In vitro performance prediction of LBFs using
lipolysis-permeation assay with the LiDo artificial membrane:
Validation and protocol optimization
By Alaa Zangana
Degree Project in Pharmaceutics and Biopharmaceutics, A9, Fall Semester 2020
Supervisors: Oliver Hedge, Christel Bergström Examiner: Per Artursson
Drug Delivery Group Department of Pharmacy
Abstract
Lipid-based formulations (LBFs) can be an effective formulation strategy to improve the bioavailability of lipophilic drugs. However, conventional single-compartment in vitro lipolysis assays often underestimate LBF performance in vivo due to the absence of absorption sink. Recently, a two-compartment setup separated by an absorptive membrane provided good in
vivo in vitro relationship (IVIVR). The goal of this study was therefore to investigate the use of
Table of Contents
1. INTRODUCTION _______________________________________________________ 3 2. MATERIALS AND METHODS ___________________________________________ 5 2.1 Materials ____________________________________________________________ 5 2.2 Preparation of “LiDo” artificial membrane __________________________________ 6 2.3 Formulation __________________________________________________________ 7 2.4 Transport studies ______________________________________________________ 7 2.5 In vitro Lipolysis Assay ________________________________________________ 8
2.5.1 In vitro lipolysis in full-scale setup ___________________________________ 8
2.5.2 In vitro lipolysis in small-scale setup _________________________________ 10 2.6 LiDo integrity during in vitro lipolysis assays ______________________________ 11 2.7 Sample processing and HPLC analysis ____________________________________ 11 2.8 Statistical analysis ____________________________________________________ 12 3. RESULTS ____________________________________________________________ 12 3.1 Transport study ______________________________________________________ 12 3.2 In vitro lipolysis-permeation assay _______________________________________ 12
3.2.1 In vitro lipolysis-permeation in full-scale setup _________________________ 12
1. Introduction
Lipid-based formulation (LBF) is an innovative strategy that can significantly enhance the oral bioavailability of lipophilic compounds. LBFs are isotropic mixtures of oils, surfactants and organic solvents in which the low soluble drug is pre-dissolved [1-4]. The major benefits associated with LBF are avoidance of rate-limiting dissolution by presenting the drug to the absorptive site in a solubilized form. Also, it increases the solubilization capacity of the GI fluids. The presence of exogenous lipid excipients stimulates pancreatic and gallbladder secretion that initiate the digestion process (lipolysis). The lipid digestion products generated are subsequently solubilized by bile salt–phospholipid–cholesterol-mixed micelles. This in turn leads to development of different lipid colloidal phases that change the intraluminal solvation capacity for poorly water-soluble drugs (PWSDs) [5]. Consequently, lipid digestion can in many cases generate supersaturation, promoting absorption or precipitation of the API in transit through the intestine [6,7].
is more suited for routine screening compared to cell-based membrane [11, 12]. However, the conditions in the donor compartment are important for the integrity of the membrane as well as digestion and absorption process. This is the basis to set up an in vitro-in vivo correlation (IVIVC). Previous lipolysis-permeation studies addressed this issue. For instance, Caco-2 cell monolayers were not compatible with porcine pancreatin used as a digestive agent [13]. Conversely, Novozym 435 (immobilized lipase) has proven compatible with Caco-2 cells leading to good agreement with in vivo data [9, 13, 14]. Not long ago, an in vitro lipolysis-permeation assay was published in which the performance of artificial membranes was examined with regards to different types of LBF digested by porcine pancreatin. The artificial membranes, hexadecane applied to polycarbonate as well as PVDF filters were not compatible during lipolysis process using porcine pancreatin (780 TBU/ml) as digestive agent. Polycarbonate treated with “LiDo” (20% soy lecithin in n-dodecane) was not compatible either. In contrast, the PVDF filter covered by “LiDo” solution could withstand porcine pancreatin for approximately 30 min. It seems promising at least for compounds predominantly absorbed by passive lipoidal diffusion, but further studies are required to assess the physical changes of membrane structure using other digestive enzymes [15].
observed that LiDo membrane is not fully compatible with all pH electrodes, possibly resulting in measurement error (unpublished data).
