UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

The effect of digestion and drug load on absorption of poorly water soluble drugs from self-nanoemulsifying… Michaelsen, Maria Høtoft 2016

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
24-ubc_2016_november_michaelsen_maria.pdf [ 5.24MB ]
Metadata
JSON: 24-1.0314313.json
JSON-LD: 24-1.0314313-ld.json
RDF/XML (Pretty): 24-1.0314313-rdf.xml
RDF/JSON: 24-1.0314313-rdf.json
Turtle: 24-1.0314313-turtle.txt
N-Triples: 24-1.0314313-rdf-ntriples.txt
Original Record: 24-1.0314313-source.json
Full Text
24-1.0314313-fulltext.txt
Citation
24-1.0314313.ris

Full Text

THE EFFECT OF DIGESTION AND DRUG LOAD ON ABSORPTION OF POORLY WATER SOLUBLE DRUGS FROM SELF-NANOEMULSIFYING DRUG DELIVERY SYSTEMS (SNEDDS) by MARIA HØTOFT MICHAELSEN B.Sc., University of Copenhagen, 2009 M.Sc., University of Copenhagen, 2012    A DISSERTATION SUBMITTED IN PARTIAL FULFILLMENT OF THE REQUIREMENTS FOR THE DEGREE OF  DOCTOR OF PHILOSOPHY in THE FACULTY OF GRADUATE AND POSTDOCTORAL STUDIES (Pharmaceutical Sciences)         THE UNIVERSITY OF BRITISH COLUMBIA (Vancouver) September 2016 © Maria Høtoft Michaelsen, 2016  U N I V E R S I T Y  O F  C O P E N H A G E N  F A C U L T Y  O F  H E A L T H  A N D  M E D I C A L  S C I E N C E S           PhD Thesis Maria Høtoft Michaelsen   The effect of digestion and drug load on absorption of poorly water soluble drugs from self-nanoemulsifying drug delivery systems (SNEDDS)         This thesis has been submitted to the Graduate School of the Faculty of Health and Medical Sciences, University of Copenhagen 2016      i  Supervisors  Professor Thomas Rades (main supervisor) Department of Pharmacy Faculty of Health and Medical Sciences University of Copenhagen, Denmark Professor Anette Müllertz (principal co-supervisor, Denmark) Department of Pharmacy Faculty of Health and Medical Sciences University of Copenhagen, Denmark Professor and Dean Kishor M. Wasan (principal co-supervisor, Canada) College of Pharmacy and Nutrition University of Saskatoon, Canada Adjunct Professor at University of British Columbia, Canada Associate Professor Anan Yaghmur (co-supervisor) Department of Pharmacy Faculty of Health and Medical Sciences University of Copenhagen, Denmark   Assessment committee  Professor Bente Gammelgaard (Chair) Department of Pharmacy Faculty of Health and Medical Sciences University of Copenhagen, Denmark Professor Jürgen Siepmann College of Pharmacy University of Lille, France Senior Principal Scientist Gitte Pommergaard Pedersen Early systemic and topical development LEO Pharma, Denmark Associate Professor Shyh-Dar Li Faculty of Pharmaceutical Sciences University of British Columbia, Canada    ii  Preface The present thesis “The effect of digestion and drug load on absorption of poorly water soluble drugs from self-nanoemulsifying drug delivery systems (SNEDDS)” is submitted to meet the requirements for obtaining a joint PhD degree at the Faculty of Health and Medical Sciences, University of Copenhagen, Denmark and the Faculty of Pharmacy, University of British Columbia, Canada. The work was initiated in April 2013 and has been performed in the Rational Oral Drug Delivery (RODD) research group, Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Denmark and in the Wasan lab at the Faculty of Pharmacy, University of British Columbia, Canada. The project was performed in affiliation with the Drug Research Academy (DRA), University of Copenhagen. This dissertation is formatted in accordance with the regulations of the University of Copenhagen and submitted in partial fulfillment of the requirements for a PhD degree awarded jointly by the University of Copenhagen and the University of British Columbia.  Versions of this dissertation will exist in the institutional repositories of both institutions. The thesis has been defended at the Faculty of Health and Medical Sciences, University of Copenhagen, Denmark August 25th 2016. Three papers are published, submitted or in preparation in peer-reviewed journals and are included as essential parts of this thesis in the appendix. Paper I: Michaelsen MH, Wasan KM, Sivak O, Müllertz A, Rades T. The Effect of Digestion and Drug Load on Halofantrine Absorption from Self-nanoemulsifying Drug Delivery System (SNEDDS). The AAPS Journal. 2016: 18(1):180-6.  Paper II: Michaelsen MH, Abdi I, Wasan KM, Müllertz A, Rades T. Lipase inhibition does not change bioavailability of fenofibrate from SNEDDS and super-SNEDDS. (Submitted). Paper III: Sassene PJ*, Michaelsen MH*, Mosgaard MD, Jensen MK, Van Den Broek E,  Wasan KM, Mu H, Rades T, Müllertz A. In vivo precipitation of poorly soluble drugs from lipid based drug delivery systems. (Submitted). *PJS and MHM are joint first authors on this publication.       iii  Acknowledgements First I would like to thank my main supervisor Thomas Rades for his supervision and competent guidance throughout my PhD. Your enthusiasm and willingness to share your knowledge has been greatly appreciated. In addition a special thanks to my co-supervisors Anette Müllertz, Kishor Wasan and Anan Yaghmur for your always positive attitude and for the scientific support and guidance. I would also like to thank Olena Sivak (UBC) and Yang Hwan Yun (UCPH) for helping with experimental work. A special thanks goes to Professor Sven Frøkjær and Dean Robert Sindelar for providing a platform for the collaboration between the University of Copenhagen and the University of British Columbia. The collaboration has been fruitful and I feel like I have been given a rare chance to be a part of a collaboration like this. I hope the collaboration between the universities will be successful in the future and provides future PhD students with the same great experience I had.  A special thanks also goes to the Faculty of Health and Medical Sciences (UCPH) for financial support during my PhD. Furthermore I would like to acknowledge Lundbeck foundation and Oticon foundation for providing travel support for conferences. I had the pleasure to collaborate with Ismahan Mahad Abdi when she was undertaking her master thesis. It was a very pleasant collaboration and we both learned a lot in the process. The best of luck for you in the future. It has been a pleasure working both in the RODD group at the University of Copenhagen and in the Wasan lab at the University of British Columbia. The research environment at both places is something special and worth to treasure since it is built on mutual trust and support. And ROODs: Thanks for all the great table soccer matches – it has been fun and I am sure that the fun we all had shines through in our scientific footsteps. A special thanks also goes to Dr. Ragna Berthelsen, Jakob Munk Plum, Dr. Line Hagner Nielsen, and Ulrich Høgstedt for proofreading the thesis. Three years have gone and a new chapter is beginning. I am proud to have been a part of such great research communities and I am happy to have met so many inspiring people. Finally I would like to thank my family and friends for their uncompromising support and encouragement. Kasper, Maja, lille Maja, mom, and dad – you mean the world to me and I love you! Maria Høtoft Michaelsen Copenhagen, May 2016     iv  List of abbreviations AUC Area Under the Curve BCS Biopharmaceutics Classification System Cmax Maximum plasma concentration DAN Fa Danazol Absolute bioavailability FA(s) FAA(s) Fatty acid(s) Free Fatty acid(s) FDA Food and Drug Administration (US) FEN Fenofibrate GI Gastrointestinal HF Halofantrine IVIVC In Vitro – In Vivo Correlation LbDDS Lipid based Drug Delivery Systems (abbreviation is used for both singular and plural form) LC Long chain MC MG MGs Medium chain Monoglyceride Monoglycerides PLM Polarized light microscopy SEDDS Self-emulsifying Drug Delivery System Seq Equilibrium solubility SMEDDS Self-microemulsifying Drug Delivery System SNEDDS Self-nanoemulsifying Drug Delivery System Super-SNEDDS Supersaturated Self-nanoemulsifying Drug Delivery System S-SEDDS Superstaurable Self-emulsifying Drug Delivery System  tmax Time to reach Cmax UWL Unstirred Water Layer XRPD X-ray powder diffraction    v  Definitions and terms used in the thesis Biorelevant media Media simulating the fluids in the gastrointestinal tract. cLogP  Calculated LogP Lipolysis end products Fatty acids, monoglycerides, and diglycerides produced by hydrolysis of triglycerides. Loading capacity Amount of drug than can be dissolved in a formulation or excipient. LogP  Partition coefficient. Microemulsion An isotropic thermodynamically stable swollen micellar solution. Mixed Micelles Micelles formed by aggregation of different components such as bile salts and phospholipids. Nanoemulsion Kinetically stable emulsion with a droplet size in the nanometer range. SMEDDS Self-emulsifying drug delivery system that forms a thermodynamically stable micro-emulsion when introduced in water. SNEDDS Self-emulsifying drug delivery system that forms a kinetically stable micro-emulsion when introduced in water.      vi  Abstract The rising numbers of poorly water soluble drugs found in the pipelines of the pharmaceutical industry elicit the need for the development of enabling formulations. Lipid-based drug delivery systems (LbDDS) are a versatile group of formulations including self-nanoemulsifying drug delivery systems (SNEDDS). SNEDDS are mixtures of oils, surfactants, and co-solvents prepared as preconcentrates and are isotropic. They are typically dosed in capsules and upon introduction into an aqueous phase and after opening the capsules, form nanoemulsions under gentle agitation. The advantage of using LbDDS, such as SNEDDS, for poorly water soluble drugs is that typically the oral bioavailability is increased compared to traditional solid dosage forms. The exact mechanism behind the increase in oral bioavailability, however, is largely unknown. The fact that the drug is dosed in solution and that digestion of the lipid excipients from the preconcentrate takes place in the gastro intestinal tract have been thought to play a role with regard to the increased oral bioavailability and trafficking of the drug from the lumen to the absorptive membrane. The purpose of the present thesis was to investigate the effect of digestion and drug load on oral absorption of poorly water soluble drugs from SNEDDS. Further, to examine whether a rat model was suitable for the in vivo pharmacokinetic experiments and to explore if a correlation between in vitro lipolysis experiments and in vivo studies in rats could be made. In order to investigate if the correlation was possible, an in vivo precipitation study in rats was performed to gain insight in the precipitation behavior of poorly water soluble drugs in vivo upon administration of SNEDDS and other LbDDS. In order to examine the effect of lipid digestion on absorption of poorly water soluble drugs from SNEDDS the lipase inhibitor orlistat (tetrahydrolipstatin) was used. Halofantrine and fenofibrate were chosen as drugs because of the different solid state characteristics of the precipitates from the dynamic in vitro lipolysis model. In the model halofantrine was found to precipitate in an amorphous form whereas fenofibrate was found to precipitate in a crystalline form. For neither halofantrine nor fenofibrate digestion appeared to have an effect on the oral bioavailability in rats, however, for halofantrine a shift in the pharmacokinetic parameters was observed. SNEDDS are typically formulated in a way that the drug load is below the equilibrium solubility (Seq), which may result in administration of several capsules in order to obtain the required dose of the drug, thus leading to problems with patient compliance. In super-SNEDDS the drug load is    vii  above Seq. The administration of halofantrine and fenofibrate in super-SNEDDS led to increased bioavailability compared to SNEDDS when evaluated in a rat model in vivo. For both halofantrine and fenofibrate there was a significant increase of Cmax for the super-SNEDDS (0.96±0.2 µg/mL and 13.34±4.7 µg/mL, respectively) compared to the SNEDDS (0.51±0.1 µg/mL and 7.36±1.7 µg/mL, respectively). For halofantrine no significant increase was seen in the bioavailability for the super-SNEDDS however, for fenofibrate the AUC of the super-SNEDDS (148.00±47.5 hr·µg/mL) was significantly larger than for the SNEDDS (88.28±20.9 hr·µg/mL).  The dynamic in vitro lipolysis model is typically used in the evaluation of digestibility and solubilization capacity of SNEDDS and other LbDDS. The model simulates the intestinal milieu with regard to composition of the intestinal fluids, pH, and lipase activity. The traditional interpretation of data from the lipolysis model is that the solubilized fraction of drug is available for absorption, whereas the precipitated fraction needs to be re-dissolved in order to be absorbed which is believed to lower the bioavailability. However, for super-SNEDDS precipitation seems to be inversely correlated with oral absorption when comparing the in vitro data with the in vivo data. The reason for the discrepancies between in vitro and in vivo data may be due to the simplification of the in vitro lipolysis model compared to the in vivo situation. The in vitro model is lacking a gastric digestion step, the gastric emptying rate is not taken into consideration, and the model is missing an absorption step. In order to examine precipitation of poorly water soluble drugs dosed in LbDDS in vivo the gastric and intestinal content of rats was examined for precipitated crystalline drug using X-ray powder diffraction and polarized light microscopy. The study revealed that precipitation only was evident in the stomach and not the intestine, thus precipitation in the intestinal in vitro model may be an artifact and may require the addition of an in vitro gastric digestion step to be predictive in vivo.  In summary, this thesis has shown that digestion, that was traditionally thought to play a major role in the increased absorption of drugs from SNEDDS, may not be crucial for this process. Studies with the lipase inhibitor orlistat have revealed that digestion in fact only plays a minor role for absorption of drugs from SNEDDS. Most probably, increased drug absorption is due to the large surface area of the nanoemulsion and that the drug is dosed in solution. In addition, the nanoemulsion acts as a reservoir of dissolved drug, replenishing the free drug in the intestinal fluids, as the drug is absorbed. Further, super-SNEDDS increase the oral bioavailability compared to SNEDDS of both halofantrine and fenofibrate in a rat model thus challenging the paradigm that    viii  lipid amount is positively correlated with absorption of poorly water soluble drugs from SNEDDS. Rats seem however, to be a reasonable model for in vivo evaluation of super-SNEDDS.  The current in vitro model does not correlate well with in vivo studies in rats and precipitation in the dynamic in vitro lipolysis model may be an artifact due to the lack of a gastric step and/or an absorption step in the model. Future studies should include additional investigations on the effect of digestion on absorption from SNEDDS using for example different concentrations of orlistat and indigestible lipid excipients. Further, the effect of drug load should be investigated using super-SNEDDS with different degrees of supersaturation. Lastly, in order to obtain in vitro in vivo correlations further development and the introduction of a gastric and/or an absorption step in the dynamic in vitro lipolysis model is needed.      ix  Resumé (abstract in Danish) I den farmaceutiske industri er der et øget antal af tungtopløselige lægemiddelstoffer i udviklingsfasen. Dette er en udfordring og nye lægemiddelformuleringer bør udvikles for at adressere den lave opløselighed og dårlige biotilgængelighed der ses hos denne gruppe af lægemiddelstoffer. Lipidbaserede formuleringer er en bred gruppe af formuleringer og herunder findes selv-nanoemulgerende drug delivery systemer (SNEDDS). SNEDDS er isotropiske blandinger af olier, surfaktanter, og co-solventer og denne blanding kaldes et præ-koncentrat. Disse formuleringer administreres ofte i gelatine kapsler og når de introduceres til et vandigt miljø under nænsom omrøring dannes en nanoemulsion. Fordelen ved at anvende lipidbaserede formuleringer, så som SNEDDS, er at den orale biotilgængelighed ofte øges sammenlignet med traditionelle formuleringer. Den eksakte mekanisme bag denne øgning af biotilgængeligheden er ukendt. Fordøjelse af lipider i formuleringen samt at lægemiddelstoffet doseres opløst menes at spille en vigtigt rolle i forbindelse med transporten af lægemiddelstof fra tarmen til epitelet på baggrund af dannede kolloid strukturer. Formålet med denne afhandling er at undersøge effekten af fordøjelse og koncentrationen af lægemiddelstof  på den orale absorption af lægemiddelstof fra SNEDDS. Desuden, undersøges det om en rotte model er brugbar i forhold til at bestemme farmakokinetiske parametre fra overmættede SNEDDS (super-SNEDDS) og om en korrelation mellem in vitro lipolyse forsøg og de opnåede in vivo farmakokinetiske parametreforsøg kunne opnås. I forsøget på at opnå denne korrelation blev udfældningen af lægemiddelstof efter dosering af lipidbaserede formuleringer i rotte tarme undersøgt. For at undersøge effekten af lipid fordøjelse på absorption of tungtopløselige lægemiddelstoffer fra SNEDDS blev lipase hæmmeren orlistat (tetrahydrolipstatin) anvendt. Halofantrin og fenofibrat blev valgt som model stoffer grundet deres forskellige fastfase egenskaber efter udfældning under in vitro lipolyse. I den dynamiske in vitro lipolyse model udfælder halofantrin i en amorf form hvorimod fenofibrat udfælder i en krystallinsk form. For hverken halofantrin eller fenofibrat spillede lipid fordøjelse en rolle for absorptionen og den oral biotilgængelighed i rotter. Dog sås en ændring i de farmakokinetiske parametre for halofantrin når orlistat var tilstede. SNEDDS formuleres ofte således at koncentrationen af lægemiddelstof er under ligevægtsopløseligheden. Dette kan resultere i administration af flere kapsler for at opnå den    x  korrekte dosis af lægemiddelstof hvilket kan lede til problemer med patient komplians. I super-SNEDDS er koncentrationen af lægemiddelstof over ligevægtsopløseligheden og prækoncentraterne er derfor overmættede. Administration af halofantrin og fenofibrat i disse super-SNEDDS øger den orale biotilgængelighed når der sammenlignes med SNEDDS efter analyse af data fra en rotte model. For både halofantrin og fenofibrat sås en signifikant forøgelse af Cmax for super-SNEDDS (henholdsvis 0.96±0.2 µg/mL og 13.34±4.7 µg/mL) sammenlignet med SNEDDS (henholdsvis 0.51±0.1 µg/mL og 7.36±1.7 µg/mL). For halofantrin sås en stigning af biotilgængeligheden for super-SNEDDS der dog ikke var signifikant. For fenofibrat sås der derimod en øget AUC for super-SNEDD (148.00±47.5 hr·µg/mL) i forhold til SNEDDS (88.28±20.9 hr·µg/mL). Den dynamiske in vitro lipolyse model bliver ofte anvendt for at evaluere fordøjelseshastigheden af SNEDDS eller andre lipidbaserede formuleringer, samt formuleringens evne til at holde lægemiddelstoffet opløst under fordøjelse Modellen simulerer det intestinale miljø med hensyn til sammensætningen af de intestinale væsker, pH og lipase aktivitet. Traditionelt anses den opløste fraktion af lægemiddelstof som den fraktion der kan absorberes, hvorimod den udfældede fraktion af lægemiddelstof skal genopløses før lægemiddelstoffet kan absorberes. Super-SNEDDS har til trods for en meget høj grad af udfældning in vitro en høj biotilgængelighed in vivo. Uoverensstemmelsen mellem in vitro og in vivo data kan skyldes den relativt simple in vitro lipolyse model, som både mangler et gastrisk fordøjelses trin samt et absorptions trin. For at undersøge i hvilken grad tungtopløselige lægemiddelstoffer doseret i lipidbaserede formuleringer udfælder in vivo blev indholdet af mavesækken samt tarmindholdet af rotter doseret med lipidbaserede formuleringer undersøgt med røntgen (X-ray powder diffraction) samt mikroskopi for at lokalisere udfældede krystaller af lægemiddelstof. Studiet afslørede at udfældning kun fandt sted i maven og ikke i tarmen. Derfor blev det konkluderet at udfældningen i den intestinale lipolyse model muligvis var en artefakt og muligvis bedre beskriver fordøjelsesprocessen i maven.  For at opsummere, så har denne afhandling afslørede ud fra studier med lipasehæmmeren orlistat at fordøjelse ikke spiller en stor rolle for absorptionen af lægemiddelstofferne halofantrin og fenofibrat fra SNEDDS. Sandsynligvis er grunden til den øgede absorption fra SNEDDS i forhold til andre lipidbaserede formuleringer, det store overfladeareal medieret af nanoemulsionen samt at lægemiddelstoffet er opløst. Desuden, vil nanoemulsionen fungere som et reservoir der kan erstatte det absorberede lægemiddelstof i de intestinale væsker. Derudover, så øger super-SNEDDS biotilgængeligheden i forhold til SNEDDS for både halofantrin og fenofibrat. Dette sætter    xi  spørgsmålstegn ved det hidtidige paradigme at mængden af lipid er positivt korreleret med absorptionen af tungtopløselige stoffer fra SNEDDS. Udfældningen i den dynamiske in vitro lipolyse model er muligvis en artefakt grundet det manglende gastriske fordøjelsestrin (herunder inkluderet manglende mavetømningstrin) samt manglende absorptionstrin. Fremtidige forsøg bør derfor fokusere på effekten af fordøjelse af SNEDDS, på absorptionen af tungtopløselige lægemiddelstoffer, ved hjælp af øgede koncentrationer af orlistat samt anvendelsen af ufordøjelige lipider i formuleringerne. Desuden, skal effekten af lægemiddelstofkoncentrationen i formuleringerne undersøges nærmere ved at bruge forskellige mætningsgrader af lægemiddelstof. Til slut, for at opnå korrelation mellem in vitro og in vivo forsøg bør in vitro lipolyse modellen udvikles yderligere og især inklusionen af et gastrisk trin samt et absorptions trin bør være i fokus.      xii  Table of Contents                                                                                                                                                                                                                                                                                     Supervisors .......................................................................................................................................................... i Assessment committee ........................................................................................................................................ i Preface ................................................................................................................................................................ ii Acknowledgements ........................................................................................................................................... iii List of abbreviations .......................................................................................................................................... iv Definitions and terms used in the thesis ............................................................................................................. v Abstract ............................................................................................................................................................. vi Resumé (abstract in Danish) .............................................................................................................................. ix Table of Contents ............................................................................................................................................. xii 1. Introduction ................................................................................................................................................... 1 1.1 Purpose of project, hypothesis, and aims ............................................................................................ 2 1.2 SNEDDS.............................................................................................................................................. 2 1.3 Model drugs ......................................................................................................................................... 3 1.4 Tetrahydrolipstatin (orlistat) ................................................................................................................ 4 1.5 The Biopharmaceutics Classification System ..................................................................................... 4 1.6 The lipid formulation classification system ......................................................................................... 5 1.7 Lipid digestion and absorption of digestion products in humans ........................................................ 6 1.8 In vivo drug solubilization and absorption from LbDDS .................................................................... 9 1.9 The digestive system in rats............................................................................................................... 12 2. The effect of digestion on absorption of drugs from SNEDDS ................................................................... 13 2.1 Methodologies investigating the effect of digestion on absorption of poorly water soluble drugs from SNEDDS ......................................................................................................................................... 14 2.2 Investigation of the effect of digestion on drug absorption ............................................................... 14 2.3 Inhibition of lipase activity ................................................................................................................ 16 2.4 Drug effect on absorption from SNEDDS ......................................................................................... 16 2.5 Proposed mechanism of drug absorption from SNEDDS when digestion is inhibited ..................... 17 2.6 Summary of results ............................................................................................................................ 18 3. The effect of drug load on absorption from SNEDDS ................................................................................ 19 3.1 Methodologies used to assess the effect of drug load on absorption from SNEDDS ....................... 19    xiii  3.2 The effect of drug load on absorption from SNEDDS ...................................................................... 20 3.3 Supersaturation of drug in SNEDDS increases oral absorption ........................................................ 21 3.4 Summary of results ............................................................................................................................ 22 4. Dynamic intestinal in vitro lipolysis in the evaluation of SNEDDS ........................................................... 23 4.1 Methodologies used to evaluate SNEDDS and super-SNEDDS in vitro .......................................... 24 4.2 Dynamic in vitro lipolysis – the effect of digestion .......................................................................... 24 4.3 Dynamic in vitro lipolysis – the effect of drug load .......................................................................... 26 4.4 Optimizing the in vitro lipolysis model ............................................................................................. 27 4.5 Summary of results ............................................................................................................................ 28 5. Drug precipitation in vivo ............................................................................................................................ 29 5.1 Methodologies used to determine in vivo precipitation ..................................................................... 29 5.2 Drug precipitation in vivo .................................................................................................................. 30 5.3 Summary of results ............................................................................................................................ 36 6. The rat as a model for evaluating super-SNEDDS in vivo and obtaining a relationship between in vitro and in vivo data ....................................................................................................................................................... 37 7. Conclusion ................................................................................................................................................... 40 8. Future Perspectives ...................................................................................................................................... 42 9. References ................................................................................................................................................... 43 10. Appendix ................................................................................................................................................... 56       1  1. Introduction A growing amount of poorly water soluble drugs is found in the pipelines of the pharmaceutical industry, most likely due to new screening methods and the use of combinatorial chemistry in drug discovery [1, 2]. This calls for advanced enabling drug delivery systems in order to address this issue, which can be seen as the most pressing drug delivery issue for oral low molecular weight drugs [2, 3]. Many of these poorly water soluble drugs belong to class II of the Biopharmaceutics Classification System (BCS) and have a low aqueous solubility and/or a slow dissolution rate leading to a poor and erratic bioavailability due to the physicochemical properties of the compounds [2]. Among other advanced enabling drug delivery systems, oral lipid based drug delivery systems (LbDDS) have attracted specific of attention over the last decades because of their ability to increase the oral bioavailability of poorly water soluble drugs. LbDDS cover a range of drug delivery systems such as oil solutions, micellar systems, emulsions, microemulsions, self-emulsifying drug delivery systems (SEDDS), self micro-emulsifying drug delivery systems (SMEDDS), and self nano-emulsifying drug delivery systems (SNEDDS) [4-7]. SNEDDS typically consist of lipids, surfactants and co-solvents and have the ability to deliver the drug in solution to the gastro-intestinal (GI) tract [8]. The processes related to the increased bioavailability of the drugs from such dosage forms are not fully understood but are thought to involve digestion of excipients and subsequent formation of various colloidal structures [6, 9].  In super-saturated SNEDDS (super-SNEDDS) the drug is introduced into the SNEDDS above its equilibrium solubility (Seq). Super-SNEDDS have attracted attention recently due to their ability to lead to an equal or higher bioavailability compared to SNEDDS in dogs and minipigs [10-13]. In conventional SNEDDS the drug amount that can be dosed is restricted by the solubility of the drug in the SNEDDS, which can result in the need of administration of several capsules in order to obtain the required dose [14].  Super-SNEDDS have the advantage that it is possible to lower the ‘pill burden’ and still obtain the required dose of drug. The present thesis presents results from studies evaluating the importance of digestion on absorption of poorly water soluble drugs from SNEDDS. Further, results evaluating the effect of drug load on absorption are presented in this thesis as well as an evaluation of the rat model in the in vivo assessment of super-SNEDDS. In vitro results using dynamic intestinal lipolysis are presented and compared with in vivo data.    2  1.1 Purpose of project, hypothesis, and aims The purpose of this PhD project is to reach a better understanding of the role of digestion and drug load in the absorption of poorly water soluble drugs from SNEDDS. Further, to investigate whether a rat model can be used for evaluating the performance of super-SNEDDS. The underlying hypotheses of this thesis are that digestion plays an important role for the absorption of poorly water soluble drugs from SNEDDS; that super-SNEDDS provide a good alternative to SNEDDS in order to increase bioavailability without increasing the total volume of formulation administered; and that the performance of SNEDDS could be assessed in a rat model. This led to the following three aims:  Assessment of the importance of digestion on absorption of poorly water soluble drugs from SNEDDS in rats.  Evaluation of the performance of super-SNEDDS compared to conventional SNEDDS in vitro and in vivo in a rat model.  In vivo evaluation of precipitation of poorly water soluble drugs from LbDDS in a rat model. 1.2 SNEDDS SNEDDS are isotropic mixtures of lipids, surfactants and co-solvents that can deliver poorly water soluble drugs in solution, thereby circumventing the dissolution step in the GI tract and increasing oral bioavailability [15, 16]. SNEDDS are prepared as preconcentrates and will upon introduction into an aqueous environment under gentle agitation form a nanoemulsion. Although nanoemulsiona are “only” kinetically stable, physical stability problems are avoided due to the fact that the emulsion only forms upon administration. Moreover, nanoemulsions are sometimes physically stable over prolonged periods of time and therefore referred to as ‘approaching thermodynamic stability’ [17].  