ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015

Digital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 207

Biopharmaceutical Evaluationof Intra-arterial Drug-DeliverySystems for Liver Cancer

Investigations in healthy pigs and liver cancerpatients

ELSA LILIENBERG

ISSN 1651-6192ISBN 978-91-554-9422-3urn:nbn:se:uu:diva-267396

Dissertation presented at Uppsala University to be publicly examined in B42, BMC,Husargatan 3, Uppsala, Friday, 22 January 2016 at 09:15 for the degree of Doctor ofPhilosophy (Faculty of Pharmacy). The examination will be conducted in English. Facultyexaminer: Professor Helen M. Burt (Faculty of Pharmacutical Sciences, University of BritishColumbia, Canada).

AbstractLilienberg, E. 2015. Biopharmaceutical Evaluation of Intra-arterial Drug-DeliverySystems for Liver Cancer. Investigations in healthy pigs and liver cancer patients. DigitalComprehensive Summaries of Uppsala Dissertations from the Faculty of Pharmacy 207.67 pp. Uppsala: Acta Universitatis Upsaliensis. ISBN 978-91-554-9422-3.

There are currently two types of intra-arterial drug-delivery system (DDS) in clinical use in thepalliative treatment of primary liver cancer. The chemotherapeutic drug doxorubicin (DOX) canbe formulated into a drug-in-lipiodol emulsion (LIPDOX) or a microparticulate drug-elutingbead system (DEBDOX). To facilitate development of future DDSs, we need to understand therelease and local distribution of drug from these DDSs into the complex, in vivo, pathologicalenvironment.

The overall aim of this project was to assess and improve understanding of the in vivo releaseof DOX from LIPDOX and DEBDOX and its local disposition in the liver. These processeswere investigated in detail in a multisampling-site, healthy pig model and in human patientswith liver cancer. The mechanisms involved in DOX disposition were studied by examiningpotential interactions between DOX and lipiodol and/or cyclosporine A (CsA) in pigs.

In this project, the main elimination pathway for DOX and its primary metabolitedoxorubicinol (DOXol) was via bile; their extensive canalicular carrier-mediated transport (e.g.ATP-binding cassette transporters ABCB1, ABCC1, ABCC2 and ABCG2) was inhibited byCsA. CsA had no effect on the carbonyl and aldo-keto reductases responsible for the metabolismof DOX into DOXol. LIPDOX released DOX more rapidly and to a greater extent into thecirculation than DEBDOX, which had only released 15% of the dose in patients after 24 hrs.The systemic exposure to DOX was lower for DEBDOX than for LIPDOX. Greater fractions ofDOXol were formed in blood and bile with LIPDOX than with DEBDOX. This may have beenbecause DOX was more widely distributed into regions with increased metabolic capacity orbecause of increased intracellular uptake when DOX was delivered in LIPDOX. The excipientlipiodol in the LIPDOX formulation did not interact with transporters, enzymes or membranesthat would explain the increased cellular uptake of DOX.

In conclusion, the release of DOX from DEBDOX is more controlled in vivo than that fromLIPDOX, indicating that DEBDOX is a more robust pharmaceutical product. The formulationsfor future optimized DDSs should therefore be more similar to DEBDOX than to LIPDOX.

Keywords: in vivo release, drug delivery systems, local delivery, drug disposition,doxorubicin, hepatocellular carcinoma, transarterial chemoembolization, transarterialchemotherapy infusion

Elsa Lilienberg, Department of Pharmacy, Box 580, Uppsala University, SE-75123 Uppsala,Sweden.

© Elsa Lilienberg 2015

ISSN 1651-6192ISBN 978-91-554-9422-3urn:nbn:se:uu:diva-267396 (http://urn.kb.se/resolve?urn=urn:nbn:se:uu:diva-267396)

List of Papers

This thesis is based on the following papers, which are referred to in the text by their Roman numerals. I Lilienberg E., Ebeling Barbier C., Nyman R., Hedeland M., Bondes-

son U., Axén N., Lennernäs H. (2014). Investigation of Hepatobili-ary Disposition of Doxorubicin Following Intrahepatic Delivery of Different Dosage Forms. Molecular Pharmaceutics, 11(1):131–144

II Dubbelboer I.R.*, Lilienberg E.*, Hedeland M., Bondesson U., Piquette-Miller M., Sjögren E., Lennernäs H. (2014). The Effects of Lipiodol and Cyclosporin A on the Hepatobiliary Disposition of Doxorubicin in Pigs. Molecular Pharmaceutics, 11(4):1301–1313

III Lilienberg E., Dubbelboer I.R., Sjögren E., Lennernäs H. Lipiodol Does Not Affect the Tissue Distribution of Intravenous Doxorubicin Infusion in Pigs. Submitted to Molecular Pharmaceutics.

IV Lilienberg E., Dubbelboer I.R., Karalli A, Axelsson R., Brismar T.B., Ebeling Barbier C., Norén A., Duraj F., Hedeland M., Bondes-son U., Sjögren E., Stål P., Nyman R, Lennernäs H. Comparison of Drug Release, Pharmacokinetics and Pharmacodynamics after Lip-iodol-based Emulsion or Drug-eluting Bead Delivery to Patients with Hepatocellular Carcinoma. In manuscript.

Reprints were made with permission from the American Chemical Society.

*These authors contributed equally to the execution of the study and writing the article.

My contribution to the papers included in this thesis was as follows: I was highly involved in the project design, planning, project progress man-agement and writing of the manuscripts for all Papers (I–IV). I was involved in the initiation of collaboration with the other researchers (Paper I and IV) and in the experimental work for all Papers. I performed the bioanalysis in Paper III and the data analysis for Paper I, III and IV.

Additional papers not included in this thesis

V Dubbelboer I.R., Lilienberg E., Ahnfelt E., Sjögren E., Lennernäs H. (2014). Treatment of intermediate stage hepatocellular carcinoma: a review of intrahepatic doxorubicin drug-delivery systems. Thera-peutic Delivery, 5(4):447–466

Contents

Introduction ..................................................................................................... 9 Parenteral drug delivery and drug distribution ......................................... 10 Elimination of drugs ................................................................................. 11 The human liver ....................................................................................... 12

The microstructure of the liver lobule ................................................. 13 The hepatocyte ..................................................................................... 15

Hallmarks of cancer ................................................................................. 15 Cell cycle control and DNA damage ................................................... 16 Tumor microenvironment .................................................................... 16

Hepatocellular carcinoma – HCC ............................................................. 18 Pathophysiology .................................................................................. 19 Treatments ........................................................................................... 19 Transarterial chemoembolization ........................................................ 20

Doxorubicin – DOX ................................................................................. 21 Hepatic disposition of DOX ................................................................ 23

Local drug delivery systems ..................................................................... 24 Lipiodol-based emulsion – LIPDOX ................................................... 24 Drug-eluting beads – DEBDOX .......................................................... 26

Aims of the thesis.......................................................................................... 28

Methods ........................................................................................................ 29 Ethics and surgical/interventional procedures .......................................... 29

The in vivo multisampling-site pig model ........................................... 29 The clinical trial in patients with liver cancer ...................................... 30

Study designs ............................................................................................ 31 Drug–excipient and drug–drug interaction studies in healthy pigs ..... 31 The in vivo release studies in healthy pigs and HCC patients ............. 32

Bioanalytical methods .............................................................................. 34 Pharmacokinetic data analysis .................................................................. 34 Statistical analysis .................................................................................... 37

Results and Discussion ................................................................................. 38 In vivo pharmacokinetics of DOX following an i.v. dose in pigs ............ 38

Plasma clearance and hepatic extraction of DOX ............................... 39 Biliary and urinary transport of DOX .................................................. 40 Tissue distribution of DOX ................................................................. 40

Drug–excipient and drug–drug interaction studies in healthy pigs .......... 41 The effects of lipiodol on the pharmacokinetics of DOX .................... 41 The effects of cyclosporin on the pharmacokinetics of DOX .............. 42

The in vivo release studies in healthy pigs and HCC patients .................. 43 Dosage and tumor characteristics ........................................................ 44 The pharmacokinetics of DOX after delivery with LIPDOX or DEBDOX ............................................................................................. 44 Release from LIPDOX ........................................................................ 46 Release from DEBDOX ...................................................................... 47 Local distribution of DOX delivered by LIPDOX and DEBDOX ...... 48 Safety and anti-tumor efficacy of the DDSs with respect to the pharmacokinetics of DOX ................................................................... 49

Conclusions ................................................................................................... 51

Future perspectives ....................................................................................... 53

Populärvetenskaplig sammanfattning ........................................................... 54

Acknowledgements ....................................................................................... 56

References ..................................................................................................... 59

Abbreviations

A Amount of drug or metabolite ABC AKR AUC

ATP-binding cassette transporter Aldo-keto reductase Area under the concentration–time curve

BCLC BCRP

Barcelona clinic liver cancer Breast cancer resistance protein

C Concentration of drug or metabolite CBR Carbonyl reductase CL Clearance Clast Last observed concentration Cmax Maximum concentration CsA cTACE CYP DDS DEBDOX DOX DOXol EH

Cyclosporin A Conventional transarterial chemoembolization (drug-in-lipiodol emulsion + embolic agent) Cytochrome P450 Drug-delivery system DOX-eluting microparticle bead DDS Doxorubicin Doxorubicinol Hepatic extraction

F fe Fi;app

Bioavailability Fraction of the dose excreted (to urine or bile) Apparent intracellular availability

HCC Hepatocellular carcinoma IC50 Concentration needed to cause 50% of the maximal

inhibition i.v. Intravenous k Rate constant Kp Tissue-to-plasma concentration partitioning

coefficient LIPDOX Lipiodol-based emulsion containing DOX MRP Multidrug resistance-associated protein Pgp P-glycoprotein RD Remaining fraction of the dose in liver and/or DDS S.E. Standard error of mean SLC Solute carrier

TACE Transarterial chemoembolization TAI Transarterial infusion chemotherapy tmax Time to reach the maximum concentration (Cmax) UPLC-MS Ultra-high performance liquid chromatography

coupled to mass spectrometry UPLC-MS/MS

Ultra-high performance liquid chromatography coupled to tandem mass spectrometry

V Volume VC/VH Vena cava orifice from the hepatic veins VF Femoral vein VH Hepatic vein VP Portal vein Vss Volume of distribution at steady state λz Terminal elimination rate constant

Introduction

There have recently been significant efforts to develop new parenteral drug-delivery systems (DDSs) with the aim of improving delivery of chemo-therapeutic drugs in the treatment of patients with cancer. The therapeutic goal is to increase intracellular tumor exposure to chemotherapeutic drugs and consequently to improve the anti-tumor efficacy. However, it has proved a difficult challenge to find a drug or DDS that is specific to tumor cells, especially since tumor cells are derived from normal cells and thus share many of their cellular properties. Many relatively sophisticated DDSs have been developed through the years to improve drug delivery to tumors; these include, for example, nanoparticles (such as liposomes), microparticles and lipid emulsions. However, only a few of these have been successful enough to reach the clinic.

The primary objective of this thesis was to increase understanding of two clinically available DDSs that are currently in use in the palliative treatment of intermediate-stage hepatocellular carcinoma (HCC; primary liver cancer). The two DDSs investigated in the thesis, the lipiodol-based emulsion (LIPDOX) and microparticle drug-eluting beads (DEBDOX), are used to deliver the anti-cancer drug doxorubicin (DOX) to the tumor. LIPDOX is an emulsion of the contrast agent lipiodol with an aqueous solution of DOX. DEBDOX is a DOX-eluting bead system (DC Bead®). These DDSs are administered locally to the liver via intra-arterial delivery. Although LIPDOX and DEBDOX are in clinical use, little is known about their target-ing and release mechanisms in vivo. Hence, to facilitate the development of future delivery strategies, this thesis specifically focused on increasing understanding of the in vivo release of DOX from these DDSs and its hepatic distribution, and the mechanisms specifically involved with these processes.

This thesis is the first part in the Target Exact Project – a liver cancer treatment optimization project that was initiated at the Department of Phar-macy, Uppsala University, Sweden.

9

Parenteral drug delivery and drug distribution Parenteral administration of drugs includes all delivery methods that are not directly administered to the gastrointestinal tract. Mostly, these take the form of injection into the vascular system, subcutaneous injection, or injection into the muscles.1 Parenteral delivery is often used to achieve a local and/or rapid effect. When a drug is administered directly into the vascular system, i.e. into veins or arteries, it reaches the circulatory system directly, and will thus be rapidly distributed in the body.1 The distribution of a drug to an intracellular site of action is driven by vascular, transvascular, and interstitial transport and then by cellular uptake.2 These processes are summarized in Figure 1.

Figure 1. Overview of the processes that govern the transport of a drug from the blood circulation into the cells. The figure shows tumor cells in unspecified tumor tissue, but the same transport processes apply to healthy tissue. These processes can, however, differ between different tissues and stages of disease as the biophysical properties of the vessel walls, cells and interstitium change. The figure is a modifica-tion of an illustration from Chauhan et al., Annu Rev Chem Biomol Eng, 2011.2

10

Vascular transport, i.e. the movement of a drug or a DDS within the blood, controls distribution of the drug into the organs and is dependent on the con-vective blood flow and the vascular morphology (the number, length, diame-ter and geometrical arrangement of the blood vessels).3 Transvascular transport, i.e. the passage of a drug or DDS across a blood vessel wall, is driven by a combination of convection and diffusion (pressure and conc-entration gradient) between the vessel and the organ interstitium. Trans-vascular transport is also dependent on the biophysical properties of the ves-sel wall and the physicochemical properties of the drug and/or DDS.2 Once in the interstitium, the drug moves towards the cells by interstitial transport. Interstitial transport is driven by a combination of convection and diffusion processes, and is dependent on the interstitial properties and the physico-chemical properties of the drug or DDS.2 Finally, cellular uptake is mediated by diffusion, carrier-mediated transport of the free drug, and/or endocytosis of the DDS. The rate of drug release from the DDS affects both the efficacy and the adverse events. Once the drug is within the different cell types it can exert both the desired on-target effects and/or unwanted off-target effects (i.e. adverse reactions). For DDSs intended for parenteral drug delivery, the design and administration of the DDS, as well as biological pathophysiologi-cal factors also are important for the local drug deposition and distribution.4

Elimination of drugs The elimination of most drugs from the body occurs through metabolism in the liver and excretion into the bile and urine. Drug-metabolizing enzymes are available to various extents throughout the body, but the liver and the kidney have the highest activity. The liver and kidney are also the most important organs for excretion of drugs.

During metabolism, hydrophilic properties are added to detoxify and/or increase the possibility of eliminating the compound. The biotransformation of a hydrophobic compound into a more water-soluble form can be described by phase I and phase II metabolism stages.5 Phase I metabolism includes non-synthetic reactions such as oxidation, reduction or hydrolysis, where a functional group is modified or added to the compound. Enzymes that cata-lyze phase I metabolic processes include cytochrome P450 proteins (CYPs), carbonyl reductases and aldo-keto reductases. Phase II metabolism involves synthetic reactions where an endogenous substance such as glucuronide, sulfate or glutathione is conjugated to the compound. Phase II metabolites are formed by transferase enzymes. These metabolites are generally more polar than phase I metabolites, and therefore more likely to be excreted into urine. Metabolites formed in the liver may be further metabolized, excreted into the bile or transported back into the blood and then excreted into the urine.6

11

Biliary and urinary excretion are processes that may involve transport proteins to various extents. Transport proteins involved in drug influx and efflux can be divided into two superfamilies: solute carriers (SLCs) and ATP-binding cassette transporters (ABCs).6, 7 The ABCs are directly driven by energy from ATP hydrolysis and therefore mediate primary active transport, mostly against a concentration gradient.8 Some SLCs are passive (facilitative) transporters that do not require any energy, and some SLCs are secondary active transporters that utilize energy from the surroundings, such as ion gradients, co-transport or counter transport. SLCs may be responsible for both influx and efflux while the ABCs that are important for drugs are mainly efflux proteins.6 Both ABCs and SLCs are highly expressed in the intestine, liver and kidney and are therefore important for the absorption, distribution and elimination of drugs. Compounds that have been transported to the bile are diverted to the intestine where they may be reabsorbed (enterohepatic cycle) or eliminated with the feces.9

The human liver The liver has a dual blood supply. Most of the blood (~75–80%) is delivered as nutrient-rich blood from the gastrointestinal tract and spleen via the portal vein (VP).10 The remainder is delivered as oxygen-rich blood directly from the aorta via the hepatic artery.10 The strategic location of the liver is ideal for elimination of exogenous compounds since compounds from the gastro-intestinal tract have to pass through it before they reach the systemic circ-ulation.