Most currently employed lipolysis-permeation system have a relatively small permeation area in relation to the volume of the donor media. This can result in a limited permeated fraction of drug [19]. On the contrary, in humans the area-to-volume (A/VD) ratio is estimated to be ~ 2
cm-1 [20]. Therefore, the (A/VD) ratio of the lipolysis-permeation setup can be increased in an
effort to improve the biorelevance of in vitro system. However, the complexity of the intestinal environment and LBF performance in vivo make it challenging to design reliable in vitro method that allows accurate prediction of LBF’s behavior. Thus, the aim of this thesis was to validate the use of the artificial” LiDo” in two two-compartment assay systems (full-scale vs. small-scale setup) performance prediction of oral lipid-based formulations (LBFs), with regard to digestive enzymes, donor volume and different pH-control methods. Novozym 435, porcine lipase, and porcine pancreatin was used as digestive agents. The poorly water-soluble drug carvedilol (MW 406.5g/mol, logP 4.1, pKa 8.2, Tm 114.1°C) was selected as the weakly basic
model compound [21]. Carvedilol was either predissolved or co-administered with varying LBF dose. The currently selected formulation was previously examined in vivo as well as in vitro using two-compartment system separated by Caco-2 cell monolayers [13]. These data were used to evaluate the in vivo relevance of the present study system.
2. Materials and Methods
2.1 Materialsalcohol) was obtained from Solveco (Sweden). Acetonitrile (³ 99.9%), dimethyl sulfoxide (DMSO ³ 99.9%), and Lucifer yellow were purchased from VWR (Sweden); while Kolliphor EL (macrogolglycerol ricinolate) and D-a-Tocopherol polyethylene glycol succinate (TPGS), soybean oil, porcine lipase (type II), porcine pancreatin (8´USP specifications), sodium chloride (NaCl), sodium hydroxide (NaOH), Bis-Tris methane, calcium chloride (CaCl2),
sodium acetate, and bovine serum albumin (BSA) were bought from Merck (Germany). Maisine CC was kindly donated by Gattefossé (France). Carvedilol was obtained from Molekula (UK). Novozym 435 (immobilized lipase) was obtained from Stream chemicals (France); and fasted state simulated intestinal fluid (FaSSIF) powder was obtained from biorelevant.com (UK). Water was purified using a Milli-Q lab water purification system (Merck, Germany).
2.2 Preparation of “LiDo” artificial membrane
Lecithin based membrane was prepared similar to a previously described study [15]. In short, Avanti’s 20% Soy PC extract was dissolved in n-dodecane supplemented with 1.5% (v/v) absolute ethanol to obtain a final concentration of 20% (w/v) lecithin. After full dissolution of lecithin extract (~3h), the solution was centrifuged (3,220 g, at 20°C for 20 minutes) to separate undissolved material. The supernatant was collected, aliquoted for single use and stored then under nitrogen at -20°C. The lipid solution was thawed at room temperature one day before experiment. Then, the permeation membrane was formed by covering polyvinylidene fluoride (PVDF) filter supports (0.45 μm pore size, porosity 0.7, thickness 100–145 μm, Immobilon-P, Merck Millipore) with 16.23 μl of LiDo solution per cm2 of filter. LiDo membrane was used
2.3 Formulation
Lipid-based formulation was prepared as a previous published study [13]. Briefly, soybean oil (triglyceride), Maisine CC (monoglyceride) and Kolliphor EL (surfactant) were pre-heated (37°C) and weighed into glass vials in the predefined fraction (0.325:0.325:0.35 w/w). Subsequently, the vials were sealed, vortex mixed and placed on a shaking plate (200 rpm) in an incubator (37°C) overnight. Prior to any experiments (24h), carvedilol was incorporated to the LBF by weighing the required amounts of LBF and carvedilol into a glass vial, vortexing the mix and placing it on shaker, at 37°C. The drug loading capacity corresponded to 66% of equilibrium solubility, at 37°C for LBFs type IIIA-LC [13]. In this study, Maisine CC was used instead of Maisine 35-1 which was used previously, since the latter is no longer commercially available. According to the manufacturer, there should be no major difference between them except that Maisine CC is more purified, leading to a lower melting point. It should be noted that carvedilol solubility could therefore differ slightly from the previously measured value.
2.4 Transport studies
Figure 1. Schematic representation of the in vitro lipolysis-permeation setups: (A) ENA (full-scale setup), (B)
PALIS (small-scale setup).
Table 1. Description and designation of lipid-based formulation (LBF) experimental conditions, physical state of
the active pharmaceutical ingredient (API) carvedilol, and relative concentrations of LBF and API to digestion medium during each experiment.
