Cellular, placental and in vivo pharmacokinetics of liposomal doxorubicin

ISSERTATIONS | SUVI KAROLIINA SOININEN | CELLULAR, PLACENTAL AND IN VIVO PHARMACOKINETICS... | No 365

uef.fi

PUBLICATIONS OF

THE UNIVERSITY OF EASTERN FINLAND Dissertations in Health Sciences

ISBN 978-952-61-2210-6 ISSN 1798-5706

Dissertations in Health Sciences

SUVI KAROLIINA SOININEN

CELLULAR, PLACENTAL AND IN VIVO PHARMACOKINETICS OF LIPOSOMAL DOXORUBICIN

SUVI KAROLIINA SOININEN

Systemic exposure to anticancer drugs may cause severe side-effects in healthy tissues, even though the drug concentra- tions remain at a low level in the tumor tissue. Nanoparticulate drug delivery systems can be one way to improve the drug efficacy and diminish the side-effects.

This thesis clarified the pharmacokinetics (PK) of anticancer drug, doxorubicin, and its passively and actively targeted nano- particulate formulations in cells in vitro, in tissues in vivo, and in human placenta ex vivo. Furthermore, cellular PK and drug

cytotoxicity were successfully characte- rized by computer-aided modeling.

SUVI KAROLIINA SOININEN

Cellular, Placental and in vivo

Pharmacokinetics of Liposomal Doxorubicin

To be presented by permission of the Faculty of Health Sciences, University of Eastern Finland for public examination in lecture hall Canthia 102, Kuopio, on Friday, September 9^th 2016, at 12 noon

Publications of the University of Eastern Finland Dissertations in Health Sciences

Number 365

School of Pharmacy Faculty of Health Sciences University of Eastern Finland

Kuopio 2016

Juvenes Print Tampere, 2016

Series Editors:

Professor Tomi Laitinen, M.D., Ph.D.

Institute of Clinical Medicine, Clinical Physiology and Nuclear Medicine Faculty of Health Sciences

Professor Hannele Turunen, Ph.D.

Department of Nursing Science Faculty of Health Sciences Professor Kai Kaarniranta, M.D., Ph.D.

Institute of Clinical Medicine, Ophthalmology Faculty of Health Sciences

Associate Professor (Tenure Track) Tarja Malm, Ph.D.

A.I. Virtanen Institute for Molecular Sciences Faculty of Health Sciences

Lecturer Veli-Pekka Ranta, Ph.D. (pharmacy) School of Pharmacy

Faculty of Health Sciences Distributor:

University of Eastern Finland Kuopio Campus Library

P.O.Box 1627 FI-70211 Kuopio, Finland http://www.uef.fi/kirjasto ISBN (print):978-952-61-2210-6

ISBN (pdf):978-952-61-2211-3 ISSN (print):1798-5706

ISSN (pdf):1798-5714 ISSN-L:1798-5706

Author’s address: School of Pharmacy Faculty of Health Sciences University of Eastern Finland KUOPIO

FINLAND

Supervisors: Docent Marika Ruponen, Ph.D.

School of Pharmacy Faculty of Health Sciences University of Eastern Finland KUOPIO

FINLAND

Professor Arto Urtti, Ph.D.

School of Pharmacy Faculty of Health Sciences University of Eastern Finland KUOPIO

FINLAND

Kati-Sisko Vellonen, Ph.D.

School of Pharmacy Faculty of Health Sciences University of Eastern Finland KUOPIO

FINLAND

Reviewers: Associate Professor Enrico Mastrobattista, Ph.D.

Department of Pharmaceutics

Utrecht Institute of Pharmaceutical Sciences Utrecht University

UTRECHT

THE NETHERLANDS

Assistant Professor Katrien Remaut, Ph.D.

Department of Pharmaceutical Sciences Ghent University

GHENT BELGIUM

Opponent: Professor Jessica Rosenholm, D. Sc. (Tech.) Pharmaceutical Sciences Laboratory Faculty of Science and Engineering Åbo Akademi University

TURKU FINLAND

Soininen, Suvi Karoliina

Cellular, Placental and in vivo Pharmacokinetics of Liposomal Doxorubicin University of Eastern Finland, Faculty of Health Sciences

Publications of the University of Eastern Finland. Dissertations in Health Sciences 365. 2016. 59 p.

ISBN (print): 978-952-61-2210-6 ISBN (pdf): 978-952-61-2211-3 ISSN (print): 1798-5706 ISSN (pdf): 1798-5714 ISSN-L: 1798-5706

ABSTRACT

Cancer is one of the major global causes of death. Generally, cancer treatment involves a combination of surgery, radiotherapy and chemotherapy. However, chemotherapy with conventional anticancer drugs has many drawbacks, one of them being the narrow therapeutic index. The systemic exposure to the drug may cause severe side-effects in healthy tissues, even though the drug concentrations remain low in the tumor tissue. Nanoparticulate drug delivery systems provide a tool to increase the drug concentrations in the target tissue and diminish the concentrations in heathy tissues. Despite the potential benefits in drug therapy, there are still many challenges to be overcome in drug targeting and development of nanoparticulate drug delivery systems. Clearly, further understanding of the kinetics of nanoparticulate systems at both the cellular and tissue levels is needed to aid in the more rational design of novel drug delivery systems.

The aim of this study was to investigate the pharmacokinetic behavior of the anticancer drug, doxorubicin, and its passively and actively targeted formulations in cells in vitro, in tissues in vivo, and in human placenta ex vivo. An avidin fusion protein expression system was used for the active targeting of biotinylated doxorubicin formulations in vitro and in vivo. Concentrations of free and liposomal doxorubicin were measured by mass spectrometry in whole cells and nuclei (target organelle), and the cytotoxicity of the formulations was evaluated. The obtained data was used to create an in silico pharmacokinetic-pharmacodynamic (PK-PD) model depicting the nuclear concentration – effect relationships.

Our studies revealed that free doxorubicin accumulated strongly in cells and their nuclei in a cell concentration dependent manner. The concentration of doxorubicin in nuclei was cell line dependent and did not necessarily correlate with the response, i.e. cytotoxicity. Free doxorubicin accumulated also in human placenta but only low levels of doxorubicin were measured in the fetal circulation ex vivo. Interestingly, the kinetics of doxorubicin delivered in pH-sensitive liposomes resembled that of the free drug in cells and in human placenta. Surface pegylation hindered the cellular uptake of liposomes and subsequently reduced the cytotoxicity of liposomal doxorubicin, although all the cell associated doxorubicin was found in nuclei. Surface pegylation prevented also the penetration of liposomes through human placenta. In active targeting, the limited cytotoxicity of the biotinylated doxorubicin conjugate was attributable to its entrapment in endo-lysosomes.

Biotinylated pH-sensitive doxorubicin liposomes restored the cytotoxic activity of the drug.

However, both biotinylated formulations used specific and unspecific cell uptake pathways.

Biotinylation did not change the pharmacokinetics (i.e. circulation time and tumor tissue uptake) of pH-sensitive liposomes in vivo. The transit-compartment PK-PD model described well the nuclear concentration – effect relationships of free and liposomal doxorubicin. This study provided new information about the cellular, placental and in vivo pharmacokinetics of nanoparticulate drug delivery systems.