The liver can be divided anatomically into eight functionally independent segments, according to Couinaud’s scheme (Figure 2). Basically, the main VP is separated into left and right branches, dividing the liver into left and right lobes that are separated by the middle hepatic vein (VH).11, 12 Separate blood supply (from the VP and hepatic artery), vascular outflow and biliary drainage systems further divide the lobes into eight segments. These blood vessels and bile ducts subdivide repeatedly within the liver and the terminal branches lead into the functional units of the liver, the hepatic lobules. The human liver contains about one million of these lobules.13

12

Figure 2. Illustration of the liver divided into eight functionally independent segments according to Couinaud´s scheme. Segments 5–8 represent the right liver lobe and segments 1–4 represent the left liver lobe. The lobes are defined by the separation of blood flow into the left and right branches of the portal vein. The lobes are also anatomically separated by the middle hepatic vein. Due to the independent blood supply for each segment they can be individually resected. Adapted by per-mission from Macmillan Publichers Ltd: ref.12,© 2014.

The microstructure of the liver lobule The liver lobule has a hexagonal shape and is composed of two major cell types: parenchymal and non-parenchymal cells. The parenchymal cells, i.e. hepatocytes, represent 60% of the cellular population in the liver and occupy almost 80% of the liver volume.14 It is the hepatocytes that are responsible for the liver metabolism and excretion of endogenous compounds and xeno-biotics such as drugs.13 The hepatocytes have many other important func-tions as well, such as the storage of proteins, the transformation of carbo-hydrates and the synthesis of proteins, cholesterol, bile salts and phospho-lipids. The non-parenchymal cells include sinusoidal endothelial cells, hepatic stellate cells, Kupffer cells (macrophages of the liver) and cholangio-cytes (biliary epithelial cells).13

The organization of the cells in the liver lobule is shown in Figure 3. The central vein is located in the center of the lobule, from whence the hepato-cytes extend radially in linear chords to the lobule periphery. These chords are usually one cell-layer thick. The portal and arterial vascular branches and the bile duct form the “portal triad”, which is located in 4–8 places in the periphery of the lobule. Portal and arterial blood is mixed in the sinusoids

13

located between the hepatic chords. From the sinusoids, the blood continues through the lobule into the central vein, from which it moves into the VHs and eventually leaves the liver to flow into the vena cava.13 Most hepato-cytes are exposed to the sinusoidal vasculature on two sides to optimize exchange. The unique physiology of liver sinusoidal endothelial cells facil-itates molecular exchange with the blood; they are highly fenestrated and lack a basement membrane. The endothelial cells are separated from the hepatocytes by the space of Disse, in which the hepatic stellate cells reside. The stellate cells are important for the storage of fat and vitamin A and for regulating the turnover of extracellular fluid in the space of Disse.15 The Kupffer cells mainly reside within the sinusoids.

Figure 3. The microstructure of the hepatic lobule. Blood enters the lobule via branches of the portal vein and hepatic artery, mixes in the sinusoids and leaves the structure through the central vein. The sinusoids surround the chords of hepatocytes that run from the central vein. The inset shows the bile canaliculi between adjacent hepatocytes. These canaliculi form a network that transfers bile in the opposite direction to the blood flow, towards the bile duct in the periphery of the lobule. Liver sinusoidal endothelial cells line the sinusoidal vessel wall. The hepatic stellate cells are situated in the space of Disse, and the Kupffer cells are mainly located in the sinusoids. The figure was reprinted with kind permission from A. Vildhede.16

14

Bile canaliculi are formed in the membrane between two adjacent hepato-cytes. Tight junctions separate the canalicular membrane from the baso-lateral membrane. These canaliculi form an interconnected network that transfers bile in the opposite direction to that of the sinusoidal blood, i.e. into the bile duct in the periphery of the lobule. The bile ducts are lined by chol-angiocytes, which have an important function that involves secretion of bicarbonates and water.17 While the hepatocytes produce the primary bile, the cholangiocytes modify the volume and composition of the bile during its transport through the network (the biliary tree). The bile components travel through the bile ducts via the common hepatic duct to the gallbladder, from which they are released into the small intestine.9

The hepatocyte Hepatocytes are cubic, polarized cells with two distinct membranes: the basolateral membrane that faces the sinusoids and adjacent cells, and the canalicular (apical) membrane that forms the bile canaliculi. Both the baso-lateral and canalicular membranes are richly covered with microvilli to in-crease the surface area for molecular exchange.18 About 68% of the hepato-cyte surface area is in contact with other hepatocytes, 22% with sinusoids and 10% with bile canaliculi.19 Transporter proteins are expressed to a high degree by both the sinusoidal and canalicular membranes of the hepatocyte, thus facilitating cellular influx and efflux of endogenous com-pounds and xenobiotics.5

Hallmarks of cancer The development of cancer is a multistep process where normal cells accu-mulate several genomic mutations over time. The many different cell types in the body enable many possibilities for different mutations and, hence, cancer is a highly heterogeneous disease. This heterogeneity is one of the reasons why finding a cure for cancer has not been successful, at least in the search for a “magic bullet” that would treat all forms of cancer. There are several common hallmarks of cancer, although the underlying mechanisms may differ between tumors. These include self-sufficiency in growth signals and insensitivity to anti-growth signals, resistance to apoptosis, induced angiogenesis, limitless replicative potential, and a tendency for tissue invasion and metastasis.20, 21 Most cancer mutations damage somatic cells, and it is only when they damage the DNA in germ cells that the mutation will be inherited by the next generation.22 Tumor cells have increased mito-sis and apoptosis activity compared to healthy cells.23

15

Cell cycle control and DNA damage The mutations associated with tumor cells and cancer are likely to be the result of DNA damage in combination with defects in DNA repair and the regulation of the cell cycle.24 The cell cycle consists of a series of highly regulated phases and irreversible events that are mandatory for prolifera-tion.25 Each phase has a specific checkpoint that must be passed before the cell can move into the next phase. In adults, most cells are dormant, outside of the cell cycle, and are thus not subject to cell division. These cells are in the G0-phase. Growth factors can stimulate these cells to move from G0 to the first gap phase, G1, in which they prepare for DNA replication. Once the cell has passed the G1 restriction point there is no return, and the cell will complete the cell cycle. After G1, the cell enters the synthesis (S) phase of DNA replication, followed by the G2 phase, in which the cell is controlled and damaged DNA is repaired. After DNA repair, the cell continues into mitosis, the M phase. In the M phase, the cell is divided and the duplicated genetic material is equally separated into two daughter cells.26 These daugh-ter cells can now either re-enter the cell cycle or become dormant in the G0 phase.

To avoid the accumulation of unwanted cells, such as those with damaged DNA, there is a well regulated system for cell death. The two main modes of cell death are apoptosis and necrosis.27 Apoptosis is ATP-dependent and is considered to be a “clean and tidy” process for killing cells. The mechanism includes special proteins (caspases) that inactivate important proteins in the cell. Inactivated cells or cell parts are typically engulfed and destroyed by phagocytes (specialized immune cells).28 Necrosis is ATP-independent and was for a long time considered to be unordered. However, there is now evidence that necrosis can also occur in a regulated fashion.29 The process of necrosis is typically described as a swelling of the cell and rupturing of the cell membrane, which causes leakage of the intracellular components, com-monly followed by inflammation.27 Programmed cell death is important as a protective mechanism against cancer. Impaired or inactivated apoptotic pro-cesses can promote cancer, since cells with mutations are not cleared from the body.

Tumor microenvironment Tumor cells are surrounded by stromal cells, endothelial cells, and macro-phages. These cells are in close contact with each other and together they create a unique environment in which the tumor can proliferate.30 Common pathophysiological changes give tumors special characteristics that con-tribute to membrane and interstitial transport barriers that drugs or DDSs have to cross:2

16

1. Growth-induced solid stress. Tumor growth occurs within a restricted area. Solid stress is the result of tissue expansion to the extent that it comes into contact with the surrounding normal cells. This solid stress accumulates in the tumor as a result of continuous increases in cell number and overproduction of the interstitial matrix. Eventually, this leads to compression of blood and lymphatic vessels, which results in reduced blood flow, increased vascular resistance and reduced fluid drainage by the lymphatic system.

2. Abnormal vascular network. Tumor cells produce a variety of pro-angiogenic factors that induce new blood vessels to grow from the sur-rounding normal vessels; these new vessels are required for the tumor to be able to grow beyond 2 mm in size. The new vessels typically lack order in branch hierarchy and they are heterogeneous and hyper-permeable to fluids and macromolecules (>40 kDa). This is because of the presence of fenestrae in the vessel walls, transendothelial channels, and large pores throughout the vasculature.31 The loss of fluid from the blood vessels leads to increased viscosity of the remaining blood, which reduces the blood flow.

3. High interstitial fluid pressure. The impaired lymphatic system, the hyperpermeable blood vessels, and the accumulation of fluid and macro-molecules in tumors all lead to high interstitial fluid pressure. These factors cause a substantial reduction in transvascular and interstitial fluid pressure gradients. Without the pressure gradient, the convective flow is reduced and the tumor distribution is mainly driven by diffusion.

4. Dense interstitial structure. For the reasons mentioned above (1–3), tumors are highly cellular and accumulate macromolecules. They also have a high proportion of stromal cells. This increases the density of the tumor structure and, consequently, also causes steric hindrances to drug transport. Since many drugs are highly bound to the matrix and cellular components, the interstitial transport rate is reduced even further.2

These intratumor transport barriers reduce the penetration of the drug into the tumor, resulting in poor, heterogeneous drug distribution.2 The impaired spatial distribution from the blood vessels has been demonstrated for DOX, mitoxantrone, and topotecan.32 However, the accumulation of macro-molecules that is characteristic of the tumor microenvironment enables passive targeting of macromolecular DDSs into this area. This indirect accumulation of drugs in the tumor is called the enhanced permeability and retention effect and was first described by Maeda in the middle of the 1980s.33

17

Hepatocellular carcinoma – HCC

Liver cancer was the sixth most common cancer form globally in 2012 (782,000 cases), and the second most lethal cancer form (745,000 deaths).34 Survival rates and prognosis are poor, with a reported yearly mortality-to-incidence ratio of 0.95.34 The highest incidence of liver cancer is observed in Asia and Africa, while the incidence in Southern Europe and Northern America is intermediate (Figure 4).34 However, it has been estimated that the incidence of liver cancer will increase globally and especially in developed countries over the next ten years.35 HCC represents 70–85% of all primary liver cancers.36

Figure 4. This world incidence map illustrates the age-standardized incidence of liver cancer in both sexes in 2012 per 100,000 individuals. Data from ref.35 Availa-ble from: http://globocan.iarc.fr, accessed on 21 October 2015.

The development of HCC is usually highly correlated with advanced chronic liver disease, i.e. liver cirrhosis. The liver cirrhosis is often associated with hepatitis C virus, hepatitis B virus, or over-consumption of alcohol. Inges-tion of aflatoxins has also been associated with a risk of developing HCC.37 Recent studies have identified obesity and diabetes type 2 as additional risk factors.38

18

Pathophysiology HCC is a primary liver cancer that originates from mutations in hepato-cytes.23 The microstructure of a HCC tumor is typically characterized by distorted chords of hepatocytes that are 2–10 cell layers thick, as opposed to the single-layer, ordered chords of hepatocytes in the healthy liver. These cell layers typically get thicker as the HCC progresses.39 The tumor cells of HCC have distinct cell membranes and are associated with atypical, often enlarged, nuclei. Some cells may appear transparent due to accumulation of lipids or glycogen. Kupffer cells are typically quantitatively reduced in HCCs, and bile pigment is present in approximately 20% of all HCCs.23, 40 The hepatic stellate cells undergo transformation into cells that produce tumor growth factors and pro-angiogenic factors and secrete extracellular matrix proteins to a higher degree.30

The majority of HCCs arise in cirrhotic livers.23 Cirrhosis is an advanced stage of fibrosis that is associated with scar tissue in the space of Disse, lost fenestrations in the sinusoidal endothelial cells, distorted hepatic vasculature, and regenerative nodules on the hepatocytes. All of these characteristics lead to impaired hepatocyte functionality and a gradual increase in intrahepatic resistance.41

HCC manifests as a highly vascularized tumor; in contrast to normal liver tissue, which has a dual blood supply, the blood in HCC tumors mainly (>90%) comes from the hepatic artery.23 The changes to the blood supply during hepatocarcinogenesis start during the early phases. As the portal venous flow diminishes, hypoxia develops in the tumor, resulting in angio-genesis from pre-existing blood vessels or development of new tumor vessels. Eventually, well or moderately differentiated HCCs are pre-dominately supplied with arterial blood. In poorly differentiated HCCs, the tumor shifts into an anaerobic metabolism and the arterial blood supply decreases.39 It is common to find a focal area in the tumor with hemorrhage or necrosis. HCCs can be single or multiple and usually range in size from less than 1 cm to 30 cm in diameter.23

Treatments The Barcelona Clinic Liver Cancer (BCLC) classification system is widely used to classify HCC into five stages:42 very early stage HCC (BCLC 0), early stage HCC (BCLC A), intermediate stage HCC (BCLC B), advanced stage HCC (BCLC C) and terminal stage HCC (BCLC D). The classification depends on prognostic predictions (tumor status such as size, number, vascu-lar invasion, and patient characteristics such as liver function and health status) and parameters known to affect the therapeutic outcome (e.g. biliru-bin and portal hypertension).

19

There are curative treatment options for patients with early HCC (BCLC 0 and A). The other BCLC stages (B–D) are considered unresectable. Patients diagnosed with BCLC stages B–C are usually offered palliative treatment options. Treatment guidelines from the European association for the study of the liver and the European organisation for research and treat-ment of cancer can be viewed in Table 1.42

Transarterial chemoembolization This thesis focuses on improving understanding of drug delivery from LIPDOX and DEBDOX, two of the DDSs used in transarterial chemo-embolization (TACE) in intermediate-stage HCC. The enhanced arterial blood supply of HCC, compared to normal liver tissue, enables a loco-regional delivery approach that, in theory, will increase local drug concentra-tions close to the tumor, while normal liver tissue is protected. In TACE, the DDS is delivered via a catheter through the hepatic artery close to the tumor, with lobar, segmental (selective) or superselective placement. The DDS is usually loaded with a single chemotherapeutic agent or a drug cocktail that is infused together with a contrast agent and embolic agents. The embolic agents (such as polyvinyl alcohol particles, microparticles, gel foam or gel sponge) are used to reduce the tumor blood flow as much as possible so as to deprive the tumor of oxygen and, at least in theory, increase the time that the drug resides in the tumor.

Generally, TACE techniques can be subdivided as follows: transarterial infusion chemotherapy (TAI), also called transarterial oily chemoemboli-zation or lipiodolization, where the DDS consists of an emulsion based on the oily contrast agent lipiodol mixed with an aqueous drug solution;42, 43

Table 1. Treatment recommendations for patients with hepatocellular carcinoma (HCC) according to guidelines from the European association for the study of the liver and the European organisation for research and treatment of cancer.42

BCLC classification Recommended treatment

BCLC 0 and A (early HCC) Resection, liver transplantation, local ablation with radiofrequency or percutaneous ethanol injection

BCLC B (intermediate stage) Transarterial chemoembolization (TACE)

BCLC C (advanced stage) Sorafenib (anti-angiogenic drug, multi-tyrosine kinase inhibitor)

BCLC D (terminal stage) Best supportive care

20

conventional TACE (cTACE), where administration of the lipiodol-based emulsion is followed by administration of embolic agents; and drug-eluting bead TACE, where the DDS consists of beads loaded with drug and func-tions as an embolic agent while simultaneously delivering the drug. TACE treatment may be repeated on several occasions, depending on the individual tumor response and the occurrence of adverse events.

The survival benefits of intra-arterial therapies have been the subject of many clinical trials, and show contradictory results.44-47 cTACE is recom-mended as the standard of care for HCC patients (BCLC B) on the basis of a meta-analysis that proved a survival benefit.46 This meta-analysis showed that the expected overall median survival (i.e length of time that the patient survives after start of treatment) was increased to about 20 months with cTACE compared to 16 months with symptomatic treatment.46 In 2007, a meta-analysis comparing cTACE, embolization only, and TAI concluded that embolization and cTACE increase the time of survival whereas no bene-fit was found for TAI.43 Many of the studies in this analysis included several chemotherapeutic drugs, and possible differences in potency of those drugs were neglected. In another randomized clinical trial where cTACE and TAI with the same drug was compared, it was concluded that embolization (cTACE) does not increase survival compared to TAI.48 The methodology in this TAI vs cTACE study was however subsequently questioned.49

The lack of strong evidence to indicate the most beneficial method for transarterial treatments (i.e. TAI, cTACE or drug-eluting bead-TACE) has resulted in a field with tremendous variations in treatment among clinical centers and interventional radiologists. These variations commonly involve the choice and dosage of the chemotherapeutics, the choice of embolic agents (if used), the selectivity of the transarterial catheter, and the frequency of repeated treatments.

Doxorubicin – DOX DOX is one of the most commonly used cytotoxic drugs in TACE.43 DOX was investigated in this thesis because it is loadable into the two DDSs in question. DOX is produced by the bacterium Streptomyces peucetius, and is a potent cytotoxic antitumor drug that has been classified as an anthracycline antibiotic.50 It is chemically described as a hydrophobic planar tetracyclic ring system connected to a more polar deoxy(amino)sugar by a glycosidic bond, see Figure 5. It is an ampholytic molecule with two pKa values: 8.2 and 10.2.51 DOX is often used as the starting material in formulations as the hydrochloride salt. DOX hydrochloride has a water solubility of 10 mg/ml at unspecified pH and has a log D7.5 of 2.42.52 It forms an orange-red solution in water.