Designation Physical state of API API conc. (mg/ml) LBF conc. (mg/ml)
F1 Pre-dissolved in LBF (25 mg/g) 0.625 25
F2 Co-administered with LBF 0.625 25
F3 Co-administered with LBF 0.625 12.5
2.5 In vitro Lipolysis Assay
2.5.1 In vitro lipolysis in full-scale setup
prepared by dissolving FaSSIF powder (2.24 mg/ml, resulting in 3.0 mM taurocholate, 0.75 mM lecithin) in either: (i) high strength lipolysis buffer (pH 6.5, 200 mM Bis-Tris, and 1.4 mM CaCl2) [25], or: (ii) normal lipolysis buffer (pH 6.5, 2 mM Bis-Tris, 1.4 mM CaCl2, and 150
(mg/ml) vs. time (min), while mass transfer to the bottom chamber was evaluated using AUC for mass flux (µg/cm2) vs. time (min). Trapezoidal rule method was used to calculate AUC.
Table 2. Assay systems and donor conditions was used in the in vitro lipolysis
Assay system VD
(cm3) A/V(cm-1D ) Digestive media pH-control Digestive agent(s)
ENA 20 1.46 FaSSIF100x - N435 PL PP
ENAnd 100 0.29 FaSSIF
1x pH-stat N435 PL
PALIS 2.15 3.31 FaSSIF100x - N435 PL PP
Abbreviations: VD (donor volume), A (membrane area), FaSSIF100x (FaSSIF in 200 mM lipolysis buffer, 100x
strength), FaSSIF1x (FaSSIF in 200 mM lipolysis buffer, 1x strength). N435 (Novozym 435), PL (porcine lipase), PP (porcine pancreatin), nd: Not defined. The lipolysis-permeation experiment using normal lipolysis buffer (FaSSIF1x) was not finished due to contamination of pH-electrode by LiDo membrane as well as time restriction.
2.5.2 In vitro lipolysis in small-scale setup
In an attempt to minimize full size ENA system, a comparator experiments were performed using small scale, Parallel Absorption and Lipolysis insert system “PALIS” (Figure 1B). LiDo membrane was mounted in custom polycarbonate inserts representing the apical side. These inserts were then transferred to wells containing 2 ml PBS (pH 7.4, 10 mM) supplemented with 4 wt% BSA. The experimental conditions are introduced in Table 2. High strength lipolysis buffer (FaSSIF100x) was used to perform lipolysis assay since no pH-stat could be used in this
system to keep pH of 6.5 on the donor side. Prior to the digestion study, the donor media was first spiked with either loaded formulation or blank LBF in a glass vial, vortexed and then introduced to each insert. Co-administered carvedilol was pre-dispersed in FaSSIF100x before it
due to volume restraints. The total exposure of carvedilol in the donor and receiver sides was assessed as mentioned earlier.
2.6 LiDo integrity during in vitro lipolysis assays
The integrity of LiDo membrane was evaluated during digestion in the donor compartment by spiking the digestion medium with Lucifer Yellow (10 µM). The fluorescence signal was determined in the samples from the receiver side using a Safire2 plate reader (Tecan, Austria).
As criteria for good barrier integrity, AUC for mass transfer of Lucifer Yellow in PALIS should be below 1 min.nmol/cm2 [15]. On the other hand, it should be below 0.5 min.nmol/cm5 in
ENA which corresponds to very high permeability.
2.7 Sample processing and HPLC analysis
2.8 Statistical analysis
Data presented as mean values with standard deviation (n=3). Statistical analysis was conducted in GraphPad Prism 9 (GraphPad Software, USA). Student’s t-test was used to compare differences between two groups. Two-way ANOVA was used to compare differences between more than two groups, followed by Tukey´s multiple comparisons analysis test. P-values less than 0.05 were regarded as statistically significant.
3. Results
3.1 Transport study
The transport study in PALIS showed that using 4% BSA in the receiver chamber provided significantly higher mass-transfer AUC of carvedilol than when TPGS was used. This trend was also observed for drug mass-transfer in a control study using ENA, although the differences were not statistically significant (Fig. A1). However, it should be noted that the transport study was initially performed in PALIS prior to performing the current study, while the control study using ENA was performed recently.