National Library of Medicine Classification: QS 532, QU 93, QV 38, QV 269, WQ 212

Medical Subject Headings: Neoplasms/Drug Therapy; Antineoplastic Agents; Doxorubicin; Nanoparticles;

Liposomes; Pharmacokinetics; Pharmacology; Intracellular Space; Placenta; Perfusion; Drug Delivery Systems

Soininen, Suvi Karoliina

Liposomaalisen doksorubisiinin solu-, istukka- ja in vivo farmakokinetiikka Itä-Suomen yliopisto, terveystieteiden tiedekunta

Publications of the University of Eastern Finland. Dissertations in Health Sciences 365. 2016. 59 s.

ISBN (nid.): 978-952-61-2210-6 ISBN (pdf): 978-952-61-2211-3 ISSN (nid.): 1798-5706 ISSN (pdf): 1798-5714 ISSN-L: 1798-5706

TIIVISTELMÄ

Syöpä on yksi merkittävimmistä syistä kuolleisuuteen maailmanlaajuisesti. Yleisesti syövän hoidossa käytetään leikkaushoitoa, radioterapiaa ja kemoterapiaa. Perinteisillä syöpälääkkeillä on kuitenkin kapea terapeuttinen leveys, minkä vuoksi lääkeaine voi aiheuttaa vakavia haittavaikutuksia terveissä kudoksissa, vaikka lääkeainekonsentraatiot syöpäkudoksessa pysyvät alhaisina. Käyttämällä nanohiukkasia lääkeaineen kantajina lääkekonsentraatiota voidaan kuitenkin nostaa kohdekudoksessa ja samalla sitä voidaan alentaa terveissä kudoksissa. Huolimatta nanohiukkasten hyödyistä lääkehoidoissa, niiden kohdentaminen on haastavaa ja laajamittainen käyttö vaatii lisätutkimuksia.

Tämän työn tarkoituksena oli tutkia syöpälääke doksorubisiinin ja sen passiivisesti ja aktiivisesti kohdennettujen formulaatioiden farmakokinetiikkaa soluissa in vitro, kudoksissa in vivo ja ihmisen istukassa ex vivo. Biotinyloitujen aktiivisesti kohdennettujen doksorubisiiniformulaatioiden kohdentamiseen käytettiin avidiinifuusioproteiinisysteemiä in vitro ja in vivo. Vapaan ja liposomaalisen doksorubisiinin konsentraatiot määritettiin massaspektrometrilla kokonaisista soluista ja solujen tumista, jotka ovat kohdesoluelimiä doksorubisiinille. Lisäksi formulaatioiden aiheuttama solutoksisuus mitattiin. Kerättyjä tuloksia käytettiin farmakokinetiikka-farmakodynamiikka (PK-PD) -mallissa kuvaamaan lääkeaineen tumakonsentraatio–vaste -suhdetta in silico.

Tutkimuksemme osoittivat, että vapaa doksorubisiini kertyi voimakkaasti soluihin ja niiden tumiin. Tuman doksorubisiinikonsentraatio riippui solun konsentraatiosta ja käytetystä solulinjasta. Doksorubisiinin tumakonsentraatio ei kuitenkaan välttämättä korreloinut sen aiheuttamaan solutoksisuuteen. Doksorubisiini kertyi myös ihmisen istukkaan ja vain pieni määrä lääkeainetta mitattiin sikiön verenkierrosta ex vivo. Kun doksorubisiini vietiin soluihin ja ihmisen istukkaan pH-sensitiivisten liposomien avulla, lääkeaineen farmakokinetiikka muistutti vapaana vietävän lääkeaineen farmakokinetiikkaa. Kun doksorubisiinin kantajina käytettiin pegyloituja liposomeja, lääkeaineen soluunotto ja pääsy ihmisen istukkakudokseen heikkenivät. Tämä selitti myös pegyloidun formulaation aiheuttaman heikon solutoksisuuden. Doksorubisiini kertyi kuitenkin yllättäen solujen tumiin, kun käytettiin pegyloituja liposomeja lääkeaineen kantajina. Aktiivisesti kohdennettu biotinyloitu doksorubisiinikonjugaatti kertyi solun endo-lysosomeihin, mikä saattoi aiheuttaa konjugaatin heikon solutoksisuuden.

Doksorubisiinin solutoksisuus voitiin kuitenkin säilyttää käyttämällä biotinyloituja pH- sensitivisiä liposomeja lääkeaineen kantajina. Biotinyloidut formulaatiot käyttivät soluunottoon spesifejä ja epäspesifejä mekanismeja. Biotinylointi ei muuttanut pH- sensitiivisten liposomien farmakokinetiikkaa (kiertoaikaa ja kertymistä syöpäkudokseen) verrattuna kontrolliin in vivo. Tietokoneavusteinen PK-PD -malli toimi luotettavasti. Työ toi uutta tietoa nanohiukkasten solu-, istukka- ja in vivo farmakokinetiikasta.

Luokitus: QS 532, QU 93, QV 38, QV 269, WQ 212

Yleinen suomalainen asiasanasto: syöpätaudit; lääkehoito; lääkkeet; lääkeaineet; nanohiukkaset;

farmakokinetiikka; farmakologia; solut; kudokset; istukka

Acknowledgements

This study was carried out in the School of Pharmacy, Faculty of Health Sciences, University of Eastern Finland, Kuopio campus during the years 2010–2016, and in the year 2009 in the Department of Pharmaceutics, University of Kuopio.

I wish to express my deepest gratitude and appreciation to the following people since this thesis could not have been accomplished without them.

First I would most sincerely like to thank my supervisor Docent Marika Ruponen (PhD Pharm) for her continuous support and help in these studies. She never counted the hours or looked at her watch when offering advice about the work. Her tenacious guidance and persistence helped me to grow as a scientist.

I wish to thank the professors and senior scientists providing supervision in different stages of my studies: Prof. Jukka Mönkkönen, Prof. Arto Urtti, Zanna Hyvönen (PhD) and Kati- Sisko Vellonen (PhD Pharm). I especially wish to thank Prof. Arto Urtti for his open attitude towards my research, and for his excellent ideas about the third publication. In general, I wish to thank him for his wisdom. I also wish to thank Zanna Hyvönen (PhD) and Kati-Sisko Vellonen (PhD Pharm) for their insightful comments and encouraging words during the times of uncertainty.

I wish to express my appreciation to the pre-examiners of this thesis, Associate Prof. Enrico Mastrobattista (PhD) from Utrecht University, The Netherlands,and Assistant Prof. Katrien Remaut (PhD) from Ghent University, Belgium. Their excellent suggestions have significantly improved this thesis. I am honored that Prof. Jessica Rosenholm [DSc (Tech.)]

from Åbo Academi University, Finland, has agreed to be the opponent in the public examination of this thesis.

I wish to extend my gratitude to Prof. Kirsi Vähäkangas with whom I had the opportunity to collaborate during the studies. She made me think a lot about embarking on an academic career. I will never forget our fruitful discussions. I also want to thank Prof. Markus Forsberg for his support in my carreer decisions. I enjoyed tremendously working in his projects, and I much admire his contagious enthusiasm towards pharmaceutical sciences. In addition, I wish to thank Prof. Seppo Ylä-Herttuala who gave me an opportunity to work in his research group during my Master’s thesis project and in a collaboration included in my PhD project. He allowed me to have access to his laboratory facilities whenever I needed them. Futhermore, I want to express my gratitude to Prof. Seppo Auriola who has followed my PhD project all the way from its very beginning. The scientific advice he gave to me helped tremendously. After getting know him, it is hard to imagine any better scientist than him. Last but not least, I wish to thank Veli-Pekka Ranta (PhD Pharm). His encouraging words during my PhD studies and ingenuity in computer based modeling helped me to complete this PhD thesis.