21

For over 30 years, DOX has been used as a chemotherapeutic agent against multiple types of cancer, e.g. bone sarcoma, Hodgkin’s disease, and ovary, breast, lung, and liver cancer.53 DOX exerts its cytotoxic effect intra-cellularly by intercalation to the DNA double strand, inhibition of topo-isomerase II (an enzyme that separates the DNA strands from each other before transcription and replication), and induction of reactive oxygen species.53, 54 These mechanisms all cause DNA damage, which leads to apop-tosis or cell cycle arrest.55-57 DOX is most effective in mitotic cells but can also be effective in other phases of the cell cycle.53 Since DOX does not specifically target tumor cells, its use is associated with several adverse effects such as nausea, vomiting, bone marrow depression, hair loss, local tissue necrosis, kidney failure, and cardiotoxicity.53 Since it causes cumula-tive, irreversible toxicity, and mainly because of the risk of cardiotoxicity, the cumulative intravenous (i.v.) dose of DOX is limited to 500 mg/m2 body surface area.57, 58 Because of this broad, dose-limiting toxicity and the poor permeability of the intestinal wall to DOX (0.1 × 10-6 cm2/s), its delivery is restricted to the parenteral route.59, 60 Following an i.v. dose, the plasma concentration–time curve for DOX is best described by a three-compartment model; it has a reported terminal half-life of 35 h and a large volume of dis-tribution (22 L/kg), which reflects extensive binding to tissue, mainly liver, kidneys, heart, and lungs.61, 62

Figure 5. The molecular structure of doxorubicin. The functional group R is a ketone. In the primary metabolite doxorubicinol, the only difference is the trans-formation of the ketone into an alcohol.

22

Hepatic disposition of DOX The major metabolizing and eliminating organ for DOX is the liver. Cellular uptake is mediated by both passive diffusion and carrier-mediated transport, followed by metabolism.63 The parent drug and metabolites are primarily eliminated through the bile. After i.v. administration of DOX, 40–50% of the dose is eliminated in the bile within a week (50% as the parent drug and 50% as metabolites, mainly doxorubicinol, DOXol).64 Renal excretion of un-changed DOX is much less, with only 12% excreted in the six days after an i.v. dose.64 Transporters and enzymes important for the hepatic disposition of DOX are summarized in Figure 6.

Figure 6. Doxorubicin (DOX) enters the cell mainly by passive diffusion, but is also a substrate for the solute carriers SLC01A2 and SLC22A16. In the cell, DOX may be metabolized into its active metabolite doxorubicinol (DOXol) or other inactive metabolites by carbonyl reductases (CBR1 and CBR3) and/or aldo-keto reductases (AKR1A1 and AKR1C3). DOX and DOXol exert their effects by causing DNA damage in the nucleus. DOX and DOXol may diffuse back into the blood or be transported into bile by transport proteins such as ABCB1 (P-glycoprotein), ABCC1 and ABCC2 (multidrug resistance-associated protein 1 and 2), ABCG2 (breast can-cer resistance protein) and Ral binding protein1 (RALBP1). The figure is based on information from Thorn et al.65

23

Although passive diffusion is thought to be the main uptake mechanism for DOX,66 the drug is a substrate for at least two transporters on the basolateral membrane of the hepatocyte: SLC22A16 and SLC01A2.67, 68 Following cellular uptake of DOX, it may be metabolized into its primary, active me-tabolite DOXol, which is 75-fold less potent than DOX. This bio-transformation is catalyzed by cytosolic carbonyl reductases (CBR1, CBR3) and aldo-keto reductases (AKR1A1 and AKR1C3) which reduce the ketone group in DOX into an alcohol, see figure 5.69, 70 CBRs and AKRs are availa-ble in many cells, but are most active in the liver and kidney.69, 71 DOX and DOXol may be further metabolized into inactive aglycones. In order to be eliminated, the aglycones need to be glucuronidized. However, these inac-tive metabolites are present in negligible amounts in human biological fluids compared to DOX and DOXol.53 DOX and DOXol are transported to the bile via transporters in the canalicular membranes of the hepatocytes. Mainly by energy-dependent ABCs such as ABCB1 (P-glycoprotein, Pgp); ABCC1 and ABCC2 (multidrug resistance-associated proteins MRP1 and MRP2); and ABCG2 (breast cancer resistance protein, BCRP) and by Ral binding proteins such as RALBP1.65

Local drug delivery systems As mentioned, the two local DDSs investigated in this thesis are LIPDOX and DEBDOX (Figure 7). A recent meta-analysis of randomized clinical trials where DOX was delivered with LIPDOX (given as cTACE) or DEB-DOX showed a similar overall survival rate, although DEBDOX was associ-ated with a higher objective tumor response rate.72 There are reports of a higher incidence of DOX-related adverse events and serious liver toxicity when using LIPDOX delivery for DOX compared to DEBDOX.73-75 It has, however, been demonstrated that DEBDOX is associated with a 6.6-times higher risk of developing liver/biliary events such as a dilated bile duct, VP narrowing, portal venous thrombosis or biloma/liver infarct than LIPDOX give as cTACE.76 The incidence of post-embolization syndrome (i.e. ab-dominal pain, nausea, vomiting and fever) is reported to be similar between the two DDSs.73

Lipiodol-based emulsion – LIPDOX Lipiodol was introduced for chemoembolization in HCC in the 1980s.77-79 It accumulates in tumor tissue, where it resides for several months after admin-istration.77, 80, 81 This is believed to be a consequence of the enhanced perme-ability and retention effect in the tumor microenvironment.82 The rationale behind using lipiodol as an intra-arterial DDS is based on the assumption

24

that lipiodol releases the drug where the emulsion is accumulated. In a rabbit tumor model, LIPDOX increased the intratumoral concentrations of DOX compared to intra-arterial infusion of a DOX solution.83 However, it has been demonstrated that the lipiodol accumulation is a poor quantative pre-dictor of local concentrations of DOX,84 which raises concerns about the potential drug-targeting advantage of lipiodol. In the following sections, the characteristics of lipiodol and the lipiodol–based emulsion are summarized briefly.

Lipiodol Lipiodol is marketed as Lipiodol Ultra Fluid® and Lipiodol® by Guerbet and is available in 47 countries.85 Although widely used globally for HCC, it was not until recently (2013-2014) that lipiodol was approved for TACE with the indication HCC in France, USA and Japan.85, 86 In Sweden it is prescribed by the physician as a compassionate-use DDS.

Lipiodol consists of di-iodinated ethyl esters of poppy seed oil, which is a heterogeneous mixture of linoleate (~60%), oleate, palmitate, stearate and linolenate.87, 88 The iodine content is 37–39% (i.e. 480 mg/ml) and this is the reason why lipiodol is visible in radiological techniques. Lipiodol is a heavy oil with a higher density than water (1.280 g/cm3

at 15°C) and has a high viscosity (25 mPa/s at 37°C, 50 mPa/S at 25°C). Intra-arterial infusion of lipiodol generates temporary, plastic embolization (i.e. adapted to vessel size) in the microvasculature.89, 90 This sinusoidal embolization causes the sinusoids to atrophy in healthy animals.91, 92 In animal models, lipiodol that passes the sinusoids is shunted forward to the portal venules where it accu-mulates and slows the sinusoidal blood flow.91-93 However, the microvascu-lar changes that occur during HCC development, i.e. development of leaky, irregular vasculature, may impair the transport of lipiodol into the portal areas.39

As lipiodol consists of lipids, many of the mechanisms involved in its dis-tribution and elimination are the same as for other lipids. In vitro studies suggest that lipiodol is incorporated into the cell membrane and transported into the cell by pinocytosis (liquid endocytosis).94, 95 Many active transport-ers (such as ABCA1, ABCB4, ABCB11, ABCD1, ABCG1, and ABCG5/8) are associated with lipid transport.96 Lipiodol is partly cleared from the blood as a result of phagocytosis by Kupffer cells.88 A study of the distribution of [14C]-lipiodol found that it is mainly eliminated to bile, while only trace amounts are found in the urine.81

Use of lipiodol-based emulsions as a drug-delivery system There is wide heterogeneity in the reported emulsion properties of LIPDOX.88, 97 The properties can vary with respect to the aqueous solution:lipiodol volume ratio, the use of densifiers, and the mixing proce-dure.88, 97 Several authors have highlighted the importance of a stable emul-

25

sion in improving the in vivo release and pharmacokinetic profile of the drug.97-100 In vitro and in vivo studies have indicated that the most stable LIPDOX emulsion is a water-in-oil emulsion with a volume ratio of one part drug (aqueous solution) and 2–4 parts lipiodol, where the specific gravity of the aqueous phase has been adjusted to that of lipiodol.93, 97, 99 However, the stability of the most stable LIPDOX emulsion is still poor. Therefore, the final formulation is traditionally prepared by the clinician minutes prior to administration by mixing with a syringe-pumping technique 10–15 times, see Figure 7A.

Figure 7. A) Preparation of the lipiodol-based emulsion. The physician receives two syringes from the hospital pharmacy, one containing lipiodol and the other contain-ing an aqueous solution of DOX. These two syringes are connected via a connector and the emulsion is prepared by pumping the solutions back and forth about 10–15 times. The figure was used with kind permission from the creator I.R. Dubbelboer. B) A syringe loaded with DC bead® mixed with a non-ionic contrast agent for visu-alization during the infusion.

Drug-eluting beads – DEBDOX The drug-eluting bead system DC Bead® was approved in 2007 as a novel DDS for the transarterial treatment of HCC.74, 101 DC Bead® (Biocompati-bles, BTG, UK) is indicated for treatment of hypervascular tumors and approved for loading with DOX. It is CE-marked (i.e. the product complies with European product directives) and, hence, is approved as a medical device instead of as a drug. Currently, it is approved in most of Europe, Rus-sia, Canada, Australia, parts of Asia, Africa and Latin America; however, it is not yet approved in the US.102

26

DC Bead® is a nondegradable, microparticulate DDS that consists of a polyvinyl alcohol-based hydrogel with associated chains of 2-acrylamido-2-methylpropane sulfonic acid sodium salt. The sulfonate sodium salt groups on these chains allow positively charged DOX molecules to replace the sodium ions during the loading process.103 The DOX-loaded DC Bead® (DEBDOX; Figure 7 B) releases the drug via a similar mechanism, by diffu-sion of ions into the bead, ion-exchange with DOX, and then diffusion of DOX out of the bead.103, 104 After superselective administration of DEBDOX, the beads cause permanent embolization of the targeted vessels at the same time as DOX is released slowly. The beads are available in several different size ranges: 70–150 µm, 100–300 µm, 300–500 µm, and 500–700 µm.102 One and three months after administration of DEBDOX100–300µm, the local tissue concentrations have been reported to be in or above the cytotoxic range for DOX (IC50 0.024–6.0 µM) for a range of 0.6 mm from the edges of the beads.105, 106

In vitro and in vivo studies have shown that the release rate of DOX is higher for smaller than for larger beads.97 The increase in release rate is probably related to the shorter diffusion path in the bead and the larger total surface area.97 In addition, smaller beads are associated with improved tox-icity and efficacy, which may be the result of better penetration and more distal embolization in tumor tissue.107

27

Aims of the thesis

The overall aim of this thesis was to assess the in vivo release of DOX from the currently used clinical formulations for palliative treatment of HCC (LIPDOX and DEBDOX) and its disposition in the surrounding tissue. A better understanding of the release mechanisms and the hepatic disposition of DOX from these DDSs in vivo is essential in order to evaluate which formulation achieves the best risk–benefit ratio, and to form the basis for future formulations and treatment strategies for this disease. Specific aims

• To examine the effects of lipiodol on the disposition of DOX in

plasma, urine and bile (Paper II) and in tissue (Paper III) in a multisampling-site, healthy pig model. Also, to investigate the mechanisms involved in increased tissue uptake of DOX.

• To study the in vivo release of DOX from the formulations LIPDOX and DEBDOX and the hepatic disposition of DOX after intra-arterial administration in a multisampling-site, healthy pig model (Paper I).

• To compare the in vivo release and hepatic disposition of DOX from the formulations LIPDOX and DEBDOX after intra-arterial admin-istration to patients diagnosed with HCC. In addition, to evaluate the efficacy, safety and urinary excretion of these formulations in the patients (Paper IV).

28

Methods

Ethics and surgical/interventional procedures

The in vivo multisampling-site pig model In Papers I, II and III, an advanced multisampling-site acute pig model was applied. This pig model was developed and validated by Petri et al. in 2006.108 The multiple sampling sites enable sampling prior to the blood reaching the liver (VP), after the liver (VH) and systemically (femoral vein; VF) as well as quantitative collection of bile and urine up to six hours after administration of a drug. These sampling sites and the large number of samples that can be collected provide a model suitable for mechanistic in vivo investigations in a healthy animal. The pig model was expected to be suitable for studies of transarterial treatment, since the pig liver is similar to that of a human in segmental anatomy, vascularization and size.109-112

The in vivo studies were approved by the local ethics committee for the use of laboratory animals in Uppsala, Sweden (C40/11). The pigs included in the studies were 10–12 weeks old, of mixed breed (Yorkshire and Swedish Landrace) and weighed on average 26.0 ± 1.3 kg in Paper I and 26.7 (range 24.2−30.3) kg in Papers II/III. The anesthesia and surgical procedures are explained in detail in Papers I and II. In brief, the pigs were sedated, anes-thetized, and connected to a respirator. They were continuously monitored for heart rate, blood gases, arterial and central venous pressure, and body temperature and electrocardiograms were run continuously, to maintain normal physiological values throughout the experiment. A midline incision of the abdominal cavity gave access to position the catheters in the VP, VH, VF, and bile duct. A tube was inserted into the bladder to collect urine. At the end of the experiment each animal was euthanized with an i.v. bolus dose of 20−30 nmol potassium chloride while still under anesthesia.

29

The clinical trial in patients with liver cancer The clinical trial described in Paper IV was performed in patients diagnosed with intermediate-stage HCC (BCLC stage B). These patients were treated with either LIPDOX as TAI or DEBDOX as TACE. The study was a multi-center study carried out at two hospitals in Sweden. The study protocol was approved by the regional ethical review board in Uppsala (Dnr 2013/227) and the study was registered in the European Clinical Trials Database (Eudra CTnr 2013-001244-56). All enrolled patients provided written informed consent. The inclusion and exclusion criteria are reported in Paper IV.

The intra-arterial procedure was equivalent at both hospitals, as described in Paper IV. Briefly, two catheters were placed intravascularly, as shown in Figure 8. As in the normal clinical procedure, a catheter was inserted into the hepatic artery via the right femoral artery for administration of the DDS. This catheter was placed lobarly for LIPDOX, in order to treat any eventual tumor satellites, and close to the tumor in the tumor-feeding arteries (super-selective) for DEBDOX. In this study, a blood sampling catheter was also placed in the vena cava orifice from the hepatic veins (VC/VH catheter) in order to obtain local measurements. This placement was accessed via the left VF through an introducer. Peripheral samples were taken directly from the introducer.

Figure 8. Catheters inserted intra-arterially during TAI or TACE treatment. One was used for drug administration and the other was used to collect local blood samples in the vena cava/hepatic vein orifice.

30

Study designs

Drug–excipient and drug–drug interaction studies in healthy pigs In Papers II and III, the mechanisms behind a potential LIPDOX-induced increased cellular uptake of DOX were investigated. This was done using a drug–excipient and drug–drug interaction approach between DOX and lipio-dol and/or cyclosporin A (CsA).

This study was designed with four parallel treatment groups (TI–TIV, Figure 9). There were two treatment phases: the reference phase (phase 1, 0−160 min) and the test phase (phase 2, 200−360 min). At times 0 and 200 min, all pigs received an i.v. infusion of DOX into an ear vein for 5 min. These two i.v. doses were given so that the pigs could serve as their own controls. Just prior to the second dose of DOX, i.e. phase 2, the treatment groups TII–TIV received an additional 5 min infusion of drug-free lipiodol emulsion (TII), a 20 min CsA infusion (TIII) or both (TIV) into the VP. If no other infusion was scheduled before the test phase, the volume was compen-sated for with physiological saline into the VP. CsA was used because it is a known inhibitor of efflux transporters important for DOX elimination and of several biotransformation enzymes important for drug metabolism.113, 114 Its main role was thus to investigate whether lipiodol interacts with transporters or enzymes important for DOX disposition. The experiment lasted for six hours, during which time samples were collected at predetermined time points from the VP, VH, and VF. Bile and urine were collected quanti-tatively (data presented in Paper II). At the end of the experiment, the pigs received a lethal dose of potassium chloride and pre-selected tissue samples (i.e. heart, liver, kidney and intestine) were collected (data presented in Paper III).

Figure 9. The study design for papers II and III.