3.2 In vitro lipolysis-permeation assay
3.2.1 In vitro lipolysis-permeation in full-scale setup
In vitro lipolysis of formulations F1–F3 (Table 1) was performed using ENA, where F1
Figure 2. Total exposure of carvedilol in the upper (donor) and bottom (receiver) chamber in ENA setup during
70 min in vitro lipolysis-permeation of F1-F3 with porcin pancreatin, porcine lipase, and Novozym 435. (A) AUCs of carvedilol concentration in the aqueous phase vs. time. (B) AUCs of carvedilol mass transfer to the receiver side
vs. time. F1: carvedilol predissolved in LBF (0.5g, 25mg/g), F2: carvedilol co-administered with LBF
(12.5mg+0.5g), F3: carvedilol co-administered with LBF (12.5mg+0.25g). Values are expressed as average values ± SD (n=3). ns: no significant differences between AUCs, *significant differences between AUCs: ***p <0,001; **p <0.01; * p <0.05.
Porcine pancreatin to digest LBFs. The total dissolved carvedilol during digestion of F1 and
F2 was similar, but significantly lower for F3 (F1 = F2 > F3, Fig. 2A). Similar results were observed in terms of the total mass-transfer of carvedilol across LiDo membrane as depicted in Figure 2B.
Porcine lipase to digest LBFs. The dissolved drug upon digestion of F2 and F1 were
comparable. However, the difference between F2 and F3 was not significant (Fig. 2A). On the other hand, dissolved drugs in the digestion chamber were not related to mass transfer (Fig. 2B).
Novozym 435 to digest LBFs. The AUCs of the total concentration vs. time were not
significantly affected when the digestion of F1–F3 was initiated by Novozym 435 (Fig. 2A).
P. Pancreatin P. Lipase Novozym 435 0 10 20 30 40 AU Cdo no r (m in × m g/ m l) F1 F2 F3 ns ✱ ✱ ns ✱✱ ns ns ns ns ENA (donor side)
P. Pancreatin P. Lipase Novozym 435 0 100 200 300 400 500 AU Cre ce iv er (m in × µ g/ cm 2 ) F1 F2 F3 ns ✱ ✱ ns ns ns ns ✱✱ ✱✱✱ ENA (receiver side)
This result was not related to the total carvedilol mass transfer to the receiver side as mass transfer from F3 was significantly higher than from F2 and F3 (Fig. 2B).
Recovery during in vitro lipolysis-permeation experiments. The recovery was generally good
(75–100%) when using either pancreatic extract or lipase as digestive agent during in vitro lipolysis-permeation assays (Fig. A2). In contrast, low recovery (60%) was observed when using Novozym 435 as digestive enzyme upon lipolysis of F1 and F2 (Fig. A1G and H).
3.2.2 In vitro lipolysis-permeation in small-scale setup
In vitro lipolysis-permeation assay of F1–F3 was also performed in the small-scale system,
PALIS. AUCs for carvedilol mass transfer vs. time is shown below in Figure 3.
Figure 3. Total exposure (AUC of carvedilol concentration vs. time) of carvedilol in the receiver compartment in
PALIS setup during 70 min in vitro lipolysis-permeation of F1–F3 with porcin pancreatin, porcine lipase, and Novozym 435. Values are expressed as average values ± SD (n=3). ns: no significant differences between AUCs, *significant differences between AUCs: ***p < 0.001 and ****p < 0.0001.
Porcine pancreatin to digest LBF. The total mass transfer obtained from pre-dissolved
carvedilol (F1) was significantly higher than coadministered carvedilol with LBF (F2 and F3), as shown in Figure 3. This trend is not consistent with what was observed in ENA (Fig. 2B).
Porcine lipase to digest LBFs. The rank order of the total average mass transfer of the assayed
formulations was (F1 > F2 > F3). However, no significant difference between F2 and F3 could be observed (Fig. 3).
Novozym 435 to digest LBFs. The total mass transfer was comparable upon digestion of F1–F3
by Novozym 435. This ranking is similar to the dissolved drug in ENA, where the differences were not statistically significant (Fig. 3).
Recovery during in vitro lipolysis-permeation experiments. In contrast to ENA, PALIS did not
allow sampling of digestive medium due to low donor volume. Therefore, recovery was assessed for only the first and last sampling point. In terms of the recovery at last sampling point, it was below 70% upon digestion of F1–F3 by Novozym 435. For lipase and pancreatin, the recovery was greater than 80% for F1, but below 80% for F2 and F3 respectively.
3.3 LiDo integrity during in vitro lipolysis assays
LiDo membrane could withstand contact with lipolysis process performed with the following enzymes: porcine pancreatin, porcine lipase and Novozym 435. AUCs for Lucifer Yellow mass-transfer obtained from both ENA and PALIS was below the critical value, see Figure A3. Furthermore, in a control experiment carried out without Lucifer Yellow, a quantifiable signal was still observed after BSA precipitation.
