Using in vitro lipolysis and SPECT/CT in vivo imaging to understand oral absorption of fenofibrate from lipid-based drug delivery systems

Article abstract of DOI:10.1016/j.jconrel.2019.11.024

Using lipid-based drug delivery systems (LbDDS) is an efficient strategy to enhance the low oral bioavailability of poorly water-soluble drugs. Here the oral absorption of fenofibrate (FF) from LbDDS in rats was investigated in pharmacokinetic, in vitro l

Full text of DOI:10.1016/j.jconrel.2019.11.024

Journal Pre-proof
Using in vitro lipolysis and SPECT/CT in vivo imaging to
understand oral absorption of fenofibrate from lipid-based drug
delivery systems
Thuy Tran, Peter Bønløkke, Cristina Rodríguez-Rodríguez,
Zeynab Nosrati, Pedro Luis Esquinas, Nrupa Borkar, Jakob Plum
Christensen, Sophie Susanna Strindberg Andersen, Stoyan
Karagiozov, Thomas Rades, Anette Müllertz, Katayoun Saatchi,
Urs O. Häfeli
PII:
S0168-3659(19)30681-9
DOI:
https://doi.org/10.1016/j.jconrel.2019.11.024
COREL 10029
Reference:
To appear in:
Journal of Controlled Release
Received date:
Revised date:
Accepted date:
19 August 2019
17 November 2019
18 November 2019
Please cite this article as: T. Tran, P. Bønløkke, C. Rodríguez-Rodríguez, et al., Using in
vitro lipolysis and SPECT/CT in vivo imaging to understand oral absorption of fenofibrate
from lipid-based drug delivery systems, Journal of Controlled Release (2019),
https://doi.org/10.1016/j.jconrel.2019.11.024
This is a PDF file of an article that has undergone enhancements after acceptance, such
as the addition of a cover page and metadata, and formatting for readability, but it is
not yet the definitive version of record. This version will undergo additional copyediting,
typesetting and review before it is published in its final form, but we are providing this
version to give early visibility of the article. Please note that, during the production
process, errors may be discovered which could affect the content, and all legal disclaimers
that apply to the journal pertain.
© 2019 Published by Elsevier.

Journal Pre-proof
Using in vitro lipolysis and SPECT/CT in vivo imaging to understand
oral absorption of fenofibrate from lipid-based drug delivery systems
Thuy Tran a, Peter Bønløkke a, Cristina Rodríguez-Rodríguez b,c,d, Zeynab Nosrati b, Pedro
Luis Esquinas e, Nrupa Borkar a, Jakob Plum Christensen a, Sophie Susanna Strindberg
Andersen a, Stoyan Karagiozov b, Thomas Rades a, Anette Müllertz a,e*, Katayoun Saatchi b*
and Urs O. Häfeli b*
a Department of Pharmacy, Faculty of Health and Medical Sciences, University of
Copenhagen, 2100 Copenhagen, Denmark.
b
Faculty of Pharmaceutical Sciences, University of British Columbia, Vancouver, BC V6T
1Z3, Canada.
c Department of Physics and Astronomy, University of British Columbia, Vancouver, BC
V6T 1Z1, Canada.
d
Medical Imaging Research Group, Department of Radiology, University of British
Columbia, Vancouver, BC V5Z 1M9, Canada.
e
Bioneer: FARMA, Faculty of Health and Medical Sciences, University of Copenhagen,
2100 Copenhagen, Denmark.
* Corresponding authors: anette.mullertz@sund.ku.dk, kathy.saatchi@ubc.ca, urs.hafeli@ubc.ca

Journal Pre-proof
ABSTRACT
Using lipid-based drug delivery systems (LbDDS) is an efficient strategy to enhance the low
oral bioavailability of poorly water-soluble drugs. Here the oral absorption of fenofibrate
(FF) from LbDDS in rats was investigated in pharmacokinetic, in vitro lipolysis, and
SPECT/CT in vivo imaging studies. The investigated formulations were soybean oil solution
(SBO), a mixture of soybean oil and monoacyl phosphatidylcholine (MAPC) (SBO-MAPC),
self-nanoemulsifying drug delivery systems with and without MAPC (SNEDDS-MAPC and
SNEDDS, respectively), and an aqueous suspension (SUSP) as a reference. Oral
bioavailability of the LbDDS ranged from 27 to 35%. A two-step in vitro lipolysis model
simulating rat gastro-intestinal digestion provided in vitro FF solubilisation data to
understand oral absorption. During the in vitro lipolysis, most FF was undissolved for SUSP
and distributed into the poorly dispersed oil phase for SBO. For the SNEDDS without
MAPC, practically all FF solubilised into the aqueous phase during the dispersion and
digestion. Adding MAPC to SBO enhanced the dispersion of the oil phase into the digestion
media while adding MAPC to SNEDDS resulted in a distribution of 29% of FF into the oil
phase at the beginning of in vitro lipolysis. FF distribution into both oil and aqueous phases
explained the higher and prolonged oral absorption of LbDDS containing MAPC. To
elucidate the relatively low bioavailability of all formulations, FF and triolein were labelled
with 123I and 125I, respectively, to study the biodistribution of drug and lipid excipients in a
dual isotope SPECT/CT in vivo imaging study. The concentration of radiolabeled drug as a
function of time in the heart correlated to the plasma curves. A significant amount of
radiolabeled drug and lipids (i.e., 28–59% and 24–60% of radiolabelled drug and lipids,
respectively) was observed in the stomach at 24 h post administration, which can be linked to
the low bioavailability of the formulations. The current study for the first time combined in
vitro lipolysis and dual isotope in vivo imaging to find the root cause of different fenofibrate
absorption profiles from LbDDS and an aqueous suspension.
KEYWORDS
Lipid-based drug delivery systems, monoacyl phosphatidylcholine, surfactant, SPECT/CT, in
vitro lipolysis, pharmacokinetics, biodistribution, fenofibrate, dual radioisotope imaging

Journal Pre-proof
1. INTRODUCTION
Modern drug discovery programs have discovered a large number of poorly water-soluble
drug candidates with low oral absorption [1]. To enable an effective oral delivery of these
molecules, lipid-based formulations (LbDDS) including self-nanoemulsifying drug delivery
systems (SNEDDS) have been used with success [2]. The use of natural surfactants has been
investigated to reduce the content of synthetic surfactants, while maintaining the
emulsification capacity and oral absorption enhancement facilitated by LbDDS [3, 4].
Previous studies on the use of monoacyl phosphatidylcholine (MAPC), which is a desirable
natural surfactant, has clarified its effect on emulsion droplet sizes and colloidal structures
formed during in vitro lipolysis, while understanding its effect on oral absorption is still a
challenge [5].
In vitro lipolysis models has been used to predict and assess the capacity of LbDDS to
enhance oral absorption [6]. During lipid digestion, the drug solubilisation capacity changes
as digestion products are formed and the drug transfers from the original LbDDS into more
polar and well-dispersed colloidal systems in the aqueous phase. Drug partitioning into the
different phases formed during in vitro lipolysis (i.e., oil, aqueous or pellet phases) is thus
commonly studied [6-9]. However, existing in vitro lipolysis models have been developed to
simulate human digestion and thus might not predict the oral bioavailability of a drug in
animals very well [10-14]. Therefore, the development of in vitro lipolysis models simulating
animal digestion conditions is necessary to better predict in vivo drug absorption at
preclinical phase.
Visualizing the in vivo biodistribution of drugs after oral administration can be an effective
strategy to understand drug absorption and the interplay between drug and excipients before
and after absorption. In vivo imaging has been used to investigate the passage of drug and
excipients in the body, but mostly with magnetic-resonance-based non-quantitative methods,
and rarely in the research field of oral LbDDS delivery [15]. As one of the most cutting-edge
in vivo imaging techniques, single photon emission computed tomography (SPECT)
combined with computed tomography (CT) can precisely quantify the amount of drug and
lipid excipients distributed in each organ of a living animal [16-20]. For this purpose, drug
and excipient are radiolabeled with two different isotopes that can then be detected at the
same time, by separating out their distinct energy windows.