31

The in vivo release studies in healthy pigs and HCC patients

Healthy pigs (Paper I) In this study, the primary aim was to investigate the effects of the intra-arterial DDSs on the pharmacokinetics of DOX in pigs. The study had a randomized, parallel design and four treatment groups, TI–TIV (Table 2). The pigs in TI were the control group and received a 55-minute i.v. infusion of DOX solution in an ear vein. The pigs in TII and TIII received a five-minute intra-arterial infusion of either LIPDOX as TAI (TII) or of DEBDOX (TIII) into the hepatic artery. In TIV, LIPDOX was administered through the hepatic artery and the bile duct simultaneously, in order to increase the local availability of the drug. The intended dose in all groups was 2 mg/kg. During the six-hour experiment, samples were collected at predetermined time-points from the VP, VH, and VF and continuously from the bile.

HCC patients (Paper IV) Data from the 14 patients who were included in the ongoing study during the period from January 2014 to March 2015 were assessed in the interim study analysis presented in Paper IV. The study was an open, prospective, multi-center study performed at the Uppsala University Hospital and the Karol-inska University Hospital, both in Sweden. The patients were divided into two study arms based on the hospital to which they were admitted, as the hospitals had different standard treatments. Six patients were enrolled at Uppsala University Hospital where LIPDOX was given as TAI, and eight

Table 2. The study design for Paper I

Group Formulation Dosea

(mg) Administration-

route

Sampling sites

Plasmab Bile

TI (n=4) Solution 47 ± 0.3 Intravenousc Yes Yes

TII (n=4) Lipiodol emulsion 3.1 ± 0.8 Intra-arterial Yes Yes

TIII (n=4) Drug-eluting beads 50 ± 0.0 Intra-arterial Yes Yes

TIV (n=2) Lipiodol emulsion 22 ± 3.5 Intra-arterial and via bile duct Yes No

aMean ± S.E., bsamples were collected from the portal, hepatic and femoral veins, cconstant infusion for 55 min given through the ear vein

32

patients were enrolled at Karolinska University Hospital where DEBDOX was given. The inclusion and exclusion criteria are reported in Paper IV.

The investigation schedule for each patient, presented in Table 3, can be summarized as follows: during and after TAI or TACE treatment, blood samples were collected at predetermined time-points from the VC/VH catheter to measure local concentrations (up to 6 h) and from a peripheral vein to measure the systemic concentrations (up to 7 days). After six hours, the VC/VH catheter was removed. Peripheral samples collected at 24 hours and 5–7 days were accessed through a needle in the arm. Urine was quanti-tatively collected from the start of treatment up to 24 hours. Safety parame-ters in the blood were measured before, 24 hours after and 5–7 days after treatment. Specifically, these parameters were: aspartate amino transferase, alanine amino transferase, alkaline phosphatase, prothrombin-international normalized ratio, bilirubin, creatinine, urea, hemoglobin, thrombocytes, leu-kocytes, albumin, sodium, potassium and C-reactive protein. In addition, the necrotic efficacy was evaluated by computed tomography or magnetic reso-nance imaging before and 4–6 weeks after treatment.

Table 3. Study flowchart for each patient. The pharmacokinetic blood samples were taken at visit 2 at pre-defined time points: before the infusion, and at 0, 5, 15, 30, and 60 min and 2, 6, and 24 h after the end of the infusion. Urine was collected during visit 2. One additional blood sample was taken at visit 3. The tumor response was assessed with imaging techniques such as computed tomography and magnetic resonance imaging.

STUDY PERIOD Visit 1

Baseline assessment

Before

Visit 2 TAI/TACE treatment

Day 1 (0–24 h)

Visit 3 Return to hospital 5–7 days

Visit 4 Return to hospital

4–6 weeks Pharmacokinetic

samples X X

Safety measurements* X X X

Imaging X X *Safety measurements: aspartate amino transferase, alanine amino transferase, alkaline phos-phatase, prothrombin-international normalized ratio, bilirubin, creatinine, urea, hemoglobin, thrombocytes, leukocytes, albumin, sodium, potassium and C-reactive protein.

33

Bioanalytical methods Two methods were used to quantify DOX and DOXol concentrations in various biological matrices: ultra-high performance liquid chromatography coupled with tandem mass spectrometry (UPLC-MS/MS), and ultra-high performance liquid chromatography coupled with mass spectrometry (UPLC-MS). Internal standards ([13C,2H3]-DOX and [13C,2H3]-DOXol) were used in all methods. The quantification of DOX and DOXol in plasma, bile and urine was carried out at the National Veterinary Institute (SVA), Uppsala, Sweden. These analyses were performed by UPLC-MS/MS as described in detail in Papers I, II and IV. Quantification in tissue homogenates was carried out at Uppsala University using a UPLC-MS system. Before injection onto the UPLC-MS system, the tissue homogenate was prepared by protein precipitation. The specific method is described in detail in Paper III.

Pharmacokinetic data analysis The pharmacokinetic parameters for DOX and DOXol in Papers I, II and IV were calculated using non-compartmental methods with WinNonLin 5.2 and 6.3 software (Pharsight, Mountain View, CA). In Papers I and II, compart-mental models were also applied. The maximum plasma concentration (Cmax) and the time to reach Cmax (tmax) were directly derived from the plasma conc-entration−time curve. The area under the concentration−time curve (AUC) for plasma and bile represents the drug exposure over time and was calc-ulated with the linear and logarithmic trapezoidal method for the ascending and descending parts to the last data point (Clast), respectively. AUC0–t was calculated from time zero to time t. AUC0–∞ was calculated by adding the residual area, which was estimated based on the slope (λz) of the two to three last measurable observations after Cmax. The terminal half-lives in the studies were calculated by ln2/λz.

Following an i.v. dose of DOX in Papers I and II, the plasma clearance (CL) of DOX was calculated as follows:

𝐶𝐶𝐶𝐶 = 𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝑖𝑖.𝑣𝑣./𝐴𝐴𝐴𝐴𝐶𝐶0−∞,𝑉𝑉𝑉𝑉 (eq.1)

where 𝐴𝐴𝐴𝐴𝐶𝐶0−∞,𝑉𝑉𝑉𝑉 is the AUC extrapolated to infinity based on the measured plasma concentrations from the VF (i.e. systemic concentrations).

34

The capacity of the liver to remove DOX from the circulation, defined as hepatic extraction (EH), was calculated based on data from the available sampling sites in the pigs before (VP) and after (VH) the liver:

𝐸𝐸𝐻𝐻 = (𝐴𝐴𝐴𝐴𝐶𝐶 𝑉𝑉𝑉𝑉 − 𝐴𝐴𝐴𝐴𝐶𝐶𝑉𝑉𝐻𝐻)/𝐴𝐴𝐴𝐴𝐶𝐶𝑉𝑉𝑉𝑉 (eq.2) where all AUCs were the observed AUCs up to six hours.

Further, the fraction of the DOX or DOXol dose excreted (fe) to bile and urine during the study period was calculated according to the following equation: 𝑓𝑓𝑒𝑒 = ∑(𝐶𝐶 ∙ 𝑉𝑉)/𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 (eq.3) where C is the concentration in the collected volume (V) of either bile or urine. The biliary excretion was assessed only in the pig studies while renal excretion was assessed in both Papers II and IV.

Tissue distribution of DOX and DOXol In Paper III, the tissue distribution of DOX and DOXol was investigated. The tissue distribution was defined as the tissue-to-plasma partitioning coef-ficient (Kp) as in equation 4.

𝐾𝐾𝑝𝑝 = 𝐶𝐶𝑡𝑡𝑖𝑖𝑡𝑡𝑡𝑡𝑡𝑡𝑒𝑒/𝐶𝐶𝑝𝑝𝑝𝑝𝑝𝑝𝑡𝑡𝑝𝑝𝑝𝑝 (eq.4) where Ctissue is the tissue concentration corrected for its blood content after six hours, as described in Paper III, and Cplasma is the VF plasma conc-entration measured at six hours in the same pig.

Mechanistic interaction between cyclosporin and DOX In Paper II, a mathematical multicompartment model was built in Win-NonLin (v 6.3) to increase the mechanistic understanding of CsA-induced alterations to DOX and DOXol pharmacokinetics (Figure 10). The model was derived from a three-compartment model with a central plasma compartment and two tissue compartments: liver and peripheral tissue. The mass transport in the model was calculated as a first order reaction according to equation 5; 𝑑𝑑𝐴𝐴/𝑑𝑑𝑑𝑑 = 𝐶𝐶 ∙ 𝑘𝑘 ∙ 𝑉𝑉 (eq.5) where A is the amount of compound, C is the concentration of compound in the compartment, k is the rate constant, and V is the volume of the compart-ment. For more details about the model, see Paper II.

35

Figure 10. Schematic overview of the multicompartment model describing the disposition of doxorubicin (DOX) and doxorubicinol (DOXol). Grey compartments with solid lines represent compartments that describe observed data (i.e. from plasma, urine and bile). Dashed arrows represent metabolism from DOX to DOXol or from DOXol to other metabolites.

The model was fitted to the observed data after administration of the first

dose of DOX, i.e. in the reference phase, and 11 model parameters were estimated. Thereafter, a sensitivity analysis was done to simulate what would happen if the estimated rate constants for metabolism and/or elimination to bile were systematically reduced by 99% for DOX and DOXol. These simu-lations were compared to the observed data, and were used to elucidate the mechanism responsible for the effects on DOX and DOXol disposition by lipiodol and/or CsA.

Bioavailability and in vivo release of DOX The bioavailability (F) is an approximation of the fraction of the dose that reaches the systemic circulation, which in this case is reached after intra-arterial hepatic drug administration. F may therefore be used to approximate the fraction released from parenteral DDSs in vivo. In Paper I, the bioavaila-bility was calculated by the following equation; 𝐹𝐹 = (𝐶𝐶𝐶𝐶 ∙ 𝐴𝐴𝐴𝐴𝐶𝐶0−6,𝐷𝐷𝐷𝐷𝐷𝐷)/𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷 (eq.6)

36

Another approach to approximate in vivo drug release is by deconvolution using WinNonLin software. Briefly, the plasma concentrations in the groups receiving the DDS were set as the response function, r(t), which was weighted against data obtained after an i.v. dose of DOX, the weighting function, w(t), which was derived from compartmental modeling of the i.v. dose. In Paper I, a two-compartment model best described the pig data col-lected over six hours after an i.v. dose. In Paper IV, human i.v. data were taken from literature and were best described by a three-compartment model.115 Deconvolution was then executed and the input function, i(t) was obtained; i(t) = r(t)//w(t). This function represents the apparent in vivo re-lease of drug from the DDSs into the local and systemic sampling sites.

Apparent intracellular availability of DOX Since metabolism of DOX to DOXol mainly occurs in the hepatocytes, the exposure of bile to DOX and DOXol was used to approximate the apparent intracellular availability of DOX, Fi;app, in Papers I and II (Equation 7). The following assumptions were made: i) that the metabolism followed linear distribution, ii) that reabsorption from the bile duct was limited, and iii) that there was no saturation of the efflux transporters and/or enzymes responsible for the biotransformation. 𝐹𝐹𝑖𝑖;𝑝𝑝𝑝𝑝𝑝𝑝 = 𝐴𝐴𝐴𝐴𝐶𝐶𝑏𝑏𝑖𝑖𝑝𝑝𝑒𝑒,𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝑝𝑝/(𝐴𝐴𝐴𝐴𝐶𝐶𝑏𝑏𝑖𝑖𝑝𝑝𝑒𝑒,𝐷𝐷𝐷𝐷𝐷𝐷 + 𝐴𝐴𝐴𝐴𝐶𝐶𝑏𝑏𝑖𝑖𝑝𝑝𝑒𝑒,𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝐷𝑝𝑝) (eq.7)

Statistical analysis In the pig studies, non-parametric tests (i.e. Kruskal-Wallis or Mann-Whitney) were used to test for significance between the treatments. In the clinical trial, Student’s t-tests were used to compare the treatments. Differences were considered to be statistically significant when p was <0.05. Post hoc analyses for multiple comparisons were performed when appli-cable. For details about the statistical analyses, see the relevant Paper.

37

Results and Discussion

In vivo pharmacokinetics of DOX following an i.v. dose in pigs DOX was rapidly and extensively distributed to tissues following five- and 55-minute i.v. infusions (ear vein; Papers II and I, respectively; Figure 11). The longer infusion time reduced the Cmax of DOX about 8-fold and the frac-tion of the formed metabolite DOXol (DOXol-to-DOX AUC0–3h ratio) about 9-fold (Figure 11, Table 4). The infusion rate-dependent metabolism of DOX into DOXol might be explained by saturation of the DOXol meta-bolism and/or excretion while plasma and tissue concentrations of DOX were high. The plasma concentration–time curves for DOXol declined in parallel with those for DOX, indicating that the pharmacokinetics of DOXol were limited by its rate of formation (Figure 11).

Figure 11. Dose-normalized plasma concentration–time curves for doxorubicin (DOX) and doxorubicinol (DOXol) following an i.v. infusion of DOX. The data were obtained from Paper I, in which four pigs received 55-min i.v. infusions of DOX, and Paper II, in which 12 pigs received 5-min i.v. infusions of DOX. The striped line represents the last datapoint measured during the 55-min infusion.

38

Plasma clearance and hepatic extraction of DOX The average plasma CL of DOX was about 48–59 mL/min/kg in pigs (Table 4). This CL was higher than in humans (13 ± 4.2 mL/min/kg) studied over seven days.116 The higher CL in pigs may have been the result of under-estimation of the residual area in these short studies (3 h and 6 h for the 5- and 55-min infusion studies, respectively). The plasma concentration–time curves in the pig studies were best described by a two-compartment model, although DOX has previously been best described by tri-exponential equa-tions with a long terminal half-life (35 h).61

The simultaneous blood sampling from the VP and VH enabled direct quantification of the EH of DOX. The EH was dependent on the infusion time; it was higher after the 55-minute infusion than after the five-minute infusion (Table 4). In Paper II (55-min infusion), it was also found that EH

Table 4. Summary of pharmacokinetic parameters for doxorubicin (DOX) and doxorubicinol (DOXol) in plasma, bile and urine following an i.v. dose of DOX into an ear vein in healthy pigs.

Pigs (Paper I) a Pigs (Paper II) b

DOX DOXol DOX DOXol

No. of pigs 4 4 12 12

Infusion time (min) 55 55 5 5

Dose (µmol) 85.7 (84.6–86.8) - 57.8 ± 4.2 -

Cmaxc (nM) 7.7 (6.5–10.3) 0.06 (0.045–0.24) 63 ± 17 0.59 ± 0.18

AUC0–3hc (µM∙min) 0.44 (0.37–0.55) 0.006 (0.005–0.027) 0.6 ± 0.11 0.07 ± 0.02

DOXol-to-DOX AUC ratioc (%) 1.4 (1.1–6.3) - 13 ± 3.6 -

CL (mL/min/kg) 59 (53–74) - 48 ± 11 -

EH (%) 41 (22–57) - 6.5 ± 13 -

fe;bile (%) 20 (6.6–23) 4.2 (2.6–5.5) 11 (4.5–16)d 2.4 (2.3–5.4)d

fe;urine (%) - - 5.8 (5.2–9.4)d 2.0 (1.8–2.6)d aData were from the femoral vein samples and are presented as medians (ranges); bdata were from the portal vein samples and are presented as means ± standard deviations; cdata were normalized after the administered dose; dmeasured for three pigs, six hours after two consecu-tive i.v. doses of DOX (dosing interval three hours), data presented as medians (ranges). AUC = area under the plasma concentration–time curve; Cmax = maximum concentration; CL = clearance; EH = hepatic extraction; fe = fraction excreted.

39

varied with time; and that it rapidly decreased once the infusion stopped. A time- and route-dependent phenomenon for EH has previously been observed for other drugs with this model (e.g. finasteride, verapamil and raloxifene).117-119 An intermediate EH of about 40% was observed for DOX during the longer infusion, which is also consistent with other data from domestic pigs and humans.120, 121 The intermediate EH value indicates that the hepatic clearance of DOX depends on the hepatic blood flow, the intrinsic metabolizing capacity of the liver, and the fraction of free drug in blood.122

Biliary and urinary transport of DOX The main elimination route for DOX was via bile, with about 4.5–23% of the dose excreted as unchanged drug during the first six hours after the infusions in pigs (Table 4). The dose-fraction of DOXol excreted was about five-fold less. The dose-fraction excreted to urine was lower than that excreted to bile: 5.8% and 2% for DOX and DOXol, respectively.

The exposure of bile to DOX and DOXol was extensive compared to that of plasma. For example, in Paper I, the bile-to-plasma AUC0–6h ratios for DOX and DOXol were 640 (410–670) and 5000 (1500–6200), respectively, given as median (range). The ten-fold higher bile-to-plasma AUC0–6h ratios for DOXol than for DOX may be explained by the higher initial concentra-tions of DOX in plasma directly after administration. Consequently this would result in a lower bile-to-plasma AUC ratio for DOX since DOXol is formed intracellularly. Within the cell, DOX (and probably also DOXol) is a substrate for several canalicular transporters (i.e. Pgp, BCRP, MRP1 and MRP2) which direct transport out of the cell to the bile rather than to the plasma.113, 123 In addition, the passive diffusion rate across lipophilic mem-branes is slower for DOXol than for DOX as a result of its more hydrophilic properties.