4. Discussion
Although TPGS is animal free and doesn’t require sample preparation prior to UV measurements, it is not as suitable for mass spectrometry (MS) as BSA. TPGS does not interfere with analysis using HPLC-UV since it does not possess chromophore groups [27]. However, BSA can be precipitated and hence, more easily enable analysis using HPLC-MS. TPGS can disturb such analysis due to ion suppression effects.
To validate the use of LiDo membrane, conditions F1-F3 (Table 1) were selected for comparison. Carvedilol was either predissolved in LBF (F1) or coadministered with varying LBF dose (F2 and F3). Digestion of F1–F3 was performed using either porcine pancreatin, porcine lipase, or Novozym 453. Unfortunately, in vitro data from lipolysis-permeation assay using normal lipolysis buffer is uncompleted. Contamination of pH electrode caused by the LiDo membrane was as previously observed still problematic even though higher donor volume (100 ml) was used in an attempt to increase distance between the electrode and the membrane. The artificial membrane suffers from the absence of a boundary separating the donor media from the barrier constituents, possibly resulting in dissolution/emulsification of membrane constituents [24]. The cause of the issue is unknown, possibly the phospholipids or the dodecane settle on the glass membrane of the probe, affecting ion transport. This in turn leads to reduced stability and response time during pH measurement. Hence, the result discussed in this section relates to data obtained from experiments performed using high strength lipolysis buffer.
lipophilic drug can be incorporated, enhancing drug solubilization [5]. A previous study demonstrated that high drug aqueous concentration occurs at the higher dose of LBF and not the lower [28]. In turn, it can clarify the higher carvedilol mass transfer obtained from F1 and F2 compared to F3 (Fig. 2B). However, Alskär et al., previously performed in vitro lipolysis-permeation experiment on the currently assayed formulations using Caco-2 cell-based model where Novozym 435 was used as digestive agent. Additionally, formulation performance was also examined in vivo using Labrador dogs, results depicted in Figure 4 [13].
Figure 4. Results obtained from a previous study by Alskär et al. In vitro lipolysis-permeation assay performed in
ENA during 40 min, using Novozym 435 as digestive enzyme. (A) AUCs of carvedilol concentration in the aqueous phase (AP) vs. time. (B) AUCs of carvedilol mass transfer to the receiver side vs. time. (C) Relative bioavailability in vivo. Values are expressed as average values ± SD (n=3). *significant differences between AUCs: ***p <0,001; **p <0.01; * p <0.05.
lipase, and co-lipase, carboxyl ester hydrolase, phospholipase A2, and amylase and peptidases as well. Lipase and co-lipase are more relevant for LBF digestion, while amylase and peptidases are probably less relevant as sugars and proteins were not present in this study. In contrast, Novozym 435 contains a single enzyme immobilized into polymeric beads which hinder accessibility to the lipid droplets [29]. In summation, the digestion of long chain LBF into FFAs seems to be more effective using pancreatin rather than Novozym 435. In addition to the enzyme activity, volume of digestive medium as well as type of buffer capacity can also affect drug mass transfer. Compared to the previously mentioned study, we used lower donor volume (20 ml) and high strength lipolysis buffer. Better agitation can probably be maintained with lower donor volume, resulting in a smaller diffusion layer, leading to higher dissolution rate [30]. High buffer capacity keeps the same pH in bulk as in the diffusion layer which leads to a constant dissolution rate and hence, high drug mass transfer can then be obtained [17]. Consequently, we found that mass transfer of carvedilol across the LiDo membrane into the receiver chamber reflected the drug plasma exposure in vivo (Fig. 4C). As described above, FFA plays an important role for improving the solubility of lipophilic weak base drugs in vitro. In contrast to the in vitro situation, it was found that LBF triggers bile salt (BS) secretion in
dogs, enhancing lipophilic drug solubility [13]. BS concentration in dogs was similar to what
was observed previously in human study after administration of small clinically relevant dose of lipid [31]. On the other hand, a likely explanation for hampered absorption (in vivo) or mass transfer (in vitro) from F3 is that the entire dose was not completely dissolved due low dose LBF. However, long chain LBF keeps more drug in the post digestion products, decreasing the free drug concentration which is available to be permeated over the absorptive membrane [32]. This explains why only 2% av the entire dose of carvedilol permeated to the receiver side in this study (Fig. A2).