Journal Pre-proof
The objective of the current study is to understand the oral absorption of the lipophilic model
drug fenofibrate (FF) from LbDDS and an aqueous suspension (used as a reference) in rats
using in vitro lipolysis and SPECT/CT in vivo imaging. First, oral absorption of FF from the
five formulations (i.e. four different LbDDS and the aqueous suspension) was evaluated in a
pharmacokinetic study in rats. Then, the formulations were assessed using an in vitro
lipolysis model simulating rat gastro-intestinal lipid digestion to understand the effect of
digestion on oral absorption. Finally, the model drug FF and triolein were radiolabeled with
123I and 125I, respectively (Fig. 1). LbDDS with both 123I-FF and 125I-triolein were dosed
orally to rats and dual-isotope SPECT/CT imaging was performed after administration and
subsequently confirmed by a biodistribution study at 24 h to visualize the distribution of both
the drug and the lipid excipients after oral administration. From in vitro lipolysis and
SPEC/CT imaging studies, information on drug solubilisation in the gastro-intestinal tract and
drug distribution in different organs is expected to provide an explanation for the varying
drug absorption obtained from different formulations.
2. MATERIALS AND METHODS
2.1 Materials
Lipoid S LPC 80 (LPC) (from soybean, containing MAPC (81%) and phosphatidylcholine
(PC) (13%) of C18:2 (48–56%), C16:0 (20–27%), C18:0 (5–8%), C18:1 (7–9%) fatty acid
(FA)) and Lipoid S PC (from soybean, containing 98.0% PC) were kindly donated by Lipoid
GmbH (Ludwigshafen, Germany). Kolliphor EL (polyoxyl 35 castor oil) (containing
glycerol-polyethylene glycol (PEG) ricinoleate, FA esters of PEG, free PEG and ethoxylate
glycerol) was kindly donated by BASF (Ludwigshafen, Germany). Maisine 35-1 (containing
monoglycerides (MG) (33.5%), diglycerides (DG) (50.9%) and triglycerides (TG) (14.7%)
predominantly of C18:1, C18:2 and C16:0 FA [21]) was kindly provided by Gattefossé
(Saint-Priest, France). FF, hydroxypropyl methyl cellulose (HPMC), soybean oil (TG of
mainly C18:2, C18:3, C18:1, C18:0 and C16:0 FA [22]), pancreatin from porcine pancreas (8
× United States Pharmacopeia (USP) specification activity), lipase from Rhizopus oryzae
(~10 U/mg), bile bovine, tris(hydroxymethyl)aminomethane (Tris), maleic acid, and 4-
bromophenyl-boronic acid (4-BPB) (≥ 95.0% pure) and organic solvents (including
dichloromethane (DCM), methanol (MeOH) and ethanol (EtOH)), as well as Bu6Sn2 and
Pd(PPh3)4 were purchased from Sigma-Aldrich (St. Louis, MO, USA). Sodium chloride was
purchased from VWR (Søborg, Denmark). Sodium hydroxide (NaOH) pellets were obtained
from Merck (Darmstadt, Germany). Water was purified using a SG Ultraclear water system

Journal Pre-proof
(SG Water GmbH, Barsbüttel, Germany). Na123I was obtained from Nordion (Vancouver,
BC, Canada) and Na125I was obtained from American Radiolabeled Chemicals Inc. (St.
Louis, MO, USA).
A1
A2
A3

Journal Pre-proof
Fig. 1. Chemical structure of (A1, A2, A3) Lipoid S LPC 80 (MAPC (80%) and PC (13%)
of C18:2 (48-56%), C16:0 (20-27%), C18:0 (5-8%), C18:1 (7-9%) FA), (B) FF, (C) 123I-FF,
(D) triolein, and (E) a potential radioiodinated structure for triolein (substitution can happen
at any of the double bonds of triolein).
2.2. Methods
2.2.1. Preparation of formulations
The composition of the investigated LbDDS before adding FF is shown in Table 1. Maisine
35-1 was melted at 50 °C and homogenized before preparing SNEDDS and SNEDDS-
MAPC. All LbDDS were prepared by weighing all excipients and FF into a glass vial to
obtain a FF concentration of 35 mg/g followed by mixing with a magnetic bar for 5 h at 50
°C. The LbDDS were equilibrated overnight at room temperature before use. An aqueous
suspension, designated as SUSP, was prepared by mixing FF in 0.5% HPMC solutions
overnight at room temperature to obtain a FF concentration of 35 mg/mL.
Table 1. Composition of blank LbDDS.
Composition (% w/w)
Formulation
Soybean oil
Maisine 35-1
Kolliphor EL Lipoid S LPC 80
SBO
SBO-MAPC
SNEDDS
100
70
20
-
-
20
20
-
-
60
30
-
30
-
SNEDDS-MAPC
20
30
2.2.2. In vivo pharmacokinetic study
The protocol used for the pharmacokinetic study was approved by the institutional ethics
committee in accordance with Danish law regulating animal experiments and in compliance
with EU directive 2010/63/EU and the NIH guidelines on animal welfare. Male Sprague-
Dawley rats (315-372 g) were purchased from Charles River (Sulzfeld, Germany). Seven
days prior to the experiment, the animals were acclimatized in standard housing conditions
with bedding materials in each cage and maintained on standard feed with free access to

Journal Pre-proof
water. The animals were fasted by removing the food 12 h before the oral dosing of the
formulations. Twenty-five rats in groups of five received one formulation each at the same
dose of 35 mg/kg FF via oral gavage. The animals were allowed free access to water during
the whole experiment and were provided standard feed at 6 h post administration. Blood
samples (200 µL) were collected from the tail vein and transferred into 0.5 mL EDTA-coated
tubes (BD, Franklin Lakes, NJ, USA) at 30 min and 1, 2, 3, 4, 6, 8, 10, and 24 h post
administration. Plasma was taken by immediately centrifuging the blood samples for 15 min
at 2,700 g and 4 °C (Spectrafuge, Labnet, NJ, USA), and stored at -20 °C until analysis. The
animals were euthanized after the study. The bioanalytical procedure to analyse fenofibric
acid level, the active drug released from fenofibrate, was previously described by Tran et al.
[4].
2.2.3. In vitro lipolysis
The experimental setup comprised of two pH-stat apparatuses, two Titrando 842, two 804 Ti
Stands, two 802 stirrers, two glass pH electrodes, two 800 Dosino dosing units, and two 10
mL autoburettes (Metrohm AG, Herisau, Switzerland) (Fig. 2). The amount of formulation
containing 114 mg FF was dispersed for 2 min in 15 mL rat simulated gastric medium (Table
2) in a thermostated vessel maintained at 37 °C using a stirrer. The pH of the dispersion in
gastric medium was manually adjusted to 4.0. After 4 min of gastric lipolysis, the transfer of
the dispersed formulation was initiated into the intestinal vessel containing 15 mL rat
simulated intestinal medium (Table 2) by a peristaltic pump at the rate of 0.93 mL/min.
Porcine pancreatin extract was added to the intestinal medium when the formulation reached
the intestinal vessel. After 10 min of intestinal lipolysis, the pump rate was reduced to 0.47
mL/min to simulate the kinetics of gastric emptying process in rats [23, 24]. The different
applied pump rates allowed a half of the formulation dispersed in the gastric medium to be
transferred into the intestinal vessel after 10 min of intestinal lipolysis, and after 30 min the
rest of the formulation was completely transferred. The pH values of the gastric and intestinal
media were maintained at 4.0 and 6.5 respectively, during the digestion by automatically
adding NaOH to neutralize ionized fatty acid released. At 15, 30, 45, 60, 75, and 90 min of
intestinal lipolysis, 3 mL samples were removed from the intestinal vessel and replaced by 3
mL of intestinal medium (Table 2) to simulate absorption. The withdrawn samples were
subsequently analysed by HPLC as described in section 2.2.4. Bile salt concentration of
gastric and intestinal medium during in vitro lipolysis is described in the Supplementary
Information (Table S1).