Typically SNEDDS are administered in soft gelatin capsules. A large number of different excipients are commonly used in the formulation process and the excipients are the determining factor for the loading capacity of the drug in the formulation [7, 9]. The dose of the drug to be administered is governed by the solubility of the drug in the preconcentrate, as the drug is usually loaded at levels below the Seq. This can result in the need of administering several capsules in order to obtain the required dose of the drug [14]. The excipients can be selected based on different rationales such as including a known p-glycoprotein (pgp) inhibiting excipient in the formulation if the drug is a    3  substrate for pgp efflux transporters in the small intestine or choosing long chain (LC) triglycerides in order to facilitate lymphatic transport of very lipophilic drugs [9, 18, 19]. In order to facilitate higher loading capacities super-SNEDDS have been developed. Super-SNEDDS contain the drug in solution above the Seq and thereby make it possible to administer the required dose with fewer capsules. Studies have shown that dosing poorly water soluble drugs in super-SNEDDS yields an equal or higher bioavailability of the drug compared to SNEDDS [10-13] and by supersaturating the drug in the pre-concentrate the thermodynamic activity of the drug is increased [20, 21]. The stability of super-saturated systems, however, may be of some concern since the drug thermodynamically is in an unfavorable state, and may thus precipitate in the formulation over time [22]. 1.3 Model drugs Halofantrine (HF) and fenofibrate (FEN) were chosen as the main model drugs in this thesis based on their low aqueous solubility and high lipophilicity. HF is a phenanthrenemethanol antimalarial compound used against multidrug resistant strains of plasmodium falciparum [23]. It is a weak base (pKa 7.64), has a log P value of 8.5, and is practically insoluble in water [24]. FEN belongs to the class of fibrates and is used in the treatment of dyslipidia [25, 26]. It is a neutral compound with a log P value of 5.2 and a solubility of 0.3 µg/mL in water [27, 28]. In in vitro dynamic lipolysis studies HF precipitates in an amorphous form whereas FEN precipitates in a crystalline form [12, 13]. Danazol (DAN) was also utilized in one study in the present thesis. DAN is a synthetic steroid used in the treatment of endometriosis [29]. It is a neutral drug and has a log P value of 4.53 and an aqueous solubility below 1 µg/mL [30, 31]. All the above mentioned model drugs belong to BCS class II and the structures of the compounds are shown in Figure 1. ClClFF FN   Halofantrine Fenofibrate Danazol Figure 1 Chemical structures of the model compounds; halofantrine (HF), fenofibrate (FEN) and danazol (DAN).    4  1.4 Tetrahydrolipstatin (orlistat) Tetrahydrolipstatin (orlistat) is a hydrogenated derivative of lipstatin isolated from Streptomyces toxitricini [32]. It is an inhibitor of pancreatic lipase, gastric lipase and carboxyl ester lipase and has previously been used in the treatment of obesity and hypercholesterolemia [32-37]. Orlistat is a highly lipophilic (cLogP 8.6), surface active compound that forms stable monolayers at oil-water interfaces and mixed micelles with bile salts. Orlistat is present at the interface between the lipids (substrate) and the intestinal fluids and the lipase and orlistat react at the oil-water interface forming an enzyme-inhibitor complex by orlistat binding to a serine residue at the active site of the lipase [36, 38-40]. Orlistat reacts with lipases at a 1:1 molar ratio and acts in a dose-dependent manner [32, 36]. The literature shows some disagreements as to whether this complex formation is reversible or not [32, 36, 41]. The structure of orlistat is shown in Figure 2.  Orlistat Figure 2 Chemical structure of orlistat. 1.5 The Biopharmaceutics Classification System Permeability and solubility are key parameters when evaluating drug absorption. Amidon and co-workers described the BCS where drugs are classified based on their permeability and solubility as either class I (high permeability, high solubility), class II (high permeability, low solubility), class III (low permeability, high solubility) or class IV (low permeability, low solubility) [42]. The classes are shown in Figure 3. According to the food and drug administration (FDA), a drug has a high permeability if more than 90% of the dose is absorbed in vivo in humans [43]. Solubility is evaluated over the pH range from 1-7.5 at 37oC and a drug is classified as highly soluble if the highest clinically relevant dose is freely soluble in 250 mL aqueous media across the mentioned pH-range [42].     5   Figure 3 Biopharmaceutics classification system displaying the four classes based on evaluation of permeability and solubility. Modified from [42]. The BCS can be used as a tool in formulation of drug delivery systems, but it also has a regulatory impact for BCS biowaivers on immediate release formulations [43, 44]. For drugs belonging to class I in the BCS and formulated as solid immediate release oral dosage forms with rapid dissolution,  biowaivers can be granted based on dissolution studies [43, 45]. 1.6 The lipid formulation classification system In 2000 Pouton presented the lipid formulation classification system (LFCS) dividing LbDDS into three classes (class I, II, and III) based on the excipients present in the systems (see Table 1). Class III was further subdivided into classes IIIa and IIIb [46]. In 2006 an updated version was introduced including an additional class of LbDDS called class IV [47]. Class I formulations are non-dispersing LbDDS and are typically used for highly lipophilic drugs. Class II formulations do not contain any water soluble surfactants and therefore the formulation retains its solvent capacity upon dispersion. Formulations belonging to class IIIa and IIIb may lose their solvent capacity upon dispersion due to the hydrophilic surfactants being dissolved or micellized in the aqueous phase. Class IV formulations are often indigestible or only slightly digestible due to the lack of lipids [47]. The LFCS can be used as a tool to categorize the plethora of LbDDS, however since the LFCS does not cover all commercially available LbDDS, Müllertz et al. have suggested alterations to the LFCS where the bulk behavior of the lipids is taken into account (according to the type of excipients (oils,    6  surfactants (high HLB), surfactants (low HLB) and co-solvents)) using Small’s lipid classification system [48, 49].  Table 1 Typical compositions of LbDDS as described by the lipid formulation classification system (LFCS). Table adapted from reference with permission from Elsevier [47].  Percentage in formulation (%, w/w) Excipients Type I Type II Type IIIa Type IIIb Type IV Oils (tri-, di-, and monoglycerides) 100 40-80 40-80 <20 - Water insoluble surfactants, HLB < 12 - 20-60 - - 0-20 Water soluble surfactants, HLB > 12 - - 20-40 20-50 30-80 Hydrophilic co-solvents - - 0-40 20-50 0-50  1.7 Lipid digestion and absorption of digestion products in humans Digestion is a dynamic process and lipid digestion has been studied extensively over the last four decades [50-52]. Understanding the digestion process is of paramount importance for understanding the fate of LbDDS after oral administration [8]. Triglycerides are the primary source of lipids in a normal western diet and a main function of the digestive system is the hydrolysis of the triglycerides to the absorbable monoglycerides (MGs) and free fatty acids (FFA). The process is initiated in the mouth with mastication and secretion of lingual lipase (although the literature is controversial about this) and continued in the stomach and in the small intestine [53, 54]. 1.7.1 Gastric lipolysis The pH of the stomach is dependent on the prandial state and is usually around 1.4-2.1 in the fasted state and elevated to 4.3-5.4 in the fed state in human adults [55]. Hutchinson et al. have found the fluid volume in the stomach in the fasted to be approximately 23.3 ± 17.3 mL and in the fed state the volume is dependent on the intake of food and fluids [56]. The gastric fluids contain gastric lipase that catalyzes the hydrolysis of triglycerides. The pH optimum of gastric lipase has been found to be between 3 and 6. Furthermore, medium chain (MC) triglycerides are digested faster than long chain (LC) triglycerides [57, 58]. Gastric lipase can hydrolyze all three ester bonds in triglycerides but does show some selectivity towards the sn-3 position [59]. Approximately 20% of the total hydrolysis of triglycerides taking place in the GI tract is ascribed to gastric digestion [60]. The gastric digestion is thought to be important for the subsequent intestinal digestion since the    7  process facilitates the emulsification of ingested triglycerides [61]. Furthermore, gastric contractions produce shear forces (especially in the antrum of the stomach) that aid in the emulsification process [53]. 1.7.2 Intestinal lipolysis The fluid volumes found in the small intestine are highly variable and the fluid is not homogenously distributed but rather separated in small pockets along the intestine [62]. Schiller and colleagues have measured the fluid volumes in healthy volunteers by magnetic resonance imaging (MRI) and found that the small intestine contained 105 ± 72 mL in the fasted state and 54 ± 41 mL in the fed state [62]. The pH in the duodenum in the fasted state is 5.6-7.0 (median 6.3) and in the fed state the pH value is 5.4-6.5 (median 6.0). In the jejenum pH is slightly increased compared to the duodenum with reported pH values between 6.5-7.8 (median 6.9) in the fasted state and 6.1 in the fed state [63]. From the stomach the chyme (partly digested food) is gradually emptied out to the duodenum and the entry of lipids in the duodenum stimulates secretions from the pancreas and gallbladder [64]. In the duodenum the chyme is thus mixed with secretions from the pancreas and gallbladder. The biliary components (bile salt, phospholipid and cholesterol) from the gall bladder are excreted as biliary mixed micelles that interact with the surface of the lipid emulsion droplets generated during the solubilization process of the ingested lipids and digestion products [59, 65]. The concentration of bile salts is typically between 2.5-5.9 mM (median of 3.25 mM) in the fasted state and increase to 3.6-24.0 mM (median of 11.8 mM) in the fed state [63]. The enzyme rich secretions from the pancreas containing pancreatic lipase and co-lipase, as well as other enzymes, hydrolyze triglycerides [59, 66]. Pancreatic lipase is active at the interface between the lipids and the intestinal fluids and to be effective is thus dependent on emulsification occurring in the stomach [67]. }. Pancreatic lipase has a pH optimum around 6.5 and requires a co-lipase in order to have optimal activity [68].  Pancreatic lipase hydrolyzes the triglycerides selectively at the sn-1 and sn-3 position to the digestion products 2-MG and two FFAs. The lipid digestion process is very efficient and more than 95% of the ingested lipids are digested and absorbed [69]. 1.7.3 Absorption of digestion products The digestion products produced upon lipase hydrolysis interact in various ways with the intestinal fluids. The release of one FA from the triglycerides results in the formation of diglycerides that are    8  only weakly polar and still exist in a dispersed phase in the intestinal fluids. MGs are more polar than diglycerides and can form various transient structures in the intestinal fluids. In the absence of bile salts, MGs and FFAs will be present at the oil-water interface due to their amphiphilic nature and will inhibit the lipase activity. However, micellar solubilization of the digestion products due to the presence of bile salt micelles eliminates the inhibition in vivo [53, 67]. In the early 1960’s Hofmann and Borgström identified the presence of two distinct phases present in the intestinal fluids after lipid digestion; a lipid rich phase and an aqueous phase containing mixed bile salt micelles [50]. Later the presence of unilamellar vesicles and mixed micelles was confirmed by Carey and co-workers [51]. Mixed micelles are micelles consisting of more than one species including bile salts as defined by Carey and Small [51]. Staggers, Hernell, and co-workers have investigated the formation of colloidal structures using ternary phase diagrams and duodenal aspirates. They hypothesized that the digestion products formed at the surface of the triglyceride rich oil droplets during digestion, bud off and form small unilamellar vesicles in an equilibrium with mixed micelles upon digestion [70, 71].  The enterocytes are separated from the lumen by the unstirred water layer (UWL), which acts as a barrier for the absorption of the digestion products [72, 73]. FFAs and MGs are absorbed via passive diffusion as monomers and not as intact mixed micelles [72, 74]. The digestion products dissociate from the mixed micelles in a manner that is not fully understood but probably is associated with the slightly lower pH found near the UWL [74-76]. The majority of the LC FFAs and MGs are re-esterificed in the endoplasmatic reticulum of the enterocyte and assembled into lipoproteins before they are secreted into the lymph whereas MC MGs and FFAs are secreted into the portal blood [77]. A schematic overview of the absorption process is shown in Figure 4.      9     Figure 4 Schematic overview of absorption of digestion products (FAs and MGs) upon digestion of triglycerides. Reprinted from [47] with permission from Elsevier. 1.8 In vivo drug solubilization and absorption from LbDDS The solubility of a drug in the GI fluids is an important determinant for the total amount of drug able to reach the systemic circulation. Especially for poorly water soluble drugs, their low aqueous solubility and/or slow dissolution becomes the rate-limiting step in the absorption process [9, 78]. A variety of approaches can be used to improve the solubility and dissolution rate of poorly water soluble drugs and the choice of formulation is often a major determinant in obtaining the desired bioavailability [47]. For slow dissolving BCS class II drugs formulated as crystalline solid formulations micronization or nanosizing can be used to improve the dissolution rate since a reduction in particle size increases specific surface area and thus the dissolution rate according to the Noyes-Whitney equation [79, 80]. Amorphous formulations with stabilizing co-excipients have proven to be a successful strategy for improving not only the dissolution rate but also the apparent solubility of class II drugs, but often suffer from low physical stability [81, 82]. When using LbDDS the drug will most often be in solution upon administration, however, improved absorption of poorly water soluble drugs from lipid suspensions compared to aqueous suspensions has also been demonstrated [5, 83]. 1.8.1 Drug solubilization in the GI tract The solubility of a drug may be influenced by the pH in the GI tract for ionizable drugs and by the presence of endogenous lipids and surfactants in the GI fluids [84, 85]. Due to the large variation in pH in the GI tract, solubility of weakly acidic and basic drugs will vary considerably during the GI transit [85]. The solubilizing capacity of the GI fluids is also well recognized and is often improved    10  upon ingestion of a (fatty) meal, a situation which is termed a “food effect”. Sunesen and co-workers for example demonstrated a solubility enhancement of DAN in simulated intestinal fluids containing bile salts, phospholipids, and digestion products (FFAs and MGs) compared to water and found that the fed state media had the greatest impact on the solubilization capacity of DAN due to micellar solubilization and solubilization in vesicles [84].  1.8.2 Absorption of poorly water soluble drugs from LbDDS Absorption of poorly water soluble drugs from LbDDS is believed to be governed by the dispersion of the LbDDS, digestion of excipients and co-administered food, and the subsequent solubilization of the drug and lipid digestion products [9]. The current hypothesis for the mechanism of absorption for poorly water soluble drugs from LbDDS is based on the lipid digestion pathway. Following administration of the LbDDS, the formulation is dispersed in the gastric fluids due to gastric contractions, and gastric lipase will start hydrolyzing any digestible components [60]. For SNEDDS the nanoemulsion droplets formed in the stomach are emptied out into the duodenum where the nanoemulsion encounters pancreatic and biliary secretions and lipolysis is continued [59, 65]. The colloidal structures that are formed upon the initial phase of digestion are further digested to swollen mixed micelles containing the drug [70, 71]. This reservoir of drug solubilized in mixed micelles facilitates drug absorption by partitioning of the drug from the micelles to the lumen allowing absorption of the free fraction of drug [9, 78]. In figure 5 a simplified graphic representation of the fate of LbDDS after ingestion is shown both when the drug is in a super-saturated state with the risk of precipitation and when the drug concentration is below Seq.    11   Figure 5 A simplified schematic overview of lipid digestion and drug solubilization in the GI tract following ingestion of a LbDDS. In the stomach the LbDDS is dispersed due to mechanical grinding and digestion. In the duodenum the drug (solubilized in various colloidal structures) will either be absorbed (if the drug concentration is below Seq) or can precipitate (if the drug concentration is below Seq) before being absorbed.    12  1.9 The digestive system in rats Animal models are used in the assessment of drugs and drug delivery systems in research, drug discovery, and drug development in order to evaluate toxicity and efficacy of new drugs and formulations [86, 87]. The general belief is that with careful interpretation results obtained in animal studies, at least to some extent, can be extrapolated to humans [86]. In the research field of LbDDS pigs, dogs, and rats are most commonly used in the evaluation of absorption of drugs from formulations [10-13, 88, 89]. Since the rat is small in size compared to humans dosing of SNEDDS and LbDDS is often performed using an oral gavage with either pre-dispersed formulation or small capsules intended for rats [90, 91]. The fluid volumes in the GI tract of the fasted rat are low hence, pre-dispersion of the formulation may be important in order for an emulsification of the LbDDS to occur in the GI tract or simply for practical reasons of administration [92]. This constitute a clear difference to the intended form of LbDDS administration in humans, which is usually done by means of a soft gelatin capsule. The GI tract of rats differs from humans in several important anatomically and physiologically aspects [93]. The pH in the stomach of rats is relatively high (3.3-5.5) compared to humans [94]. In contrast the pH in the rat small intestine has been reported to be 6.5-6.8  i.e. rather similar to the human intestinal conditions [94]. Unlike humans, rats do not have a gall bladder resulting in continuous secretion of large volumes of dilute bile [93, 94]. Tanaka et al. have measured the concentrations of bile in the small intestine under fasted conditions and found it to be 51 mM which is much higher than what has been reported in humans (see above) [95]. Others have reported values between 9-20 mM which still is high compared to humans and may affect solubilization of poorly water soluble drugs [96]. On the other hand, the catalytic activity of rat pancreatic lipase in the intestine is lower than in humans [97-99].       13  2. The effect of digestion on absorption of drugs from SNEDDS This chapter is based on Paper I : The Effect of Digestion and Drug Load on Halofantrine Absorption from Self-nanoemulsifying Drug Delivery Systems (SNEDDS) and Paper II: Lipase inhibition does not change bioavailability of fenofibrate from SNEDDS and super-SNEDDS. The effect of digestion on absorption from LbDDS is still not clear, and the literature regarding this topic is somewhat contradictory. In a study by de Smidt et al., the effect of digestion and vehicle dispersion on absorption of penclodimine from a series of formulations containing medium chain (MC) triglycerides and tocophersolan (a synthetic water soluble vitamin E derivative) both in varying ratios was evaluated using 1% of the lipase inhibitor orlistat [100]. It was found that the effect of orlistat was related to the droplet size of the emulsion: for the formulation with the lowest particle size (below 200 nm) there was no change in oral bioavailability of penclomidine when adding orlistat. On the other hand there was a slight decrease in oral bioavailability when adding orlistat to the formulations with an intermediate droplet size (710-730 nm). The largest effect of digestion on absorption of penclomidine was seen with the administration of penclomidine in crude MC triglycerides. There was a two-fold drop in the oral bioavailability when orlistat was present and thus for the crude MC emulsion an effect of digestion on absorption of penclomidine was found [100]. In a different study using penclomidine as a model drug no significant differences were found in the bioavailability upon administration of penclomidine in digestible soybean oil or triolein emulsions compared to an indigestible mineral oil emulsion [101]. However, a trioctanoin emulsion resulted in a higher bioavailability of penclomidine compared to the mineral oil emulsion. All formulations had a droplet size of 200-400 nm and the conclusion was, that drug release was partly due to digestion and partly due to diffusional drug release [101]. In contrast, in a study applying the compound 1-cyclopropyl-4-phenyl-6-chlor-2(1H)-quinazolinone (SL-5 12) as a model compound for a poorly water soluble drug dissolved in MC triglycerides and a poorly digestible synthetic lipid N-α-methylbenzyllinoleamide (MBLA) the absorption of the model compound was significantly lower when dosed in the indigestible MBLA compared to the digestible MC triglycerides and the conclusion was that the observed difference was due to the difference in digestibility of the MC triglycerides and MBLA [102]. It is thus of great interest to get a deeper understanding the role digestion plays for absorption of poorly water soluble drugs from SNEDDS and LbDDS.    14  2.1 Methodologies investigating the effect of digestion on absorption of poorly water soluble drugs from SNEDDS In order to investigate the effect of digestion on absorption of poorly water soluble drugs from SNEDDS, and still maintaining GI motility and secretions, orlistat was used to inhibit the digestion in vivo[103, 104]. The concentration of orlistat used in Paper I and Paper II was based on a previous paper by de Smidt et al. where 1 % (w/w) orlistat was used [100]. Two poorly water soluble drugs, HF and FEN were chosen as model compounds due to their high solubility in triglycerides. Furthermore, previous studies have shown that HF precipitates in an amorphous form during in vitro intestinal lipolysis, whereas FEN precipitates in a crystalline form during in vitro intestinal lipolysis [12, 13]. For the in vivo evaluation a rat model was chosen due to the fact that rats are one of the most commonly used species for in vivo evaluation of LbDDS and SNEDDS, but also due to the lower cost and easy accessibility compared to e.g. dogs and pigs. The SNEDDS used in these studies were based on a mixture of LC triglycerides, diglycerides, and mono glycerides, the indigestible surfactant Kolliphor RH 40 and absolute ethanol as a co-solvent. The relative composition is shown in Table 2. The chosen SNEDDS are well characterized with regard to the emulsion droplets size (43.5±0.5 nm) [105] and the same formulation compositions were used for both HF and FEN in order to be able to compare the effects of digestion on absorption for both compounds without confounding formulation effects. Table 2 Composition of SNEDDS. Excipient % (w/w) Soybean oil 27.5 Maisine 35-1 27.5 Kolliphor RH 40 35 Absolute ethanol 10 2.2 Investigation of the effect of digestion on drug absorption In Figure 6, the plasma concentration-time profiles of HF (left) and FEN (right) administered in SNEDDS are shown.  The plasma curves for HF dosed in SNEDDS with and without orlistat indicate that there is an effect of inhibiting digestion on the absorption of HF for example with regards to the Cmax.    15   Figure 6 Plasma concentration curves of HF (left) and FEN (right) following oral administration of SNEDDS with (black) and without orlistat (blue) to rats. Treatments included SNEDDS with HF or FEN (●) and SNEDDS with HF or FEN and 1% orlistat (). n=5-6. However, when looking at the pharmacokinetic data (Table 3) only the Tmax is significantly different between the HF containing SNEDDS with and without orlistat (2.8±1.2 hours and 6.3±1.2 hours respectively) (P<0.05). Cmax seems to be reduced for the SNEDDS containing orlistat (0.3±0.05 µg/mL) compared to the SNEDDS without orlistat (0.51±0.1 µg/mL), however there is no statistical difference (P>0.05).  The area under the curve (AUC) for FEN dosed in SNEDDS with and without orlistat are comparable with no significant differences (P>0.05)1. Table 3 Pharmacokinetic parameters of HF and FEN following oral administration to male SD rats in SNEDDS with and without orlistat. Drug HF FEN Parameter SNEDDS  SNEDDS + Orlistat SNEDDS  SNEDDS + Orlistat Cmax (µg/mL) 0.51±0.1 0.30±0.05 7.36±1.7 6.54±1.1 Tmax (hours) 2.8±1.2a 6.3±1.2a 2.33±1.4 2.92±1.9 t½ 13.5±2.1 21.2±3.9 4.17±0.6 4.97±1.2 AUC(0-30) (hr·µg/mL) 6.26±1.4 6.09±1.4 88.28±20.9 66.3±14.9 Fa (%)* 22.5±6.3 21.9±6.5 N/A N/A Data represents mean±SEM, n=5-6. Numbers with the same letters indicate significant differences (ANOVA with Šídák’s multiple comparison test with α 0.05).  *Absolute bioavailability Fa (%) was determined as follows: Fa=100*(AUCPO·Div)/(AUCiv·DPO)                                                             1 For HF and FEN dosed in super-SNEDDS with and without orlistat the effect of digestion on absorption is comparable to the SNEDDS for both HF and FEN (data not shown, see appendix I and II).    16  2.3 Inhibition of lipase activity In the studies with HF and FEN 1% (w/w) orlistat was used to inhibit the lipase activity. As mentioned, orlistat was chosen on the basis of a publication by de Smidt et al. and in vitro testing using a dynamic in vitro intestinal lipolysis model where the effect of orlistat on lipolysis was assessed [100]. Porcine pancreatic lipase is used in the in vitro model, but is has been shown that orlistat is effective on lipases secreted by the rat [106]. In the rat, the pancreatic duct responsible for the secretion of lipase flows directly into the common bile duct and since the rat has no gall bladder there is a continuous dilute secretion of bile salts into the duodenum [94]. In addition, there is a slow basal flow of lipase continuously being secreted without stimulation [107]. Orlistat inhibits lipase at a 1:1 molar ratio [108] also meaning that with a continuous flow of lipase the inhibitor can potentially be saturated over time and digestion will then subsequently occur. However, whether the depletion of orlistat is a problem with the chosen methodology is unknown and the increased Tmax and decreased Cmax observed with HF dosed in SNEDDS with orlistat is likely to be based on an effect related to the lack of digestion when orlistat is present (Figure 6). 2.4 Drug effect on absorption from SNEDDS The interplay between drug and formulation has an effect on oral absorption and bioavailability of the drug from SNEDDS. The excipients in the SNEDDS have an effect on the emulsification and subsequent droplet size of the nanoemulsion, but also the digestability and the physicochemical characteristics of the drug plays a role with regard to droplet size of the formulation [109, 110]. Due to its high lipophilicity (logP  8.5) HF has the potential of being absorbed lymphatically especially when the drug is administered with a formulation containing LC triglycerides. LC triglycerides promote lymphatic absorption since this type of absorption is dependent on the assembly of chylomicrons in the enterocyte [111-113]. In a study by Porter et al. the authors assessed lymphatic transport of HF in a triple cannulated anesthetized rat model. They used a micellar lipid formulation, an emulsified lipid formulation, and a lipid solution resembling the different lipid digestion stages. It was found that the lymphatic transport of HF was enhanced by using the micellar formulation resembling the later stages of lipid digestion. Although the maximum rates of the lymphatic transport of HF were similar, the peak transport for the micellar system occurred earlier than for the lipid solution [114]. The effect described in the study by Porter et al. is very similar to the effects seen with HF dosed in the SNEDDS with and without orlistat described in this thesis (Figure 6, left). The prolonged Tmax and seemingly lowered Cmax may therefore be due to a    17  decrease in the lymphatic transport when digestion is inhibited or delayed. The same shift of pharmacokinetic parameters is not seen with the SNEDDS containing FEN; however this is expected since FEN not is a lymphatically transported drug due to its lower log P value. In addition FEN is a prodrug that is metabolized to the less lipophilic fenofibric acid in vivo [115].  2.5 Proposed mechanism of drug absorption from SNEDDS when digestion is inhibited The in vivo data strongly suggest that digestion does not play a major role for absorption of HF and FEN from SNEDDS with regards to oral bioavailability or AUC in rats (Figure 6 and Table 3). Several studies have shown that oil solutions and crude emulsions are more dependent on digestion in order to make the drug available for absorption than finer emulsions [100, 116, 117]. When taking the obtained results into account one can therefore speculate whether the increased absorption and bioavailability often observed with SNEDDS is merely a question of keeping the drug in solution during the GI transit and that a direct partitioning from the nanoemulsion droplets facilitated by their large surface area gives rise to increased absorption even when digestion is inhibited (Figure 7 (1)). The traditional view on the importance of digestion may still hold true when digestion is taking place (Figure 7 (2)), but it appears that digestion not is a prerequisite for absorption of poorly water soluble drugs from SNEDDS.    18   Figure 7 Simplified schematic drawing of the proposed mechanism of absorption from SNEDDS without digestion (1) and with digestion (2).  2.6 Summary of results In these studies, the effect of digestion on absorption of the two poorly water soluble drugs HF and FEN was investigated in vivo using the lipase inhibitor orlistat. For the very lipophilic HF a change in pharmacokinetic parameters with a lower Tmax and a seemingly lower Cmax but with no change in oral bioavailability was observed. For the less lipophilic FEN, there was no change in the pharmacokinetic parameters when inhibiting digestion compared to digestion occurring. The differences seen between the two drugs are probably drug specific due to differences in physicochemical properties. However, for both HF and FEN digestion does not seem to be a prerequisite for absorption from SNEDDS.      19  3. The effect of drug load on absorption from SNEDDS This chapter is based on Paper I: The Effect of Digestion and Drug Load on Halofantrine Absorption from Self-nanoemulsifying Drug Delivery System (SNEDDS) and Paper II: Lipase inhibition does not change bioavailability of fenofibrate from SNEDDS and super-SNEDDS. The dose of a drug to be administered in SNEDDS has traditionally been limited by the solubility of the drug in the SNEDDS preconcentrate. Since the solubility of the drug in the preconcentrates often is relatively low compared to the dose, this can result in administration of multiple capsules to the patient in order to obtain the required dose, which can  lead to poor patient compliance [14, 118]. This limitation of SNEDDS has been addressed by the development of super-SNEDDS. In super-SNEDDS, the drug is introduced above the Seq resulting in an increase of the thermodynamic activity of the drug [119]. However, since supersaturated drugs are in a thermodynamically unfavorable state, supersaturated systems have the risk of drug precipitation during storage, in order for the drug to return to a more favorable thermodynamic state [120]. The concept of super-saturation has been studied extensively for transdermal drug delivery systems where it has been shown that the thermodynamic activity of the drug in the formulation is more important than the absolute concentration [20, 21, 121]. Thomas et al. have studied the in vivo performance of super-SNEDDS loaded with HF, FEN, and simvastatin at 150% of Seq in dogs and minipigs. For HF and simvastatin there was a significant increase of oral bioavailability and AUC when dosed in super-SNEDDS compared to SNEDDS in dogs, whereas for FEN the AUC and bioavailability in mini-pigs for SNEDDS and super-SNEDDS were comparable [11-13]. 3.1 Methodologies used to assess the effect of drug load on absorption from SNEDDS In order to investigate the effect of drug load on the absorption from SNEDDS, super-SNEDDS with a drug load of 150% Seq of HF or FEN were prepared and the performance was evaluated in vivo in rats against SNEDDS with a drug load of 75% of Seq. The excipients used were a mixture of LC triglycerides, MGs and diglycerides, Kolliphor RH40 and absolute ethanol. The composition of the super-SNEDDS can be seen in Table 2 (Chapter 2). Thomas et al. have previously shown an increase in the performance of super-SNEDDS loaded with HF compared to SNEDDS in beagle dogs and comparable plasma curves of SNEDDS and super-SNEDDS loaded with HF in minipigs [11-13]. However, since studies in larger animals such as dogs and minipigs are more expensive it    20  was of interest to see whether the before mentioned effects of super-SNEDDS could be evaluated also in a rat model. 