The assay with porcine lipase as digestive agent showed similar ranking of the dissolved carvedilol in the AP as previously observed in vitro (Fig. 2A and Fig 4A), but not the drug mass transfer (Fig 2B and Fig. 4B). These new results may not be entirely reliable since dissolved carvedilol in ENA reflected ranking of drug mass transfer in PALIS (Fig. 3). A possibility exists that the receiver samples obtained from F3 were more concentrated than expected due to too low amount of acetonitrile used for precipitation of BSA, compared to F1 and F2. This experiment needs to be repeated to further evaluate this result. Unlike outcomes with aforementioned digestive enzymes, assays with Novozym 435 as digestive agent demonstrated equal amounts of dissolved drug in the AP. It can be linked to incomplete lipid digestion since Novozym 435 was not properly dispersed in the donor medium due to aggregation. For that reason, more drug could be located in the oil phase, which was not quantified. Thus, this could explain the low recovery when F1 and F2 were digested using Novozym 435 (Fig. A2G–H). On the other hand, higher mass-transfer for F3 compared to F1 and F2 can be explained by that actual free concentration during lipolysis of F3 was higher than during lipolysis of F1 and F2, resulting in higher permeation of drug across the membrane.
Apart from this, no issues with barrier integrity were evident. Looking at AUC for mass-transfer from control experiment (without Lucifer Yellow) (Fig. A3A), it seems that rather BSA led to unspecified signal, resulting in very low calculated AUC value. Thus, no Lucifer Yellow was permeated across LiDo membrane during lipolysis-permeation experiments.
Pancreatin has been used previously and LiDo membrane could withstand pancreatin for only 30 min [15]. The enzyme concentration used in the previous study was however much higher than the one used in this work. This can explain why LiDo membrane was tolerated in the lipolysis process performed with pancreatin in this study. In summation, LiDo membrane retained its integrity during lipolysis process with all three digestive agents used in this study.
For any additional in vitro studies, the digestive enzyme and its activity should be chosen carefully. Higher BS concentration can be tested with enzymes of low lipolytic activity since BS secretions played an important role in vivo to improve the solubility of the lipophilic weak base drug. This can facilitate the accessibility of the enzyme to the oil/water interface. However,
in vitro system should be more biorelevant to humans. Therefore, it might be necessary to
consider species differences in terms of the bile secretion. For instance, rodents have higher average bile concentration in their GI fluids. Hence, the latter can be suboptimal in vivo model for LBF studies. Further in vitro studies should therefore, when possible, focus on comparison with in vivo studies in larger mammals with GI conditions comparable to humans. In addition to neutral and basic compounds already investigated (fenofibrate and carvedilol, respectively), lipophilic acids should also be tested in the currently investigated in vitro systems. It should also be noted for future studies, that compounds belonging to BCS class II are more relevant to study. They are more likely to benefit from LBF, since their limited solubility in GI fluids are the main obstacle for good bioavailability, with exception for metabolic degradation.
5. Conclusion
systems, the enzymes, enzyme activity, and BS concentration should be chosen carefully to mimic in vivo conditions. This study represents as small part in an ongoing effort to characterize and optimize the studied systems’ suitability for prediction of in vivo performance of LBF. Therefore, a larger number of drugs and formulations need to be studied before any solid conclusions of suitability can be drawn.
6. ACKNOWLEDGEMENTS
7. References
[1] Kuentz M, Holm R, Elder DP. Methodology of oral formulation selection in the pharmaceutical industry. European Journal of Pharmaceutical Sciences. 25 maj 2016;87:136–63.
[2] Neslihan Gursoy R, Benita S. Self-emulsifying drug delivery systems (SEDDS) for improved oral delivery of lipophilic drugs. Biomedicine & Pharmacotherapy. 01 april 2004;58(3):173–82.
[3] Pouton CW. Formulation of poorly water-soluble drugs for oral administration: Physicochemical and physiological issues and the lipid formulation classification system. European Journal of Pharmaceutical Sciences. 01 november 2006;29(3):278–87.
[4] Hauss DJ. Oral lipid-based formulations. Advanced Drug Delivery Reviews. 30 juli 2007;59(7):667–76.
[5] Feeney OM, Crum MF, McEvoy CL, Trevaskis NL, Williams HD, Pouton CW, m.fl. 50years of oral lipid-based formulations: Provenance, progress and future perspectives. Advanced Drug Delivery Reviews. 01 juni 2016; 101:167–94.