Journal Pre-proof
Table 2. Composition of digestion medium used for the in vitro lipolysis model.
Gastric medium
Intestinal medium
15
Initial total volume (mL)
Bile bovine (mM)
Phosphatidylcholine (mM)
Sodium chloride (mM)
Calcium chloride (mM)
Tris maleate (mM)
pH
15
0.08
25.00 [25-27]
2.75 [25, 28]
73.0
0.02
34.2
-
1.40
2.0
2.0
4.0 ± 0.1 [28]
Rhizopus oryzae
12.0 [29]
6.5 ± 0.1 [28]
Porcine pancreatin
22.5 [30]
Lipase source
Lipase activity (TBU/mL)
Fig. 2. In vitro lipolysis model simulating lipid digestion in rats. The formulation dispersed in
simulated gastric medium is transferred to the intestinal medium using a peristaltic pump.
2.2.4. Determination of FF solubilisation during in vitro lipolysis
Two samples of one mL were taken at 0, 15, 30, 60, 75, and 90 min of intestinal digestion
and immediately treated with lipolysis inhibitor (5 µL of 1.0 M 4-BPB (in MeOH) per mL of

Journal Pre-proof
sample). One sample was thoroughly mixed to determine the total amount of FF per mL. The
other sample was centrifuged for 30 min at 19,100 g and 37 °C in a 5417R Eppendorf
centrifuge (Hauppauge, NY, USA) to separate the oil, aqueous and pellet phases to determine
the amount of FF distributed into the pellet and aqueous phases. The amount of FF was
determined using a validated HPLC method previously described by Tran et al. [4].
2.2.5. Synthesis of 123I-FF
123I-FF was synthesised based on the method described by Breyer and co-workers (Fig. 3)
[31]. The stannylation and radioiodination processes were performed as follows.
2.2.5.1. Synthesis of stannylated FF (Bu3Sn-FF)
Bu6Sn2 (0.4 mL, 0.79 mmol) and Pd(PPh3)4 (99 mg) were successively added to a solution of
I-FF (1) (Fig. 3) in anhydrous toluene (182 mg (0.4 mmol) in 6 mL) under nitrogen. The
reaction mixture was heated at 80 °C for 18 h. The resulting mixture was cooled to room
temperature and the dark solid formed was filtered through a silica pad and concentrated
under reduced pressure. The residue was purified twice: (1) flash chromatography on silica
using DCM/MeOH as an eluent in gradient to isolate the target product alongside with some
impurities, (2) radial chromatography on a 1 mm plate using DCM/MeOH in gradient to
1
obtain Bu3S-FF (2) (122 mg, 49%) (Fig. 3). H NMR (400 MHz, CDCl3): δ 7.75 (d, 2H, J
8.9Hz), 7.65 (d, 2H, J 7.8Hz), 7.55 (d, 2H, J 8.1Hz), 6.84 (d, 2H, J 9.0Hz), 5.06 (sept, 1H, J
6.2Hz), 1.63 (s, 6H), 1.52 (m, 6H), 1.32 (m, 6H), 1.18 (d, 6H, J 6.2Hz), 1.07 (t, 6H, J 15.9,
7.6Hz), 0.87 (t, 9H, J 14.8, 7.5Hz). ESI-MS: M/z = 637 [M + Na]+, 616 [M + H]+.
2.2.5.2. Radioiodination process to form 123I-FF
Bu3S-FF (100 µL, 15 mg/mL EtOH) was injected into a Reacti-Vial™ containing Na123I (185
MBq). H2O2 (100 µL, 3% in HCl (1 M)) was added and the vial was mixed using an
Eppendorf shaker at 900 rpm for 40 min. NaHSO3 (150 µL, 0.1 M) was added and mixed to
quench the reaction. The vial was vented with a filtered syringe, decapped and the contents
neutralized with NaHCO3 (150 µL, 0.1 M). The 123I-FF was then extracted into 500 µL of
DCM. The organic phase was transferred into a clean tube and DCM evaporated to leave 123I-
FF (70% labeling efficiency). Quality control was done by TLC against cold I-FF using
hexane:MeOH:ethyl acetate mixture (at a ratio of 85:10:5) as eluent (Rf = 0.45).

Journal Pre-proof
13
1H and C NMR spectra were recorded at 400 and 100 MHz, respectively, on a Bruker AC
400 Ultrashield 10 spectrometer. Chemical shifts are expressed in ppm ( scale). Standard
abbreviations are used to report peak multiplicities. Coupling constants are reported in Hertz
(Hz). Mass spectroscopy was performed on an AB SciexQTrap® 5500 hybrid linear ion-trap
triple quadrupole mass spectrometer equipped with a Turbo Spray source (AB SCIEX,
Concord, ON, Canada). The mass spectrometer was operated in positive or negative
ionization mode and data were acquired using Analyst 1.5.2 software (SCIEX, Framingham,
MA, USA) on a Microsoft Windows XP Professional operating platform.
Fig. 3. Synthesis of radiolabeled 123I-FF. Adapted from Breyer et al. [31].
2.2.6. Synthesis of 125I-triolein
125I-triolein was synthesised in a hood approved for iodination based on the method described
by Lubran et al. [32]. In brief, Na125I (185 MBq in 500 µL NaOH (0.1 N)) was added to 3 mL
ether in a 25 mL glass-stoppered cone shaped flask. KI (1 mL, 1 mg/mL), KIO3 (0.5 mL, 2
mg/mL) and H2SO4 (200 µL, 2 N) were added in this order and the system was closed. The
flask was shaken until all iodine was liberated into the etheric layer as indicated by the yellow
colour. The colourless aqueous phase was then removed. To the organic layer 1 mL of freshly
prepared chlorine dissolved in ether was added. To prepare this chlorine solution, KMnO4 (50
mg) was placed in 5 mL diethyl ether under inert atmosphere and 500 µL of concentrated
HCl was added to the mixture, followed by shaking the sealed system until the organic layer
turned light yellow. The reaction flask was shaken until the turbid yellow mixture turned
clear.