Tissue distribution of DOX DOX was extensively distributed in all the investigated tissues (Kp >140) after six hours, with the following rank order: kidney >liver >heart > intes-tine (Paper III). This is in agreement with the high volume of distribution (Vss= 130 ± 98 L) in the same pigs (Paper II). Explanations for this extensive tissue distribution include the amphiphilic properties of DOX, and also its high affinity for DNA, which would result in intracellular accumulation. Indeed, it has been reported that DOX concentrations in the cell nucleus are 50-fold higher than those in the cytoplasmic compartment, while intracellu-lar concentrations are generally 10- to 500-fold higher than in the extracellu-lar space.53

40

Drug–excipient and drug–drug interaction studies in healthy pigs

The effects of lipiodol on the pharmacokinetics of DOX Lipiodol did not affect the disposition of DOX and DOXol after an i.v. infu-sion of DOX into an ear vein in healthy pigs. Plasma, bile and tissue concen-trations were no different from those in control pigs (Figure 12). The fe into urine and the individual DOXol-to-DOX AUC ratio were not altered either. These results indicate a lack of direct interaction between lipiodol and DOX. Hence, there is no evidence of an interaction with transporters, enzymes and/or biological membranes important for the hepatobiliary disposition of DOX and DOXol.

Figure 12. The effect of lipiodol (LIP) and cyclosporin A (CsA) on doxorubicin (DOX; left columns) and doxorubicinol (DOXol; right columns) plasma concentra-tions, amounts accumulated in bile, and tissue concentrations in healthy pigs. DOX was given as an i.v.dose at 0 and 200 min. LIP and/or CsA were administered before the second dose of DOX (indicated by the red striped line). *p <0.05, **p <0.01.

41

We suggest that the in vivo increase in tissue uptake of DOX (mainly by the tumor) when DOX is emulsified in lipiodol is related more to the encap-sulation of DOX in the formulation than to excipient–drug interactions on a membrane, enzyme or transport level.83, 99, 124 The complexity of the biologi-cal systems makes it difficult to study the detailed effects of formulation properties on drug distribution with in vitro models.125 Therefore, more systematic in vivo approaches are needed to elucidate the mechanisms be-hind the increased uptake of DOX by tumor tissue.

The effects of cyclosporin on the pharmacokinetics of DOX When CsA was administered with DOX to pigs, the plasma exposure to DOX was unaffected but that to DOXol was increased (Figure 12, Table 5). In addition, there was a clear reduction in the biliary excretion of both DOX and DOXol. CsA treatment (TIII and TIV) caused an increase in liver, kid-ney and intestine concentrations of DOX compared to control (TI and TII, p <0.05). These effects of CsA on DOX and DOXol disposition in plasma and bile are in accordance with those in other short (<4 h) in vivo studies.126-129

Table 5. Pharmacokinetic parameter changes after infusion of cyclosporin (CsA) to pigs (groups TIII and TIV) in combination with doxorubicin administration, com-pared to control groups without CsA (TI and TII). No change (±30%) is indicated by ↔, a 30–60% change is indicated by ↑ or ↓, a 60–80% decrease is indicated by ↓↓ and a 80–100% decrease is indicated by ↓↓↓.

Doxorubicin Doxorubicinol Plasma

AUC ↔ ↑↑ Cmax ↔ ↑↑

Bile fe, bile ↓↓↓ ↓↓↓

AUCbile/AUCplasma ↓↓↓ ↓↓↓ Fiapp ↑ N.A.

Urine fe, urine ↓↓ ↔

Tissues Liver conc. ↑ N.D.

Kidney conc. ↑↑ ↔ Heart conc. ↔ N.D.

Intestine conc. ↑ N.D. AUC = area under the concentration–time curve; conc. = concentration; Cmax = maximum concentration; fe = fraction excreted; Fi;app = apparent intracellular availability; N.A. = not applicable; N.D. = not detected.

42

We suggest that DOX and DOXol use the same canalicular transport pro-teins, since the inhibition of biliary transport by CsA was similar for the parent drug and the metabolite.

CsA is a known inhibitor of the efflux transporters Pgp, MRP2 and BCRP, and of influx transporters (e.g. OATPs) and metabolic enzymes (e.g. CYPs).113, 123, 130 The unchanged plasma exposure to DOX following CsA treatment suggests that the interaction with DOX was not caused by inhibi-tion of basolateral influx transporters. For other drugs where CsA is known to interact with basolateral uptake transporters, CsA treatment increased plasma exposure to and decreased hepatic clearance of the parent drug.131, 132 This theory is further supported by simulations made during sensitivity analysis of the multi-compartment model; it was necessary for the metabo-lism of DOX to DOXol to be unaffected in the model in order to fit the observed increase in plasma concentrations of DOXol. In analogy with the lack of reports, it suggests that CsA does not inhibit AKRs and CBRs.

Pgp, MRP2 and BCRP are also expressed in tissues other than the liver,7 which may explain the observed increase in DOX concentrations in kidney and intestine, and the decrease in urine. Reduced efflux of DOX to bile, urine and the intestinal lumen will increase the concentrations in liver, kidney and intestine. Higher intracellular concentrations of DOX increase the availability of substrate for metabolism into DOXol. However, the DOXol concentrations in kidney and the renal excretion of DOXol were unchanged after inhibition by CsA. This observation might be explained by varying expression and activity of cytosolic CBRs and AKRs in the different tissues.69, 133 The high intracellular concentrations of DOX caused by CsA inhibition may have saturated these enzymes in the kidney but not in the liver or intestine, and/or they may have induced other metabolic pathways leaving DOXol concentrations in the kidney unchanged.69

The in vivo release studies in healthy pigs and HCC patients This section discusses the in vivo release and local distribution of DOX after administration with LIPDOX and DEBDOX. There is an important differ-ence in the administration approach between these two DDSs. LIPDOX is administered selectively to one liver lobe without injecting additional embo-lic agents, resulting in transient embolization. DEBDOX is administered superselectively to the tumor-feeding arteries close to the tumor, where it provides complete embolization and slowly releases DOX.

43

Dosage and tumor characteristics The tumor characteristics and dosage of DOX in the DDSs for the preclinical and clinical study subjects are summarized in Table 6.

In healthy pigs (Paper I), it was impossible to administer the intended dose of LIPDOX (50 mg of DOX) as an unexpected stasis occurred in the artery after the infusion. As a consequence, a lower dose was given. Pigs seem to be more sensitive to the highly viscous lipiodol than humans, maybe because of differences in blood vessel anatomy and/or physiological adapt-ations. It has, for example, been reported that younger pigs (aged three months) have a narrower VP with a thinner vessel wall than humans.134 The droplet sizes in relation to the blood vessel size may also be an important factor for the increased vessel resistance, since larger droplets have higher embolic capacities. Despite these difficulties in administering LIPDOX, the DDS released DOX sufficiently well to be able to quantify the conc-entrations in plasma. In liver cancer patients (Paper IV), the treated tumors were larger at screening in the LIPDOX group than in the DEBDOX group (p <0.05).

The pharmacokinetics of DOX after delivery with LIPDOX or DEBDOX The dose-normalized plasma concentrations of DOX and DOXol in pigs and patients (Paper IV) following LIPDOX and DEBDOX delivery are displayed in Figure 13. The Cmax of DOX was 10–11 times higher after LIPDOX than after DEBDOX in both species. For both DDSs, the Cmax was followed by a rapid distribution phase. The plasma exposure to DOX (AUC0–6h in pig and AUC0–24h in patients) was 2–3 times higher after LIPDOX than after DEB-DOX. The higher exposure to LIPDOX was observed both locally and

Table 6. Baseline characteristics and dosage of intra-arterial LIPDOX and DEB-DOX treatment in the in vivo release studies with healthy pigs and human liver cancer patients. Pigs (Paper I) Patients (Paper IV)

LIPDOX DEBDOX LIPDOX DEBDOX No of included subjects 4 4 6 8

Baseline tumor size (mm)a No tumor No tumor 72 (46–112) 26 (8–75)

Child-Pugh (A/B/na) - - 6/0/0 5/2/1 Dose of DOX (mg)a 2.3 (2.3–5.4) 50 (50–50) 50 (50–50) 90 (22.5–150) No. of treated tumors - - 6 9 aMedian (range). DEBDOX = microparticle DOX-eluting beads; DOX = doxorubicin; LIPDOX = lipiodol-based emulsion containing DOX; na = not available

44

systemically. This relationship between the two DDSs is consistent with previously reported data in patients.74 In summary, LIPDOX released DOX to the systemic circulation to a greater extent than DEBDOX during the study.

It has been suggested that DEBDOX is associated with fewer DOX-related adverse events because of the lower systemic exposure.73, 74 How-ever, the recommended clinical dose for DEBDOX is three times higher than that for LIPDOX (150 mg vs 50 mg), which compensates for the difference in systemic exposure. For example, using the local bioavailability of DOX in patients, this corresponds to about 16 mg of DOX released from both DDSs after six hours. The Cmax of DOX is, however, still lower for DEBDOX, which might affect the frequency of adverse events.

Figure 13. Dose-normalized plasma concentration–time curves after administration of doxorubicin (DOX) in a lipiodol-based emulsion (LIPDOX) or DOX-eluting microparticle beads (DEBDOX) to pigs [(A) DOX and (B) doxorubicinol (DOXol)] and to patients [(C) DOX and (D) DOXol]. Measurements were taken for six hours in pigs and for seven days in patients; the insets display the first six hours in patients.

45

Release from LIPDOX In most of the pigs and patients, 50% of the dose was released from LIPDOX into the circulation within the first 30 min (Figure 14, Table 7).

The rapid release from LIPDOX suggests that the emulsion breaks down once it is infused into the circulation, just as it does when it comes into contact with elution media in vitro.135-137 Indeed, the sustained delivery of DOX is highly correlated with the physicochemical properties of the emul-sion.100 In Papers I and IV, the physical stability of the emulsion was opti-mized with a densifier and the most optimal volume ratio between lipiodol and the aqueous DOX solution (1:3.3 aqueous DOX in lipiodol) was used.97,

99, 138 Despite these measures, the emulsion still starts to separate 10 min after preparation in vitro.99

Figure 14. The cumulative fraction of the dose of doxorubicin (DOX) released (approximated by deconvolution) as a function of time at the systemic (A) and local (B) sampling sites over six hours following intra-arterial injection of a lipiodol-based emulsion containing DOX (LIPDOX) or DOX-eluting microparticle beads (DEBDOX) in pigs and liver cancer patients. The results for liver cancer patients over seven days are presented in graph (C). LIPDOX released DOX more rapidly than DEBDOX.

46

The release mechanism for DOX from the emulsion involves partitioning between the water and oil phases as well as diffusion through the oil phase.137 The fraction of DOX that partitions into the lipid phase of the emulsion has the potential to follow lipiodol into the tissues. Inter-individual differences in tumor size and vessel anatomy may be of greater importance for LIPDOX than for DEBDOX because of the unstable nature of the emul-sion, since the fixed position of DEBDOX and the reduced blood flow will enhance the probability that DOX will reach the arteries adjacent to the administration site.2

Release from DEBDOX Approximately 32% of the dose was released from DEBDOX into the local compartment in pigs after six hours. In patients, the corresponding fraction of the dose released was 13% locally after six hours and 15% systemically after 24 hours (Figure 14, Table 7). This difference in in vivo release between pigs and patients (32% vs 13%) may be the result of species differ-ences, as well as the absence of tumors or cirrhotic livers in the pigs. DOX could therefore have been distributed more freely in pigs, as they lacked the transport barriers associated with the tumor microenvironment.2 Comparison of the half-lives5–7d of DOX between the DDSs indicates that the release rate

Table 7. The bioavailability (in this case, the release) and local distribution of dox-orubicin (DOX) following intra-arterial delivery in a lipiodol-based emulsion (LIPDOX) or DOX-eluting microparticle beads (DEBDOX) to healthy pigs (Paper I) and liver cancer patients (Paper IV).

Systemic Local Species LIPDOX DEBDOX LIPDOX DEBDOX

Bioavailability/ release

F (%) a,b Pig 29 (22–44) 14 (9.4–18) 85 (70–94) 24 (18–29)

Fdeconv, 6h (%) a,d Pig 38 (28–56) 14 (11–21) 110 (92–120) 32 (24–38)

Fdeconv, 6h (%) c,d Human 51 ± 7.6 11 ± 5.6 69 ± 25 13 ± 6.5

Fdeconv, 24h (%) c,d Human 51 ± 7.6 15 ± 8.5 NA NA

Local distribution

Fi;app a Pig 32 (22–43) 18 (17–21) NA NA

AUClast,Doxol/DOX c Human 1.2 ± 0.53 0.54 ± 0.21 1.3 ± 0.70 0.11 ± 0.03

RD, 6h (%) c Human NA NA 31 ± 25 87 ± 6.4 aData are shown as medians (interquartile ranges); bF was obtained using equation 6; cdata are shown as means ± standard deviations; ddata were derived from deconvolution. Deconvoluted values are only presented up to 24 h due to few observations at 5–7 days. AUC = area under the concentration–time curve; F = bioavailability; Fi;app = apparent intracellular availability; NA = not applicable; RD = remaining dose in the liver, calculated by 100-Flocal for each pa-tient.

47

from DEBDOX is lower than the elimination rate of DOX (approximately 3-fold). The mechanism for release of DOX from DEBDOX is based on ion exchange and is then driven by diffusion out of the beads followed by diffu-sion through a stagnant layer surrounding the beads to reach its target site.

The fixed position of DEBDOX in the artery after the administration may have implications for the release of DOX, especially as this DDS is supposed to deliver DOX over a significantly longer period than LIPDOX. Both in vitro and in vivo studies have shown a slow release rate from DEBDOX, which suggests that the interaction between DOX and the functional unit (sulfonate ions) of the DDS is strong.73, 105, 106, 136, 139-141 For example, it has been reported that 48% and 89% of the dose is released in healthy pigs after one and three months, respectively.106 In addition, a foreign body response has been observed around DEBDOX in vivo, leaving a fibrotic capsule around the beads.105 The static environment created around the beads may restrict the release of DOX, raising concerns that the loaded dose might not be completely released.88, 97

Local distribution of DOX delivered by LIPDOX and DEBDOX For both DDSs, the fraction of the dose released at the local sampling site was higher than that at the systemic sampling site in both pigs and humans, with a 2- to 3-fold difference between the sites in pigs and a smaller differ-ence in patients (1.1- to 1.3-fold; Table 6). The concentrations at the systemic sampling sites were lower because DOX mixes with the circulation and is distributed to other tissues.

It is important to know the local distribution, and hence concentrations, at the target site in order to predict the anti-tumor effects and adverse events. In Papers I and IV, we approximated the availability in the liver and the hepatocytes with different approaches.

Intracellular availability of DOX In pigs, the biliary excretion of DOX and DOXol was used to approximate the apparent intracellular availability (Fi;app) of DOX, which appeared to be higher for LIPDOX than for DEBDOX (Table 6). The biliary flow was not affected by either of the formulations (median flow of 14–18 µL/min/kg). In patients, the local plasma DOXol-to-DOX AUC ratio was 2.7-fold higher after LIPDOX delivery than after DEBDOX delivery (p <0.001; Table 6). The higher ratio for LIPDOX than for DEBDOX may be explained by dif-ferences in the local distribution of DOX and the subsequent metabolism. One hypothesis for explaining these two independent observations is that increased intracellular delivery of DOX from LIPDOX would lead to more substrate available for DOXol formation. This is consistent with observa-tions made by Furuta et al. in 1988.83 In the liver cancer patients, who had cirrhotic livers to a greater or lesser extent, the increase in the DOXol-to-

48

DOX AUC ratio could have been the result of wider distribution of DOX after LIPDOX administration than after DEBDOX. Thus, DOX from LIPDOX could reach the healthier liver regions and/or other tissues with a greater metabolic capacity than the areas in close vicinity to the infused DEBDOX. The liver and kidney have the greatest capacity (highest intrinsic clearance) of the main sites in the body for conversion of DOX to DOXol.69,

71 The metabolic capacity of CBRs and AKRs in the cirrhotic liver and in patients with HCC is not yet fully understood but it is expected to be reduced since many CYP enzymes that are important for the metabolism of drugs other than DOX are down-regulated in the cirrhotic liver and HCC patients.142 Dose of DOX remaining in the liver In patients, the fraction of the dose that was released to the local sampling site was used to evaluate how much of the dose still remained in the liver after six hours (RD; see equation 2 in Paper IV). The RD at six hours was 31 ± 25% for LIPDOX and 87 ± 6.4 % for DEBDOX. This fraction might still be bound to the DDS and/or be distributed to the different compartments of the liver (i.e. extra- and intracellular compartments in tumor or nontumor tissue). The first option is less likely for LIPDOX with its fast release, although DOX may follow LIPDOX into the cells. For DEBDOX, on the other hand, the slow release and strong interaction between DOX and the beads suggest that most of the dose is still bound to the DDS after six hours.