3.2 The effect of drug load on absorption from SNEDDS The plasma concentration versus time curves of HF and FEN after oral administration of SNEDDS and super-SNEDDS to rats are shown in Figure 8 and the pharmacokinetic parameters are displayed in Table 4. The doses used for SNEDDS and super-SNEDDS were 6.7 mg/kg and 8 mg/kg for HF and FEN, respectively. Since the drug dose was kept constant, the super-SNEDDS were dosed with half the amount of vehicle compared to the SNEDDS.  Figure 8 Plasma concentration curves of HF (left) and FEN (right) following oral administration of SNEDDS (black) and super-SNEDDS (blue) to rats. Treatments included SNEDDS with HF or FEN (●) and super-SNEDDS with HF or FEN (). n=5-6. From the data in Figure 8 (left) it is apparent that dosing of HF in super-SNEDDS yields a significantly higher Cmax compared to HF dosed in SNEDDS (0.96±0.2 µg/mL and 0.51±0.1 µg/mL, respectively) (P<0.05). The AUC and concomitant bioavailability are seemingly increased for the super-SNEDDS however the increases are not significant (P>0.05). For FEN (Figure 8, right), there is a significant increase in both Cmax and AUC for the super-SNEDDS (13.34±4.7 µg/mL and 148.00±47.5 hr·µg/mL respectively) compared to the SNEDDS (7.36±1.7 µg/mL and 88.28±20.9 hr·µg/mL respectively) (P<0.05).       21  Table 4 Pharmacokinetic parameters of HF and FEN following oral administration to male SD rats in SNEDDS and super-SNEDDS. Drug HF FEN Parameter SNEDDS  super-SNEDDS SNEDDS  super-SNEDDS Cmax (µg/mL) 0.51±0.1a 0.96±0.2 a 7.36±1.7b 13.34±4.7b Tmax (hours) 2.8±1.2 1.3±0.1 2.33±1.4 4.33±2.3 t½ 13.5±2.1 13.5±1.7 4.17±0.6 5.08±1.2 AUC(0-30) (hr·µg/mL) 6.26±1.4 9.14±0.79 88.28±20.9 b 148.00±47.5b Fa (%)* 22.5±6.3 32.9±3.6 N/A N/A Data represents mean±SEM, n=5-6. Numbers with the same letters indicate significant differences (ANOVA with Šídák’s multiple comparison test with α 0.05). *Absolute bioavailability Fa (%) was determined as follows: Fa=100*(AUCPO·Div)/(AUCiv·DPO) 3.3 Supersaturation of drug in SNEDDS increases oral absorption Traditionally, it has been thought that the lipid amount was positively correlated with the absorption of poorly water soluble drugs when dosed in LbDDS such as SNEDDS. Larsen et al. studied the absorption of DAN from Labrafil M2125CS and found a correlation between lower drug loads and higher oral bioavailabilities [89]. However, recent studies with super-saturated SNEDDS indicate that supersaturating the drug, and hence, increasing the thermodynamic activity of the drug in the SNEDDS increases the oral bioavailability (Figure 8 and Table 4) [10-13]. Gao et al. have used a slightly different approach and have prepared supersaturable SEDDS (S-SEDDS) with poorly water soluble drugs containing polymers as precipitation inhibitors [122-124]. Their strategy was to generate super-saturation upon dispersion of the SEDDS, but with drug concentrations below Seq in the preconcentrate. They found that the addition of 5% HPMC as a precipitation inhibitor to the SEDDS resulted in a 20-fold increase in Cmax and a 10-fold increase in oral bioavailability compared to the SEDDS without HPMC in rats [122]. Whether super-SNEDDS induce supersaturation in vivo in the GI tract is unknown however, Williams et al. have suggested that the increased bioavailability of poorly water soluble drugs when administered in SNEDDS is due to supersaturation in vivo [125]. It would be of great interest to combine the concepts of super-SNEDDS and S-SEDDS in future studies. Micellar solubilization could also be a possible explanation for the superior performance of the super-SNEDDS compared to the SNEDDS presented in Figure 8 and Table 4. High amounts of surfactants have shown to decrease absorption of poorly water soluble drugs probably due to    22  entrapment of the drug in micelles resulting in a reduction of free drug [126]. In a study by Berthelsen et al., the in vivo performance of FEN administered in varying ratios of the surfactant Kolliphor RH40 was studied. It was found that the oral bioavailability of FEN in 25% (w/v) Kolliphor RH40 was lower than in 15% (w/v) Kolliphor RH40. The authors speculate that increasing surfactant concentration increased micellar solubilization and hence decreased the free drug concentration available for absorption [126]. In the SNEDDS, the surfactant to drug ratio is higher compared to the super-SNEDDS and thus some of the drug in the SNEDDS could be trapped in Kolliphor RH40 micelles and hence, lead to a lower concentration of free drug and thus a lower drug absorption compared to the super-SNEDDS. 3.4 Summary of results Super-SNEDDS loaded with 150% of Seq of either HF and FEN demonstrated an increased in vivo performance when tested against SNEDDS in a rat model. For HF, only Cmax was significantly increased with a possible increase in bioavailability although this effect was not significant. For FEN, both Cmax and AUC were increased when dosed in super-SNEDDS compared to SNEDDS. The mechanism behind the increased absorption of HF and FEN in the super-SNEDDS is probably due to the increased free drug levels close to the absorptive barrier and an increase of thermodynamic activity.      23  4. Dynamic intestinal in vitro lipolysis in the evaluation of SNEDDS This chapter is based on Paper I: The Effect of Digestion and Drug Load on Halofantrine Absorption from Self-nanoemulsifying Drug Delivery System (SNEDDS) and Paper II: Lipase inhibition does not change bioavailability of fenofibrate SNEDDS and super-SNEDDS. The dynamic in vitro lipolysis model is an established model for simulating the complex processes taking place in the GI tract upon administration of LbDDS and has been used in the assessment of digestibility and ability of LbDDS to keep the drug solubilized during the GI transit [127-129]. Most often the dynamic in vitro lipolysis model is mimicking the conditions in the small intestine. A biorelevant medium simulating the intestinal fluids with regard to the presence of bile salts, phospholipids, and pH is used as the continuous phase. For initiating lipolysis, pancreatic lipase of porcine origin is typically added since porcine pancreatic lipase is sn-1 and sn-3 specific and therefore comparable with human pancreatic lipase [61, 127-130]. Upon hydrolysis of triglycerides by pancreatic lipase, the released FAs are titrated with sodium hydroxide, and in order to be able to control the rate of lipolysis, Zangenberg et al. developed the dynamic in vitro lipolysis model where calcium chloride is added continuously [127, 128]. In other similar digestion models calcium chloride is added as a fixed dose before initiation of the experiment [30, 131]. The FAs released upon digestion of triglycerides inhibit the pancreatic lipase and the calcium chloride binds to the FAs resulting in precipitation of the formed calcium soaps [127, 132]. Calcium ions also have been implicated in the mechanism of action of pancreatic lipase as a modulator of the velocity of hydrolysis [133]. It is possible to sample from the digestion medium during the course of the lipolysis to determine in which phases the drug is found after centrifugation and to analyze the solid state form of the drug in the pellet [128]. Typically, the solubilized drug is assumed to be the fraction available for absorption whereas the precipitated fraction found in the pellet would need to be re-dissolved in order to be absorbed [4]. However, Sassene et al. emphasized, the importance of the solid state characteristics of the precipitated drug when analyzing the precipitate after dynamic in vitro lipolysis of a SMEDDS containing cinnarizine. During the digestion, 59% of the drug precipitated but the dissolution rate of the cinnarizine in the pellet was found to be 10 times higher than of a cinnarizine spiked pellet sample. XRPD and PLM analysis revealed that no crystalline cinnarizine was present in the pellet suggesting that the precipitated cinnarizine was amorphous [134]. Examining the pellets by 13C-NMR, 1H-NMR and DSC, interactions between the nitrogens of cinnarizine and the carboxylic groups of the FAs generated during the digestion were found    24  resulting in an amorphous cinnarizine/FA precipitate  [135]. The solid state characteristics of other poorly water soluble drugs have also been investigated and HF, carvedilol, cinnarizine, and simvastatin were found to precipitate in an amorphous form in the dynamic in vitro lipolysis model, whereas loratidine, FEN, and DAN were found to precipitate in a crystalline form [4, 11-13, 22, 134, 136]. 4.1 Methodologies used to evaluate SNEDDS and super-SNEDDS in vitro A biorelevant intestinal medium was used to mimic the intestinal fluid. The medium was composed of bile salts, phospholipids, buffer and sodium chloride for adjusting the ionic strength. The pH was adjusted to 6.5 (see Table 5 for composition). Table 5 Composition of biorelevant intestinal medium simulating intestinal fluid.   The composition and pH of the biorelevant intestinal medium is based on values found in humans [63, 127]. In order to improve the simulation of the conditions in the rat intestine, the volume of the simulated intestinal medium used in the lipolysis model was scaled to be approximately 6 times larger than the fluid volume found in the fasted rat intestine and the amount of formulation used was scaled accordingly and was thus six times larger than the dose used in the in vivo studies [92]. This is done in order to increase the chance of getting a good correlation with the in vivo data. However, since literature on the exact composition of the intestinal fluids in rats is scarce, the composition of the biorelevant medium in the model was kept as a human intestinal model. 4.2 Dynamic in vitro lipolysis – the effect of digestion The effect of digestion on the solubilizing capacity of SNEDDS containing HF and FEN was investigated in vitro using the dynamic in vitro lipolysis, as described above. In Figure 9 the Excipient Concentration (mM) Bovine bile salt 2.5 Phospholipids 0.26 Tris 2 Maleic acid 2 Sodium chloride 50    25  amount of solubilized and precipitated HF in SNEDDS (a and b) and FEN in super-SNEDDS (c and d) with and without orlistat is shown.   Figure 9 Relative amount of drug found in the aqueous phase (grey) and the pellet phase (black) after 60 minutes of dynamic in vitro intestinal lipolysis of a) SNEDDS containing HF, b) SNEDDS containing HF with 1 % orlistat, c) super-SNEDDS containing FEN, and d) super-SNEDDS containing FEN with 1 % orlistat. * indicate significant difference (α=0.05). n=3 and data represents mean +/- SEM. It is evident that upon digestion of SNEDDS and super-SNEDDS containing HF and FEN respectively (Figure 9 a and c) drug precipitation increases over time. Upon addition of 1% orlistat, precipitation is lowered for both SNEDDS and super-SNEDDS. SNEDDS (and super-SNEDDS) have the ability to keep the drug in solution in the nanoemulsion droplets. After initiation of digestion by addition of lipase, the digestible components of the formulation are hydrolyzed and various colloidal structures evolve. Small angle x-ray scattering (SAXS) and cryogenic transmission electron microscopy (cryo-TEM) have been used to gain insight into the structural changes in the colloids occurring upon digestion in the small intestine [137-141].  The biorelevant intestinal medium contains micelles consisting of bile salts and phospholipids both also present as endogenous compounds in vivo [140]. With the introduction of lipolytic end products such as FFAs and MGs, vesicles and other colloidal structures start evolving [140]. Fatouros and co-workers have investigated the structural development of SNEDDS during in vitro digestion using SAXS and found that lamellar phases appear shortly after initiation of lipolysis with hexagonal phases evolving over time [137]. More recently, Warren et al. investigated the real-time formation of the colloidal structures developing during digestion of LC-SNEDDS by coupling a lipolysis model to a synchrotron-SAXS set-up. They also detected lamellar and inverse hexagonal phases (H2-phases) during 50 minutes of lipolysis [141]. Since the initial emulsion droplets only persist for a short time frame in vivo, understanding of the different structures evolving during digestion is important to    26  understand changes in solubilization capacity during digestion. The colloidal structures have also been implicated in the trafficking of the drugs from the lumen to the absorptive site [9, 78]. If the colloidal structures formed during the digestion process have a lower solubilization capacity for the drug compared to the initial formulation (preconcentrate), the drug is at risk of precipitating. The addition of 1% orlistat reduces digestion significantly and thus precipitation is also decreased. When digestion is inhibited the emulsion droplets are thought to persist and the loss of solubilization capacity is avoided (Figure 9).     4.3 Dynamic in vitro lipolysis – the effect of drug load In Figure 10, the amount of solubilized and precipitated HF (a and b) and FEN (c and d) in super-SNEDDS is shown.   Figure 10 Relative amount of drug found in the aqueous phase (grey) and the pellet phase (black) after 60 minutes of dynamic in vitro intestinal lipolysis of a) SNEDDS containing HF, b) super-SNEDDS containing HF, c) SNEDDS containing FEN, and d) super-SNEDDS containing FEN. * indicate significant difference (α=0.05). n=3 and data represents mean +/- SEM. For both HF and FEN, when dosed in super-SNEDDS, the amount of precipitated drug after 60 minutes of lipolysis is significantly higher than from the SNEDDS. The continuing lipid digestion is reflected in a gradual decrease in solubilization of HF and FEN for both SNEDDS and super-SNEDDS. The amount of surfactants and type of formulation used has been shown to play a large role for the solubilization of FEN due its high lipophilicity, and since it is a neutral drug and therefore does not have a pH dependent solubility. In contrast, pH plays a larger role for HF solubility since it is a weak base, despite of its very high lipophilicity [140]. However, since pH was kept constant at 6.5 in the dynamic in vitro lipolysis experiments to reflect the intestinal conditions, the pH dependent solubility for HF is not reflected in these experiments. Since the dose was kept    27  constant for both HF and FEN in SNEDDS and super-SNEDDS, the lipid to drug ratio is higher for the SNEDDS than for the super-SNEDDS and thus the solubilization capacity is also higher for the SNEDDS. During digestion, the SNEDDS will also yield more digestion products that can facilitate the solubilization of HF and FEN.  Thomas et al. have also looked into the in vitro behavior of super-SNEDDS loaded with HF and FEN using digestion models [12, 13]. For HF they found that SNEDDS and super-SNEDDS provided similar solubilization capacity in vitro [12]. For FEN they found that SNEDDS had a much larger solubilization capacity than super-SNEDDS keeping approximately three times as much FEN solubilized after 60 minutes of lipolysis [13]. The differences between the data presented in Figure 10 and the data presented by Thomas et al. can most likely be ascribed to the differences in the in vitro models, for example with respect to the volume of media used, formulation to media ratio, and the addition of a gastric step in the FEN study [12, 13].   4.4 Optimizing the in vitro lipolysis model Most in vitro lipolysis models reflect the intestinal milieu found in humans with regard to the composition of the simulated intestinal fluids and lipase activity [4]. Anby et al. proposed some modifications of the human intestinal in vitro lipolysis model to align it with the events taking place in the digestive tract of rats and obtained IVIVC between an in vivo study in rats and the modified in vitro rat lipolysis model [99]. Instead of using pancreatic lipase of porcine origin they used ex vivo pancreatic-biliary secretions from rats. However, since this for practical reasons is a complicated procedure they also measured the lipase activity in the pancreatic-biliary secretions and found the activity to be  200 TBU/mL. This is much lower than the activity used in models mimicking digestion in humans ( 550 USP U/mL) [99, 127]. Also Tønsberg et al. have investigated the lipase activity in rats and found a similar activity of 153 U/mL [142]. In the set-up used for the experiments in Figure 9 and 10, the lipase activity was not adjusted based on the lipase activity values reported in rats [99, 142]. The volume however, was adjusted based on the lower volumes found in the GI tract of the rat instead of the traditionally higher volumes used in human intestinal models. In order to titrate the system, the model was scaled to approximately 6 times the fluid volumes found in the rat to a final volume of 30 mL and the doses of the formulations were also increased to 6 times the amount used in vivo [92].     28  4.5 Summary of results Precipitation from SNEDDS and super-SNEDDS can be lowered when digestion is inhibited, since digestion is affecting the solubilization capacity of the drug in SNEDDS and super-SNEDDS. The drug precipitation from super-SNEDDS is more pronounced since the drug is in an unfavorable thermodynamic state but also due to the lower lipid to drug ratio resulting in a lower solubilization capacity compared to the drug in SNEDDS. The in vitro model provides a tool for investigation of the digestibility and solubilization capacity of SNEDDS in vitro however, species specific modifications should be considered in order to have predictable models.      29  5. Drug precipitation in vivo  This chapter is based on Paper III: In vivo precipitation of poorly soluble drugs from lipid based drug delivery systems The dynamic in vitro lipolysis model simulates the GI environment and is a method used to evaluate the ability of LbDDS to keep the drug in solution during digestion [4]. Details about the model were discussed in Chapter 4. The current hypothesis regarding this model is that the solubilized fraction is readily available for absorption and several studies have shown a positive correlation between the amount of solubilized drug in the dynamic in vitro lipolysis model and the oral bioavailability in vivo [4, 9, 22]. Cuiné et al. investigated a series of microemulsions consisting of LC glycerides and Kolliphor EL containing DAN both in vitro and in vivo in beagle dogs [143]. Here a rank order correlation between the amount of solubilized DAN in vitro using a lipolysis model and the bioavailability in vivo was found [143]. However, the positive correlation between solubilized drug and bioavailability was questioned by Thomas et al. when introducing super-SNEDDS [11-13]. In in vitro studies precipitation was more pronounced for super-SNEDDS compared to SNEDDS for both HF, FEN and simvastatin and yet this was not reflected in a lower bioavailability [11-13]. The solid state properties of the amorphous precipitated HF were used to explain the findings. Even for the crystalline precipitating FEN the increased precipitation seen for the super-SNEDDS compared to the SNEDDS was not reflected in the in vivo studies where the bioavailability of the SNEDDS and super-SNEDDS were comparable [13]. To which extent precipitation of poorly soluble drugs occurs in vivo when dosed in LbDDS is to date unknown and is therefore highly interesting to investigate in order to get predictive in vitro models. 5.1 Methodologies used to determine in vivo precipitation In order to assess precipitation from LbDDS in vivo studies in rats were carried out. LbDDS I is a SNEDDS and was included in order to have a relevant SNEDDS in the study.  LbDDS II is a type IV formulation in the LFCS and is composed purely of surfactants. Previous studies have shown that this type of formulation is poor in keeping drugs in solution upon dispersion [47]. The compositions of the formulations are shown in Table 6. In order to be able to detect crystalline precipitates in the GI tract, FEN and DAN were chosen as model drugs since both drugs have been shown to precipitate crystalline in vitro [4, 13]. In this experiment it is important to know the gastric emptying time in order to be able to predict the best time points for finding precipitated drug in the    30  GI tract. A study with paracetamol dissolved in LbDDS I and II was used to evaluate the gastric emptying time (data not shown) since paracetamol is readily absorbed when reaching the small intestine [144, 145]. The rats were dosed with 300 µL LbDDS I or II containing either FEN or DAN and euthanized after 5 or 90 minutes for LbDDS I and 5 or 30 minutes for LbDDS II. The GI content was subsequently removed from the stomach and the small intestine and analyzed for crystalline precipitated drug by X-ray powder diffraction (XRPD) and polarized light microscopy (PLM). Table 6 The relative composition of LbDDS I and II. LbDDS I LbDDS II Excipient % (w/w) Excipient % (w/w) Sesame oil 20.6 Transcutol HP 50 Kolliphor RH40 45 Kolliphor EL 50 Oleic acid 15.4 - - Brij 97 9 - - Absolute ethanol 10 - -  5.2 Drug precipitation in vivo XRPD can detect order in a crystal lattice resulting in diffractograms with drug specific reflections [146]. PLM can detect birefringence even if only very few crystals are present in the sample. In Figure 11, the diffractograms (a) and micrographs (b) of the GI content after oral administration of LbDDS I containing DAN are shown. No crystalline DAN was detected in the stomach or the small intestine of the rats at any time points. The birefringence seen on the micrographs corresponds to the inherent birefringence also seen in blank samples (see appendix 3).    31   Figure 11 XRPD diffractograms (a) and PLM micrographs (b) of GI content of rats after administration of DAN in 300 µL LbDDS I. I: reference diffractogram/micrograph of crystalline DAN, II: stomach (5minutes), III: stomach (90 minutes), IV: intestine (5 minutes), and V: intestine (90 minutes). In Figure 12 diffractograms (a) and micrographs (b) of the GI content after oral administration of LbDDS I containing FEN are shown. After 90 minutes crystalline FEN was detected in the stomach by XRPD and PLM, whereas no crystalline FEN was detected in the small intestine.       32   Figure 12 XRPD diffractograms (a) and PLM micrographs (b) of GI content of rats after administration of FEN in 300 µL LbDDS I. I: reference diffractogram/micrograph of crystalline FEN, II: stomach (5minutes), III: stomach (90 minutes), IV: intestine (5 minutes), and V: intestine (90 minutes). In Figure 13 diffractograms (a) and micrographs (b) of the GI content after oral administration of DAN in LbDDS II are shown. Crystalline precipitated DAN was detected by XRPD and PLM in the stomach after 5 minutes and traces could still be detected after 30 minutes whereas no precipitate was detected in the small intestine. From the micrographs (b) it is evident that DAN crystals were present, however, the morphology seems to have changed when comparing micrographs II (stomach after 5 minutes)  and III (stomach after 30 min) with micropgraph I (crystalline DAN).    33   Figure 13 XRPD diffractograms (a) and PLM micrographs (b) of GI content of rats after administration of DAN in 300 µL LbDDS II. I: reference diffractogram/micrograph of crystalline DAN, II: stomach (5minutes), III: stomach (30 minutes), IV: intestine (5 minutes), and V: intestine (30 minutes). In Figure 14 diffractograms (a) and micrographs (b) of GI content after oral administration of FEN in LbDDS II are shown. FEN was detected by XRPD and PLM in the gastric content after 5 minutes and after 30 minutes higher amounts of precipitated FEN were present. The FEN crystals seen in the micrographs (IIb and IIIb) seem to have a changed morphology compared to crystalline FEN used as the starting material (Ib).    34   Figure 14 XRPD diffractograms (a) and PLM micrographs (b) of GI content of rats after administration of FEN in 300 µL LbDDS II. I: reference diffractogram/micrograph of crystalline FEN, II: stomach (5minutes), III: stomach (30 minutes), IV: intestine (5 minutes), and V: intestine (30 minutes). Prior to evaluating LbDDS in vivo, the solubilization capacity will most often have been determined in in vitro digestion studies where also the drug precipitation is assessed. However, to date no studies has been conducted evaluating to which extent poorly water soluble drugs dosed in LbDDS precipitate in vivo. No precipitation was detected in the stomach or small intestine after oral administration of DAN in LbDDS I (Figure 11). This was unexpected since LbDDS I contains 20.6% sesame oil and hence, upon administration will be subjected to digestion by gastric and pancreatic lipases. Previous studies with SNEDDS such as LbDDS I, have shown that upon digestion of the lipids the formulation loses its solubilizing capacity and hence the risk of precipitation of the drug is increased [88]. Precipitation was not seen in either the diffractograms or the micrographs such that possible issues with limit of detection (LOD) can (within reason) be ruled out as a reason for the lack of detection of crystalline material. The LOD for XRPD is around 1% of drug in the examined matrix however, for PLM the LOD is much lower since only a single crystal has to be present in the investigated sample order to detect it [147]. For FEN administered in LbDDS I precipitation was detected after 90 minutes in the stomach and this was probably due to a loss in solubilization capacity after gastric digestion of the sesame oil in the formulation and delayed gastric emptying as described above (Figure 12). No precipitation was detected in the small intestine which again was surprising since precipitation was found during intestinal in vitro lipolysis (data not shown).    35  When administering LbDDS II with either FEN or DAN (Figure 13 and Figure 14) precipitation was detected in the stomach 5 minutes after administration and was still present 30 minutes after administration to the rats. Since LbDDS II is a type IV formulation according to the LFCS, precipitation from this type of formulation was expected. However, unexpectedly no precipitation was observed in the small intestine for either FEN or DAN. LbDDS II consists of the surfactant Kolliphor EL and the co-solvent Transcutol HP. Trancutol HP is a hydrophilic co-solvent that is expected to partition into the GI fluids of the rat and hence, the formulation quickly loses its solubilization capacity of the drug [5, 148, 149].  For both FEN and DAN precipitation was absent in the small intestine for both LbDDS I and LbDDS II. Since most in vitro dynamic lipolysis models simulate the milieu in the small intestine these data could be used to explain the lack of IVIVC seen with the model. The lack of precipitation in the small intestine could be due to several factors such as gastric emptying and the high amount of bile salts found in the rat intestine. The gastric emptying is especially interesting since it is a gradual process. Absorption of most drugs takes place in the small intestine, and the lack of precipitation could be due to the fact that the drug is quickly absorbed when reaching the small intestine and that due to the gradual stomach emptying supersaturation is thus not achieved. Further, the dynamic in vitro lipolysis model is mimicking digestion in humans, whereas precipitation was investigated in rats. There are several differences between the GI tract in humans and rats such as the lack of a gall bladder in rats resulting in a continuous flow of dilute bile [93, 94].  Studies of precipitation in humans are rare, but Hens et al. performed a study with posaconazole (a weak base) suspensions with different pH values (1.6 and 7.1 respectively) and a solution (pH 1.6) in humans in order to investigate the gastrointestinal interplay between dissolution, supersaturation and precipitation of the drug [150]. In the stomach, posaconazole was mostly dissolved, when dosed in the acidic suspension, whereas more precipitate was found when the neutral suspension was administered. The solution kept the drug dissolved during the transit time in the stomach. However, the authors found significant intestinal precipitation after intragastric administration of both suspensions and the solutions [150]. Since posaconazole is a weak base the solubility behavior of the drug is pH dependent and the drug is more soluble at lower pH values [151]. It should be stressed however, that no lipids were added to the formulations in this clinical study. It would be interesting to investigate precipitation from LbDDS in humans in order to compare the results obtained in rats in the current study. Based on the presented results (Figure 11-14) it is evident that    36  precipitation occurs in the stomach of the rat. The reason for the lack of precipitation in the small intestine is probably related to the gradual transfer of formulation from the stomach to the intestine and thus the gastric emptying rate is a determinant for the concentration of drug in the intestine. Therefore in order to obtain IVIVC it can be speculated that a combined gastric-intestinal lipolysis model with a gradual gastric emptying step should be used to assess LbDDS in vitro. 5.3 Summary of results The precipitation behavior of FEN and DAN was investigated in vivo in two different LbDDS. No precipitation of DAN was evident upon administration of LbDDS I in the GI tract of rats. FEN precipitation was found to occur in the stomach after 90 minutes and it was concluded that the precipitation was a result of a loss of solubilization capacity due to digestion of the lipid excipient. Both FEN and DAN precipitation was detected in the stomach after 5 minutes and after 30 minutes upon administration of the drugs in LbDDS II. LbDDS II is a LFCS type IV formulation and a loss of solubilization capacity occurs due to partitioning of the hydrophilic co-solvent into the aqueous phase. In vivo precipitation studies in rats or other mammals could be used to highlight the importance of the gradual gastric emptying and digestion of LbDDS on precipitation behavior of poorly water soluble drugs. The current in vitro models need to be improved in terms of mimicking the entire GI transit in order to reflect the in vivo situation which will be discussed in more detailed in the next chapter.      37  6. The rat as a model for evaluating super-SNEDDS in vivo and obtaining a relationship between in vitro and in vivo data In this chapter the presented results (Chapter 2-5) will be discussed in relation to each other.  Dogs and pigs have some advantages compared to rats when using them in the in vivo evaluation of SNEDDS, the most prominent being that SNEDDS and other LbDDS can be dosed in relevant lipid doses for humans in pigs as well as dogs [11, 13, 88]. Super-SNEDDS have previously only been evaluated in dogs and pigs while the data presented in the present thesis for the first time evaluates super-SNEDDS in rats [10-13]. Thomas et al. have evaluated the effect of drug load on oral absorption from SNEDDS and super-SNEDDS with both HF and FEN in dogs and pigs, respectively [12, 13]. When comparing the data from the in vivo study with HF dosed in SNEDDS and super-SNEDDS tested in rats (Figure 8 and Table 4) with the study performed in beagle dogs the results are comparable. HF dosed in super-SNEDDS resulted in a significantly higher Cmax, AUC, and absolute bioavailability compared to SNEDDS [12]. In the rat study HF dosed in super-SNEDDS resulted in a significant increase in Cmax and a seemingly higher AUC and bioavailability, although not significant, compared to SNEDDS [10]. Thus, the results seem to be comparable across species for halofantrine dosed in SNEDDS and super-SNEDDS. However, when comparing the in vivo study with FEN dosed in SNEDDS and super-SNEDDS tested in rats (Figure 8 and Table 4) with the study performed by Thomas and coworkers in minipigs some differences are seen. In minipigs, the Cmax, AUC, and relative bioavailability of FEN dosed in SNEDDS and super-SNEDDS were comparable [13]. In rats, on the other hand, there was a significant increase in Cmax and AUC for FEN dosed in super-SNEDDS compared to SNEDDS. Since the formulations used in the FEN studies were comparable, the differences in the pharmacokinetic parameters are likely to be species dependent. Whilst the absorption of drugs across the intestinal barrier (at least in the absence of active transport) is thought to be similar across species, interspecies variation in pH, first-pass metabolism, and other physiological/anatomical differences may give rise to species specific drug absorption and bioavailability [86]. Since FEN is a neutral molecule, the observed differences are not due to pH dependent solubility but increased solubilization due to high bile concentrations in the rat or effects related to the delayed gastric emptying time in pigs could potentially be responsible for the observed differences [86].      38  When evaluating SNEDDS and other LbDDS in vitro, the goal is to be able to predict the behavior in vivo either by establishing a level A correlation (full correlation between in vitro and in vivo data as defined by FDA) or by obtaining a rank-order correlation [4, 152]. However, the establishment of a predictive lipolysis model is a complex task in which many physiological parameters have to be taken into account [152]. The physiological factors are especially hard to mimic in models, since the physiological response changes continuously depending on e.g. the prandial state, gastric emptying rate, and intestinal residence time [152]. Using the solubilized fraction of drug in the dynamic in vitro lipolysis model as the fraction available for absorption and not considering the precipitate has proven to be misleading in some instances. When looking at the plasma curves of super-SNEDDS with HF and FEN compared to the SNEDDS in Figure 8 and comparing them to the in vitro lipolysis data in Figure 10 it is clear that in vitro solubilization is not a suitable parameter in the evaluation of bioavailability, when using the in vitro lipolysis model. The same trend is seen in the papers from Thomas et al. despite some differences in the in vitro model where precipitation from super-SNEDDS is significantly larger than from SNEDDS in vitro but the bioavailability of super-SNEDDS is equal or better compared to SNEDDS [11-13]. In fact, in the case of super-SNEDDS, increased precipitation was inversely correlated to the bioavailability. The reason for these conflicting results may be an oversimplification of the dynamic in vitro lipolysis model compared to the actual in vivo conditions. One major difference between the in vitro model and the in vivo situation is that that drugs (and lipids) are absorbed in vivo. Due to the lack of an absorption step in the dynamic in vitro lipolysis model, in vivo precipitation may be overestimated in vitro. Different strategies such as biopharmaceutical modeling, using mimicked digested systems in perfusion models, using silicone disks to absorb FAs, and ex-vivo absorption models have been used in order to provide an absorption step to the in vitro lipolysis model in the quest for obtaining IVIVC [116, 153-155]. Crum et al. have addressed this issue by coupling an intestinal perfusion model with an in vitro intestinal lipolysis set-up [156]. This allow the authors to follow the digestion process in vitro while assessing the impact of in vitro digestion on absorption of lipid digestion products and drug in vivo. They compared a type IIIA, a type IIIB, and a type IV formulation containing FEN and found that while the dynamic in vitro lipolysis showed significant differences in solubilization capacity upon digestion, absorption when using the newly developed model was similar for all formulations [156]. The in vitro data produced by Crum and coworkers using the newly developed model showed good    39  correlation with in vivo data of FEN in similar formulations obtained in pigs [156, 157]. The model proposed by Crum et al. is very elaborate and is not a model that can be used in regular assessment of SNEDDS or other LbDDS.  To which extent precipitation from LbDDS occurs in vivo is debatable. When evaluating data from in vitro lipolysis the different phases are analyzed for drug content as seen in Figure 9 and Figure 10. However, when taking the data from Chapter 5 into account the current paradigm of how to evaluate lipolysis data is questioned. In vivo the LbDDS traverse both the gastric and the intestinal environment whereas most dynamic in vitro lipolysis models only mimic the conditions in the small intestine with regard to bile salts and phospholipids and utilize pancreatic lipase of porcine origin [4]. Since almost 20% of the digestion taking place in vivo is attributed to gastric lipase combined gastric-intestinal models have been developed [158]. There is a lack of commercially available gastric lipases and therefore microbial lipases have been utilized instead [13, 158, 159]. Christophersen et al. used a lipase from Candida. antarctica in lieu of a recombinant/aspirated gastric lipase in a gastro-intestinal lipolysis model and found a correlation with the in vivo bioavailability of SNEDDS loaded with cinnarizine in beagle dogs in the fasted state [158, 160]. As already discussed precipitation may be overestimated in vivo due to absorption. But since precipitation in Chapter 5 was only apparent in the stomach further development of gastro-intestinal lipolysis models may provide more insight into the role of precipitation on bioavailability and may help to establish IVIVC. In the development of a gastro-intestinal lipolysis model species specific differences in the composition of the media (both gastric and intestinal) should be taken into account with regard to bile salt levels, pH, and lipase activity. Most studies at present compare human in vitro models with in vivo studies performed in animals. Also the role of the gradual gastric emptying should be explored in the further development of gastro-intestinal lipolysis models, since this could be part of the explanation for the lack of intestinal precipitation in Chapter 5.    40  7. Conclusion The overall purpose of this thesis was to evaluate the effect of digestion and drug load on oral absorption of poorly water soluble drugs from SNEDDS. In addition, the purpose was to evaluate the rat as an in vivo model for testing super-SNEDDS and evaluate the predictivity of the dynamic in vitro lipolysis model compared to the rat studies.   SNEDDS and other LbDDS provide a relatively simple formulation strategy to improve the oral bioavailability of poorly water soluble drugs belonging to class II of the BCS. The exact mechanism behind the increase in oral bioavailability often seen for SNEDDS and other LbDDS compared to conventional solid dosage forms is largely related to the fact that the drug is dosed in solution. Digestion was though to play a major role for absorption from SNEDDS due to various colloidal species being formed in the digestion process. However, the studies performed in this thesis indicate that digestion may not be crucial if the dispersion of the SNEDDS results already in vary fine particles. Thus, the increased oral bioavailability often seen with SNEDDS may simply be related to the large surface area provided by the nanoemulsion containing the drug in solution. The nanoemulsion droplets function as a reservoir from where the drug can partition into intestinal fluids for further absorption. If the initial dispersion of the LbDDS leads to coarse emulsions the digestion process merely serves to create smaller particles that then fulfill the same role as the nano-droplets formed from SNEDDS. Super-SNEDDS provide a good alternative to SNEDDS for increasing the oral bioavailability even further without increasing the number of capsules. The exact mechanism behind the benefits associated with super-SNEDDS is unknown however; it is likely to be linked to the increased thermodynamic activity of the drug in the nanoemulsion droplets (i.e. also in the absence of digestion). The increased drug concentration in the nanoemulsion droplets provides a concentration gradient between drug in these droplets and the free drug in the lumen leading to an increase in the absorbable fraction. The current dynamic in vitro lipolysis model may over-predict precipitation resulting in difficulties establishing IVIVC. Precipitation of poorly water soluble drugs dosed in LbDDS in vivo in rats is predominantly occurring in the stomach and not in the small intestine as previously thought. The current models mainly simulate intestinal lipolysis but with the present knowledge the need for improving the current models has been elucidated.    41  Rats seem to be a feasible animal model to use in the evaluation of super-SNEDDS in spite of the physiological and anatomical differences between rodents and humans. However, in order to obtain IVIVC the in vitro lipolysis models should be changed into representative models of the digestive system of rats.      42  8. Future Perspectives In order to understand the mechanism behind the increased drug absorption seen when using LbDDS, such as SNEDDS, a better understanding of the GI processes such as gastric emptying, GI transit time, and GI fluid composition in humans and animals is needed. In the present thesis the role of digestion of SNEDDS on absorption of poorly water soluble drugs was shown to be less important than previously though. However, further studies with increasing doses of orlistat and by preparing SNEDDS with indigestible lipids, such as olestra, could potentially further elucidate the mechanisms behind the increased absorption often seen when administering poorly water soluble drugs in SNEDDS.  Super-SNEDDS have proven to be an interesting strategy for improving absorption of poorly water soluble drugs and more research in this area is needed. Studies with different degrees of supersaturation in vitro and in vivo could be useful in learning more about the effect of super-saturation on oral absorption. Also, investigating maximum degrees of supersaturation and how the ability for supersaturating a given drug is related to the intrinsic properties of the drug compound could prove valuable in order to assess which drugs are suitable for supersaturation in SNEDDS. In order to obtain a better of IVIVC the current lipolysis model should be further developed. To date, an absorption step is lacking in the models which may prove to be essential in obtaining predictable models. The importance of adding an absorption step which can be used in an everyday laboratory setting should be stressed since some of the elaborate solutions that have previously been proposed may be too complex to use on a general basis. Also using combined GI models simulating both the gastric and intestinal compartment would be relevant especially considering the data obtained from the in vitro precipitation study. The in vitro models should be species specific since most in vitro data is compared with in vivo studies in animals. Therefore species specific models should be developed taking differences in the intestinal fluid composition and lipase activity into account, but also differences in gastric emptying and transit time should be considered.        43  9. References 1. Fahr, A. and X. Liu, Drug delivery strategies for poorly water-soluble drugs. Expert opinion on drug delivery, 2007. 4(4): p. 403-416.  2. Lipinski, C.A., et al., Experimental and computational approaches to estimate solubility and permeability in drug discovery and development settings. Advanced Drug Delivery Reviews, 2012. 64, Supplement: p. 4-17.  3. O'Driscoll, C.M. and B.T. Griffin, Biopharmaceutical challenges associated with drugs with low aqueous solubility—The potential impact of lipid-based formulations. Advanced Drug Delivery Reviews, 2008. 60(6): p. 617-624.  4. Larsen, A.T., P. Sassene, and A. Müllertz, In vitro lipolysis models as a tool for the characterization of oral lipid and surfactant based drug delivery systems. International Journal of Pharmaceutics, 2011. 417(1–2): p. 245-255.  5. Mu, H., R. Holm, and A. Müllertz, Lipid-based formulations for oral administration of poorly water-soluble drugs. International Journal of Pharmaceutics, 2013. 453(1): p. 215-224.  6. Thomas, N., T. Rades, and A. Müllertz, Recent developments in oral lipid-based drug delivery. Journal of Drug Delivery Science and Technology, 2013. 23(4): p. 375-382.  7. Pouton, C.W. and C.J.H. Porter, Formulation of lipid-based delivery systems for oral administration: Materials, methods and strategies. Advanced Drug Delivery Reviews, 2008. 60(6): p. 625-637.  8. Humberstone, A.J. and W.N. Charman, Lipid-based vehicles for the oral delivery of poorly water soluble drugs. Advanced Drug Delivery Reviews, 1997. 25(1): p. 103-128.  9. Porter, C.J.H., N.L. Trevaskis, and W.N. Charman, Lipids and lipid-based formulations: optimizing the oral delivery of lipophilic drugs. Nat Rev Drug Discov, 2007. 6(3): p. 231-248.  10. Michaelsen, M., et al., The Effect of Digestion and Drug Load on Halofantrine Absorption from Self-nanoemulsifying Drug Delivery System (SNEDDS). The AAPS Journal, 2016. 18(1): p. 180-186.  11. Thomas, N., et al., Supersaturated Self-Nanoemulsifying Drug Delivery Systems (Super-SNEDDS) Enhance the Bioavailability of the Poorly Water-Soluble Drug Simvastatin in Dogs. The AAPS Journal, 2013. 15(1): p. 219-227.  12. Thomas, N., et al., In vitro and in vivo performance of novel supersaturated self-nanoemulsifying drug delivery systems (super-SNEDDS). Journal of Controlled Release, 2012. 160(1): p. 25-32.    44  13. Thomas, N., et al., In Vitro Lipolysis Data Does Not Adequately Predict the In Vivo Performance of Lipid-Based Drug Delivery Systems Containing Fenofibrate. The AAPS Journal, 2014. 16(3): p. 539-549.  14. Christensen, J.Ø., et al., Solubilisation of poorly water-soluble drugs during in vitro lipolysis of medium- and long-chain triacylglycerols. European Journal of Pharmaceutical Sciences, 2004. 23(3): p. 287-296.  15. Gershanik, T. and S. Benita, Self-dispersing lipid formulations for improving oral absorption of lipophilic drugs. European Journal of Pharmaceutics and Biopharmaceutics, 2000. 50(1): p. 179-188.  16. Charman, W.N., et al., Effect of Food and a Monoglyceride Emulsion Formulation on Danazol Bioavailability. The Journal of Clinical Pharmacology, 1993. 33(4): p. 381-386.  17. Tadros, T., et al., Formation and stability of nano-emulsions. Advances in Colloid and Interface Science, 2004. 108–109: p. 303-318.  18. Dahan, A. and A. Hoffman, Use of a Dynamic in Vitro Lipolysis Model to Rationalize Oral Formulation Development for Poor Water Soluble Drugs: Correlation with in Vivo Data and the Relationship to Intra-Enterocyte Processes in Rats. Pharmaceutical Research, 2006. 23(9): p. 2165-2174.  19. Cornaire, G., et al., Impact of excipients on the absorption of P-glycoprotein substrates in vitro and in vivo. International Journal of Pharmaceutics, 2004. 278(1): p. 119-131.  20. Pellett, M.A., et al., The penetration of supersaturated solutions of piroxicam across silicone membranes and human skin in vitro. Journal of Controlled Release, 1997. 46(3): p. 205-214.  21. Pellett, M.A., A.F. Davis, and J. Hadgraft, Effect of supersaturation on membrane transport: 2. Piroxicam. International Journal of Pharmaceutics, 1994. 111(1): p. 1-6.  22. Khan, J., T. Rades, and B. Boyd, The Precipitation Behavior of Poorly Water-Soluble Drugs with an Emphasis on the Digestion of Lipid Based Formulations. Pharmaceutical Research, 2015. 33(3): p. 548-562.  23. Bryson, H.M. and K.L. Goa, Halofantrine. Drugs, 1992. 43(2): p. 236-258.  24. Khoo, S.-M., et al., A physicochemical basis for the extensive intestinal lymphatic transport of a poorly lipid soluble antimalarial, halofantrine hydrochloride, after postprandial administration to dogs. Journal of Pharmaceutical Sciences, 2002. 91(3): p. 647-659.  25. Munoz, A., J.P. Guichard, and P. Reginault, Micronised fenofibrate. Atherosclerosis, 1994. 110: p. S45-S48.     45  26. Ling, H., J.T. Luoma, and D. Hilleman, A review of currently available fenofibrate and fenofibric acid formulations. Cardiology Research, 2013. 4(2): p. 47-55.  27. Vogt, M., K. Kunath, and J.B. Dressman, Dissolution enhancement of fenofibrate by micronization, cogrinding and spray-drying: Comparison with commercial preparations. European Journal of Pharmaceutics and Biopharmaceutics, 2008. 68(2): p. 283-288.  28. Do, T.T., et al., The conflict between in vitro release studies in human biorelevant media and the in vivo exposure in rats of the lipophilic compound fenofibrate. International Journal of Pharmaceutics, 2011. 414(1–2): p. 118-124.  29. Telimaa, S., J. Puolakka, and A. Kauppila, Placebo-controlled comparison of danazol and high-dose medroxyprogesterone acetate in the treatment of endometriosis. Gynecological Endocrinology, 1987. 1(1): p. 13-23.  30. Porter, C.J., et al., Susceptibility to lipase-mediated digestion reduces the oral bioavailability of danazol after administration as a medium-chain lipid-based microemulsion formulation. Pharmaceutical research, 2004. 21(8): p. 1405-1412.  31. Bakatselou, V., R.C. Oppenheim, and J.B. Dressman, Solubilization and wetting effects of bile salts on the dissolution of steroids. Pharmaceutical research, 1991. 8(12): p. 1461-1469.  32. Weibel, E.K., et al., Lipstatin, an inhibitor of pancreatic lipase, produced by Streptomyces toxytricini. I. Producing organism, fermentation, isolation and biological activity. The Journal of antibiotics, 1987. 40(8): p. 1081-1085.  33. Padwal, R.S. and S.R. Majumdar, Drug treatments for obesity: orlistat, sibutramine, and rimonabant. The Lancet. 369(9555): p. 71-77.  34. Mittendorfer, B., et al., Orlistat Inhibits Dietary Cholesterol Absorption. Obesity Research, 2001. 9(10): p. 599-604.  35. McClendon, K.S., D.M. Riche, and G.I. Uwaifo, Orlistat: current status in clinical therapeutics. Expert Opinion on Drug Safety, 2009. 8(6): p. 727-744.  36. Borgström, B., Mode of action of tetrahydrolipstatin: a derivative of the naturally occurring lipase inhibitor lipstatin. Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism, 1988. 962(3): p. 308-316.  37. Krebs, M., et al., Prevention of in vitro lipolysis by tetrahydrolipstatin. Clinical Chemistry, 2000. 46(7): p. 950-954.  38. Hadváry, P., et al., The lipase inhibitor tetrahydrolipstatin binds covalently to the putative active site serine of pancreatic lipase. Journal of Biological Chemistry, 1991. 266(4): p. 2021-2027.     46  39. Tiss, A., et al., Surface behaviour of bile salts and tetrahydrolipstatin at air/water and oil/water interfaces. Chemistry and Physics of Lipids, 2001. 111(1): p. 73-85.  40. Tiss, A., et al., Transfer of orlistat through oil–water interfaces. Chemistry and Physics of Lipids, 2002. 119(1–2): p. 41-49.  41. Tiss, A., et al., Inhibition of human pancreatic lipase by tetrahydrolipstatin: Further kinetic studies showing its reversibility. Journal of Molecular Catalysis B: Enzymatic, 2009. 58(1–4): p. 41-47.  42. Amidon, G., et al., A Theoretical Basis for a Biopharmaceutic Drug Classification: The Correlation of in Vitro Drug Product Dissolution and in Vivo Bioavailability. Pharmaceutical Research, 1995. 12(3): p. 413-420.  43. Yu, L.X., et al., Biopharmaceutics Classification System: The Scientific Basis for Biowaiver Extensions. Pharmaceutical Research. 19(7): p. 921-925.  44. Polli, J.E., et al., Summary Workshop Report: Bioequivalence, Biopharmaceutics Classification System, and Beyond. The AAPS Journal, 2008. 10(2): p. 373-379.  45. Food and D. Administration, Guidance for industry: waiver of in vivo bioavailability and bioequivalence studies for immediate-release solid oral dosage forms based on a biopharmaceutics classification system. Food and Drug Administration, Rockville, MD, 2000.  46. Pouton, C.W., Lipid formulations for oral administration of drugs: non-emulsifying, self-emulsifying and ‘self-microemulsifying’ drug delivery systems. European Journal of Pharmaceutical Sciences, 2000. 11, Supplement 2: p. S93-S98.  47. Pouton, C.W., Formulation of poorly water-soluble drugs for oral administration: Physicochemical and physiological issues and the lipid formulation classification system. European Journal of Pharmaceutical Sciences, 2006. 29(3–4): p. 278-287.  48. Müllertz, A., et al., New perspectives on lipid and surfactant based drug delivery systems for oral delivery of poorly soluble drugs. Journal of Pharmacy and Pharmacology, 2010. 62(11): p. 1622-1636.  49. Small, D.M., A classification of biologic lipids based upon their interaction in aqueous systems. Journal of the American Oil Chemists Society, 1968. 45(3): p. 108-119.  50. Hofmann, A.F. and B. Borgström, The Intraluminal Phase of Fat Digestion in Man: The Lipid Content of the Micellar and Oil Phases of Intestinal Content Obtained during Fat Digestion and Absorption. Journal of Clinical Investigation, 1964. 43(2): p. 247-257.  51. Carey, M.C. and D.M. Small, The characteristics of mixed micellar solutions with particular reference to bile. The American Journal of Medicine, 1970. 49(5): p. 590-608.    47  52. Mattson, F.H. and R.A. Volpenhein, The Digestion and Absorption of Triglycerides. Journal of Biological Chemistry, 1964. 239(9): p. 2772-2777.  53. Carey, M.C., D.M. Small, and C.M. Bliss, Lipid digestion and absorption. Annual Review of Physiology, 1983. 45(1): p. 651-677.  54. Kawai, T. and T. Fushiki, Importance of lipolysis in oral cavity for orosensory detection of fat. American Journal of Physiology - Regulatory, Integrative and Comparative Physiology, 2003. 285(2): p. R447-R454.  55. Dressman, J.B., et al., Upper gastrointestinal (GI) pH in young, healthy men and women. Pharmaceutical research, 1990. 7(7): p. 756-761.  56. Hutchinson, A., J.R. Maltby, and C.R. Reid, Gastric fluid volume and pH in elective inpatients. Part I: Coffee or orange juice versus overnight fast. Canadian journal of anaesthesia, 1988. 35(1): p. 12-15.  57. Hamosh, M., et al., Fat Digestion in the Newborn: Characterization of Lipase in Gastric Aspirates of Premature and Term Infants. Journal of Clinical Investigation, 1981. 67(3): p. 838-846.  58. Liao, T.H., P. Hamosh, and M. Hamosh, Fat Digestion by Lingual Lipase: Mechanism of Lipolysis in the Stomach and Upper Small Intestine. Pediatr Res, 1984. 18(5): p. 402-409.  59. N’Goma, J.-C.B., et al., Understanding the lipid-digestion processes in the GI tract before designing lipid-based drug-delivery systems. Therapeutic Delivery, 2011. 3(1): p. 105-124.  60. Carriere, F., et al., Secretion and contribution to lipolysis of gastric and pancreatic lipases during a test meal in humans. Gastroenterology, 1993. 105(3): p. 876-888.  61. Gargouri, Y., et al., Importance of human gastric lipase for intestinal lipolysis: an in vitro study. Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism, 1986. 879(3): p. 419-423.  62. Schiller, C., et al., Intestinal fluid volumes and transit of dosage forms as assessed by magnetic resonance imaging. Alimentary Pharmacology & Therapeutics, 2005. 22(10): p. 971-979.  63. Bergström, C.A.S., et al., Early pharmaceutical profiling to predict oral drug absorption: Current status and unmet needs. European Journal of Pharmaceutical Sciences, 2014. 57: p. 173-199.  64. Hildebrand, P., et al., Effects of a cholecystokinin receptor antagonist on intestinal phase of pancreatic and biliary responses in man. Journal of Clinical Investigation, 1990. 85(3): p. 640-646.     48  65. Maldonado-Valderrama, J., et al., The role of bile salts in digestion. Advances in Colloid and Interface Science, 2011. 165(1): p. 36-46.  66. Carrière, F., et al., In vivo and in vitro studies on the stereoselective hydrolysis of tri- and diglycerides by gastric and pancreatic lipases. Bioorganic & Medicinal Chemistry, 1997. 5(2): p. 429-435.  67. Mu, H. and C.-E. Høy, The digestion of dietary triacylglycerols. Progress in Lipid Research, 2004. 43(2): p. 105-133.  68. Borgström, B. and C. Erlanson, Pancreatic Lipase and Co-Lipase. European Journal of Biochemistry, 1973. 37(1): p. 60-68.  69. Hyun, S.A., G.V. Vahouny, and C.R. Treadwell, Portal absorption of fatty acids in lymph- and portal vein-cannulated rats. Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism, 1967. 137(2): p. 296-305.  70. Hernell, O., J.E. Staggers, and M.C. Carey, Physical-chemical behavior of dietary and biliary lipids during intestinal digestion and absorption. 2. Phase analysis and aggregation states of luminal lipids during duodenal fat digestion in healthy adult human beings. Biochemistry, 1990. 29(8): p. 2041-2056.  71. Staggers, J.E., et al., Physical-chemical behavior of dietary and biliary lipids during intestinal digestion and absorption. 1. Phase behavior and aggregation states of model lipid systems patterned after aqueous duodenal contents of healthy adult human beings. Biochemistry, 1990. 29(8): p. 2028-2040.  72. Wilson, F.A., V.L. Sallee, and J.M. Dietschy, Unstirred Water Layers in Intestine: Rate Determinant of Fatty Acid Absorption from Micellar Solutions. Science, 1971. 174(4013): p. 1031-1033.  73. Fagerholm, U. and H. Lennernäs, Experimental estimation of the effective unstirred water layer thickness in the human jejunum, and its importance in oral drug absorption. European Journal of Pharmaceutical Sciences, 1995. 3(5): p. 247-253.  74. Westergaard, H. and J. Dietschy, The mechanism whereby bile acid micelles increase the rate of fatty acid and cholesterol uptake into the intestinal mucosal cell. Journal of Clinical Investigation, 1976. 58(1): p. 97.  75. Simmonds, W.J., The Role of Micellar Solubilization in Lipid Absorption. Immunol Cell Biol, 1972. 50(4): p. 403-421.  76. Hoffman, N.E., The relationship between uptake in vitro of oleic acid and micellar solubilization. Biochimica et Biophysica Acta (BBA) - Biomembranes, 1970. 196(2): p. 193-203.     49  77. Kiyasu, J.Y., B. Bloom, and I.L. Chaikoff, The Portal Transport Of Absorbed Fatty Acids. Journal of Biological Chemistry, 1952. 199(1): p. 415-419.  78. MacGregor, K.J., et al., Influence of lipolysis on drug absorption from the gastro-intestinal tract. Advanced Drug Delivery Reviews, 1997. 25(1): p. 33-46.  79. Merisko-Liversidge, E., G.G. Liversidge, and E.R. Cooper, Nanosizing: a formulation approach for poorly-water-soluble compounds. European Journal of Pharmaceutical Sciences, 2003. 18(2): p. 113-120.  80. Noyes, A.A. and W.R. Whitney, The Rate of Solution of Solid Substances in Their Own Solutions. Journal of the American Chemical Society, 1897. 19(12): p. 930-934.  81. Löbmann, K., et al., Amino acids as co-amorphous stabilizers for poorly water soluble drugs – Part 1: Preparation, stability and dissolution enhancement. European Journal of Pharmaceutics and Biopharmaceutics, 2013. 85(3, Part B): p. 873-881.  82. Löbmann, K., et al., Amino acids as co-amorphous stabilizers for poorly water-soluble drugs – Part 2: Molecular interactions. European Journal of Pharmaceutics and Biopharmaceutics, 2013. 85(3, Part B): p. 882-888.  83. Carrigan, P.J. and T.R. Bates, Biopharmaceutics of Drugs Administered in Lipid-Containing Dosage Forms I: GI Absorption of Griseofulvin from an Oil-in-Water Emulsion in the Rat. Journal of Pharmaceutical Sciences, 1973. 62(9): p. 1476-1479.  84. Sunesen, V.H., et al., In vivo in vitro correlations for a poorly soluble drug, danazol, using the flow-through dissolution method with biorelevant dissolution media. European Journal of Pharmaceutical Sciences, 2005. 24(4): p. 305-313.  85. Dressman, J.B., et al., Estimating drug solubility in the gastrointestinal tract. Advanced Drug Delivery Reviews, 2007. 59(7): p. 591-602.  86. Lin, J.H., Species similarities and differences in pharmacokinetics. Drug Metabolism and Disposition, 1995. 23(10): p. 1008-1021.  87. Tang, C. and T. Prueksaritanont, Use of In Vivo Animal Models to Assess Pharmacokinetic Drug-Drug Interactions. Pharmaceutical Research, 2010. 27(9): p. 1772-1787.  88. Larsen, A.T., et al., Oral bioavailability of cinnarizine in dogs: Relation to SNEDDS droplet size, drug solubility and in vitro precipitation. European Journal of Pharmaceutical Sciences, 2013. 48(1–2): p. 339-350.  89. Larsen, A., et al., Lipid-based Formulations for Danazol Containing a Digestible Surfactant, Labrafil M2125CS: In Vivo Bioavailability and Dynamic In Vitro Lipolysis. Pharmaceutical Research, 2008. 25(12): p. 2769-2777.    50  90. Turner, P.V., et al., Administration of substances to laboratory animals: routes of administration and factors to consider. Journal of the American Association for Laboratory Animal Science, 2011. 50(5): p. 600-613.  91. Turner, P.V., et al., Administration of substances to laboratory animals: equipment considerations, vehicle selection, and solute preparation. Journal of the American Association for Laboratory Animal Science, 2011. 50(5): p. 614-627.  92. McConnell, E.L., A.W. Basit, and S. Murdan, Measurements of rat and mouse gastrointestinal pH, fluid and lymphoid tissue, and implications for in-vivo experiments. Journal of Pharmacy and Pharmacology, 2008. 60(1): p. 63-70.  93. DeSesso, J.M. and C.F. Jacobson, Anatomical and physiological parameters affecting gastrointestinal absorption in humans and rats. Food and Chemical Toxicology, 2001. 39(3): p. 209-228.  94. Kararli, T.T., Comparison of the gastrointestinal anatomy, physiology, and biochemistry of humans and commonly used laboratory animals. Biopharmaceutics & Drug Disposition, 1995. 16(5): p. 351-380.  95. Tanaka, Y., et al., Regional differences in the components of luminal water from rat gastrointestinal tract and comparison with other species. Journal of Pharmacy & Pharmaceutical Sciences, 2012. 15(4): p. 510-518.  96. Hagio, M., et al., Improved analysis of bile acids in tissues and intestinal contents of rats using LC/ESI-MS. Journal of Lipid Research, 2009. 50(1): p. 173-180.  97. Hamosh, M. and R.O. Scow, Lingual Lipase and Its Role in the Digestion of Dietary Lipid. Journal of Clinical Investigation, 1973. 52(1): p. 88-95.  98. Høy, a.C.-E., Structured Lipids Improve Fat Absorption in Normal and Malabsorbing Rats. The Journal of Nutrition, 2000. 130(11): p. 2802-2808.  99. Anby, M.U., et al., 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. Molecular Pharmaceutics, 2014. 11(11): p. 4069-4083.  100. de Smidt, P.C., M.A. Campanero, and I.F. Trocóniz, Intestinal absorption of penclomedine from lipid vehicles in the conscious rat: contribution of emulsification versus digestibility. International Journal of Pharmaceutics, 2004. 270(1–2): p. 109-118.  101. Myers, R.A. and V.J. Stella, Systemic bioavailability of penclomedine (NSC-338720) from oil-in-water emulsions administered intraduodenally to rats. International Journal of Pharmaceutics, 1992. 78(1–3): p. 217-226.    51  102. Yoshiya, Y., et al., Biopharmaceutical studies of lipid-containing oral dosage forms: Relationship between drug absorption rate and digestibility of vehicles. International Journal of Pharmaceutics, 1979. 3(1): p. 23-31.  103. Grundy, D., The effect of surgical anaesthesia on antral motility in the ferret. Experimental physiology, 1990. 75(5): p. 701-708.  104. Bueno, L., J.-P. Ferre, and Y. Ruckebusch, Effects of anesthesia and surgical procedures on intestinal myoelectric activity in rats. The American journal of digestive diseases, 1978. 23(8): p. 690-695.  105. Thomas, N., et al., Influence of lipid composition and drug load on the In Vitro performance of self-nanoemulsifying drug delivery systems. Journal of Pharmaceutical Sciences, 2012. 101(5): p. 1721-1731.  106. Fernandez, E. and B. Borgström, Effects of tetrahydrolipstatin, a lipase inhibitor, on absorption of fat from the intestine of the rat. Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism, 1989. 1001(3): p. 249-255.  107. Erlanson-Albertsson, C., A. Larsson, and R. Duan, Secretion of Pancreatic Lipase and Colipase from Rat Pancreas. Pancreas, 1987. 2(5): p. 531-535.  108. Hadvary, P., H. Lengsfeld, and H. Wolfer, Inhibition of pancreatic lipase in vitro by the covalent inhibitor tetrahydrolipstatin. Biochem. J, 1988. 256: p. 357-361.  109. Mohsin, K., Design of Lipid-Based Formulations for Oral Administration of Poorly Water-Soluble Drug Fenofibrate: Effects of Digestion. AAPS PharmSciTech, 2012. 13(2): p. 637-646.  110. Kalepu, S., M. Manthina, and V. Padavala, Oral lipid-based drug delivery systems – an overview. Acta Pharmaceutica Sinica B, 2013. 3(6): p. 361-372.  111. Holm, R., et al., Structured Triglyceride Vehicles for Oral Delivery of Halofantrine: Examination of Intestinal Lymphatic Transport and Bioavailability in Conscious Rats. Pharmaceutical Research, 2002. 19(9): p. 1354-1361.  112. Sylvén, C. and B. Borgström, Intestinal absorption and lymphatic transport of cholesterol in the rat: influence of the fatty acid chain length of the carrier triglyceride. Journal of Lipid Research, 1969. 10(4): p. 351-355.  113. Charman, W.N.A. and V.J. Stella, Estimating the maximal potential for intestinal lymphatic transport of lipophilic drug molecules. International Journal of Pharmaceutics, 1986. 34(1): p. 175-178.  114. Porter, C.J.H., S.A. Charman, and W.N. Charman, Lymphatic transport of halofantrine in the triple-cannulated anesthetized rat model: Effect of lipid vehicle dispersion. Journal of Pharmaceutical Sciences, 1996. 85(4): p. 351-356.    52  115. Hanafy, A., et al., Pharmacokinetic evaluation of oral fenofibrate nanosuspensions and SLN in comparison to conventional suspensions of micronized drug. Advanced Drug Delivery Reviews, 2007. 59(6): p. 419-426.  116. Boyd, B.J., C.J.H. Porter, and W.N. Charman, Using the polymer partitioning method to probe the thermodynamic activity of poorly water-soluble drugs solubilized in model lipid digestion products. Journal of Pharmaceutical Sciences, 2003. 92(6): p. 1262-1271.  117. Nielsen, F.S., K.B. Petersen, and A. Müllertz, Bioavailability of probucol from lipid and surfactant based formulations in minipigs: Influence of droplet size and dietary state. European Journal of Pharmaceutics and Biopharmaceutics, 2008. 69(2): p. 553-562.  118. Thi, T.D., et al., Formulate-ability of ten compounds with different physicochemical profiles in SMEDDS. European Journal of Pharmaceutical Sciences, 2009. 38(5): p. 479-488.  119. Brouwers, J., M.E. Brewster, and P. Augustijns, Supersaturating drug delivery systems: The answer to solubility-limited oral bioavailability? Journal of Pharmaceutical Sciences, 2009. 98(8): p. 2549-2572.  120. Bevernage, J., et al., Evaluation of gastrointestinal drug supersaturation and precipitation: Strategies and issues. International Journal of Pharmaceutics, 2013. 453(1): p. 25-35.  121. Davis, A.F. and J. Hadgraft, Effect of supersaturation on membrane transport: 1. Hydrocortisone acetate. International Journal of Pharmaceutics, 1991. 76(1–2): p. 1-8.  122. Gao, P., et al., Development of a supersaturable SEDDS (S-SEDDS) formulation of paclitaxel with improved oral bioavailability. Journal of Pharmaceutical Sciences, 2003. 92(12): p. 2386-2398.  123. Gao, P., et al., Enhanced Oral Bioavailability of a Poorly Water Soluble Drug PNU‐91325 by Supersaturatable Formulations. Drug development and industrial pharmacy, 2004. 30(2): p. 221-229.  124. Gao, P., et al., Characterization and optimization of AMG 517 supersaturatable self-emulsifying drug delivery system (S-SEDDS) for improved oral absorption. Journal of Pharmaceutical Sciences, 2009. 98(2): p. 516-528.  125. Williams, H., et al., Lipid-Based Formulations and Drug Supersaturation: Harnessing the Unique Benefits of the Lipid Digestion/Absorption Pathway. Pharmaceutical Research, 2013: p. 1-17.  126. Berthelsen, R., et al., Kolliphor Surfactants Affect Solubilization and Bioavailability of Fenofibrate. Studies of in Vitro Digestion and Absorption in Rats. Molecular Pharmaceutics, 2015. 12(4): p. 1062-1071.     53  127. Zangenberg, N.H., et al., A dynamic in vitro lipolysis model: I. Controlling the rate of lipolysis by continuous addition of calcium. European Journal of Pharmaceutical Sciences, 2001. 14(2): p. 115-122.  128. Zangenberg, N.H., et al., A dynamic in vitro lipolysis model: II: Evaluation of the model. European Journal of Pharmaceutical Sciences, 2001. 14(3): p. 237-244.  129. Thomas, N., et al., Characterising lipid lipolysis and its implication in lipid-based formulation development. The AAPS journal, 2012. 14(4): p. 860-871.  130. Capolino, P., et al., In vitro gastrointestinal lipolysis: replacement of human digestive lipases by a combination of rabbit gastric and porcine pancreatic extracts. Food Digestion, 2011. 2(1-3): p. 43-51.  131. Kaukonen, A.M., et al., Drug Solubilization Behavior During in Vitro Digestion of Simple Triglyceride Lipid Solution Formulations. Pharmaceutical Research. 21(2): p. 245-253.  132. Scow, R.O., P. Desnuelle, and R. Verger, Lipolysis and lipid movement in a membrane model. Action of lipoprotein lipase. Journal of Biological Chemistry, 1979. 254(14): p. 6456-63.  133. Alvarez, F.J. and V.J. Stella, The Role of Calcium Ions and Bile Salts on the Pancreatic Lipase-Catalyzed Hydrolysis of Triglyceride Emulsions Stabilized with Lecithin. Pharmaceutical Research, 1989. 6(6): p. 449-457.  134. Sassene, P.J., et al., Precipitation of a poorly soluble model drug during in vitro lipolysis: characterization and dissolution of the precipitate. Journal of pharmaceutical sciences, 2010. 99(12): p. 4982-4991.  135. Sassene, P.J., et al., Elucidating the Molecular Interactions Occurring during Drug Precipitation of Weak Bases from Lipid-Based Formulations: A Case Study with Cinnarizine and a Long Chain Self-Nanoemulsifying Drug Delivery System. Molecular Pharmaceutics, 2015. 12(11): p. 4067-4076.  136. Stillhart, C., D. Dürr, and M. Kuentz, Toward an Improved Understanding of the Precipitation Behavior of Weakly Basic Drugs from Oral Lipid-Based Formulations. Journal of Pharmaceutical Sciences, 2014. 103(4): p. 1194-1203.  137. Fatouros, D., et al., Structural Development of Self Nano Emulsifying Drug Delivery Systems (SNEDDS) During In Vitro Lipid Digestion Monitored by Small-angle X-ray Scattering. Pharmaceutical Research, 2007. 24(10): p. 1844-1853.  138. Fatouros, D., et al., Colloidal Structures in Media Simulating Intestinal Fed State Conditions with and Without Lipolysis Products. Pharmaceutical Research, 2009. 26(2): p. 361-374.    54  139. Fatouros, D.G., B. Bergenstahl, and A. Mullertz, Morphological observations on a lipid-based drug delivery system during in vitro digestion. European Journal of Pharmaceutical Sciences, 2007. 31(2): p. 85-94.  140. Kleberg, K., et al., Biorelevant media simulating fed state intestinal fluids: Colloid phase characterization and impact on solubilization capacity. Journal of Pharmaceutical Sciences, 2010. 99(8): p. 3522-3532.  141. Warren, D.B., et al., Real time evolution of liquid crystalline nanostructure during the digestion of formulation lipids using synchrotron small-angle X-ray scattering. Langmuir, 2011. 27(15): p. 9528-9534.  142. Tønsberg, H., et al., Effect of bile on the oral absorption of halofantrine in polyethylene glycol 400 and polysorbate 80 formulations dosed to bile duct cannulated rats. Journal of Pharmacy and Pharmacology, 2011. 63(6): p. 817-824.  143. Cuiné, J.F., et al., Increasing the proportional content of surfactant (Cremophor EL) relative to lipid in self-emulsifying lipid-based formulations of danazol reduces oral bioavailability in beagle dogs. Pharmaceutical research, 2007. 24(4): p. 748-757.  144. Tarling, M.M., et al., A model of gastric emptying using paracetamol absorption in intensive care patients. Intensive Care Medicine. 23(3): p. 256-260.  145. Satoshi, H., et al., The measurement of gastric emptying in conscious rats by monitoring serial changes in serum acetaminophen level. Journal of Pharmacological and Toxicological Methods, 1994. 31(3): p. 161-165.  146. Shah, B., V.K. Kakumanu, and A.K. Bansal, Analytical techniques for quantification of amorphous/crystalline phases in pharmaceutical solids. Journal of Pharmaceutical Sciences, 2006. 95(8): p. 1641-1665.  147. Chung, F.H., Quantitative interpretation of X-ray diffraction patterns of mixtures. I. Matrix-flushing method for quantitative multicomponent analysis. Journal of Applied Crystallography, 1974. 7(6): p. 519-525.  148. Williams, H.D., et al., 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. Journal of Pharmaceutical Sciences, 2012. 101(9): p. 3360-3380.  149. Mosgaard, M.D., et al., Development of a high-throughput in vitro intestinal lipolysis model for rapid screening of lipid-based drug delivery systems. European Journal of Pharmaceutics and Biopharmaceutics, 2015. 94: p. 493-500.  150. Hens, B., et al., Supersaturation and Precipitation of Posaconazole upon Entry in the Upper Small Intestine in Humans. Journal of Pharmaceutical Sciences, 2015: p. n/a-n/a.    55  151. Walravens, J., et al., Effect of pH and comedication on gastrointestinal absorption of posaconazole. Clinical pharmacokinetics, 2011. 50(11): p. 725-734.  152. Lu, Y., S. Kim, and K. Park, In vitro–in vivo correlation: Perspectives on model development. International Journal of Pharmaceutics, 2011. 418(1): p. 142-148.  153. Yeap, Y.Y., N.L. Trevaskis, and C.J.H. Porter, The Potential for Drug Supersaturation during Intestinal Processing of Lipid-Based Formulations May Be Enhanced for Basic Drugs. Molecular Pharmaceutics, 2013. 10(7): p. 2601-2615.  154. Stillhart, C., et al., Biopharmaceutical Modeling of Drug Supersaturation During Lipid-Based Formulation Digestion Considering an Absorption Sink. Pharmaceutical Research, 2014. 31(12): p. 3426-3444.  155. Dahan, A. and A. Hoffman, The effect of different lipid based formulations on the oral absorption of lipophilic drugs: The ability of in vitro lipolysis and consecutive ex vivo intestinal permeability data to predict in vivo bioavailability in rats. European Journal of Pharmaceutics and Biopharmaceutics, 2007. 67(1): p. 96-105.  156. Crum, M.F., et al., A new in vitro lipid digestion – in vivo absorption model to evaluate the mechanisms of drug absorption from lipid-based formulations. Pharmaceutical Research, 2015. 33(4): p. 970-982.  157. Griffin, B.T., et al., Comparison of in vitro tests at various levels of complexity for the prediction of in vivo performance of lipid-based formulations: Case studies with fenofibrate. European Journal of Pharmaceutics and Biopharmaceutics, 2014. 86(3): p. 427-437.  158. Christophersen, P.C., et al., Fed and fasted state gastro-intestinal in vitro lipolysis: In vitro in vivo relations of a conventional tablet, a SNEDDS and a solidified SNEDDS. European Journal of Pharmaceutical Sciences, 2014. 57: p. 232-239.  159. Mercuri, A., et al., The Effect of Composition and Gastric Conditions on the Self-Emulsification Process of Ibuprofen-Loaded Self-Emulsifying Drug Delivery Systems: A Microscopic and Dynamic Gastric Model Study. Pharmaceutical Research, 2011. 28(7): p. 1540-1551.  160. Christiansen, M.L., et al., Cinnarizine food-effects in beagle dogs can be avoided by administration in a Self Nano Emulsifying Drug Delivery System (SNEDDS). European Journal of Pharmaceutical Sciences, 2014. 57: p. 164-172.      56  10. Appendix       Appendix I Paper I: Michaelsen MH, Wasan KM, Sivak O, Müllertz A, Rades T. The Effect of Digestion and Drug Load on Halofantrine Absorption from Self-nanoemulsifying Drug Delivery System (SNEDDS). The AAPS Journal. 2016: 18(1):180-6. Research ArticleThe Effect of Digestion and Drug Load on Halofantrine Absorptionfrom Self-nanoemulsifying Drug Delivery System (SNEDDS)Maria Høtoft Michaelsen,1,2 Kishor M. Wasan,2,3 Olena Sivak,2 Anette Müllertz,1,4,5 and Thomas Rades1Received 12 June 2015; accepted 29 September 2015Abstract. A super-saturated self-nanoemulsifying drug delivery system (super-SNEDDS), containing thepoorly water-soluble drug halofantrine (Hf) at 150% of equilibrium solubility (Seq), was comparedin vitro and in vivo with a conventional SNEDDS (75% of Seq) with respect to bioavailability anddigestibility. Further, the effect of digestion on oral absorption of Hf from SNEDDS and super-SNEDDSwas assessed by incorporation of the lipase inhibitor tetrahydrolipstatin (orlistat) into the SNEDDS. TheSNEDDS contained soybean oil/Maisine 34-I (1:1), Kolliphor RH40, and ethanol at a ratio of 55:35:10,w/w percent. For the dynamic in vitro lipolysis, the precipitation of Hf at 60 min was significantly largerfor the super-SNEDDS (66.8±16.4%) than for the SNEDDS (18.5±9.2%). The inhibition of the in vitrodigestion by orlistat (1% (w/w)) lowered drug precipitation significantly for both the super-SNEDDS(36.8±1.7%) and the SNEDDS (3.9±0.7%). In the in vivo studies, the super-SNEDDS concept provedvalid in a rat model with a significantly larger Cmax for the super-SNEDDS (964±167 ng/mL) than for theSNEDDS (506±112 ng/mL). The bioavailability of Hf dosed in super-SNEDDS (32.9±3.6%) andSNEDDS (22.5±6.3%) did not change significantly with co-administration of orlistat (45.5±7.3% and 21.9±6.5%, respectively). However, the pharmacokinetic parameters changed; the tmax of the super-SNEDDS(1.3±0.1 h) and SNEDDS (2.8±1.2 h) were significantly lower when dosed with orlistat (6.0±1.3 and 6.3±1.2 h, respectively). These findings suggest that the role of lipid digestion for the absorption of drugsfrom SNEDDS may be less important than previously thought.KEY WORDS: absorption; digestion; halofantrine; orlistat; SNEDDS; super-SNEDDS.INTRODUCTIONThe physico-chemical properties of highly lipophiliccompounds often result in poor and variable bioavailabil-ity (1,2). This calls for enabling drug delivery systems inorder to decrease variability in absorption and to increasebioavailability. One strategy to solve this problem is toutilize lipid- and surfactant-based drug delivery systems,such as self-nanoemulsifying drug delivery systems(SNEDDS). SNEDDS preconcentrates are water-freeisotropic systems consisting of oil, surfactant, co-solvent,and drug, which upon mild agitation in aqueous environ-ments (such as in the gastrointestinal tract) form ananoemulsion. The delivery of lipophilic drugs inSNEDDS often results in an increased oral absorptionand hence an improved bioavailability (3).Traditionally, drugs are loaded into SNEDDS atconcentrations below their equilibrium solubility (Seq)(4). The potentially resulting low drug load, however,can be an obstacle, and a reason not to develop aSNEDDS for a given drug, for example if the requireddose cannot be obtained in one single capsule. Incontrast, the recently introduced super-saturatedSNEDDS (super-SNEDDS) contain the drug at concen-trations above Seq and thereby both the dose and thethermodynamic activity of these delivery systems areincreased. Recent studies in dogs and minipigs haveshown that the bioavailability of a drug in super-SNEDDS is equal to or better than that in conventionalSNEDDS when the same dose is given. Thus, the numberof capsules to be ingested to obtain the desired dose canbe decreased (5–7). However, due to the lower cost andthe ease of handling of rats compared to dogs andminipigs, studies in rats are often preferred.Among the many variables that are thought to beaffecting bioavailability of a drug dosed in SNEDDS andsuper-SNEDDS, the most prominent are the physico-chemical properties of the drug, the type and compositionof the lipid vehicle used, and the emulsification tendencyand digestibility of the SNEDDS (3,8,9). Lipid digestion1Department of Pharmacy, Faculty of Health and Medical Sciences,University of Copenhagen, Universitetsparken 2, DK-2100,Copenhagen, Denmark .2 Faculty of Pharmaceutical Sciences, University of British Columbia,2405 Wesbrook Mall, Vancouver, British Columbia, Canada V6T1Z3.3 College of Pharmacy and Nutrition, University of Saskatchewan,110 Science Place, Saskatoon, Saskatchewan, Canada S7N 2Z4.4 Bioneer:FARMA, Department of Pharmacy, Faculty of Healthand Medical Sciences, University of Copenhagen, DK-2100,Copenhagen, Denmark .5 To whom correspondence should be addressed. (e-mail:anette.mullertz@sund.ku.dk)The AAPS Journal (# 2015)DOI: 10.1208/s12248-015-9832-71550-7416/15/0000-0001/0 # 2015 American Association of Pharmaceutical Scientistsand the physico-chemical behavior of the digestionproducts have been extensively studied since the early1960s (10). Digestion is a dynamic process, and studieshave shown that various colloidal phases with differentmorphologies are formed during lipid digestion. Specifi-cally, the digestion of SNEDDS is thought to play animportant role for the absorption of the drugs due to thealtered solubilizing capacity of the various and transientcolloidal structures (11,12). For example, Fatouros andcoworkers subjected a SNEDDS composed of long-chaintriacylglycerides to in vitro lipolysis and the formedstructures were analyzed by small-angle X-ray scattering(SAXS) (13). Initially, oil droplets coexist with micelles,but after 60 min of lipolysis (corresponding to 90%digestion of the triacylglycerides), predominantly mixedmicelles were observed along with few unilamellarvesicles. The traditional theory on drug absorption fromSNEDDS thus revolves around the solubilization of thedrug in the formed colloidal structures and the formationof mixed micelles, promoting increased absorption. Thegeneral belief is that the mixed micelles containing thedrug will diffuse to the unstirred water layer at theepithelial membrane and, due to the pH gradient foundthere, they disintegrate. The digestion products, themixed micelles are composed of, will be absorbed andthe bile salts re-circulated. The drug will be released fromthe mixed micelles upon their disintegration and hence beabsorbed as free drug molecules.One strategy to evaluate the role of digestion on drugabsorption is the use of tetrahydrolipstatin (orlistat).Orlistat is a potent inhibitor of different lipases such aspancreatic lipase, gastric lipase, and carboxyl ester lipase,but not phospholipase A2 (14). In vitro studies withorlistat have shown that when it interacts with porcinepancreatic lipase, the interaction is irreversible; however,the same may not be true for all lipases (14). A morerecent study with human pancreatic lipase has shown thatthe inhibitory effect exerted by orlistat may in fact bereversible (15). Orlistat has previously been shown toinhibit rat pancreatic lipase effectively and can thereforebe used as a tool to inhibit lipolysis (digestion) ofnanoemulsion droplets in vivo in rats (16). Orlistat islipophilic and forms a stable monolayer at the oil-waterinterface; the availability of orlistat at the interface is ofgreat importance to the inhibition of lipases, whereasorlistat in the core of an emulsion droplet will act merelyas a reservoir (17). Hence, orlistat is expected to bepresent at the oil-water interface of the emulsion droplets,where it forms a complex with the lipase at its active siteand blocks the activity of the lipase (18). In a previousstudy applying orlistat as a tool, de Smidt et al. found thatthe digestibility of medium-chain triacylglycerides andsurfactant-based formulations did not significantly influ-ence the absorption of penclomidine in rats. However,when using crude medium-chain triglyceride as a vehicle,the absorption was to a higher degree dependent ondigestion (9).The aims of the present study were twofold: toevaluate if the super-SNEDDS concept is also valid in ratsand to elucidate the effect of digestion on drug absorptionfrom SNEDDS using the lipase inhibitor orlistat as a tool.MATERIALS AND METHODSMaterialsHalofantrine hydrochloride was purchased from APACPharmaceutical LLC (Hangzhou, China), and thehalofantrine base was subsequently prepared fromhalofantrine hydrochloride as previously described (19).Lipoid E 80 was kindly donated by Lipoid (Ludwigshafen,Germany). Soybean oil (long-chain (LC) glycerides), glycer-ol, ethylenediaminetetraacetic acid tripotassium saltdihydrate (EDTA), 4-bromophenyl-boronic acid (BBBA),bile extract, tris-(hydroxymethyl)aminomethane (tris), maleicacid, calcium chloride, sodium hydroxide, and porcinepancreatic lipase were obtained from Sigma-Aldrich (St.Louis, MO, USA). Maisine 35-1 (a mixture of LC mono-,di-, and triglycerides) was kindly donated by Gattefossé (St.Priest, France), and Kolliphor RH 40 was donated by BASF(Ludwigshafen, Germany). Euthanyl® (sodium pentobarbital240 mg/mL) was supplied by Bimeda-MTC (Cambridge,Ontario, Canada). Purified water was obtained from aMillipore Milli-Q Ultra Pure water purification system(Billerica, MA, USA). tert-Butyl methyl ether (TBME),UPLC-grade acetonitrile, and glacial acetic acid were allfrom Fisher Scientific (Waltham, MA, USA). All otherchemicals were of analytical grade.Preparation of FormulationsPreconcentrates were prepared as previously described(20). The SNEDDS were composed of soybean oil (27.5%(w/w)), Maisine 35-1 (27.5% (w/w)), Kolliphor RH40 (35%(w/w)), and absolute ethanol (10% (w/w)). Briefly, the LCglycerides (soybean oil) and molten Maisine 35-1 (heated to50°C) were mixed. Kolliphor RH 40 was heated to 50°C andadded to the lipid mixture. Ethanol was added and themixture was left until homogenous.The drug was weighed into the preconcentrates followedby sonication for 5 min for the SNEDDS and 30 min for thesuper-SNEDDS. The super-SNEDDS were then subjected toheating at 60°C for 5 h and then left at 37°C overnight toequilibrate, and a clear solution was obtained. The super-SNEDDS stayed physically stable and Hf stayed chemicallystable for at least 6 months (5). The drug was added at 75%equilibrium solubility (Seq) and 150% Seq for the SNEDDSand super-SNEDDS, respectively. The equilibrium solubilitywas previously determined by Thomas et al. (5). Orlistat wasdissolved in ethanol and added to the formulations, giving afinal concentration of 1% (w/w). Immediately before thein vivo experiments, the SNEDDS were emulsified usingMilli-Q water at a 25% (w/w) lipid/water ratio.The intravenous o/w emulsion contained 0.1% (w/w) Hf,20% (w/w) soybean oil, 2% (w/w) lecithin, 2.5% (w/w)glycerol, and 75.4% (w/w) purified water. The emulsion wasprepared by dissolving Hf and lecithin in soybean oil undergentle heating (50°C). Glycerol was added to the water, andthe solution was heated to 50°C before the two phases werecombined. The mixture was homogenized for 3 min. Toreduce the droplet size of the resulting emulsion further, theemulsion was homogenized on ice for 5 min using anultrasonicator with a microtip at power output 5 (SonicsMichaelsen et al.Vibra-Cell, Sonics and Materials, Newtown, USA). The finalemulsion was filtered through a sterile 0.45-μm filter.In Vitro LipolysisDynamic in vitro lipolysis was carried out as previouslydescribed (20–22). Briefly, the SNEDDS was weighed into avessel containing 25 mL fasted state intestinal lipolysismedium (2.5 mM bovine bile salt, 0.26 mM phospholipid,2 mM Tris, 2 mM maleic acid, and 50 mM sodium chloride).After equilibration for 3 min, the pH was adjusted to 6.5 byan automated pH-stat (Metrohm Titrino 744, Tiamo version1.3, Switzerland). The in vitro lipolysis was initiated by theaddition of 5 mL freshly prepared pancreatic lipase (pH 6.5,37°C). The pancreatic lipase was prepared by weighing thelipase into a polypropylene tube and adding water. Themixture was centrifuged (7 min, 4000 rpm, 37°C) and the pHof the supernatant was adjusted to 6.5, resulting in a totallipase activity of 550 u/mL. The rate of lipolysis wascontrolled by addition of calcium in the form of CaCl2(0.6 M; 0.045 mmol/min) throughout the lipolysis (60 min).The liberated fatty acids generated during the lipolysis werecontinuously titrated with NaOH (0.4 M) to maintain the pHat 6.5. After 60 min, the lipolysis was terminated and a backtitration was performed at pH 9 to determine the exactamount of liberated free fatty acids.One milliliter of digestion medium was withdrawn attime zero and after 60 min of lipolysis. The lipase activity wasquenched with 5 μL 4-BBBA (1 M in methanol) followed byultracentrifugation (50,000 rpm, 50 min at 37°C).Quantitative Analysis of Samples from LipolysisSamples were analyzed for Hf content after 60 min oflipolysis in the aqueous and pellet phases after ultracentrifu-gation and dilution, using an isocratic high-performanceliquid chromatography (HPLC) method. The systemconsisted of a Dionex ASI-100 Automated sample injector,P680 HPLC pump, and a PDA-100 photo diode arraydetector (Thermo Fisher Scientific, Waltham, MA, USA).The column used was a Waters x-bridge C8 column (Waters,Milford, MA, USA). The mobile phases consisted of aceto-nitrile and purified water containing 0.2% SDS and glacialacetic acid in the ratio 80:20 (v:v). A constant flow of 0.8 mL/min was employed with an injection volume of 10 μL. Thechromatograms were analyzed using Thermo ScientificDionex Chromeleon 7 Chromatography Data System soft-ware (Thermo Fisher Scientific, Waltham, MA, USA).In Vivo StudyThe animal protocols used in this study were approvedby the University of British Columbia’s Animal CareCommittee and conform to the Canadian Council on AnimalCare guidelines. Male Sprague Dawley (SD) rats (270–310 g)with jugular vein cannulation were obtained from CharlesRiver Laboratories (Wilmington, MA, USA). The animalswere kept on standard feed and had free access to waterduring the experiment. Before initiation of the experiment,the rats were fasted for 10–15 h. The rats received 6.7 mg/kgHf and 140 or 70 mg/kg of lipid from the SNEDDS and super-SNEDDS, respectively, by oral gavage. Two hundred twenty-five microliters of blood samples was drawn from the jugularvein at 0 and 30 min and 1, 1.5, 2, 4, 6, 8, 12, 24, and 30 h afteradministration and collected into Eppendorf tubes containingEDTA. The withdrawn blood was replaced by an equalvolume of normal saline containing 100 units heparin to avoidhypovolemia and to flush the jugular vein catheter. Theplasma was subsequently harvested by centrifugation (10 min,5000 rpm) and stored at −80°C until analysis.The group receiving intravenous treatment received adose of 1.7 mg/kg Hf in the jugular vein. Blood samples wereadditionally collected at 0.5, 2, 10, and 20 min. The animalswere euthanized 30 h after dosing. The doses chosen werebased on existing literature (23,24).Quantitative Analysis of Plasma SamplesQuantification of Hf in the plasma samples obtained fromthe in vivo study was performed using a Waters Acquity UPLCsystem equipped with a binary solvent delivery system and aPDA detector (Waters, Milford, MA, USA). The analysis wasbased on a previously validated method described byHumberstone et al. (25). Briefly, 100 μL of plasma was aliquotedto a polypropylene tube and the plasma proteins wereprecipitated with 1 mL of acetonitrile and the samples vortexedfor 2 min. Four milliliters of TBME was added and thesamples were again vortexed for 2 min. The samples werethen centrifuged for 5 min (4000 rpm, 4°C) and thesupernatant transferred to a polypropylene tube contain-ing 100 μL of 5 mM HCl in acetonitrile. The content wasthen evaporated to dryness under a stream of nitrogen at45°C for 90 min (TurboVap LV, Zymark, MA, USA) andthe residue reconstituted in acetonitrile. Twenty-fivemicroliters of the reconstituted residue was injected ontoa Waters x-bridge C8 column (Waters, Milford, MA,USA) maintained at 30°C. A constant flow rate of1 mL/min and a mobile phase consisting of a mixture ofacetonitrile and water containing 0.2% SDS and 0.2%glacial acetic acid in the ratio 85:15 were employed, andthe data was collected at 257 nm. Standard curves wereprepared daily for the quantification of Hf in the plasmasamples. The chromatograms were analyzed using Em-power 3 Chromatography Data Software (Waters, Milford,MA, USA).Pharmacokinetic AnalysisThe pharmacokinetic parameters were analyzed usingWinNonlin Professional Version 6.3 (Pharsight Corporation,Mountain View, CA, USA). The area under the curve (AUC)was determined using the linear trapezoidal model from t=0until the last plasma sample was taken at t=30 h. Themaximum plasma concentration of Hf (Cmax), the time wherethe maximum concentration occurred (tmax), and the timerequired to reduce the plasma concentration to half of theCmax (t½) were determined from the individual curves. Thedata was normalized to dose. The absolute bioavailability Fawas calculated as follows:Fa ¼AUCpo⋅Div AUCiv⋅Dpo  ⋅100Importance of Drug Load and Digestion on Absorption from SNEDDSStatistical AnalysisThe data sets are expressed as mean±standard error ofthe mean (SEM). Statistical analysis was performed usingGraphPad Prism 6.03 (GraphPad Software, San Diego, CA,USA). For the in vitro data, ANOVA followed by Tukey’sposttest was used (α=0.05), and for the in vivo data, ANOVAfollowed by either Šídák or Dunnet’s posttest was used toanalyze statistical differences between groups (α=0.05).RESULTS AND DISCUSSIONThe aims of the present study were to evaluate theabsorption of Hf from SNEDDS and super-SNEDDS in ratsand to elucidate the effect of digestion on absorption of Hfusing the lipase inhibitor orlistat. Hf as the free base wasemployed as a model drug.Dynamic In Vitro LipolysisThe titration curves from the in vitro lipolysis of theSNEDDS and super-SNEDDS with and without orlistat areshown in Table I and Fig. 1. When orlistat is present, theactivity of pancreatic lipase is almost completely inhibited.The difference between the titration curves for the SNEDDSand the super-SNEDDS without orlistat can be ascribed tothe amount of lipid present in the vessel. Since the same doseof drug was added in all the lipolysis experiments, the lipidcontent was 50% for the super-SNEDDS compared to theSNEDDS (the SNEDDS have a drug load of 75% of the Seq,whereas the super-SNEDDS are super-saturated and contain150% of the Seq).The relative amount of Hf found in either the pellet orthe aqueous phase after in vitro lipolysis at 0 and 60 min isdepicted in Fig. 2. When comparing the SNEDDS and thesuper-SNEDDS, the amount precipitated from the super-SNEDDS after 60 min of lipolysis is significantly higher thanthat of the SNEDDS (P<0.1). Before initiation of lipolysis,the solubilization capacity can be ascribed to the dispersion ofthe undigested formulation in the presence of the bile saltsand phospholipids. During lipolysis, the continuous digestionof the SNEDDS and super-SNEDDS and the formation ofdigestion products (free fatty acids and monoglycerides)result in a decrease in the solubilization capacity andprecipitation of Hf. Since the lipid to drug ratio is higher forthe SNEDDS than for the super-SNEDDS, the solubilizingcapacity of the SNEDDS is higher. Upon hydrolysis, thehigher amounts of lipids will generate more digestionproducts and this will facilitate drug solubilization due tothe colloidal structures that are formed. From Fig. 2, it is clearthat addition of orlistat to the SNEDDS results in less Hfprecipitated at 60 min (P<0.1). The same trend can be seenfor the super-SNEDDS with and without orlistat (P<0.1). Theamount of precipitation for the super-SNEDDS with orlistatis higher than that for the SNEDDS with orlistat even thoughthere is no digestion. In agreement with a previous study, theHf precipitating during in vitro lipolysis was amorphous.Interestingly, the drug precipitates in the presence of orlistatwere also found to be amorphous by X-ray powder diffraction(XRPD) (data not shown) (7). Precipitation in the presenceof orlistat can be ascribed to the physical instability of thesuper-saturated system where the drug precipitates over timebecause of the energy induced by the stirring.In Vivo StudyThe mean plasma concentration profiles of Hf adminis-tered orally to rats are presented in Fig. 3, and thepharmacokinetic parameters are provided in Table II. AllSNEDDS were dosed at 6.7 mg Hf/kg body weight of the ratsresulting in 140 mg lipid/kg for the SNEDDS and 70 mg lipid/kg for the super-SNEDDS.Super-SNEDDS EffectAdministration of Hf in super-SNEDDS results in asignificantly larger Cmax compared to the SNEDDS (P<0.05).There were no differences in the time to reach the maximumconcentration (tmax) and the half-life (t½) between theSNEDDS and the super-SNEDDS. Although the differencesin AUC between the SNEDDS and super-SNEDDS have notreached significance, there is a tendency of a higher AUC forthe super-SNEDDS and hence also a higher relative bioavail-ability of the drug than for the SNEDDS. The samedifferences can be observed for the SNEDDS and super-SNEDDS when administered with orlistat.The exact mechanism behind the increased absorption ofHf from SNEDDS and super-SNEDDS is still unknown.Traditionally, it is thought that the amount of lipids present inthe gastrointestinal fluids is positively correlated to theabsorption due to increased solubilization capacity. However,Table I. Amount of Free Fatty Acids (FFA) generated fromhydrolysis of the lipids from the formulations during dynamic in vitrolipolysis after back litrationμmol digested (total) % digested (total)SNEDDS 228.8±23.1a 55.4±5.7aSNEDDS+orlistat 1.3±2.2a 0.3±0.5aSuper-SNEDDS 180.1±4.6b 80.4±2.2bSuper-SNEDDS+orlistat 13.5±26.0b 6.2±11.7bThe data represents mean±SEM, n=3. The same letters indicatesignificant differencesSNEDDS self-nanoemulsifying drug delivery systemFig. 1. Amount of free fatty acids (FFA) generated from hydrolysisof the lipids in the SNEDDS (grey circles), the super-SNEDDS (blackcircles), the SNEDDS with orlistat (grey diamonds), and the super-SNEDDS with orlistat (black diamonds) during dynamic in vitrolipolysis. The data represents mean±SD, n=3Michaelsen et al.although super-SNEDDS in the current study only providehalf the amount of lipids with respect to the drug concentra-tion compared to the SNEDDS, the overall performancemeasured from the PK parameters is still similar or evenimproved. Thus, the solubilizing capacity of the digestedlipids does not seem to be the only factor important for theabsorption of HF. Based on the XRPD in vitro examination,the solid-state form of the drug precipitate during lipolysis isamorphous. It can thus be speculated that the increasedprecipitation of the drug in an amorphous form from thesuper-SNEDDS is positively correlated to increased absorp-tion, due to promotion of local supersaturation close to theintestinal lumen upon re-dissolution of the amorphousprecipitate. This is in agreement with a previous in vivo studyon dogs (5).The GI physiology of dogs and that of rats differgreatly, but the obtained data indicates that the rat is alsoa suitable model for evaluation of the performance ofsuper-SNEDDS.Orlistat EffectConsidering the effect of orlistat, as seen in Fig. 3 andTable II, tmax is significantly longer for both the SNEDDS (6.3±1.2 h) and the super-SNEDDS (6.0±1.0 h) with orlistatcompared to the SNEDDS (2.8±1.2 h) and super-SNEDDS(1.3±0.1 h) without orlistat. Orlistat reduces the Cmax for boththe SNEDDS and super-SNEDDS; however, this reductiondoes not reach significant levels (P=0.13).Thus, orlistat changes the pharmacokinetic parameterswhen added to both SNEDDS and super-SNEDDS, but thetotal bioavailability does not change significantly (P>0.05).The super-SNEDDS containing orlistat have a larger AUCthan the other groups, but due to the relatively largevariability, there is no statistical difference.In a previous study applying orlistat as a tool, de Smidtet al. found that the digestibility of medium-chaintriacylglycerides and surfactant-based formulations did notsignificantly influence the absorption of penclomidine in rats,which is in agreement with the data presented here. However,when using medium-chain triglyceride solution as a vehicle,the absorption was to a higher degree dependent on digestion(9). It is well known that medium-chain lipids to a lesserFig. 2. Relative amount of Hf found in the pellet (black) and the aqueous phase (grey) at time zero and after 60 min ofin vitro lipolysis for the SNEDDS and super-SNEDDS, with and without orlistat. The data has been normalized andrepresents the mean±SEM, n=3. *indicates significant differences between the data points obtained after 60 minFig. 3. Plasma concentration of Hf following oral administration as aSNEDDS and b super-SNEDDS to SD rats. Treatment includedSNEDDS (black circles), SNEDDS with orlistat (black squares),super-SNEDDS (black circles), and super-SNEDDS with orlistat(black squares). The data represents mean±SEM, n=6Importance of Drug Load and Digestion on Absorption from SNEDDSdegree are dependent on digestion compared to LC lipids(26); therefore, it is difficult to compare these data directlywith the those of the current study, in which LC lipids wereused. Extended tmax can in some cases be ascribed to gastro-retentivity; however, since orlistat previously has beenreported to accelerate gastric emptying, the increased tmaxcan therefore be attributed to the effect of digestion onabsorption from the SNEDDS (27).Comparison of In Vivo and In Vitro DataIn vitro lipolysis is often used to assess lipid formulationssuch as SNEDDS (28). Traditionally, it has been assumed thatprecipitation during in vitro lipolysis will lead to decreasedin vivo absorption. On these terms, the current in vitro datacannot fully explain the obtained in vivo data; the increasedprecipitation in the case of the super-SNEDDS would lead tothe expectation that the SNEDDS would perform betterin vivo than the super-SNEDDS, which is not the case. Eventhough in vitro lipolysis can be useful for predicting thein vivo behavior, the model does not take absorption intoconsideration and also the extent of in vivo precipitation is, tothe authors’ knowledge, still unknown. This is in accordancewith a previous study by Thomas et al., also studying super-SNEDDS with Hf, but in a dog model. In this case, thefindings were explained by the solid-state properties ofprecipitated Hf, which precipitates in an amorphous form(5). From the current study, it was shown that amorphousprecipitation of Hf is most likely not due to a complexformation with digestion products (since digestion iscompletely blocked in the presence of orlistat). The amor-phous nature of the precipitate is thus either an inherentproperty of the drug or due to complex formation with othercomponents being present in the digestion medium. In thiscontext, it is worth mentioning that Humberstone et al.suggested the possibility of the formation of a Hf-taurocholate complex (29).It is evident from the effect of orlistat on the pharmaco-kinetic parameters (tmax, Cmax, and t½) that absorption of Hffrom SNEDDS is affected by digestion, whereas the bioavail-ability is not changed significantly. The traditional mechanismproposed for absorption of poorly water-soluble drugs fromSNEDDS involves the formation of different colloidalstructures and mixed micelles, which will facilitate transportof the drug to the epithelial membrane (30). The formation ofthese structures is highly dependent on the action of thedigestive lipases. In the current study, the in vitro data showsalmost complete blockage of the digestion by addition of 1%(w/w) orlistat to the SNEDDS, which suggest that this alsocan be the case in vivo. It is therefore possible that colloidalstructures are not generated when orlistat is present, andtherefore, some other mechanisms of absorption will beimportant. It can be speculated that the absorption is merelyfacilitated by the nanoemulsion droplets, serving as areservoir containing the drug in a predissolved form.Partitioning of Hf from the nanoemulsion droplets to theexisting bile salt micelles will, in this case, enable absorption.However, it is also possible that the in vivo lipase activityis not fully inhibited by the amount of orlistat dosed; orlistat isinhibiting the pancreatic lipase at a stoichiometry of around1:1 (31). In the rat, pancreatic lipase is continuously secreted(32). It is therefore also possible that the orlistat inhibition isovercome by the continuous secretion of pancreatic lipase.This will lead to a reduced and/or delayed digestion of theSNEDDS, which can also result in the observed changes ofthe pharmacokinetic parameters, e.g., the delayed tmax. In thiscase, orlistat will not lead to changes in the absorptionmechanism but merely in the pharmacokinetics.CONCLUSIONThe present study has demonstrated that the super-SNEDDS principle is also effective in rats; a significantlyhigher Cmax of Hf was found when dosed in a super-SNEDDScompared to a conventional SNEDDS with the same dose ofdrug.The pharmacokinetic parameters change when co-dosingSNEDDS and super-SNEDDS with orlistat; the Cmax de-creases and the absorption phase seems to be extended,especially for the super-SNEDDS. However, these changesdo not influence the overall bioavailability. Thus, orlistat canpossibly be used as a tool to change pharmacokinetics ofdrugs dosed in SNEDDS. It can be speculated that theincreased absorption of poorly soluble drugs, often seen fromSNEDDS, is less than previously thought due to digestion ofthe lipids. It is possible that the partitioning directly from thenanoemulsion droplet to the bile salt micelles facilitates theabsorption; however, further studies are needed in order tounderstand the mechanism of absorption from SNEDDS.ACKNOWLEDGMENTSThe authors would like to thank Yang Hwan Yun forhelp with the in vitro lipolysis data. Funding for this projectwas provided by the Canadian Institute of Health Research toKMW. A special thanks to University of Copenhagen,Table II. Pharmacokinetic Parameters of HF Following Oral Administration to Male SD Rats in Different SNEDDSParameter SNEDDS SNEDDS+orlistat Super-SNEDDS Super-SNEDDS+orlistatCmax (ng/mL) 506±112a 295±53 964±167a 793±123tmax (h) 2.8±1.2a 6.3±1.2a 1.3±0.1b 6.0±1.0bt½ 13.5±2.1 21.2±3.9 13.5±1.7 34.2±17.3AUC(0–30) (h·ng/mL) 6262±1367 6092±1409 9142±787 12946±1590Fa (%)a 22.5±6.3 21.9±6.5 32.9±3.6 45.5±7.3Data represents mean±SEM, n=6. Numbers with different letters indicate significant differences (ANOVA with Šídák’s multiple comparisontest with α 0.05). Values with the same letters indicate significant differences. a and b indicate the differences found with Šídák’s posttestSNEDDS self-nanoemulsifying drug delivery system, AUC area under the curveaAbsolute bioavailability Fa (%) was determined as follows: Fa=100*(AUCPO·Div)/(AUCiv·DPO)Michaelsen et al.Denmark, and University of British Columbia, Canada, forproviding a platform for collaboration.Compliance with Ethical StandardsThe animal protocols used in this study were approved by theUniversity of British Columbia’s Animal Care Committeeand conform to the Canadian Council on Animal Careguidelines.REFERENCES1. Stegemann S, Leveiller F, Franchi D, de Jong H, Lindén H.When poor solubility becomes an issue: from early stage to proofof concept. Eur J Pharm Sci. 2007;31(5):249–61.2. Amidon G, Lennernäs H, Shah V, Crison J. A theoretical basisfor a biopharmaceutic drug classification: the correlation ofin vitro drug product dissolution and in vivo bioavailability.Pharm Res. 1995;12(3):413–20. English.3. Humberstone AJ, Charman WN. Lipid-based vehicles for theoral delivery of poorly water soluble drugs. Adv Drug Deliv Rev.1997;25(1):103–28.4. Christensen JØ, Schultz K, Mollgaard B, Kristensen HG,Mullertz A. Solubilisation of poorly water-soluble drugs duringin vitro lipolysis of medium- and long-chain triacylglycerols. EurJ Pharm Sci. 2004;23(3):287–96.5. Thomas N, Holm R, Müllertz A, Rades T. In vitro and in vivoperformance of novel supersaturated self-nanoemulsifying drugdelivery systems (super-SNEDDS). J Control Release.2012;160(1):25–32.6. Thomas N, Holm R, Garmer M, Karlsson J, Müllertz A, RadesT. Supersaturated self-nanoemulsifying drug delivery systems(super-SNEDDS) enhance the bioavailability of the poorlywater-soluble drug simvastat in in dogs . AAPS J.2013;15(1):219–27. English.7. Thomas N, Richter K, Pedersen T, Holm R, Müllertz A, RadesT. In vitro lipolysis data does not adequately predict the in vivoperformance of lipid-based drug delivery systems containingfenofibrate. AAPS J. 2014;16(3):539–49. English.8. Bloedow DC, Hayton WL. Effects of lipids on bioavailability ofsulfisoxazole acetyl, dicumarol, and griseofulvin in rats. J PharmSci. 1976;65(3):328–34.9. de Smidt PC, Campanero MA, Trocóniz IF. Intestinal absorptionof penclomedine from lipid vehicles in the conscious rat:contribution of emulsification versus digestibility. Int J Pharm.2004;270(1–2):109–18.10. Hofmann AF, Borgström B. The intraluminal phase of fatdigestion in man: the lipid content of the micellar and oil phasesof intestinal content obtained during fat digestion and absorp-tion. J Clin Investig. 1964;43(2):247–57.11. Yeap YY, Trevaskis NL, Quach T, Tso P, Charman WN, PorterCJH. Intestinal bile secretion promotes drug absorption fromlipid colloidal phases via induction of supersaturation. MolPharm. 2013;10(5):1874–89.12. Yano K, Masaoka Y, Kataoka M, Sakuma S, Yamashita S.Mechanisms of membrane transport of poorly soluble drugs: roleof micelles in oral absorption processes. J Pharm Sci.2010;99(3):1336–45.13. Fatouros D, Deen GR, Arleth L, Bergenstahl B, Nielsen F,Pedersen J, et al. Structural development of self nano emulsifyingdrug delivery systems (SNEDDS) during in vitro lipid digestionmonitored by small-angle X-ray scattering. Pharm Res.2007;24(10):1844–53. English.14. Weibel EK, Hadvary P, Hochuli E, Kupfer E, Lengsfeld H.Lipstatin, an inhibitor of pancreatic lipase, produced by Strepto-myces toxytricini. I. Producing organism, fermentation, isolationand biological activity. J Antibiot (Tokyo). 1987;40(8):1081–5.15. Tiss A, Lengsfeld H, Carrière F, Verger R. Inhibition of humanpancreatic lipase by tetrahydrolipstatin: further kinetic stud-ies showing its reversibility. J Mol Catal B Enzym.2009;58(1–4):41–7.16. Fernandez E, Borgström B. Effects of tetrahydrolipstatin, alipase inhibitor, on absorption of fat from the intestine of the rat.Biochim Biophys Acta. 1989;1001(3):249–55.17. Tiss A, Lengsfeld H, Hadváry P, Cagna A, Verger R. Transfer oforlistat through oil–water interfaces. Chem Phys Lipids.2002;119(1–2):41–9.18. Tiss A, Ransac S, Lengsfeld H, Hadvàry P, Cagna A, Verger R.Surface behaviour of bile salts and tetrahydrolipstatin at air/water and oil /water interfaces. Chem Phys Lipids.2001;111(1):73–85.19. Porter CJH, Charman SA, Charman WN. Lymphatic trans-port of halofantrine in the triple-cannulated anesthetized ratmodel: effect of lipid vehicle dispersion. J Pharm Sci.1996;85(4):351–6.20. Thomas N, Müllertz A, Graf A, Rades T. Influence of lipidcomposition and drug load on the in vitro performance of self-nanoemulsifying drug delivery systems. J Pharm Sci.2012;101(5):1721–31.21. Zangenberg NH, Müllertz A, Kristensen HG, Hovgaard L. Adynamic in vitro lipolysis model: I. Controlling the rate oflipolysis by continuous addition of calcium. Eur J Pharm Sci.2001;14(2):115–22.22. Zangenberg NH, Müllertz A, Gjelstrup Kristensen H, HovgaardL. A dynamic in vitro lipolysis model: II: evaluation of themodel. Eur J Pharm Sci. 2001;14(3):237–44.23. Holm R, Porter CH, Müllertz A, Kristensen H, Charman W.Structured triglyceride vehicles for oral delivery of halofantrine:examination of intestinal lymphatic transport and bioavailabilityin conscious rats. Pharm Res. 2002;19(9):1354–61. English.24. Karpf D, Holm R, Kristensen H, Müllertz A. Influence of thetype of surfactant and the degree of dispersion on the lymphatictransport of halofantrine in conscious rats. Pharm Res.2004;21(8):1413–8. English.25. Humberstone AJ, Currie GJ, Porter CJH, Scanlon MJ, CharmanWN. A simplified liquid chromatography assay for the quantita-tion of halofantrine and desbutylhalofantrine in plasma andidentification of a degradation product of desbutylhalofantrineformed under alkaline conditions. J Pharm Biomed Anal.1995;13(3):265–72.26. Chow BPC, Shaffer EA, Parsons HG. Absorption of triglycer-ides in the absence of lipase. Can J Physiol Pharmacol.1990;68(4):519–23.27. Borovicka J, Schwizer W, Guttmann G, Hartmann D, KosinskiM, Wastiel C, et al. Role of lipase in the regulation ofpostprandial gastric acid secretion and emptying of fat inhumans: a study with orlistat, a highly specific lipase inhibitor.Gut. 2000;46(6):774–81.28. Dahan A, Hoffman A. Use of a dynamic in vitro lipolysis modelto rationalize oral formulation development for poor watersoluble drugs: correlation with in vivo data and the relationshipto intra-enterocyte processes in rats . Pharm Res.2006;23(9):2165–74. English.29. Humberstone AJ, Porter CJH, Charman WN. A physicochem-ical basis for the effect of food on the absolute oral bioavailabil-ity of halofantrine. J Pharm Sci. 1996;85(5):525–9.30. MacGregor KJ, Embleton JK, Lacy JE, Perry EA, SolomonLJ, Seager H, et al. Influence of lipolysis on drug absorptionfrom the gastro-intestinal tract. Adv Drug Deliv Rev.1997;25(1):33–46.31. Hadvary P, Lengsfeld H, Wolfer H. Inhibition of pancreaticlipase in vitro by the covalent inhibitor tetrahydrolipstatin.Biochem J. 1988;256:357–61.32. Erlanson-Albertsson C, Larsson A, Duan R. Secretion ofpancreatic lipase and colipase from rat pancreas. Pancreas.1987;2(5):531–5.Importance of Drug Load and Digestion on Absorption from SNEDDS      Appendix II Paper II: Michaelsen MH, Abdi I, Wasan KM, Müllertz A, Rades T. Lipase inhibition does not change bioavailability of fenofibrate from SNEDDS and super-SNEDDS. (In preparation).  1  Lipase inhibition does not change bioavailability of fenofibrate from 1 SNEDDS and super-SNEDDS 2 Running title: 3 Maria Høtoft Michaelsen1,2*, Ismahan Mahad Abdi1*, Kishor M. Wasan2,3, Anette Müllertz1,4,5, and 4 Thomas Rades1. 5 *These authors have contributed equally to this study. 6 1 Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, 7 Universitetsparken 2, DK-2100, Copenhagen, Denmark .  8 2 Faculty of Pharmaceutical Sciences, University of British Columbia, 2405 Wesbrook Mall, 9 Vancouver, British Columbia, Canada V6T 1Z3.  10 3 College of Pharmacy and Nutrition, University of Saskatchewan, 104 Clinic Place, Saskatoon, 11 Saskatchewan, Canada S7N 2Z4.  12 4 Bioneer:FARMA, Department of Pharmacy, Faculty of Health and Medical Sciences, University 13 of Copenhagen, DK-2100, Copenhagen, Denmark .  14 5 To whom correspondence should be addressed. (e-mail: anette.mullertz@sund.ku.dk) 15 2  Abstract 16 The effect of drug load and digestion on absorption was evaluated in vitro and in vivo for a super-17 saturated self-nanoemulsifying drug delivery system (super-SNEDDS) containing 150% fenofibrate 18 at equilibrium solubility (Seq) and a self-nanoemulsifying drug delivery system (SNEDDS) 19 containing 75% fenofibrate at Seq. The lipase inhibitor orlistat (tetrahydrolipstatin) was incorporated 20 into the SNEDDS at 1 % (w/w) in order to assess the effect of digestion on absorption. In vitro 21 lipolysis revealed that adding orlistat decreased digestion from 78.0 ± 5.4 % and 49.5 ± 0.9 % to 22 15.0 ±1.1 % and 12.0 ± 0.5 % for the SNEDDS and the super-SNEDDS respectively during the 60 23 minute lipolysis. Precipitation was significantly larger for the super-SNEDDS (87.01 ±2.9 %) than 24 for the SNEDDS (20.04 ± 0.8%) and the addition of orlistat reduced drug precipitation significantly 25 for the super-SNEDDS. Solid state analysis using XRPD and PLM revealed that fenofibrate 26 precipitated in a crystalline form from both SNEDDS and super-SNEDDS during in vitro lipolysis. 27 Fenofibrate precipitated from the SNEDDS in the form of large diamond shaped crystals whereas 28 the fenofibrate from the super-SNEDDS precipitated in form of elongated needleshaped crystals. In 29 the in vivo studies performed in rats the super-SNEDDS had a significantly higher Cmax (13.34 ± 30 4.66 µg/mL and 7.7 ± 1.7 µg/mL respectively) and AUC (148.0 ± 47.5 h µg/mL and 88.3 ± 20.9 h 31 µg/mL respectively) than the SNEDDS . Orlistat did not affect the absorption of fenofibrate from 32 super-SNEDDS and SNEDDS. These findings suggest that the super-SNEDDS concept is valid for 33 fenofibrate dosed in SNEDDS in a rat model and that digestion does not play a major role for the 34 absorption of fenofibrate from SNEDDS. 35 Key words: absorption, digestion, fenofibrate, orlistat, supersaturation, super-36 SNEDDS, SNEDDS  37 3  Introduction  38 Administration of drugs by the oral route is often preferred due to the ease of manufacturing of oral 39 dosage forms and high patient compliance. However, with an increasing number of poorly water 40 soluble drugs from the drug discovery pipelines of the pharmaceutical industry, oral drug delivery 41 faces significant challenges. Drugs belonging to class II of the biopharmaceutics classification 42 system (BCS) often have a poor and variable absorption resulting in an inconsistent bioavailability 43 due to the physicochemical properties of the drugs leading to low aqueous solubility or poor 44 dissolution rates [1, 2]. Low aqueous solubility and poor dissolution rates can be overcome 45 however, by “enabling” drug delivery systems. Lipid based drug delivery systems such as self-46 nanoemulsifying drug delivery systems (SNEDDS) belong to this type of enabling formulations and 47 have the advantage that the drug is dissolved in the SNEDDS preconcentrate and thus is 48 administered already in solution. Hence, the dissolution step, required if the drug is administered in 49 a solid form, is avoided. SNEDDS are isotropic mixtures of oils, surfactants and co-solvents that 50 form nanoemulsions upon gentle agitation in an aqueous environment, such as in the 51 gastrointestinal tract, and they are typically administered in soft gelatin capsules. Administration of 52 BCS class II drugs in SNEDDS thus typically results in increased bioavailability [3].  53 In supersaturated SNEDDS (super-SNEDDS) the drug is in solution in the SNEDDS preconcentrate 54 above its equilibrium solubility (Seq) and supersaturation is a way of increasing the thermodynamic 55 activity of the drug [4, 5]. Super-SNEDDS provide an alternative to conventional SNEDDS where 56 the dose of a drug to be administered is limited by the solubility of the drug in the formulation and 57 is often between 50% and 90% of Seq in the SNEDDS preconcentrate [6]. This can result in the need 58 for administration of several capsules in order to administer the required dose. In recent studies by 59 Thomas et al. super-SNEDDS have been found to be superior compared to SNEDDS yielding the 60 same or higher bioavailabilities for halofantrine, simvastatin, and fenofibrate compared to SNEDDS 61 in mini pigs and dogs [7-9]. In agreement with this, in a rat study by Michaelsen et al. super-62 SNEDDS with halofantrine also yielded a higher bioavailability compared to SNEDDS with the 63 same dose administered but only half the formulation volume [10].  64 Lipid digestion and the physico-chemical behavior of the digestion products have been studied 65 thoroughly since the early 1960s [11]. Since SNEDDS are at least in part composed of digestible 66 excipients, digestion is thought to be one of the key factors affecting the bioavailability of a drug 67 when dosed in a SNEDDS. Upon digestion of the digestible components in a SNEDDS, a variety of 68 4  colloidal structures with varying solubilizing capacity for the drug will form [12-15]. The current 69 hypothesis regarding absorption from SNEDDS revolves around the formation of mixed colloidal 70 species such as vesicles and micelles that enhance drug solubilization in the gastro intestinal (GI) 71 tract and hence decrease the risk of drug precipitation. The drug solubilized in the colloidal 72 structures is in a rapidly reestablished equilibrium with the free drug, which gives an absorption 73 advantage compared to traditional dissolution from a conventional solid oral dosage form. The 74 current thinking is that the mixed micelles disintegrate due to the lower pH found near the unstirred 75 water layer of the epithelial membranes of the GI tract, and that the fraction solubilized in the mixed 76 micelles therefore is as easy accessible as the free fraction [16, 17].  77 Orlistat (tetrahydrolipstatin) is known as a potent inhibitor of pancreatic lipase, gastric lipase, and 78 carboxyl ester lipase and has been studied extensively throughout the last decades [18-21]. Orlistat 79 is known to inhibit triglyceride digestion when dissolved in the triglycerides and is thought to 80 inhibit the lipase at a 1:1 molar ratio [22]. In the present study Orlistat was applied as a tool to 81 inhibit the digestion of the lipids present in the SNEDDS in order to evaluate the effect of lipid 82 digestion on absorption of the drug. In a previous study where orlistat was co-dosed with 83 halofantrine in SNEDDS the bioavailability of halofantrine was not affected. However, Cmax 84 decreased and the absorption phase was extended. The importance of digestion in absorption from 85 SNEDDS was questioned and an alternative hypothesis regarding absorption from SNEDDS was 86 proposed. When digestion is inhibited the colloidal structures formed from the digestion products 87 will not be generated and the lipid emulsion droplets more or less stay intact. Absorption may 88 therefore be facilitated by diffusion from the lipid droplets acting as a reservoir. Also the 89 amorphous nature of the halofantrine precipitate studied after dynamic in vitro lipolysis was 90 speculated to play a role with regard to the absorption [10]. In a study by de Smidt et al. also using 91 orlistat to investigate the effect of digestion on absorption of penclomidine, similar findings were 92 reported when using emulsified lipid based formulations, but when dosing crude medium chain 93 triglycerides digestion was found to be important for absorption [23].  94 The aim of this study was two-fold: first, to further evaluate if the super-SNEDDS concept is valid 95 in a rat model and secondly, to investigate whether digestion plays a role on the absorption of the 96 crystalline precipitating poorly water soluble drug fenofibrate from SNEDDS and super-SNEDDS. 97 Materials and methods 98 5  Materials 99 Fenofibrate, fenofibric acid, clofibric acid, soybean oil (long-chain (LC) glycerides), 4-100 bromophenyl-boronic acid (BBBA), bovine bile extract, tris-(hydroxymethyl) aminomethane (tris), 101 maleic acid, calcium chloride, sodium hydroxide, and porcine pancreatic lipase were obtained from 102 Sigma-Aldrich (St. Louis, MO, USA). Maisine 35-1 (a mixture of LC mono-, di-, and triglycerides) 103 was kindly donated by Gattefossé (St. Priest, France), and Kolliphor RH 40 was donated by BASF 104 (Ludwigshafen, Germany). Orlistat was purchased from Molekula (Newcastle Upon Tyne, UK). 105 Purified water was obtained from a Millipore Milli-Q Ultra Pure water purification system 106 (Billerica, MA, USA). All other chemicals were of analytical grade and were used as received 107 unless specified otherwise. 108 Preparation of formulations 109 SNEDDS preconcentrates were prepared as previously described [9, 10, 24].  The composition of 110 the SNEDDS was soybean oil (27.5%), Maisine 35-1 (27.5%), Kolliphor RH40 (35%) and absolute 111 ethanol (10%). Briefly, molten Maisine 35-1 was mixed with soybean oil (1:1). Molten Kolliphor 112 RH40 was then added to the lipids and mixed until homogenous. After cooling to ambient 113 temperature ethanol was added. The preconcentrates were stirred until they appeared visually 114 homogenous. Drug loaded SNEDDS were prepared by adding fenofibrate into glass vials and 115 subsequently adding the preconcentrate. The equilibrium solubility of fenofibrate in the SNEDDS 116 preconcentrate is 88.5 mg/g [8]. Fenofibrate was added to the SNEDDS and the super-SNEDDS at 117 75% and 150% of the equilibrium solubility (Seq), respectively. The SNEDDS were stirred with a 118 magnetic stirring bar at room temperature to aid the drug dissolution process. The super-SNEDDS 119 were ultrasonicated for 30 min and then heated to 60oC for 3 hours and left overnight to cool at 120 37oC. Complete dissolution of the drug in the super-SNEDDS was confirmed by polarized light 121 microscopy. Orlistat was used in a final concentration of 1 % (w/w) and dissolved in the 122 preconcentrates at room temperature upon stirring. 123 Dynamic in vitro lipolysis 124 Dynamic in vitro lipolysis was carried out under fasted state conditions as previously described [10, 125 24-26]. The SNEDDS were weighed into a thermostated vessel containing 25 mL intestinal 126 lipolysis medium (2.5 mM bovine bile salt, 0.26 mM phospholipid, 2 mM Tris buffer, 2 mM maleic 127 acid, and 50 mM sodium chloride). Lipolysis was initiated by the addition of 5 mL freshly prepared 128 6  pancreatic lipase to the intestinal lipolysis medium. The pancreatic lipase was prepared by weighing 129 the crude porcine pancreatic lipase into a polypropylene tube and adding Milli-Q water. The 130 mixture was vortexed until homogenous and then centrifuged (7 min, 6500 x g, 37°C). The pH of 131 the supernatant was adjusted to 6.5 resulting in a final lipase activity of 550 USP U/mL. The in 132 vitro lipolysis was controlled by continuous addition of calcium (0.6 M; 0.045 mmol/min of a 133 solution with CaCl2) throughout the 60 min run-time of the experiment. The released free fatty acids 134 were continuously titrated with NaOH (0.4 M) by an automated pH-stat (Metrohm Titrino 744, 135 Tiamo Version 1.3, Switzerland) to keep the pH constant at 6.5. Samples (1 mL) were withdrawn 136 from the thermostated vessel at the time points 0, 15, 30, 45 and 60 min. Lipase activity was 137 immediately inhibited by the addition of 5 µL 4-BBBA (1 M in methanol) followed by 138 centrifugation (18,255 x g, 30 min at 37°C). This resulted in separation of the clear supernatant 139 from the pellet. The phases were quantified for their drug content by HPLC following appropriate 140 dilution with methanol (see below). The solid state properties of fenofibrate in the pellet were 141 analyzed by XRPD and PLM after 60 min lipolysis (see below). The lipolysis was terminated after 142 60 min and a back titration was performed at pH 9 to determine the exact amount of released free 143 fatty acids. 144 Quantification of lipolysis samples 145 The amount of fenofibrate in the aqueous and pellet phases after centrifugation and dilution was 146 determined by isocratic high-performance liquid chromatography (HPLC) using a Phenomenex 147 Kinetex 5u C18 100A column (100 x 4.6 mm) protected by a Phenomenex guard column 148 (Phenomenex, Torrance, CA, USA). The mobile phase consisted of 85% (v/v) methanol and 15% 149 (v/v) purified water, the flow rate was 0.9 mL/min, and the injection volume was 20 µL. The 150 detection wavelength was 288 nm, and the drug retention time was approximately 2.4 min. All 151 samples were analyzed at room temperature using a Dionex ASI-100 automated sample injector, 152 P680 HPLC pump, and PDA-100 photodiode array detector (Thermo Fisher Scientific, Waltham, 153 MA, USA).  The chromatograms were evaluated using Thermo Scientific Dionex Chromeleon 7 154 Chromatography Data System (Thermo Fisher Scientific, Waltham, MA, USA). 155 X-ray powder diffraction 156 The solid state properties of the pellets were analyzed after 60 min of lipolysis by X-ray powder 157 diffraction (XRPD) using a X´Pert Pro X-ray diffractometer (MPD PW3040/60 XRD, PANalytical, 158 7  Almelo, The Netherlands). The samples were scanned from 5°-35° 2θ with a step size of 0.0260 ° 159 2θ and a speed of 0.0336o/s. As a radiation source CuKα radiation (l.542 Å) was used and Kβ 160 radiation was eliminated using a nickel filter.  The system was operated with a voltage of 40 kV and 161 a current of 30 mA. Each measurement lasted approximately 10 min. The data was collected using 162 X’Pert data collector version 2.2 and were analyzed with X’Pert highscore plus version 2.2.4 (both 163 from PANalytical B.V.). 164 Polarized light microscopy 165 Microscopic analysis of the pellet was conducted using a Zeiss Axiolab microscope (Carl Zeiss, 166 Göttingen, Germany). The pellets were observed under cross polarizers. Polarized light micrographs 167 were obtained by using a moticam 10.0 MP digital camera and Motic Images Plus 2.0 Software 168 (Hong Kong, China). Super-SNEDDS and SNEDDS were observed with a magnification of 40x 169 and 5x, respectively. 170 In vivo study 171 The in vivo study protocols were approved by the animal welfare committee appointed by the 172 Danish ministry of justice and conform to the NIH guidelines on animal welfare. The study was 173 carried out in compliance with EC directive 86/609/EEC and with the Danish law regulating animal 174 experiments. Male Sprague Dawley rats (268-290 g) obtained from Harlan (Horst, Netherlands) 175 were divided randomly into 6 groups of 6 animals each. The animals were allowed to acclimatize 176 for 7 days prior to the experiments and were kept on a standard feed with free access to water.  Prior 177 to the study the rats were fasted for 6 hours.  The animals had free access to water during the 178 experiment and were allowed food 4 hours after initiation of the experiment. All animals received 8 179 mg/kg fenofibrate by oral gavage. All formulations were pre-emulsified with MilliQ water before 180 the dosing (10 % (w/w) formulation). Blood samples were withdrawn from the tail vein at 30 min 181 and 1, 2, 3, 4, 6, 8, 10, 24 and 30 hours after administration and collected into 0.5 mL EDTA coated 182 tubes.  Plasma was separated from the samples immediately by centrifugation at (15 min, 4 ° C, 183 1651 x g) and stored at − 80 ° C until analysis. The rats were euthanized 30 hours after dosing. 184 Bioanalysis 185 The preparation of plasma samples was based on a method modified from Berthelsen et al. [27]. 186 This method was originally adapted from Hanafy et al. [28]. The process was carried out as follows; 187 8  50 µL plasma was mixed with 20 µL internal standard (clofibric acid in methanol, 200 µg/mL) 188 followed by addition of 100 µL acetonitrile. The mixture was briefly vortexed and subsequently 189 ultrasonicated in an ultrasonic water bath for 10 min. The samples were stored at − 20 °C for 10 190 min, followed by centrifugation (14 min, 20,227 x g) at 0 ° C. The clear supernatant was transferred 191 into HPLC vials and analyzed by HPLC. The samples for the standard calibration curve were 192 prepared similar to the plasma samples with spiked blank plasma. The standard curves were run in 193 the concentration range 0.06− 20.0 µg/mL and were linear over the entire range.  The concentration 194 of fenofibric acid in the plasma samples was determined by isocratic HPLC using a Phenomenex 195 Kinetex 5u XB-C18 100A column (100 x 4.6 mm) (Phenomenex, Torrance, CA, USA) using a 196 Dionex ASI-100 automated sample injector, P680 HPLC pump, and PDA-100 photodiode array 197 detector (Thermo Fisher Scientific, Waltham, MA, USA). The chromatograms were evaluated using 198 the Thermo Scientific Dionex Chromeleon 7 Chromatography Data System (Thermo Fisher 199 Scientific, Waltham, MA, USA). The mobile phase consisted of 68% methanol and 32% formic 200 acid (0.1%). The flow rate was 0.6 mL/min and the injection volume 20 µL. Fenofibric- and 201 clofibric acid were measured at 287 nm with approximate retention times of 8 and 4 min, 202 respectively. 203 Pharmacokinetic analysis 204 All pharmacokinetic parameters were calculated using WinNonlin Professional Version 6.3 205 (Pharsight Coporation, Mountainview, CA, USA). The area under the curve (AUC) was determined 206 using the linear trapezoidal model from t = 0 h to t = 30 h. AUC0-30h, maximum plasma 207 concentration (Cmax), time to reach the maximum plasma concentration (Tmax), and the half-life (t½) 208 were determined from the individual plasma curves. All data was normalized to dose. 209 Statistical analysis 210 Statistical analysis was performed using GraphPad Prism 6.03 (GraphPad Software, San Diego, 211 CA, USA). Data is expressed as mean ± standard deviation (SD) or standard error of the mean 212 (SEM). The data was analyzed using ANOVA followed by Šídáks posttest (α=0.05). 213 Results 214 Dynamic in vitro lipolysis 215 9  SNEDDS and super-SNEDDS were subjected to dynamic in vitro lipolysis in order to determine the 216 degree of digestion and to evaluate the capacity of the SNEDDS to solubilize fenofibrate throughout 217 the 60 min lipolysis. In Figure 1 the percentage of released fatty acids from SNEDDS and super-218 SNEDDS with and without orlistat is shown. The total percentage of released fatty acids after 60 219 min in vitro lipolysis is 78.0  ± 5.4 % and 49.5  ± 0.9 % for SNEDDS and super-SNEDDS 220 respectively. The addition of orlistat leads, as expected, to a significant reduction in the fatty acids 221 released from both SNEDDS and super-SNEDDS to 15.0 % ± 1.1 and to 12.0 % ± 0.5, respectively. 222  223 Figure 1. Percentage of released free fatty acids from the SNEDDS (▲), super-SNEDDS (), 224 SNEDDS with orlistat (●), and super-SNEDDS with orlistat ( ) during 60 min dynamic in vitro 225 lipolysis. Data represents mean ± SD, n=3. 226 Figure 2 illustrates the relative distribution of fenofibrate in the aqueous phase (grey) and the pellet 227 (black).  In Figure 2a the drug distribution of fenofibrate in the aqueous and pellet phase for 228 SNEDDS is shown. During the course of lipolysis precipitation of fenofibrate is gradually 229 increasing and a significant difference between 0 min and 30-60 min is found (p<0.05). The same is 230 seen for the super-SNEDDS (Figure 2b) however, precipitation during the 60 min lipolysis is more 231 pronounced. There is a significant difference between the precipitation of fenofibrate at 0 min and 232 15-60 min (p<0.0001) for the super-SNEDDS. For the SNEDDS and super-SNEDDS with orlistat 233 precipitation remains constant throughout the lipolysis and there is no significant difference when 234 comparing 0 min and 60 min. 235 10  Comparing the SNEDDS and the super-SNEDDS (Figure 2a and 2b), it can be seen that 236 significantly more fenofibrate precipitates during lipolysis of the super-SNEDDS. For the SNEDDS 237 with and without orlistat no difference is seen between the amounts of precipitated fenofibrate after 238 60 min of lipolysis. For the super-SNEDDS with and without orlistat, there is a significant 239 difference between the amount of precipitated fenofibrate after 60 min of lipolysis. 240  241 Figure 2. The relative amount of fenofibrate in the pellet (black) and the aqueous phase (grey) 242 during lipolysis of SNEDDS (a), super-SNEDDS (b), SNEDDS with orlistat (c) and super- 243 SNEDDS with orlistat (d). Values with the same letters indicate significant differences between the 244 groups and * indicate significant difference from precipitated amount at 0 minutes.  (ANOVA with 245 Šídák’s multiple comparison test with α= 0.05). Data has been normalized and represents mean ± 246 SEM, n=3. 247 Solid state characterization of pellets 248 11  The solid state properties of the pellets after dynamic in vitro lipolysis were investigated by XRPD 249 and PLM. Pure fenofibrate corresponding to the crystalline form I (Figure 3a) exhibited several 250 characteristic reflections between 13 and 32 o2θ with the most pronounced reflections found at 251 approximately 21-22 o2θ [29]. The characteristic diffraction pattern of fenofibrate form I was also 252 visible in the diffractograms of the pellets from the super-SNEDDS (figure 3c) and the spiked 253 sample (figure 3g). In the pellets from the SNEDDS, the SNEDDS with orlistat and the super-254 SNEDDS with orlistat, no crystalline fenofibrate was detected. To confirm the presence of 255 crystalline fenofibrate PLM was utilized. Crystalline precipitate was confirmed to be present in both 256 the SNEDDS and the super-SNEDDS. The morphology of the crystals were however smaller and 257 more needlelike and elongated for the super-SNEDDS than for the SNEDDS. No precipitation was 258 observed for the SNEDDS and super-SNEDDS containing orlistat. 259  260 Figure 3. XRPD diffractograms and polarized light micrographs of crystalline fenofibrate (a), pellets 261 obtained from SNEDDS (b), pellets obtained from super-SNEDDS (c), pellets obtained from SNEDDS with 262 Orlistat (d), pellets obtained from super-SNEDDS with Orlistat (e), pellets obtained from blank SNEDDS (f)  263 and blank pellets spiked with fenofibrate (g). All samples were taken after 60 min in vitro lipolysis and 264 centrifugation.   265 In vivo study 266 To investigate the absorption of fenofibrate from SNEDDS and super-SNEDDS an in vivo rat study 267 was conducted in order to evaluate the effect of drug load and digestion on absorption. Following 268 12  oral administration, fenofibrate is metabolized to fenofibric acid in the enterocytes. The plasma 269 concentration curves of fenofibric acid are depicted in figure 4. Fenofibrate was dosed at 8 mg/kg 270 body weight resulting in in 34.3 mg lipid/kg for the SNEDDS and 17.2 mg lipid/kg for the super-271 SNEDDS. 272  273  274 Figure 4. Plasma concentration of fenofibrate following oral administration of SNEDDS (a) and 275 super-SNEDDS (b). The treatments included SNEDDS ( ), SNEDDS with orlistat ( ), super-276 SNEDDS (), and super-SNEDDS with orlistat (▲). Data represents mean ± SEM, n= 6.   277 All SNEDDS showed a fast absorption and a complete elimination of the drug 30 hours after 278 administration. A noncompartmental model was used to estimate the pharmacokinetic parameters 279 including, Cmax, Tmax, t½ and AUC0-30h of fenofibrate dosed in SNEDDS and super-SNEDDS with 280 and without orlistat (table 1). Cmax and AUC for the super-SNEDDS (13.3 ± 4.7 and 148 ± 47.5 281 h*µg/mL respectively) are significantly higher than for the SNEDDS (7.4 ± 1.7 µg/mL and 88.3 ± 282 20.9 h*µg/mL respectively) (p<0.05). No significant differences are observed when comparing 283 SNEDDS and super-SNEDDS to the SNEDDS and super-SNEEDS containing orlistat (p>0.05). 284 The Tmax of super-SNEEDS appears shorter in the presence of orlistat; however, there was no 285 significant difference (p>0.05).  286 Table 1: Pharmacokinetic parameters of fenofibrate following oral administration of fenofibrate to 287 male Sprague Dawley rats in SNEDDS and super-SNEDDS. 288 13   SNEDDS SNEDDS with orlistat Super-SNEDDS Super-SNEDDS with orlistat Cmax (µg/mL) 7.36±1.7a 6.54±1.1 13.34±4.66a,b 13.73±3.1 Tmax (h) 2.33±1.4 2.92±1.9 4.33±2.3 2.67±0.8 t½ (h) 4.17±0.6 4.97±1.2 5.08±1.2 4.78±0.3 AUC (h* µg/mL) 88.28±20.9a 66.3±14.9 148.00±47.5a, b 136.96±27.5 Data represents mean±SD, n = 6. Values with the same letters indicate significant differences (ANOVA with Šídák’s 289 multiple comparison test with α 0.05).  290 Discussion 291 An increased absorption when dosing poorly water soluble drugs in lipid based drug delivery 292 systems such as SNEDDS compared to traditional dosage forms is well documented and super-293 SNEDDS have proven to provide a good strategy to increase the absorption even further from 294 SNEDDS [7, 8, 24]. Digestion has been speculated to play a critical role for absorption of these 295 types of systems. However, the exact mechanism of the observed increase in drug absorption for 296 both SNEDDS and super-SNEDDS is still unknown. 297 The current study set out to investigate the effect of super-saturation and digestion on the absorption 298 of the poorly water soluble drug fenofibrate. Fenofibrate was chosen as the model drug, since it 299 precipitates crystalline during in vitro lipolysis. In a previous study where the amorphous 300 precipitating drug halofantrine was dosed to rats in SNEDDS and super-SNEDDS, it was 301 speculated that the reason for the observed effects in vivo such as decreased Cmax and a prolonged 302 absorption phase when digestion was inhibited were, at least in part, due to the solid state 303 characteristics of the precipitate [10]. 304 In vitro lipolysis 305 In vitro lipolysis is frequently employed during development of lipid formulations such as 306 SNEDDS in order to evaluate the degree of digestibility and the ability of the digesting SNEDDS to 307 keep the drug solubilized [30]. One common assumption is that the better the formulation can keep 308 the drug solubilized during the digestion process the higher the bioavailability [31]. From Figure 2 a 309 and b it is evident that fenofibrate is much more prone to precipitate from the super-SNEDDS 310 compared to the SNEDDS. After 60 minutes of lipolysis 20.1 ± 0.8 % fenofibrate has precipitated 311 from the SNEDDS whereas 87.1 ± 2.9% has precipitated from the super-SNEDDS. This is partly 312 due to the fact that the system is in an unfavorable thermodynamic state but also because of the loss 313 of solubilization capacity during the digestion. During the lipolysis the lipid components of the 314 14  SNEDDS will be digested, resulting in the formation of various and transient colloidal structures 315 composed of digestion products and components of the biorelevant media (bile salts and 316 phospholipids). Lipid digestion products are usually more polar than undigested lipids hence the 317 solubilization capacity decrease during the course of the lipolysis. Since the dose of fenofibrate has 318 been kept constant in these experiments the amount of lipids present for the super-SNEDDS is only 319 half of that of the SNEDDS, which plays a role for the subsequent solubilization capacity during 320 digestion due to the lower level of colloidal structures generated.  321 Orlistat is known as a potent lipase inhibitor of both gastric, pancreatic and carboxyl ester lipase 322 and from the in vitro lipolysis results with orlistat it is clear that orlistat is inhibiting digestion.  323 [19]. In the presence of orlistat the level of precipitated fenofibrate is constant throughout the 324 digestion process since the lipase activity is inhibited and the formulations thus retain their 325 solubilizing capacity.  326 In vivo study 327 Based on the assumption that only the solubilized fraction of fenofibrate is available for absorption 328 it would be expected that the SNEDDS would lead to a higher bioavailability of fenofibrate than the 329 super-SNEDDS. However,  recent studies employing super-SNEDDS, as a novel drug delivery 330 system for poorly water soluble drugs, have shown that they give rise to an equal or higher 331 bioavailability compared to SNEDDS [7, 8, 24]. These studies have been performed in both dogs 332 and minipigs and more recently the performance of super-SNEDDS has been studied in a rat model 333 [10]. In the study by Thomas et al. SNEDDS, super-SNEDDS, and SNEDDS-suspensions with 334 fenofibrate were tested in minipigs and all formulations yielded comparable plasma curves despite 335 large differences in the in vitro data [8]. From the in vivo data in the present study it is evident that 336 the super-SNEDDS give rise to both a higher Cmax but also AUC compared to the SNEDDS, which 337 is in contrast to the aforementioned study by Thomas et al. [8]. To which extent the precipitation of 338 fenofibrate dosed in SNEDDS occurs in vivo is unknown [32]. However, from the PLM 339 micrographs it is evident that the morphology of the precipitates from the SNEDDS and the super-340 SNEDDS differ. The fenofibrate in the SNEDDS precipitates as larger diamond shaped crystals, 341 whereas fenofibrate precipitates from the super-SNEDDS consist of more elongated needle-shaped 342 crystals. This difference in crystal morphology may be important for the interpretation of the in vivo 343 studies if precipitation in fact occurs in vivo. Hanafy et al. show in a study with micronized 344 fenofibrate (5 µm in mean particle size) and solid lipid nanoparticles (58 nm in mean particle size) 345 15  loaded with fenofibrate that particle size is important for the subsequent bioavailability [28]. From 346 the Noyes-Whitney equation it is clear that a reduction in particle size will give an increase in 347 dissolution rate [33]. The smaller fenofibrate crystals seen in the precipitate for the super-SNEDDS 348 may therefore quickly be re-dissolved and hence play a positive role with regard to the increased 349 Cmax and AUC.  350 Digestion is believed to play a major role for the absorption of poorly water soluble drugs from 351 SNEDDS [34-36]. However, in the present study no significant differences are seen when co-dosing 352 fenofibrate in SNEDDS or super SNEDDS with or without the lipase inhibitor orlistat. In a recent 353 study by the authors, inhibiting the digestion with SNEDDS loaded with halofantrine, it was also 354 found that digestion did not play a role with regard to absorption, however, inhibition of digestion 355 changed the pharmacokinetic parameters [10]. The colloidal structures formed upon digestion have 356 been thought to play a major role with regard to the trafficking of drug from the lumen to the 357 absorptive membrane, however, the present data indicate that formation of colloidal structures such 358 as vesicles and mixed micelles may not have a significant influence on fenofibrate absorption. In a 359 study by Yeap et al. cinnarizine absorption from a model colloidal system containing long chain 360 fatty acids was investigated using amiloride as a fatty acid absorption inhibitor in an intestinal 361 perfusion set-up in rats [16]. They found that in the presence of amiloride, cinnarizine absorption 362 decreased significantly and thus concluded that formulations containing absorbable lipids would be 363 more effective in enhancing absorption of poorly water soluble drugs. In contrast, de Smidt et al. 364 showed that digestion did not play a critical role for absorption of penclomidine from lipid based 365 formulations containing medium chain lipids co-dosed with orlistat [23]. It can be speculated that 366 the increased absorption seen from SNEDDS compared to conventional dosage forms is merely a 367 question of keeping the drug dissolved in the nanoemulsion droplets. When digestion is inhibited 368 the nanoemulsion droplets will continue to exist throughout the gastrointestinal transit and thus 369 provide a constant solubilizing compartment for the drug. Partitioning of fenofibrate from the 370 nanoemulsion droplets to the endogenous colloidal structures such as bile salt micelles will thus 371 facilitate absorption.  372 Conclusion 373 The present study demonstrates that super-SNEDDS loaded with fenofibrate have a superior in vivo 374 performance compared to SNEDDS in a rat model. The super-SNEDDS had both a significantly 375 higher Cmax and AUC compared to the SNEDDS. This increased performance of super-SNEDDS 376 16  compared to SNEDDS has been shown previously in a range of animal models with different poorly 377 water soluble drugs. In the studies utilizing orlistat it was found that absorption of the poorly water 378 soluble drug fenofibrate from SNEDDS is not dependent on digestion. It is proposed that the 379 partitioning of the drug from the intact nanoemulsion droplets to the endogenous bile salt micelles 380 facilitates the absorption, however, further studies are needed in order to understand the underlying 381 mechanism behind absorption from SNEDDS. 382 Acknowledgements 383 Funding for this project was given by the Canadian Institute of Health Research granted to KMW. 384 A special thanks goes to University of British Columbia, Canada and University of Copenhagen for 385 providing a platform for collaboration. 386 References 387 1. Stegemann, S., et al., When poor solubility becomes an issue: From early stage to proof of concept. 388 European Journal of Pharmaceutical Sciences, 2007. 31(5): p. 249-261. 389 2. Amidon, G., et al., A Theoretical Basis for a Biopharmaceutic Drug Classification: The Correlation of 390 in Vitro Drug Product Dissolution and in Vivo Bioavailability. Pharmaceutical Research, 1995. 12(3): 391 p. 413-420. 392 3. Humberstone, A.J. and W.N. Charman, Lipid-based vehicles for the oral delivery of poorly water 393 soluble drugs. Advanced Drug Delivery Reviews, 1997. 25(1): p. 103-128. 394 4. Brouwers, J., M.E. Brewster, and P. Augustijns, Supersaturating drug delivery systems: The answer 395 to solubility-limited oral bioavailability? Journal of Pharmaceutical Sciences, 2009. 98(8): p. 2549-396 2572. 397 5. Gao, P., et al., Characterization and optimization of AMG 517 supersaturatable self-emulsifying 398 drug delivery system (S-SEDDS) for improved oral absorption. Journal of Pharmaceutical Sciences, 399 2009. 98(2): p. 516-528. 400 6. Christensen, J.Ø., et al., Solubilisation of poorly water-soluble drugs during in vitro lipolysis of 401 medium- and long-chain triacylglycerols. European Journal of Pharmaceutical Sciences, 2004. 23(3): 402 p. 287-296. 403 7. Thomas, N., et al., Supersaturated Self-Nanoemulsifying Drug Delivery Systems (Super-SNEDDS) 404 Enhance the Bioavailability of the Poorly Water-Soluble Drug Simvastatin in Dogs. The AAPS 405 Journal, 2013. 15(1): p. 219-227. 406 8. Thomas, N., et al., In Vitro Lipolysis Data Does Not Adequately Predict the In Vivo Performance of 407 Lipid-Based Drug Delivery Systems Containing Fenofibrate. The AAPS Journal, 2014. 16(3): p. 539-408 549. 409 9. Thomas, N., et al., In vitro and in vivo performance of novel supersaturated self-nanoemulsifying 410 drug delivery systems (super-SNEDDS). Journal of Controlled Release, 2012. 160(1): p. 25-32. 411 10. Michaelsen, M., et al., The Effect of Digestion and Drug Load on Halofantrine Absorption from Self-412 nanoemulsifying Drug Delivery System (SNEDDS). The AAPS Journal, 2016. 18(1): p. 180-186. 413 11. Hofmann, A.F. and B. Borgström, The Intraluminal Phase of Fat Digestion in Man: The Lipid Content 414 of the Micellar and Oil Phases of Intestinal Content Obtained during Fat Digestion and Absorption. 415 Journal of Clinical Investigation, 1964. 43(2): p. 247-257. 416 17  12. Fatouros, D., et al., Structural Development of Self Nano Emulsifying Drug Delivery Systems 417 (SNEDDS) During In Vitro Lipid Digestion Monitored by Small-angle X-ray Scattering. Pharmaceutical 418 Research, 2007. 24(10): p. 1844-1853. 419 13. Fatouros, D.G., B. Bergenstahl, and A. Mullertz, Morphological observations on a lipid-based drug 420 delivery system during in vitro digestion. European Journal of Pharmaceutical Sciences, 2007. 31(2): 421 p. 85-94. 422 14. Fatouros, D., et al., Colloidal Structures in Media Simulating Intestinal Fed State Conditions with and 423 Without Lipolysis Products. Pharmaceutical Research, 2009. 26(2): p. 361-374. 424 15. Warren, D.B., et al., Real time evolution of liquid crystalline nanostructure during the digestion of 425 formulation lipids using synchrotron small-angle X-ray scattering. Langmuir, 2011. 27(15): p. 9528-426 9534. 427 16. Yeap, Y.Y., N.L. Trevaskis, and C.J. Porter, Lipid absorption triggers drug supersaturation at the 428 intestinal unstirred water layer and promotes drug absorption from mixed micelles. Pharmaceutical 429 research, 2013. 30(12): p. 3045-3058. 430 17. Shiau, Y.F., Mechanism of intestinal fatty acid uptake in the rat: the role of an acidic microclimate. 431 The Journal of Physiology, 1990. 421: p. 463-474. 432 18. Weibel, E.K., et al., Lipstatin, an inhibitor of pancreatic lipase, produced by Streptomyces toxytricini. 433 I. Producing organism, fermentation, isolation and biological activity. The Journal of antibiotics, 434 1987. 40(8): p. 1081-1085. 435 19. Borgström, B., Mode of action of tetrahydrolipstatin: a derivative of the naturally occurring lipase 436 inhibitor lipstatin. Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism, 1988. 962(3): 437 p. 308-316. 438 20. Fernandez, E. and B. Borgström, Effects of tetrahydrolipstatin, a lipase inhibitor, on absorption of 439 fat from the intestine of the rat. Biochimica et Biophysica Acta (BBA) - Lipids and Lipid Metabolism, 440 1989. 1001(3): p. 249-255. 441 21. Sternby, B., et al., Degree of in vivo inhibition of human gastric and pancreatic lipases by Orlistat 442 (Tetrahydrolipstatin, THL) in the stomach and small intestine. Clinical Nutrition, 2002. 21(5): p. 395-443 402. 444 22. Hadváry, P., et al., The lipase inhibitor tetrahydrolipstatin binds covalently to the putative active site 445 serine of pancreatic lipase. Journal of Biological Chemistry, 1991. 266(4): p. 2021-2027. 446 23. de Smidt, P.C., M.A. Campanero, and I.F. Trocóniz, Intestinal absorption of penclomedine from lipid 447 vehicles in the conscious rat: contribution of emulsification versus digestibility. International Journal 448 of Pharmaceutics, 2004. 270(1–2): p. 109-118. 449 24. Thomas, N., et al., Influence of lipid composition and drug load on the In Vitro performance of self-450 nanoemulsifying drug delivery systems. Journal of Pharmaceutical Sciences, 2012. 101(5): p. 1721-451 1731. 452 25. Zangenberg, N.H., et al., A dynamic in vitro lipolysis model: I. Controlling the rate of lipolysis by 453 continuous addition of calcium. European Journal of Pharmaceutical Sciences, 2001. 14(2): p. 115-454 122. 455 26. Zangenberg, N.H., et al., A dynamic in vitro lipolysis model: II: Evaluation of the model. European 456 Journal of Pharmaceutical Sciences, 2001. 14(3): p. 237-244. 457 27. Berthelsen, R., et al., Kolliphor Surfactants Affect Solubilization and Bioavailability of Fenofibrate. 458 Studies of in Vitro Digestion and Absorption in Rats. Molecular Pharmaceutics, 2015. 12(4): p. 1062-459 1071. 460 28. Hanafy, A., et al., Pharmacokinetic evaluation of oral fenofibrate nanosuspensions and SLN in 461 comparison to conventional suspensions of micronized drug. Advanced Drug Delivery Reviews, 462 2007. 59(6): p. 419-426. 463 29. Heinz, A., et al., Understanding the solid-state forms of fenofibrate – A spectroscopic and 464 computational study. European Journal of Pharmaceutics and Biopharmaceutics, 2009. 71(1): p. 465 100-108. 466 18  30. Larsen, A.T., P. Sassene, and A. Müllertz, In vitro lipolysis models as a tool for the characterization of 467 oral lipid and surfactant based drug delivery systems. International Journal of Pharmaceutics, 2011. 468 417(1–2): p. 245-255. 469 31. Dahan, A. and A. Hoffman, Use of a Dynamic in Vitro Lipolysis Model to Rationalize Oral 470 Formulation Development for Poor Water Soluble Drugs: Correlation with in Vivo Data and the 471 Relationship to Intra-Enterocyte Processes in Rats. Pharmaceutical Research, 2006. 23(9): p. 2165-472 2174. 473 32. Sassene, P., et al., In vivo precipitation of poorly water soluble drugs from lipid based drug delivery 474 systems. In preparation, 2016. 475 33. Noyes, A.A. and W.R. Whitney, The Rate of Solution of Solid Substances in Their Own Solutions. 476 Journal of the American Chemical Society, 1897. 19(12): p. 930-934. 477 34. Iwanaga, K., et al., Disposition of Lipid-Based Formulation in the Intestinal Tract Affects the 478 Absorption of Poorly Water-Soluble Drugs. Biological and Pharmaceutical Bulletin, 2006. 29(3): p. 479 508-512. 480 35. Yeap, Y.Y., et al., Intestinal Bile Secretion Promotes Drug Absorption from Lipid Colloidal Phases via 481 Induction of Supersaturation. Molecular Pharmaceutics, 2013. 10(5): p. 1874-1889. 482 36. Yano, K., et al., Mechanisms of membrane transport of poorly soluble drugs: Role of micelles in oral 483 absorption processes. Journal of Pharmaceutical Sciences, 2010. 99(3): p. 1336-1345. 484      Appendix III Paper III: Sassene PJ*, Michaelsen MH*, Mosgaard MD, Jensen MK, Van Den Broek E,  Wasan KM, Mu H, Rades T, Müllertz A. In vivo precipitation of poorly soluble drugs from lipid based drug delivery systems. (Submittedto Molecular Pharmaceutics). *PJS and MHM are joint first authors on this publication.   In vivo precipitation of poorly soluble drugs from lipid based drug delivery systems P.J. Sassene *1, M.H. Michaelsen*1,2, M.D. Mosgaard1, M.K. Jensen1, E. Van Den Broek1, K.M. Wasan2,3, H. Mu1, T. Rades1 and A. Müllertz1.   *These authors contributed equally to the work.  1:Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, Copenhagen, 2100, Denmark.  2:Faculty of Pharmaceutical Sciences, University of British Columbia, 2146 East Mall, Vancouver, BC, Canada V6T 1Z3. 3. College of Pharmacy and Nutrition, University of Saskatchewan, E3122-104 Clinic Place, Saskatoon, SK, Canada S7N 2Z4.               Corresponding author: Professor Thomas Rades, thomas.rades@sund.ku.dk, Department of Pharmacy, Faculty of Health and Medical Sciences, University of Copenhagen, Universitetsparken 2, Copenhagen, 2100, Denmark. Abstract Objective: Precipitation of poorly water-soluble drugs from lipid-based drug delivery systems (LbDDS) has been studied extensively during in vitro lipolysis, but has never been shown in vivo. The aim of this study was therefore to investigate if drug precipitation can occur from LbDDS during transit of the gastrointestinal-tract in vivo. Methods: Rats were administered 300 µL of either of two LbDDS (LbDDS I and LbDDS II) loaded with danazol or fenofibrate.  The rats were euthanized at various time points after administration of both LbDDS containing either drug and the contents of the stomach and proximal part of the small intestine were harvested. The contents were analyzed for crystalline drug by x-ray powder diffraction (XRPD) and polarized light microscopy (PLM). Results: No drug precipitation was evident in the stomach or the intestine after administration of LbDDS I containing danazol at the tested time points. Fenofibrate precipitation was absent in the stomach initially after administration of LbDDS I, but was evident in the stomach 90 min after dosing. No crystalline fenofibrate was observed in the intestine. Danazol and fenofibrate precipitation was evident in the stomach following administration of LbDDS II containg either drug, but not in the intestine at the tested time point. Conclusion: Drug precipitation from LbDDS was observed in the stomach, but not in the intestine, which is contrary to what in vitro lipolysis data of the tested formulations suggests. Thus precipitation of drugs from LbDDS seems to an artifact of the in vitro lipolysis model. Introduction Lipid-based drug delivery systems (LbDDS) have been used throughout the last three decades to improve the oral bioavailability of poorly water-soluble drugs [1-8]. A general concern with LbDDS is the inherent digestibility of most lipids throughout the gastrointestinal (GI) passage [9-11]. Digestion will modify the lipid excipients and thus the composition of the LbDDS, potentially altering the solubilizing capacity of the formulation for the drug [12]. Attempts have been made to capture the effect of digestion on drug solubilization by in vitro lipolysis models simulating the digestive environment of the GI-tract. Most of these models are only assessing the intestinal digestion as the majority of lipid digestion occurs in the small intestine [7, 11, 13-21]. The current hypothesis is that the fraction of drug solubilized in the aqueous phase, is the fraction available for absorption [2, 12, 22]. Several studies have shown a correlation between the amount of drug solubilized and the bioavailability. A study by Larsen et al. 2008 investigated the bioavailability in rats of danazol in Labrafil M2125CS at different concentrations relative to the saturation solubility of danazol in the formulation. The authors found a correlation between lower drug loadings and higher bioavailabilities. They attributed this finding to less drug precipitating from the formulations during transit in GI-tract. This interpretation was supported by in vitro lipolysis data, showing more precipitation with higher drug loads [2]. This paradigm has, however, recently been questioned by Thomas et al. (2012), who found extensive precipitation of the poorly water soluble drug halofantrine during in vitro digestion of a supersaturated self nano-emulsifying drug delivery system (super-SNEDDS), but a superior bioavailability compared to a conventional SNEDDS dosed with halofantrine showing less precipitation in vitro (same dose, twice the amount of vehicle) [23]. The discrepancy between in vitro drug precipitation and in vivo bioavailability was explained by the solid state of the precipitate, since halofantrine was shown to precipitate in an amorphous form. Hence the solid state of the precipitated drug may also play an important role in drug absorption and has therefore been explored thoroughly throughout recent years [7, 10, 24-28]. Another study by Thomas and coworkers (Thomas et al. (2014)), investigated a super-SNEDDS loaded with fenofibrate as model drug [29]. In this case the same trend as in Thomas et al. (2012) was found; A more extensive precipitation from the super-SNEDDS than from the conventional SNEDDS, but a better in vivo bioavailability. Fenofibrate was, however, shown to precipitate in a crystalline form during the in vitro lipolysis and the solid state of the precipitated drug could therefore not be responsible for the superior bioavailability. Other studies with fenofibrate loaded LbDDS show the same discrepancy between in vivo bioavailability and drug precipitation in vitro [30, 31]. These studies indicate that the current approach to interpret in vitro lipolysis data might be inadequate or that the in vitro lipolysis models are not predictive and drug precipitation in vivo might not occur to the same extent as currently believed.  Even though in vitro drug precipitation has been studied extensively, no studies, to the best of the authors knowledge, have been looking at drug precipitation from LbDDS in vivo. However, in order to study in vivo precipitation of drugs from LbDDS, the location of the LbDDS in the GI-tract is essential. Thus the rate of gastric emptying of the tested LbDDS is crucial and can be determined in man by magnetic resonance imaging (MRI) and telemetry using radiolabelled material [32-34]. Both methods are cumbersome and require specialized and expensive equipment. The rate of gastric emptying has, however, also been studied via the plasma concentration-time curve of paracetamol following oral administration, since paracetamol is readily absorbed in the proximal small intestine [35]. The paracetamol plasma concentration-time curve has been shown to correlate well with telemetry studies and does therefore provide an alternative to more elaborate methods [35]. Thus the aim of the current study was firstly, to investigate the gastric emptying rate following oral administration of two LbDDS loaded with paracetamol. Secondly, to investigate if precipitation of the two poorly soluble drugs danazol and fenofibrate (both previously shown to precipitate crystalline during in vitro lipolysis) occurred in vivo following oral administration of the drugs dissolved in LbDDS to rats. Drug precipitation was investigated by euthanizing rats at predetermined time points relative to the gastric emptying rate to ensure samples containing LbDDS in the stomach and LbDDS in the small intestine.  The content of the stomach and proximal small intestine was collected and analyzed by x-ray powder diffraction (XRPD) and polarized light microscopy (PLM) to identify the presence of crystalline drug.   Materials and methods Materials Sesame oil (S3547), oleic acid (75096), 4-bromobenzene boronic acid (4-BBBA) (B75956, ≥95.0%), porcine pancreatin (P1625), tris(hydroxymethyl)aminomethane (T8 760-2), bovine serum albumin (A7906), porcine bile extract (B8631), sodium taurodeoxycholate (T0875), maleic acid (M0375), protease inhibitor cocktail (P8340), and fenofibrate (F-6020) were purchased from Sigma Aldrich (St Louis, MO, USA). Egg phospholipid was purchased from Lipoid (Ludwigshafen, Germany) and both NaCl (≥99.5%) and CaCl2 (99.0-102.0%) were obtained from Merck (Darmstadt, Germany). Kolliphor EL and Kolliphor RH40 were donated by BASF (Ludwigshafen, DE). Brij 97 was purchased from Croda (Rawcliffe Bridge, UK). Transcutol HP was generously donated by Gattefossé (Saint Priest, France). Danazol USP was obtained from Unikem A/S (Copenhagen, Denmark). Paracetamol was donated by Fagron (Copenhagen, Denmark).  The water used in this study was purified by a SG ultra clear ultraviolet (UV) water purifier (SG Wasseraufbereitung und Regenerierstation, Barsbüttel, Germany). All other chemicals were of analytical or chromatographic grade. Preparation of formulations Two previously developed LbDDS were prepared in order to assess the precipitation [3, 11]. LbDDS I was composed of sesame oil (20.6 % w/w), Kolliphor RH40 (45 % w/w), oleic acid (15.4 % w/w), Brij 97 (9 % w/w) and absolute ethanol (10 % w/w). The Kolliphor RH 40, Brij 97 and oleic acid were heated until melted and hereafter mixed with the rest of the excipients except ethanol. The mixture was inspected for homogeneity and after evaluation of the mixture ethanol was added. Danazol or fenofibrate were added at 80% of their saturation solubilities (20.7 mg/g and 69.5 mg/g respectively). For the gastric emptying rate studies paracetamol was added at 100% equilibrium solibility (52.2 mg/g).  LbDDS II was composed of Transcutol HP and Kolliphor EL (1:1). Danazol and fenofibrate were added at 80 % of their saturation solubilities (52.4 mg/g and 151 mg/g respectively). For the gastric emptying rate studies paracetamol was again added at 100% equilibrium solibility (100.1 mg/g). Pharmacokinetic studies of paracetamol Male Sprague Dawley rats (10 weeks, approximately 300 g) were purchased from Harlan (Horst, NL). All animals were acclimatized for 6-10 days prior to the experiments and allowed free access to water and kept on a standard diet. The animal care and experiments were approved by the Animal Welfare Committee appointed by the Danish Ministry of justice and conducted and conform to the EC Directive 2010/63/EU, the Danish law regulating animal experiments and NIH guidelines for the care and use of laboratory animals. Gastric emptying The rate of gastric emptying for LbDDS I and LbDDS II was determined by loading the formulations with paracetamol. Before initiation of the experiment the rats were fasted for 10-15 hours with access to water. Each group consisted of 6 rats was dosed with 300 µL LbDDS I or LbDDS II by oral gavage equivalent to 52.2 mg/kg and 100.1 mg/kg of paracetamol, respectively. 200 µL blood samples were collected from the tail vein at 15, 30, 45, 60 minutes and 1.5, 2, 2.5, 3, 4 and 6 hours after administration into tubes containing EDTA (Microvette, Nümbrecht, Germany). The blood samples were centrifuged at 6,500 rpm at 4oC (3600 g at rmax) (Eppendorf, Hamburg, Germany), and the plasma was analyzed for paracetamol by an enzymatic colorimetric assay kit (Cambridge Life Sciences, Cambridge, UK). Precipitation in vitro and in vivo  In vitro precipitation studies In vitro intestinal lipolysis studies of LbDDS I and II, containing danazol or fenofibrate were conducted to ensure that drug precipitation occurred in vitro from the tested formulations. The digestions were conducted according to Zangenberg et al. (2001) with minor modifications and samples were treated as previously described by Sassene et al (2010) [10, 14]. Digestions were conducted by adding 1.5 g of LbDDS containing danazol or fenofibrate to 300 ml medium simulating the environment of the small intestine (Table 1). Samples were withdrawn after 0 min, 25 min, 50 min and 80 min of digestion and were treated with 4-bromobenzeneboronic acid (4-BBBA) to quench the hydrolysis. Samples were subsequently centrifuged for 10 min at 15,000 RPM (2.0×104 g (rmax)) in a Heraeus Sepatech Biofuge 15 centrifuge (Osterode, Germany) to separate the digestion phases, and the pellet phase was isolated and analyzed by XRPD (see below). Table 1. Composition of the in vitro lipolysis medium Composition Initial concentration Bile salts 5 mM Phospholipid 1.25 mM Trizma maleate 2 mM Sodium chloride 150 mM     Ca2+ 0.045 mM/min Pancreatic lipase 600 USPunits/mL  In vivo precipitation studies Rats were fasted overnight prior to dosing of 300 µl of LbDDS I or LbDDS II containing either danazol or fenofibrate by oral gavage. The study design is shown in Table 2.  After euthanasia the stomach and small intestine were isolated and removed. The content of the stomach and proximal small intestine (10 cm onwards from the pylorus) was scraped with a spatula into polyethylene tubes and centrifuged in a Thermo Scientific Megafuge 16 R (Osterode, Germany) at 4,000 rpm (2.9×103 g (rmax)) for 10 min. The pellet was subsequently analyzed for crystalline precipitate of danazol or fenofibrate using XRPD and PLM (see below). Diffractograms and micrographs of only one rat representative of each group will be presented throughout the paper, to minimize the data load. Protease inhibitor was added to the supernatant, to avoid degradation of rat lingual lipase and immediately stored on ice. The supernatant was used to determine the lipase activity in the rat’s stomach.  Table 2. The design of the in vivo precipitation study: Number of rats and time of euthanasia. Formulation Euthanasia (5 min) Euthanasia (30 min) Euthanasia (90 min) LbDDS I (danazol) 3 - 3 LbDDS I (fenofibrate) 3 - 3 LbDDS II (danazol) 3 3 - LbDDS II (fenofibrate) 3 3 -  X-ray powder diffraction The pellets obtained from the gastric and intestinal samples were analyzed by XRPD for crystalline precipitate of danazol and fenofibrate. The diffractograms were measured using a Panalytical X’pert Pro diffractometer with a PIXcel detector (PANalytical B.V, Almelo, NL). Samples were placed on aluminum sample holders and scanned from 5 - 35o (2θ) using a Cu Kα (λ=1.54187 Å) radiation source. A nickel filter was used to eliminate Kβ radiation. The scanning speed and step size were 0.0336o/s and 0.0260o 2θ respectively. The applied voltage was 45 kV and the current 40 mA. Data was collected using X’Pert data collector version 2.2 and analyzed using X’pert Highscore plus version 2.2.4 (PANalytical B.V, Almelo, NL). Polarized light microscopy Microscope analysis was performed on the pellets obtained from the gastric and intestinal samples using a Zeiss Axiolab microscope (Carl Zeiss, Göttingen, Germany) equipped with cross polarizers.  The micrographs were captured using a Moticam digital camera and Motic Images Plus software version 2.0 from Motic (Wetzlar, Germany). Lipase activity assay Glyceryl tributyrate (0.5 g) was added to a thermostated glass vessel (37°C) and connected to a Titrando lipolysis system (842 with a 804 Ti stand, Metrohm, Herisau, Switzerland) with a stirrer and pH-electrode. Preheated medium (37°C) as listed in Table 3 was added (14.5 mL) and allowed to equilibrate for 100 seconds before 100µl aspirated rat gastric fluid (supernatant) was added to initiate the experiment. NaOH (0.025 M) was added automatically to the reaction vessel (by an automatic titrator, 800 Dosino, Metrohm, Herisau, Switzerland) throughout the experiment to neutralize the release of free fatty acids in order to maintain the pH at 4.5. The assay was run for 10 min before a backtitration to pH 9 was conducted, for determination of the un-ionized fatty acids. Three blank experiments (no rat gastric fluid) were conducted to test how much NaOH was needed to elevate the pH from 4.5 to 9 in the absence of digestion products. Table 3. Composition of the lipase assay medium Component Concentration (mM) BSA (bovine serum albumin) 1.5 µM NaCl 150 mM Sodium taurodeoxycholate  2 mM  Results Gastric emptying rate of LbDDS I and II The paracetamol plasma concentration-time curves following oral administration of LbDDS I and II are shown in Figure 1. For LbDDS I, a plasma concentration of 35 ± 7 µg/ml  was observed 15 min after administration, which indicated a rapid onset of appearance of paracetamol in plasma. The concentration remained constant throughout the first hour of the experiment, but increased slightly to 42 ± 5 µg/ml after 1.5 hours, which were the Cmax and Tmax, respectively. The paracetamol concentration then declined slowly throughout the rest of the experiment and ended up at 14 ± 4 µg/ml, 6 hours after administration. LbDDS II did also display a rapid onset of paracetamol appearance in plasma with a concentration of 89 ± 23 µg/ml, 15 min after administration. The plasma concentration increased to 99 ± 24 µg/ml, 30 min after administration, which were the Cmax and Tmax, respectively. The plasma concentration then slowly declined throughout the rest of the experiment and ended up at 44 ± 6  µg/ml, 6 hours after administration.  Figure 1. The plasma concentration-time curve of paracetamol following oral administration of 300 µl of LbDDS I ()or LbDDS II (). (mean ± SEM, n=6) Drug precipitation during in vitro lipolysis The diffractograms of the pellets collected after 80 min intestinal in vitro lipolysis of LbDDS I and II loaded with danazol or fenofibrate are shown in Figure 2. It was evident that precipitation of danazol and fenofibrate occurred during digestion of both LbDDS, since reflections corresponding to those of crystalline danazol and fenofibrate, were present in the formulations loaded with danazol and fenofibrate, respectively. Thus precipitation from these formulations is expected to occur in the intestine in vivo, if the in vitro lipolysis model is predictive. The formulations are therefore suitable for the in vivo precipitation study.  Figure 2. The XRPD diffractograms of precipitated danazol and fenofibrate following in vitro digestion of LbDDS I and LbDDS II. I: crystalline danazol, II: pellet from LbDDS I containing danazol, III: pellet from LbDDS II containing danazol, IV: crystalline fenofibrate, V: pellet from LbDDS I containing fenofibrate, VI: pellet from LbDDS II containing fenofibrate. Diffractograms of the pellets obtained during digestion of the LbDDS without drugs showed no reflections corresponding to those from danazol or fenofibrate (data not shown). It is evident that crystalline drug is present in all samples. In vivo precipitation studies In vivo precipitation studies of blank formulations It was initially investigated if crystalline matter was present in the GI-content of the rats, which could influence the detection of precipitated drug. This was studied by analyzing the GI-content following administration of LbDDS I and LbDDS II without drug by XRPD and PLM. No reflections were present in the diffractograms of either the gastric or intestinal contents for any of the formulations, indicating lack of detectable crystalline material (Figure 3a). Thus no reflections from the formulation or the inherent GI-content of the rats are expected to overlap with the reflections from either crystalline danazol or fenofibrate. The PLM micrographs did show a slight birefringence for both the gastric and the intestinal samples for both LbDDS (Figure 3b), indicating that small amounts of crystalline matter were present inherently in the GI-content of the rats.  Figure 3ab. XRPD diffractograms (a) and PLM micrographs (b) of the blank (LbDDS without drug) GI-contents. I: stomach LbDDS I, II: intestine LbDDS I, III: stomach LbDDS II, IV: intestine LbDDS II.  In vivo drug precipitation from LbDDS I The diffractograms of the GI-content following administration of LbDDS I loaded with danazol are shown in Figure 4a. The reflections observed in the diffractogram of crystalline danazol are in agreement with previously published diffractograms of danazol [12]. Danazol precipitation does, however, seem to be absent or very limited in the stomach and the intestine at the tested time points (5 min and 90 min), since no reflections corresponding to those of crystalline danazol were evident. This was also confirmed by the PLM micrographs (Figure 4b, II-V. The morphology of the crystals did furthermore not correspond to those of crystalline danazol (Figure 4b, I) and are likely to originate from the crystalline matter inherently present in the GI-tract of the rats (Figure 3b).   Figure 4ab. XRPD diffractograms (a) and PLM micrographs (b) of the gastrointestinal contents of rats folowing oral administration of 300 µl LbDDS I loaded with danazol. I: crystalline danazol, II: stomach 5 min, III: stomach 90 min, IV: intestine 5 min, V: intestine 90 min. No precipitation of danazol is evident in the stomach or the intestine.  The diffractograms of the GI-content following administration of LbDDS I loaded with fenofibrate are shown in Figure 5a. The reflections observed in the diffractogram of crystalline fenofibrate are in agreement with previously published diffractograms of fenofibrate [29]. Fenofibrate precipitation did not occur in the stomach within the first 5 min, but was evident after 90 min, since reflections corresponding to those of crystalline fenofibrate were present (Figure 5a). No drug precipitation was evident in the small intestine at the tested time points (5 min and 90 min). The PLM micrographs supported the XRPD findings. Only little birefringence (corresponding to the crystals observed in the blank GI-content in Figure 3b) was observed in the gastric sample after 5 min incubation and in both the intestinal samples (Figure 5b, II, IV, V), indicating lack of crystalline fenofibrate. In contrast, the micrograph of the stomach sample after 90 min incubation showed birefringence corresponding to that of dot-shaped crystalline fenofibrate  (crystal size to small to assign a habit) (Figure 5b, I).   Figure 5ab. XRPD diffractograms (a) and PLM micrographs (b) of the gastrointestinal contents of rats following oral administration of 300 µl LbDDS I loaded with fenofibrate. I: crystalline fenofibrate, II: stomach 5 min, III: stomach 90 min, IV: intestine 5 min, V: intestine 90 min.  No precipitation of fenofibrate was evident 5 min after administration either in the stomach or the intestine. Fenofibrate precipitation occured after 90 min in the stomach, but not in the intestine.  In vivo drug precipitation from LbDDS II  The diffractograms of the GI-content following administration of LbDDS II loaded with danazol are shown in Figure 6a. The reflections observed in the stomach samples taken at 5 and 30 min after administration of LbDDS II, correspond to those observed for crystalline danazol. Thus danazol precipitation is evident in the stomach. No reflections were observed in the intestinal samples at the tested time points (5 min and 30 min). The PLM micrographs did again support the XRPD findings (Figure 6b). Crystals were evident in the stomach samples 5 and 30 min after administration of LbDDS II. Though the crystal form was the same, as confirmed by the XRPD micrographs, the morphology of the precipitated danazol crystals changed from dot-shaped in the original powder, dissolved in the LbDDS I (crystal size to small to assign a habit), to rod-shaped crystals. Few crystals were present in the intestinal samples, but they were not rod-shaped and did most likely not originate from danazol.    Figure 6ab. XRPD diffractograms (a) and PLM micrographs (b) of the gastrointestinal contents of rats following oral administration of 300 µl LbDDS II loaded with danazol. I: crystalline danazol, II: stomach 5 min, III: stomach 30 min, IV: intestine 5 min, V: intestine 30 min.  Danazol precipitation was evident in the stomach both 5 min and 30 min after administration. No drug precipitation was evident in the intestine at any of the tested time points. The diffractograms of the GI-content following administration of LbDDS II loaded with fenofibrate are shown in Figure 7a. The reflections observed in the gastric samples 5 and 30 min after administration of LbDDS II, correspond to those observed for crystalline fenofibrate, indicating drug precipitation in the stomach. No reflections were observed in the intestinal samples at the tested time points (5 min and 30 min). The PLM micrographs did again support the XRPD findings (Figure 7b). Crystals were evident in the stomach samples 5 and 30 min after administration of LbDDS II. Again the crystal form seemed to be the same, indicating that the repetitive pattern of the crystal lattice is the same, but the morphology of the precipitated fenofibrate changed from dot-shaped (crystal size to small to assign a habit) to rod-shaped crystals. Few crystals were present in the intestinal samples. They were, however, not rod-shaped and were expected to originate from inherent material present in the intestinal lumen.    Figure 7ab. XRPD diffractograms (a) and PLM micrographs (b) of the gastrointestinal contents of rats following oral administration of 300 µl LbDDS II loaded with fenofibrate. I: crystalline fenofibrate, II: stomach 5 min, III: stomach 30 min, IV: intestine 5 min, V: intestine 30 min.   Fenofibrate precipitation was evident in the stomach both 5 min and 30 min after administration. No drug precipitation was evident in the intestine at any of the tested time points.  The pH and lipase activity in the fasted rat stomach The pH in the gastric juice of the fasted rats was measured to be 5.1 ± 0.6 (n=8) and the lipase activity was 12.0 ± 6.5 TBU/ml. Thus lipid digestion of the LbDDS might occur during in vivo incubation of the formulation in the stomach.   Discussion Gastric emptying rate The rate of gastric emptying of LbDDS I and LbDDS II can be extracted from the plasma concentration-time profiles of paracetamol, since paracetamol is rapidly absorbed upon arrival to the proximal small intestine [35]. Thus the paracetamol concentration in plasma will increase with increasing amounts of LbDDS in the intestine and decrease with decreasing amounts. The somewhat constant plasma concentration of paracetamol from 15 min to 3 hours after administration of LbDDS I therefore indicate that the formulation is gradually emptied into the duodenum. This was to be expected, since LbDDS I contains sesame oil and oleic acid, which both are dietary lipids known to delay gastric emptying [36, 37]. The formulation would therefore be present in the intestine throughout the first 3 hours after administration. The time point where most LbDDS I was expected to be in the intestine would be at Tmax= 90 min. Thus, since the highest concentration of drug is present in the intestine at Tmax, the likelihood of seeing drug precipitation in the intestine from LbDDS I would consequently also be highest around Tmax. The rats used for the in vivo precipitation studies of danazol and fenofibrate from LbDDS I, were therefore euthanized 90 min after dosing. For LbDDS II, the plasma concentration-time curves of paracetamol showed a Tmax at 30 min. Thus the retention of gastric emptying seemed less affected by LbDDS II, than LbDDS I. LbDDS II consists of Transcutol HP, which is a non-digestible co-solvent and Kolliphor EL which is polyethoxylated castor oil and only very slightly digestible. Previous studies have shown that only 0.36 mmol of fatty acid is released during intestinal in vitro lipolysis of Kolliphor EL [38]. Thus the caloric content of LbDDS II is significantly less than for LbDDS I, and LbDDS I was therefore also expected to have a slower gastric emptying rate. The rats used for the in vivo precipitation study of danazol and fenofibrate from LbDDS II, were euthanized 30 min after dosing, since the initial rate of gastric emptying was faster than for LbDDS I. Precipitation of both formulations was also analyzed 5 min after administration to assess whether drug precipitation occurred during dispersion in the gastric fluid. Drug precipitation in vitro and in vivo.  The solid-state characteristics of the gastrointestinal content following oral administration of LbDDS I containing danazol were shown in Figure 4a,b. No precipitation of danazol was evident in either the stomach or the intestine at the tested time points. The lack of crystalline drug in the stomach 5 min after administration could be expected, since LbDDS I previously has been shown to be able to maintain drug solubilized during dispersion in a biorelevant medium [10]. Drug precipitation from LbDDS I is triggered by digestion and limited lipolysis was expected to occur within the 5 min incubation in the stomach, since the lipase activity only was 12.0 TBU/ml [10]. The absence of drug precipitation in the stomach at 90 min suggests that the intragastric digestion was insufficient to reduce the solubilizing capacity of the formulation to the point where danazol precipitation occurred. Surprisingly, no precipitation of danazol was evident in the intestinal samples at the tested time points. This was unexpected, since the in vitro lipolysis data showed that danazol precipitated in a crystalline form during digestion of LbDDS I (Figure 2). The lack of precipitation in vivo could be due to drug and formulation transiting beyond the initial 10 cm of the small intestine, where the luminal content were collected. This was, however, unlikely since preliminary studies analyzing the content of the entire small intestine did not show any precipitation either. Another concern could be that danazol precipitation was below the limit of detection (LOD) of the analytical methods used in this study. The LOD of the XRPD apparatus is 1% drug concentration in the investigated solid matter (data not shown). The pellet collected from the fasted rat intestines was weighing approximately 100 mg. Thus if danazol were to precipitate in the intestine then it would be less than 1 mg or less than 16% of the administered dose. The LOD of PLM is, however, much lower than the XRPD as only very few crystals need to be present in order for precipitation to be detectable. Thus if danazol precipitation did occur from LbDDS I at all, it is expected to be very limited. Ideally, drug precipitation in vivo should be co-investigated by a chromatographic technique, such as high performance liquid chromatography (HPLC). However, the GI-content of the rats, even though they were fasted, did still contain wooden fibers (originating from the floor of their cage) to which LbDDS loaded with drug could adsorp. An extraction of drug by organic solvent, for HPLC analysis, could therefore extract drug from the pellet phase, which actually was still in solution in the LbDDS. Thus HPLC analysis was not suitable for drug quantification in the present study.  Fenofibrate precipitation from LbDDS I did not occur in the stomach at 5 min after administration, but was evident after 90 min (Figure 5). This suggests that the gastric digestion of LbDDS I is reducing the solubilizing capacity of the formulation. The only digestible component in LbDDS I is sesame oil, which makes up 20.6% of the formulation. Sesame oil consists of triacylglycerides, primarily containing (>80%) oleic and linoleic acid. Thus the extent of intragastric digestion of LbDDS I can be calculated if it is assumed that the volume of gastric fluid in rats is 1 ml (1 to 2 ml of gastric fluid is secreted per hour) and only two fatty acids from the triacylglycerides in sesame oil need to be the liberated in order to obtain complete lipid digestion [39]. Based on these assumptions, LbDDS I will be completely digested within 11 min incubation in the stomach. It should, however, be noted that most lipases display a higher activity towards shorter chain than long chain triacylglycerides [40, 41]. Thus the rate of sesame oil digestion might be lower than the digestion rate of tributyrin, which was the substrate used to determine the lipase activity. It is furthermore important to keep in mind that gastric lipase activity in man has been shown to be highly dependent on the surface tension in the interphase between oil and water [42]. No data on the surface tension dependency of rat lingual lipase activity can be found in the literature. The type of surfactants present at the interphase can furthermore affect the rate of digestion and differs between the lipase activity assay and the in vivo gastric digestion of LbDDS I [43]. Thus using the lipase activity to determine the absolute extent of digestion in vivo is associated with some uncertainty but does, however, indicate that LbDDS I will undergo gastric digestion in vivo. It is therefore likely that the solubilizing capacity of LbDDS I will change during gastric incubation. For LbDDS I, no detectable fenofibrate precipitation was present in the intestine at the tested time points, which again was surprising, since extensive precipitation was observed in the intestinal in vitro lipolysis study (Figures 2 and 5). If precipitation occurred below the LOD of the XRPD, then it would again be less than 1 mg or 5% of the dose. Thus the amounts of fenofibrate precipitating, if any, should have no significant effect on bioavailability. Precipitation of both danazol and fenofibrate was evident in the stomach following administration of LbDDS II at the tested time points (5 and 30 min) (Figures 6 and 7). Previous in vitro studies also suggest drug precipitation from this formulation during initial dispersion prior to enzymatic degradation [11]. This was most likely caused by partitioning of the hydrophilic co-solvent Transcutol HP to the aqueous phase resulting in a reduced solubilization capacity of the formulation. Despite the extensive precipitation of danazol and fenofibrate in the stomach no precipitated drug was present in the intestine at the tested time points (Figure 6 and 7). Thus the precipitated drug seems to resolubilize in the intestinal environment. Again if precipitation did occur below the LOD of the XRPD apparatus, then only 6% and 2% of the dosed danazol or fenofibrate did precipitate, respectively.  The lack of precipitation in the intestine regardless of formulation and drug suggests that the traditional intestinal in vitro lipolysis model is inadequate in predicting the in vivo bioavailability of LbDDS. This finding is very much in line with previous studies showing no correlation between in vivo bioavailability of fenofibrate, simvastatin and halofantrine and drug precipitation in vitro [23, 28-31]. Thus our study provides an explanation for the discrepancy between the in vivo and in vitro results, as no solid drug seems to be present in the intestine. The lack of drug precipitation from LbDDS I during incubation in the intestine is probably attributed to several different factors. Gradual gastric emptying of formulation to the intestine allowing the small fraction of emptied drug to stay solubilized seems to be a plausible explanation. The gradual gastric emptying is, however, likely to work in combination with drug absorption providing an additive effect to avoid drug precipitation. Thus the continuous removal of drug and lipolysis products from the intestinal lumen by absorption before the solubilizing capacity of the formulation has been reduced by digestion to the point where drug precipitates, is expected to play a major role in avoiding drug precipitation.  Resolubilization of precipitated drug from LbDDS II in the intestine is probably also a combination of gradual gastric emptying and intestinal absorption. It is furthermore worth noticing that the previously mentioned pharmacokinetic studies with simvastatin, fenofibrate and halofantrine in LbDDS were conducted in different species, e.g. dogs, rats and minipigs, and no clear correlation between in vivo bioavailability and in vitro solubilization could be found [23, 28, 29, 31]. Thus the lack of precipitation in the intestine in the present study is not likely to be due to the rat model itself, even though bile is continuously secreted [44].  The doses of danazol (approximately 20 and 50 mg/kg for LbDDS I and II, respectively) and fenofibrate (approximately 70 and 150 mg/kg for LbDDS I and II, respectively) used in this study are in general higher than the doses used in other rat studies of these drugs (28 mg/kg for danazol and 9 mg/kg for fenofibrate) [2, 45]. Thus the likelihood of intestinal precipitation should be higher in our study, than if clinically relevant doses were administered. Intestinal drug precipitation from LbDDS does therefore not seem to be an issue in vivo, but an artifact of the current in vitro lipolysis models. Conclusion In terms of the triacylglyceride containing, digestible LbDDS I, there was no gastric precipitation at 5 min of either danazol og fenofibrate. At 90 min fenofibrate, but not danazol, showed gastric precipitation. Thus intragastric digestion of a LbDDS can be sufficient to cause drug precipitation. No solid, crystalline danazol or fenofibrate was present in the intestine following administration of LbDDS I, which was surprising, since the in vitro intestinal lipolysis data suggested massive drug precipitation. For the surfactant and co-solvent based LbDDS II, both danazol and fenofibrate precipitated crystalline in the stomach, but no crystalline drug was detectable in the intestine, even though massive drug precipitation again had been shown during intestinal in vitro lipolysis. Thus the discrepancy previous studies have seen between no significant difference in bioavailability of LbDDS loaded with fenofibrate, simvastatin and halofantrine, despite differences in drug precipitation in vitro, can be explained by the lack of precipitation in vivo. Thus the traditional intestinal in vitro lipolysis model seems to overpredict the extent of intestinal drug precipitation in vivo and data should therefore be handled with care. In order to make the in vitro lipolysis model more predictive incorporation of a gastric digestion step with gradual emptying to the intestine, as well as an absorptive step simulating drug removal is essential. Acknowledgements  The study was conducted as part of the Oral Biopharmaceutics Tools (OrBiTo) project (http://www.orbitoproject.eu) funded by the Innovative Medicines Initiative (IMI) Joint Undertaking under Grant Agreement No 115369. This project was also partly funded by the Canadian Institute of Health Research to KMW.    References  [1] A. Mullertz, 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. [2] 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. [3] A.T. Larsen, A. Ogbonna, R. Abu-Rmaileh, B. Abrahamsson, J. Ostergaard, A. Mullertz, SNEDDS Containing Poorly Water Soluble Cinnarizine; Development and in Vitro Characterization of Dispersion, Digestion and Solubilization, Pharmaceutics, 4 (2012) 641-665. [4] R. Holm, C.J. Porter, G.A. Edwards, A. Mullertz, H.G. Kristensen, W.N. Charman, Examination of oral absorption and lymphatic transport of halofantrine in a triple-cannulated canine model after administration in self-microemulsifying drug delivery systems (SMEDDS) containing structured triglycerides, Eur.  J. Pharm. Sci., 20 (2003) 91-97. [5] S.M. Khoo, R.J. Prankerd, G.A. Edwards, C.J.H. Porter, W.N. Charman, A physicochemical basis for the extensive intestinal lymphatic transport of a poorly lipid soluble antimalarial, halofantrine hydrochloride, after postprandial administration to dogs, J. Pharm. Sci., 91 (2002) 647-659. [6] C.T. Phan, P. Tso, Intestinal lipid absorption and transport, Front. Biosci., 6 (2001) D299-319. [7] C. Stillhart, D. Durr, M. Kuentz, Toward an Improved Understanding of the Precipitation Behavior of Weakly Basic Drugs from Oral Lipid-Based Formulations, J. Pharm. Sci., 103 (2014) 1194-1203. [8] A. Dahan, A. Hoffman, The effect of different lipid based formulations on the oral absorption of lipophilic drugs: The ability of in vitro lipolysis and consecutive ex vivo intestinal permeability data to predict in vivo bioavailability in rats, Eur. J. Pharm. Biopharm., 67 (2007) 96-105. [9] C.J. Porter, A.M. Kaukonen, B.J. Boyd, G.A. Edwards, W.N. Charman, Susceptibility to lipase-mediated digestion reduces the oral bioavailability of danazol after administration as a medium-chain lipid-based microemulsion formulation, Pharm. Res., 21 (2004) 1405-1412. [10] P.J. Sassene, M.M. Knopp, J.Z. Hesselkilde, V. Koradia, A. Larsen, T. Rades, A. Mullertz, Precipitation of a poorly soluble model drug during in vitro lipolysis: characterization and dissolution of the precipitate, J. Pharm. Sci., 99 (2010) 4982-4991. [11] H.D. Williams, P. Sassene, K. Kleberg, J.C. Bakala-N'Goma, M. Calderone, V. Jannin, A. Igonin, A. Partheil, D. Marchaud, E. Jule, J. Vertommen, M. Maio, R. Blundell, H. Benameur, F. Carriere, A. Mullertz, C.J. Porter, C.W. Pouton, 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. [12] A.T. Larsen, P. Sassene, A. Mullertz, 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. [13] N.H. Zangenberg, A. Mullertz, H.G. Kristensen, L. Hovgaard, A dynamic in vitro lipolysis model. II: Evaluation of the model, Eur.  J. Pharm. Sci., 14 (2001) 237-244. [14] N.H. Zangenberg, A. Mullertz, H.G. Kristensen, L. Hovgaard, A dynamic in vitro lipolysis model. I. Controlling the rate of lipolysis by continuous addition of calcium, Eur.  J. Pharm. Sci., 14 (2001) 115-122. [15] J.C. Bakala-N'Goma, H.D. Williams, P.J. Sassene, K. Kleberg, M. Calderone, V. Jannin, A. Igonin, A. Partheil, D. Marchaud, E. Jule, J. Vertommen, M. Maio, R. Blundell, H. Benameur, A. Mullertz, C.W. Pouton, C.J. Porter, F. Carriere, Toward the establishment of standardized in vitro tests for lipid-based formulations. 5. Lipolysis of representative formulations by gastric lipase, Pharm. Res., 32 (2015) 1279-1287. [16] P. Sassene, K. Kleberg, H.D. Williams, J.C. Bakala-N'Goma, F. Carriere, M. Calderone, V. Jannin, A. Igonin, A. Partheil, D. Marchaud, E. Jule, J. Vertommen, M. Maio, R. Blundell, H. Benameur, C.J. Porter, C.W. Pouton, A. Mullertz, Toward the establishment of standardized in vitro tests for lipid-based formulations, part 6: effects of varying pancreatin and calcium levels, Aaps J., 16 (2014) 1344-1357. [17] H.D. Williams, M.U. Anby, P. Sassene, K. Kleberg, J.C. Bakala-N'Goma, M. Calderone, V. Jannin, A. Igonin, A. Partheil, D. Marchaud, E. Jule, J. Vertommen, M. Maio, R. Blundell, H. Benameur, F. Carriere, A. Mullertz, C.W. Pouton, C.J. Porter, Toward the establishment of standardized in vitro tests for lipid-based formulations. 2. The effect of bile salt concentration and drug loading on the performance of type I, II, IIIA, IIIB, and IV formulations during in vitro digestion, Mol. Pharm., 9 (2012) 3286-3300. [18] H.D. Williams, P. Sassene, K. Kleberg, M. Calderone, A. Igonin, E. Jule, J. Vertommen, R. Blundell, H. Benameur, A. Mullertz, C.W. Pouton, C.J. Porter, L. Consortium, 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, Pharma. Res., 30 (2013) 3059-3076. [19] A. Dahan, A. Hoffman, Use of a dynamic in vitro lipolysis model to rationalize oral formulation development for poor water soluble drugs: Correlation with in vivo data and the relationship to intra-enterocyte processes in rats, Pharm. Res., 23 (2006) 2165-2174. [20] C.J. Porter, A.M. Kaukonen, A. Taillardat-Bertschinger, B.J. Boyd, J.M. O'Connor, G.A. Edwards, W.N. Charman, Use of in vitro lipid digestion data to explain the in vivo performance of triglyceride-based oral lipid formulations of poorly water-soluble drugs: studies with halofantrine, J. Pharm.  Sci., 93 (2004) 1110-1121. [21] M.D. Mosgaard, P. Sassene, H. Mu, T. Rades, A. Mullertz, Development of a high-throughput in vitro intestinal lipolysis model for rapid screening of lipid-based drug delivery systems, Eur.  J. Pharm. Biopharm., 94 (2015) 493-500.  [22] C.J.H. Porter, N.L. Trevaskis, W.N. Charman, Lipids and lipid-based formulations: optimizing the oral delivery of lipophilic drugs, Nat. Rev. Drug Discov., 6 (2007) 231-248. [23] N. Thomas, R. Holm, A. Mullertz, T. Rades, In vitro and in vivo performance of novel supersaturated self-nanoemulsifying drug delivery systems (super-SNEDDS), J. Control. Release, 160 (2012) 25-32. [24] P.J. Sassene, M.D. Mosgaard, K. Lobmann, H. Mu, F.H. Larsen, T. Rades, A. Mullertz, Elucidating the Molecular Interactions Occurring during Drug Precipitation of Weak Bases from Lipid-Based Formulations: A Case Study with Cinnarizine and a Long Chain Self-Nanoemulsifying Drug Delivery System, Mol. Pharm., 12 (2015) 4067-4076. [25] J. Khan, A. Hawley, T. Rades, B.J. Boyd, In Situ Lipolysis and Synchrotron Small-Angle X-ray Scattering for the Direct Determination of the Precipitation and Solid-State Form of a Poorly Water-Soluble Drug During Digestion of a Lipid-Based Formulation, J. Pharm. Sci., (2015). [26] J. Khan, T. Rades, B. Boyd, The Precipitation Behavior of Poorly Water-Soluble Drugs with an Emphasis on the Digestion of Lipid Based Formulations, Pharm. Res., (2015). [27] Z. Misic, D.S. Jung, G. Sydow, M. Kuentz, Understanding the interactions of oleic acid with basic drugs in solid lipids on different biopharmaceutical levels., J. Excip. Food Chem., 5 (2014) 2. [28] N. Thomas, R. Holm, M. Garmer, J.J. Karlsson, A. Mullertz, T. Rades, Supersaturated Self-Nanoemulsifying Drug Delivery Systems (Super-SNEDDS) Enhance the Bioavailability of the Poorly Water-Soluble Drug Simvastatin in Dogs, Aaps J., 15 (2013) 219-227. [29] N. Thomas, K. Richter, T.B. Pedersen, R. Holm, A. Mullertz, 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. [30] B.T. Griffin, M. Kuentz, M. Vertzoni, E.S. Kostewicz, Y. Fei, W. Faisal, C. Stillhart, C.M. O'Driscoll, C. Reppas, J.B. Dressman, Comparison of in vitro tests at various levels of complexity for the prediction of in vivo performance of lipid-based formulations: Case studies with fenofibrate, Eur. J. Pharm. Biopharm., 86 (2014) 427-437. [31] R. Berthelsen, R. Holm, J. Jacobsen, J. Kristensen, B. Abrahamsson, A. Mullertz, Kolliphor Surfactants Affect Solubilization and Bioavailability of Fenofibrate. Studies of in Vitro Digestion and Absorption in Rats, Mol. Pharm., 12 (2015) 1062-1071. [32] L. Marciani, P.A. Gowland, R.C. Spiller, P. Manoj, R.J. Moore, P. Young, A.J. Fillery-Travis, Effect of meal viscosity and nutrients on satiety, intragastric dilution, and emptying assessed by MRI, Am. J. Physiol. Gastrointest. Liver Physiol., 280 (2001) G1227-1233. [33] J. Moore, F. Moody, F. Datz, P. Christian, Gastric-Emptying of Liquids and Solids Following Gastric Partition in the Morbidly Obese, Gastroenterol., 80 (1981) 1234-1234. [34] J.G. Moore, P.E. Christian, R.E. Coleman, Gastric-Emptying of Varying Meal Weight and Composition in Man - Evaluation by Dual Liquid-Phase and Solid-Phase Isotopic Method, Digest. Dis. Sci., 26 (1981) 16-22. [35] R.C. Heading, J. Nimmo, L.F. Prescott, P. Tothill, Dependence of Paracetamol Absorption on Rate of Gastric Emptying, Brit. J. Pharmacol., 47 (1973) 415-421. [36] G.A. Kossena, W.N. Charman, C.G. Wilson, B. O'Mahony, B. Lindsay, J.M. Hempenstall, C.L. Davison, P.J. Crowley, C.J.H. Porter, Low dose lipid formulations: Effects on gastric emptying and biliary secretion, Pharm. Res., 24 (2007) 2084-2096. [37] H.H. Holzer, C.M. Turkelson, T.E. Solomon, H.E. Raybould, Intestinal Lipid Inhibits Gastric-Emptying Via Cck and a Vagal Capsaicin-Sensitive Afferent Pathway in Rats, Am. J. Physiol-Gastr. L., 267 (1994) G625-G629. [38] J.F. Cuine, C.L. McEvoy, W.N. Charman, C.W. Pouton, G.A. Edwards, H. Benameur, C.J. Porter, Evaluation of the impact of surfactant digestion on the bioavailability of danazol after oral administration of lipidic self-emulsifying formulations to dogs, J. Pharm. Sci., 97 (2008) 995-1012. [39] D.A. Brodie, R.W. Marshall, O.M. Moreno, Effect of Restraint on Gastric Acidity in Rat, Am. J. Physiol., 202 (1962) 812-&. [40] Y. Gargouri, G. Pieroni, C. Riviere, J.F. Sauniere, P.A. Lowe, L. Sarda, R. Verger, Kinetic assay of human gastric lipase on short- and long-chain triacylglycerol emulsions, Gastroenterol., 91 (1986) 919-925. [41] F. Carriere, H. Moreau, V. Raphel, R. Laugier, C. Benicourt, J.L. Junien, R. Verger, Purification and biochemical characterization of dog gastric lipase, Eur. J. Biochem., 202 (1991) 75-83. [42] S. Nury, G. Pieroni, C. Riviere, Y. Gargouri, A. Bois, R. Verger, Lipase kinetics at the triacylglycerol-water interface using surface tension measurements, Chem. Phys. Lipids, 45 (1987) 27-37. [43] Y. Gargouri, R. Julien, A.G. Bois, R. Verger, L. Sarda, Studies on the Detergent Inhibition of Pancreatic Lipase Activity, J. Lipid Res., 24 (1983) 1336-1342. [44] 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. [45] A.R. Patel, P.R. Vavia, Preparation and in vivo evaluation of SMEDDS (Self-Microemulsifying Drug Delivery System) containing fenofibrate, Aaps J., 9 (2007) E344-E352.  

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.24.1-0314313/manifest

Comment

Related Items