Journal Pre-proof
Triolein (500 µL) was dissolved in ether (1 mL) and mixed into the activated iodinating
solution. The flask was left for 2 h with occasional stirring. The reaction was quenched by
adding 7 mL of iodide thiosulphate solution which was made by mixing equal amounts of a
5% w/v KI and a 5% w/v Na2S2O4 solution. The aqueous phase was removed and the organic
phase washed twice with the same solvent, followed by three washes with water. The ether
was evaporated to leave the 125I-radiolabeled oil (15% labeling efficiency).
2.2.7. Biodistribution study
2.2.7.1. SPECT/CT imaging
2.2.7.1.1. Animal experiment procedure for SPECT/CT imaging
The study was performed in accordance with the Animal Care Committee of the University
of British Columbia under the approved protocol A16-0150 and the NIH guidelines on
animal welfare. Fifteen male Sprague-Dawley rats (Charles River, Canada) were randomly
arranged in five groups of three rats, each group received one formulation via oral gavage.
The fasting and dosing protocol of the biodistribution study was similar to the
pharmacokinetic study (section 2.2.4) except that no blood samples were taken from the
living animal and LbDDS was prepared with radioactive 123I-FF and 125I-triolein. 123I-FF and
125I-triolein were mixed with LbDDS and 123I-FF was mixed with SUSP overnight using an
Eppendorf Thermomixer R at the rate of 900 rpm at room temperature. Between 11.0–48.4
MBq of 123I-FF and 4.3–32.3 MBq of 125I-triolein was administered to each rat.
During the SPECT/CT imaging, the rats were anaesthetized using isoflurane (1–3% for
maintenance, up to 5% for induction) in oxygen from a precision vaporizer and received a
subcutaneous injection of lactated Ringer’s solution (6 mL per hour of imaging) for hydration
prior to the SPECT/CT imaging scan. The SPECT/CT imaging was performed on a
VECTor/CT preclinical small animal scanner (MILabs, Utrecht, The Netherlands).
Respiratory rate and body temperature of the animals were monitored constantly during the
scans while isoflurane dose and bed temperature were accordingly adjusted. All animals
recovered after each scan. The animals were euthanized by CO2 after the final scan at 24 h
post administration before cardiac puncture blood sampling and tissue collection.
2.2.7.1.2. Dual-isotope SPECT/CT parameters and image reconstruction
The whole-body SPECT data from 123I and 125I were simultaneously acquired in list-mode on
a VECTor/CT scanner equipped with an ultra-high resolution rat and mouse (UHRRM)

Journal Pre-proof
multi-pinhole collimator [33]. Dynamic whole-body scans over 1 h (3 frames, 20 min/frame)
were acquired at 30 min and 5 h after the oral administration to obtain details on passage of
the drug and lipids every 20 min. For the results graphed in Figures 6-8, only the first frames
were used of the 30 min and 5 h time points. Static one-hour scans were performed at 10 and
24 h after oral administration of the formulation to provide high quality images when the
activity of 123I decreases. Following each SPECT acquisition, a CT scan was performed to
obtain anatomical information and correct the imaging data for photon attenuation to obtain
fully quantitative images.
The list-mode data were used to generate projection images corresponding to the main
photopeaks of 123I and 125I. For 123I, a photopeak energy window was set at 160 keV (20%
width, 144 keV–176 keV); two scatter windows (10% width) were set on each side of the
photopeak (ranging from 126–144 keV and 176–192 keV for the lower and upper scatter
windows, respectively). The 123I projection data was reconstructed using a pixel-based OSEM
method (16 subsets and 10 iterations) [34] with CT-based attenuation [35] and triple-energy
window (TEW) scatter corrections [36]. The counts in the reconstructed image (referred to as
R160 keV) were converted into units of activity concentration (in MBq/mL) by applying a
phantom-based calibration factor [35, 37].
Since the 125I photopeak at 25 keV overlaps with one of the photopeaks from 123I, a method to
remove the 123I signal from the 125I photopeak was developed to estimate the 125I in vivo
biodistribution using SPECT/CT. The method was divided into three steps. First, the
parameter 푘 that represents the ratio of reconstructed photopeak signals of 123I at 25 keV and
123I at 160 keV was estimated using a uniform phantom filled with 123I and scanned with
VECTor. The parameter 푘 is independent of the attenuation and scattering conditions as it
was determined from quantitative reconstructed images (i.e., images with attenuation and
scatter corrections). Second, an image was reconstructed from the corresponding animal data
using a photopeak centered at 25 keV (100% width, 12.5 – 37.5 keV) with triple-energy
window scatter (scatter windows ranging from 7.5 - 12.5 keV and 37.5 - 42.5 keV for the
lower and upper windows, respectively) and attenuation corrections using POSEM (16
subsets, 10 iterations). The resulting image (referred to as R125+123) corresponds to the
biodistribution image from both the 125I and the 123I X-ray photopeaks. Finally, the
reconstructed 125I biodistribution (R125) was estimated by applying the following expression
to the 푅125+123 and the 푅160 images on a voxel by voxel basis:

Journal Pre-proof
R125 [counts/voxel] = R125+123[counts/voxel] - k × R160 keV [counts/voxel]
(Eq. 1)
The biodistribution of radioactivity was calculated based on SPECT/CT images and
presented as standardized uptake values (SUV) (Eq. 2).
Concentration in volume of interest [MBq/mL]
SUV [g/mL] =
× Body weight [g]
(Eq. 2)
Injected dose [MBq]
The animal body weight and injected dose were taken into account when normalizing SUV in
each organ. This allows a direct comparison of different radioisotope distributions, such as in
our case 123I and 125I, without further calculations. Further details regarding the dual-imaging
with the two iodine radioisotopes were described by Nosrati et al. [38].
2.2.7.2. Biodistribution study using gamma counter measurement after animal euthanasia
After animal euthanasia at 25 h post administration, blood samples were immediately
collected by cardiac puncture and animal organs (stomach, intestine, liver, thyroid, thymus,
and bladder) removed from the body. The stomach content and tissue were separated and all
organ tissues carefully divided. The radioactivity of the samples was measured using a
Sodium-Iodide (NaI) Packard Cobra II auto-gamma counter (Canberra-Packard Canada Ltd,
Mississauga, ON, Canada). The biodistribution of 123I and 125I per organ is presented as
percent of injected dose (%ID) (Eq. 3).
Measured organ activity (based on -counter measurement ) [MBq]
%ID =
× 100
(Eq. 3)
Total activity given to the rat (based on imaging analysis) [MBq]
2.2.8. Data and statistical analysis
The pharmacokinetic profiles of SNEDDS are presented as fenofibric acid concentration in
plasma (µg/mL) vs. time (h). The corresponding area under the curve (AUC) and elimination
half-life were calculated based on non-compartmental analysis using Pharsight Phoenix®
WinNonlin® 6.4 software (Certara, Mountain View, CA, USA). Statistical comparisons of the
data were performed with one-way ANOVA followed by the Tukey mean comparison test
(significant level α = 0.05) using OriginPro 9.1 software (Origin Corporation, Northampton,

Journal Pre-proof
MA, USA). Pharmacokinetic data are presented as the mean ± standard error of the mean
(SEM) (n = 5). All other data are presented as the mean ± standard deviation (SD) (n = 3).
3. RESULTS AND DISCUSSION
3.1. In vivo pharmacokinetic study in rats
The oral absorption of FF from the LbDDS and SUSP (Table 1) were assessed in a
pharmacokinetic study in rats (Fig. 4). The corresponding pharmacokinetic parameters are
summarized in Table 3. The absolute bioavailability of the five formulations ranged from 24
to 35% (Table 3), calculated based on the data obtained after injecting intravenously a 10
mg/mL FF nanoemulsion [39]. Among the five formulations, SUSP exhibited the lowest
AUC0–6 h, AUC0–24 h and Cmax and the longest Tmax (Table 3), corresponding to the slowest
and lowest drug absorption. Among the four LbDDS, the Cmax values of LbDDS containing
surfactants (i.e., SBO-MAPC, SNEDDS, and SNEDDS-MAPC) were significantly higher
than SUSP, while there was no significant difference between SBO and SUSP. Drug plasma
concentration during the first 2 h post administration was ranked as follows: SNEDDS >
SNEDDS-MAPC > SBO-MAPC > SBO > SUSP. The drug plasma concentration for
SNEDDS increased fast before reaching a high Cmax value and then rapidly decreased (Fig.
4). Compared to SNEDDS, the drug plasma concentration of SBO-MAPC and SNEDDS-
MAPC decreased more slowly after reaching Cmax. As a result, SBO-MAPC, SNEDDS-
MAPC, and SNEDDS exhibited significantly higher AUC0–6 h values than SUSP, while only
SBO-MAPC and SNEDDS-MAPC did so for AUC0–24 h (Table 3).
140
120
100
80
60
40
20
0
0
2
4
6
8
10 12 14 16 18 20 22 24
Time (h)
Fig. 4. Plasma concentration of fenofibric acid over 24 h following oral administration of the
five formulations. The data are presented as mean + SEM (n = 5).