Safety and anti-tumor efficacy of the DDSs with respect to the pharmacokinetics of DOX In Paper IV, the efficacy and safety of the two loaded DDSs were not the primary objective. However, it was of interest to relate those parameters to the release and pharmacokinetics of DOX. The number of included patients was too low for adequate comparison between the DDSs regarding safety and efficacy. The responses and adverse events are presented in Table 8.

The short-term necrotic effect of the treatment tended to be better with DEBDOX, as expected.73, 74 This is because DEBDOX causes complete embolization from a fixed location, which induces ischemia, in contrast to LIPDOX which exerts transient embolization and is more widely spread in the liver tissue. Unloaded beads have approximately the same necrotic effect as DEBDOX, with a longer time to progression of the tumor when DOX is present.143 The stasis caused by DEBDOX and the subsequent blood clotting and necrosis may affect the release and pharmacokinetics of DOX. It has been shown in explants from DEBDOX-receiving patients that the DOX concentrations are higher in necrotic areas than in non-necrotic areas after one month.105 This may be related to the low cellularity in necrotic areas,

49

which will result in less restricted diffusion than in the tumor micro-environment.144

Table 8 indicates that the incidence of post-embolization syndrome (i.e. abdominal pain, nausea, vomiting and fever) was similar between LIPDOX and DEBDOX recipients, while liver failure only occurred in LIPDOX recipients. This observation might also be related to the degree of emboliza-tion and the wider spread of LIPDOX in the liver. However the observations are few and therefore one should be careful with interpretations before more patients are included in the study.

Table 8. Treatment responses according to mRECIST and adverse events reported during the study period in Paper IV. The mRECIST classification was assessed by computed tomography or magnetic resonance imaging prior to and 4–6 weeks after transarterial treatment of patients with hepatocellular carcinoma with the lipiodol-based emulsion containing doxorubicin (LIPDOX) or doxorubicin-eluting micro-particle beads (DEBDOX). LIPDOX DEBDOX mRECISTa

Progressive disease 0 (0%) 0 (0%) Stable disease 2 (50%) 0 (0%) Partial response 1 (25%) 5 (62.5%) Complete response 1 (25%) 3 (37.5%) Adverse events and causality to treatmentb

Probable Post-embolization syndromec: 3 Liver failure: 2

Post-embolization syndromec: 5 Infection in necrotic part of tumor: 1

Possible

Death: 1 Back pain: 1

Unlikely Rectal bleeding: 1 Confusion: 1 Headache: 1

Urinary tract infection:1

aThe Modified Response Evaluation Criteria in Solid Tumors include tumor size and the extent of necrosis in the tumor in the evaluation.145 bReported in absolute numbers; the rela-tionship with treatment is indicated as probable, possible or unlikely. cPost-embolization syndrome is an expected adverse event following embolization and includes abdominal pain, nausea, vomiting and fever.

50

Conclusions

The main objective of this thesis was to evaluate two current clinical formu-lations (LIPDOX and DEBDOX) of the chemotherapeutic drug DOX from a biopharmaceutical perspective. Specifically, the focus was on assessing the effects of these DDSs on the disposition of DOX in vivo. Understanding the in vivo release and disposition of a drug from these DDSs is an important step in the development and optimization of new DDSs. With novel DDSs there are opportunities for improving the risk–benefit ratio for patients with HCC by prolonging their overall survival and/or improving their quality of life. The main conclusions drawn from the investigations in this thesis were: • DOX and DOXol were extensively distributed to tissue and bile in pigs.

The hepatic extraction of DOX was time- and infusion-dependent. The intermediate EH (~40%) following a 55-minute i.v. infusion suggests that the hepatic elimination of DOX is dependent on the hepatic blood flow, intrinsic metabolizing capacity, and fraction of free drug.

• The excipient lipiodol did not interact with transporters or enzymes

important for DOX disposition in vivo. Nor did it alter the membrane fluidity to enhance the cellular uptake of DOX. Instead it is suggested that the increased tissue accumulation of the drug delivered in lipiodol could be related to the properties of the formulation (drug + excipients) rather than to a direct interaction between excipient and drug.

• The transport protein inhibitor CsA inhibited carrier-mediated excretion

of DOX to bile and urine. The excretion of DOXol to bile was also in-hibited. The site of inhibition of biliary transport was assumed to be at the canalicular membrane rather than at the basolateral membrane, since plasma concentrations of DOX were unchanged but those for DOXol were increased. An inhibition on the basolateral membrane would have affected the plasma concentrations of the parent drug.131, 132 Multi-compartment modeling and sensitivity analysis confirmed that the transport inhibition by CsA occurred at the canalicular membrane. In addition, it is suggested that CsA does not inhibit enzymes responsible for the metabolism of DOX to DOXol (i.e. AKRs and CBRs).

51

• DOX was released faster from LIPDOX than from DEBDOX in vivo and was distributed more widely in the body, which could be associated with a higher risk of DOX-related adverse events. However, the three-times-higher recommended dose of DOX with DEBDOX (150 mg vs 50 mg) may lead to a similar systemic plasma exposure to DOX as LIPDOX.

• DEBDOX only released 15% of the dose to the circulation during 24

hours, which confirms its relatively extended release and prolonged tu-mor exposure. Future in vivo studies need to investigate whether this re-lease rate for DOX is too slow and whether the total dose might not be released.

• Local distribution of DOX from LIPDOX and DEBDOX differed in

patients, with 31% and 87% of the dose remaining in the DDS or liver after six hours, respectively. DOX from LIPDOX is likely to have been distributed in the liver because of its fast, extensive release character-istics, while most of the DOX from DEBDOX is likely to still reside within the DDS after six hours.

• DOXol was formed to a greater extent after delivery by LIPDOX than

after DEBDOX in both pigs and patients. This can be explained by i) more extensive intracellular delivery from LIPDOX, since DOX could be co-transported with the lipid part of the emulsion, and/or ii) wider distribution of DOX from LIPDOX into regions with higher metabolic capacity and activity.

• The plasma concentration–time curve for DOX was similar in both

healthy pigs and liver cancer patients. This indicates the usefulness of the advanced multisampling-site pig model when assessing in vivo release and distribution of DOX from intra-arterial DDSs for treatment of HCC. However, the pig model was less suitable for the lipid-based emulsion because the formulation induced unexpected early stasis of the blood vessel on administration. This has to be taken into consideration if a high dose is required.

52

Future perspectives

Finding a treatment for patients with cancer is challenging. It is well known that formulation of an active drug into a DDS can facilitate the targeting of the tumor by the drug while reducing the likelihood of adverse events. It therefore makes sense to develop efficient delivery systems for chemo-therapeutic drugs known to be potent against cancer. In this thesis, DDSs used clinically in palliative treatment for primary liver cancer were studied in detail in vivo to learn which approach delivers the drug most efficiently.

The studies conducted in this thesis concluded that future optimized novel drug delivery strategies should be more similar to DEBDOX than to LIPDOX. This is because DEBDOX is a more robust pharmaceutical prod-uct from which the drug does not immediately separate on contact with blood causing a high extrahepatic DOX exposure. With controlled release, as in DEBDOX, the inter-individual response and frequency of drug-related adverse events may be easier to predict. However, it is important to further investigate the impact of a number of pathophysiological factors on the slow release of DOX from DEBDOX. The local drug–tumor exposure relation-ship has not been established and further in-depth investigation is required.

This thesis is part of a liver cancer treatment-optimization project at the Department of Pharmacy, Uppsala University, Sweden. Apart from the work in this thesis, the project also includes studies of the in vitro release of DOX from LIPDOX and DEBDOX as well as physiologically-based biopharma-ceutical models based on the in vitro and in vivo studies. These studies will create a knowledge base for the future development of novel DDSs with optimal in vivo drug release. The primary goal in the future is to design an injectable formulation that can be given repeatedly and is easily monitored using imaging techniques, and that can release a combination of drugs effec-tively in the vicinity of the tumor tissue.

53

Populärvetenskaplig sammanfattning

Patienter med primär levercancer (cancer som uppstår i levern) har en dålig prognos. Cancerformen är den näst dödligaste i världen. Risken att insjukna i levercancer ökar om man har levercirrhos, Hepatit B eller C, är alkoholist eller om man fått i sig en viss typ av mögelgifter (aflatoxiner). Nyligen framkom det även att risken ökar om man har diabetes eller fetma. Om levercancern inte går att operera bort kan man få livsuppehållande behand-ling med syfte att bromsa sjukdomsförloppet och öka patientens livskvalitet. Levercancer i intermediärt stadium behandlas med cellgiftet doxorubicin i Sverige. Doxorubicin ges i en läkemedelsbärare som sprutas in direkt i levern via leverartären, det blodkärl som försörjer merparten av levercancer-tumören med syre och näringsämnen. Läkemedelsbäraren används för att öka chanserna för doxorubicin att nå tumören där den kan utöva sin effekt och minska risken att doxorubicin sprids till resten av kroppen och ge allvar-liga biverkningar. Man kan även dosera ett blodproppsbildande ämne efter det att man doserat läkemedelsbäraren för att stoppa blodflödet till tumören. Det gör att tiden som cellgiftet befinner sig lokalt nära tumören ökar och att näring- och syretillförseln till tumören avstannar. Idag finns även små partik-lar i mikrostorlek laddade med doxorubicin tillgängliga i sjukvården. Dessa mikropartiklar fungerar då både som läkemedelsbärare och blodproppsbil-dare i en och samma produkt.

Syftet med denna avhandling var att undersöka två läkemedelsbärare som idag används vid behandling av levercancer. Den ena läkemedelsbäraren består av en vatten-i-olja emulsion (ungefär som en salladsdressing) och den andra består av mikropartiklar. Båda bärarna är laddade med doxorubicin när man ger det till patient. Emulsionen baseras på oljan lipiodol som är ett kon-trastmedel som man kan följa på röntgen. Där kan man se att oljan ansamlas i tumörer upp till flera månader efter behandlingen. Syftet med denna av-handling var främst att ta reda på hur bra dessa läkemedelsbärare släpper ifrån sig och fördelar doxorubicin lokalt i och runt tumören. Detta för att kunna avgöra vilken strategi som är bäst vid utveckling och optimering av nya bärarsystem som skulle kunna förbättra behandlingen mot levercancer så att patienten lever längre och får en bättre livskvalitet.

Frisläppningen och fördelningen av doxorubicin från emulsionen och mikropartiklarna studerades både i friska grisar och i patienter med lever-cancer genom att ta blodprover vid flera tidpunkter från flera ställen (både lokalt och ute i blodomloppet) och genom att samla urin. I grisar samlades

54

också galla. Detta för att bestämma läkemedelshalten av doxorubicin och dess nedbrytningsprodukt doxorubicinol i kroppen.

Den lipiodol-baserade emulsionen släppte doxorubicin mycket snabbare och mer omfattande ut i blodomloppet än vad mikropartiklarna gjorde både i grisar och i patienter. Emulsionen släppte ifrån sig ca 50% av den laddade dosen till blodomloppet inom 30 minuter medan mikropartiklarna enbart släppte ifrån sig 15% av den laddade dosen under de första 24 timmarna efter dosering i patienter. Det är tveksamt om mikropartiklarna kan släppa ifrån sig allt doxorubicin som de laddats med i kroppen och därför behövs ytterligare forskning och fördjupning på området för att förstå detta. En större andel doxorubicinol bildades när doxorubicin gavs i emulsionen än vad det gjorde när det gavs i mikropartiklarna. Två hypoteser kan förklara detta. Antingen sprider emulsionen doxorubicin till en större del av levern och/eller andra regioner i kroppen som har en annan förmåga att bryta ned doxorubicin, eller så ökar emulsionen upptaget av doxorubicin in till cellerna där nedbrytningen sker. Den tumördödande effekten var jämförbar mellan läkemedelsbärarna, man kan dock ana en liten fördel med mikropartiklarna.

De mekanismer som är involverade i fördelningen av doxorubicin inuti kroppen och i elimineringen av doxorubicin från kroppen undersöktes i gri-sar. Detta gjordes genom att injicera doxorubicin i blodomloppet efter att grisen behandlats med lipiodol och/eller cyclosporin. Cyclosporin är ett lä-kemedel som hämmar flera proteiner som transporterar läkemedel in och ut ur kroppens celler. Det hämmar också enzymer som är viktiga för fördel-ningen och elimineringen för flera läkemedel. Vi kom fram till att doxoru-bicin och dess nedbrytningsprodukt doxorubicinol framför allt elimineras via gallan genom transportproteiner som är placerade i cellmembranet mellan levercellerna och gallgångarna. Cyclosporin hade däremot ingen effekt på de enzymer som bryter ned doxorubicin till doxorubicinol. Tidigare studier har visat att läkemedelshalterna i tumören ökar om man administrerar doxoru-bicin samtidigt som lipiodol. I våra interaktionsstudier såg vi att lipiodol inte påverkar fördelningen av doxorubicin när det ges separat. Lipiodol har alltså inte någon effekt på transportproteiner, enzymer eller membran viktiga för doxorubicins fördelning. Därmed måste det ökade upptaget av doxorubicin som man ser när man administrerar det tillsammans med lipiodol bero på formuleringens egenskaper (d.v.s. läkemedelsbärare + doxorubicin) istället.

Sammanfattningsvis har denna avhandling bidragit med ny kunskap om hur dessa två läkemedelsbärare, som är i klinisk användning, släpper ifrån sig och fördelar doxorubicin. Dessutom har möjliga mekanismer för ökat cellupptag av doxorubicin studerats. Framtida läkemedelsbärare ämnade för livsuppehållande behandling mot levercancer borde likna mikropartiklarna mer än emulsionen eftersom mikropartiklarna släpper ifrån sig doxorubicin mer kontrollerat. Detta gör att förutsättningarna är bättre för att kunna för-utse hur bra effekt behandlingen kommer att ha och för att förutse hur många eventuella biverkningar läkemedelet kommer att ge.

55

Acknowledgements

The work in this thesis was carried out at the Department of Pharmacy, Fac-ulty of Pharmacy, Uppsala University, Sweden. Jag vill börja med att tacka IF:s stiftelse och Apotekare CD Carlssons stif-telse (Apotekarsocieteten), Anna Mara Lundins stipendiefond (Smålands Nation) och Absorption Systems Travel Grant (GPEN), för stipendier som gjort det möjligt för mig att åka på internationella kurser och konferenser. Det finns många som jag vill tacka för bidrag till arbetet med denna avhand-ling och för hjälp och stöd som jag har fått på vägen; Hans Lennernäs – Min handledare, för att du antog mig som doktorand. För att du trott på mig och för att du peppat mig när det har varit tufft. Din energi och entusiasm har varit ovärderlig. Särskilt stort tack för all hjälp och stöd du gett mig under sista tiden av avhandlingsskrivandet. Niklas Axén – Min bihandledare, för givande diskussioner om projektet i stort, framför allt i början av min doktorandtid. Dina fikabrödsleveranser har varit mycket uppskattade avbrott under långa dagar på lab. Stort tack också för hjälpen under avhandlingsarbetet. Erik Sjögren – Min bihandledare, för att du alltid ställer upp med tid och hjälp när man ber om det. Det har varit ett nöje att arbeta med dig och ta del av din kunskap i mina senare projekt.

Ilse Dubbelboer och Emelie Ahnfelt – Mina fina projektkollegor och vänner. Tack för er energi, för givande forskningsdiskussioner, för samtal om livet, för roliga resor, för stärkande promenader och för en himla massa skratt! För att ni har lärt mig att inte ha ångest och att gå hem i tid. Jag kommer att sakna att jobba tillsammans med er! My co-writers – Ulf Bondesson, Mikael Hedeland, Rickard Nyman, Char-lotte Ebeling Barbier, Agneta Norén, Frans Duraj, Micheline Piquette-Miller, Per Stål, Torkel Brismar, Amar Karalli, och Rimma Axelsson, for excellent collaborations. Ett särskilt tack till Ulf och Mikael för ovärderlig hjälp med analyser och kommentarer på SOP:ar författade för det cytosta-

56

tika-lab som jag varit med om att bygga upp. Rickard och Charlotte ska ha ett stort tack för att ni har varit så välkomnande på Interventionsradiologen på Akademiska Sjukhuset i Uppsala. Ett stort tack även till Per och Amar på Karolinska Sjukhuset i Huddinge som även där har utgjort grunden för ett väldigt bra och trevligt samarbete. Ni har lärt mig otroligt mycket. All personal på Karolinska och Akademiska Sjukhuset som hjälpt till med provtagning och gett utrymme för mig i ert dagliga arbete. Däribland särskilt Pia Loqvist, för all hjälp på Karolinska Sjukhuset under studiens gång. Deltagande studiepatienter. Anders Nordgren och Monika Hall, för experimentell hjälp med kirurgi un-der försöken i artikel I och II. Utan er hade det aldrig blivit något! Elisabeth Fredriksson, för allt arbete du lagt ner på att sitta i mörker med de kemiska analyserna. Man kan vara säker på att ens prover är i trygga händer hos dig. Mina kollegor på institutionen för Farmaci. Särskilt till Ulla Wästberg-Galik, Eva Nises-Ahlgren, Pernilla Larsson och Heléne Lyngå för all administrativ hjälp. Till Lucia Lazorova för guidning genom avfall, beställningar och ke-mikaliehantering för en nybliven förvirrad doktorand. Till Lotta Wahlberg för ditt stora hjärta och för stöd med riskbedömning under uppbyggnad av cytostatika-labbet. Till mina doktorandkollegor genom åren. Ni har varit ett stort bidrag till varför det har varit så kul på jobbet (och ibland även efter jobbet). Jag kom-mer sakna er allihop. Jag kommer framförallt minnas otaliga fikaraster, AW, quiz-nights, pre-lab drinks, post-lab drinks, disputationsfester och knallyn. Speciellt kommer jag minnas: Joel Hellrup, min eminenta rumbo, som (näst-an) varje dag har livat upp stämningen med ett sjungande ”god morgon!” Anna Vildhede, för att förlora inte är ett alternativ för dig. Du är en inspirat-ion! Linda Alskär, för din starka personlighet och ditt goda humör. Fortsätt kämpa för oss kvinnor! Sara ”the Queen” Andersson för att man aldrig har tråkigt med dig. Doktorander i biofarmacigruppen (nu som då); Ilse, Emelie, Helena Thörn, David Dahlgren, Carl Roos och Johanna Eriksson. Jag är tacksam för att ni alla har varit en del av min grupp (bästa gruppen). Snart är det min tur att bjuda på lakrits. Ett stort tack till er som hjälpt till med kor-rekturläsning till denna avhandling. Till er andra forna och nuvarande doktorander på BMC; Anna Sommansson, Erika Brolin, Sofia Gustafsson och Ida Netterberg, som har lyst upp min tillvaro med luncher, pyssel eller ett glas vin när det har behövts.