[6] Williams HD, Sassene P, Kleberg K, Calderone M, Igonin A, Jule E, m.fl. Toward the Establishment of Standardized In vitro Tests for Lipid-Based Formulations, Part 3: Understanding Supersaturation Versus Precipitation Potential During the In vitro Digestion of Type I, II, IIIA, IIIB and IV Lipid-Based Formulations. Pharm Res. 01 december 2013;30(12):3059–76.
[7] Tanaka Y, Tay E, Nguyen T-H, Porter CJH. Quantifying In vivo Luminal Drug Solubilization -Supersaturation-Precipitation Profiles to Explain the Performance of Lipid Based Formulations. Pharm Res. 01 mars 2020;37(3):1–17.
[8] Crum MF, Trevaskis NL, Williams HD, Pouton CW, Porter CJH. A new in vitro lipid digestion – in vivo absorption model to evaluate the mechanisms of drug absorption from lipid-based formulations. Pharm Res. 01 april 2016;33(4):970–82.
[9] Keemink J, Mårtensson E, Bergström CAS. Lipolysis-Permeation Setup for Simultaneous Study of Digestion and Absorption in vitro. Mol Pharmaceutics. 04 mars 2019;16(3):921– 30.
[10] Bevernage J, Brouwers J, Annaert P, Augustijns P. Drug precipitation–permeation interplay: Supersaturation in an absorptive environment. European Journal of Pharmaceutics and Biopharmaceutics. 01 oktober 2012;82(2):424–8.
[11] Berben P, Brouwers J, Augustijns P. The artificial membrane insert system as predictive tool for formulation performance evaluation. International Journal of Pharmaceutics. 15 februari 2018;537(1):22–9.
[12] Berben P, Brouwers J, Augustijns P. Assessment of Passive Intestinal Permeability Using an Artificial Membrane Insert System. JPharmSci. 01 januari 2018;107(1):250–6.
[13] Alskär LC, Parrow A, Keemink J, Johansson P, Abrahamsson B, Bergström CAS. Effect of lipids on absorption of carvedilol in dogs: Is coadministration of lipids as efficient as a lipid-based formulation? Journal of Controlled Release. 28 juni 2019; 304:90–100.
[14] Keemink J, Bergström CAS. Caco-2 Cell Conditions Enabling Studies of Drug Absorption from Digestible Lipid-Based Formulations. Pharm Res. 26 februari 2018;35(4):74.
[15] Hedge OJ, Bergström CAS. Suitability of Artificial Membranes in Lipolysis-Permeation Assays of Oral Lipid-Based Formulations. Pharm Res. juni 2020;37(6):99.
[17] Amaral Silva D, Al-Gousous J, Davies NM, Bou Chacra N, Webster GK, Lipka E, m.fl. Simulated, biorelevant, clinically relevant or physiologically relevant dissolution media: The hidden role of bicarbonate buffer. European Journal of Pharmaceutics and Biopharmaceutics. 01 september 2019; 142:8–19.
[18] Heider M, Hause G, Mäder K. Does the commonly used pH-stat method with back titration really quantify the enzymatic digestibility of lipid drug delivery systems? A case study on solid lipid nanoparticles (SLN). European Journal of Pharmaceutics and Biopharmaceutics. 01 december 2016; 109:194–205.
[19] Mudie DM, Shi Y, Ping H, Gao P, Amidon GL, Amidon GE. Mechanistic analysis of solute transport in an in vitro physiological two-phase dissolution apparatus. Biopharmaceutics & Drug Disposition. 2012;33(7):378–402.
[20] Mudie DM, Shi Y, Ping H, Gao P, Amidon GL, Amidon GE. Mechanistic analysis of solute transport in an in vitro physiological two-phase dissolution apparatus. Biopharmaceutics & Drug Disposition. 2012;33(7):378–402.
[21] Fagerberg JH, Tsinman O, Sun N, Tsinman K, Avdeef A, Bergström CAS. Dissolution Rate and Apparent Solubility of Poorly Soluble Drugs in Biorelevant Dissolution Media. Mol Pharmaceutics. 04 oktober 2010;7(5):1419–30.