Journal Pre-proof
Table 3. Pharmacokinetic parameters of fenofibric acid following oral administration of the
five formulations. The data are presented as mean ± SEM (n = 5).
Formulations
AUC0-6 h(μgh/mL)
SBO
SBO-MAPC SNEDDS
478 ± 58 a 540 ± 58 b
SNEDDS-MAPC
535 ± 21 c
SUSP
347 ± 59
266 ± 13 a,b,c
1025 ± 90a,b
68 ± 7a,b,c
1173 ± 164 1515 ± 53a 1374 ± 114
1524 ± 72b
125 ± 7c
AUC0-24 h(μgh/mL)
Cmax (μg/mL)
Tmax (h)
90 ± 11
124 ± 5a
122 ± 15b
5.8 ± 1.7
5.9 ± 0.3
5.0 ± 0.9
6.2 ± 1.1
5.0 ± 1.3
5.3 ± 0.3
4.4 ± 0.7a
5.1 ± 0.3
9.2 ± 0.5a
5.3 ± 0.3
T1/2 (h)
Absolute
bioavailability *
27 ± 4
35 ± 1
32 ± 3
35 ± 2
24 ± 2
a,b,c Same superscript indicates significant differences in AUC0-6 h, AUC0-24 h Cmax, Tmax or T1/2
,
between the formulations (with p < 0.05)
* Absolute bioavailability was calculated based on the intravenous injection of 10 mg/mL FF
nanoemulsion [39]
3.2. In vitro lipolysis
The dispersion and digestion of the formulations and drug solubilisation in simulated gastro-
intestinal media was studied in the in vitro lipolysis model (Fig. 5) to understand FF oral
absorption from the five formulations. The emulsification of LbDDS upon dispersion was
visually observed and the solubilisation of FF in digested formulations was analysed at
several time points. While most FF was solubilised in either the oil phase or the aqueous
phase during 90 min of digestion of LbDDS (97–98% for SBO, SBO-MAPC, and SNEDDS
and 91% for SNEDDS-MAPC), less than 3% of the drug was solubilised in the case of SUSP.
This explained the enhanced rate and extent of absorption from all LbDDS compared to
SUSP. During dispersion and digestion in the gastric and subsequently the intestinal medium,
SBO formed an unstable coarse emulsion of which two phases rapidly separated when
stopping agitation. LbDDS containing surfactants, i.e., SBO-MAPC, SNEDDS, and
SNEDDS-MAPC, formed more stable colloidal systems in both gastric and intestinal media
than SBO, which related to faster FF absorption from LbDDS containing surfactants
compared to SBO.
During the in vitro intestinal digestion, the distribution of FF into the oil and aqueous phases
are different between the four LbDDS (Fig. 4). FF concentration in the aqueous phase during
90 min of in vitro lipolysis can be ranked as: SNEDDS > SNEDDS-MAPC > SBO-MAPC ≈
SBO > SUSP (Fig. 5), similar as the ranking for drug plasma concentration during the first 2h

Journal Pre-proof
post administration (Fig. 4). Although similar amount of FF distributed into the oil phase of
SBO-MAPC and SBO, the oil phase of SBO-MAPC emulsified much better than SBO in the
simulated digestion media. Thus, the in vivo absorption rate of the formulations could be
related to not only drug solubilisation but also emulsification capacity of the LbDDS.
Similar trends were found between the in vitro lipolysis and the pharmacokinetic study
explaining the difference of Cmax, Tmax, and AUC values between formulations. With poor
emulsification capacity and the high fraction of drug distributing into the oil phase, SBO
exhibited low Cmax and AUC0-24 h values, or a low, but prolonged, oral absorption in rats (Fig.
4) compared to the other formulations. Incorporating MAPC into SBO enhanced the
emulsification and drug concentration in the aqueous phase during dispersion and digestion,
so that higher AUC and Cmax were obtained corresponding to enhanced rate and extent of
absorption. The presence of MAPC in SBO-MAPC and SNEDDS-MAPC allowed for a
balance between drug distributed into aqueous and oil phases, which were well-emulsified, to
obtain high drug plasma concentration, which increased faster than from SBO and more
prolonged than from SNEDDS, and therefore resulted in high AUC0-24 h value.
Fig. 5. Distribution of FF into the oil, aqueous and pellet phases in the intestinal phase during
90 min of in vitro lipolysis simulating rat conditions of (A) SBO, (B) SBO-MAPC, (C)
SNEDDS, (D) SNEDDS-MAPC and (E) SUSP. The data are presented as mean + SD (n = 3).

Journal Pre-proof
3.3. Biodistribution study
3.3.1. SPECT/CT imaging
In vitro lipolysis explained the differences in terms of rate and extent of oral drug absorption
when comparing formulations, but cannot explain the low absolute bioavailability of all
formulations. The biodistribution of drug and lipids by using SPECT/CT imaging, as a non-
invasive approach, was investigated with this objective. 123I-FF and 125I-triolein were used to
label the drug and the lipid excipients, respectively. The biodistribution of drug and
excipients can be obtained independently in the same animal during 24 h post administration,
and thereby provide information on their passage through the gastro-intestinal tract and after
being absorbed. New methods of radiolabelling fenofibrate and lipid excipients were
investigated and acceptable yields of the iodination with 123I (approximately 70% labelling
efficiency for 123I-FF) and 125I (approximately 15% labelling efficiency for 125I-triolein) were
obtained. The high stability of radiolabeled 123I-FF was confirmed in human serum by Breyer
et al. [31]. 123I-FF was shown to convert completely to radiolabeled fenofibric acid in
homogenized mouse liver [31]. Both radioisotopes could be detected and quantified at
acceptable statistics during 24 h after oral administration to rats.
Similarly to the FF metabolism [40, 41], the ester bond of 123I-FF is expected to be
hydrolysed to 123I-fenofibric acid, which is metabolised in the liver to form 123I-labeled
glucuronide metabolites. A biodistribution study of 123I-FF in mice after intravenous injection
showed that most of the drug was accumulated in the liver, the target organ of FF or
fenofibric acid, between 5 min and 6 h post administration [31]. Based on the metabolism of
triolein [42], 125I-triolein is suggested to form 125I-oleic acid and 125I-monooleate during the
digestion, which both are absorbed into the enterocytes and delivered to the lymph and
systemic blood as triglycerides in 125I-chylomicrons. Fig. 6 shows SBO containing 123I-
FF/125I-triolein as representative dynamic SPECT/CT images for the five formulations at 30
min and 5, 9.5 and 24 h post administration. The biodistribution of 123I and 125I indicates the
biodistribution of the drug and lipids following oral administration.