57

Ett STORT tack till vänner och familj som stöttat mig utanför jobbet, jag hade aldrig klarat detta utan er och framför allt, det hade inte varit lika roligt längs vägen:

Nadia Rouhani, Sofie Engberg, Christa Nilsson och Emma Ong Pålsson. Med vänner som er är jag lyckligt lottad! Jag kommer alltid minnas rolighet-er, resor, middagar och mys som jag upplevt med er och jag hoppas på mycket mer. Jag ser fram emot nästa gång! Speciellt tack till Nadia, Christa och Emma för alla gånger ni har öppnat era hem för mig. Vänner från Snerikekören, för att ni har hjälpt mig att tänka på annat än jobb, åtminstone på måndagkvällar och ibland (ofta) för fest eller konsert. I er har jag hittat vänner för livet! Framförallt vill jag tacka hela ALTSO; Lisa Onsbacke, Maria Lindahl, Rebecka Almerén, Veronika Bäckman, Kristin Menkens och Amanda Runsiö – det är så himla roligt att kläcka idéer, arra, sjunga eller bara hänga tillsammans med er! Vilket härligt gäng jag har hittat i er. Jag vill också tacka Camilla Mercurio, Magdalena Reiman, Emma Fredbäck och Emma Rosén för att ni är fantastiska. Mina äldsta vänner; Therèse Greene, Sophie Leitner, Klara Sinder och Malin Hammarbäck, för så många äventyr, så många skratt och så många minnen! Tack för att ni bara finns ett telefonsamtal bort! Viktor Sivertsson för att du alltid har stöttat mig och varit en fin vän.

Mamma och pappa, för att ni alltid finns där, för stort och litet. Jag vet att ni hade varit nöjda med mig vad jag än tagit mig för. Ni har sporrat mig att vilja göra mitt bästa och att nå i mål. Tack också för alla gånger som ni låtit gamm-volvon stå till mitt förfogande, det har underlättat forskningen. Arvid med familj, Gustav och Hanna, och Anders, att umgås med min familj är bland det mest avslappnande jag vet. Tack för alla stunder med er!

David Andersson, för allt stöd under de senaste månaderna. Jag är så lycklig för att du är en del av mitt liv. Du får hela mig att le. Jag älskar dig!

Elsa, Uppsala, November 2015

58

References

1. Aulton, M.E. & Taylor, K. Aulton's pharmaceutics : the design and manufacture of medicines, Edn. 4th. (Churchill Livingstone/Elsevier, Edinburgh; 2013).

2. Chauhan, V.P., Stylianopoulos, T., Boucher, Y. & Jain, R.K. Delivery of molecular and nanoscale medicine to tumors: transport barriers and strategies. Annual review of chemical and biomolecular engineering 2, 281-298 (2011).

3. Jain, R.K. Vascular and interstitial barriers to delivery of therapeutic agents in tumors. Cancer metastasis reviews 9, 253-266 (1990).

4. Yang, C. & Burt, H.M. Drug-eluting stents: factors governing local pharmacokinetics. Advanced drug delivery reviews 58, 402-411 (2006).

5. Doring, B. & Petzinger, E. Phase 0 and phase III transport in various organs: combined concept of phases in xenobiotic transport and metabolism. Drug metabolism reviews 46, 261-282 (2014).

6. Pang, K.S., Rodrigues, A.D. & Peter, R.M. Enzyme- and transporter-based drug-drug Interactions : progress and future challenges. (Springer, New York; 2010).

7. International Transporter, C. et al. Membrane transporters in drug development. Nat Rev Drug Discov 9, 215-236 (2010).

8. Zhang, X.C., Han, L. & Zhao, Y. Thermodynamics of ABC transporters. Protein & cell (2015).

9. Roberts, M.S., Magnusson, B.M., Burczynski, F.J. & Weiss, M. Enterohepatic circulation: physiological, pharmacokinetic and clinical implications. Clinical pharmacokinetics 41, 751-790 (2002).

10. Vollmar, B. & Menger, M.D. The hepatic microcirculation: mechanistic contributions and therapeutic targets in liver injury and repair. Physiological reviews 89, 1269-1339 (2009).

11. Majno, P. et al. Anatomy of the liver: an outline with three levels of complexity--a further step towards tailored territorial liver resections. Journal of hepatology 60, 654-662 (2014).

12. Siriwardena, A.K., Mason, J.M., Mullamitha, S., Hancock, H.C. & Jegatheeswaran, S. Management of colorectal cancer presenting with synchronous liver metastases. Nature reviews. Clinical oncology 11, 446-459 (2014).

13. Godoy, P. et al. Recent advances in 2D and 3D in vitro systems using primary hepatocytes, alternative hepatocyte sources and non-parenchymal liver cells and their use in investigating mechanisms of hepatotoxicity, cell signaling and ADME. Archives of toxicology 87, 1315-1530 (2013).

14. Kmiec, Z. Cooperation of liver cells in health and disease. Advances in anatomy, embryology, and cell biology 161, III-XIII, 1-151 (2001).

15. Hellerbrand, C. Hepatic stellate cells--the pericytes in the liver. Pflugers Archiv : European journal of physiology 465, 775-778 (2013).

59

16. Vildhede, A. in Department of Pharmacy (Uppsala University, Uppsala: Acta Universitatis Upsaliensis; 2015).

17. Rao, R.K. & Samak, G. Bile duct epithelial tight junctions and barrier function. Tissue barriers 1, e25718 (2013).

18. Evans, W.H. A biochemical dissection of the functional polarity of the plasma membrane of the hepatocyte. Biochimica et biophysica acta 604, 27-64 (1980).

19. Hammad, S. et al. Protocols for staining of bile canalicular and sinusoidal networks of human, mouse and pig livers, three-dimensional reconstruction and quantification of tissue microarchitecture by image processing and analysis. Archives of toxicology 88, 1161-1183 (2014).

20. Hanahan, D. & Weinberg, R.A. Hallmarks of cancer: the next generation. Cell 144, 646-674 (2011).

21. Hanahan, D. & Weinberg, R.A. The hallmarks of cancer. Cell 100, 57-70 (2000).

22. Weinberg, R.A. The biology of cancer. (Garland Science, New York; 2007). 23. Carr, B.I. (ed.) Hepatocellular carcinoma: Diagnosis and treatment., Edn. 2nd.

(Humana Press, Philidelphia, USA; 2010). 24. Lord, C.J. & Ashworth, A. The DNA damage response and cancer therapy.

Nature 481, 287-294 (2012). 25. Liu, D.Z. & Ander, B.P. Cell cycle inhibition without disruption of

neurogenesis is a strategy for treatment of aberrant cell cycle diseases: an update. TheScientificWorldJournal 2012, 491737 (2012).

26. Novak, B., Tyson, J.J., Gyorffy, B. & Csikasz-Nagy, A. Irreversible cell-cycle transitions are due to systems-level feedback. Nature cell biology 9, 724-728 (2007).

27. Nikoletopoulou, V., Markaki, M., Palikaras, K. & Tavernarakis, N. Crosstalk between apoptosis, necrosis and autophagy. Biochimica et biophysica acta 1833, 3448-3459 (2013).

28. Snustad, D.P. & Simmons, M.J. Principles of genetics, Edn. 4th. (Wiley, Hoboken, NJ; 2006).

29. Vandenabeele, P., Galluzzi, L., Vanden Berghe, T. & Kroemer, G. Molecular mechanisms of necroptosis: an ordered cellular explosion. Nature reviews. Molecular cell biology 11, 700-714 (2010).

30. Heindryckx, F. & Gerwins, P. Targeting the tumor stroma in hepatocellular carcinoma. World journal of hepatology 7, 165-176 (2015).

31. Fang, J., Nakamura, H. & Maeda, H. The EPR effect: Unique features of tumor blood vessels for drug delivery, factors involved, and limitations and augmentation of the effect. Advanced drug delivery reviews 63, 136-151 (2011).

32. Patel, K.J., Tredan, O. & Tannock, I.F. Distribution of the anticancer drugs doxorubicin, mitoxantrone and topotecan in tumors and normal tissues. Cancer chemotherapy and pharmacology 72, 127-138 (2013).

33. Maeda, H., Tsukigawa, K. & Fang, J. A Retrospective 30 Years After Discovery of the EPR Effect of Solid Tumors: Next-Generation Chemotherapeutics and Photodynamic-therapy-Problems, Solutions, Prospects. Microcirculation (2015).

34. Ferlay, J. et al. Cancer incidence and mortality worldwide: Sources, methods and major patterns in GLOBOCAN 2012. International journal of cancer. Journal international du cancer 136, E359-386 (2015).

35. Ferlay, J. et al., Vol. 2015. (ed. I.A.f.R.o. Cancer) (International Agency for Research on Cancer, http://globocan.iarc.fr; 2013), accessed 21 Oct 2015.

60

36. Jemal, A. et al. Global cancer statistics. CA: a cancer journal for clinicians 61, 69-90 (2011).

37. Schutte, K., Bornschein, J. & Malfertheiner, P. Hepatocellular carcinoma--epidemiological trends and risk factors. Digestive diseases 27, 80-92 (2009).

38. Marengo, A., Rosso, C. & Bugianesi, E. Liver Cancer: Connections with Obesity, Fatty Liver, and Cirrhosis. Annual review of medicine (2015).

39. Yoshimitsu, K. Transarterial chemoembolization using iodized oil for unresectable hepatocellular carcinoma: perspective from multistep hepatocarcinogenesis. Hepatic medicine : evidence and research 6, 89-94 (2014).

40. Tsuboyama, T. et al. Hepatocellular carcinoma: hepatocyte-selective enhancement at gadoxetic acid-enhanced MR imaging--correlation with expression of sinusoidal and canalicular transporters and bile accumulation. Radiology 255, 824-833 (2010).

41. Schuppan, D. & Afdhal, N.H. Liver cirrhosis. Lancet 371, 838-851 (2008). 42. EASL & EORTC EASL-EORTC clinical practice guidelines: management of

hepatocellular carcinoma. Journal of hepatology 56, 908-943 (2012). 43. Marelli, L. et al. Transarterial therapy for hepatocellular carcinoma: which

technique is more effective? A systematic review of cohort and randomized studies. Cardiovascular and interventional radiology 30, 6-25 (2007).

44. Oliveri, R.S., Wetterslev, J. & Gluud, C. Transarterial (chemo)embolisation for unresectable hepatocellular carcinoma. Cochrane database of systematic reviews (2011).

45. Lo, C.M. et al. Randomized controlled trial of transarterial lipiodol chemoembolization for unresectable hepatocellular carcinoma. Hepatology 35, 1164-1171 (2002).

46. Llovet, J.M., Bruix, J. & Grp, B.C.L.C. Systematic review of randomized trials for unresectable hepatocellular carcinoma: Chemoembolization improves survival. Hepatology 37, 429-442 (2003).

47. Llovet, J.M. et al. Arterial embolisation or chemoembolisation versus symptomatic treatment in patients with unresectable hepatocellular carcinoma: a randomised controlled trial. Lancet 359, 1734-1739 (2002).

48. Okusaka, T. et al. Transarterial chemotherapy alone versus transarterial chemoembolization for hepatocellular carcinoma: a randomized phase III trial. Journal of hepatology 51, 1030-1036 (2009).

49. Forner, A. & Trinchet, J.C. Transarterial therapies in HCC: does embolization increase survival? Journal of hepatology 51, 981-983 (2009).

50. Malla, S., Niraula, N.P., Singh, B., Liou, K. & Sohng, J.K. Limitations in doxorubicin production from Streptomyces peucetius. Microbiological research 165, 427-435 (2010).

51. Moffat, A.C. Clarke's analysis of drugs and poisons : in pharmaceuticals, body fluids and postmortem material, Edn. 3rd. (Pharmaceutical Press, London ; Chicago; 2004).

52. Rivory, L.P., Avent, K.M. & Pond, S.M. Effects of lipophilicity and protein binding on the hepatocellular uptake and hepatic disposition of two anthracyclines, doxorubicin and iododoxorubicin. Cancer chemotherapy and pharmacology 38, 439-445 (1996).

53. Tacar, O., Sriamornsak, P. & Dass, C.R. Doxorubicin: an update on anticancer molecular action, toxicity and novel drug delivery systems. The Journal of pharmacy and pharmacology 65, 157-170 (2013).

61

54. Gewirtz, D.A. A critical evaluation of the mechanisms of action proposed for the antitumor effects of the anthracycline antibiotics adriamycin and daunorubicin. Biochemical pharmacology 57, 727-741 (1999).

55. Stravopodis, D.J. et al. Grade-dependent effects on cell cycle progression and apoptosis in response to doxorubicin in human bladder cancer cell lines. International journal of oncology 34, 137-160 (2009).

56. Lee, S.M. et al. Gamma-irradiation and doxorubicin treatment of normal human cells cause cell cycle arrest via different pathways. Molecules and cells 20, 331-338 (2005).

57. Minotti, G., Menna, P., Salvatorelli, E., Cairo, G. & Gianni, L. Anthracyclines: molecular advances and pharmacologic developments in antitumor activity and cardiotoxicity. Pharmacological reviews 56, 185-229 (2004).

58. Jain, D. Cardiotoxicity of doxorubicin and other anthracycline derivatives. J Nucl Cardiol 7, 53-62 (2000).

59. Kim, J.E. et al. The limited intestinal absorption via paracellular pathway is responsible for the low oral bioavailability of doxorubicin. Xenobiotica; the fate of foreign compounds in biological systems 43, 579-591 (2013).

60. Neuhoff, S., Artursson, P., Zamora, I. & Ungell, A.L. Impact of extracellular protein binding on passive and active drug transport across Caco-2 cells. Pharmaceutical research 23, 350-359 (2006).

61. Robert, J., Bui, N.B. & Vrignaud, P. Pharmacokinetics of Doxorubicin in Sarcoma Patients. European journal of clinical pharmacology 31, 695-699 (1987).

62. Mross, K. et al. Pharmacokinetics and metabolism of epidoxorubicin and doxorubicin in humans. Journal of clinical oncology : official journal of the American Society of Clinical Oncology 6, 517-526 (1988).

63. Sugano, K. et al. Coexistence of passive and carrier-mediated processes in drug transport. Nat Rev Drug Discov 9, 597-614 (2010).

64. Danesi, R., Fogli, S., Gennari, A., Conte, P. & Del Tacca, M. Pharmacokinetic-pharmacodynamic relationships of the anthracycline anticancer drugs. Clinical pharmacokinetics 41, 431-444 (2002).

65. Thorn, C.F. et al. Doxorubicin pathways: pharmacodynamics and adverse effects. Pharmacogenetics and genomics 21, 440-446 (2011).

66. Regev, R. & Eytan, G.D. Flip-flop of doxorubicin across erythrocyte and lipid membranes. Biochemical pharmacology 54, 1151-1158 (1997).

67. Okabe, M. et al. Characterization of the organic cation transporter SLC22A16: a doxorubicin importer. Biochemical and biophysical research communications 333, 754-762 (2005).

68. Durmus, S. et al. In vivo disposition of doxorubicin is affected by mouse Oatp1a/1b and human OATP1A/1B transporters. International journal of cancer. Journal international du cancer 135, 1700-1710 (2014).

69. Kassner, N. et al. Carbonyl reductase 1 is a predominant doxorubicin reductase in the human liver. Drug metabolism and disposition: the biological fate of chemicals 36, 2113-2120 (2008).

