 1, 3, 6 and 24 hours) in 20 µL of supernatant
(3 minutes, RT, 5 000 x g) obtained from the incubation mixture. Internal standards were added
to the collected samples, and the samples were treated with 180 µL of ice-cold MeOH, vortexed
and centrifuged (2 minutes, 4 °C, 10 000 x g). Final supernatants were filtered through a
0.22 µm porosity filter and immediately analyzed. In addition, DEX (100 µM) was incubated
in blank rabbit plasma at 37 °C to examine conversion of DEX to ADR-925 in this material in
vitro using similar experimental design as described for tissue homogenates.
Determination of DEX and ADR-925 concentrations in biological samples
A previously described HPLC-MS method (Kovarikova et al., 2013) was adapted and validated
to determine DEX and ADR-925 concentrations in samples obtained from in vitro and in vivo
experiments. However, validation of this method (according to FDA guidelines) for
12

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
simultaneous determination of both DEX and ADR-925 besides the intermediate hydrolytic
products B and C was not possible and therefore the intermediates B and C were not followed
in this study. Briefly, the analyses were performed on an LC 20A Prominence (Shimadzu,
Duisburg, Germany) chromatographic system coupled on-line with LCQ Advantage Max mass
spectrometer (Thermo Finnigan, San Jose, USA). Synergi Polar-RP column (150 mm × 3 mm,
4 μm; Phenomenex, Torrance, USA) was used as a stationary phase. The mobile phase
composed of components A (5 % MeOH in 2 mM ammonium formate solution, without pH
adjustment) and B (MeOH) was employed in the following gradient: 0–7 minutes (0 % B), 7–
8 minutes (0–50 % B), 8–22 minutes (50 % B), 22.0–22.1 minutes (50–0 % B), 22.1–
30.0 minutes (0 % B). A mobile phase flow rate of 0.15 mL/min, a column oven temperature
of 25 °C, the injection volume of 10 μL and an autosampler temperature of 4 °C were used for
analysis. ESI in positive mode was used, and quantification was performed in SRM mode: DEX
+
+
+
([M+H] at m/z 269 → m/z 155), I.S.DEX ([M+H] at m/z 255 → m/z 141), ADR-925 ([M+K] at
+
m/z 343 → m/z 299) and I.S.ADR-925 ([M+K] at m/z 345 → m/z 301).
Both internal standards were added to all biological samples before treatment. Pellets (4.8 × 106
NVCM) were precipitated with ice-cold MeOH (300 μL) and homogenized using a Bio–vortex
mixer (BioSpec Products, Bartlesville, USA) for 30 s and then in an ultrasonic bath. The
samples were centrifuged (10 minutes, 4 °C, 16,800 × g), and the resulting supernatant was
filtered through a 0.45 μm porosity filter (Milex – Hv, Merck-Millipore, Darmstadt, Germany)
and analyzed. Cell culture medium samples (50 μL) were diluted with MeOH (20 ×) and
directly analyzed. Rabbit plasma (100 μL) was precipitated with ice-cold MeOH (600 µL),
vortexed, and after centrifugation (10 minutes, 4 °C, 16,800 × g), the supernatants were
collected and analyzed. Urine samples were centrifuged, diluted (50-100 x) with a MeOH/water
mixture (1:1, v/v) and directly analyzed. Tissues were pulverized under liquid nitrogen, and
samples (≈ 100 mg) of each tissue were precipitated with ice-cold methanol (500 µL) and
13

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
homogenized using an Ultraturax homogenizer (IKA, Staufen, Germany). After centrifugation
(10 minutes, 4 °C, 16,800 × g), the supernatants were collected and analyzed.
For each biological material, the method was validated according to the U.S. Food and Drug
(FDA) Administration Guidance for Industry: Bioanalytical Method Validation (2001). The
linearity of the methods were proven within following concentration ranges: 4–80 pmol/106
NVCM and 7–70 pmol/106 NVCM for DEX and ADR-925, respectively; 8–100 μM in cell
culture medium and ADS buffer (both compounds); 2–340 μM (DEX) and 1–100 μM (ADR-
925) in plasma; 15–1500 μM in urine (both compounds) and 3.7–200 nmol/g and 1.6–
16 nmol/g per wet tissue weight for DEX and ADR-925, respectively. The samples containing
higher concentrations of the analytes were appropriately diluted to fit the relevant concentration
range. The applicability of this approach was proved by testing the dilution integrity. Precision
and accuracy for all biological materials were ≤ 15 and 20 % for all concentrations and LLOQ,
respectively, and thus meet FDA requirements.
Pharmacokinetic analysis of data from in vivo experiments
Standard non-compartmental analysis was performed using Kinetica software (version 4.0,
Thermo Fisher Scientific Inc., USA). The nonlinear mixed-effects modeling approach was used
with pharmacokinetic parameters estimated by the Stochastic Approximation Expectation
Maximization algorithm coupled with the Markov Chain Monte Carlo procedure (SAEM-
MCMC) as implemented in the software Monolix (version 4; Lixoft, France). The consecutive
steps of the pharmacokinetic analysis are as follows:
Naive pooled data (NPD) approach. The concentration-time data of 20 animals subjected to
destructive sampling after a single dose of DEX i.p. were naive pooled, and the geometric mean
concentration-time profiles of plasma DEX and ADR-925 were subjected to standard non-
compartmental analysis (NCA). The same approach was used to calculate model-independent
14

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
parameters of plasma ADR-925 from 96 assayed concentrations after i.v. infusion of the
metabolite to 9 animals. The cumulative amounts of DEX and ADR-925 excreted in urine by
all animals were plotted against time, the parameters of one-phase association curve were fitted
to the data, and the mean total amounts excreted as unchanged drug and/or ADR-925 were
calculated as asymptotes. The fractions of the dose excreted as unchanged drug and/or the
metabolite and renal clearances were computed.
The standard two-stage (STS) method. To improve the accuracy of the PK naive pooled data
biased by animals terminated at the time of unfinished metabolization and elimination
processes, PK parameters after both DEX and ADR-925 administration were estimated by NCA
using data of the subgroup of the animals with entire 12-hour sampling (n = 5 for each
compound). The total urinary excretions and renal clearances were calculated without
extrapolation.
Population modeling. Logarithms of molar DEX and ADR-925 concentrations in plasma and
of their urinary molar quantities improved the precision of parameter estimates and were used
throughout. Inter-animal variability was evaluated with the use of an exponential error model,
and residual variability was evaluated with an additive error model on the logarithmic scale.
First, separate models were built up that described DEX plasma concentration and the amount
of DEX excreted in urine after its i.p. injection (Model 1), and for plasma and urinary data of
ADR-925 after i.v. infusion of the metabolite (Model 2). Finally, a joint model was elaborated
that simultaneously described DEX and ADR-925 plasma concentrations and urinary amounts
after the i.p. injection of the parent compound (Model 3).
The compartmental models were parameterized in terms of volume of distribution (central and
peripheral), inter-compartmental clearance, total and renal clearances, and, in case of the joint
Model 3, the metabolic clearance of DEX to ADR-925. The models for DEX assumed first-
15

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
order absorption without a lag time, and the absolute bioavailability F = 1 was assumed based
on the rapid appearance of DEX in plasma after administration.
Urinary recovery of the i.v. dose of ADR-925 was shown to be complete in this study.
Therefore, the metabolite volumes of distribution were identifiable in Model 3, which included
the data on urinary excretion of ADR-925. The predictive performance of the models was
evaluated using goodness-of-fit plots and the distribution of observations with respect to the
model-predicted median profiles and 90 % prediction intervals, i.e., the visual predictive checks
(VPC). Finally, empirical Bayes estimates of each parameter were obtained, and secondary
pharmacokinetic characteristics were calculated.
In the case of tissue concentrations, the mean value and variance of the AUC were calculated
with the help of the package PK, version 1.3-3, named Basic Non-Compartmental
Pharmacokinetics (Jaki et al., 2010).
Data analysis and presentation
Data are shown as the mean ± SD unless stated otherwise. Pearson and Spearman correlation
analyses, linear regression analyses, and statistical significance differences (using One-Way
ANOVA with Holm-Sidak’s or ANOVA on Ranks with Dunn’s post hoc test) were determined
according to the data character using software SigmaStat 3.5 (SPSS, USA).
16

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Results
Distribution of DEX/ADR-925 in isolated rat cardiomyocytes and conversion of DEX to
ADR-925
DEX and ADR-925 concentrations were determined in cell culture medium and inside NVCM
after their incubation with DEX (100 µM, up to 24 hours). In the culture medium with NVCM,
DEX concentrations gradually fell to ≈ 30 µM after 24 hours (Fig. 2A), while ADR-925
concentrations steadily increased from the 3rd hour to exceed the concentrations of the parent
compound between the 12th and 24th hours, reaching ≈ 45 µM at the end of the experiment (Fig.
2A). Almost identical results were obtained in culture media without NVCM (Fig. 2A), which
indicates that DEX conversion to ADR-925 is related to spontaneous hydrolysis rather than
cardiomyocyte metabolic activity. Inside NVCM, DEX was determined after 3 hours of
incubation, with its peak value in the 6th hour. The concentrations of DEX gradually declined
thereafter but were detectable until the end of the incubation (Fig. 2B). In the same experiment,
ADR-925 concentrations inside the cells were first detected in the 9th hour and then
progressively increased until the end of experiment (24 hours), where it achieved concentrations
approximately twice as high as those of DEX (Fig. 2B). Direct incubation of NVCM with ADR-
925 (in the same concentrations as DEX, 100 µM) resulted in a slower penetration of this agent
into the intracellular compartment compared to DEX, but the concentrations were considerably
higher than those formed in the previous experiment after incubation of the cells with DEX.
The concentrations of ADR-925 then gradually increased in time to match the concentrations
of DEX determined in the previous experiment in the 9th hour (Fig. 2B), and a further increase
of ADR-925 concentration continued until the end of the incubation. Intracellular
concentrations of ADR-925 were almost double throughout this experiment compared to the
concentrations of ADR-925 determined after DEX incubation.
17

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
In further experiments, we wanted to examine whether and to what extent active transport
contributes to the penetration of the relatively polar metabolite ADR-925 into NVCM.
Therefore, these cells were incubated with ADR-925 (100–1000 µM) at 4 °C or 37 °C for 24
hours. The result of this experiment showed that the penetration of ADR-925 into the cells is
temperature-dependent, as the concentrations inside the cells were significantly lower at 4 °C
compared to 37 °C, which indicates contribution of an active type of transport (Fig. 2C). On the
other hand, the experiment performed with the higher concentration did not show any ceiling
effect, which suggests that the transport is not saturable in the tested concentration range.
To directly assess the role of intracellular enzymes in DEX metabolism to ADR-925, 100 μM
DEX was incubated for 24 hours with NVCM homogenate or the corresponding ADS buffer.
The results showed that NVCM homogenate (Fig. 2D) did not have any significant effect on
the formation of ADR-925 from DEX, and the concentrations were comparable with those
determined in the buffer (Fig. 2D).
DEX metabolism and disposition in rabbits
Plasma concentrations and urinary excretion of DEX and ADR-925
The rate of absorption of DEX (60 mg/kg) administered as a single intraperitoneal bolus to
rabbits was high (Fig. 3A). The highest plasma concentrations of DEX were observed in
individual time profiles between the 5th and 30th minute after the injection with the median time
interval of 10 minutes. The geometric mean concentration of plasma DEX peaked at the
10th minute, and its decrease from the cmax (337  was apparently biphasic, leaving only
1.7 % of the cmax concentration at the end of the twelve-hour sampling interval.
The metabolite was detected in all plasma samples collected at the 5th minute, but its
concentrations (Fig. 3A) were relatively lower compared to the parent compound. Its geometric
mean of concentration increased gradually until the 45th minute, reaching the cmax (35.6 .
18

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
A broad plateau was then observed on the mean concentration-time profile of ADR-925
between the 45th minute and the 5th hour. After this time interval, the mean concentration of the
metabolite showed a monotonous decline in parallel with that of DEX. The residual metabolite
concentrations at the 12th hour corresponded to 14 % of the cmax. The slow formation of ADR-
925 in plasma observed in the curve of the mean concentrations followed with a plateau was
not the effect of pooling of the data from different animals as it was also found on the
concentration-time profiles of individual animals (data not shown). The median time to the
maximum concentration was 2 hours (0.5–8 hours) in 13 rabbits that were sampled over a
sufficiently long interval to reach the cmax
.
In another experiment, ADR-925 (60 mg/kg) was infused intravenously to rabbits over
30 minutes to follow the distribution and elimination kinetics of the metabolite from plasma.
The highest concentrations of ADR-925 in plasma were observed at the end of the infusion
(30 minutes) with the exception of one animal (20 minutes). The geometric mean of the cmax
was found to be 557 . Similar to DEX, the decline of the mean plasma concentrations of
ADR-925 exhibited a biphasic pattern and, after 12 hours, only 0.2 % of the cmax remained in
plasma (Fig. 3B).
The cumulative urinary excretion of DEX and ADR-925 followed for 12 hours after a bolus
dose of 60 mg/kg DEX i.p. is shown in Fig. 4A. The mean total amounts of DEX and ADR-
925 excreted via the urine, estimated as asymptotes by nonlinear regression, were 309 mol
and 103 mol, respectively. The amounts of the parent drug, metabolite, and their sum in the
urine estimated from all animals (NPD) represent 35.5 % (DEX), 11.9 % (ADR-925) and
47.4 % (DEX + ADR-925) of the administered dose, respectively. Similarly, in the subgroup
of 5 animals with urine collected over the entire 12 hours, the total recoveries were
19

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
35.7 ± 11.1 % (DEX), 11.0 ± 4.3 % (ADR-925) and 46.6 ± 10.9 % (DEX + ADR-925) of the
dose.
A vast majority of the intravenously infused dose of ADR-925 was excreted unchanged in urine,
supporting the view that renal excretion is a predominant elimination pathway for this
metabolite (Fig. 4B). The mean total urinary excretion and recovery of ADR-925 estimated
from all naive-pooled animals were 664 mol and 89.5 %, respectively. Nonlinear regression
may have underestimated the total amount of the drugs excreted by the kidneys due to the
influence of data from animals with shorter intervals of collection. In the subgroup of the
animals within the 12-hour collection interval, the mean urinary recovery of the infused dose
of the metabolite achieved 98 ± 6 %, indicating that the excretion was practically complete.
Non-compartmental analysis
Pharmacokinetic parameters calculated by non-compartmental analysis (NCA) are summarized
in Table 1. Due to the batch design of the experiment, i.e., the variable length of the sampling
interval in subgroups of animals, NCA could be applied on the geometric mean of
concentration-time profile obtained from all naive-pooled animals (n = 20) and on individual
profiles of animals with 12-hour sampling (n = 5). Blood sampling over 12 hours was adequate
for the description of plasma pharmacokinetics of DEX and ADR-925 in both studies. After a
bolus dose of DEX, the extrapolated parts of the AUCs were less than 4 % (DEX) and 10 %
(ADR-925), with the exception of one individual concentration-time profile of ADR-925 where
the extrapolated part of the AUC was 23 %. After i.v. infusion of ADR-925, extrapolation
beyond the 12th hour accounted for less than 1 % of the AUC.
In general, there was good agreement between the results of the NCA analysis of naive-pooled
concentration-time data from all animals and the two-stage approach used in five animals
sampled over 12 hours (Table 1). Two exceptions were the cmax of DEX and tmax of the
20

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
metabolite after DEX injection. Obviously, the single-point pharmacokinetic characteristics
cmax and tmax are more variable between animals, and their estimates are more prone to bias. In
addition, the broad plateau on the ADR-925 mean concentration-time profile explains the
discrepancy regarding its tmax
.
The fraction of the DEX dose metabolized to ADR-925 (11 %) was calculated assuming that
the absolute bioavailability of DEX equals one (F = 1) and that renal excretion is the sole
elimination pathway for the metabolite after DEX injection, which is supported by our data
after infusion of ADR-925. As judged from the ratio of the mean AUCs (parent/metabolite)
of 3.2, DEX was the quantitatively prevailing compound in plasma.
The renal clearance of ADR-925 after DEX injection was approximately two-fold lower
compared to infusion of ADR-925. Of note, the cmax of plasma ADR-925 was 15-fold higher at
the end of the i.v. infusion of ADR-925, and the concentrations remained markedly higher
during the first post-infusion hour compared to the maximum concentrations of ADR-925
metabolically formed after DEX injection. The finding of the concentration-dependent renal
clearance complicated the comparison of the MRT and t1/2 values of ADR-925 in both studies.
The MRTi.v. of ADR-925 (2.5 hours), calculated as (MRTmetabolite - MRTparent) after DEX
injection, was thus higher than the MRTi.v. of ADR-925 after infusion of ADR-925 (1.6 hours),
obtained as the difference (MRTinf – Tinf/2).
Population pharmacokinetic analysis
A schematic representation of the separate models for DEX and ADR-925 is given in Fig. 5A
and Fig. 5B, respectively. After i.p. injection, disposition of DEX in plasma was best described
by a two-compartment open model with first-order absorption and elimination without a time
lag (Model 1). The final model for ADR-925 was a two-compartment model with zero-order
input via i.v. infusion and first-order elimination (Model 2). In the joint parent-metabolite model
21

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
(Model 3; Fig. 5C), formation of ADR-925 from plasma DEX was modeled as a first-order
process, and the amount of the metabolite recovered in urine was taken as representative of its
overall quantity produced by metabolism. The central and peripheral volumes of distribution of
the metabolite were added to achieve the best description of the combined data from DEX and
ADR-925.
The statistical summary of Bayesian post hoc estimates of individual pharmacokinetic
parameters produced by Models 1, 2 and 3 are presented in the Table 2. The relative standard
errors not exceeding 24 % document an adequate precision of estimates for all parameters.
Values of clearances and volumes of distribution agreed reasonably well with those from NCA.
The joint parent-metabolite model also allowed for determination of the metabolic clearance of
DEX (0.045 L·h-1·kg-1) and the estimation of secondary pharmacokinetic parameters for DEX
presented as a geometric mean: Cltot = 0.26 L·h-1·kg-1, Vss = 0.74 L/kg, t1/2α = 11.5 minutes and
t1/2β = 131.4 minutes.
The visual predictive performance and goodness-of-fit of the models was good (Fig. 6 and
Suppl. Fig. 1). The observed plasma DEX and ADR-925 concentrations fit well within the 5–
95 % of prediction intervals. Population mean predicted plasma concentrations adequately
describe the central tendencies in concentration-time data. The goodness-of-fit is evident from
the scatter plots of observed vs. individual predicted plasma concentrations, calculated by
Bayesian estimation (Fig. 6 and Suppl. Figs. 1C, D).
Tissue concentrations of DEX and ADR-925 in rabbits
DEX and its metabolite ADR-925 were also determined in the selected tissues (LV
myocardium, liver and soleus muscle) at different time points after DEX administration to
rabbits (Fig. 7A–C). Very similar DEX and ADR-925 concentrations (including the overall
concentration-time profile) were determined in the LV myocardium and in the liver. The
22

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
concentrations of DEX in both tissues were more than double compared to the soleus muscle
30 minutes after drug administration, and they decreased faster than in the muscle. Both in the
myocardium and liver, the decrease in DEX concentrations was not accompanied by a
proportional increase of ADR-925. In fact, the concentrations of the metabolite remained
relatively steady throughout the experiment particularly in the liver, with tendency towards a
slow rise in the heart. These data argue against distinct intracellular trapping and cumulation of
the relatively polar ADR-925 in this tissue. A different situation was demonstrated in soleus
muscle, where the concentrations of the metabolite increased more significantly compared to
the decline of DEX. Different concentrations of both drugs in the soleus muscle compared to
other tissues could be caused by lower perfusion of the peripheral muscle compared to the
myocardium and the liver.
The DEX concentrations determined in plasma at the end of the study showed a very good
relationship with the corresponding concentrations found in the tissues in the same time point
(Suppl. Fig. 2A-C). However, this was not the case for ADR-925 (Suppl. Fig. 2D-F), where the
concentrations in the myocardium and liver tissue showed no significant correlation with the
plasma concentrations (p > 0.05), and only poor and a surprisingly negative correlation was
found in the soleus muscle (p = 0.021, R = -0.513).
The total exposure of the tissues to DEX and ADR-925 estimated as AUC (Table 3) revealed
that after DEX administration, the LV myocardium and liver tissue were exposed to 5.7 and 4.8
times (respectively) higher concentrations of the parent drug compared to ADR-925, whereas
the ratio was lower in the soleus muscle (3.5 times). The ratios in the former tissues are
significantly higher than that found in plasma (3.2 times).
To directly estimate tissue metabolism of DEX to ADR-925, we incubated LV myocardial and
liver homogenates with DEX at 37 °C and determined concentrations of both of these
23

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
compounds (Fig. 8A, B). Only a slight decrease of DEX concentrations occurred after 6 hours
of incubation with myocardial homogenates, which corresponds with low concentrations of the
metabolite. Importantly, the amount of the myocardial tissue in the homogenate had no impact
on the results, and the concentrations were very similar as in ADS buffer without tissue
homogenate. In contrast, the same experiments performed with liver homogenates documented
a significantly accelerated decrease of DEX concentration with a concomitant increase of ADR-
925. Because there was also an apparent difference between the samples prepared with a
different amount of the liver tissue, it seems that this activity is attributable to content of the
tissue and biotransformation enzymes. To estimate whether DEX may also be converted to
ADR-925 in plasma more effectively than by simple chemical hydrolysis in the buffer, DEX
was incubated in rabbit plasma at 37 °C (Fig. 9). These experiments demonstrated that DEX is
converted to ADR-925 in plasma significantly more effectively than in ADS buffer. After 6
hours of incubation, plasma concentrations of DEX decreased by more than 60 %, while in
ADS buffer it was approx. 20 %, and the concentrations of ADR-925 rose concomitantly.
Interestingly, DEX concentrations in plasma dropped down after 24 hours to negligible values,
while ADR-925 concentrations approached 100 μM, suggesting almost complete conversion of
DEX to ADR-925 under these conditions. This was not the case in ADS buffer, where the
decline of DEX concentrations was approx. 50 % at the same time point with a similar increase
of ADR-925 concentrations. Although intermediate metabolites B and C could not be
determined in these experiments, their theoretical portion can be approximated based on mass
balance as a resting portion on the top of the sum of DEX and ADR-925 (Fig. 9B).
24

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Discussion
This study adds several important pieces into the puzzle of the pharmacokinetics of DEX and
its putative active metabolite ADR-925 under well-established experimental conditions in
which DEX is known to provide reliable protection from ANT cardiotoxicity. We have
documented a slow and progressive rise of intracellular concentrations of ADR-925 in NVCM
after DEX incubation. We have proven that ADR-925 can directly, albeit relatively slowly,
penetrate from the cell culture medium into cardiomyocytes, and the resulting intracellular
concentrations are significantly higher than after incubation of the cells with equimolar DEX
concentrations. Our data suggest the contribution of an active transmembrane transport, but
particular transporter for ADR-925 remains unknown. According to its chemical structure, one
could speculate on involvement of a member of the solute carrier family (SLC) transporters,
especially for organic cation/carnitine transporter family (OCTN, SLC22) (Klaassen and
Aleksunes, 2010; Tamai, 2013), from which OCTN2 was described to be present in the mouse
and human heart (Grube et al., 2006). Of note, we showed that significant non-enzymatic
hydrolysis of DEX to ADR-925 already occurs in the culture media, and the rate is faster than
in the buffer of corresponding pH. Hence, during 12-hour experiments, the cells are exposed to
relatively high concentrations of ADR-925, and in longer experiments, the exposure to the
metabolite predominates over DEX. It is noteworthy that NVCM lysate had no significant
impact on conversion of DEX to ADR-925. This may imply that a significant part of ADR-925
intracellular concentrations found after incubation with DEX can arise from penetration of
ADR-925 from the medium, where it originates by non-enzymatic hydrolysis of the parent
compound. The hydrolysis has been described to be promoted by presence of metal-ions (esp.
Fe2+, Ca2+, Cu2+ and Mg2+) (Buss and Hasinoff, 1995; Buss and Hasinoff, 1997), although in
the rather higher concentrations then they are present in the present cell media. On the other
hand, Schroeder et al. (2005) previously published that suspension of neonatal cardiomyocytes
25

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
accelerated the decrease of DEX concentrations in the medium, which may appear to be in
contrast to our findings. However, the interpretation of these results is complicated by short
incubation of the cells with DEX (2 hours), resulting in a relatively small decline of DEX
concentration, and by a lack of information on ADR-925 formation.
In contrast to DEX, incubation of isolated rat cardiomyocytes with ADR-925 has been
previously reported to be unable to protect the cells from doxorubicin toxicity, and one possible
explanation was poor penetration of the putative active metabolite into the cells due to its
anionic hydrophilic nature (Hasinoff et al., 2003b). However, our data obtained also from rat
cardiomyocytes clearly contradict this hypothesis and further contribute to questioning ADR-
925 as an active metabolite responsible for DEX-induced cardioprotection (Sterba et al., 2013).
Plasma pharmacokinetic profiles of DEX in rabbits determined in this study are surprisingly
close to those reported from clinical trials. Basic pharmacokinetic parameters calculated in this
study practically match those reported in patients (in order: our NCA data vs. human data -
NCA data if possible): biological half-life of elimination (t1/2β) 2.0 hours vs. ≈ 2–4 hours
(Earhart et al., 1982; Vogel et al., 1987; Hochster et al., 1992; Rosing et al., 1999), total
clearance 0.26 vs. ≈ 0.19–0.35 L·h-1·kg-1 (Holcenberg et al., 1986; Vogel et al., 1987; Rosing
et al., 1999), volume of distribution (Vss) 0.74 vs. ≈ 0.45–1.1 L/kg (Holcenberg et al., 1986;
Wiseman and Spencer, 1998), and the recovery of administered DEX dose in urine (as
DEX + ADR-925) 47 % vs. 40–48 % (Earhart et al., 1982; Rosing et al., 1999; Tetef et al.,
2001). DEX pharmacokinetic in humans was described to be different in pediatric patients and
in adults (Holcenberg et al., 1986), and to be linear in broad range of DEX concentrations
(Hochster et al., 1992). Our findings in rabbits thus differ considerably from rodents where the
elimination of DEX has been reported as 4–6 times faster (Hasinoff et al., 1998; Hasinoff and
Aoyama, 1999a). This suggests that rabbit is a valuable animal model to study the
cardioprotective effects of DEX not only from a pharmacodynamic (Herman and Ferrans, 1998;
26

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Simunek et al., 2004; Popelova et al., 2009; Jirkovsky et al., 2012) but also from a
pharmacokinetic point of view.
The plasma profile of the putative active metabolite (ADR-925) after DEX administration to
rabbits was also relatively close to reports from the single human study published so far
(Schroeder et al., 2003). In both, the metabolite appeared in plasma in the first sample taken
after drug administration and formed a plateau for several hours (≈ 5 hours in rabbits and ≈ 4
hours in humans) with a relatively slow decline until the end of the study. Although the
administered doses were not the same in these studies, the range of plasma concentrations of
ADR-925 in the plateau phase was similar (≈ 40 μM after 1000 mg/m2 of DEX in rabbits and
30 μM after 1500 mg/m2 in patients). Unfortunately, the pharmacokinetic analysis for ADR-
925 was not performed in the clinical study, and thus the pharmacokinetic parameters cannot
be directly compared. The profile of ADR-925 was also studied in rats after DEX administration
(40 mg/kg, i.v.) (Schroeder and Hasinoff, 2002). In the latter study, ADR-925 was detected in
plasma soon after DEX administration, with a peak of plasma concentrations in the 80th minute
(≈ 80 μM) and tended to decrease faster than in rabbits and humans. The short duration of this
study (3 hours) did not allow for appropriate evaluation of metabolite elimination. However,
the same authors evaluated the pharmacokinetics of ADR-925 after an i.v. bolus of either ADR-
925 or the intermediate metabolites (B/C) of DEX to rats (Schroeder and Hasinoff, 2005). They
observed a rapid decline of ADR-925 with t1/2 ≈ 20 min (as estimated from the graphical data)
after administration of ADR-925, while an even shorter t1/2 (8.3 minutes) was calculated from
elimination constants for ADR-925 after administration of intermediate metabolites. This
contrasts our data from rabbits, where the t1/2 ≈ 100 minutes. Our study has shown that ADR-
925 is almost completely eliminated unchanged by the kidney to urine. We have also created a
population-based pharmacokinetic model taking into account both parent DEX and ADR-925.
27

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Our study has also systematically described the time course of DEX and ADR-925 tissue
concentrations, especially in LV myocardium as the target tissue for cardioprotection. The data
from patients are naturally unavailable and besides scarce data from conference abstracts
(Mhatre et al., 1983; McPherson et al., 1984), very little information has been published from
animal experiments in extenso so far (Schroeder and Hasinoff, 2005; Schroeder et al., 2008).
As a part of these two studies, myocardial ADR-925 concentrations were described after DEX
administration (40 mg/kg, i.v.) to rats in the single arbitrary selected time point (after 2.8 or 2-
2.5 hours, respectively). However, only the former study also provided concentrations of the
parent compound. The concentrations of ADR-925 tended to be significantly higher in these rat
studies than in our experiments with rabbits, but direct comparison is difficult due to different
dose/route of administration and only discrete information of tissue concentrations in the former
studies. As plasma pharmacokinetics of DEX and ADR-925 is markedly different in rodents
than in humans and rabbits, tissue concentrations of both compounds may also be affected.
We have studied the metabolism of DEX to ADR-925 in rabbit tissue homogenates to reveal
that unlike liver, myocardial tissue does not accelerate DEX hydrolysis and the formation of
ADR-925. This is consistent with our in vitro results obtained from DEX incubation with
NVCM and their lysates and with the study published by Hasinoff et al. (1991). Although the
former study struggled with difficulties such as unavailability of appropriate bioanalytical
methodology and the incubations were rather short (4 hours), they also found a lack of ADR-
925 increase after DEX incubation with porcine heart homogenate, contrary to incubation with
the liver homogenate. This may be evidently connected with the high amount and activity of
dihydropyrimidase, which catalyzes the first hydrolytic step, in the liver in contrast to
myocardial tissue (Hasinoff et al., 1991). Interestingly, we found that DEX was considerably
hydrolyzed to ADR-925 by incubation in fresh rabbit plasma under physiological temperature.
Hence, it is possible that DEX may be significantly converted to ADR-925 in the circulation,
28

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
which may contribute to early detection of ADR-925 after DEX administration in all studies
performed so far. The plasma instability of DEX could not confound data in our study because
its pre- and analytical parts were designed and validated accordingly. Altogether, these data
may imply that the ADR-925 concentrations found inside rat cardiomyocytes and rabbit LV
myocardium are determined largely by distribution of ADR-925 or its intermediate hydrolytic
products from the circulation, where it could originate both by non- or enzymatic hydrolysis
and/or by release from metabolically active organs such as the liver. Differences between
correlations of plasma and tissue concentrations of DEX and ADR-925 may be a consequence
of different distributions into the intracellular compartments (slower for ADR-925), which
correspond with the more hydrophilic nature of ADR-925.
One of the limitations of the present study is the lack of direct information on the two
intermediate hydrolytic metabolites B and C. This was caused by difficulties with their
appropriate analytical determination besides DEX and ARD-925. However, both intermediate
metabolites have been tested previously in vitro for cardioprotective properties against ANT
cardiotoxicity with negative outcomes (Hasinoff et al., 2003) and therefore they were out of our
primal interest. Indirect information on portion of B and C from the plasma hydrolysis
experiments suggested their minor appearance.
In conclusion, the present study filled in several important gaps in the pharmacokinetic profile
and tissue disposition of the clinically used cardioprotectant DEX to its putative active
metabolite ADR-925 in a rabbit model, which shows pharmacokinetic similarities to humans.
Importantly, the finding that ADR-925 could directly enter cardiomyocytes opens a venue for
direct determination of whether ADR-925 is actually the active metabolite of DEX that is
responsible for its cardioprotective effects, as this is now disputed in the literature (Hasinoff
and Herman, 2007; Sterba et al., 2013; Henninger and Fritz, 2017).
29

JPET Fast Forward. Published on December 22, 2017 as DOI: 10.1124/jpet.117.244848
This article has not been copyedited and formatted. The final version may differ from this version.
JPET #244848
Acknowledgement
The authors thank Mrs. Klára Lindrová for excellent technical assistance during the
experiments.
30