Journal Pre-proof
Fig. 6. Three-dimensional (3D) volume-rendered SPECT/CT images of radiolabeled drug (in
red) and radiolabeled lipids (in blue) distributed in rats at 30 min, 5 h, 9.5 h, and 24 h after
oral administration of SBO containing 123I-FF/125I-triolein. Regions where both radiolabeled
drug and lipids reside are in pink. The SBO system is shown as a representative example of
the investigated LbDDS, as activity distributions (see Figures 7 and 8) were not statistically
different. 3D SPECT/CT images taken after oral administration of other formulations are
mentioned in Supplementary Information (Figures S1-S4).
Based on the SPECT/CT image analysis, the movement of the radiolabelled drug and lipid
excipient to and through the different organs could be followed over 24 h and is
quantitatively shown in Fig. 7.

Journal Pre-proof
Fig. 7. Standardized uptake value (SUV) of (left) radiolabeled drug and (right) radiolabeled
lipids at 30 min to 24 h after oral administration in rats in stomach, intestine, heart, and liver
based on SPECT/CT images. The data are presented as mean + SD (n = 2–3).
SPECT/CT images show that both radiolabeled drug and lipids were found in the stomach
and the intestine at 30 min after oral administration (Fig. 6), and then slowly moved

Journal Pre-proof
through the gastro-intestinal tract. The activities of 123I and 125I in the stomach, intestine,
heart and liver were calculated from the SPECT/CT images and the previously performed
phantom studies and calibration curves [38, 43] (Fig. 7).
Given in percentage of the administered dose, 56–71% of radiolabelled drug and 54–83%
of radiolabelled lipids were found in the stomach 30 min after oral gavage, while between
4–9% of radiolabeled drug and 5–10% of radiolabeled lipid were found in the entire
intestine (Fig. 8). As shown in Fig. 7, the amount of radioactivity from both isotopes in the
stomach decreased over time with a faster gastric clearance rate during the first 9.5 h and
slower clearance from 9.5–24 h post administration. At 24 h, significant radioactivity from
both isotopes remained in the stomach, i.e., 28–59% of radiolabeled drug and 24–60% of
radiolabeled lipid (Fig. 8). 5–9% of radiolabeled drug and 3% of radiolabeled lipid, was
found in the intestine at the 24 h time point. The AUC values of SUV of radiolabeled drug
and lipid over 24 h in the stomach and entire intestine are similar for the five formulations
(Fig. 7).
Fig. 8. Percentage of (A) radiolabeled drug and (B) radiolabeled lipids compared to
injected dose (ID) in stomach and intestine at 30 min and 24 h after oral administration
based on SPECT/CT image analysis. The data are presented as mean + SD (n = 2–3).
The direct equivalent of measuring blood drug concentrations in conventional
pharmacokinetic studies is to choose the heart as the organ of interest in SPECT/CT
studies and follow its radioactivity intensity. The SUV profile reflects mostly the heart’s
blood volume and allows for easy concentration calculations. In this study, both the 123I

Journal Pre-proof
and the 125I showed a maximum SUV in the heart at 5 h post administration (Fig. 7), which
correlates well with the Tmax in the pharmacokinetic study (Fig. 5). SUSP and SBO
exhibited the lowest uptake of radiolabeled drug in the heart in the imaging study,
similarly to their lowest mean drug plasma concentration versus time profile in the
pharmacokinetic study. For SNEDDS, the high level found in the heart correlates with the
high uptake in the liver, both of which may relate to the fast drug absorption seen in the
pharmacokinetic study. In the pharmacokinetic study, SNEDDS did not perform
significantly better than SBO and SUSP; but the drug uptake profiles in the heart showed
that only SNEDDS exhibited significantly higher AUC compared to SBO and SUSP (Fig.
7). The difference between the imaging and the pharmacokinetic study is likely due to the
lower number of time points during the imaging. Specifically, only four time points were
allowed during 24 h because the duration of total anaesthesia was limited to ensure animal
well-being. The activities could thus not be measured between 3-6 h, which did not allow
to demonstrate and confirm the prolonged high drug plasma concentrations from SBO-
MAPC and SNEDDS-MAPC in the pharmacokinetic study.
Interestingly, the SUV versus time profile of radiolabeled drug and lipid found in the heart
were similar to the SUV profile in the intestine. The absorption of the two radiolabeled
compounds seemed proportional to the amounts present in the intestine, which is only the
case if drug and lipid amounts were lower than their maximum absorption capacity in the
intestine. This might indeed have been the case since fenofibrate is known to be absorbed
fast and well [12].
Regarding the biodistribution of lipid excipients (Fig. 7 and 8), more lipid was found in
the stomach and the intestine for SBO compared to other the LbDDS containing
surfactants and lower lipid levels. A slower gastric emptying rate can be expected from
SBO, as lipids are emptied more slowly from a poorly emulsified lipid phase than from a
fine oil-in-water emulsion [44]. The lipid SUV in the liver was also higher for LbDDS
containing higher triglyceride levels (i.e., SBO and SBO-MAPC) than LbDDS containing
lower triglyceride levels (i.e., SNEDDS and SNEDDS-MAPC).
There was a trend that drug uptake in the heart and the liver (Fig. 7) was higher for the
SNEDDS as compared to the SUSP and SBO, however, this was not statistically
significant. Further, in all 5 formulations, there was no correlation between the distribution

Journal Pre-proof
of the drug and the lipid, indicating that they are not co-distributed in the body. One
possible explanation might be the high variability between animals, which could be due to
the low number of animals used and the inconsistent amount of drug measured in the
stomach. For future studies, the inter-subject variability should be reduced by increasing
the number of time points and/or animals used. To reduce variations related to gastric
emptying, the amount of food and bedding materials in the rat stomach should be reduced
by adjusting feeding and housing strategies (e.g., changing food type, circadian rhythm,
bedding materials or housing in wire-bottom cages).
3.3.2. Final organ activity measurement and biodistribution
After the final 24 h SPECT/CT imaging time point, the animals were sacrificed and the
radioactivity in relevant organs were measured in a calibrated gamma counter to confirm
the quantitative imaging data and to provide supplementary information from low activity
organs. Organ activity concentrations and organ activities were then calculated the %ID/g
and the %ID/organ for different organs, urine and feces (Fig. 9). At 25 h after oral gavage,
32–73% of radiolabeled drug and 16–34% of radiolabeled lipid were found in the stomach
content. At the same time point, 0.5–1.3% of radiolabeled drug and 0.4–2.2% of
radiolabeled lipid were found in the stomach tissue, and 2–4% of radiolabeled drug and
0.1–0.4% of radiolabeled lipid were found in the whole small intestine (both content and
tissue). No significant differences were observed between the five formulations for both
123I-FF and 125I-triolein. Both gamma counter biodistribution analysis and SPECT/CT
imaging allow the conclusion that a large amount of drug and lipid excipients remained in
the stomach after 24 h post administration. Small differences between the two methods can
be attributed to the inhomogeneous distribution of radioactivity in the whole organ.
Although the organ tissues and stomach content were thoroughly homogenized, errors
might come especially from the small sample volumes used for the gamma counter
measurements.