I wish to thank also Pauliina Lehtolainen-Dalkilic (PhD), Hanna Lesch (PhD), Minna Kaikkonen (PhD), Thomas Wirth (PhD), Ann-Marie Määttä (PhD), Prof. David Selwood (PhD), Galina Wirth (MSc), Vesa Karttunen (PhD Pharm), Jenni Repo (MSc Pharm), Pyry Välitalo (PhD Pharm) and Aki Heikkinen (PhD Pharm) for their major contribution to this thesis work. In particular, I would like to thank Jenni Repo (MSc Pharm) with whom we have shared times of despair and success, but mainly the latter. She has been a true inspiration for me with her excellent ideas and her lovely smile.

I wish to extend my appreciation to all of my fellow scientists in the University of Eastern Finland and in the University of Helsinki. Some of them are mentioned here; Ari Kauppinen (PhD), Björn Peters (PhD Pharm), Tuomas Ervasti [D.Sc. (Tech.)], Muluneh Fashe (PhD Pharm), Hristo Zlatev (MD), Sofia Sousa (PhD Pharm), Jenni Hakkarainen (PhD), Lucia Podracka (MSc), Eva Ramsay (MSc Pharm), Ana Cecilia A X De-Oliveira (MD, PhD), Emmanuelle Morin-Picardat (PhD) and Polina Ilina (PhD). It has been pleasure to work with all of you. My special thanks goes to my colleagues Sami Poutiainen (PhD) and Riikka Laitinen (PhD) who kept believing in me. In their presence, I have been able to show my true colors. I wish also to thank Maiju Järvinen (MSc Pharm) and Lucia Podracka (MSc) for their loyal friendship. I further wish to thank Ewen MacDonald (PhD) for revising the English language of this thesis, and the administrative and technical personnel of the School of Pharmacy, University of Eastern Finland.

Finally, I would like to thank my family and relatives. I have shared countless discussions with my beloved aunt Doctor Kaija Huttunen about the present status and the future of the national healthcare system. I cannot thank her enough for her endless interest towards my research studies. I wish also to thank my mother- and father-in-law for their warm support.

Especially I wish to thank my parents and sister and her family for their constant love and support. In addition, I thank my grandmother for reminding of how far we have come. I express my deepest appreciation to my husband for his patience and love.

This thesis work was financially supported by the University of Eastern Finland, the Finnish Funding Agency for Innovation and the European Regional Development Fund.

All the funders are gratefully acknowledged.

Kuopio, August 2016

Karoliina Soininen

List of the original publications

This dissertation is based on the following original publications (I–III):

I Soininen SK, Lehtolainen-Dalkilic P, Karppinen T, Puustinen T, Dragneva G, Kaikkonen MU, Jauhiainen M, Allart B, Selwood DL, Wirth T, Lesch HP, Määttä AM, Mönkkönen J, Ylä-Herttuala S, Ruponen M. Targeted delivery via avidin fusion protein: intracellular fate of biotinylated doxorubicin derivative and cellular uptake kinetics and biodistribution of biotinylated liposomes. European Journal of Pharmaceutical Sciences 47: 848-856, 2012.

II Soininen SK*, Repo JK*, Karttunen V, Auriola S, Vähäkangas KH, Ruponen M.

Human placental cell and tissue uptake of doxorubicin and its liposomal formulations. Toxicology Letters239: 108-114, 2015.

III Soininen SK, Vellonen K-S, Heikkinen AT, Auriola S, Ranta V-P, Urtti A,

Ruponen M. Intracellular PK/PD relationships of free and liposomal doxorubicin:

quantitative analyses and PK/PD modeling. Molecular Pharmaceutics 13: 1358- 1365, 2016.

*Equal contribution.

The publications have been adapted with the permission of the copyright owners.

Contents

1 INTRODUCTION ... 1

2 REVIEW OF THE LITERATURE ... 2

2.1

Nanoparticles in cancer therapy ...

2.1.1 Passive targeting... 4

2.1.2 Active targeting ... 6

2.2

Intracellular kinetics of nanoparticles ...

2.2.1 Cell adhesion ... 9

2.2.2 Cell internalization ... 10

2.2.3 Intracellular distribution ... 11

2.2.4 Methodology to study the intracellular distribution of drugs and nanoparticles ... 11

2.3

Nanoparticles and human placenta ...

3 AIMS OF THE STUDY ... 19

4 OVERVIEW OF THE MATERIALS AND METHODS ... 21

4.1

Cell lines and a tumor model ...

4.2

Drugs and formulations ...

4.3

Studies on drug uptake and distribution ...

4.3.1 Qualitative analysis ... 22

4.3.2 Quantitative analyses ... 22

4.4

Cytotoxicity analysis ...

4.5

Intracellular pharmacokinetics-pharmacodynamics model ...

5 RESULTS AND DISCUSSION ... 27

5.1

Cell uptake of doxorubicin and its formulations ...

5.2

Intracellular distribution of doxorubicin and its formulations ...

5.3

Human placental tissue uptake and transplacental transfer of doxorubicin and its liposomal formulations ...

5.4

Biodistribution of doxorubicin and its liposomal formulations in vivo ....

5.5

Cytotoxicity of doxorubicin and its formulations ...

6 CONCLUSIONS ... 35

7 SUMMARY AND FUTURE PERSPECTIVES ... 36

8 REFERENCES ... 38 APPENDICES

Abbreviations

α^vβ³ Integrin

α^vβ⁵ Integrin

τ Mean transit time of the response

ABC Accelerated blood clearance

ACUPA S,S-2-[3-(5-amino-1-carboxypentyl)-ureido]-pentanedioic acid AD-PEG-hTf Adamantine-polyethylene glycol conjugated to human transferrin AIDS Acquired immune deficiency syndrome

AUC Area under the curve

B-DOX Biotinylated doxorubicin conjugate

BeWo Human choriocarcinoma

BL-DOX Biotinylated pH-sensitive liposomal doxorubicin

BT4C Rat glioma

BT4CAvidinFusionProtein Rat glioma expressing the avidin fusion protein Caki-2 Human kidney clear cell carcinoma

CD Cluster of differentiation

CE Capillary electrophoresis

CHEMS Cholesteryl hemisuccinate

Chol Cholesterol

C^max Peak concentration

C^threshold Threshold drug concentration in the nucleus to elicit any response

DAPI 4',6-diamidino-2-phenylindole

DOPE 1,2-dioleoyl-sn-glycero-3-phosphoethanolamine DSPE-PEG²⁰⁰⁰ 1,2-diastearoyl-sn-glycero-3-phosphoethanolamine-N-

[methoxy(polyethylene glycol)-2000]

DSPE-PEG²⁰⁰⁰ Biotin 1,2-diastearoyl-sn-glycero-3-phosphoethanolamine-N- [biotinyl(polyethylene glycol)-2000]

EC⁵⁰ Drug concentration to elicit 50% of maximum response EGFR Epidermal growth factor receptor

EMA European Medicines Agency

EPR Enhanced permeation and retention

ER Endoplasmic reticulum

F Fetal

FDA US Food and Drug Administration

GRAF-1 Guanosine 5'-triphosphatase regulator associated with focal adhesion kinase 1

HER2 Human epidermal growth factor receptor 2 HPLC High performance liquid chromatography

HPLC-MS/MS High performance liquid chromatography tandem mass spectrometry

HSPC Hydrogenated soy phosphatidylcholine IC⁵⁰ Half maximal inhibitory concentration

Ig Immunoglobulin

i.v. Intravenous

K Response as the initial cell kill rate constant

K^max The maximum initial response as the cell kill rate constant LC-MS Liquid chromatography mass spectrometry

LDLR Low-density lipoprotein receptor L-DOX pH-sensitive liposomal doxorubicin

M Maternal

mAb Monoclonal antibody

MPS Mononuclear phagocyte system

MW Molecular weight

ND Not determined

NP Nanoparticle

p53 Tumor antigen

PBS Phosphate buffered saline

PEG Polyethylene glycol

PK Pharmacokinetics

PK-PD Pharmacokinetic-Pharmacodynamic

PD Pharmacodynamics

PL-DOX Pegylated liposomal doxorubicin PSMA Prostate-specific membrane antigen

QSAR Quantitative structure-activity relationship

RB94 Retinoblastoma 94

RGD Arginine-Glycine-Asparagine

RRM2 Ribonucleotide reductase M2

scFv Single chain variable domain fragments

t^1/2 Half-life

TPP Thiamine pyrophosphate

TR Transferrin receptor

U-118 MG Human glioblastoma-astrocytoma U-87 MG Human glioblastoma-astrocytoma

V Cell viability

VEGF Vascular endothelial growth factor receptor

VH Variable heavy

WHO World Health Organization

1 Introduction

Cancer is one of the major global causes of death. This is largely attributable to the aging and growth of the world´s population. According to the World Health Organization (WHO), in 2012, there were 8.2 million cancer related deaths, whichcorresponded to 13% of all deaths. Our understanding on cancer diagnostics, treatments and prevention has significantly improved during the last five decades. However, in global terms the incidences of cancer and cancer mortality are rising, even though in the developed countries the cancer mortality rate has declined. The WHO predicts that there will be an annual increase in new cancer cases from 14 million in 2012 to 22 million within the next two decades.

The therapeutic index of conventional anticancer drugs is usually narrow. The high systemic concentrations of the drug evoke dose-limiting toxicity, but nonetheless the drug dose reaching the tumor tissue may remain at a low level (Lammers et al., 2012). With targeted nanoparticulate drug delivery systems, the drug can be selectively delivered into the solid tumor tissue. This is due to the prolonged circulation time of nanoparticles in blood and their enhanced accumulation into the tumor tissue through the fenestrated walls of the tumor´s vasculature. The lack of lymphatic drainage in tumor tissue also improves the storage of nanoparticles in the tissue for long time. This has been referred to as the enhanced permeation and retention (EPR) effect (Maeda, 2001; Matsumura and Maeda, 1986). In addition, while nanoparticles may enhance the efficacy of the drug, they can reduce the drug concentrations in the healthy tissues, and subsequently, some of the severe side-effects of the drug. The properties of nanoparticles can be further modified e.g. by biodegradable materials to promote the controlled drug release at the tumor tissue, or by targeting ligands to enhance the cancer cell uptake.

There has been a huge investment of time and funds for the preclinical development of new nanomedicines. For instance, the National Nanotechnology Initiative in the United States budgeted $1.5 billion towards nanotechnology for the fiscal year 2016. However, only a low number of applications have reached the clinical phase, and even fewer have been clinically approved. This might be due to many issues that have limited the introduction of nanoparticles into oncology. For example, the industrial production, including scalability, reproducibility from batch to batch, and cost and quality control, may be challenging. The production of complex nanoparticle structures is particularly demanding and, therefore, simple formulations are preferable. In addition, the stability of formulation has to be well controlled during their storage as well as when they are in the biological environment.

Nanoparticles may also induce immunogenicity (Sharma et al., 2012; Szebeni et al., 2011).

For instance, the activation of the complement system, allergic-like reactions or other adverse reactions have been observed in some patients (Moghimi and Farhangrazi, 2013;

Moghimi, 2014). Furthermore, the translation of the results from bench to bedside has been challenging, e.g. the efficacy of the nanoparticulate systems might not be adequate even though the severe side-effects may be diminished. In addition, the regulatory guidelines are still nonstandardized (Ehmann et al., 2013). Therefore, a better understanding of the pharmacokinetics of nanoparticulate systems at both the cellular and tissue levels would possibly improve the rational design of nanomedicines and speed up the development of these drugs.

In this thesis work, doxorubicin loaded nanoparticulate drug delivery systems for passive and active targeting were developed. Their pharmacokinetics were evaluated in cells in vitro, in tissues in vivo and in human placenta ex vivo. Finally, the pharmacokinetic- pharmacodynamic (PK-PD) relationships of the liposomal formulations were clarified at the target organelle level in silico.

2 Review of the Literature

2.1 NANOPARTICLES IN CANCER THERAPY

Nanostructures. In the area of pharmaceutics, nanoparticles are typically 1 to 1000 nm in size (figure 1). They can be obtained by three different processes; self-assembly (Busseron et al., 2013), nucleation (Thanh et al., 2014) or condensation (Birringer et al., 1984) from one or more components of organic or inorganic material at the atomic bond or molecular levels.

Drugs may be incorporated within the nanoparticles in several ways, such as by encapsulation or by conjugation with the nanoparticle. For example, in matrix systems (e.g.

dendrimers) the drug is dispersed throughout the particle structure while in nanocapsules (e.g. liposomes) the drug is encapsulated in an aqueous or oily cavity surrounded by a single polymeric membrane (Prabhu et al., 2015). In drug-nanoparticle conjugates, a covalent conjugation chemistry with biodegradable linkers is commonly used.

Figure 1. Schematic illustration of nanoparticles. (A) polymeric nanocapsules (Ø 50–200 nm), (B) polymeric micelles (Ø 5–100 nm) and (C) polymeric dendrimers (Ø 1–200 nm), (D) liposomes (Ø 50–200 nm) and inorganic (E) metallic nanoparticles (Ø 2–40 nm) and (F) carbon nanotubes (Ø 1–2 nm, length 1–30 µm). Low molecular weight drug (pink) is (A) retained inside the nanocapsule, (C, E and F) covalently conjugated with the nanoparticle, (B, C, D and F) encapsulated in the nanoparticle or (D) lipid bilayer -embedded.

Liposomes. Liposomes are spherical vesicles consisting of at least one phospholipid bilayer enclosing an aqueous compartment (Bangham et al., 1965). The phospholipids can be either saturated or unsaturated, and cholesterol is commonly incorporated into the lipid bilayer to increase its mechanical strength. Liposomes are classified into small (< 100 nm in diameter) and large (100 to 800 nm) unilamellar vesicles, and multilamellar vesicles (> 500 nm) depending on their size and the number of lipid bilayers (Ulrich, 2002). Liposomes can carry hydrophilic drugs which are encapsulated in the aqueous interior or lipophilic drugs incorporated into the lipid bilayer. Liposomes are considered to be biocompatible, biodegradable and relatively non-toxic (Çağdaş et al., 2014).

Liposomes can be prepared by different methods, such as the thin-film hydration method (Bangham technique), sonication, extrusion, reverse phase evaporation or freeze- thawing (Akbarzadeh et al., 2013; Szoka and Papahadjopoulos, 1978). Drugs can be loaded into them either passively while they are being produced or actively after their preparation.

For instance, the active loading methods can utilize a pH-gradient or an ammonium sulfate gradient across the lipid bilayer (Haran et al., 1993; Mayer et al., 1986).

Interactions of nanomaterials in the body. Nanoparticles can be administered in patients by many routes, e.g. intravenously (i.v.), intramuscularly, subcutaneously or per oral. The i.v.

administration is the most preferable route since nanoparticles do not have good vascular bioavailability in the bloodstream after extravascular delivery, for instance after oral or intramuscular administration. After i.v. injection, there is a rapid blood clearance of nanoparticles typically encountered. This is due to plasma protein binding on the surface of nanoparticles, in a phenomenon called opsonization. The proteins, or protein corona, on the nanoparticle surface allows the cells of the mononuclear phagocyte system (MPS) (or the reticulo endothelial system) to recognize the nanoparticles, and remove them rapidly from the blood circulation by phagocytosis and by other endocytic uptake routes. The phago/endocytosed nanoparticles are sequestered from the blood circulation into the MPS organs (spleen, liver, lymph nodes and lung) (Owens and Peppas, 2006). Small sized nanoparticles (< 8 nm in diameter) can also be excreted from the blood circulation through kidneys (Yu and Zheng, 2015).

In order to avoid the formation of a protein corona, the surface of nanoparticles can be modified with hydrophilic polymers that reduce interactions with the proteins. Most commonly the nanoparticle surface has been shielded with polyethylene glycol (PEG) (Abuchowski et al., 1977; Owens and Peppas, 2006) after whichthe circulation half-life (t^1/2) of these so-called “sterically” stabilized pegylated nanoparticles is extended. In addition to nanoparticles, some therapeutic proteins have been pegylated in an attempt to prolong their t^1/2. For instance, pegylated L-asparaginase displayed a circulation t^1/2 of 357 h compared to 20 h with the unmodified form in 25 patients (Ho et al., 1986). However, some problems have been accosiated with the pegylation. It has been shown to reduce the cell uptake efficiency of nanoparticles (Biswas et al., 2013; Hama et al., 2015; Manshian et al., 2015; Wang and Thanou, 2010). In addition, an accelerated blood clearance (ABC) of pegylated nanoparticles has been detected after repeated administrationsin animal models (Dams et al., 2000; Ishida et al., 2006a). It is still unclear whether the ABC phenomenon concerns pegylated empty or drug containing nanoparticles and if so, with which dosing regimens (Abu Lila et al., 2013).

Nanoparticles in oncology. The pegylated liposomal doxorubicin (trade name DOXIL^® in USA and CAELYX^® in Europe) was the first nanoparticle to receive clinical approval in Europe (table 1). The liposomes are formed by a lipid bilayer (mean diameter of ~80–100 nm) with the anticancer drug, doxorubicin, encapsulated into the aqueus interior (Gabizon et al., 2003). Doxorubicin is encapsulated in the liposomes by the ammonium sulfate gradient method, and it forms a rod-like precipitate inside the liposomes (Haran et al., 1993). The surface of the liposomes is pegylated and therefore their blood circulation time is extended.

This leads to longer exposure time with approx. 300-fold higher area under the curve (AUC) values and a longer t^1/2 (60–90 h) than conventional free doxorubicin (distribution phase t^1/2 of 5–10 min and terminal phase t^1/2 of 30 h) in human patients (Gabizon et al., 1994; Gabizon et al., 2003; Gabizon et al., 2012; Greene et al., 1983). The major benefit of DOXIL^® over free doxorubicin is the reduction of severe side-effect, cardiomyopathy, which was one of the main factors favouring its clinical approval (http://www.ema.europa.eu/ema/).

Other commercially available nanoparticles for cancer therapy are listed in table 1. The approval of nanoparticles has been based on two aspects;improved efficacy (e.g. Abraxane^® and Mepact^®) and fewer side-effects (e.g. Myocet^® and Oncaspar^®) when compared with their conventional counterparts (http://www.ema.europa.eu/ema/). There are also many therapeutic nanoparticles undergoing clinical evaluation. Some of the potential candidates in phase III trials are the heat-sensitive liposomal doxorubicin (ThermoDox^®) and paclitaxel albumin-stabilized and micellar formulations (Paclical^®, Taxol^® and ABI-007), as well as inorganic carbon nanoparticle lymph node tracers for cancer diagnostics/prognostics and crystallized hafnium oxide nanoparticles intended to enhance radiotherapy [clinicaltrials.gov and (Alphandery et al., 2015)].

2.1.1 Passive targeting

Nanoparticles accumulate into the solid tumor tissue via the enhanced permeation and retention (EPR)effect (Maeda, 2001; Matsumura and Maeda, 1986) (figure 2). The EPR effect is also commonly referred to as passive targeting. For instance, all the commercially available nanoparticles (table 1) take advantage of the EPR effect when gaining access to the solid tumor tissue.

The EPR effect is determined by the differences present in the vasculatures of normal and tumor tissues. The normal capillary vasculature wall consists of a continuous endothelial cell layer with relatively tight endothelia and a basement membrane, and therefore, large sized nanoparticles cannot pass through the endothelium by passive diffusion (figure 2). Unlike the normal vasculature, there are small fenestrae (< 0.5 µm) and intercellular gaps (~1.5 µm) in the abnormally formed tumor vasculature endothelium. In addition, the basement membrane is discontinuous or even absent on the basal side of the endothelial cell layer (Dudley, 2012; Huang et al., 2015). Therefore, as soon as nanoparticles reach the tumor site, they can extravasate through the fenestrae and pass through the gaps into the tissue space (figure 2). Typically nanoparticles 10 to 100 nm in diameter effectively accumulate in the tumor tissue, but nanoparticles up to 400 nm in diameter can also extravasate through the tumor vasculature (Cheng et al. 2012).Nanoparticles are retained in the tissue because of the high interstitial pressure and dysfunctional lymphatic drainage (figure 2).

Although passive targeting allows an efficient accumulation of nanoparticles into the solid tumor tissue, the nanoparticle distribution in the tissue is poor. This is due to the heterogeneous nature of the tumor vasculature. For instance, in different regions of the same tumor tissue, the endothelial structure, fenestrae and gap sizes vary, and in some regions, the vasculature can be absent (Minchinton and Tannock, 2006). This leads to an uneven accumulation of nanoparticles within the tumor tissue, or even to a lack of uptake.

In addition, the distribution of nanoparticles is further prevented by the extracellular matrix which is rich in collagen fibers and glycosaminoglycans (Bertrand et al., 2014). For instance, recently it was shown that the distance of drug extravasation from the blood vessels did not exceed 50 µm after liposomal drug delivery (Yokoi et al., 2015). However, the intratumoral mobility of nanoparticles has been improved with enzyme or small molecule drug treatments inhibiting the synthesis of extracellular matrix proteins or disrupting the extracellular matrix structures in the tumor tissue (Diop-Frimpong et al., 2011; Kohli et al., 2014; McKee et al., 2006).

Figure 2. Passive targeting of nanoparticles. Nanoparticles can extravasate through the tumor vasculature endothelium via intercellular gaps (~1.5 µm) and fenestrae (~0.5 µm). Due to the absence of lymphatic drainage, nanoparticles are retained in the tumor tissue. Adapted from (Dudley, 2012; Fanciullino and Ciccolini, 2009; Minchinton and Tannock, 2006).

Table 1. Passively targeted nanoparticles in clinical use. Sources of the information www.ema.europa.eu/ema/, www.fda.gov and (Allen and Cullis, 2004; Krishnan and Rajasekaran, 2014; Rivera Gil et al., 2010).

Formulation Toxic agent Trade name Cancer type Current status (approval year) Non-pegylated

liposomes Cytarabine DepoCyt^® Lymphomatous meningitis FDA-approved (1999)

Daunorubicin DaunoXome^® AIDS-related Kaposi’s

sarcoma FDA-approved

(1996)

Doxorubicin Myocet^® Metastatic breast cancer EMA-, Canada- approved (2000) Muramyl

tripeptide phosphatidyl ethanolamine

Mepact^®

(mifamurtide) Nonmetastatic resectable

osteosarcoma EMA-approved

(2009)

Vincristine sulfate Marqibo^® Philadelphia chromosome–

negative acute

lymphoblastic leukemia

FDA-approved (2012)

Irinotecan Onivyde™ Advanced (metastatic)

pancreatic cancer FDA-approved (2015)

Pegylated

liposome Doxorubicin DOXIL^®/CAELYX^® Recurrent ovarian cancer, metastatic breast cancer (except in the USA), AIDS- related Kaposi’s sarcoma and multiple myeloma

FDA-, EMA- approved (1995)

Polymeric

nanoparticle Paclitaxel Abraxane^® Metastatic breast cancer, Metastatic adenocarcinoma of the pancreas,Non-small cell lung cancer

FDA-, EMA- approved (2005)

Polymeric

micelle Paclitaxel Genexol^® PM Metastatic breast cancer,

urothelial carcinoma South Korea- approved (2006)

Polymeric

conjugates L-asparaginase Oncaspar^®

(pegaspargase) Acute lymphoblastic

leukemia FDA-approved

(1994) Neocarzinostatin Zinostatin

Stimalmer^® Primary unresectable

hepatocellular carcinoma Japan- approved (1994)

PEG = polyethylene glycol; FDA = US Food and Drug Administration; EMA = European Medicines Agency; AIDS = Acquired Immune Deficiency Syndrome.

2.1.2 Active targeting

The aim of active drug targeting is to enhance the uptake of a drug by the target cells in a specific manner. The enhanced cell uptake of actively targeted nanomedicines is based on the specific ligand-receptor, antibody-antigen or other molecular interactions on the cell surface (Steichen et al., 2013); some of these are listed in table 2.

The prequisites of active targeting include the selection of a suitable target protein and development of controlled formulation. First of all, the target protein, either receptor or antigen, has to be specifically expressed on the surface of the tumor cell or tumor vascular endothelium in order to facilitate selective targeting into the tissue. Some commonly overexpressed receptors and antigens in cancer cells include transferrin receptor (Galbraith et al., 1980), folate receptor (Parker et al., 2005) and human epidermal growth factor receptor (EGFR) (Grandis and Tweardy, 1993), and in endothelial cells vascular endothelial growth factor receptor (VEGF) (Holmes et al., 2007), integrins (α^vβ³ and α^vβ⁵) (Desgrosellier and Cheresh, 2010) and prostate-specific membrane antigen (PSMA) (Ghosh and Heston, 2004) (table 2). The level of receptor expression must be verified from specimens taken from the patient before the proper personalized therapy can be selected. In addition, the receptor needs to be expressed homogenously on the surface of the cancer cells or tumor vasculature endotheliumwith a long-term existence. It is also beneficial if the receptor recirculates back to the cell surface after internalization and is not shed to the blood circulation.

The selected receptor defines the ligand of choice. Monoclonal antibodies (mAb), fragments of the antibodies, proteins (e.g. growth factors, transferrin and affibody), peptides (e.g. Arginine-Glycine-Asparagine peptides and aptides), small molecules [e.g.

folic acid and thiamine pyrophosphate (TPP)], carbohydrate moieties (e.g. mannose, glucose and galactose) and nucleic-acid based aptamers (e.g. aptamer 10) are frequently studied ligands in targeted nanoparticles (Bertrand et al., 2014) (table 2). Often the ligand is attached to the distal end of a PEG-chain on the surface of the nanoparticle, which may disturb the orientation and/or the conformation of the ligand. For instance,the PEG-chains caused sterical hindrance between the ligand and receptor, when the ligand-conjugated PEG-chains were shorter than the surrounding PEG-chains (Stefanick et al., 2013). In addition, PEG-conjugated endothelium targeting peptides were shown to be hidden inside the surrounding PEG-layer (Lehtinen et al., 2012). Furthermore, the ligand-conjugation may disturb the steric stabilization role of PEG. It was previously demonstrated that the specific transferrin receptor mediated cell uptake of pegylated silica nanoparticles was reduced due to the formation of a protein corona on the surface of the nanoparticles (Salvati et al., 2013).

Even though none of the actively targeted nanoparticles has yet achieved clinical approval, antibody-drug conjugates are available for clinical use (table 3). In those conjugates, a cytotoxic drug, such as the microtubule-disrupting compound, auristatin, or the DNA-disrupting drug, calicheamicin (Jerjian et al., 2016), is conjugated either to actively targeting ligand (e.g. interleukin-2) or more commonly to mAb (e.g. anti-cluster of differentiation (CD) 19, -CD20, -CD22, -CD30, -CD56 and -CD79b). The field of antibody- drug conjugates is constantly expanding with 90 currently recruiting or ongoing clinical phase studies (clinicaltrials.gov).

Table 2. Examples of target proteins expressed in cancer tissues or cells, and ligands and antibodies used in the active targeting. The information is collected from the Internet (clinicaltrials.gov) and from the references indicated in the table.

Target protein Target cancer tissue or cells Targeting ligand/antibody Reference Transferrin receptor Colon, breast, kidney, lung,

stomach and ovarian cancers Transferrin (Zuckerman et al., 2014) Folate receptor Ovarian, renal, lung, breast and

pancreatic cancers Folic acid (Kue et al.,

2016) Human epidermal

growth factor receptors (EGFR and HER2)

Non-small cell lung, colorectal, ovarian, gastric, breast and bladder cancers and glioma

Epidermal growth factor (Mamot et al., 2012)

Cholecystokin receptors Pancreas, medullary thyroid, lung, breast, ovarian, gastrointestinal tract and colon cancers

Cholecystokinin and gastrin (Kue et al., 2016)

Prostate specific

membrane antigen Prostate cancer and high-grade

prostatic intraepithelial neoplasia Antibody (Hrkach et al., 2012) Mucin-1 Breast, lung, prostate and

colorectal cancers Monoclonal antibodies and

fragments of antibodies (Tang et al., 2010) Integrins (αvβ3 and

αvβ5) Breast, prostate, pancreatic, ovarian, cervical cancers and melanoma, glioblastoma and tumor endothelia

Tripeptide motif of Arginine-

Glycine-Asparagine (RGD) (Kue et al., 2016)

Vascular cell adhesion

molecule 1 Gastric, lung and breast cancers, melanoma, leukemia and renal cell carcinoma

Very late antigen-4 and

peptides (Nahrendorf

et al., 2006)

Vascular endothelial

growth factor receptor Neovascular endothelial cells Growth factors (Olsson et al., 2006;

Shi et al., 2015) Tumor endothelial

marker 1 Neovascular endothelial cells Monoclonal antibodies and

fragments of antibodies (Facciponte et al., 2014) Aminopeptidase N

receptor Perivascular cells of tumor blood vessels, stromal cells surrounding prostatic carcinoma cells, non- keratinising type cercical

squamous cell cancer, renal clear cell cancer

Peptides and monoclonal

antibodies (Wickstrom

et al., 2011)

Antigens expressed in tumor-associated macrophages

Breast, prostate, ovary, cervix, stomach, lung, brain and bladder cancers, and Hodgkin lymphoma

Monoclonal antibodies (Ries et al., 2014; Tang et al., 2013) Fibroblast activation

protein Pancreatic, gastrointestinal and

breast cancers Monoclonal antibodies and fragments of antibodies, fibroblast activation protein reactive T cells

(Tran et al., 2013)

Endothelial cell tyrosine

kinase receptor Monocytes in cancer tissue Monoclonal antibodies and

fragments of antibodies (Forget et al., 2014)

Table 3. Examples of actively targeted drug conjugates and nanoparticles. The information is collected from the Internet (clinicaltrials.gov; www.ema.europa.eu/ema/; www.fda.gov) and (Bertrand et al., 2014; van der Meel et al., 2013).

TR = transferrin receptor; p53 = tumor antigen; scFv = single chain variable domain fragments;

RB94 = retinoblastoma 94; HER2 = human epidermal growth factor receptor 2; CD =cluster of differentiation; VH = variable heavy; EGFR = epidermal growth factor receptor; NP = nanoparticle; RRM2 = ribonucleotide reductase M2; AD-PEG-hTf = adamantine-polyethylene glycol conjugated to human transferrin; ACUPA = S,S-2-[3-(5-amino-1-carboxypentyl)-ureido]- pentanedioic acid.

Formul. Code/

Name Toxic agent Target

[ligand] Cancer type Current status (approval year) Drug

conjugates Ontak^® Diphtheria

toxin Interleukin-2 receptor [Interleukin-2]

T-cell lymphoma FDA-approved (1999)

Zevalin^® Yttrium-90-

tiuxetan CD20 antigen

[Ibritumomab] B cell non-Hodgkin

lymphoma FDA-approved

(2002), EMA- approved (2004) SGN-35 Monomethyl

auristatin E CD30 antigen

[Brentuximab] Hodgkin lymphoma,systemic anaplastic large cell

lymphoma

FDA-approved (2011), EMA- approved (2012) CMC-544 Calicheamicin

γ1 CD22 antigen

[Inotuzumab] Non-Hodgkin lymphoma,

acute lymphoblastic leukemia Phase I, II, III Non-PEGylated

liposomes

MBP-426 Oxaliplatin TR

[Transferrin] Advanced/metastatic solid tumors and gastric, gastroesophageal junction, esophageal adenocarcinomas

Phase I, II

SGT-53 p53 plasmid

DNA TR [scFv] Recurrent glioblastoma, metastatic pancreatic cancer, advanced solid tumors

Phase I, II

SGT-94 RB94 plasmid

DNA TR [scFv] Solid tumors Phase I

MM-302 Doxorubicin HER2 [scFv] Breast cancer Phase I, II, III Lipovaxin-

MM Melanoma

antigens and interferon γ

CD209 [Domain antibody VH]

Melanoma Phase I

PEGylated

liposomes Anti-EGFR immuno- liposomes

Doxorubicin EGFR

[Cetuximab] Solid tumors Phase I

2B3-101 Doxorubicin Glutathione transporter [Glutathione]

Brain Metastases,

lung cancer, breast cancer, melanoma, malignant gliom a, meningeal carcinomatosis

Phase I, II

PEGylated polymeric NPs

CALAA-01 RRM2 siRNA Transferrin

[AD-PEG-hTf] Solid tumors Phase I BIND-014 Docetaxel Prostate-

specific membrane antigen [ACUPA]

Metastatic and castration- resistant prostate cancers, solid tumors, urothelial- and cholangiocarcinomas, non- small cell lung and

cervical cancers,

squamous cell carcinoma of head and neck

Phase I, II

2.2 INTRACELLULAR KINETICS OF NANOPARTICLES

After accessing the interstitial space inside the tumor, the nanoparticles must diffuse to the deeper parts of the tissue before thay can reach the target cells. Subsequently, the drug should be released from the nanoparticles either into the extracellular space or within the target cells. It has been estimated that about 60% of all drug target sites are located on the cell surface (Overington et al., 2006). However, many drugs have an intracellular target site i.e. acting at intracellular organelles, such as mitochondria, nucleus, early or late endosome, lysosome, endoplasmic reticulum or Golgi (Rajendran et al., 2010). For instance, many anticancer drugs, including DNA-intercalators, alkylating agents and topoisomerase inhibitors, have an intracellular target site located in the nucleus (Minotti et al., 2004;

Pommier, 2013; Witte et al., 2005).

The release of the drug from the nanoparticles can be triggered by either internal or external stimuli. A low pH and enzymatic activity are examples of internal stimuli, while heat (Needham et al., 2013; Yatvin et al., 1978), light (Lajunen et al., 2016) or ultrasound (Oerlemans et al., 2013)can be an external stimulus. It is well known that the temperature in tumor tissue is slightly elevated (by 0.3–2 °C) (Chung et al., 2011; Keyserlingk et al., 2000) and the pH is lower (6.5–7.2) than the normal physiological pH (7.4) (Vander Heiden et al., 2009; Warburg, 1956). In addition, the enzymatic activity in tumor tissue (e.g.

phospholipases, glycosidases and matrix metalloproteinases) and lysosomes (e.g.

glycosidases, sulfatases and cathepsin proteases) has been utilized for the degradation of the nanoparticles and subsequently triggering the drug release (Lee et al., 2011; Zhu et al., 2012). Intracellular late endosomes and lysosomes have also lower pH (4.5–5.5) (Forgac, 2007), and therefore, many pH-responsive nanoparticles or drug conjugates, e.g. with pH- sensitive polymers or lipids, or acid-labile linkers, have been developed for cancer therapy (Kratz, 2007; Li et al., 2015; Xu et al., 2015).

2.2.1 Cell adhesion

Macromolecular nanoparticles are internalized by the cells via energy-dependent endocytosis mechanisms (Anderson et al., 1978). However, in order to be endocytosed, nanoparticles need to attach to structures on the cell membrane through molecular interactions. These interactions can be considered as either attractive or repulsive forces (Min et al., 2008) and as being either specific or unspecific in nature. The specific ligand- receptor interactions are necessary for active targeting, whereas the unspecific interactions, such as hydrophobic and electrostatic, are important for the nontargeted drug delivery systems (Zhang et al., 2015).

The physicochemical properties of nanoparticles largely determine the efficiency of cell adhesion. Surface charge is a major physicochemical factor that has an influence on whether or not nanoparticles can bind to the cell. The cell membrane is negatively charged due to the presence of a large number of cell membrane associated anionic glycoproteins (Varki et al., 2009). For this reason, positively charged (cationic) nanoparticles can adhere to the cell membrane more avidly than their negatively charged or neutral counterparts.

Cationic compounds, polyethyleneimine and polyamidoamine, are both commonly used to increase the cell adhesion of nanoparticles, especially in the field of nonviral gene therapy (Felgner et al., 1987; Hwang et al., 2015; Ilina et al., 2012; Nabel et al., 1993). In addition, other physicochemical properties of nanoparticles, such as shape, hydrophilicity and surface roughness, can modulate the nanoparticle-cell interactions (Nel et al., 2009). For example, the electrostatic repulsive forces and steric hindrance between the nanoparticles and the cell surface are increased by the attachment of hydrophilic polymers, such as PEG, surfactants or poly(aspartic acid), onto the nanoparticle surface (Min et al., 2008; Zhang et al., 2015). Subsequently, this leads to reduced cell uptake efficiency, as has been demonstrated with pegylated silver nanoparticles (Manshian et al., 2015). The shape can be

also an important factor determining the cell adhesion properties. A rod-like shape of a nanoparticle improved the cell adhesion in comparison with cylindrical and spherical shapes (Gratton et al., 2008; Tan et al., 2013).

2.2.2 Cell internalization

Nanoparticles enter the cells via cell-type specific endocytosis pathways, classified into phagocytosis and pinocytosis (figure 3). Pinocytosis occurs in almost all cell types (Doherty and McMahon, 2009) whereas phagocytosis is present only in specialized mammalian cells (e.g. macrophages, neutrophils and dendritic cells) (Metchnikoff, 1893). Pinocytosis refers to a wide diversity of different endocytosis pathways, such as macropinocytosis, clathrin-, caveolae- and follitin-mediated endocytosis, and other clathrin-independent endocytosis pathways (figure 3).

Although the presence of endocytosis pathways have been known for a long time (Anderson et al., 1976; Brown and Goldstein, 1986), and the molecular mechanisms studied (Lu et al., 2009a), the internalization pathways of nanoparticles are still not well characterized in the literature. Recent studies have suggested that the cell internalization of nanoparticles takes place through multiple co-existing endocytosis pathways (Gilleron et al., 2013; Kuhn et al., 2014), and is predominantly defined by the size of the nanoparticles (Champion et al., 2008; Chono et al., 2007). For instance, liposomal nanoparticles were shown to enter cells via a biphasic mechanism using clathrin-mediated endocytosis and macropinocytosis in a cell-type specific manner (Gilleron et al., 2013). In addition, a combination of several uptake mechanisms was observed to be involved in the uptake of polystyrene nanoparticles into macrophages and epithelial cells (Kuhn et al., 2014). These studies indicate that the cell uptake of nanoparticles not only occurs through multiple overlapping mechanisms, but is also cell-type dependent. The influence of size is based on the endocytosis pathways expressed in the target cells. It is generally thought that larger nanoparticles (< 5 µm) can be internalized through phagocytosis (Champion et al., 2008) or macropinocytosis (Mercer and Helenius, 2009), whereas the clathrin-mediated pathway is limited to ~100 nm sized nanoparticles (Miller et al., 2015) and caveolae-mediated up to 80 nm (Parton and Simons, 2007). For example, the most optimal particle size for the efficient cell uptake of mesoporous silica nanoparticles was shown to be 50 nm in the range of 30 to 280 nm in human cervical cancer cells (Lu et al., 2009b).

Figure 3. Cell internalization of nanoparticles. Endocytosis pits, spherical and tubular in shape, are formed from the cell membrane invaginations. The approximate diameters of the structures are indicated in the figure (Champion et al., 2008; Mercer and Helenius, 2009; Miller et al., 2015; Parton and Simons, 2007; Swanson, 2008). Different proteins [e.g. dynamin, clathrin, caveolin and guanosine 5'-triphosphatase regulator associated with focal adhesion kinase 1 (GRAF1)] are involved with the endocytosis pathways. Modified from (Johannes et al., 2015;

Mercer and Greber, 2013).

2.2.3 Intracellular distribution

After endocytosis, the nanoparticles are directed to the early and late endosomes (figure 4).

Later, the late endosomes are transformed into mature lysosomes, which lead to the rapid accumulation of nanoparticles in the lysosomal space (Huang et al., 2016; Owen et al., 2013). Inside lysosomes, the nanoparticles can be degraded by lysosomal enzymes or exocytosed back to the extracellular space (figure 4). In exocytosis, the lipid membrane of an intracellular vesicle, exosome, becomes fused with the cell membrane, allowing the escape of nanoparticles from the cells (Fan et al., 2016). It is also possible that the nanoparticles are transported within the vesicular structures through the polarized cell epithelium or endothelium from the apical to the basolateral site or vice versa (Tuma and Hubbard, 2003).

Typically the drug is released in endo-lysosomes, after which the drug can be distributed according to its physicochemical properties (figure 4). For example, the pH-gradient between the cytosol (pH ~7) and the lysosomes (pH > 4.5) is known to lead to the lysosomal accumulation of weakly basic drugs, such as chloroquine and doxorubicin (Duvvuri et al., 2004; Hostetler et al., 1985). The drug becomes entrapped in the lysosomal compartment because the protonated form of the drug has poor permeability (Goldman et al., 2009).

Cationic and lipophilic compounds are sequestered in mitochondria due to the mitochondrial membrane potential (Logan et al., 2016). It is also possible that the highly lipophilic drugs, e.g. diazepam,trazodone or verapamil (Cheymol, 1993), may partition into the lipidic phase, i.e. lipidic membranes or lipid droplets (Fujimoto and Parton, 2011). In addition to the physicochemical properties, the amount of possible drug binding sites is crucial for intracellular drug distribution. There are many potential intracellular binding sites for drugs e.g. proteins, DNA or RNA. For example, the drug examined in this thesis, doxorubicin, binds to DNA (Perez-Arnaiz et al., 2014) whereas cocaine and thiouracil can be bound to the natural pigment, melanin (Aubry, 2002). Therefore, these sites might act as subcellular depots for drugs. The intracellular drug distribution can be also affected by many other cell type specific factors, such as the expression of efflux transporters or the presence of metabolizing enzymes. Due to these reasons, the subcellular localization of drugs might be highly compartmentalized; in fact the subcellular localization of drugs and nanoparticles is still poorly understood.

2.2.4 Methodology to study the intracellular distribution of drugs and nanoparticles An awareness of the intracellular drug distribution and concentrations is highly important in order to understand the pharmacological on-site and off-site effects, i.e. efficacy and toxicity. However, the clarification of intracellular drug distribution is challenging due to the lack of sensitive detection methods. For example, there are no standardized methods capable of undertaking the intracellular quantification of drugs. A summary of the available methods is shown in table 4.

Figure 4. Schematic illustration of intracellular distribution of nanoparticles (NP). NPs use energy-dependent endocytosis mechanisms for internalization by cells. After the endocytosis, nanoparticles are located inside the vesicular structures, endosomes. Endosomes can fuse with lysosomes, where the nanoparticles can be degraded (Gilleron et al., 2013). In order to make contact with the molecular target site, the drug has to be released from the NPs. After its release, the intracellular distribution of the drug (in the free form) is controlled by its physicochemical properties, and the specific binding rates, affinities and capacities towards the intracellular binding sites (Eytan, 2005), such as DNA, RNA or proteins. Endoplasmic reticulum

= ER.

2.2.4.1 Fluorescence microscopy imaging

Imaging with fluorescence microscopy is the most common method used for investigating the intracellular distribution of drugs and nanoparticles. This technique allows the imaging of intact cells, even at a single molecule level. The principle of two-dimensional wide-field microscopy imaging and three-dimensional confocal microscopy imaging is that first the native molecule is exited with the specific wavelength of light, and after the relaxation, the emitted light (i.e. emission spectra) is measured.

The popularity of imaging is partly due to the wide variety of fluorescent probes available, such as rhodamine B, calcein and AlexaFluor 514, which can be conjugated with or encapsulated in nanoparticles with relative ease. In addition, fluorescent markers for cell organelles, such as early endosomes, Golgi, lysosomes, mitochondria and nucleus, make it possible to detect fluorescently labeled nanoparticles within the specific organelles (Zeng et al., 2014). However, in the case of small drug molecules, the conjugation of fluorescent probes is not feasible since the probe can change the physicochemical properties and the intracellular kinetics of the drug. Some small molecule drugs, such as doxorubicin and daunorubicin (Karukstis et al., 1998), are intrinsically fluorescent and therefore readily amenable to fluorescence based techniques.