[22]Mu L, Feng S-S. PLGA/TPGS Nanoparticles for Controlled Release of Paclitaxel: Effects of the Emulsifier and Drug Loading Ratio. Pharm Res. 01 november 2003;20(11):1864–72. [23]Kataoka M, Masaoka Y, Yamazaki Y, Sakane T, Sezaki H, Yamashita S. In vitro System
to Evaluate Oral Absorption of Poorly Water-Soluble Drugs: Simultaneous Analysis on Dissolution and Permeation of Drugs. Pharm Res. 01 oktober 2003;20(10):1674–80. [24] Berben P, Bauer-Brandl A, Brandl M, Faller B, Flaten GE, Jacobsen A-C, m.fl. Drug
permeability profiling using cell-free permeation tools: Overview and applications. European Journal of Pharmaceutical Sciences. 01 juli 2018; 119:219–33.
[25] Alvebratt C, Keemink J, Edueng K, Cheung O, Strømme M, Bergström CAS. An in vitro dissolution–digestion–permeation assay for the study of advanced drug delivery systems. European Journal of Pharmaceutics and Biopharmaceutics. 01 april 2020; 149:21–9. [26] Fuchs A, Dressman JB. Composition and Physicochemical Properties of Fasted-State
Human Duodenal and Jejunal Fluid: A Critical Evaluation of the Available Data. JPharmSci. 01 november 2014;103(11):3398–411.
[27] Christiansen A, Backensfeld T, Kühn S, Weitschies W. Investigating the Stability of the Nonionic Surfactants Tocopheryl Polyethylene glycol Succinate and Sucrose Laurate by HPLC–MS, DAD, and CAD. Journal of Pharmaceutical Sciences. 01 maj
2011;100(5):1773–82.
[28] Lennernäs H, Regårdh C-G. Evidence for an Interaction Between the β-Blocker Pafenolol and Bile Salts in the Intestinal Lumen of the Rat Leading to Dose-Dependent Oral Absorption and Double Peaks in the Plasma Concentration–Time Profile. Pharm Res. 01 juni 1993;10(6):879–83.
[29] Phan S, Salentinig S, Hawley A, Boyd BJ. Immobilised Lipase for In vitro Lipolysis Experiments. Journal of Pharmaceutical Sciences. 2015;104(4):1311–8.
[30] Sugano K. Theoretical comparison of hydrodynamic diffusion layer models used for dissolution simulation in drug discovery and development. International Journal of Pharmaceutics. 03 november 2008;363(1):73–7.
[31] Carrière F, Renou C, Lopez V, de Caro J, Ferrato F, Lengsfeld H, m.fl. The specific activities of human digestive lipases measured from the in vivo and in vitro lipolysis of test meals. Gastroenterology. 01 oktober 2000;119(4):949–60.
8. Appendix
Table A1. The masses of model compound, formulation and enzymes were required to perform the experiments
using ENA. Abbreviations: N435 (Novozym), PL (Porcine lipase), PP (Porcine pancreatin).
Designation Assay system Co-administere d carvedilol (mg) Pre-dissolved carvedilol (mg) Formulation used (g) Enzyme used Enzyme activity (TPU/ml) F1 ENA - 12.5 0.5 N 435 50 - 12.5 0.5 PP 50 - 12.5 0.5 PL 50 F2 ENA 12.5 - 0.5 N 435 50 12.5 - 0.5 PP 50 12.5 - 0.5 PL 50 F3 ENA 12.5 - 0.25 N 435 50 12.5 - 0.25 PP 50 12.5 - 0.25 PL 50
Table A2. The masses of model compound, formulation and enzymes were required to perform the experiments
using 6-wells plate as assay system. Abbreviations: N435 (Novozym), PL (Porcine lipase), PP (Porcine pancreatin).
Figure A1. Calculated AUCs of carvedilol mass-transfer to the receiver compartments vs time profiles from ENA
and PALIS. Flux of carvedilol was evaluated with phosphate buffered saline (pH 7.4, 10 mM) supplemented with either 4% (w/w) bovine serum albumin (BSA) or 0.2% (w/v) d-a-Tocopherol polyethylene glycol succinate (TPGS). Values are expressed as average values ± SD (n=3). * significant differences between AUCs: * p < 0.05,
ns: no significant differences between AUCs.
Figure A2. Mass balance where percentage carvedilol distribution in the system was calculated to assess the total
Figure A3. The total Lucifer Yellow mass transfer into the receiver compartment vs. time in both (A) ENA and
(B) PALIS. Dashed line represents criterion where membrane integrity ceases. Abbreviation: Ctrl (control), PP (porcine pancreatin), PL (porcine lipase), and N435 (Novozym 435).