70. Bains, O.S., Grigliatti, T.A., Reid, R.E. & Riggs, K.W. Naturally occurring variants of human aldo-keto reductases with reduced in vitro metabolism of daunorubicin and doxorubicin. The Journal of pharmacology and experimental therapeutics 335, 533-545 (2010).

71. Bains, O.S. et al. A correlation between cytotoxicity and reductase-mediated metabolism in cell lines treated with doxorubicin and daunorubicin. The Journal of pharmacology and experimental therapeutics 347, 375-387 (2013).

62

72. Xie, Z.B. et al. Systematic review comparing the safety and efficacy of conventional and drug-eluting bead transarterial chemoembolization for inoperable hepatocellular carcinoma. Hepatology research : the official journal of the Japan Society of Hepatology 45, 190-200 (2015).

73. Lammer, J. et al. Prospective randomized study of doxorubicin-eluting-bead embolization in the treatment of hepatocellular carcinoma: results of the PRECISION V study. Cardiovascular and interventional radiology 33, 41-52 (2010).

74. Varela, M. et al. Chemoembolization of hepatocellular carcinoma with drug eluting beads: efficacy and doxorubicin pharmacokinetics. Journal of hepatology 46, 474-481 (2007).

75. Huang, K., Zhou, Q., Wang, R., Cheng, D. & Ma, Y. Doxorubicin-eluting beads versus conventional transarterial chemoembolization for the treatment of hepatocellular carcinoma. Journal of gastroenterology and hepatology 29, 920-925 (2014).

76. Guiu, B. et al. Liver/biliary injuries following chemoembolisation of endocrine tumours and hepatocellular carcinoma: Lipiodol vs. drug-eluting beads. Journal of hepatology 56, 609-617 (2012).

77. Yumoto, Y. et al. Hepatocellular carcinoma detected by iodized oil. Radiology 154, 19-24 (1985).

78. Konno, T. Targeting cancer chemotherapeutic agents by use of lipiodol contrast medium. Cancer 66, 1897-1903 (1990).

79. Konno, T. et al. Effect of arterial administration of high-molecular-weight anticancer agent SMANCS with lipid lymphographic agent on hepatoma: a preliminary report. European journal of cancer & clinical oncology 19, 1053-1065 (1983).

80. Park, C., Choi, S.I., Kim, H., Yoo, H.S. & Lee, Y.B. Distribution of Lipiodol in hepatocellular carcinoma. Liver 10, 72-78 (1990).

81. Iwai, K., Maeda, H. & Konno, T. Use of oily contrast medium for selective drug targeting to tumor: enhanced therapeutic effect and X-ray image. Cancer research 44, 2115-2121 (1984).

82. Maeda, H. Vascular permeability in cancer and infection as related to macromolecular drug delivery, with emphasis on the EPR effect for tumor-selective drug targeting. Proceedings of the Japan Academy. Series B, Physical and biological sciences 88, 53-71 (2012).

83. Furuta, T., Kanematsu, T., Kakizoe, S. & Sugimachi, K. Selective effect of doxorubicin suspended in lipiodol on VX2 carcinoma in rabbits. Journal of surgical oncology 39, 229-234 (1988).

84. Gaba, R.C. et al. Ethiodized oil uptake does not predict doxorubicin drug delivery after chemoembolization in VX2 liver tumors. Journal of vascular and interventional radiology : JVIR 23, 265-273 (2012).

85. www.guerbet.com, Vol. 2015 press release (guerbet group, www.guerbet.com; 2015), accessed in 21/10/2015.

86. http://www.guerbet-us.com/ (2015), accessed in 21/10/2015. 87. Ozcan, M.M. & Atalay, C. Determination of seed and oil properties of some

poppy (Papaver somniferum L.) varieties. Grasas Aceites 57, 169-174 (2006). 88. Idee, J.M. & Guiu, B. Use of Lipiodol as a drug-delivery system for

transcatheter arterial chemoembolization of hepatocellular carcinoma: a review. Critical reviews in oncology/hematology 88, 530-549 (2013).

89. de Baere, T. et al. Circulatory alterations induced by intra-arterial injection of iodized oil and emulsions of iodized oil and doxorubicin: experimental study. Radiology 194, 165-170 (1995).

63

90. Kan, Z., Ivancev, K., Hagerstrand, I., Chuang, V.P. & Lunderquist, A. In vivo microscopy of the liver after injection of Lipiodol into the hepatic artery and portal vein in the rat. Acta radiologica 30, 419-425 (1989).

91. Kan, Z. & Wallace, S. Sinusoidal embolization: impact of iodized oil on hepatic microcirculation. Journal of vascular and inter-ventional radiology : JVIR 5, 881-886 (1994).

92. Kan, Z. et al. Distribution and effect of iodized poppyseed oil in the liver after hepatic artery embolization: experimental study in several animal species. Radiology 186, 861-866 (1993).

93. Kan, Z., Wright, K. & Wallace, S. Ethiodized oil emulsions in hepatic microcirculation: in vivo microscopy in animal models. Academic radiology 4, 275-282 (1997).

94. Bhattacharya, S. et al. Human liver cancer cells and endothelial cells incorporate iodised oil. British journal of cancer 73, 877-881 (1996).

95. Kim, H. et al. Effects of iodinated fatty acid ester on human hepatocellular carcinoma cells. Yonsei medical journal 32, 16-23 (1991).

96. Tarling, E.J., Vallim, T.Q. & Edwards, P.A. Role of ABC transporters in lipid transport and human disease. Trends in endocrinology and metabolism: TEM (2013).

97. Dubbelboer, I.R. et al. Treatment of intermediate stage hepatocellular carcinoma: a review of intrahepatic doxorubicin drug-delivery systems. Therapeutic delivery 5, 447-466 (2014).

98. Lin, S.Y., Wu, W.H. & Lui, W.Y. In vitro release, pharmacokinetic and tissue distribution studies of doxorubicin hydrochloride (Adriamycin HCl) encapsulated in lipiodolized w/o emulsions and w/o/w multiple emulsions. Die Pharmazie 47, 439-443 (1992).

99. Choi, J.W. et al. Comparison of drug release and pharmacokinetics after transarterial chemoembolization using diverse lipiodol emulsions and drug-eluting beads. PloS one 9, e115898 (2014).

100. Boulin, M. et al. Improved stability of lipiodol-drug emulsion for transarterial chemoembolisation of hepatocellular carcinoma results in improved pharmacokinetic profile: Proof of concept using idarubicin. European radiology (2015).

101. Lewis, A.L. et al. Doxorubicin eluting beads - 1: effects of drug loading on bead characteristics and drug distribution. Journal of materials science. Materials in medicine 18, 1691-1699 (2007).

102. http://bead.btg-im.com/, Vol. 2015 (Biocompatibles BTG, 2015), accessed 21 Oct 2015.

103. Lewis, A.L. & Holden, R.R. DC Bead embolic drug-eluting bead: clinical application in the locoregional treatment of tumours. Expert opinion on drug delivery 8, 153-169 (2011).

104. Biondi, M., Fusco, S., Lewis, A.L. & Netti, P.A. Investigation of the mechanisms governing doxorubicin and irinotecan release from drug-eluting beads: mathematical modeling and experimental verification. Journal of materials science. Materials in medicine (2013).

105. Namur, J. et al. Embolization of hepatocellular carcinoma with drug-eluting beads: doxorubicin tissue concentration and distribution in patient liver explants. Journal of hepatology 55, 1332-1338 (2011).

106. Namur, J. et al. Drug-eluting beads for liver embolization: concentration of doxorubicin in tissue and in beads in a pig model. Journal of vascular and interventional radiology : JVIR 21, 259-267 (2010).

64

107. Padia, S.A. et al. Safety and efficacy of drug-eluting bead chemoembolization for hepatocellular carcinoma: comparison of small-versus medium-size particles. Journal of vascular and interventional radiology : JVIR 24, 301-306 (2013).

108. Petri, N. et al. First-pass effects of verapamil on the intestinal absorption and liver disposition of fexofenadine in the porcine model. Drug metabolism and disposition: the biological fate of chemicals 34, 1182-1189 (2006).

109. Helke, K.L. & Swindle, M.M. Animal models of toxicology testing: the role of pigs. Expert opinion on drug metabolism & toxicology 9, 127-139 (2013).

110. Upton, R.N. Organ weights and blood flows of sheep and pig for physiological pharmacokinetic modelling. Journal of pharmacological and toxicological methods 58, 198-205 (2008).

111. Court, F.G. et al. Segmental nature of the porcine liver and its potential as a model for experimental partial hepatectomy. The British journal of surgery 90, 440-444 (2003).

112. Arvidsson, D., Svensson, H. & Haglund, U. Laser-Doppler flowmetry for estimating liver blood flow. The American journal of physiology 254, G471-476 (1988).

113. Matsson, P., Pedersen, J.M., Norinder, U., Bergstrom, C.A. & Artursson, P. Identification of novel specific and general inhibitors of the three major human ATP-binding cassette transporters P-gp, BCRP and MRP2 among registered drugs. Pharmaceutical research 26, 1816-1831 (2009).

114. Knox, C. et al. DrugBank 3.0: a comprehensive resource for 'omics' research on drugs. Nucleic acids research 39, D1035-1041 (2011).

115. Johnson, P.J. et al. Pharmacokinetics and toxicity of intraarterial adriamycin for hepatocellular carcinoma: effect of coadministration of lipiodol. Journal of hepatology 13, 120-127 (1991).

116. Camaggi, C.M. et al. Epirubicin and doxorubicin comparative metabolism and pharmacokinetics. A cross-over study. Cancer chemotherapy and pharmacology 21, 221-228 (1988).

117. Lundahl, A., Hedeland, M., Bondesson, U. & Lennernas, H. In vivo investigation in pigs of intestinal absorption, hepatobiliary disposition, and metabolism of the 5alpha-reductase inhibitor finasteride and the effects of coadministered ketoconazole. Drug metabolism and disposition: the biological fate of chemicals 39, 847-857 (2011).

118. Thorn, H.A., Yasin, M., Dickinson, P.A. & Lennernas, H. Extensive intestinal glucuronidation of raloxifene in vivo in pigs and impact for oral drug delivery. Xenobiotica; the fate of foreign compounds in biological systems 42, 917-928 (2012).

119. Thorn, H.A., Lundahl, A., Schrickx, J.A., Dickinson, P.A. & Lennernas, H. Drug metabolism of CYP3A4, CYP2C9 and CYP2D6 substrates in pigs and humans. European journal of pharmaceutical sciences : official journal of the European Federation for Pharmaceutical Sciences 43, 89-98 (2011).

120. August, D.A., Verma, N., Vaertan, M.A., Shah, R. & Brenner, D.E. An evaluation of hepatic extraction and clearance of doxorubicin. British journal of cancer 72, 65-71 (1995).

121. Garnick, M.B., Ensminger, W.D. & Israel, M. A clinical-pharmacological evaluation of hepatic arterial infusion of adriamycin. Cancer research 39, 4105-4110 (1979).

65

122. Rowland, M. & Tozer, T.N. Clinical pharmacokinetics and pharmacodynamics: concepts and applications, Edn. 4th ed. (Philadelphia : Wolters Kluwer Health/Lippincott William & Wilkins, c2011., United States; 2011).

123. Law, V. et al. DrugBank 4.0: shedding new light on drug metabolism. Nucleic acids research 42, D1091-1097 (2014).

124. Liang, B. et al. Effect of transcatheter intraarterial therapies on the distribution of Doxorubicin in liver cancer in a rabbit model. PloS one 8, e76388 (2013).

125. Kwon, I.K., Lee, S.C., Han, B. & Park, K. Analysis on the current status of targeted drug delivery to tumors. Journal of controlled release : official journal of the Controlled Release Society 164, 108-114 (2012).

126. van Asperen, J., van Tellingen, O., Tijssen, F., Schinkel, A.H. & Beijnen, J.H. Increased accumulation of doxorubicin and doxorubicinol in cardiac tissue of mice lacking mdr1a P-glycoprotein. British journal of cancer 79, 108-113 (1999).

127. van Asperen, J., van Tellingen, O. & Beijnen, J.H. The role of mdr1a P-glycoprotein in the biliary and intestinal secretion of doxorubicin and vinblastine in mice. Drug metabolism and disposition: the biological fate of chemicals 28, 264-267 (2000).

128. Booth, C.L., Brouwer, K.R. & Brouwer, K.L. Effect of multidrug resistance modulators on the hepatobiliary disposition of doxorubicin in the isolated perfused rat liver. Cancer research 58, 3641-3648 (1998).

129. Vlaming, M.L. et al. Carcinogen and anticancer drug transport by Mrp2 in vivo: studies using Mrp2 (Abcc2) knockout mice. The Journal of pharmacology and experimental therapeutics 318, 319-327 (2006).

130. Preissner, S. et al. SuperCYP: a comprehensive database on Cytochrome P450 enzymes including a tool for analysis of CYP-drug interactions. Nucleic acids research 38, D237-243 (2010).

131. Watanabe, T., Kusuhara, H., Maeda, K., Shitara, Y. & Sugiyama, Y. Physiologically based pharmacokinetic modeling to predict transporter-mediated clearance and distribution of pravastatin in humans. The Journal of pharmacology and experimental therapeutics 328, 652-662 (2009).

132. Bergman, E. et al. Enterohepatic disposition of rosuvastatin in pigs and the impact of concomitant dosing with cyclosporine and gemfibrozil. Drug metabolism and disposition: the biological fate of chemicals 37, 2349-2358 (2009).

133. Wu, C. et al. BioGPS: an extensible and customizable portal for querying and organizing gene annotation resources. Genome biology 10, R130 (2009).

134. Wang, P.J. et al. Biomechanical study of hepatic portal vein in humans and pigs and its value in liver transplantation. Transplantation proceedings 41, 1906-1910 (2009).

135. Yamashita, Y.T., T.; Fukushima, S.; Nishida, M.; Nakano, M. Experimental Study of Hepatic Artery Administration of Various Formulations of Lipiodol with Anticancer Agents. Kumamoto Med. J. 40, 53-65 (1987).

136. Lewis, A.L. et al. DC bead: in vitro characterization of a drug-delivery device for transarterial chemoembolization. Journal of vascular and interventional radiology : JVIR 17, 335-342 (2006).

137. Bjerregaard, S., Soderberg, I., Vermehren, C. & Frokjaer, S. Formulation and evaluation of release and swelling mechanism of a water-in-oil emulsion using factorial design. International journal of pharmaceutics 193, 1-11 (1999).

66

138. Shin, S.W. The current practice of transarterial chemoembolization for the treatment of hepatocellular carcinoma. Korean journal of radiology : official journal of the Korean Radiological Society 10, 425-434 (2009).

139. Lewis, A.L. et al. Pharmacokinetic and safety study of doxorubicin-eluting beads in a porcine model of hepatic arterial embolization. Journal of vascular and interventional radiology : JVIR 17, 1335-1343 (2006).

140. Gonzalez, M.V. et al. Doxorubicin eluting beads-2: methods for evaluating drug elution and in-vitro:in-vivo correlation. Journal of materials science. Materials in medicine 19, 767-775 (2008).

141. Biondi, M., Fusco, S., Lewis, A.L. & Netti, P.A. New insights into the mechanisms of the interactions between doxorubicin and the ion-exchange hydrogel DC Bead for use in transarterial chemoembolization (TACE). Journal of biomaterials science. Polymer edition 23, 333-354 (2012).

142. Chen, H. et al. Expression of P450 and nuclear receptors in normal and end-stage Chinese livers. World journal of gastroenterology : WJG 20, 8681-8690 (2014).

143. Malagari, K. et al. Prospective randomized comparison of chemoembolization with doxorubicin-eluting beads and bland embolization with BeadBlock for hepatocellular carcinoma. Cardiovascular and interventional radiology 33, 541-551 (2010).

144. Koh, D.M. & Collins, D.J. Diffusion-weighted MRI in the body: applications and challenges in oncology. AJR. American journal of roentgenology 188, 1622-1635 (2007).

145. Lencioni, R. & Llovet, J.M. Modified RECIST (mRECIST) assessment for hepatocellular carcinoma. Seminars in liver disease 30, 52-60 (2010).

67

Acta Universitatis UpsaliensisDigital Comprehensive Summaries of Uppsala Dissertationsfrom the Faculty of Pharmacy 207

Editor: The Dean of the Faculty of Pharmacy

A doctoral dissertation from the Faculty of Pharmacy, UppsalaUniversity, is usually a summary of a number of papers. A fewcopies of the complete dissertation are kept at major Swedishresearch libraries, while the summary alone is distributedinternationally through the series Digital ComprehensiveSummaries of Uppsala Dissertations from the Faculty ofPharmacy. (Prior to January, 2005, the series was publishedunder the title “Comprehensive Summaries of UppsalaDissertations from the Faculty of Pharmacy”.)

Distribution: publications.uu.seurn:nbn:se:uu:diva-267396

ACTAUNIVERSITATIS

UPSALIENSISUPPSALA

2015