Journal Pre-proof
Fig. 9. Organ distribution of (A) radiolabeled drug and (B) radiolabeled lipids measured in
the gamma counter 25 h after oral administration and expressed as a percentage of the
injected dose. The data are presented as the mean + SD (n = 2–3)
The significant amount of radiolabelled drug found in the gastro-intestinal tract after 24 h
can be related to the low absolute bioavailability found in the pharmacokinetic study (i.e.,
24 to 35%). This can be either a slow gastric emptying observed in the biodistribution
study due to the constantly abundant presence of substances in the rat stomach or the re-
ingestion of excreted formulation. In the literature, food was reported to be found in rat
stomach after overnight fasting [45]. In addition, during fasting (12 h prior to
administration and 6 h post administration) for both the pharmacokinetic and
biodistribution studies, the animals were kept in cages with bedding materials mixed with
their faeces, which might contain excreted formulation. Hungry animals seem to have a
tendency to fill their stomach with these materials, which we found at the 24 h time point
during biodistribution in their stomach. Since most radiolabelled drug and lipid was still
inside the stomach at that time, it appears that drug and lipid stayed with the bedding
materials for a long time.
Besides the stomach and the intestine, radiolabeled drug and lipid were mainly found in
the liver, thyroid, blood, faeces and urine by the gamma counter measurements. The
distribution of 123I and 125I in the thyroid increased over time indicating that free iodine
was released from radiolabelled drug and lipid. Despite a visible radioactivity
accumulation in the thyroid (Fig. 6), which increased over the 24 h study period, the
maximum radioactivity in the thyroid, i.e., at 24 h post administration, contributed less

Journal Pre-proof
than 2% of the total administered radioactivity dose as detected from gamma counter
measurements (Fig. 9).
4. CONCLUSION
Results from both two-step in vitro lipolysis and dual isotope SPECT/CT imaging studies
correlated well with the pharmacokinetic study. Combining these techniques provided a
complete explanation of the fate of drug and co-administered excipients with important and
precise details on drug solubilisation and biodistribution after oral administration from
different formulations. While the aqueous suspension, SUSP, led to the slowest and lowest
drug absorption due to low drug solubilisation, SNEDDS with 60% Kolliphor exhibited the
fastest absorption of FF and in accordance also displayed a high drug concentration in the
aqueous phase during in vitro lipolysis. SNEDDS and SBO containing MAPC exhibited
enhanced oral bioavailability with prolonged drug absorption profile, which was explained by
the enhanced emulsification with drug solubilised in both aqueous and oil phases during lipid
digestion. To explain the limited oral bioavailability of LbDDS (less than 34%) despite
absorption enhancement, SPECT/CT imaging was used and shown that significant amounts
of drug and lipid formulation were found in the animals’ stomach even 24 h post
administration. The study is a proof-of-concept for the use of SPECT/CT imaging in drug and
formulation absorption. A study where the animals are conditioned to become diurnal and/or
housed in wire-bottom cages to obtain an empty stomach before administration and avoid re-
ingestion of formulation should be performed to reduce artefacts.
ACKNOWLEDGMENT
Financial support from the University of Copenhagen, the Phospholipid Research Center
(Heidelberg, Germany), the Lundbeck Foundation, and the Canadian Foundation for
Innovation (CFI, project number 25413) are kindly acknowledged. We are also grateful for
the support of technicians and veterinarians at the Department of Pharmacy, University of
Copenhagen, Denmark and the Centre for Comparative Medicine, University of British
Columbia, Canada.

Journal Pre-proof
REFERENCES
[1] S. Stegemann, F. Leveiller, D. Franchi, H. De Jong, H. Lindén, When poor solubility
becomes an issue: From early stage to proof of concept, Eur J Pharm Sci, 31 (2007) 249-
261.
[2] A. Müllertz, A. Ogbonna, S. Ren, T. Rades, New perspectives on lipid and surfactant
based drug delivery systems for oral delivery of poorly soluble drugs, J Pharm
Pharmacol, 62 (2010) 1622-1636.
[3] T. Tran, X. Xi, T. Rades, A. Müllertz, Formulation and characterization of self-
nanoemulsifying drug delivery systems containing monoacyl phosphatidylcholine, Int J
Pharm, 502 (2016) 151-160.
[4] T. Tran, S.D.V.S. Siqueira, H. Amenitsch, A. Müllertz, T. Rades, In vitro and in vivo
performance of monoacyl phospholipid-based self-emulsifying drug delivery systems, J
Control Release, 255 (2017) 45-53.
[5] T. Tran, S. Siqueira, H. Amenitsch, T. Rades, A. Mullertz, Monoacyl phosphatidylcholine
inhibits the formation of lipid multilamellar structures during in vitro lipolysis of self-
emulsifying drug delivery systems, Eur J Pharm Sci, 108 (2017) 62-70.
[6] A.T. Larsen, P. Sassene, A. Müllertz, In vitro lipolysis models as a tool for the
characterization of oral lipid and surfactant based drug delivery systems, Int J Pharm,
417 (2011) 245-255.
[7] J.Ø. Christensen, K. Schultz, B. Mollgaard, H.G. Kristensen, A. Mullertz, Solubilisation
of poorly water-soluble drugs during in vitro lipolysis of medium-and long-chain
triacylglycerols, Eur J Pharm Sci, 23 (2004) 287-296.
[8] A.T. Larsen, A.G. Ohlsson, B. Polentarutti, R.A. Barker, A.R. Phillips, R. Abu-Rmaileh,
P.A. Dickinson, B. Abrahamsson, J. Ostergaard, A. Müllertz, Oral bioavailability of
cinnarizine in dogs: Relation to SNEDDS droplet size, drug solubility and in vitro
precipitation, Eur J Pharm Sci, 48 (2013) 339-350.
[9] A. Larsen, R. Holm, M.L. Pedersen, A. Mullertz, Lipid-based formulations for danazol
containing a digestible surfactant, Labrafil M2125CS: in vivo bioavailability and
dynamic in vitro lipolysis, Pharm Res, 25 (2008) 2769-2777.
[10] N. Thomas, K. Richter, T.B. Pedersen, R. Holm, A. Müllertz, T. Rades, In vitro lipolysis
data does not adequately predict the in vivo performance of lipid-based drug delivery
systems containing fenofibrate, AAPS J, 16 (2014) 539-549.

Journal Pre-proof
[11] T.T. Do, M. Van Speybroeck, R. Mols, P. Annaert, J. Martens, J. Van Humbeeck, J.
Vermant, P. Augustijns, G. Van den Mooter, The conflict between in vitro release studies
in human biorelevant media and the in vivo exposure in rats of the lipophilic compound
fenofibrate, Int J Pharm, 414 (2011) 118-124.
[12] M.F. Crum, N.L. Trevaskis, H.D. Williams, C.W. Pouton, C.J.H. Porter, A new in vitro
lipid digestion – in vivo absorption model to evaluate the mechanisms of drug absorption
from lipid-based formulations, Pharm Res, 33 (2015) 970-982.
[13] M.H. Michaelsen, K.M. Wasan, O. Sivak, A. Müllertz, T. Rades, The effect of digestion
and drug load on halofantrine absorption from self-nanoemulsifying drug delivery
system (SNEDDS), AAPS J, 18 (2016) 180-186.
[14] R. Berthelsen, R. Holm, J. Jacobsen, J. Kristensen, B. Abrahamsson, A. Mꢀllertz,
Kolliphor surfactants affect solubilization and bioavailability of fenofibrate. studies of in
vitro digestion and absorption in rats, Mol Pharm, 12 (2015) 1062-1071.
[15] W. Weitschies, C.G. Wilson, In vivo imaging of drug delivery systems in the
gastrointestinal tract, Int J Pharm, 417 (2011) 216-226.
[16] C.M. Gomes, A.J. Abrunhosa, P. Ramos, E.K. Pauwels, Molecular imaging with SPECT
as a tool for drug development, Adv Drug Deliv Rev, 63 (2011) 547-554.
[17] R. Misri, K. Saatchi, U.O. Häfeli, Nanoprobes for hybrid SPECT/MR molecular
imaging, Nanomedicine (Lond), 7 (2012) 719-733.
[18] K. Saatchi, P. Soema, N. Gelder, R. Misri, K. McPhee, J.H. Baker, S.A. Reinsberg, D.E.
Brooks, U.O. Häfeli, Hyperbranched polyglycerols as trimodal imaging agents: design,
biocompatibility, and tumor uptake, Bioconjug Chem, 23 (2012) 372-381.
[19] K. Saatchi, U.O. Häfeli, Radiolabeling of biodegradable polymeric microspheres with
[
99mTC(CO)3]+ and in vivo biodistribution evaluation using MicroSPECT/CT imaging,
Bioconjug Chem, 20 (2009) 1209-1217.
[20] R. Misri, K. Saatchi, S.S. Ng, U. Kumar, U.O. Häfeli, Evaluation of 111In labeled
antibodies for SPECT imaging of mesothelin expressing tumors, Nucl Med Biol, 38
(2011) 885-896.
[21] H.D. Williams, P. Sassene, K. Kleberg, J.C. Bakala‐N'Goma, M. Calderone, V. Jannin,
A. Igonin, A. Partheil, D. Marchaud, E. Jule, Toward the establishment of standardized
in vitro tests for lipid‐based formulations, part 1: method parameterization and
comparison of in vitro digestion profiles across a range of representative formulations, J
Pharm Sci, 101 (2012) 3360-3380.

Journal Pre-proof
[22] M.C. Michalski, C. Genot, C. Gayet, C. Lopez, F. Fine, F. Joffre, J.-L. Vendeuvre, J.
Bouvier, J.-M. Chardigny, K. Raynal-Ljutovac, Multiscale structures of lipids in foods as
parameters affecting fatty acid bioavailability and lipid metabolism, Progress in lipid
research, 52 (2013) 354-373.
[23] D.N. Lorenz, Gastric emptying of milk in rat pups, Am J Physiol, 248 (1985) R732-738.
[24] J. Jordi, F. Verrey, T.A. Lutz, Simultaneous assessment of gastric emptying and
secretion in rats by a novel computed tomography-based method, Am J Physiol, 306
(2014) G173-G182.
[25] R. Holm, A. Müllertz, H. Mu, Bile salts and their importance for drug absorption, Int J
Pharm, 453 (2013) 44-55.
[26] R. Coleman, S. Iqbal, P.P. Godfrey, D. Billington, Membranes and bile formation.
Composition of several mammalian biles and their membrane-damaging properties,
Biochem J, 178 (1979) 201-208.
[27] M. Hagio, M. Matsumoto, M. Fukushima, H. Hara, S. Ishizuka, Improved analysis of
bile acids in tissues and intestinal contents of rats using LC/ESI-MS, J Lipid Res, 50
(2009) 173-180.
[28] T.T. Kararli, Comparison of the gastrointestinal anatomy, physiology, and biochemistry
of humans and commonly used laboratory animals, Biopharm Drug Dispos 16 (1995)
351-380.
[29] P. Sassene, M. Michaelsen, M. Mosgaard, M. Jensen, E. Van Den Broek, K. Wasan, H.
Mu, T. Rades, A. Mꢀllertz, In vivo precipitation of poorly soluble drugs from lipid-based
drug delivery systems, Mol Pharm, 13 (2016) 3417-3426.
[30] M.U. Anby, T.-H. Nguyen, Y.Y. Yeap, O.M. Feeney, H.D. Williams, H. Benameur,
C.W. Pouton, C.J. Porter, An in vitro digestion test that reflects rat intestinal conditions
to probe the importance of formulation digestion vs first pass metabolism in danazol
bioavailability from lipid based formulations, Mol Pharm, 11 (2014) 4069-4083.
[31] S. Breyer, A. Semmler, T. Miller, A. Hill, S. Geissler, U. Haberkorn, W. Mier,
Radioiodinated dechloro-4-iodofenofibrate: a hydrophobic model drug for molecular
imaging studies, Int J Pharm, 431 (2012) 78-83.
[32] M. Lubran, J.D. Pearson, A screening test for steatorrhoea using 131I-labelled triolein, J
Clin Pathol, 11 (1958) 165-169.
[33] F. van der Have, B. Vastenhouw, R.M. Ramakers, W. Branderhorst, J.O. Krah, C. Ji,
S.G. Staelens, F.J. Beekman, U-SPECT-II: An ultra-high-resolution device for molecular
small-animal imaging, J Nucl Med, 50 (2009) 599-605.

Journal Pre-proof
[34] W. Branderhorst, B. Vastenhouw, F.J. Beekman, Pixel-based subsets for rapid multi-
pinhole SPECT reconstruction, Phys Med Biol, 55 (2010) 2023-2034.
[35] C. Wu, J.R. de Jong, H.A. Gratama van Andel, F. van der Have, B. Vastenhouw, P.
Laverman, O.C. Boerman, R.A. Dierckx, F.J. Beekman, Quantitative multi-pinhole
small-animal SPECT: Uniform versus non-uniform Chang attenuation correction, Phys
Med Biol, 56 (2011) N183-193.
[36] K. Ogawa, Y. Harata, T. Ichihara, A. Kubo, S. Hashimoto, A practical method for
position-dependent compton-scatter correction in single photon emission CT, IEEE
Trans Med Imaging, 10 (1991) 408-412.
[37] C. Wu, F. van der Have, B. Vastenhouw, R.A.J.O. Dierckx, A.M.J. Paans, F.J. Beekman,
Absolute quantitative total-body small-animal SPECT with focusing pinholes, Eur J Nucl
Med Mol Imaging, 37 (2010) 2127-2135.
[38] Z.P. Nosrati, P. Esquinas, C. Rodriguez-Rodriguez, T. Tran, T. Esposito, K. Saatchi,
U.O. Häfeli, Simultaneous SPECT imaging of I-125 and I-123, a practical approach to
assessing a drug and its carrier at the same time with preclinical dual imaging, (In
preparation).
[39] N. Borkar, Oral delivery of fenofibrate-loaded lipid matrix particles, Master Thesis,
Department of Pharmacy, University of Copenhagen, Copenhagen, Denmark, 2012.
[40] D.B. Miller, J.D. Spence, Clinical pharmacokinetics of fibric acid derivatives (fibrates),
Clin Pharmacokinet, 34 (1998) 155-162.
[41] M.J. Chapman, Pharmacology of fenofibrate, Am J Med, 83 (1987) 21-25.
[42] C.M. Lo, P. Tso, 4 - Physicochemical basis of the digestion and absorption of
triacylglycerol, in: Designing Functional Foods, Woodhead Publishing, 2009, pp. 94-
125.
[43] S. Lee, J. Gregor, S.J. Kennel, D.R. Osborne, J. Wall, GATE Validation of standard dual
energy corrections in small animal SPECT-CT, PLOS ONE, 10 (2015) e0122780.
[44] M.I. Friedman, I. Ramirez, M.G. Tordoff, Gastric emptying of ingested fat emulsion in
rats: implications for studies of fat-induced satiety, Am J Physiol, 270 (1996) R688-692.
[45] P. Jeffrey, M. Burrows, A. Bye, Does the rat have an empty stomach after an overnight
fast?, Lab Anim, 21 (1987) 330-334.
Graphical abstract: