Targeted drug delivery to breast cancer using polymeric nanoparticle micelles

by

Karyn Susana Ho

A thesis submitted in conformity with the requirements for the degree of Doctor of Philosophy

Department of Chemical Engineering and Applied Chemistry Institute of Biomaterials and Biomedical Engineering

University of Toronto

© Copyright by Karyn Susana Ho, 2012

ii

Targeted drug delivery to breast cancer using polymeric nanoparticle micelles

Karyn Susana Ho

Doctor of Philosophy

Department of Chemical Engineering and Applied Chemistry Institute of Biomaterials and Biomedical Engineering

University of Toronto

2012

Abstract

Broad distribution and activity limit the utility of anti-cancer compounds by causing

unacceptable systemic toxicity and narrow therapeutic indices. To improve tumour

accumulation, drug-loaded macromolecular assemblies have been designed to replace

conventional surfactant-based formulations. Their nanoscale size enhances tumour accumulation

via hyperpermeable vasculature and reduced lymphatic drainage. Incorporating targeting ligands

introduces cell specificity through receptor-specific binding and uptake, enabling drugs to reach

intracellular targets. In this work, the targeting properties of polymer nanoparticle micelles of

poly(2-methyl-2-carboxytrimethylene carbonate-co-D,L-lactide)-graft-poly(ethylene glycol)-

furan (poly(TMCC-co-LA)-g-PEG) were verified using in vitro and in vivo models of breast

cancer.

To select a relevant mouse model, the vascular and lymphovascular properties of two tumour

xenograft models were compared. Greater accumulation of a model nanocarrier was observed in

orthotopic mammary fat pad (MFP) tumours than size matched ectopic subcutaneous tumours,

suggesting that the organ environment influenced the underlying pathophysiology.

Immunostaining revealed greater vascular thickness, density and size, and thinner basement

iii

membranes in MFP tumours, likely contributing to greater blood perfusion and vascular

permeability.

Based on these observations, MFP tumour-bearing mice were used to characterize the

pharmacokinetics and biodistribution of a taxol drug, docetaxel, encapsulated in poly(TMCC-co-

LA)-g-PEG nanoparticles. The nanoparticle formulation demonstrated longer docetaxel

circulation in plasma compared to the conventional surfactant-based formulation. As a result,

greater docetaxel retention was uniquely measured in tumour tissue, extending exposure of

tumour cells to the active compound and suggesting potential for increased anti-cancer efficacy.

Furthermore, active targeting of antibody-modified nanoparticles to live cells was shown to be

selective and receptor-specific. Binding isotherms were used to quantify the impact of antibody

density on binding strength. The equilibrium binding constant increased linearly with the

average number of antibodies per particle, which is consistent with a single antibody-antigen

interaction per particle. This mechanistic understanding enables binding behaviour to be

adjusted in a predictive manner and guides rational nanoparticle design.

These studies validate poly(TMCC-co-LA)-g-PEG nanoparticles as a platform for targeted

delivery to cancer on both a tissue and cellular level, forming a compelling justification for

further pre-clinical evaluation of this system for safety and efficacy in vivo.

iv

Acknowledgements

I would first like to thank my closest collaborators and co-authors: Peter Poon, Ahmed Aman, Yoko Kosaka, Meng Shi, Yakov Lapitsky, Shawn Owen, Armand Keating, Rima Al-awar, Simone Helke, Xinghua Wang, and Yusuke Katayama. You not only greatly accelerated my progress, but without you this work would not have been possible. I would also like to thank the following people for always being generous with their time and insight: Warren Chan, Carol Lee, Ralph DaCosta, Brian Wilson, Tracy Liu, Ian Corbin, Gang Zheng, Ray Reilly, Robert Prud’homme, James Jonkman, Miria Bartolini, and Alicia Viloria-Petit. To the remaining people who served on my various committees, thank you for driving me to think about my project more deeply: Vladimir Torchilin, Shirley Wu, Milica Radisic, and Brad Saville.

A special thank you to my supervisor, Molly Shoichet, for the initial vision and opportunity to pursue this project: Molly, you gave me advice, resources, and manpower to ensure I had the best chance of success in my degree. Above all, you were flexible and encouraging of special skills that will be invaluable to my future professional development. I am grateful that you share my excitement in all my endeavours.

My family has always motivated me to make the most of my potential. Mom, your unconditional support and pride give me confidence to go after my biggest dreams. You have sacrificed so much to provide for all of us, and I can’t imagine where we would be without you. Dad, it goes without saying that your enthusiasm for education inspires me to love learning. Karen, your nearly daily presence makes me feel like a piece of home is always with me. I will never forget that my life in science would never have taken off without your insistence. Karmen, from sharing a room to sharing an apartment, I am thankful for your constant companionship. You show us all that perseverance leads to great things.

To my labmates, thank you for making my time here so enjoyable, and for showing me every day how to be a better scientist. I am especially thankful for the friends I made here who did not stop at making a mark on me professionally. You were a sounding board for anything I needed, and you made my days in Toronto brighter: Ryan Wylie, Douglas Baumann, Michael Conrad, Howard Kim, Nic Leipzig, Katarina Vulic, and Malgosia Pakulska. To my best friend, Catherine Kang, when we met I could not have imagined how full my life would become by having you in it. Our friendship has touched and illuminated every aspect of my life and broadened my ambitions in countless ways. To Crissa Koh, thank you for pushing me to meet your high standards and work ethic in our final year at UBC. Your energy helped me get here and encouraged me to see more of the world.

Karyn Susana Ho

v

Table of Contents

Acknowledgements ........................................................................................................................ iv  

Table of Contents ............................................................................................................................. v  

List of Tables ................................................................................................................................... x  

List of Figures ................................................................................................................................. xi  

1   Rationale ..................................................................................................................................... 1  

1.1   Hypothesis and objectives ................................................................................................... 2  

2   Introduction ................................................................................................................................. 4  

2.1   Barriers to conventional anti-cancer drug delivery strategies ............................................. 4  

2.2   Targeting the solid tumour microenvironment .................................................................... 6  

2.2.1   Hyperpermeable tumour vasculature and tissue level accumulation ....................... 6  

2.2.2   Overexpressed cancer genes and receptor mediated cell uptake ............................. 8  

2.3   Polymeric nanoparticles and design elements in nanomedicine .......................................... 9  

2.3.1   Controlling self-assembly and particle size ........................................................... 10  

2.3.2   Extending circulation time and drug activity ......................................................... 11  

2.3.3   Increasing drug loading and micelle stability ........................................................ 12  

2.3.4   Achieving bioactive surface modification ............................................................. 14  

2.4   Poly(TMCC-co-LA)-g-PEG graft copolymer for nanoparticle drug delivery .................. 15  

2.4.1   Material properties ................................................................................................. 15  

2.4.2   Chemical modification .......................................................................................... 17  

2.5   Antibody-modified nanoparticles demonstrate antigen-specific activity .......................... 18  

2.5.1   Herceptin immunonanoparticle live cell equilibrium binding ............................... 18  

2.5.2   Herceptin immunonanoparticle cell cytotoxicity ................................................... 20  

2.6   Conclusions ....................................................................................................................... 24  

vi

2.7   Thesis scope ....................................................................................................................... 25  

2.8   References ......................................................................................................................... 26  

3   Pathophysiological assessment of human tumour xenografts as models of EPR in breast cancer ........................................................................................................................................ 33  

3.1   Abstract .............................................................................................................................. 33  

3.2   Background ........................................................................................................................ 34  

3.3   Methods ............................................................................................................................. 36  

3.3.1   Materials ................................................................................................................ 36  

3.3.2   Cell maintenance and preparation ......................................................................... 36  

3.3.3   Tumour xenograft models ..................................................................................... 37  

3.3.4   Dye injections and tissue collection ...................................................................... 37  

3.3.5   Immunostaining ..................................................................................................... 38  

3.3.6   Image acquisition and analysis .............................................................................. 38  

3.4   Results and discussion ....................................................................................................... 39  

3.4.1   Orthotopic cell transplantation influences tumour growth rate and size variation ................................................................................................................. 39  

3.4.2   MFP tumours exceed SC tumours in model nanocarrier accumulation ................ 41  

3.4.3   Elements of tumour vascular pathophysiology observed in tumour models ........ 42  

3.5   Conclusions ....................................................................................................................... 49  

3.6   List of abbreviations used .................................................................................................. 49  

3.7   Authors’ contributions ....................................................................................................... 50  

3.8   Acknowledgements ........................................................................................................... 50  

3.9   References ......................................................................................................................... 50  

4   Drug-loaded nanoparticles for targeted delivery to a mouse model of breast cancer ............... 54  

4.1   Abstract .............................................................................................................................. 54  

4.2   Introduction ....................................................................................................................... 54  

4.3   Experimental ...................................................................................................................... 57  

vii

4.3.1   Materials ................................................................................................................ 57  

4.3.2   DTX concentration measurement .......................................................................... 57  

4.3.3   Free DTX and DTX-NP formulation ..................................................................... 58  

4.3.4   Cell maintenance and preparation ......................................................................... 58  

4.3.5   Tumour xenograft model ....................................................................................... 59  

4.3.6   Pharmacokinetics and biodistribution ................................................................... 59  

4.3.7   Plasma preparation ................................................................................................ 60  

4.3.8   Tissue preparation .................................................................................................. 60  

4.4   Results ............................................................................................................................... 60  

4.4.1   Pharmacokinetics ................................................................................................... 60  

4.4.2   Biodistribution ....................................................................................................... 63  

4.5   Discussion .......................................................................................................................... 65  

4.6   Conclusions ....................................................................................................................... 68  

4.7   Acknowledgements ........................................................................................................... 68  

4.8   References ......................................................................................................................... 69  

5   Antibody-modified nanoparticles for active and tunable binding to cancer cells .................... 73  

5.1   Abstract .............................................................................................................................. 73  

5.2   Introduction ....................................................................................................................... 74  

5.3   Experimental ...................................................................................................................... 76  

5.3.1   Materials ................................................................................................................ 76  

5.3.2   Nanoparticle synthesis ........................................................................................... 77  

5.3.3   Cell lines and maintenance .................................................................................... 77  

5.3.4   Flow cytometric analysis ....................................................................................... 78  

5.4   Theory ................................................................................................................................ 78  

5.4.1   Multivalent Binding ............................................................................................... 79  

5.4.2   Monovalent Binding .............................................................................................. 80  

viii

5.5   Results and Discussion ...................................................................................................... 82  

5.6   Conclusions ....................................................................................................................... 87  

5.7   Acknowledgements ........................................................................................................... 87  

5.8   References ......................................................................................................................... 87  

5.9   Supplementary information ............................................................................................... 91  

6   Discussion ................................................................................................................................. 92  

6.1   Validated pre-clinical tumour models ............................................................................... 92  

6.2   Quantitative pharmacokinetics and biodistribution ........................................................... 94  

6.3   Long circulating polymer nanoparticles ............................................................................ 95  

6.4   Successful passive tumour targeting .................................................................................. 96  

6.5   Quantitative live cell binding ............................................................................................ 96  

6.6   Mechanistic predictions of binding behaviour .................................................................. 97  

6.7   Conclusions ....................................................................................................................... 98  

6.8   Achievement of objectives ................................................................................................ 99  

6.9   References ....................................................................................................................... 100  

7   Limitations and recommendations for future work ................................................................ 101  

7.1   Mixed cell populations in tumor xenograft models ......................................................... 101  

7.2   Cellular uptake as a function of antibody density ........................................................... 102  

7.3   Tumour penetration as a function of antibody density .................................................... 103  

7.4   Evaluating alternative targeting ligands .......................................................................... 104  

7.5   Safety and efficacy .......................................................................................................... 104  

7.6   References ....................................................................................................................... 105  

Copyright Acknowledgements .................................................................................................... 109  

8   Appendices .............................................................................................................................. 111  

8.1   List of abbreviations ........................................................................................................ 111  

8.2   List of parameters and mathematical notation ................................................................. 112  

ix

8.3   Additional data ................................................................................................................ 114  

8.4   References ....................................................................................................................... 116  

x

List of Tables

Table 2.1 Summary of polymer and micelle properties. .............................................................. 15  

Table 2.2 Flow cytometry data showing apoptotic mechanism of cell death for SKBR-3 cells

treated with 1.75 µg/mL DOX or DOX equivalent after 24 hours of incubation. ......................... 24  

Table 3.1 Immunostaining protocol details listed by antigen ........................................................ 38  

Table 4.1 Pharmacokinetic parameters calculated for DTX formulations after bolus IV

administration of 1.5 mg/kg DTX to tumour-bearing mice ........................................................... 62  

xi

List of Figures

Figure 2.1 Broad drug distribution and activity can cause severe toxicity in healthy tissues.

Resulting side effects are commonly seen in (A) blood (anemia, immune suppression), (B) the

digestive tract (nausea, vomiting), and (C) hair follicles (hair loss). .............................................. 5  

Figure 2.2 Solid tumour pathophysiology is permissive to active and passive targeting of

nanoscale carriers. Features include disorganized blood vessel architecture, leading to gaps in

the blood vessel wall that allow unregulated transport of macromolecules and other nanoscale

materials into tumour tissue. Nanoparticles modified with antibodies are depicted here

selectively extravasating across these gaps. Certain oncogenes may also be overexpressed as cell

surface markers, providing a target for antibody binding and a means for cellular uptake. ........... 8  

Figure 2.3 Long plasma circulation times promote passive targeting of nanocarriers via EPR by

increasing the number of passes through hyperpermeable tumour vasculature. PEGylated

nanoparticles are depicted here passing through gaps between disorganized endothelial cells in a

tumour blood vessel. ...................................................................................................................... 12  

Figure 2.4 Active recognition and binding of cancer cells can be achieved using targeting

ligands against an overexpressed oncogene on the cell surface to provide a means of selective

cellular uptake. As an example, an antibody-modified nanoparticle is depicted here binding a

cell through an antigen-specific interaction. .................................................................................. 14  

Figure 2.5 Synthesis of poly(TMCC-co-LA)-g-PEG through ring opening polymerization.

Highlighted: hydrophobic backbone (green); PEG graft (grey); furan functional group (red).

(Reprinted with permission from John Wiley and Sons [78]) ....................................................... 16  

Figure 2.6 Covalent antibody attachment to prepared nanoparticles via Diels-Alder chemistry.

Diels-Alder coupling was selected to preserve antibody bioactivity because it proceeds under

mild reaction conditions and in aqueous media. (Reprinted with permission from John Wiley and

Sons [78]) ...................................................................................................................................... 18  

xii

Figure 2.7 Flow cytometry analysis of Herceptin immunonanoparticle binding to cell lines at

equilibrium. SKBR-3 (HER2 overexpressing), MCF-7 (normal HER2 expression), and MDA-

MB-468 (HER2 negative) cell lines were incubated with fluorescently labeled nanoparticles to

confirm selectivity of Herceptin immunonanoparticles for HER2 overexpressing cells, and

specific antibody-antigen interactions versus a series of controls for non-specific adsorption. ... 20  

Figure 2.8 Confocal images showing cellular uptake of various nanoparticle formulations in

SKBR-3 cells after 6 h of incubation at 37 °C. Blue indicates cell nuclei (A-D), green indicates

fluorescently labeled polymer (A-B), and red indicates DOX autofluorescence (C-D). (A) shows

unmodified nanoparticles, where no uptake or binding is observed. (B) depicts Herceptin

nanoparticles, where binding and receptor-mediated uptake are both observed. (C) shows that

DOX nanoparticles are also internalized and accumulate intracellularly via hydrophobic

interactions between DOX and the cell membrane. (D) illustrates that even greater uptake is

observed with Herceptin-DOX-nanoparticles, as both uptake mechanisms are present. .............. 21  

Figure 2.9 Average fluorescence measurements from confocal images of DOX-conjugated

nanoparticles inside SKBR-3 cells after 6 h of incubation at 37 °C. These data show that while

both formulations undergo rapid internalization, Herceptin enhances cell uptake. ...................... 22  

Figure 2.10 Cell growth relative to controls for DOX treated SKBR-3 breast cancer cells versus

HMEC-1 healthy endothelial cells after 72 h of treatment at 5.0 µg/mL DOX or DOX equivalent

(IC50 of free DOX against SKBR-3 cells). This measure of effective cytotoxicity reflects a

combination of cytotoxic and cytostatic effects. ........................................................................... 23  

Figure 3.1 MFP and SC tumour sizes. Tumour volumes were calculated based on caliper

measurements post-dissection of the major and minor axes and thickness (n = 4-6). SC tumours

required longer development times to become size matched to MFP tumours. Greater variability

was also observed at longer times, particularly in SC tumours, where several animals had smaller

tumours than the cohort examined the week before. ..................................................................... 40  

Figure 3.2 FITC-Dextran accumulation in tumour tissue normalized to liver tissue control. High

molecular weight dextran (2 MDa, ~80 nm) was injected IV into tumour animals as a model

nanocarrier and allowed to distribute prior to sacrifice. 3 week old MFP tumours showed higher

accumulation of the nanocarrier than 5 week old SC tumours at a 90% confidence interval. All

xiii

data are shown as the mean of n = 4 animals ± SD. Lines connecting bars denote statistical

significance, P < 0.10. ................................................................................................................... 42  

Figure 3.3 CD31 and collagen IV immunostaining. Mean blood vessel wall thickness visualized

through (A) CD31 (endothelial cells) and (B) collagen IV (basement membrane). Both are

abnormally thick as compared to healthy liver control tissue, which is denoted by the dashed line.

(C) shows that mean blood vessel density assayed using CD31 staining is greatest in 3 week old

MFP tumours. (D) indicates mean vascular area as a measure of blood vessel size and capacity.

Their small size categorizes them as microvasculature. All data are shown as the mean of n = 4

animals ± SD. Starred lines connecting bars denote statistical significance, P < 0.05. ............... 44  

Figure 3.4 CD31 and αSMA co-staining. Representative images of pericytes (αSMA, violet) that

are detached from blood vessels (CD31, brown) in: (A) 3 week MFP, (B) 4 week MFP, and (C) 5

week SC tumours. Several blood vessels are highlighted with black arrows; blue staining

represents cell nuclei. (D) shows that pericytes are exclusively associated with blood vessels in

healthy liver control tissue. Scale bars represent 200 µm. ........................................................... 46  

Figure 3.5 LYVE-1 immunostaining. (A) shows mean lymphatic vessel density, and (B) shows

mean vessel area, both of which are indicators of lymphovascular capacity. Both measures were

found to have unequal variance between groups, and therefore although the groups were not

equivalent, ANOVA could not be used to verify their differences. While 3 week old MFP

tumours had the highest mean lymphatic vessel density, 5 week old SC tumours had greater

mean vessel size, both of which contribute to overall lymphatic drainage capacity. All data are

shown as the mean of n = 4 animals ± SD. Representative images of collagen (violet) positive

but CD31 (brown) negative fluid filled spaces are shown in: (C) 3 week MFP and (D) 5 week SC

tumours. Several of these spaces are highlighted with black arrows; blue staining represents cell

nuclei. Scale bars represent 200 µm. ............................................................................................ 48  

Figure 4.1 Poly(TMCC-co-LA)-g-PEG, shown here with a furan group at the PEG terminus, is

an amphiphilic co-polymer that self assembles into polymeric nanoparticle micelles with a core-

shell structure on dialysis against water. DTX and the polymer are first co-dissolved in organic

solvent before dialysis. During dialysis, DTX partitions into the hydrophobic core, thereby

encapsulating it. The polymeric nanoparticles have functional groups available for further

xiv

modification: carboxylic acid groups on the poly(TMCC-co-LA) backbone and furan moieties on

the PEG corona. ............................................................................................................................. 56  

Figure 4.2 Pharmacokinetic profiles of free DTX (o) and DTX-NP (�) in tumour-bearing mice.

The plasma profiles differ significantly by 2 h post injection. The DTX-NP formulation reached

its terminal elimination phase earlier, and coupled with a slower terminal elimination rate, the

enhanced plasma retention continued to amplify over time. Points shown are the mean of n=5

animals, with error bars representing their standard deviation. Starred points represent

statistically different group means (p < 0.05). ............................................................................... 61  

Figure 4.3 Biodistribution profiles of free DTX (o) and DTX-NP (�) in (A) liver, (B) spleen,

(C) lung, (D) kidney, (E) heart, and (F) tumour tissue. Points shown are the mean of n=5

animals, with error bars representing their standard deviation. Starred points represent

statistically different group means (p < 0.05). ............................................................................... 64  

Figure 5.1 Functionalizing immunonanoparticles with greater numbers of targeting antibodies

enhances their ability to associate with target cells. This effect can result from (A) increases in

binding events per particle (multivalent binding) or (B) increases in possible binding

configurations with a single interaction (monovalent binding). Illustrated here are

immunonanoparticles with Ω = 3 attached antibodies. In (A), the number of antibodies bound to

cell receptors, α, is shown as α = 3 (left) and α = 2 (right). In (B) the number of antibodies

bound to cell receptors, α, is shown as α =1 for all nanoparticles. The mechanism is an important

consideration in immunonanoparticle design, as it dictates how binding strength will increase as

the antibody conjugation density increases. .................................................................................. 75  

Figure 5.2 Fractional coverage of Herceptin immunonanoparticles bound to HER2

overexpressing SKBR-3 cells as a function of immunonanoparticle concentration. The arrow

indicates ascending order of antibody conjugation density: Herceptin immunonanoparticles

bearing 1.9 (�), 3.2 (r), 5.9 (�), and 9.4 (n) antibodies; inset shows fractional coverage for

free Herceptin (p). ........................................................................................................................ 83  

Figure 5.3 (A) eqK and (B) GΔ increase in absolute value as the number of Herceptin

antibodies per nanoparticle increases, thereby indicating greater binding affinity. The open

symbols (¯) represent the values calculated for free Herceptin, which denotes a monovalent

xv

case, and the closed symbols (u) represent Herceptin immunonanoparticles. The trends in (A)

and (B) follow the theoretical behaviour of monovalent immunonanoparticle binding. .............. 84  

Figure 5.4 Comparison of the experimental and theoretical fractional coverages (θ) of SKBR-3

cells by free Herceptin (p) and Herceptin immunonanoparticles bearing 1.9 (�), 3.2 (r), 5.9

(�), and 9.4 (n) antibodies exhibiting monovalent binding. The experimentally derived θ values

closely match the theoretically predicted θ values, with R2 = 0.99. .............................................. 86  

Figure 5.5 In the case of monovalent binding, increasing the number of antibodies per particle,

Ω, results in an increase in the number of possible binding configurations where only one

interaction occurs. (A) The number of possible combinations of monovalent binding events

increases with the number of conjugated antibodies due to an amplified number of possible

rotational binding orientations for N bound nanoparticles (first term of Equation 5). (B) Also,

given a lattice of M potential binding sites, the number of distinct lattice configurations in which

the particles can bind is given by a binomial coefficient (second term of Equation 5). ............... 91  

Figure 8.1 Plasma concentration profile for docetaxel in tumour-bearing mice. Orange symbols

show Herceptin-docetaxel-nanoparticles, black symbols show docetaxel nanoparticles, white

symbols show free docetaxel. ...................................................................................................... 114  

Figure 8.2 Tissue concentration profile for docetaxel in tumour-bearing mice. Orange symbols

show Herceptin-docetaxel-nanoparticles, black symbols show docetaxel nanoparticles, white

symbols show free docetaxel. ...................................................................................................... 115  

1

1 Rationale Broad distribution and activity underlie the dose-limiting systemic toxicity associated with anti-

cancer drug therapy. Further compounding the problem, promising drug candidates are often

bulky and polycyclic hydrophobic compounds with poor aqueous solubility. As a result, they are

formulated in mixtures of surfactants and organic solvents which exert their own non-specific

toxicity. Therefore, targeting strategies that replace surfactant-based formulations and deliver a

greater portion of the injected dose to cancer cells have the potential to significantly enhance

treatment safety and efficacy.

To improve selectivity, several unique cancer features have been identified as potential targets,

including abnormal vascular structure and pathological overexpression of cell surface receptors.

Drug-loaded nanoscale assemblies have previously been shown to accumulate selectively in

tumour tissue via enhanced permeability and retention (EPR) that results from hyperpermeable

tumour vasculature and insufficient lymphatic drainage. Antibody-modified nanoparticles have

further demonstrated receptor-specific binding of cancer cells, inducing endocytosis and

providing a mechanism for selective drug uptake.

An amphiphilic graft copolymer, poly(TMCC-co-LA)-g-PEG, was previously designed to take

advantage of these features. Poly(TMCC-co-LA)-g-PEG nanoparticles form through self-

assembly into ordered nanostructures having a hydrophobic core for drug loading, and a

hydrophilic shell for extended blood circulation and solubility in aqueous media. Furan

functional groups on the free PEG termini allow covalent attachment of maleimide-modified

antibodies under mild reaction conditions via Diels-Alder chemistry, promoting binding activity

of the final construct. The successful application of poly(TMCC-co-LA)-g-PEG nanoparticles to

cancer targeting on a tissue and cellular level would establish their utility in biological

applications.

2

1.1 Hypothesis and objectives

To establish the utility of poly(TMCC-co-LA)-g-PEG nanoparticles in cancer targeting

applications, we hypothesized that:

Poly(TMCC-co-LA)-g-PEG nanoparticle micelles will provide specific anti-cancer drug delivery

at both tissue and cellular levels

To prove this hypothesis, this work was divided into three primary objectives:

1. To investigate the influence of cell injection site on nanocarrier targeting to tumour

xenografts developed in mice.

In Chapter 3, we characterized the pathophysiology of human tumour xenografts

developed in mice after introducing a breast cancer cell line orthotopically (mammary fat

pad) or ectopically (subcutaneous), both of which are common pre-clinical models.

Based on the improved tumour uptake of a model nanocarrier and the underlying vascular

and lymphovascular properties, the orthotopic model was selected for further in vivo

studies.

(2) To demonstrate that poly(TMCC-co-LA)-g-PEG nanoparticles improve pharmacokinetics

and biodistribution over a surfactant-based drug formulation.

In Chapter 4, we loaded a taxol drug, docetaxel, into poly(TMCC-co-LA)-g-PEG

nanoparticles and measured the resulting plasma and tissue levels in orthotopic tumour-

bearing mice. We established that the nanoparticle formulation results in greater blood

circulation and tumour retention over the conventional ethanolic polysorbate 80

formulation. These results suggest potential for improved anti-tumour efficacy based on

extended exposure of cancer cells to a high drug concentration.

(3) To verify that antibody-modified poly(TMCC-co-LA)-g-PEG nanoparticles bind

selectively to cells overexpressing a target surface antigen.

In Chapter 2, we used live breast cancer cells overexpressing human epidermal growth

factor receptor 2 (HER2) to verify receptor-specific binding of nanoparticles modified

with an anti-HER2 antibody (trastuzumab, trade name Herceptin). In Chapter 5, this

3

system was extended to quantify binding strength as a function of antibody conjugation

density. Empirical binding behaviour was consistent with a theoretical model of

monovalent binding (one antibody-antigen interaction per particle), demonstrating that

binding strength can be predicted and controlled.

4

2 Introduction Portions of this chapter are derived from the following manuscript:

Ho KS, Shoichet MS (2013). Design Considerations of Polymeric Nanoparticle Micelles for Targeted Chemotherapeutic Delivery. Current Opinion in Chemical Engineering, 2(1): 53-59.

Reprinted with permission from Elsevier.

Nanoparticle drug delivery systems present exciting opportunities for safer and more effective

anti-cancer drug therapy. By engineering intelligent biomaterials for these applications, it is

possible to develop platform technologies that can be used to target and destroy more cancer

cells, and with greater specificity. Toxic chemotherapeutics given in their free form distribute

broadly throughout the body, but by redirecting more of the drug dose towards tumour sites,

reduced systemic side effects are expected, making anti-cancer treatment safer than conventional

approaches.

2.1 Barriers to conventional anti-cancer drug delivery strategies

Conventional chemotherapy is used to treat cancer using systemic application of toxic

compounds at their maximum tolerated dose, resulting in unacceptable side effects due to their

broad distribution and activity [1-3]. This approach suffers from both the inherent difficulties in

formulating anti-cancer drugs for their free administration [4, 5], and from physical and chemical

barriers to targeting active compounds to tumour tissue [6].

Simple intravenous administration of free chemotherapeutic drugs results in broad

biodistribution because this conventional approach lacks a mechanism for tumour specific

accumulation. Moreover, these toxic compounds commonly exploit rapid cell division or other

poorly selective criteria to obtain selectivity for cancer cells, but their non-specific activity kills

many other healthy cell types (Figure 2.1) [7, 8]. Consequently, patients can experience severe

side effects resulting from toxicity against healthy tissues, restricting maximum drug doses and

treatment efficacy [9].

5

Figure 2.1 Broad drug distribution and activity can cause severe toxicity in healthy tissues. Resulting

side effects are commonly seen in (A) blood (anemia, immune suppression), (B) the digestive tract

(nausea, vomiting), and (C) hair follicles (hair loss).

Additionally, anticancer drugs are often bulky hydrophobic molecules, and to overcome their

poor aqueous solubility they are administered in surfactant-containing formulations.

Unfortunately, these surfactants (e.g., polysorbate 80 and cremophor EL) are each associated

with their own side effects, further limiting the therapeutic window [4, 5]. Their use also triggers

special handling in the clinic, and patient pre-treatment to improve surfactant tolerance during

drug therapy [5].

Degradation remains an important challenge in systemic drug delivery, as free drugs are not

protected from enzymes present in the bloodstream [10]. As a result, drugs may lose activity

before they are delivered to diseased cells. Certain byproducts of these reactions have

themselves been shown to cause non-specific toxicity [11]. Low stability of chemotherapeutic

drugs in blood circulation limits their utility in cancer treatment.

Physical barriers to chemical treatment also apply to solid tumour systems. Due to the increased

metabolic rate associated with cancer cells, they are often in acidic and hypoxic environments;

both efficient chemical killing and sensitization of cells to radiation killing demand high rates of

cell division which cannot be supported by low oxygen levels [6]. Tumour vasculature is also

A   B   C  

6

poorly organized and heterogeneously arranged, leading to areas of poor perfusion [12]. As a

result, mass transport can also be an issue, as drug molecules themselves must cross large

distances to penetrate tumour tissues completely [13]. Increased interstitial fluid pressure is

another factor, further limiting transcapillary fluid flow and convective transport, which can be

particularly restrictive in the case of large molecules like antibodies and protein drugs that rely

on convection more than on diffusion [6]. Furthermore, the majority of anti-cancer drugs require

cellular uptake to reach their sites of action [14].

While problematic, it is important to note that systemic circulation can be utilized to deliver anti-

cancer drugs to sites that have not been defined in advance, such as to metastases or cancer cells

that invaded beyond the margin of surgical resection. However, compound degradation in the

bloodstream and non-specific uptake represent important challenges.

2.2 Targeting the solid tumour microenvironment

Rapid cell growth is not the only distinguishing feature of cancer; other cancer features can offer

completely different approaches to cancer targeting. Abnormal vascular structure and changes in

gene expression are two features of tumour pathophysiology that can be exploited in drug

targeting.

2.2.1 Hyperpermeable tumour vasculature and tissue level accumulation

In healthy tissue, pro-angiogenic and anti-angiogenic factors balance one another to maintain an

organized vascular system [6, 15]. Initially, these balancing stimuli cause tumours to be

restricted in size by their available blood supply. To surpass a critical size, tumours must first

induce an imbalance in vascular regulation and develop new blood vessels to provide nutrients to

rapidly dividing cancer cells [16, 17]. In addition to neovascularization, they can also remodel

existing blood vessels [15, 18]. However, the resulting vasculature associated with tumours is

disorganized, structurally flawed, and immature [6].

Disorganized tumour architecture is characterized by several pathological features that ultimately

lead to large fenestrations in the blood vessel walls [19, 20]. In normal tissue, endothelial cells

are arranged in a monolayer with tight junctions between cells, creating a continuous and

selective barrier that regulates transport of macromolecules [19]. By contrast, tumour

vasculature is comprised of endothelial cells that are poorly aligned, multi-layered, and

7

discontinuous [17]. This structure forms the basis for pathological fenestrations. Consequently,

large molecules that would ordinarily be retained by healthy blood vessels (> 2 nm) have been

shown to cross these gaps (> 100 nm) [21, 22]. Additionally, the pericytes that normally

stabilize the endothelium are often detached or absent, leading to poor control over vessel

permeability and blood flow [18]. The basement membrane may also be abnormally thick or

absent, resulting in changes in mass transport across the extracellular matrix [20].

Further compounding the enhanced transport of fluid and material into tumour tissue, functional

lymphatic drainage is often impaired [16]. As a result, tumours experience inadequate removal

of plasma components and cellular waste material, leading to slow convective material transport,

increased interstitial fluid pressure, and lymphedema [15].

To exploit these features, it is possible to reformulate drugs with macromolecules or nanoscale

drug delivery systems to channel a greater portion of the drug dose through hyperpermeable

vasculature into tumour tissue [17]. While free drugs are small enough to pass through healthy

blood vessels, nanocarriers have improved selectivity for leaky tumour vasculature. This size-

based approach, also called passive targeting, provides a basis for targeting nanoscale materials

to cancer on a tissue level (Figure 2.2). Poor lymphatic drainage in tumours further promotes

nanocarrier retention in the extracellular space [16]. Together these phenomena are called

enhanced permeability and retention (EPR) [17, 23, 24].

To observe these effects in vivo, this underlying pathophysiology will need to be replicated. In

the case of tumour xenograft models, this may be contingent on the host organ environment

selected [25, 26].

8

Figure 2.2 Solid tumour pathophysiology is permissive to active and passive targeting of nanoscale

carriers. Features include disorganized blood vessel architecture, leading to gaps in the blood vessel wall

that allow unregulated transport of macromolecules and other nanoscale materials into tumour tissue.

Nanoparticles modified with antibodies are depicted here selectively extravasating across these gaps.

Certain oncogenes may also be overexpressed as cell surface markers, providing a target for antibody

binding and a means for cellular uptake.

2.2.2 Overexpressed cancer genes and receptor mediated cell uptake

Tumours are comprised of a heterogeneous cell population that includes cancer cells, stromal

cells, and other supporting cell types. Even within the tumour cell population several phenotypes

are present [27]. Tumour cells have subtle differences compared to healthy cells, as certain

genes are turned down (tumour suppressor genes) and others are turned up (oncogenes) [28, 29].

Selected oncogenes result in increased expression of markers on the cancer cell surface. Binding

these markers is called active targeting, and it allows us to target cancer, or potentially

subpopulations within cancer, on a cellular level (Figure 2.2) [30]. Notably, by using a

nanoscale drug delivery vehicle, it becomes possible to access intracellular targets because

nanoparticles are 2-3 orders of magnitude smaller than cells [14]. It also becomes possible to

target specific subpopulations within solid tumours based on phenotypes that are associated with

9

higher instances of relapse [31]. Through these combined approaches, there is potential to

provide better cancer specificity than traditional chemotherapy, which may translate into

improved tumour remission rates and lower toxic side effects for patients.

Overexpression of the human epidermal growth factor receptor 2 (HER2) oncogene characterizes

20-30% of breast cancers [32, 33]. This cell surface marker can be present in levels exceeding

100-fold increases over normal cells [34], and this altered phenotype has important implications

on cell behavior. HER2 is one of a family of receptor tyrosine kinases that can form

homodimers or heterodimers, resulting in cross-phosphorylation and sending a proliferative

signal to the cell [33, 35, 36]. Notably, HER2 does not require activation by a ligand, and is

therefore constitutively active, leading to a continuous proliferative signal [35]. Consequently

this oncogene is associated with chemoresistance, higher metastatic potential, and a poorer

prognosis for patients [31]. By unique association with aggressive disease and metastasis, HER2

cell surface overexpression is an ideal candidate for active targeting strategies [37].

Unlike other potential target receptors that have specificity for a natural ligand, such as folate

and folate receptor, HER2 has no known ligand. To take advantage of this promising target, a

humanized monoclonal antibody, Herceptin (trastuzumab), has been developed to bind the HER2

extracellular domain. Herceptin has been approved for clinical use and derives its activity

through two main mechanisms: inhibition of cell growth and concomitant apoptosis, and

stimulation of natural killer cells through antibody-dependent cellular cytotoxicity (ADCC) [38,

39]. However, Herceptin has only shown limited success as a standalone treatment [32, 39].

Fortunately, when administered with chemotherapeutic agent, the success rate of the

combination therapy increases [32, 40, 41]. This presents an exciting opportunity to combine

nanoscale systems to reformulate anti-cancer drugs for tissue level targeting with the Herceptin

antibody for cellular level targeting.

2.3 Polymeric nanoparticles and design elements in nanomedicine

Polymeric nanoparticles represent one approach to nanomedicine that takes advantage of

hyperpermeable tumour vasculature to improve drug distribution via the EPR effect.

Engineering the composition of amphiphilic copolymers gives control over many aspects of the

resulting micelles that form in aqueous systems [42]. Greater selectivity and concomitant

10

reduction in systemic toxicity create opportunities to broaden the therapeutic window and

improve the clinical outcomes of cancer treatment [3, 8, 43]. By utilizing the bloodstream for

distribution, there is also potential to reach both primary and secondary tumours. The polymer

can also provide protection against non-specific drug uptake and enzyme mediated drug

degradation in the bloodstream [10]. Targeting ligands are also commonly attached to add cell-

specific targeting and receptor-mediated uptake [14, 44, 45]. Polymeric systems promise

flexible chemical modification strategies, simple and tunable self-assembly into ordered

structures, and control over physical properties, all through rational design of their composition

[42].

Nanoparticles have been produced using a variety of biodegradable polymers as the core forming

segment to give a versatile suite of materials for drug delivery. Polyesters (eg. poly(lactic acid)

(PLA)) [46], poly(amino acids) (eg. poly(aspartic acid)) [47], and poly(oxypropylene) (eg.

poloxamers or Pluronics) [48] are amongst the most well studied materials for cancer drug

delivery.

2.3.1 Controlling self-assembly and particle size

Nanoparticle micelles can form spontaneously when amphiphilic block or graft co-polymers are

introduced into aqueous environments [49]. The hydrophobic polymer segments form the

micelle core and have the ability to physically load hydrophobic chemotherapeutic drugs [42].

Their size and shape influence their pharmacokinetics and biodistribution properties.

Nanoparticles under 10 nm can be quickly cleared in capillary beds and lymph nodes, while

those above 200 nm are rapidly removed from circulation via splenic filtration [50]. Size and

shape also impact particle transport, immune recognition, and cell uptake. Indeed, particle

curvature and aspect ratio determine their transport behaviour in the bloodstream [51], and

influence cellular internalization processes [52]. Additionally, discs and rod-shaped nanocarriers

have shown improved blood circulation properties over spherical particles [51, 53, 54], leading

to increased interest in developing drug carriers that circulate a particular geometry and break

into smaller nanocarriers for improved tumour accumulation, penetration, and cell uptake [54,

55].

Size also impacts passive targeting because accumulation in tumour tissue via EPR depends on

extravasation through gaps in hyperpermeable tumour vasculature, putting restrictions on

11

nanoparticle size. While the ideal size range is a topic of debate, a generally accepted range is

50-150 nm. Nonetheless, several ongoing studies argue the utility of nanocarriers outside this

range. Generally, larger nanoparticles can carry greater drug loads because of the larger

available volume for encapsulation [42]. However, unless large nanoparticles are flexible and

easily compressed, they may encounter difficulty crossing tumour vasculature [56]. Another

balancing consideration is tumour penetration, because intratumoural distribution of large

macromolecular assemblies is driven primarily by convection, leaving larger nanoparticles

trapped close to blood vessels [2]. Ultrasmall (<10 nm) gold nanoparticles show more uniform

distribution because they are able to diffuse through tissue [57].

Self-assembly is influenced by several factors, including the respective lengths of the core and

shell forming blocks, which influence the critical micelle concentration, an important measure of

nanoparticle stability [58]. The nanoparticle preparation method also determines the size and

shape of the micelles that form, although the resulting drug loading and micelle structure may be

kinetically unstable [42].

2.3.2 Extending circulation time and drug activity

A long circulation half life is a pre-requisite to tumour accumulation via passive targeting;

multiple passes through hyperpermeable tumour vasculature are required to observe EPR [49, 51,

59], and this means that drug-loaded nanoparticles must be designed to evade rapid drug

degradation and non-specific uptake (Figure 2.3). To accomplish this, the most common strategy

has been to incorporate poly(ethylene glycol) (PEG) as the hydrophilic block in the copolymers

used to prepare nanoparticles. The PEGylation strategy has many benefits, including stabilizing

nanoparticles against aggregation, providing a neutral surface charge, and limiting adsorption of

proteins and opsonins that would invoke clearance by the immune system [60]. Ideally, the

length and density of PEG in each micelle would be adequate to create a brush layer, shielding

the core from interactions with blood proteins [61, 62]. Polymer nanoparticles are well suited to

dense PEGylation because stable incorporation of PEG can be achieved simply by adjusting the

hydrophobic segment length in parallel; PEG incorporation is limited in liposomal systems,

where high PEG-lipid content tends to form small curved micelles instead of stable membrane

structures [62].

12

Figure 2.3 Long plasma circulation times promote passive targeting of nanocarriers via EPR by

increasing the number of passes through hyperpermeable tumour vasculature. PEGylated nanoparticles

are depicted here passing through gaps between disorganized endothelial cells in a tumour blood vessel.

While PEGylation extends the circulation of the nanocarrier, encapsulated drug activity must

also be protected. Metabolic processes are a major concern in conventional drug delivery

strategies where the free drug is in direct contact with blood. When drugs are loaded in the

nanoparticle core, degradation is inhibited by physically preventing enzymes from accessing the

encapsulated material, improving the pharmacokinetic profile. The inverse effect is that the drug

is also physically prevented from accessing cells while circulating, limiting non-specific toxicity.

All of these benefits have potential to increase accumulation and specificity of active drug

compounds at tumour sites.

2.3.3 Increasing drug loading and micelle stability

An important advantage of incorporating drugs into nanoparticles is that the polymer provides a

hydrophobic space to solubilize drug compounds. Typically, poor drug solubility in aqueous

media necessitates their formulation in surfactants and organic solvents, which cause side effects

of their own. Biocompatible polymers that form stable and drug-loaded nanoparticles are

therefore an attractive alternative from a formulations perspective [49, 63].

Drug loading in polymeric nanoparticles has been achieved in many ways, usually falling into

the broad categories of covalent attachment to the polymer, or physical entrapment via

hydrophobic interactions in the nanoparticle core. Considering that nanoparticles may represent

13

less than 1% of the total volume in a colloidal suspension, even less of which corresponds to the

hydrophobic core, high and stable drug loading is important [42]. Polymeric nanoparticles have

potential to attain higher drug loading than liposomes, where lipophilic drugs partition primarily

into the lipid membrane, further restricting the available space [64]. To enhance drug loading,

pH gradients (citrate) have been used to actively load drug compounds through precipitation

[65], and alternative core materials have been co-encapsulated into liposomes [66]. However,

these approaches are restricted to compounds that are relatively hydrophilic. Many promising

drug candidates derive their potency from strong interactions with biological lipid membranes

(cells and cell nuclei), which often are often accompanied by elevated hydrophobicity [49].

One way to ensure stable drug loading is to use covalent attachment to the polymer. This

polymer-drug conjugate method is useful when using a drug that contains easily modified

functional groups, such as free amines, carboxylic acids, or hydroxyl groups [67]. The

corresponding functionality required for attachment can either then be designed into the polymer,

or the drug can first be modified to fit the required chemical linkage already designed into a

particular polymer. This approach has found interesting applications in triggered release, where

the polymer-drug conjugate is a prodrug that is cleaved through a labile linker when exposed to

specific conditions, such as low pH or enzymatic degradation [67]. Drugs or polymers that are

relatively hydrophilic may also be combined using this strategy, where hydrophobic interactions

alone would not effectively keep the drug from partitioning out of the core.

Highly hydrophobic drugs do not require chemical modification to be stably incorporated into a

hydrophobic nanoparticle core. In these cases, the drug may simply be loaded during

nanoparticle preparation and will preferentially partition into the core. This effect may be further

increased by integrating components that encourage greater loading, such as covalently attached

drug molecules to promote stacking during encapsulation. With physical drug encapsulation, it

is critical that the nanoparticles are stable; because their disassembly would trigger immediate

drug release, intact nanoparticles are vital to targeting strategies. Ideally, the polymer being used

would have a low critical micelle concentration (CMC), which confers thermodynamic stability

even under the considerable dilution that occurs immediately on injection into the blood

compartment, which only intensifies with time as the polymer distributes [68]. However, many

polymer systems have high kinetic stability, especially those with high glass transition

temperatures, which means they exhibit slow rates of disassembly even when diluted below their

14

CMC [42]. High stability and drug loading are both important features of polymeric systems that

lead to their utility in targeting applications.

2.3.4 Achieving bioactive surface modification

In addition to chemotherapeutic agents, nanoparticles can be modified with targeting ligands,

such as the native ligand to a receptor [69, 70], receptor antagonists [71], peptides [72],

aptamers [73, 74], and antibodies [75] or their fragments [76]. Targeting ligands may exert their

own therapeutic effects, contributing to treatment efficacy beyond their role in targeting and

specificity. In selecting an appropriate coupling chemistry, the goal is to achieve high coupling

efficiency without sacrificing binding activity and specificity of the ligand [30, 77]. Especially

where chemical modifications are made on assembled nanoparticles, reactions and processing

conditions can disrupt micelle structure or negatively impact drug activity. The required

reagents, potential byproducts, temperature, solvent, and necessary purification steps must all be

given careful consideration. Ideally, the reaction should proceed under mild conditions, in an

aqueous environment, and require minimal post-processing. With desired chemical functional

groups in mind, polymers can be chosen or synthesized to provide platforms for simple surface

modification protocols. By preserving binding activity, selective nanoparticle uptake by a target

cell population is enabled through receptor-mediated endocytosis (Figure 2.4).

Figure 2.4 Active recognition and binding of cancer cells can be achieved using targeting ligands against

an overexpressed oncogene on the cell surface to provide a means of selective cellular uptake. As an

15

example, an antibody-modified nanoparticle is depicted here binding a cell through an antigen-specific

interaction.

2.4 Poly(TMCC-co-LA)-g-PEG graft copolymer for nanoparticle drug delivery

With these broad criteria in mind, poly(TMCC-co-LA)-g-PEG, an amphiphilic graft copolymer,

was previously developed to take advantage of tumour targeting on both a cellular and tissue

level. This novel material has a block-like structure and carboxylic acid groups on the

hydrophobic poly(TMCC-co-LA) backbone and furan groups on the free PEG termini present

versatile options for site-specific modification of the nanoparticle core and surface, respectively.

As a result, targeting ligands can be displayed fully on the nanoparticle surface, maximizing their

presentation to target receptors, while compounds that negatively impact the surface properties

of the nanoparticle can be confined to the core.

2.4.1 Material properties

In the studies that follow, we used a 20 kDa polymer backbone modified with either a 3.4 kDa or

10 kDa PEG graft (Figure 2.5) [78]. The molecular weight and composition of this material

make it biodegradable and bioeliminable, making it an attractive candidate for biological

applications [79]. We have also previously shown that this polymer self-assembles into spherical

nanoparticle micelles of tunable size through a simple dialysis process [78, 80], during which

small molecule hydrophobic agents can be encapsulated in the core. The properties of the

micelles used in this thesis are summarized in Table 2.1.

Table 2.1 Summary of polymer and micelle properties.

    Chapter  4   Chapters  2  and  5      Hydrodynamic  micelle  diameter   83   87   nm  PDI   0.43   0.28      Backbone  molecular  weight   20   20   kDa  Backbone  TMCC  content   13   13   mol  %  PEG  molecular  weight   10   3.4   kDa  Injected  polymer   34   -­‐   mg/kg  Injected  docetaxel   1.5   -­‐   mg/kg  Docetaxel  loading   4.2   -­‐   wt  %    

16

The kinetic stability of nanoparticles of poly(TMCC-co-LA)-g-PEG nanoparticles loaded with

dyes has been shown in biologically relevant media [68]. Furthermore, we can easily control the

respective lengths of the hydrophobic and hydrophilic segments, which have been shown to

affect nanoparticle size, stability, and PEG coverage [42, 58]. We have also demonstrated

control over the composition of the hydrophobic backbone (ratio of TMCC:LA) by controlling

monomer feed ratios [58]. Beyond influencing nanoparticle size and stability, the backbone

composition also determines the hydrophobicity of the nanoparticle core. TMCC is relatively

hydrophilic compared to LA, and so controlling the relative amount of TMCC in the backbone

influences its overall hydrophobicity. Another option is to control the extent of benzyl

deprotection of the TMCC units after polymerization. Manipulating these parameters may

enable the nanoparticle core properties to be adjusted to optimize drug loading according to the

hydrophobicity of the drug compound. While these features are attractive from a design

standpoint, graft copolymers for anti-cancer drug delivery are not well characterized at biological

interfaces. The aims of this thesis were built around the biological assessment of this material as

a targeted drug delivery vehicle to model tumours and cancer cells.

Figure 2.5 Synthesis of poly(TMCC-co-LA)-g-PEG through ring opening polymerization. Highlighted:

hydrophobic backbone (green); PEG graft (grey); furan functional group (red). (Reprinted with permission

from John Wiley and Sons [78])

17

2.4.2 Chemical modification

The application of polymers to targeted drug delivery systems is motivated largely by our ability

to tune their composition easily on the bench so we can engineer and tune their properties. With

poly(TMCC-co-LA)-g-PEG, it is possible to carry out covalent modification easily and

specifically, either on the carboxylic acids present on the hydrophobic backbone, or at the furans

present on the free ends of the hydrophilic graft. This presents opportunities to deliver multiple

agents in a common vehicle, such as combining drug therapies, or consolidating therapy with a

diagnostic contrast agent.

Notably, poly(TMCC-co-LA)-g-PEG was designed with antibody modification in mind.

Antibodies represent a powerful tool for active targeting, as they are highly specific and can bind

their targets strongly and avidly, and can be produced to target a wide variety of epitopes to suit

different applications. However, antibodies are also large biological molecules that are sensitive

to harsh reaction conditions that can lead to denaturation and loss of function. Therefore,

antibody attachment is ideally carried out on prepared nanoparticles to reduce the number of

handling steps, and the reaction must proceed under mild conditions to prevent damage to the

antibody, the carrier and the encapsulated compounds [81]. To accommodate these

requirements, the furan functional group was included in our polymer composition to attach

antibodies to the surface of prepared nanoparticles via Diels-Alder chemistry (Figure 2.6).

Diels-Alder reactions proceed at mild temperatures and pH and in aqueous media without

requiring a catalyst or producing byproducts. The resulting bond is stable under physiological

conditions. An interesting feature of the linker between the PEG chain and the furan group is

that it contains a serum-stable amide bond which can rapidly hydrolyze under acidic conditions

in the presence of hydrolases in lysosomes [82]. As a result, cellular trafficking of the

nanoparticle can be independent of the targeting antibody or other molecules added to the

nanoparticle surface.

18

Figure 2.6 Covalent antibody attachment to prepared nanoparticles via Diels-Alder chemistry. Diels-

Alder coupling was selected to preserve antibody bioactivity because it proceeds under mild reaction

conditions and in aqueous media. (Reprinted with permission from John Wiley and Sons [78])

2.5 Antibody-modified nanoparticles demonstrate antigen-specific activity

Poly(TMCC-co-LA)-g-PEG nanoparticles were designed to conserve antibody activity after

covalent attachment. To verify that binding and specificity were retained in the final

immunonanoparticles, we tested them using live cell assays.

2.5.1 Herceptin immunonanoparticle live cell equilibrium binding

This section contains data from the following manuscript:

Shi M, Wosnick JH, Ho K, Keating A, and Shoichet MS (2007). Immuno-polymeric nanoparticles by Diels-Alder chemistry. Angewandte Chemie International Edition, 46(32): 6126-6131.

Reprinted with permission from John Wiley and Sons.

To demonstrate the utility of Herceptin-modified nanoparticles, we performed a selectivity test

using a series of human cells lines with various levels of HER2 expression: SKBR-3 (HER2

19

overexpression); MCF-7 (normal HER2 expression); and MDA-MB-468 (HER2 negative). By

incubation with fluorescently labeled nanoparticles, equilibrium binding was assessed on live

cells using flow cytometry. High levels of Herceptin immunonanoparticle binding was observed

in the SKBR-3 cell line, with a four-fold increase in fluorescence over the isotype control (Figure

2.7). The same formulation gave a much lower fluorescence signal on incubation with MCF-7

cells, which confirms that binding levels are a function of antigen expression levels. Likewise,

no binding was observed with MDA-MB-468 cells where HER2 was absent.

To verify that these observations resulted from specific antibody-antigen binding events, and not

due to non-specific adsorption from other differences in expression profiles, several controls

were run in parallel. These controls demonstrated little binding in absence of Herceptin

(unmodified nanoparticles, and IgG1κ isotype control) or when HER2 receptors were blocked by

pre-incubation with free Herceptin (Figure 2.7). Based on these results, Herceptin

immunonanoparticles conserved binding activity and selectivity of Herceptin for the HER2

antigen, demonstrating their utility in active targeting applications.

20

Figure 2.7 Flow cytometry analysis of Herceptin immunonanoparticle binding to cell lines at equilibrium.

SKBR-3 (HER2 overexpressing), MCF-7 (normal HER2 expression), and MDA-MB-468 (HER2 negative)

cell lines were incubated with fluorescently labeled nanoparticles to confirm selectivity of Herceptin

immunonanoparticles for HER2 overexpressing cells, and specific antibody-antigen interactions versus a

series of controls for non-specific adsorption.

2.5.2 Herceptin immunonanoparticle cell cytotoxicity

This section contains data from the following manuscript:

Shi M, Ho K, Keating A, and Shoichet MS (2009). Doxorubicin-conjugated immuno-nanoparticles for intracellular anticancer drug delivery. Advanced Functional Materials, 19(11): 1689-1696.

Reprinted with permission from John Wiley and Sons.

To validate the use of Herceptin immunonanoparticles as drug delivery vehicles, we sequentially

modified assembled nanoparticles via Diels-Alder chemistry: first with Herceptin as a targeting

ligand, then with doxorubicin (DOX) as a cytotoxic drug compound. These Herceptin-DOX-

nanoparticles (NP-aHER2-DOX) were then used to treat HER2 overexpressing SKBR-3 cells.

21

Using confocal microscopy, their uptake was observed versus unmodified nanoparticles, and

nanoparticles modified with only Herceptin (NP-aHER2) or only DOX (NP-DOX). DOX-

containing particles were visualized via red DOX autofluorescence, and the remaining

formulations were modified with green Alexa-fluor 488. Unmodified nanoparticles had no

mechanism for active cellular uptake, and as expected, no membrane binding or intracellular

accumulation was observed (Figure 2.8A). The remaining formulations were surface-modified

with DOX and/or Herceptin, giving rise to substantial intracellular accumulation within 6 h of

incubation with SKBR-3 cells. These results verified that surface modification with the

Herceptin antibody conferred not only specific binding activity to our nanoparticles, but also

selectivity in uptake by inducing receptor mediated endocytosis (Figure 2.8B). Interestingly, the

placement of DOX on the nanoparticle surface also conferred non-specific uptake to DOX

nanoparticles (Figure 2.8C), likely via hydrophobic interactions with the cell membrane that the

free compound also known to exert. Fluorescence was also observed when both Herceptin and

DOX were present on the nanoparticle surface (Figure 2.8D). Moreover, the average

fluorescence intensity associated with NP-aHER2-DOX was greater than NP-DOX, suggesting

that Herceptin provides a means to direct greater amounts of DOX into HER2 positive cells via

targeting (Figure 2.9).

Figure 2.8 Confocal images showing cellular uptake of various nanoparticle formulations in SKBR-3 cells

after 6 h of incubation at 37 °C. Blue indicates cell nuclei (A-D), green indicates fluorescently labeled

polymer (A-B), and red indicates DOX autofluorescence (C-D). (A) shows unmodified nanoparticles,

where no uptake or binding is observed. (B) depicts Herceptin nanoparticles, where binding and

A Unmodified NP B NP-aHER2 C NP-DOX D NP-aHER2-DOX

22

receptor-mediated uptake are both observed. (C) shows that DOX nanoparticles are also internalized

and accumulate intracellularly via hydrophobic interactions between DOX and the cell membrane. (D)

illustrates that even greater uptake is observed with Herceptin-DOX-nanoparticles, as both uptake

mechanisms are present.

Figure 2.9 Average fluorescence measurements from confocal images of DOX-conjugated nanoparticles

inside SKBR-3 cells after 6 h of incubation at 37 °C. These data show that while both formulations

undergo rapid internalization, Herceptin enhances cell uptake.

To test whether the increased uptake of NP-aHER2-DOX translated into greater reductions in

cell number, we compared them to NP-DOX and free DOX using a colourimetric assay of live

cell metabolism of a tetrazolium salt (MTS assay) after a 72 h treatment period. To investigate

cell selectivity, we tested these formulations using both target cells (SKBR-3) and healthy

endothelial cells (HMEC-1). NP-aHER2-DOX induced greater reductions in viable cell number

in the target cell population than in the healthy cell population (Figure 2.10); this measure of

effective cytotoxicity may reflect a combination of cytotoxic and cytostatic effects. This

formulation similarly outperformed NP-DOX, indicating that Herceptin modification enhances

the effective cytotoxicity. The most remarkable difference in formulations is that although free

DOX produced the greatest cell number reduction, only the nanoparticle formulations

demonstrated selectivity for the target cell population after 72 h.

23

Figure 2.10 Cell growth relative to controls for DOX treated SKBR-3 breast cancer cells versus HMEC-1

healthy endothelial cells after 72 h of treatment at 5.0 µg/mL DOX or DOX equivalent (IC50 of free DOX

against SKBR-3 cells). This measure of effective cytotoxicity reflects a combination of cytotoxic and

cytostatic effects.

To elucidate the mechanism of viable cell reduction with our nanoparticle formulations, we used

flow cytometry to assess treated cells stained for propidium iodide (PI) to label dead cells, and

Annexin V to mark apoptotic cells. In this way we categorized cell number reductions via the

following types of observations: (1) treatments where cells stalled in the cell cycle remain

distributed similarly to controls (mainly PI and Annexin V negative); (2) treatments that

triggered early apoptosis (PI negative and Annexin V positive); (3) cells that are in late apoptosis

(PI and Annexin V positive); or cells undergoing necrosis (PI positive and Annexin V negative).

After 24 h of incubation with each formulation, the percentage of the cell population represented

by each category is listed in Table 2.2. Both Herceptin and DOX retained their original

mechanism of action when conjugated to nanoparticles alone: Herceptin reduces viable cell

counts through a cytostatic mechanism after binding HER2, whereas DOX causes apoptosis after

cell uptake and nuclear transport. However, when both are conjugated to a common particle,

NP-aHER-DOX demonstrated an even greater percentage of cells undergoing apoptosis than NP-

DOX. This observation suggests that although the improved effective cytotoxicity of NP-

24

aHER2-DOX may have been enhanced by a cytostatic mechanism of action, increased DOX

uptake via HER2 mediated endocytosis also led to synergistic levels of apoptotic cell death.

Table 2.2 Flow cytometry data showing apoptotic mechanism of cell death for SKBR-3 cells treated with

1.75 µg/mL DOX or DOX equivalent after 24 hours of incubation.

    Percentage  of  cell  population  Treatment   Live   Early  apoptotic   Late  apoptotic   Necrotic  

Untreated  cells   84.5   8.9   3.6   3.0  Free  aHER2   84.2   4.6   8.1   3.1  Free  DOX   60.9   12.7   19.5   6.9  NP-­‐aHER2   85.6   4.2   7.2   3.0  NP-­‐DOX   77.5   7.2   8.8   6.5  NP-­‐aHER2-­‐DOX   68.8   9.5   15.4   6.3  

2.6 Conclusions

Broad activity and distribution of chemotherapeutic drugs administered in their conventional free

forms has unacceptable side effects on healthy tissues. Using nanomedicine to reformulate anti-

cancer agents in targeted delivery platforms has previously shown utility in drug delivery

applications. Polymers offer tunable composition and properties, and nanoparticles prepared

from poly(TMCC-co-LA)-g-PEG have shown promise as kinetically stable structures that have

demonstrated receptor-specific binding, uptake, and cytotoxicity when modified with an

antibody for active targeting. This versatile platform is expected to be useful in biological

applications while providing simple methods for site-specific covalent modifications.

To advance our understanding of our material, we tested poly(TMCC-co-LA)-g-PEG

nanoparticles at a variety of biological interfaces that are relevant to their utility in drug targeting

applications. In the studies that follow, we challenged our nanoparticles to deliver a small

molecule drug in a relevant mouse model of breast cancer and proved that our material improved

the retention of a drug compound in tumour tissue through passive targeting. Beyond simple

confirmation of selective cell targeting, we also demonstrated that it is possible to predict and

tune binding strength using a mathematical model that we validated empirically. These studies

justify further pre-clinical evaluation of our system as a candidate for targeted drug formulations.

25

2.7 Thesis scope

To assess the ability of poly(TMCC-co-LA)-g-PEG nanoparticles to target cancer on both tissue

and cellular levels, this thesis was divided into three sections:

1. Animal models of disease are essential to the pre-clinical evaluation of treatment

strategies, but to be effective they must capture the underlying pathophysiology of human

diseases. Human tumour xenografts represent the most widely used model of solid

cancers because they are simple to replicate. To be useful in evaluating drug distribution

in the context of nanomedicine, these tumour models must demonstrate the

hyperpermeable vasculature and insufficient lymphatic drainage that enable EPR. In

Chapter 3, accumulation of a model nanocarrier and immunostaining of vascular and

lymphovascular structures were used to compare tumours that developed after cell

injection into ectopic (subcutaneous) and orthotopic (mammary fat pad) sites. This study

was used to select orthotopic tumour xenografts as the more relevant model for

subsequent in vivo testing.

2. The pharmacokinetics and biodistribution of drug formulations illustrate the transport and

metabolism of drug compounds after administration. This information can then be used

to determine dosage, scheduling, areas for improvement, and potential for systemic

toxicity. In Chapter 4, we encapsulated a taxol drug, docetaxel, in the core of

poly(TMCC-co-LA)-g-PEG nanoparticles to evaluate pharmacokinetic parameters and

measure tissue uptake in orthotopic tumour-bearing mice. After intravenous injection,

our nanoparticles were shown to prolong the plasma circulation time of docetaxel and to

extend docetaxel retention at the tumour site when compared to the conventional

ethanolic polysorbate 80 formulation. As a result, our nanoparticle formulation is likely

to enhance efficacy at matched dosages, as cytotoxicity depends both on concentration

and exposure time. These data validated the utility of our polymeric nanoparticles in

passive tumour targeting applications.

3. In introductory work for this thesis, we verified that poly(TMCC-co-LA)-g-PEG

nanoparticles could be successfully modified with Herceptin to introduce selective

binding and uptake in live HER2 overexpressing cells. In Chapter 5, we were interested

in tuning the strength of binding between these immunonanoparticles and cells because

26

this behaviour may influence cell uptake and tumour penetration. We confirmed that

nanoparticle binding strength is a function of antibody conjugation density by preparing a

series of nanoparticles having different average numbers of Herceptin antibodies per

particle. By modeling the resulting binding isotherms and calculating the equilibrium

binding constants, Keq, for each formulation, we found that Keq increased in a manner

consistent with monovalent binding (one antibody-antigen interaction per particle). This

theoretical model, which is broadly applicable to targeted nanoparticle systems, enables

binding strength of our Herceptin-modified nanoparticles to be adjusted in a predictive

manner.

In this work, we validated poly(TMCC-co-LA)-g-PEG nanoparticles as a platform capable of

targeted delivery to cancer on both a tissue and cellular level. This forms a compelling

justification for further pre-clinical evaluation of our system for safety and efficacy in vivo.

2.8 References

[1] Arias JL. Drug Targeting Strategies in Cancer Treatment: An Overview. Mini-Rev Med Chem 2011;11:1-17.

[2] Danhier F, Feron O, Preat V. To exploit the tumor microenvironment: Passive and active tumor targeting of nanocarriers for anti-cancer drug delivery. J Control Release 2010;148:135-146.

[3] De Souza R, Zahedi P, Allen CJ, Piquette-Miller M. Polymeric drug delivery systems for localized cancer chemotherapy. Drug Deliv 2010;17:365-375.

[4] Engels FK, Mathot RAA, Verweij J. Alternative drug formulations of docetaxel: a review. Anti-Cancer Drug 2007;18:95-103.

[5] Hennenfent KL, Govindan R. Novel formulations of taxanes: a review. Old wine in a new bottle? Ann Oncol 2006;17:735-749.

[6] Cairns R, Papandreou I, Denko N. Overcoming physiologic barriers to cancer treatment by molecularly targeting the tumor microenvironment. Mol Cancer Res 2006;4:61-70.

[7] Jordan MA, Wilson L. Microtubules as a target for anticancer drugs. Nat Rev Cancer 2004;4:253-265.

[8] Torchilin VP. Drug targeting. Eur J Pharm Sci 2000;11:S81-S91.

[9] Felici A, Verweij J, Sparreboom A. Dosing strategies for anticancer drugs: the good, the bad and body-surface area. Eur J Cancer 2002;38:1677-1684.

27

[10] Panyam J, Labhasetwar V. Biodegradable nanoparticles for drug and gene delivery to cells and tissue. Adv Drug Deliver Rev 2003;55:329-347.

[11] Salvatorelli E, Menna P, Cascegna S, Liberi G, Calafiore AM, Gianni L, et al. Paclitaxel and docetaxel stimulation of doxorubicinol formation in the human heart: Implications for cardiotoxicity of doxorubicin-taxane chemotherapies. J Pharmacol Exp Ther 2006;318:424-433.

[12] Jain RK, Stylianopoulos T. Delivering nanomedicine to solid tumors. Nat Rev Clin Oncol 2010;7:653-664.

[13] Minchinton AI, Tannock IF. Drug penetration in solid tumours. Nat Rev Cancer 2006;6:583-592.

[14] Panyam J, Labhasetwar V. Targeting intracellular targets. Curr Drug Deliv 2004;1:235-247.

[15] Kerbel RS. Tumor angiogenesis: past, present and the near future. Carcinogenesis 2000;21:505-515.

[16] Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249-257.

[17] Iyer AK, Khaled G, Fang J, Maeda H. Exploiting the enhanced permeability and retention effect for tumor targeting. Drug Discov Today 2006;11:812-818.

[18] Morikawa S, Baluk P, Kaidoh T, Haskell A, Jain RK, McDonald DM. Abnormalities in pericytes on blood vessels and endothelial sprouts in tumors. Am J Pathol 2002;160:985-1000.

[19] Dreher MR, Liu WG, Michelich CR, Dewhirst MW, Yuan F, Chilkoti A. Tumor vascular permeability, accumulation, and penetration of macromolecular drug carriers. J Natl Cancer I 2006;98:335-344.

[20] Jain RK. Normalization of tumor vasculature: An emerging concept in antiangiogenic therapy. Science 2005;307:58-62.

[21] Bae KH, Chung HJ, Park TG. Nanomaterials for cancer therapy and imaging. Mol Cells 2011;31:295-302.

[22] Hashizume H, Baluk P, Morikawa S, McLean JW, Thurston G, Roberge S, et al. Openings between defective endothelial cells explain tumor vessel leakiness. Am J Pathol 2000;156:1363-1380.

[23] 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. Adv Drug Deliver Rev 2011;63:136-151.

[24] Matsumura Y, Maeda H. A New Concept for Macromolecular Therapeutics in Cancer-Chemotherapy - Mechanism of Tumoritropic Accumulation of Proteins and the Antitumor Agent Smancs. Cancer Res 1986;46:6387-6392.

28

[25] Kerbel RS, Cornil I, Theodorescu D. Importance of Orthotopic Transplantation Procedures in Assessing the Effects of Transfected Genes on Human Tumor-Growth and Metastasis. Cancer Metast Rev 1991;10:201-215.

[26] Killion JJ, Radinsky R, Fidler IJ. Orthotopic models are necessary to predict therapy of transplantable tumors in mice. Cancer Metast Rev 1998;17:279-284.

[27] Evan GI, Vousden KH. Proliferation, cell cycle and apoptosis in cancer. Nature 2001;411:342-348.

[28] Hanahan D, Weinberg RA. The hallmarks of cancer. Cell 2000;100:57-70.

[29] Hanahan D, Weinberg RA. Hallmarks of Cancer: The Next Generation. Cell 2011;144:646-674.

[30] Marcucci F, Lefoulon F. Active targeting with particulate drug carriers in tumor therapy: fundamentals and recent progress. Drug Discov Today 2004;9:219-228.

[31] Ross JS, Fletcher JA. The HER-2/neu oncogene in breast cancer: Prognostic factor, predictive factor, and target for therapy. Stem Cells 1998;16:413-428.

[32] Nahta R, Esteva FJ. Herceptin: mechanisms of action and resistance. Cancer Lett 2006;232:123-138.

[33] Witton CJ, Reeves JR, Going JJ, Cooke TG, Bartlett JM. Expression of the HER1-4 family of receptor tyrosine kinases in breast cancer. J Pathol 2003;200:290-297.

[34] Ross JS, Fletcher JA, Bloom KJ, Linette GP, Stec J, Symmans WF, et al. Targeted therapy in breast cancer - The HER-2/neu gene and protein. Mol Cell Proteomics 2004;3:379-398.

[35] Hudelist G, Kostler WJ, Attems J, Czerwenka K, Muller R, Manavi M, et al. Her-2/neu-triggered intracellular tyrosine kinase activation: in vivo relevance of ligand-independent activation mechanisms and impact upon the efficacy of trastuzumab-based treatment. Brit J Cancer 2003;89:983-991.

[36] Grant S, Qiao L, Dent P. Roles of ERBB family receptor tyrosine kinases, and downstream signaling pathways, in the control of cell growth and survival. Front Biosci 2002;7:D376-D389.

[37] Wang SC, Hung MC. HER2 overexpression and cancer targeting. Semin Oncol 2001;28:115-124.

[38] Yan L, Hsu K, Beckman RA. Antibody-based therapy for solid tumors. Cancer J 2008;14:178-183.

[39] Baselga J, Norton L, Albanell J, Kim YM, Mendelsohn J. Recombinant humanized anti-HER2 antibody (Herceptin (TM)) enhances the antitumor activity of paclitaxel and doxorubicin against HER2/neu overexpressing human breast cancer xenografts. Cancer Res 1998;58:2825-2831.

29

[40] Lee S, Yang WT, Lan KH, Sellappan S, Klos K, Hortobagyi G, et al. Enhanced sensitization to Taxol-induced apoptosis by Herceptin pretreatment in ErbB2-overexpressing breast cancer cells. Cancer Res 2002;62:5703-5710.

[41] Leyland-Jones B. Trastuzumab: hopes and realities. Lancet Oncol 2002;3:137-144.

[42] Allen C, Maysinger D, Eisenberg A. Nano-engineering block copolymer aggregates for drug delivery. Colloid Surface B 1999;16:3-27.

[43] Wolinsky JB, Colson YL, Grinstaff MW. Local drug delivery strategies for cancer treatment: Gels, nanoparticles, polymeric films, rods, and wafers. J Control Release 2012;159:14-26.

[44] Gao Y, Chen LL, Gu WW, Xi Y, Lin LP, Li YP. Targeted Nanoassembly Loaded with Docetaxel Improves Intracellular Drug Delivery and Efficacy in Murine Breast Cancer Model. Mol Pharmaceut 2008;5:1044-1054.

[45] Shi M, Ho K, Keating A, Shoichet MS. Doxorubicin-Conjugated lmmuno-Nanoparticles for Intracellular Anticancer Drug Delivery. Adv Funct Mater 2009;19:1689-1696.

[46] Allemann E, Brasseur N, Benrezzak O, Rousseau J, Kudrevich SV, Boyle RW, et al. Peg-Coated Poly(Lactic Acid) Nanoparticles for the Delivery of Hexadecafluoro Zinc Phthalocyanine to Emt-6 Mouse Mammary-Tumors. J Pharm Pharmacol 1995;47:382-387.

[47] Yokoyama M, Miyauchi M, Yamada N, Okano T, Sakurai Y, Kataoka K, et al. Characterization and anticancer activity of the micelle-forming polymeric anticancer drug adriamycin-conjugated poly(ethylene glycol)-poly(aspartic acid) block copolymer. Cancer Res 1990;50:1693-1700.

[48] Kabanov AV, Batrakova EV, Meliknubarov NS, Fedoseev NA, Dorodnich TY, Alakhov VY, et al. A New Class of Drug Carriers - Micelles of Poly(Oxyethylene)-Poly(Oxypropylene) Block Copolymers as Microcontainers for Drug Targeting from Blood in Brain. J Control Release 1992;22:141-157.

[49] Torchilin VP. Targeted polymeric micelles for delivery of poorly soluble drugs. Cell Mol Life Sci 2004;61:2549-2559.

[50] Champion JA, Katare YK, Mitragotri S. Particle shape: A new design parameter for micro- and nanoscale drug delivery carriers. J Control Release 2007;121:3-9.

[51] Caldorera-Moore M, Guimard N, Shi L, Roy K. Designer nanoparticles: incorporating size, shape and triggered release into nanoscale drug carriers. Expert Opin Drug Del 2010;7:479-495.

[52] Jiang W, Kim BYS, Rutka JT, Chan WCW. Nanoparticle-mediated cellular response is size-dependent. Nature Nanotechnology 2008;3:145-150.

[53] Ferrari M. Nanogeometry: beyond drug delivery. Nat Nanotechnol 2008;3:131-132.

30

[54] Geng Y, Dalhaimer P, Cai SS, Tsai R, Tewari M, Minko T, et al. Shape effects of filaments versus spherical particles in flow and drug delivery. Nature Nanotechnology 2007;2:249-255.

[55] Wong C, Stylianopoulos T, Cui JA, Martin J, Chauhan VP, Jiang W, et al. Multistage nanoparticle delivery system for deep penetration into tumor tissue. P Natl Acad Sci USA 2011;108:2426-2431.

[56] Moghimi SM, Hunter AC, Murray JC. Long-circulating and target-specific nanoparticles: Theory to practice. Pharmacol Rev 2001;53:283-318.

[57] Huang KY, Ma HL, Liu J, Huo SD, Kumar A, Wei T, et al. Size-Dependent Localization and Penetration of Ultrasmall Gold Nanoparticles in Cancer Cells, Multicellular Spheroids, and Tumors in Vivo. Acs Nano 2012;6:4483-4493.

[58] Lu J, Shoichet MS. Self-Assembled Polymeric Nanoparticles of Organocatalytic Copolymerizated D,L-Lactide and 2-Methyl 2-Carboxytrimethylene Carbonate. Macromolecules 2010;43:4943-4953.

[59] Khalid MN, Simard P, Hoarau D, Dragomir A, Leroux JC. Long circulating poly(ethylene glycol)-decorated lipid nanocapsules deliver docetaxel to solid tumors. Pharmaceut Res 2006;23:752-758.

[60] Nicholas AR, Scott MJ, Kennedy NI, Jones MN. Effect of grafted polyethylene glycol (PEG) on the size, encapsulation efficiency and permeability of vesicles. Bba-Biomembranes 2000;1463:167-178.

[61] Photos PJ, Bacakova L, Discher B, Bates FS, Discher DE. Polymer vesicles in vivo: correlations with PEG molecular weight. J Control Release 2003;90:323-334.

[62] Tirosh O, Barenholz Y, Katzhendler J, Priev A. Hydration of polyethylene glycol-grafted liposomes. Biophys J 1998;74:1371-1379.

[63] Gaucher G, Dufresne MH, Sant VP, Kang N, Maysinger D, Leroux JC. Block copolymer micelles: preparation, characterization and application in drug delivery. J Control Release 2005;109:169-188.

[64] Kita K, Dittrich C. Drug delivery vehicles with improved encapsulation efficiency: taking advantage of specific drug-carrier interactions. Expert Opin Drug Del 2011;8:329-342.

[65] Gubernator J. Active methods of drug loading into liposomes: recent strategies for stable drug entrapment and increased in vivo activity. Expert Opin Drug Del 2011;8:565-580.

[66] Ashley CE, Carnes EC, Phillips GK, Padilla D, Durfee PN, Brown PA, et al. The targeted delivery of multicomponent cargos to cancer cells by nanoporous particle-supported lipid bilayers. Nat Mater 2011;10:389-397.

[67] Mahato R, Tai WY, Cheng K. Prodrugs for improving tumor targetability and efficiency. Adv Drug Deliver Rev 2011;63:659-670.

31

[68] Lu J, Owen SC, Shoichet MS. Stability of Self-Assembled Polymeric Micelles in Serum. Macromolecules 2011;44:6002-6008.

[69] Hilgenbrink AR, Low PS. Folate receptor-mediated drug targeting: From therapeutics to diagnostics. J Pharm Sci-Us 2005;94:2135-2146.

[70] Li X, Tian X, Zhang J, Zhao X, Chen XH, Jiang YH, et al. In vitro and in vivo evaluation of folate receptor-targeting amphiphilic copolymer-modified liposomes loaded with docetaxel. Int J Nanomed 2011;6:1167-1184.

[71] Kiziltepe T, Ashley JD, Stefanick JF, Qi YM, Alves NJ, Handlogten MW, et al. Rationally engineered nanoparticles target multiple myeloma cells, overcome cell-adhesion-mediated drug resistance, and show enhanced efficacy in vivo. Blood Cancer Journal 2012;2:e64.

[72] Yao NH, Xiao WW, Wang XB, Marik J, Park SH, Takada Y, et al. Discovery of Targeting Ligands for Breast Cancer Cells Using the One-Bead One-Compound Combinatorial Method. J Med Chem 2009;52:126-133.

[73] Berezovski MV, Lechmann M, Musheev MU, Mak TW, Krylov SN. Aptamer-facilitated biomarker discovery (AptaBiD). J Am Chem Soc 2008;130:9137-9143.

[74] Fan Z, Shelton M, Singh AK, Senapati D, Khan SA, Ray PC. Multifunctional Plasmonic Shell-Magnetic Core Nanoparticles for Targeted Diagnostics, Isolation, and Photothermal Destruction of Tumor Cells. Acs Nano 2012;6:1065-1073.

[75] Kocbek P, Obermajer N, Cegnar M, Kos J, Kristl J. Targeting cancer cells using PLGA nanoparticles surface modified with monoclonal antibody. J Control Release 2007;120:18-26.

[76] Tang Y, Wang J, Scollard DA, Mondal H, Holloway C, Kahn HJ, et al. Imaging of HER2/neu-positive BT-474 human breast cancer xenografts in athymic mice using In-111-trastuzumab (Herceptin) Fab fragments. Nucl Med Biol 2005;32:51-58.

[77] Shi M, Lu J, Shoichet MS. Organic nanoscale drug carriers coupled with ligands for targeted drug delivery in cancer. J Mater Chem 2009;19:5485-5498.

[78] Shi M, Wosnick JH, Ho K, Keating A, Shoichet MS. Immuno-polymeric nanoparticles by Diels-Alder chemistry. Angew Chem Int Edit 2007;46:6126-6131.

[79] Etrych T, Kovar L, Subr V, Braunova A, Pechar M, Chytil P, et al. High-molecular-weight Polymers Containing Biodegradable Disulfide Bonds: Synthesis and In Vitro Verification of Intracellular Degradation. J Bioact Compat Pol 2010;25:5-26.

[80] Shi M, Shoichet MS. Furan-functionalized co-polymers for targeted drug delivery: caracterization, self-assembly and drug encapsulation. Journal of Biomaterials Science, Polymer Edition 2008;19:1143-1157.

[81] Nobs L, Buchegger F, Gurny R, Allemann E. Current methods for attaching targeting ligands to liposomes and nanoparticles. J Pharm Sci-Us 2004;93:1980-1992.

32

[82] Trouet A, Masquelier M, Baurain R, Deprezdecampeneere D. A Covalent Linkage between Daunorubicin and Proteins That Is Stable in Serum and Reversible by Lysosomal Hydrolases, as Required for a Lysosomotropic Drug-Carrier Conjugate - Invitro and Invivo Studies. P Natl Acad Sci-Biol 1982;79:626-629.

33

3 Pathophysiological assessment of human tumour xenografts as models of EPR in breast cancer

This chapter is derived from the following manuscript:

Ho KS, Poon PC, Owen SC, and Shoichet MS (2013) Blood vessel hyperpermeability and pathophysiology in human tumour xenograft models of breast cancer: a comparison of ectopic and orthotopic tumours. BMC Cancer, 12: 579.

Reprinted with permission from BioMed Central.

3.1 Abstract

Background Human tumour xenografts in immune compromised mice are widely used as cancer

models because they are easy to reproduce and simple to use in a variety of pre-clinical

assessments. Developments in nanomedicine have led to the use of tumour xenografts in testing

nanoscale delivery devices, such as nanoparticles and polymer-drug conjugates, for targeting and

efficacy via the enhanced permeability and retention (EPR) effect. For these results to be

meaningful, the hyperpermeable vasculature and reduced lymphatic drainage associated with

tumour pathophysiology must be replicated in the model. In pre-clinical breast cancer xenograft

models, cells are commonly introduced via injection either orthotopically (mammary fat pad,

MFP) or ectopically (subcutaneous, SC), and the organ environment experienced by the tumour

cells has been shown to influence their behaviour. Methods To evaluate xenograft models of

breast cancer in the context of EPR, both orthotopic MFP and ectopic SC injections of MDA-

MB-231-H2N cells were given to NOD scid gamma (NSG) mice. Animals with matched

tumours in two size categories were tested by injection of a high molecular weight dextran as a

model nanocarrier. Tumours were collected and sectioned to assess dextran accumulation

compared to liver tissue as a positive control. To understand the cellular basis of these

observations, tumour sections were also immunostained for endothelial cells, basement

membranes, pericytes, and lymphatic vessels. Results SC tumours required longer development

times to become size matched to MFP tumours, and also presented wide size variability and

ulcerated skin lesions 6 weeks after cell injection. The 3 week MFP tumour model demonstrated

greater dextran accumulation than the size matched 5 week SC tumour model (for P < 0.10).

Immunostaining revealed greater vascular density and thinner basement membranes in the MFP

tumour model 3 weeks after cell injection. Both the MFP and SC tumours showed evidence of

34

insufficient lymphatic drainage, as many fluid-filled and collagen IV-lined spaces were

observed, which likely contain excess interstitial fluid. Conclusions Dextran accumulation and

immunostaining results suggest that small MFP tumours best replicate the vascular permeability

required to observe the EPR effect in vivo. A more predictable growth profile and the absence of

ulcerated skin lesions further point to the MFP model as a strong choice for long term treatment

studies that initiate after a target tumour size has been reached.

Keywords

Tumour xenograft models, orthotopic transplantation, ectopic transplantation, enhanced

permeability and retention, breast cancer, blood vessel hyperpermeability, nanomedicine,

targeting

3.2 Background

Pre-clinical development of anti-cancer therapeutics relies on availability of relevant and

reproducible in vivo tumour models. Human tumour xenograft models in immunodeficient mice

are widely used to assess pharmacokinetics, biodistribution, and treatment efficacy because they

are inexpensive and easy to replicate [1]. However, their utility in evaluating potential treatment

strategies depends on their capacity to recapitulate human disease conditions.

Progress in nanomedicine seeks to shift distribution of therapeutic compounds to tumour tissue

by targeting hyperpermeable tumour vasculature [2, 3]. Tumours are restricted in size until they

can trigger greater blood vessel density through angiogenesis and blood vessel remodeling [4, 5].

Compared to normal tissue, tumour tissue has been demonstrated to be more permissive to

extravasation of macromolecules as a result of abnormal blood vessel structure [3]. Moreover,

tumour tissue is subject to poor lymphatic drainage, leading to greater retention of material in the

extravascular space. These combined phenomena are called enhanced permeability and retention

(EPR) and form the basis for improved selectivity of nanoscale drug delivery for solid tumour

targeting [2, 4, 6].

Several pathological features of tumour vasculature lead to its utility in targeting applications.

Pathological tumour vessels are dynamic, and can result both from angiogenesis and remodeling

of existing vessels [5, 7]. Endothelial cells that comprise tumour blood vessels have poor

35

organization, leading to gaps between cells, multiple endothelial cell layers, and unusual

tortuosity and branching [8, 9]. These openings allow unregulated movement of macromolecules

and nanoscale carriers across tumour vessel walls and into the surrounding tissue [10]. In

response, the associated basement membrane is also often thickened or absent [9, 11]. This

apparent dichotomy stems from a dynamic interaction between increased and multilayered

collagen deposition in the basement membrane [10, 12-14] and increased expression of matrix

metalloproteinases (MMPs) that can result in collagen degradation [15]. Further enhancing the

aberrant permeability of tumour blood vessels, the pericytes that normally cover and stabilize the

outer vessel wall can also be missing or detached, leading to a more immature vessel structure

[5]. The absence of these contractile support cells may lead to further increased vessel

permeability and weakened control over blood flow [7]. Lymphatic vessels are closely

associated with and derived from the blood vessel network. They are responsible for

transporting waste out of tissues, but tumours are often deficient in lymphatic drainage, leading

to increased accumulation of macromolecular material in tumour tissue [4]. Each of these

features contributes to the pathophysiology that enables the EPR effect.

To validate the use of human tumour xenografts in mouse models of breast cancer to investigate

tumour targeting via EPR, we studied MDA-MB-231-H2N cells transplanted in NOD scid

gamma (NSG) mice and compared two common cell injection sites in the context of EPR

permissive pathology. Owing to its simplicity, tumour cells are often introduced ectopically as

subcutaneous (SC) injections, regardless of their native tissue type [16-18]. Cells injected

orthotopically (eg. breast cancer cells into mammary tissue) are subject to biological cues present

in the relevant organ environment [17]. Allowing tumour cells to grow in their orthotopic

environment influences growth rate, blood and lymphatic vessel development, metastatic

potential, interstitial pressure, and response to therapy [5, 17-20]. We hypothesized that the

orthotopic environment may also influence the permeability of the resulting tumour vasculature.

Groups of animals were compared as cohorts of matching tumour size because size, and not

elapsed time, is a standard prognostic measure used to assess breast cancer stage [21].

Currently, the benefit of using either SC or MFP in xenograft models of breast cancer in

assessing targeting through the EPR effect is not well characterized. To investigate vessel

permeability, orthotopic and ectopic tumour-bearing NSG mice were given intravenous

injections of a fluorescently labeled high molecular weight dextran (FITC-Dextran, 2 MDa, ~80

36

nm [8]) as a model nanocarrier. After allowing the dextran to circulate, animals were sacrificed,

their tumours removed, measured using calipers, and fixed with paraformaldehyde. Tumours

were cryosectioned and examined for dextran accumulation. Tissue sections were also

immunostained for markers of vascular endothelial cells (CD31), basement membrane (collagen

IV), pericytes (alpha smooth muscle actin (αSMA)), and lymphatic vessels (lymphatic vessel

endothelial hyaluronan receptor (LYVE-1)).

3.3 Methods

3.3.1 Materials

All cell culture materials were purchased from Gibco-Invitrogen (Burlington, ON, Canada).

MDA-MB-231-H2N cells and NOD scid gamma (NSG) mice were generous gifts from Dr.

Robert Kerbel (Sunnybrook Research Institute, Toronto, ON, Canada), which were then

maintained or bred in-house. Lysine-fixable dextran-FITC (MW 2 MDa) was purchased from

Invitrogen (Burlington, ON, Canada). Slides and cover slips were purchased from Fisher

Scientific (Ottawa, ON, Canada). Primary antibodies were purchased from Abcam (Cambridge,

MA, USA) for CD31 (ab28364), LYVE-1 (ab14917), collagen IV (ab19808), and αSMA

(ab5694). Immunostaining reagents (rabbit IgG Elite ABC kit, avidin/biotin kit, enzyme

substrates, Vectashield mounting medium) were purchased from Vector Labs (Burlington, ON,

Canada). Entellan hard mounting medium was purchased from EMD Millipore (Billerica, MA,

USA). All other materials were purchased from Sigma-Aldrich (Mississauga, ON, Canada) and

used as received unless otherwise noted.

3.3.2 Cell maintenance and preparation

MDA-MB-231-H2N cells were maintained in RPMI 1640 culture medium, supplemented with

10% heat-inactivated fetal bovine serum (FBS), 50 units/mL penicillin and 50 mg/mL

streptomycin under a humidified 5% CO2 environment. To prepare cell suspensions for injection,

adherent cells were first rinsed with phosphate buffered saline, pH 7.4 (PBS), and then incubated

briefly with trypsin-ethylenediamine tetraacetic acid (trypsin-EDTA, 0.25%/0.038%). Once the

cells were suspended, enzymatic digestion was inhibited with FBS, and the cells were pelleted

and washed 3 times in PBS before resuspension at the desired concentration. Cells were kept on

ice prior to injection.

37

3.3.3 Tumour xenograft models

The protocols used in these in vivo studies were approved by the University Health Network

Animal Care Committee and performed in accordance with current institutional and national

regulations. Animals were housed in a 12 h light and 12 h dark cycle with free access to food and

water. NSG mice were bred in-house, and 7-9 week old female mice were selected for tumour

xenotransplantation.

Mice in all experimental groups were inoculated with 106 MDA-MB-231-H2N cells suspended

in 50 µL of sterile PBS. Prior to injection, mice were anaesthetized with isoflurane-oxygen. To

form ectopic SC tumours, anaesthetized mice were injected with cells under the skin in the right

dorsal flank. To form orthotopic mammary fat pad (MFP) tumours, the surgical area was

depilated and swabbed with 70% ethanol and betadine before making an incision in the skin of

the lower abdomen to the right of the midline, uncovering the mammary fat pad in the right

inguinal region where cells were injected into the fat pad. The incision was then sutured closed

and lactated Ringer’s solution and buprenorphine were given post-operatively for recovery and

pain management. Solid tumours were allowed to form over a period of 3-5 weeks. Cohorts of

tumour-bearing animals were divided into two groups to proceed onwards for testing; the first

group was tested once their tumours reached an average diameter along the major axis of 7 mm

as measured through the skin using calipers, and the second group tested the following week.

3.3.4 Dye injections and tissue collection

Once tumours reached their target size, mice were injected with 0.5 mg of FITC-dextran in 200

µL of PBS via intravenous (IV) tail vein injection [8]. After 1 h, animals were sacrificed by CO2

asphyxiation and tissue samples (tumour and liver) were collected by dissection; tumour samples

were directly measured for diameter along both the major and minor axes (L and W) and

thickness (H) using calipers (ellipsoid volume calculated as 𝜋 6×𝐿×𝑊×𝐻 [22]), and each

sample was placed separately in cassettes and submerged in 4% paraformaldehyde for 24 h at 4

°C. Tissue samples were then cryoprotected in 30% sucrose in PBS and stored at 4 °C. Tissue

samples were cryosectioned in 10 µm sections, and pairs of slices 50 µm apart were mounted

onto slides, and stored at -80 °C. For fluorescence analysis, slides were rehydrated in PBS and

coverslipped using Vectashield mounting medium.

38

3.3.5 Immunostaining

Three slides (six tissue sections) from each tumour were selected for each set of stains such that

each slide contained sections a minimum of 300 µm away from the previous slide. Thawed

slides were hydrated and washed in PBS and incubated with 0.3% H2O2 in methanol for 20 min

before being washed in PBS again and blocked in 1.5% normal goat serum (NGS) in PBS (see

Table 3.1 for details). Avidin and biotin blocking reagents were applied sequentially for 15 min

each before incubating with the primary antibody at 4 °C overnight (dilutions noted in Table

3.1). The following day, slides were washed in PBS and incubated with a biotinylated secondary

goat anti-rabbit IgG (1:200 dilution as instructed in kit) , followed by incubation with avidin-

biotinylated enzyme complex (ABC reagent) (times noted in Table 3.1). Rinsed sections were

then developed using 3,3’-diaminobenzidine (DAB) enzyme substrate for 1-10 min (brown

product). If applicable, slides were then co-stained by repeating the above procedure beginning

at the NGS blocking step, and developed in VIP enzyme substrate for 5-7 min (violet product).

All slides were counter stained by applying 0.5% methyl green for 10 min (blue-green nuclear

stain), washed in distilled water, dried in 1-butanol, and transferred to xylene before being

coverslipped using Entellan hard mounting medium.

Table 3.1 Immunostaining protocol details listed by antigen

  CD31   Collagen  IV   αSMA   LYVE-­‐1  

NGS  blocking  incubation  time   1  h   1  h     1  h   20  min  

Primary  antibody  dilution   1:200   1:1000   1:1000   1:1000  

Secondary  antibody  incubation  time   1  h   30  min   30  min   30  min  

ABC  reagent  incubation  time   1  h   30  min   30  min   30  min  

3.3.6 Image acquisition and analysis

All fluorescence images were acquired with a fixed exposure time for each channel using an

Olympus BX50 with a UPlanSApo 10×/0.40 objective, Photometrics CoolSNAP HQ2

monochrome camera, and motorized stage (Olympus Canada Inc., Richmond Hill, ON, Canada).

Images were tiled together using Metamorph, and analyzed using ImageJ. Brightfield images

39

were acquired using an Aperio ScanScope XT (Aperio, Vista, CA, USA) for whole slide

scanning at 20× magnification and analyzed using ImageScope Microvessel Analysis. Statistical

significance between groups was first tested with Bartlett’s test for equality of variance (P <

0.05). Where variances were equivalent, one-way ANOVA was applied, followed by a corrected

unpaired t-test; differences are denoted by square bracket symbols connecting the differing

groups (P < 0.05, unless otherwise noted).

3.4 Results and discussion

3.4.1 Orthotopic cell transplantation influences tumour growth rate and size variation

Tumour size of human tumour xenograft models grown in mice both orthotopically (MFP) and

ectopically (SC) was monitored weekly through the skin in live animals using calipers.

Following cell injection, MFP tumours reached a target size of 7 mm in diameter across the

major axis by 3 weeks post-injection whereas SC tumours took an additional 2 weeks to reach

this size. Differences in growth rate were expected, as each injection site provides a different

microenvironment. Cohorts of animals were selected based on tumour size matching instead of

development time because size is one of three standard measurements that determines breast

cancer prognosis [21]. After resection, tumours were measured directly using calipers and the

volumes were calculated based on measurements of the major and minor axes and thickness

(Figure 3.1). The difference in time needed to achieve size matched populations for MFP and

SC tumour models suggests that the organ environment influences the growth rate of xenografted

cells.

To investigate effects associated with tumour size, 4 animals from each tumour type were

randomly selected for dextran injection and tumour resection once the 7 mm major axis diameter

was reached, with the remaining animals evaluated the following week. The 7 mm target size

was selected to allow adequate vascular pathology to develop, as neovascularization of the

tumour is most pronounced over serveral days immediately after a palpable mass (20 mm3) has

formed [23]. Notably, the week after this target size was reached, tumour size variability

increased in both tumour sites (P < 0.05 by Bartlett’s test of equality of variances).

Unexpectedly, several tumours in the SC group at 6 weeks post-cell injection were smaller than

those observed the previous week (Figure 3.1). Additionally, several replicates were of similar

40

size or larger size, resulting in a broad size distribution of the resulting tumours. In this group, 4

out of 6 animals developed hard fibrotic tissue leading to an ulcerated skin lesion by this time (an

indication for humane sacrifice). These lesions, which made these animals unsuitable for further

study beyond this time, were not observed in any other group. MFP tumours were grown from

cells injected directly into the centre of the MFP, surrounding transplanted cells with endogenous

support cells and forming a biological barrier against contact with the skin, which may have

prevented ulceration. These injected cells also had access to the pre-existing vascular network

and biological signaling molecules present in the MFP. Overall, the MFP tumours were more

consistent in size than the SC tumours. Given the large variability in the 6 week SC tumour

group, these samples were not further analyzed. Instead, 5 week SC tumours were compared

with 3 and 4 week MFP tumours, which were similar in size.

Figure 3.1 MFP and SC tumour sizes. Tumour volumes were calculated based on caliper measurements

post-dissection of the major and minor axes and thickness (n = 4-6). SC tumours required longer

development times to become size matched to MFP tumours. Greater variability was also observed at

longer times, particularly in SC tumours, where several animals had smaller tumours than the cohort

examined the week before.

0  

50  

100  

150  

200  

250  

300  

MFP  3  wks    MFP  4  wks    SC  5  wks    SC  6  wks  

Tumou

r  volum

e  (m

m3 )  

41

3.4.2 MFP tumours exceed SC tumours in model nanocarrier accumulation

Prior to sacrifice, a high molecular weight FITC-dextran, used as a model nanocarrier, was

injected to assess blood vessel permeability. Data were normalized to liver tissue collected as a

positive control: liver endothelial cells have natural fenestrations (123 ± 24 nm diameter) [24]

for transfer of substrates from the blood to hepatocytes, making the liver an ideal organ for

observing nanocarrier uptake. In mice, blood flow through the liver is also estimated at 23% of

cardiac output, making it one of the best perfused organs on a per gram basis [25].

Based on fluorescence images of tissue sections, relatively poor dextran uptake was observed in

tumour tissue compared to liver tissue across all groups. A threshold was defined to exclude

background signal detected in blank tumour and liver tissue and the remaining areas,

representing levels above this threshold, were quantified. Less than 1% of the positive signal

area observed in the liver control was observed in tumour slices (Figure 3.2). This can partially

be explained by relatively low blood flow through tumour tissue, which has previously been

reported to be up to 5-fold lower than in liver [26]. The remaining discrepancy between the

dextran accumulation between tumour and liver samples suggests that the model tumour

vasculature was less permissive to dextran uptake than the fenestrated liver endothelium, and/or

that the lymphatic drainage in the model tumour prevented stable dextran accumulation.

Interestingly, dextran uptake in 3 week old MFP tumours was higher than size matched 5 week

old SC tumours at 90% confidence (P = 0.08 by one-way ANOVA), suggesting that the

orthotopic MFP environment encouraged EPR permissive vasculature and/or lymphovasculature.

42

Figure 3.2 FITC-Dextran accumulation in tumour tissue normalized to liver tissue control. High molecular

weight dextran (2 MDa, ~80 nm) was injected IV into tumour animals as a model nanocarrier and allowed

to distribute prior to sacrifice. 3 week old MFP tumours showed higher accumulation of the nanocarrier

than 5 week old SC tumours at a 90% confidence interval. All data are shown as the mean of n = 4

animals ± SD. Lines connecting bars denote statistical significance, P < 0.10.

3.4.3 Elements of tumour vascular pathophysiology observed in tumour models

To better understand the underlying vascular pathophysiology present in both tumour models,

tumour slices were immunostained to provide information on the blood and lymphatic vessels

present. Tissue was stained for CD31, an endothelial cell marker, to locate and characterize

blood vessels. In normal blood vessels, an intact monolayer of endothelials cells is expected,

whereas hyperpermeable tumour blood vessels are characterized by multiple layers of

discontinuous endothelial cells that may sprout outwards or project into the vessel lumen [10,

13]. The CD31 staining revealed greater vessel wall thickness across all groups when compared

to liver tissue (represented by a dashed line) which was used as a healthy tissue control (Figure

3.3A). This observation suggests that blood vessels present in all models, whether they are

0  

0.2  

0.4  

0.6  

0.8  

1  

1.2  

MFP  3  wks   MFP  4  wks   SC  5  wks  

Percen

t  pixels  a

bove  th

reshold  

(normalize

d  to  liver)  

FITC-­‐Dextran  P  <  0.10  

43

existing vessels that have been remodeled or newly formed vessels, have the abnormal multi-

layered endothelial cell structure associated with solid tumours. The vessel thickness was

highest in the 3 week old MFP tumours, indicating a greater level of endothelial cell

disorganization in this group. It is possible that this led to the increased permeability observed in

the 3 week MFP tumours using a relatively large model nanocarrier (~80 nm), an effect that is

more pronounced in other studies utilizing models such as albumin (~7 nm) [27, 28].

Separate sections were also co-stained for collagen IV to visualize the thickness of the associated

basement membrane. The basement membrane forms a physical barrier that inhibits transport of

high molecular weight materials across blood vessel walls [15, 29]. In tumour pathophysiology,

opposing phenomena have been observed: the basement membrane can thicken, thin, or even be

absent. In the MFP and SC tumour models, the basement membrane was thickened compared to

healthy liver blood vessels (Figure 3.3B). This observation is consistent with the xenografted

MDA-MB-231-H2N cell line being poorly invasive like its parental line, MDA-MB-231 [30].

Conversely, a more metastatic cell line is often capable of using MMPs to degrade the basement

membrane to enable cell migration through neighbouring blood vessels [30]. The 5 week old SC

tumours were observed to have the highest basement membrane thickness, indicating the greatest

mass transport barrier against nanocarrier delivery.

44

Figure 3.3 CD31 and collagen IV immunostaining. Mean blood vessel wall thickness visualized through

(A) CD31 (endothelial cells) and (B) collagen IV (basement membrane). Both are abnormally thick as

compared to healthy liver control tissue, which is denoted by the dashed line. (C) shows that mean blood

vessel density assayed using CD31 staining is greatest in 3 week old MFP tumours. (D) indicates mean

vascular area as a measure of blood vessel size and capacity. Their small size categorizes them as

microvasculature. All data are shown as the mean of n = 4 animals ± SD. Starred lines connecting bars

denote statistical significance, P < 0.05.

0  

0.5  

1  

1.5  

2  

2.5  

MFP  3  wks    MFP  4  wks    SC  5  wks  

Mean  vessel  wall  thickne

ss  (µ

m)   CD31  

*  *  

0  

0.5  

1  

1.5  

2  

2.5  

MFP  3  wks    MFP  4  wks    SC  5  wks  

Mean  vessel  wall  thickne

ss  (µ

m)   Collagen  IV  

*  *  

0  

50  

100  

150  

200  

MFP  3  wks    MFP  4  wks    SC  5  wks  

Mean  vessel  den

sity  (#/m

m2 )   CD31  

*  *  

0  

50  

100  

150  

200  

MFP  3  wks    MFP  4  wks    SC  5  wks  

Mean  vascular  area  (µm

2 )   CD31  

*  

A   B  

C   D  

45

CD31 staining also revealed differences in vascular density, with the 3 week old MFP tumours

having a significantly greater vessel density than the other groups (Figure 3.3C). The decrease in

vascular density from 3 weeks to 4 weeks in the MFP model suggests that the tumour cell growth

may be too rapid for the corresponding new blood vessels to form. The thick basement

membranes observed in the tumour tissue may also contribute to this deficiency as the basement

membrane must be degraded before vascular branching can occur [15]. Although the 3 week old

MFP and 5 week old SC tumours were size matched, the MFP model had greater blood vessel

density, which may be attributed to greater vascular density in the MFP. Together these

observations suggest that remodeling blood vessels already present in the transplantation site are

important in establishing relevant tumour vasculature. The relatively poor vascular density in SC

tumours may also explain the poor engraftment after 6 weeks, as a lack of blood flow may inhibit

further growth and lead to necrosis.

The mean vascular area was also quantified, giving an indication of the size, and therefore the

capacity of the blood vessels present in each tumour type. The vascular area in 3 week old MFP

tumours was significantly higher than the 4 week old MFP tumours (Figure 3.3D), indicating that

in addition to decreasing vessel density with increasing tumour size, there is on average a lower

capacity for blood in the vessels present. Having a greater density and capacity for blood

perfusion enhances the likelihood for delivery of materials to the 3 week old MFP tumours

through systemic circulation.

At the same time, all of the evaluated models are likely underperfused as their small size

categorizes them as microvasculature [31]. This low overall capacity for blood flow impacts

their utility in assessing nanocarrier accumulation via EPR, and likely results in regions of

hypoxia and heterogeneous drug distribution.

CD31 was also co-stained with αSMA to visualize differences in pericyte association with blood

vessels. Pericytes are important blood vessel support cells that help to regulate blood flow and

vessel permeability, but are often detached in tumour pathophysiology. This was indeed

observed across all tumour models (Figure 3.4A-C), where pericytes (violet) were distributed

throughout tumour tissue instead of associating exclusively with blood vessels (brown) and

forming uniform layers around the endothelial cell layer, as observed in healthy liver tissue

(Figure 3.4D).

46

Figure 3.4 CD31 and αSMA co-staining. Representative images of pericytes (αSMA, violet) that are

detached from blood vessels (CD31, brown) in: (A) 3 week MFP, (B) 4 week MFP, and (C) 5 week SC

tumours. Several blood vessels are highlighted with black arrows; blue staining represents cell nuclei.

(D) shows that pericytes are exclusively associated with blood vessels in healthy liver control tissue.

Scale bars represent 200 µm.

LYVE-1 staining was used to detect lymphatic vessels in tumour tissue. Lymphatic vessels

provide a network to drain protein rich interstitial fluid back into circulation. By the nature of

their function, these vessels are porous to allow macromolecules to be transported [32], and

therefore nanocarrier accumulation in tumour tissue may increase when their expression is

impaired. Mouse models of lymphatic impairment can be generated by surgically ablating

lymphatic vessels in the tail, resulting in lymphedema. In these models, the surrounding tissue

attempts to restore homeostasis by generating new lymphatic vessels and dilating the remaining

lymphatic vessels, suggesting that both density and diameter impact drainage capacity [33].

A   B  

C   D  

47

LYVE-1 stained sections were used to quantify lymphatic vessel size and density (Figure 3.5A-

B). Both of these measures gave different variances between groups (P < 0.05 by Bartlett’s test

of equality of variances) meaning that the groups tested were not equivalent. While the mean

lymphatic vessel density was highest in the 3 week old MFP tumours, the 5 week old SC

tumours demonstrated the highest mean lymphatic vessel area. These factors counterbalance one

another, as density and capacity each contribute to overall drainage.

There is evidence that both the MFP and SC tumour models yielded poor lymphatic drainage

compared to healthy tissue. Accumulation of interstitial fluid in cases of lymphedema has been

shown to lead to the deposition of collagen [34]. Visual examination of the tumour slices

revealed a high density of collagen IV-lined spaces that were CD31 negative, which likely

represent fluid-filled cavities in the tumour tissue (Figure 3.5C-D). These likely contain excess

interstitial fluid resulting from a combination of increased vascular permeability and deficient

lymphatic drainage.

48

Figure 3.5 LYVE-1 immunostaining. (A) shows mean lymphatic vessel density, and (B) shows mean

vessel area, both of which are indicators of lymphovascular capacity. Both measures were found to have

unequal variance between groups, and therefore although the groups were not equivalent, ANOVA could

not be used to verify their differences. While 3 week old MFP tumours had the highest mean lymphatic

vessel density, 5 week old SC tumours had greater mean vessel size, both of which contribute to overall

lymphatic drainage capacity. All data are shown as the mean of n = 4 animals ± SD. Representative

images of collagen (violet) positive but CD31 (brown) negative fluid filled spaces are shown in: (C) 3 week

MFP and (D) 5 week SC tumours. Several of these spaces are highlighted with black arrows; blue

staining represents cell nuclei. Scale bars represent 200 µm.

0  

5  

10  

15  

20  

25  

MFP  3  wks    MFP  4  wks    SC  5  wks  

Mean  vessel  den

sity  (#

/mm

2 )   LYVE-­‐1  

0  

50  

100  

150  

200  

250  

MFP  3  wks    MFP  4  wks    SC  5  wks  

Mean  vessel  area  (µm

2 )  

LYVE-­‐1  A   B  

C   D  

49

Taken together, the data gathered through CD31 and collagen IV immunostaining suggest that,

of the models tested, the 3 week MFP tumour best replicates the vascular permeability required

to observe the EPR effect in vivo. However, the blood vessels visualized are sparse and small,

contributing to low accumulation of the model nanocarrier used in this study. Both MFP and SC

tumours showed evidence of excess interstitial fluid accumulation, suggesting poor lymphatic

drainage in both models. While MFP tumours demonstrated greater lymphatic vessel density,

SC tumours had greater lymphatic vessel size, both of which contribute to drainage, making it

difficult to easily differentiate the two models in terms of drainage capacity. MFP tumours also

demonstrated greater utility for long-term treatment studies, as their growth is more consistent at

large tumour sizes, and no skin ulcerations were observed.

3.5 Conclusions

This study provides insight into the vascular properties of human tumour xenograft models of

breast cancer in both MFP (orthotopic) and SC (ectopic) environments, two common pre-clinical

models. When both animal models were challenged with a high molecular weight dextran as a

model nanocarrier, there was higher accumulation in MFP tumours 3 weeks after cell injection.

Further adding to the evidence that MFP tumour vasculature has greater permeability to

macromolecules – a pathological feature relevant to nanocarrier accumulation via EPR – CD31

and collagen IV immunostaining revealed greater vascular density and size, as well as thinner

basement membranes, in MFP tumours collected 3 weeks after cell injection. Both models

demonstrated poor dextran accumulation compared to the liver as a positive control, suggesting

that although several pathological features were observed, low vascular density and small blood

vessel size led to relatively poor tumour perfusion. Both the MFP and SC tumour models

showed evidence of poor lymphatic drainage, as several CD31 negative and collagen IV positive

fluid filled cavities were observed. The MFP environment offered several practical benefits,

including shorter development times to reach a target tumour size, more consistent growth

profiles, and the absence of ulcerated skin lesions observed in SC tumour animals.

3.6 List of abbreviations used

α-SMA Alpha smooth muscle actin

DAB 3,3’-diaminobenzidine

EPR Enhanced permeability and retention

50

FBS Fetal bovine serum

LYVE-1 Lymphatic vessel endothelial hyaluronan receptor

MFP Mammary fat pad

MMP Matrix metalloproteinase

NGS Normal goat serum

NSG mice NOD scid gamma mice

PBS Phosphate buffered saline, pH 7.4

SC Subcutaneous

3.7 Authors’ contributions

KSH designed the study and protocols, performed animal experiments, immunostained tissue,

collected images, maintained and prepared cells for transplantation, executed the data analysis,

and prepared the manuscript. PP was responsible for the breeding the mouse colony, performing

cell injections, monitoring tumour growth, and assisted in designing protocols, performing the

animal experiments, immunostaining tissue, and collecting images. SCO participated in

designing the study and protocols, and assisted in performing SC cell injections. MSS

participated in study design and was involved in writing the manuscript. All authors read and

approved the final manuscript.

3.8 Acknowledgements

We thank: Drs. Robert Kerbel (Sunnybrook Health Science Centre), Armand Keating and Yoko

Kosaka (Princess Margaret Hospital) for their help and advice in establishing the mouse tumour

model. We are grateful to the Canadian Institutes of Health Research (CIHR to MSS) for funding

of this research.

3.9 References

[1] Teicher BA. Human tumor xenografts and mouse models of human tumors: re-discovering the models. Expert Opin Drug Dis 2009;4:1295-1305.

[2] Torchilin V. Tumor delivery of macromolecular drugs based on the EPR effect. Adv Drug Deliver Rev 2011;63:131-135.

[3] 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. Adv Drug Deliver Rev 2011;63:136-151.

51

[4] Carmeliet P, Jain RK. Angiogenesis in cancer and other diseases. Nature 2000;407:249-257.

[5] Kerbel RS. Tumor angiogenesis: past, present and the near future. Carcinogenesis 2000;21:505-515.

[6] Matsumura Y, Maeda H. A New Concept for Macromolecular Therapeutics in Cancer-Chemotherapy - Mechanism of Tumoritropic Accumulation of Proteins and the Antitumor Agent Smancs. Cancer Res 1986;46:6387-6392.

[7] Morikawa S, Baluk P, Kaidoh T, Haskell A, Jain RK, McDonald DM. Abnormalities in pericytes on blood vessels and endothelial sprouts in tumors. Am J Pathol 2002;160:985-1000.

[8] Dreher MR, Liu WG, Michelich CR, Dewhirst MW, Yuan F, Chilkoti A. Tumor vascular permeability, accumulation, and penetration of macromolecular drug carriers. J Natl Cancer I 2006;98:335-344.

[9] Jain RK. Normalization of tumor vasculature: An emerging concept in antiangiogenic therapy. Science 2005;307:58-62.

[10] Hashizume H, Baluk P, Morikawa S, McLean JW, Thurston G, Roberge S, et al. Openings between defective endothelial cells explain tumor vessel leakiness. Am J Pathol 2000;156:1363-1380.

[11] Sorensen AG, Batchelor TT, Zhang WT, Chen PJ, Yeo P, Wang MY, et al. A "Vascular Normalization Index" as Potential Mechanistic Biomarker to Predict Survival after a Single Dose of Cediranib in Recurrent Glioblastoma Patients. Cancer Res 2009;69:5296-5300.

[12] Hida K, Hida Y, Shindoh M. Understanding tumor endothelial cell abnormalities to develop ideal anti-angiogenic therapies. Cancer Sci 2008;99:459-466.

[13] McDonald DM, Baluk P. Significance of blood vessel leakiness in cancer. Cancer Res 2002;62:5381-5385.

[14] Baluk P, Morikawa S, Haskell A, Mancuso M, McDonald DM. Abnormalities of basement membrane on blood vessels and endothelial sprouts in tumors. Am J Pathol 2003;163:1801-1815.

[15] Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamic niche in cancer progression. J Cell Biol 2012;196:395-406.

[16] Kerbel RS, Cornil I, Theodorescu D. Importance of Orthotopic Transplantation Procedures in Assessing the Effects of Transfected Genes on Human Tumor-Growth and Metastasis. Cancer Metast Rev 1991;10:201-215.

[17] Killion JJ, Radinsky R, Fidler IJ. Orthotopic models are necessary to predict therapy of transplantable tumors in mice. Cancer Metast Rev 1998;17:279-284.

[18] Lunt SJ, Kalliomaki TMK, Brown A, Yang VX, Milosevic M, Hill RP. Interstitial fluid pressure, vascularity and metastasis in ectopic, orthotopic and spontaneous tumours. Bmc Cancer 2008;8.

52

[19] Wilmanns C, Fan D, Obrian CA, Bucana CD, Fidler IJ. Orthotopic and Ectopic Organ Environments Differentially Influence the Sensitivity of Murine Colon-Carcinoma Cells to Doxorubicin and 5-Fluorouracil. International Journal of Cancer 1992;52:98-104.

[20] Francia G, Cruz-Munoz W, Man S, Xu P, Kerbel RS. Mouse models of advanced spontaneous metastasis for experimental therapeutics. Nat Rev Cancer 2011;11:135-141.

[21] Edge SB, Compton CC. The American Joint Committee on Cancer: the 7th Edition of the AJCC Cancer Staging Manual and the Future of TNM. Ann Surg Oncol 2010;17:1471-1474.

[22] Tomayko MM, Reynolds CP. Determination of Subcutaneous Tumor Size in Athymic (Nude) Mice. Cancer Chemoth Pharm 1989;24:148-154.

[23] Schiffelers RM, Metselaar JM, Fens MHAM, Janssen APCA, Molema G, Storm G. Liposome-encapsulated prednisolone phosphate inhibits growth of established tumors in mice. Neoplasia 2005;7:118-127.

[24] Cogger VC, McNerney GP, Nyunt T, DeLeve LD, McCourt P, Smedsrod B, et al. Three-dimensional structured illumination microscopy of liver sinusoidal endothelial cell fenestrations. J Struct Biol 2010;171:382-388.

[25] Davies B, Morris T. Physiological Parameters in Laboratory-Animals and Humans. Pharmaceut Res 1993;10:1093-1095.

[26] Hori K, Saito S, Takahashi H, Sato H, Maeda H, Sato Y. Tumor-selective blood flow decrease induced by an angiotensin converting enzyme inhibitor, temocapril hydrochloride. Jpn J Cancer Res 2000;91:261-269.

[27] Chen B, Pogue BW, Zhou XD, O'Hara JA, Solban N, Demidenko E, et al. Effect of tumor host microenvironment on photodynamic therapy in a rat prostate tumor model. Clin Cancer Res 2005;11:720-727.

[28] Tong RT, Boucher Y, Kozin SV, Winkler F, Hicklin DJ, Jain RK. Vascular normalization by vascular endothelial growth factor receptor 2 blockade induces a pressure gradient across the vasculature and improves drug penetration in tumors. Cancer Res 2004;64:3731-3736.

[29] Kong G, Braun RD, Dewhirst MW. Hyperthermia enables tumor-specific nanoparticle delivery: Effect of particle size. Cancer Res 2000;60:4440-4445.

[30] Abdelkarim M, Vintonenko N, Starzec A, Robles A, Aubert J, Martin M-L, et al. Invading Basement Membrane Matrix Is Sufficient for MDA-MB-231 Breast Cancer Cells to Develop a Stable In Vivo Metastatic Phenotype. PLoS ONE 2011;6:e23334.

[31] Perles-Barbacaru AT, van der Sanden BPJ, Farion R, Lahrech H. How stereological analysis of vascular morphology can quantify the blood volume fraction as a marker for tumor vasculature: comparison with magnetic resonance imaging. J Cerebr Blood F Met 2012;32:489-501.

53

[32] Rockson SG. Diagnosis and management of lymphatic vascular disease. J Am Coll Cardiol 2008;52:799-806.

[33] Schneider M, Ny A, Ruiz de Almodovar C, Carmeliet P. A new mouse model to study acquired lymphedema. PLoS Med 2006;3:e264.

[34] Szuba A, Rockson SG. Lymphedema: anatomy, physiology and pathogenesis. Vasc Med 1997;2:321-326.

 

54

4 Drug-loaded nanoparticles for targeted delivery to a mouse model of breast cancer

This chapter is derived from the following manuscript:

Ho KS, Aman AM, Al-awar RS, and Shoichet MS (2012) Amphiphilic micelles of poly(2-methyl-2-carboxytrimethyle carbonate-co-D,L-lactide)-graft-poly(ethylene glycol) deliver anti-cancer drugs to solid tumours. Biomaterials, 33 (7), 2223–2229.

Reprinted with permission from Elsevier.  

4.1 Abstract

Drug delivery to solid tumours remains a challenge because both tumour physiology and drug

solubility are unfavourable. Engineered materials can provide the basis for drug reformulation,

incorporating active compounds and modulating their pharmacokinetic and biodistribution

behaviour. To this end, we encapsulated docetaxel, a poorly soluble taxane drug, in a self-

assembled polymeric nanoparticle micelle of poly(2-methyl-2-carboxytrimethylene carbonate-

co-D,L-lactide)-graft-poly(ethylene glycol) (poly(TMCC-co-LA)-g-PEG). This formulation was

compared with its conventional ethanolic polysorbate 80 formulation in terms of plasma

circulation and biodistribution in an orthotopic mouse model of breast cancer. Notably, the

polymeric nanoparticle formulation achieved greater tumour retention, resulting in prolonged

exposure of cancer cells to the active drug. This behaviour was unique to the tumour tissue. The

active drug was eliminated at equal or greater rates in all other tissues assayed when delivered in

the polymeric nanoparticles vs. the free drug formulation. Thus, these polymeric nanoparticles

are promising vehicles for solid tumour drug delivery applications, offering greater tumour

exposure while eliminating the need for toxic solvents and surfactants in the dosing formulation.

4.2 Introduction

Solid tumours, such as breast cancer, present several physical barriers against effective drug

delivery, as therapeutic agents must cross into, and remain, at the tumour site despite high

interstitial pressures and low vascular densities [1-3]. Additionally, many anti-cancer drugs have

non-specific modes of action, so when coupled with a broad systemic distribution, the resulting

impact on healthy cells leads to dose-limiting toxicity [4]. Nanoparticle targeting exploits a

unique physiological feature of solid tumours resulting from rapid malignant growth:

hyperpermeable vasculature and poor lymphatic drainage lead to enhanced permeability and

55

retention (EPR) of large molecules and small particles on the nanometer scale, providing a

means for selective tissue accumulation [5,6]. Well-designed nanoscale drug delivery systems

have the potential to increase the therapeutic index of small molecule drugs by extending drug

circulation while boosting solid tumour specificity and accumulation through the EPR effect. To

take advantage of EPR, several technologies have been developed, including liposomes [7],

dendrimers [8], and polymeric nanoparticles [9].

Polymeric nanoparticles, comprised of a hydrophobic core and hydrophilic corona, are

particularly compelling for the encapsulation and delivery of hydrophobic and poorly water

soluble chemotherapeutic drugs. Many of these polymers are block copolymers of hydrophobic

poly(aspartic acid) or poly(lactide-co-glycolide) and hydrophilic poly(ethylene glycol) [10,11].

Several parameters have been investigated in terms of circulation half-life, including the length

and density of the PEG block [12] and size and shape of the nanoparticles [13,14]. Interestingly

few polymeric nanoparticles have been designed with functional groups.

The copolymer poly(2-methyl-2-carboxytrimethylene carbonate-co-D,L-lactide)-graft-

poly(ethylene glycol) (poly(TMCC-co-LA)-g-PEG) was designed to have either a PEG-furan or

a PEG-azide for facile click modification of the nanoparticle surface by either Diels-Alder [15]

or Huisgen 1,3-dipolar cycloaddition [16], respectively. By combining the hydrophobic

backbone of poly(TMCC-co-LA) with hydrophilic PEG, the resulting amphiphilic copolymer

spontaneously self-assembles into nanoscale core-shell micelles on contact with water through a

simple dialysis process [15]. Interestingly, there is consistently only one PEG per poly(TMCC-

co-LA) backbone, giving this polymer a block-like structure. Moreover, the polymeric

nanoparticles have demonstrated stability in blood serum proteins in vitro [17]. We hypothesized

that these poly(TMCC-co-LA)-g-PEG nanoparticle micelles would be beneficial in vivo, where

the hydrophobic inner core would load a hydrophobic chemotherapeutic drug for delivery, and

the PEG corona would reduce protein binding and thereby allow longer circulation and greater

tumour accumulation before elimination by the reticuloendothelial system (RES) [9].

Biologically active anti-cancer drugs are often hydrophobic, bulky, and polycyclic, leading to

poor aqueous solubility and limited utility [18]. Consequently, such compounds are often

formulated with organic co-solvents and surfactants, each with their own systemic toxicities.

Docetaxel (DTX) is a small molecule taxane drug that falls into this category: it demonstrates

56

excellent clinical activity against breast cancer but requires a high concentration of polysorbate

80 (PS80 or tween 80) to solubilize relevant concentrations for dosing. Unfortunately, dosing

these levels of PS80 causes hypersensitivity reactions, necessitating pre-treatment with

corticosteroids and further reducing the mean tolerable dose [19-21].

To take advantage of the potency of DTX without being limited by current formulations, we

endeavoured to encapsulate it in the poly(TMCC-co-LA)-g-PEG nanoparticles, taking advantage

of its solubility in the hydrophobic core of our polymeric nanoparticles (Figure 4.1). Importantly,

this methodology required neither chemical modification nor the use of toxic co-solvents in the

final formulation. Success here allowed us to test, for the first time, these polymeric

nanoparticles in terms of the in vivo circulation and biodistribution of DTX vs. standard

formulations.

Figure 4.1 Poly(TMCC-co-LA)-g-PEG, shown here with a furan group at the PEG terminus, is an

amphiphilic co-polymer that self assembles into polymeric nanoparticle micelles with a core-shell

structure on dialysis against water. DTX and the polymer are first co-dissolved in organic solvent before

dialysis. During dialysis, DTX partitions into the hydrophobic core, thereby encapsulating it. The

polymeric nanoparticles have functional groups available for further modification: carboxylic acid groups

on the poly(TMCC-co-LA) backbone and furan moieties on the PEG corona.

To understand the pharmacokinetic behaviour and biodistribution of DTX-loaded nanoparticles

(DTX-NP), we compared their performance vs. free DTX in the conventional ethanolic PS80

formulation after dose matched IV injection in tumour-bearing mice. Solid orthotopic tumours

were established by transplanting human breast cancer cells into the mammary fat pads of female

mice. Ultra performance liquid chromatography-coupled with mass spectrometry (UPLC-MS)

detects the unaltered therapeutic compound without labelling, making it possible to distinguish

Self-assembly

via dialysis

DTX

57

the active compound from inactive fragments, metabolites, or uncoupled tags. This analytical

technique is quantitative, sensitive to nM levels, and compatible with small sample volumes [22].

While radiolabeling is commonly used to quantitatively track a drug in tissues and plasma after

dosing [23], it relies on a tag as a reporter, and degradation products from the tagged compound

can result in misleading data. The UPLC-MS method used here allowed us to directly measure

the concentrations of unmetabolized DTX in plasma and tissues after intravenous (IV) injection.

Blood samples were drawn via tail vein or cardiac puncture. Using UPLC-MS, we quantified

DTX concentrations in the plasma fraction over an 8 h time course and calculated

pharmacokinetic (PK) parameters for each formulation.

4.3 Experimental

4.3.1 Materials

All cell culture materials were purchased from Gibco-Invitrogen (Burlington, ON, Canada).

MDA-MB-231-H2N cells and NOD scid gamma (NSG) mice were generous gifts from Dr.

Robert Kerbel (Sunnybrook Research Institute, Toronto, ON, Canada), which were then

maintained or bred in-house. Dialysis membranes were acquired from Spectrum Laboratories

(Rancho Dominguez, CA, USA). Docetaxel was obtained through LC Laboratories (Woburn,

MA, United States). Poly(TMCC-co-LA) was synthesised as previously described [15,24].

Furan-PEG-NH2 was prepared from 10 kDa Boc-NH-PEG-NHS obtained from Rapp Polymere

(Tuebingen, Germany), and grafted to the polymer backbone as previously described [15,24].

The resulting grafted copolymer is shown in Figure 4.1. Heparinized capillary tubes were

purchased through Sarstedt (Montreal, QC, Canada). All other materials were purchased from

Sigma-Aldrich (Mississauga, ON, Canada) and used as received unless otherwise noted.

4.3.2 DTX concentration measurement

Chromatographic separations were carried out on an ACQUITY UPLC BEH C18 (2.1 × 50 mm,

1.7 µm) column using ACQUITY UPLC system. The mobile phase was 0.1% formic acid in

water (solvent A) and 0.1% formic acid in acetonitrile (solvent B). The column was equilibrated

for 1 min in 95% solvent A as the starting point for the gradient, dropping to 5% over 4.5 min,

holding for 0.5 min, and moving back to 95% in 0.5 min. A Waters Xevo QTof MS equipped

with an atmospheric pressure ionization source was used for MS analysis. For quantification,

58

stock standard solutions of the active DTX compound were added to the final appropriate matrix

for comparison: 2:1 acetonitrile:water, v/v for nanoparticle suspensions, or precipitated plasma

or pooled tissue homogenates as appropriate. Under these conditions the polymer nanoparticles

are dissolved, resulting in a combined measurement of both encapsulated and released DTX

present in the original sample. The instrument was sensitive to DTX concentrations as low as 5

ng/mL. All values shown are the average of 5 samples with error bars representing their standard

deviation. Group means were compared by one-way ANOVA followed by a corrected unpaired

t-test; differences are denoted by starred symbols (p < 0.05). MassLynx 4.1 was used for peak

area analysis and WinNonLin was used to obtain pharmacokinetic parameters in a non-

compartmental model.

4.3.3 Free DTX and DTX-NP formulation

An aqueous suspension of DTX-NP was prepared by self-assembly via a simple dialysis process.

First, 15 mg of poly(TMCC-co-LA)-g-PEG and 6 mg of DTX were dissolved together in 1.425

mL of dimethylformamide (DMF). The solution pH was then adjusted with 75 µL of 500 mM

borate buffer, pH 9.0. This mixture was then transferred to a 12-14 kDa molecular weight cut off

membrane and dialysed a minimum of four times against distilled water over 24 h at room

temperature. This process yielded polymeric nanoparticles loaded with 4.2 wt% DTXwith a z-

average diameter of 80nm as measured by dynamic light scattering (Malvern, Zetasizer). Just

prior to injection, suspensions of DTX-NP were adjusted for physiological salt content by

addition of 10× phosphate buffered saline, pH 7.4 (PBS). Free DTX was prepared by first

dissolving DTX in a mixture of ethanol and PS80 before final concentration adjustment with

PBS (10% ethanol, 7.5% PS80, 82.5% PBS) directly prior to injection.

4.3.4 Cell maintenance and preparation

MDA-MB-231-H2N cells were maintained in RPMI 1640 culture medium, supplemented with

10% heat-inactivated fetal bovine serum (FBS), 50 units/mL penicillin and 50 mg/mL

streptomycin under a humidified 5% CO2 environment. To prepare cell suspensions for

injection, adherent cells were first rinsed with PBS, and then incubated briefly with trypsin-

ethylenediamine tetraacetic acid (trypsin-EDTA, 0.25%/0.038%). Once the cells were

suspended, enzymatic digestion was inhibited with FBS, and the cells were pelleted and washed

59

3 times in PBS before resuspension at the desired concentration. Cells were kept on ice prior to

injection.

4.3.5 Tumour xenograft model

The protocols used in these in vivo studies were approved by the University Health Network

Animal Care Committee and performed in accordance with current institutional and national

regulations. Animals were housed in a 12 h light and 12 h dark cycle with free access to food and

water. NSG mice were bred in-house, and 7-9 week old female mice were selected for tumour

xenotransplantation. To form orthotopic mammary fat pad tumours, mice were inoculated with

106 MDA-MB-231-H2N cells suspended in 50 µL of sterile PBS via the following surgical

procedures. Prior to surgery, mice were anaesthetized with isoflurane-oxygen. The surgical area

was depilated and swabbed with betadine before making an incision in the skin of the lower

abdomen to the right of the midline, uncovering the mammary fat pad in the right inguinal region

where cells were injected into the fat pad. The incision was then sutured closed and lactated

Ringer’s solution and buprenorphine were given post-operatively for recovery and pain

management. Solid tumours were allowed to form over a period of 3-4 weeks. Cohorts of

tumour-bearing animals proceeded onwards for testing once their tumours reached an average

diameter of 7 mm as measured through the skin using calipers.

4.3.6 Pharmacokinetics and biodistribution

DTX-NP and free DTX were compared by giving IV doses of 1.5 mg/kg DTX or DTX

equivalent as 200 µL tail vein injections into tumour-bearing mice. Groups of 15 mice were

randomly assigned to each formulation. These groups were subdivided into 3 groups of 5 mice

with terminal end points at 2, 4, and 8 h. Each of these subgroups was placed on a staggered

blood sampling schedule such that each mouse was sampled for blood via the tail vein no more

than twice; blood samples were collected using heparinized capillary tubes and immediately

centrifuged to collect the plasma fraction. At each terminal time point, animals were sacrificed

by CO2 asphyxiation, and blood was collected via cardiac puncture using heparinized needles

and the plasma fraction was immediately isolated by centrifugation. Tissue samples (heart, lung,

liver, kidney, spleen, tumour) were also collected by dissection and placed separately in vials.

All plasma and tissue samples were snap frozen immediately after collection and kept on dry ice

before transfer to -80 ºC for long term storage.

60

4.3.7 Plasma preparation

To prepare samples for UPLC-MS, plasma samples were thawed and immediately combined

with twice their volume in acetonitrile to induce protein precipitation. The supernatant was

transferred to an MS vial and stored at 4 ºC until analysis. The plasma concentration of DTX was

calculated by comparison against blank plasma samples that were spiked with a known

concentration of DTX (125-2000 ng/mL as a two-fold dilution series).

4.3.8 Tissue preparation

Tissue samples were first thawed, accurately weighed, and transferred to vials containing beads

for homogenization (zirconia beads for spleen samples, stainless steel beads for all remaining

tissue samples). Based on the weight, a multiple of that amount was recorded and added in

distilled water (2× for each spleen vial, or 2-4× to a minimum 600 mg total weight in each of the

remaining vials) to facilitate homogenization. Samples were then vigorously agitated 3 times in

30 s on/30 s off intervals using a bead beater instrument to mechanically disrupt the tissues.

Aliquots of tissue homogenate were transferred into tubes containing double the volume in

acetonitrile for protein precipitation. The supernatant was transferred to an MS vial and stored at

4 ºC until analysis. The tissue concentration of DTX was calculated by comparison against blank

tissue homogenate samples that were spiked with a known concentration of DTX (125-2000

ng/mL as two-fold dilutions) and adjusted for the applied dilution factor.

4.4 Results

4.4.1 Pharmacokinetics

Following IV injection, drug compounds distribute through the body and are in turn metabolized

and eliminated, and these processes can be modelled with pharmacokinetic parameters using the

plasma profile. Both polymeric and conventional formulations exhibited a sharp initial drop in

plasma concentration (Figure 4.2), with nearly 90% of the detectable DTX dose leaving

circulation within 10 min. The steep initial decrease in plasma DTX concentration observed for

both polymeric nanoparticle and standard formulations is characteristic of bolus dosing followed

by rapid distribution to surrounding tissues [25]. Metabolic processes likely also contributed

because only the intact compound was measured. Remarkably, the plasma profiles diverged

significantly at 2 h post injection, with the DTX-NP formulation stabilizing at its terminal

61

elimination phase by 1 h, while the free DTX formulation continued its initial rapid distribution

phase until 2 h. By 2 h post injection, the performance of the DTX-NP formulation exceeded that

of the free DTX formulation by producing a greater than 8-fold plasma concentration difference

(3.62% vs. 0.43% initial dose remaining), widening to a 14-fold difference by 8 h (1.71% vs.

0.12% initial dose remaining).

Figure 4.2 Pharmacokinetic profiles of free DTX (o) and DTX-NP (�) in tumour-bearing mice. The

plasma profiles differ significantly by 2 h post injection. The DTX-NP formulation reached its terminal

elimination phase earlier, and coupled with a slower terminal elimination rate, the enhanced plasma

retention continued to amplify over time. Points shown are the mean of n=5 animals, with error bars

representing their standard deviation. Starred points represent statistically different group means (p <

0.05).

The improved circulation properties of the polymeric nanoparticle formulation were also reflected

in the formulation’s pharmacokinetic parameters (

Table 4.1). Even at early times, the modeled initial plasma concentration, Co, which accounts for

the instantaneous dilution due to distribution, was maintained at higher levels for the DTX-NP

62

formulation vs. free DTX, despite the doses being matched at 1.5 mg/kg. For each formulation,

the volume of distribution, Vd, was calculated to reflect the theoretical volume over which the

DTX is evenly distributed after injection. The calculated Vd for the DTX-NP formulation was

half of that for free DTX, further indicating greater retention of active DTX in plasma circulation

when delivered by polymeric nanoparticles.

Table 4.1 Pharmacokinetic parameters calculated for DTX formulations after bolus IV administration of

1.5 mg/kg DTX to tumour-bearing mice

Formulation Pharmacokinetic parameter Units Free DTX DTX np Co Initial plasma concentration ng mL-1 1.47 × 103 1.87 × 103 Vd Volume of distribution mL kg-1 4.59 × 103 2.17 × 103 t1/2, λ Lambda half life h 3.32 5.33 AUCall Area under the curve (to t = 8 h) h ng mL-1 1.49 × 103 3.52 × 103 AUC∞ Area under the curve (to t = ∞) h ng mL-1 1.57 × 103 5.31 × 103

AUMC∞ Area under the first moment curve (to t = ∞) h2 ng mL-1 2.55 × 103 3.82 × 104

Cl Clearance mL h-1 kg-1 958 282

The terminal portions of each plasma profile further distinguished the two groups. Owing to the

1.6-fold longer lambda half life, t1/2,λ, for the DTX-NP group, the profiles continued to diverge as

more time elapsed. The increasing concentration differences at later times profoundly impacted

the pharmacokinetic measures of drug exposure: AUC (area under the curve) and AUMC (area

under the first moment curve). Indeed, the AUC for concentration vs. time for the DTX-NP

group showed a greater than 2-fold increase over the 8 h observation period, and a greater than 3-

fold increase when the duration was extended to infinite time relative to free DTX. Plasma

concentrations at later times had an amplified impact on the AUMC for concentration × time vs.

time and this value increased by an order of magnitude for the DTX-NP formulation vs. the

conventional free DTX formulation. The clearance, Cl, is a measure of the blood volume that is

processed and completely cleared of the injected compound over time. Cl decreased by more

than 3-fold when DTX was formulated in polymeric nanoparticles vs. conventional PS80. This

dramatic decrease suggests that encapsulated DTX is more slowly metabolised and excreted by

63

the body. The fold changes in AUC, AUMC, Vd, and Cl values reported here are all consistent

with the ranges published elsewhere for polymeric and liposomal DTX delivery systems [26-31].

In addition to demonstrating similar biodistribution, our polymeric nanoparticles have the

advantage of having functional groups available for facile water-based chemistry, allowing

further modification [15]. Overall, our pharmacokinetic analysis indicates that with the same

initial DTX dose, greater drug exposure was achieved when the drug was formulated in

polymeric nanoparticles vs. conventional surfactants, which have the added disadvantage of

being cytotoxic and dose-limiting. As a result, the enhanced drug circulation time increased the

number of passes through the hyperpermeable tumour vasculature and likely promoted tumour

accumulation.

4.4.2 Biodistribution

To evaluate how encapsulation in our poly(TMCC-co-LA)-g-PEG nanoparticles affects tissue

distribution of DTX, a panel of organs from the same experimental groups were harvested at the

sacrificial time points. These samples were later homogenized and assayed for DTX content by

UPLC-MS. Nanoparticle formulations often accumulate in the organs rich in RES cells, such as

the liver and spleen. Remarkably, there was no significant enhancement of DTX levels in the

RES organs resulting from nanoparticle encapsulation (Figure 4.3A and B). While accumulation

in RES organs was expected, the observation that uptake was not increased in the liver or spleen,

relative to free DTX formulations, suggests that the PEG layer on the nanoparticle surface was

successful in moderating the RES response [27].

64

Figure 4.3 Biodistribution profiles of free DTX (o) and DTX-NP (�) in (A) liver, (B) spleen, (C) lung, (D)

kidney, (E) heart, and (F) tumour tissue. Points shown are the mean of n=5 animals, with error bars

representing their standard deviation. Starred points represent statistically different group means (p <

0.05).

Organs that are active in filtration, such as the lungs and kidneys, are also common sites for

nanoparticle accumulation. Filtration through the lungs resulted in high initial entrapment of

DTX for both the conventional and nanoparticle formulations, followed by rapid elimination,

with no significant differences between group means (Figure 4.3C). The lungs often act as a

filter for particles as the first capillary bed encountered after tail vein injection [32].

Consequently, the lungs did show the highest DTX concentration of all the organs tested at 2 h

post injection, but there were no significant differences between group means, and the

concentrations rapidly declined at similar rates. This suggests that a portion of each formulation

became transiently entrapped in the lung tissue, possibly due to larger nanoparticles or

aggregates, but these particles (of free drug or drug-loaded nanoparticles) subsequently cleared.

The kidneys also acted as filters for the injected DTX, which was observed in particular with the

D   E   F  

C  B  A  

65

DTX-NP formulation, with significantly elevated DTX accumulation in the kidneys throughout

the 8 h period of observation (Figure 4.3D). Although increased concentrations of DTX in the

kidneys were detected for the DTX-NP formulation, the elimination rate constant was higher

(1.3-fold increase), resulting in a projected convergence to a level equal to that of the free drug

by 24 h.

The next tissue analysed was the heart, where little accumulation was expected. The nanoparticle

formulation trended towards a reduced heart accumulation at the early 2 h time point (Figure

4.3E); although the difference was not significant, a large variance was observed for the free

DTX formulation at the 2 h post injection measurement, and a general upward trend for heart

accumulation at this early time.

Significantly greater tumour retention was observed when DTX was encapsulated in

nanoparticles starting at 4 h, and maintained at 8 h post injection (Figure 4.3F). Indeed, when the

tumour accumulation data were fitted with a first order decay, DTX-NP had a greater than five-

fold lower elimination rate constant than free DTX, demonstrating significantly greater

accumulation of DTX as a result of its delivery in the nanoparticles. This divergence of tissue

accumulation was uniquely observed in the tumour and demonstrates the benefit of DTX

delivery in nanoparticles.

4.5 Discussion

We designed the amphiphilic poly(TMCC-co-LA)-g-PEG to self-assemble into nanoparticle

micelles, where the hydrophobic biodegradable core of poly(TMCC-co-LA) allows for

hydrophobic drug encapsulation and the hydrophilic corona of PEG permits longer circulation

time by reducing protein adsorption and cellular recognition. PEG has been shown to be critical

design parameter for longer circulation: early particle formulations without PEG demonstrated

that particulate drug delivery systems were completely eliminated from circulation within

seconds to minutes [33]. The goal of longer circulation is to achieve greater and selective tumour

accumulation. In fact, additional passes through the hyperpermeable vasculature associated with

solid tumours generally enhances tumour accumulation [34,35]. Enhanced circulation of our

DTX-NP was verified using standard PK parameters, demonstrating that greater exposure was

achieved with this formulation, even in this model where there was a single injection of a fixed

dose. Importantly, enhanced circulation may also increase systemic exposure and general

66

toxicity because drug activity is not limited to cancer cells. As a result, there is a compromise

between these opposing factors that requires a balance between high tumour accumulation and

low systemic distribution [36]. Importantly, we observed both longer circulation and greater

tumour accumulation of these polymeric nanoparticles.

The overall distribution profile of the DTX-NP formulation is encouraging, as the poly(TMCC-

co-LA)-g-PEG nanoparticles established a strong contrast between accumulation in diseased

tumour tissues and clearance from healthy tissues, likely due to their engineered material

properties. For example, the low critical micelle concentration, measured at 3 µg/mL [15],

exceeds the injected concentration by three orders of magnitude, which likely allowed a

significant portion of polymeric nanoparticles to circulate intact, instead of rapidly disassembling

after dilution in blood. The serum stability of poly(TMCC-co-LA)-g-PEG nanoparticles has also

been confirmed in vitro using biologically relevant media, further suggesting high stability of the

nanoparticles and their encapsulated load in circulation [17]. PEG itself has several important

properties. Firstly, the 10 kDa molar mass exceeds the typical 1-5 kDa range that is commonly

used, lowering the PEG density required to reach the more effective brush regime for enhanced

circulation [37,38]. Liposomal systems are limited to lower PEG molecular weights because

longer PEG chains compromise liposomal stability. Polymeric systems, such as our poly(TMCC-

co-LA)-g-PEG nanoparticles, can stably incorporate higher molar mass PEG by manipulating the

molar mass of the hydrophobic region [12]. Secondly, each poly(TMCC-co-LA) chain is

modified with an average of one PEG chain [15], leading to excellent coverage of the

nanoparticle core, which is further reflected by the nearly neutral surface charge of the

assembled nanoparticles [16]. Thirdly, the amide bond between PEG and poly(TMCC-co-LA) is

one of the more serum-stable bonds [39], ensuring lasting nanoparticle coverage after injection.

While all particle systems are ultimately cleared, ideally adequate tumour accumulation is

reached prior to clearance. This process is normally triggered by degradation or erosion of the

polymer comprising the nanoparticle [32]. The poly(TMCC-co-LA)-g-PEG nanoparticles are

subject to eventual erosion because the polymer chains are not cross-linked. The polymer chains

are of sufficiently low molar mass (<30 kDa) to be cleared by the kidneys [40] and are

themselves biodegradable. These design elements each provide qualities that favour solid tumour

drug delivery while minimizing systemic accumulation.

67

Thorough biodistribution analysis allowed us to quantify the final concentration of DTX in

different organs. Notably, DTX did not accumulate in RES organs (liver or spleen) over free

DTX controls, suggesting that the inclusion of PEG successfully reduced the expected RES

response to foreign particles. The DTX-NP formulation did increase the kidney accumulation

over free DTX, pointing to a partial shift to renal clearance, where free DTX is mainly

metabolized and excreted by biliary clearance [41,42].

Analysis of the heart demonstrated lower variability of DTX accumulation when administered

via polymeric nanoparticles vs. as free DTX. In clinical use, taxanes such as DTX are commonly

paired with doxorubicin to treat metastatic breast cancer, but cardiotoxicity is a primary side

effect of this drug combination. When used in combination therapy, their administration is

staggered, thereby lowering concurrent levels of both drugs to reduce this interaction and

reducing cardiotoxicity [43]. Consistently lower initial accumulation in the heart (as was

observed with DTX-NP) may allow increased flexibility in the dosing schedule without the risk

of introducing severe cardiotoxicity.

The tumour specificity of the DTX-NP formulation is of particular interest. The greater retention

of DTX in the tumour when delivered in our polymeric nanoparticles is consistent with the EPR

effect observed with other particulate systems. Because this behaviour was uniquely observed in

the tumour tissue, there is potential for specific cytotoxic impact on cancer cells while the drug

compound is eliminated from the rest of the body. Our system also compared favourably with

radiolabeled liposomal DTX formulations: the latter delivered DTX and/or its metabolites to

subcutaneous tumours at levels between 2 and 8% initial dose/g, 6 h post injection, depending in

the extent of PEGylation [27], whereas our DTX-NP delivered 5% initial dose/g at 6 h post

injection based on fitted values from a first order decay for DTX-NP. Similarly, folate targeted

PEGylated liposomes achieved 7% initial dose/g of non-degraded DTX 4 h post injection to

intradermal tumours measured by LC-MS [28], vs. our 6% initial dose/g for DTX-NP at 4 h.

Indeed, our polymeric nanoparticle formulation delivered an active anti-cancer drug to solid

tumours with improved retention over the free drug alone, and this has important implications for

anti-tumour efficacy. By extending the drug exposure of diseased cells, the required cumulative

dose will likely decrease, as DTX cytotoxicity depends on both concentration and contact time

[44]. Moreover, the dose used in this study (1.5 mg/kg) is consistent with DTX levels

68

administered in metronomic dosing schedules in ovarian cancer models, where several small

doses were given at high frequency [45]. This strategy is in contrast with mean tolerated dose

approaches, where the highest tolerated dose is given at low frequency to allow healthy tissues to

recover between treatments. Metronomic dosing was shown to reduce both systemic toxicity and

cumulative dose while improving anti-tumour efficacy, under the premise that the cancer cells

have less recovery time between dosing. In combination with more sustained drug levels at each

dose, the polymeric nanoparticle formulation offers great potential to demonstrate improved

efficacy as a result of greater targeting. Moreover, the nanoparticle DTX delivery system

obviates the use of conventional surfactant-based delivery systems which themselves are

cytotoxic and cause systemic toxicity.

4.6 Conclusions

Nanoparticle formulations for anti-cancer drugs are designed to couple specific tumour tissue

accumulation with quick elimination from healthy organs. Our DTX-loaded poly(TMCC-co-

LA)-g-PEG nanoparticles achieved this with increased DTX accumulation in the tumour and

simultaneous DTX clearance from other organs to which it distributed over an 8 h period of

observation. The pharmacokinetic profile of the polymeric nanoparticle formulation vs. free

DTX demonstrated improved circulation properties at later times, which likely contributed to the

favourable tumour accumulation of DTX when delivered via our engineered formulation. The

specific retention of DTX in tumour tissue suggests that this polymeric nanoparticle delivery

strategy will be efficacious against solid tumours.

4.7 Acknowledgements

We thank: Drs. Robert Kerbel (Sunnybrook Health Science Centre), Armand Keating and Yoko

Kosaka (Princess Margaret Hospital) for their help and advice in establishing the mouse tumour

model; and Mr. Peter Poon (University of Toronto) for help in growing the tumours and

obtaining serum and tissue samples. We are grateful to the Canadian Institutes of Health

Research (CIHR to MSS, RSA) the Ontario Centres of Excellence (OCE to MSS), the Ontario

Institute for Cancer Research (OICR to RSA), the Natural Sciences and Engineering Research

Council (NSERC to KSH), and the Government of Ontario/DuPont Canada Scholarship in

Science and Technology (OGSST to KSH) for funding of this research.

69

4.8 References

[1] Cairns R, Papandreou I, Denko N. Overcoming physiologic barriers to cancer treatment by molecularly targeting the tumor microenvironment. Mol Cancer Res 2006;4:61-70.

[2] Dreher MR, Liu WG, Michelich CR, Dewhirst MW, Yuan F, Chilkoti A. Tumor vascular permeability, accumulation, and penetration of macromolecular drug carriers. J Natl Cancer I 2006;98:335-344.

[3] Minchinton AI, Tannock IF. Drug penetration in solid tumours. Nat Rev Cancer 2006;6:583-592.

[4] Felici A, Verweij J, Sparreboom A. Dosing strategies for anticancer drugs: the good, the bad and body-surface area. Eur J Cancer 2002;38:1677-1684.

[5] Matsumura Y, Maeda H. A new concept for macromolecular therapeutics in cancer-chemotherapy - mechanism of tumoritropic accumulation of proteins and the antitumor agent smancs. Cancer Res 1986;46:6387-6392.

[6] Torchilin V. Tumor delivery of macromolecular drugs based on the EPR effect. Adv Drug Deliver Rev 2011;63:131-135.

[7] Mamot C, Drummond DC, Hong K, Kirpotin DB, Park JW. Liposome-based approaches to overcome anticancer drug resistance. Drug Resist Update 2003;6:271-279.

[8] Majoros IJ, Williams CR, Baker JR. Current dendrimer applications in cancer diagnosis and therapy. Curr Top Med Chem 2008;8:1165-1179.

[9] Allen C, Maysinger D, Eisenberg A. Nano-engineering block copolymer aggregates for drug delivery. Colloid Surface B 1999;16:3-27.

[10] Kataoka K, Bae Y. Intelligent polymeric micelles from functional poly(ethylene glycol)-poly(amino acid) block copolymers. Adv Drug Deliver Rev 2009;61:768-784.

[11] Sahoo SK, Acharya S. PLGA nanoparticles containing various anticancer agents and tumour delivery by EPR effect. Adv Drug Deliver Rev 2011;63:170-183.

[12] Photos PJ, Bacakova L, Discher B, Bates FS, Discher DE. Polymer vesicles in vivo: correlations with PEG molecular weight. J Control Release 2003;90:323-334.

[13] Geng Y, Dalhaimer P, Cai SS, Tsai R, Tewari M, Minko T, et al. Shape effects of filaments versus spherical particles in flow and drug delivery. Nat Nanotechnol 2007;2:249-255.

[14] Sailor MJ, Park JH, von Maltzahn G, Zhang LL, Schwartz MP, Ruoslahti E, et al. Magnetic iron oxide nanoworms for tumor targeting and imaging. Adv Mater 2008;20:1630-1635.

[15] Shi M, Wosnick JH, Ho K, Keating A, Shoichet MS. Immuno-polymeric nanoparticles by Diels-Alder chemistry. Angew Chem Int Edit 2007;46:6126-6131.

70

[16] Lu J, Shoichet MS. Self-assembled polymeric nanoparticles of organocatalytic copolymerizated D,L-lactide and 2-methyl 2-carboxytrimethylene carbonate. Macromolecules 2010;43:4943-4953.

[17] Lu J, Owen SC, Shoichet MS. Stability of self-assembled polymeric micelles in serum. Macromolecules 2011;44:6002–6008.

[18] Torchilin VP. Targeted polymeric micelles for delivery of poorly soluble drugs. Cell Mol Life Sci 2004;61:2549-2559.

[19] Aapro M, Bruno R. Early clinical-studies with docetaxel. Eur J Cancer 1995;31A:S7-S10.

[20] Engels FK, Mathot RAA, Verweij J. Alternative drug formulations of docetaxel: a review. Anti-Cancer Drug 2007;18:95-103.

[21] Hennenfent KL, Govindan R. Novel formulations of taxanes: a review. old wine in a new bottle? Ann Oncol 2006;17:735-749.

[22] Xu RNX, Fan LM, Rieser MJ, El-Shourbagy TA. Recent advances in high-throughput quantitative bioanalysis by LC-MS/MS. J Pharmaceut Biomed 2007;44:342-355.

[23] Gupta N, Price PM, Aboagye EO. PET for in vivo pharmacokinetic and pharmacodynamic measurements. Eur J Cancer 2002;38:2094-2107.

[24] Shi M, Ho K, Keating A, Shoichet MS. Doxorubicin-conjugated immuno-nanoparticles for intracellular anticancer drug delivery. Adv Funct Mater 2009;19:1689-1696.

[25] Gunaratna C. Drug metabolism and pharmacokinetics in drug discovery: a primer for bioanalytical chemists, part II. Current Separations 2001;19:87-92.

[26] Esmaeili F, Dinarvand R, Ghahremani MH, Ostad SN, Esmaily H, Atyabi F. Cellular cytotoxicity and in-vivo biodistribution of docetaxel poly(lactide-co-glycolide) nanoparticles. Anti-Cancer Drug 2010;21:43-52.

[27] Khalid MN, Simard P, Hoarau D, Dragomir A, Leroux JC. Long circulating poly(ethylene glycol)-decorated lipid nanocapsules deliver docetaxel to solid tumors. Pharm Res-Dord 2006;23:752-758.

[28] Li X, Tian X, Zhang J, Zhao X, Chen XH, Jiang YH, et al. In vitro and in vivo evaluation of folate receptor-targeting amphiphilic copolymer-modified liposomes loaded with docetaxel. Int J Nanomed 2011;6:1167-1184.

[29] Mu CF, Balakrishnan P, Cui FD, Yin YM, Lee YB, Choi HG, et al. The effects of mixed MPEG-PLA/Pluronic (R) copolymer micelles on the bioavailability and multidrug resistance of docetaxel. Biomaterials 2010;31:2371-2379.

[30] Upadhyay KK, Mishra AK, Bhatt AN, Castro E, Chuttani K, Dwarakanath BS, et al. In vitro and in vivo evaluation of docetaxel loaded biodegradable polymersomes. Macromol Biosci 2010;10:503-512.

71

[31] Yuan ZG, Chen DW, Zhang ST, Zheng ZD. Preparation, characterization and evaluation of docetaxel-loaded, folate-conjugated PEG-liposomes. Yakugaku Zasshi 2010;130:1353-1359.

[32] Moghimi SM, Hunter AC, Murray JC. Long-circulating and target-specific nanoparticles: theory to practice. Pharmacol Rev 2001;53:283-318.

[33] Gref R, Minamitake Y, Peracchia MT, Trubetskoy V, Torchilin V, Langer R. Biodegradable long-circulating polymeric nanospheres. Science 1994;263:1600-1603.

[34] McNeeley KM, Annapragada A, Bellamkonda RV. Decreased circulation time offsets increased efficacy of PEGylated nanocarriers targeting folate receptors of glioma. Nanotechnology 2007;18:Article 385101/385101-385111.

[35] Senthilkumar M, Mishra P, Jain NK. Long circulating PEGylated poly(D,L-lactide-co-glycolide) nanoparticulate delivery of docetaxel to solid tumors. J Drug Target 2008;16:424-435.

[36] Karmali PP, Simberg D. Interactions of nanoparticles with plasma proteins: implication on clearance and toxicity of drug delivery systems. Expert Opin Drug Del 2011;8:343-357.

[37] Nicholas AR, Scott MJ, Kennedy NI, Jones MN. Effect of grafted polyethylene glycol (PEG) on the size, encapsulation efficiency and permeability of vesicles. Bba-Biomembranes 2000;1463:167-178.

[38] Tirosh O, Barenholz Y, Katzhendler J, Priev A. Hydration of polyethylene glycol-grafted liposomes. Biophys J 1998;74:1371-1379.

[39] Parr MJ, Ansell SM, Choi LS, Cullis PR. Factors influencing the retention and chemical-stability of poly(ethylene glycol)-lipid conjugates incorporated into large unilamellar vesicles. Bba-Biomembranes 1994;1195:21-30.

[40] Etrych T, Kovar L, Subr V, Braunova A, Pechar M, Chytil P, et al. High-molecular-weight polymers containing biodegradable disulfide bonds: synthesis and in vitro verification of intracellular degradation. J Bioact Compat Pol 2010;25:5-26.

[41] Bardelmeijer HA, Roelofs ABGH, Hillebrand MJX, Beijnen JH, Schellens JHM, van Tellingen O. Metabolism of docetaxel in mice. Cancer Chemoth Pharm 2005;56:299-306.

[42] Sparreboom A, VanTellingen O, Scherrenburg EJ, Boesen JJB, Huizing MT, Nooijen WJ, et al. Isolation, purification and biological activity of major docetaxel metabolites from human feces. Drug Metab Dispos 1996;24:655-658.

[43] Salvatorelli E, Menna P, Cascegna S, Liberi G, Calafiore AM, Gianni L, et al. Paclitaxel and docetaxel stimulation of doxorubicinol formation in the human heart: implications for cardiotoxicity of doxorubicin-taxane chemotherapies. J Pharmacol Exp Ther 2006;318:424-433.

[44] Hill BT, Whelan RDH, Shellard SA, Mcclean S, Hosking LK. Differential cytotoxic effects of docetaxel in a range of mammalian tumor-cell lines and certain drug-resistant sublines in-vitro. Invest New Drug 1994;12:169-182.

72

[45] Kamat AA, Kim TJ, Landen CN, Lu CH, Han LY, Lin YG, et al. Metronomic chemotherapy enhances the efficacy of antivascular therapy in ovarian cancer. Cancer Res 2007;67:281-288.

73

5 Antibody-modified nanoparticles for active and tunable binding to cancer cells

This chapter is derived from the following manuscript:

Ho K, Lapitsky Y, Shi M, and Shoichet MS (2009). Tunable immunonanoparticle binding to cancer cells: thermodynamic analysis of targeted drug delivery vehicles. Soft Matter, 5(5): 1074-80.

Reprinted with permission from The Royal Society of Chemistry.

5.1 Abstract

Tumour cells are often associated with altered surface receptor profiles, and these changes can

provide a basis for targeted delivery of anti-cancer agents. Functionalizing a colloidal drug

delivery vehicle, such as a polymeric nanoparticle, with several targeting ligands has

qualitatively been shown to increase the effective affinity of the nanoparticle for its target

receptor over the affinity of the free ligand. However, whether this increase results from multiple

simultaneous interactions per particle (multivalent binding) or increased configurations for single

binding events per particle (monovalent binding) is unclear. A quantitative approach was

required to distinguish between these possible mechanisms. In this study, human epidermal

growth factor receptor 2 (HER2) overexpressing cancer cells (SKBR-3) were used as the target

for anti-HER2 (trastuzumab, HerceptinTM) immunonanoparticles. We varied the antibody

conjugation density on the immunonanoparticles and measured their cellular binding by a flow

cytometric immunoassay. Using this method, we were able to directly assay the targeted cells

and quantify immunonanoparticle binding strength, allowing us to better understand whether

immunonanoparticles were bound by monovalent or multivalent interactions. The binding data

for each formulation were fitted to Langmuir isotherms, and based on the theory presented

herein, it was concluded that the system studied behaved in a manner consistent with monovalent

binding. Understanding this property of immunonanoparticle binding is useful in drug delivery

applications, where manipulating the strength of such interactions is essential to controlling their

targeting capacity on both tissue and cellular levels. The models developed here can be used to

quantitatively predict binding strength for rational immunonanoparticle design.

74

5.2 Introduction

The development of targeted drug carriers is driven by limitations identified with the free

administration of anti-cancer agents, including short plasma half lives, systemic toxicity, and

mass transport barriers restricting accumulation at tumour sites [1–4]. The altered phenotype of

cancer cells often includes changes to their surface receptor profiles, providing a basis for active

targeting using monoclonal antibodies that recognize and bind specific receptors with elevated

expression levels [1,5,6]. Covalent attachment of such antibodies to polymeric drug carriers

allows their guided transport to the surfaces of targeted cells, while the polymer is designed to

protect drug bioactivity, increase circulation time, and shield healthy cells from cytotoxic agents

[2–4,7,8]. Moreover, binding can enhance retention at tumour sites and can introduce a means

for rapid receptor-mediated internalization of drug-loaded carriers into the intracellular

compartment, a common site of action for cytotoxic drugs [2,4,9,10].

In the case of colloidal drug-loaded polymer aggregates, including self-assembled polymeric

nanoparticles, every polymer chain can participate in drug delivery, but the direct modification

of each polymer chain with a targeting antibody becomes unnecessary; unmodified polymer

chains can be targeted as members of a modified aggregate, and fewer targeting antibodies are

then required overall [11]. The number of antibodies per aggregate can be controlled by varying

the reaction conditions during their attachment (e.g., reaction time, feed ratio of antibody to

polymer) [6]. Enhanced binding strength of immunonanoparticles over free antibody can occur

through two possible mechanisms: the introduction of multiple simultaneous interactions per

particle (multivalent binding, Figure 5.1A) or the increase in possible configurations for single

binding events per particle (monovalent binding, Figure 5.1B).

75

Figure 5.1 Functionalizing immunonanoparticles with greater numbers of targeting antibodies enhances

their ability to associate with target cells. This effect can result from (A) increases in binding events per

particle (multivalent binding) or (B) increases in possible binding configurations with a single interaction

(monovalent binding). Illustrated here are immunonanoparticles with Ω = 3 attached antibodies. In (A),

the number of antibodies bound to cell receptors, α, is shown as α = 3 (left) and α = 2 (right). In (B) the

number of antibodies bound to cell receptors, α, is shown as α =1 for all nanoparticles. The mechanism

is an important consideration in immunonanoparticle design, as it dictates how binding strength will

increase as the antibody conjugation density increases.

Multivalent binding events would greatly enhance binding through avidity, where the

presentation of multiple tethers to the cell surface maintains association and cell-particle

proximity after a single dissociation event, thereby promoting re-attachment [12]. A more

moderate increase in binding strength is associated with monovalent binding [13]. The dramatic

increase in binding strength associated with avidity is a phenomenon that would be most

beneficial for antibodies that have poor affinity with their targets [8,14,15]. Conversely, in cases

where the binding affinity is very high, there is decreased utility in treating solid tumours, where

complete tumour penetration can be limited by strong association with cells directly adjacent to

tumour vasculature [7,16,17]. Heterogeneity of targeting within the tumour mass leads to

incomplete eradication of tumour cells, as certain cells will receive either no drug or levels of

drug below the therapeutic index [16]. Furthermore, binding strength influences the cell-

associated fraction at a given particle concentration, and this information helps assess whether

the particle dosage administered delivers drug levels within the therapeutic index [18].

Whether multivalent interactions can occur is predicated on both the density of the targeted

receptor on the cell surface and the density of binding sites on the drug carrier. Quantitative

A B

76

avidity measurements have been performed using receptors immobilized onto a hard synthetic

substrate [19–21]. However, interactions with cells can be dramatically different, in part because

lateral movement of receptors within the cell membrane can result in transient increases in local

receptor density; this phenomenon has previously been shown with solutions of free antibody

[22]. Having a flexible polymeric spacer, such as poly(ethylene glycol) (PEG), between the

nanoparticle core and the targeting molecule, provides the latter with greater free volume (and

thus greater likelihood) to interact with cellular receptors [23].

A simple, quantitative method to investigate the binding interactions between cells and

immunonanoparticles by directly assaying the targeted cells is lacking. We developed a flow

cytometric immunoassay to assess equilibrium cell binding of an anti-human epidermal growth

factor receptor 2 (anti-HER2) immunonanoparticle system. HER2 is a cell surface receptor that

becomes overexpressed in 20–30% of breast cancer cases. It is used as an indication for

treatment with Herceptin (trastuzumab) [24–27], which is the anti-HER2 monoclonal antibody

conjugated to the polymeric nanoparticles in this study. SKBR-3 cells were chosen as an in vitro

model of HER2 overexpression, and express 1 × 106 receptors/cell [28]; using this model, we

have shown previously that the binding of Herceptin-immunonanoparticles is receptor specific,

with little non-specific adsorption [6]. Here we varied the number of conjugated antibodies per

nanoparticle, measured dose responsive binding, and, by fitting binding isotherms to each,

quantified how the cell-particle binding varies with the density of conjugated antibodies.

Thermodynamic analysis of these variations elucidates the nature of the binding events between

our anti-HER2 immunonanoparticles and HER2 overexpressing SKBR-3 cells in culture. This

reveals a linear scaling between the immunonanoparticle binding strength and the number of

conjugated antibodies, thereby providing quantitative guidelines for tuning cell-particle

interactions in the design of colloidal vehicles for targeted drug delivery.

5.3 Experimental

5.3.1 Materials

All cell culture materials were purchased from Gibco-Invitrogen (Burlington, ON, Canada).

SKBR-3 cells were obtained through ATCC (Manassas, VA, USA). Dialysis membranes were

acquired from Spectrum Laboratories (Rancho Dominguez, CA, USA). The Herceptin antibody

was purchased through Hoffmann-La Roche Limited (Mississauga, ON, Canada). The polymeric

77

nanoparticles were synthesized as previously described [6,29]. All other materials were

purchased from Sigma-Aldrich (Mississauga, ON, Canada) and used as received unless

otherwise noted.

5.3.2 Nanoparticle synthesis

An aqueous suspension of poly(2-methyl-2-carboxytrimethylene carbonate-co-D,L-lactide)-

graft-poly(ethylene glycol)-furan (poly(TMCC-co-LA)-g-PEG-furan) nanoparticles was

prepared by dialysis, as reported previously [6,29]. Briefly, the polymer was first dissolved in a

mixture of 95 vol% dimethylformamide (DMF) and 5 vol% 500 mM borate buffer, pH 9.0 at a

final concentration of 10 mg/mL; the solution was then dialysed a minimum of four times against

distilled water at room temperature over 24 h using a 12–14 kDa molecular weight cut off

membrane. This procedure yielded nanoparticles with a mean hydrodynamic diameter of 80 nm

as measured by dynamic light scattering (Brookhaven 90Plus Particle Sizer, Brookhaven

Instruments, Holtsville, NY, USA) with the hydrophobic poly(TMCC-co-LA) backbone

comprising the nanoparticle core, and the hydrophilic, flexible, furan-terminated PEG grafts

comprising the nanoparticle shell. Site-specific chemical modification of carbohydrates on the Fc

region of the Herceptin antibody provided a maleimide functional group, allowing covalent

attachment of Herceptin to the PEG-furan termini through Diels–Alder chemistry [6,29].

Specifically, a furan-functionalized nanoparticle solution (4 mg in 1 mL of distilled water) was

mixed with maleimide modified Herceptin (100 mg in 120 µL of 100 mM β-

morpholinoethanesulfonic acid (MES) buffer, pH 5.5) and incubated at 37 ºC. By adjusting the

reaction time (20 min, 1, 2, and 4 h), the antibody conjugation density was varied to have an

average of 1.9 ± 0.3, 3.2 ± 0.5, 5.9 ± 0.2, and 9.4 ± 0.9 antibodies/nanoparticle, based on a 95%

confidence interval. The average values were estimated as previously reported, based on the

hydrodynamic particle diameter, by comparing the fluorescence intensity of the Alexa Fluor®

430-labelled Herceptin and immunonanoparticles made by reaction with this fluorescent

Herceptin [6]. The resulting immunonanoparticles were then purified using a Sephacryl S-

300HR column equilibrated in phosphate buffered saline, pH 7.4 (PBS).

5.3.3 Cell lines and maintenance

SKBR-3 cells were maintained in McCoy’s 5A culture medium, supplemented with 10% heat-

inactivated fetal bovine serum (FBS), 50 units/mL penicillin and 50 mg/mL streptomycin under a

78

humidified 5% CO2 environment. To prepare cell suspensions, adherent cells were first rinsed

with PBS, then incubated briefly with trypsin-ethylenediamine tetraacetic acid (trypsin-EDTA,

0.25%/0.038%). Once the cells were suspended, enzymatic digestion was inhibited with FBS,

and the cells were pelleted and resuspended at the desired concentration.

5.3.4 Flow cytometric analysis

To quantify immunonanoparticle binding, a fluorescently labeled secondary antibody was used

for detection. Subsequent intensity measurements could then be carried out on a cell by cell basis

by fluorescence activated cell sorting (FACS), where the intensity values are proportional to the

number of bound immunonanoparticles. To do this, SKBR-3 cells were first suspended, as

described above, in PBS at a final concentration of 1 × 106 cells/mL and distributed into 200 µL

aliquots in 1.7 mL centrifuge tubes. The cells were then incubated for 30 min at 4 ºC to inhibit

endocytosis (cellular uptake), and resuspended in 200 µL of immunonanoparticle solution at

varying concentrations in triplicate. The cells were again incubated for 30 min at 4 ºC to reach

equilibrium binding [22,30,31], and washed with 1 mL of cold PBS, pelleted, and resuspended in

50 µL of FACS buffer (PBS supplemented with 1% FBS and 2 mM EDTA). Rabbit antihuman

immunoglobulin G-fluorescein isothiocyanate (IgG-FITC) secondary antibody was diluted 1/200

in PBS, and 10 µL was added to each of the cell suspensions. After 30 min incubation at 4 ºC,

the cells were washed with 1 mL cold FACS buffer, resuspended in 550 µL of fresh FACS

buffer with 0.6 mg/mL propidium iodide (PI), and transferred to 5 mL FACS tube for analysis.

Data acquisition was performed on a FACS Calibur (Becton Dickinson, Mississauga, ON,

Canada) and analysis was performed using CellQuest software (Becton Dickinson). The first

10,000 events were recorded, and the live cell population was gated for analysis of FITC

fluorescence intensity. All values shown are the average of triplicate samples with error bars

representing their standard deviation.

5.4 Theory

As a first order approximation, the binding of polymeric immunonanoparticles (diameter ~ 100

nm) to much larger cancer cells (diameter ~ 10 mm) can be quantified using the Langmuir

binding isotherm, where the binding strength is quantified using a single equilibrium binding

constant, Keq:

79

NPeq

NPeq

CKCK

+=1

θ (1)

Here, θ represents the fraction of the occupied binding surface on the cell and NPC is the solution

nanoparticle concentration. The number of spaces available for nanoparticle binding is

dependent on the density of the targeted receptor, and in the case of very high receptor

expression levels, is sterically limited by the volumes occupied by bound nanoparticles [32].

The available potential binding space can be further limited in cases where receptor clustering

occurs. Indeed, cancer cells with varying levels of receptor overexpression have previously been

shown to have similar saturation concentrations, likely because further binding is sterically

hindered [18]. The isotherm accounts for these variations by defining θ as a fraction of

saturation, the upper limits of which are influenced by receptor density and distribution, as well

as particle size.

The binding constant, Keq (which is often also reported as its reciprocal, the dissociation

constant, Kd), is related to the molar Gibbs free energy of binding ( GΔ ) via:

⎟⎟⎠

⎞⎜⎜⎝

⎛ Δ−=RTGKeq exp (2)

The variations in eqK and GΔ with the number of antibodies on the nanoparticle surface is

influenced by two primary effects: (1) multivalent cell-nanoparticle interactions (avidity) and (2)

an increase in the number of possible monovalent binding configurations for a single cell-

nanoparticle pair. The thermodynamic analysis of these two mechanisms is outlined below.

5.4.1 Multivalent Binding

Multivalent binding enhances binding stability by establishing multiple tethers to the binding

surface; when a single interaction is disrupted, the remaining interactions maintain cell-particle

proximity, thereby promoting subsequent re-attachment. In the case where the number of

antibodies on the nanoparticle surface, Ω, affects the valency of the cell-particle interaction

(Figure 5.1A), the molar Gibbs free energy of nanoparticle binding is roughly proportional to the

average number of antigen-antibody interactions, α [33]:

80

)1(~)( GG ΔΩΔ α (3)

Where )(ΩΔG is the molar Gibbs free energy of the α-valent antibody-antigen binding and

)1(GΔ is the molar Gibbs free energy of binding for a single antigen-antibody pair. This

linearity reflects the additivity of the antigen-antibody interactions. However, deviations from

this relationship can exist due to cooperative and anti-cooperative interactions between the

coupled antigen-antibody pairs [33]. Combining this result with Equation 2 indicates an

exponential dependence between eqK and the number of interacting nanoparticles:

α)1(~)( eqeq KK Ω (4)

Thus, multivalent interactions can result in an increase of several orders of magnitude in the

immunonanoparticle binding strength as the number of conjugated antibodies is increased, with α

being a function of Ω. In the event of infrequent multivalent interactions (1 < α < 2), this

increase will be less dramatic, but the value for eqK should increase exponentially.

5.4.2 Monovalent Binding

In contrast to multivalent binding, for monovalent interaction, the amplified binding affinity may

reflect an increase in the number of unique configurations for a single nanoparticle with Ω

conjugated antibodies to bind to the cell surface (Figure 5.1B and Figure 5.5, Supplementary

Information). In this case, the binding strength can be approximated theoretically by defining a

canonical partition function, Q(M,N,T), for N nanoparticles binding to M binding sites [34]:

( ) ⎟⎟

⎠

⎞⎜⎜⎝

⎛ −×

−×Ω=

TkN

NMNMTMNQ

B

N εexp!!

!),,( (5)

where the first term accounts for the internal degrees of freedom of N nanoparticles bound to the

cells, the second term accounts for the number of lattice configurations in which these particles

81

bind to M sites, and the third term accounts is the Boltzmann factor for N nanoparticles binding

to the cells with the molecular energy, ε (Figure 5.5, Supplementary Information).

The chemical potential of the cell-bound nanoparticles ( Aµ ) can be calculated using the

following relationship [34]:

TM

BA NTMNQTk

,

),,(ln⎟⎠

⎞⎜⎝

⎛∂

∂−=µ (6)

This yields the expression:

⎟⎠

⎞⎜⎝

⎛−

+Ω−=θ

θεµ

1lnln TkTk BBA (7)

where θ is equal to N/M. Because at equilibrium this chemical potential is equal to the chemical

potential of the nanoparticles in solution (i.e., NPBSS CTk ln0, += µµ ), θ can be solved as a

function of the free nanoparticles in solution, NPC :

NPeq

NPeq

CKCK)(1)(Ω+

Ω=θ (8a)

where

⎟⎟⎠

⎞⎜⎜⎝

⎛ −⋅Ω=Ω

TkK

B

deq

εµ 0,exp)( (8b)

From these expressions for variations in eqK and GΔ with Ω are obtained as:

)1()( eqeq KK Ω=Ω (9)

Ω−Δ=ΩΔ ln)1()( RTGG (10)

82

Thus, in the absence of multivalent interactions (and the presence of monovalent interactions),

eqK is predicted to increase linearly with Ω, and GΔ to vary logarithmically through the

proportionality constant, RT.

5.5 Results and Discussion

Over the range of Herceptin and nanoparticle concentrations studied, Herceptin

immunonanoparticles (bearing between 1.9 and 9.4 antibodies per particle) bound to the SKBR-3

cells in a dose dependent manner. This binding was detected on FACS using a FITC-conjugated

secondary antibody against Herceptin. Secondary antibody detection of primary binding events

is a common technique used for FACS analysis and has been shown to have greater sensitivity

than directly labeling the primary antibody [35].

The binding assay was performed at 4 °C to inhibit cellular internalization of Herceptin

immunonanoparticles [36]. By excluding cellular uptake, cellular interactions included only

binding and dissociation events, thereby allowing the measurement of the equilibrium binding

isotherm. Importantly, surface bound immunonanoparticles were accessible to the secondary

antibody used in the FACS analysis and thus did not require permeabilization of the cell

membrane for detection. Because the fluorescence intensity is proportional to the number of

immunonanoparticles bound to cells, the binding constant eqK and the saturation fluorescence

intensity MAXI can be fitted using the Langmuir model via:

NPeq

NPeqMAX

NP CKCKI

CI+

=1

)( (11)

where )( NPCI is the concentration-dependent measured fluorescence intensity. The fractional

coverage, θ, can then be calculated by dividing the measured fluorescence intensity by the

saturation fluorescence intensity. The equilibrium binding of single Herceptin antibodies and

Herceptin immunonanoparticles to SKBR-3 cells closely follow the Langmuir isotherms

indicated by the solid lines (Figure 5.2). The R2 value for each fitted line exceeds 0.95.

83

Figure 5.2 Fractional coverage of Herceptin immunonanoparticles bound to HER2 overexpressing SKBR-

3 cells as a function of immunonanoparticle concentration. The arrow indicates ascending order of

antibody conjugation density: Herceptin immunonanoparticles bearing 1.9 (�), 3.2 (r), 5.9 (�), and 9.4

(n) antibodies; inset shows fractional coverage for free Herceptin (p).

The fitted eqK values increase linearly with the antibody conjugation density from 0.11 nM-1 for

single Herceptin antibodies (which are likely similar to those that would be obtained from

immunonanoparticles bearing 1.0 antibody per particle) to 1.03 nM-1 for the

immunonanoparticles bearing 9.4 antibodies per particle (Figure 5.3A). This 10-fold increase in

eqK corresponds to a nearly 10-fold increase in the number of antibodies available per

immunonanoparticle. These variations agree well with the model for monovalent binding

behaviour, described by Equation 9, where the cell-particle affinity increases due to an amplified

number of unique binding states. Likewise, the variations in GΔ (Figure 5.3B) are in remarkable

agreement with logarithmic scaling (Equation 10), where a fitted proportionality constant

(1.01RT) is within 1% of the theoretical value (RT) obtained, along with a )1(GΔ of – 51.5

84

kJ/mol. No threshold antibody density for binding was observed, which is consistent with

previous reports on targeted particles having flexible spacers between the targeting molecule and

the core [32, 37, 38]. This immunonanoparticle approach yields the flexibility to use a particular

IgG antibody for a targeting application with eqK up to an order of magnitude greater than its

original value in a tunable manner.

Figure 5.3 (A) eqK and (B) GΔ increase in absolute value as the number of Herceptin antibodies per

nanoparticle increases, thereby indicating greater binding affinity. The open symbols (¯) represent the

values calculated for free Herceptin, which denotes a monovalent case, and the closed symbols (u)

represent Herceptin immunonanoparticles. The trends in (A) and (B) follow the theoretical behaviour of

monovalent immunonanoparticle binding.

85

The variations in eqK and GΔ do not support the occurrence of multivalent binding, which is

predicted to give rise to a much more dramatic, near-exponential increase in eqK and a near-

linear increase in GΔ in the case where all antibodies on a bound immunonanoparticle

participate in binding (α = Ω, Equations 3 and 4). Even in the case where only a fraction of

antibodies are bound (α < Ω), if they occur with great enough frequency to influence the average

system behaviour, the increased binding valency should strengthen binding over the predicted

values for increased binding configurations given by the monovalent binding model. Instead, the

monovalent binding model accurately described the magnitude of the increases in eqK without

the need to account for contributions due to multivalent interactions. These findings support the

increase in the number of possible binding configurations associated with monovalent binding as

the main driver of the enhanced binding strength.

These observations are consistent with the small fraction of the nanoparticle surface that comes

in contact with the cell upon binding. Herceptin is an IgG class antibody, an isotype which

occupies an area with a 30 nm diameter [39]. The 80 nm immunonanoparticles tested are highly

curved compared to the cell surfaces and likely give rise to a small cell-particle contact area; the

small contact area compared to the area occupied by each antibody makes it improbable for

multiple antibodies to be localized at the cell-particle interface, even with the antibody mobility

provided by the flexible PEG spacer as an attachment point to the particle core. Hence, the

amplified binding of nanoparticles targeted using large targeting molecules (e.g., whole

antibodies or antibody fragments) [40,41] is likely caused by an increase in the number of

monovalent binding states, and not the multivalent interactions to which it has formerly been

attributed. Nanoparticles that are densely covered with smaller targeting ligands (e.g., hundreds

or thousands of low molecular weight molecules per particle) [9, 19, 38, 42] may still exhibit

multivalent cell-particle interactions, which, unlike the system described here, can lead to a non-

linear eqK versus Ω dependence.

The agreement between the data in Figure 5.3 and monovalent binding model suggests that

immunonanoparticle binding strength can be predictably tuned by adjusting the number of

conjugated Herceptin antibodies according to Equation 9. This agreement between the predicted

fractional coverage (calculated from the fitted )1(GΔ -value) and that obtained experimentally

86

using either free Herceptin or the Herceptin immunonanoparticles is further illustrated in Figure

5.4. The experimental θ values are closely correlated to the theoretical predictions (R2 = 0.99),

supporting Equation 9 as a useful quantitative guideline for designing immunonanoparticles for

targeted drug delivery to tumour sites.

Figure 5.4 Comparison of the experimental and theoretical fractional coverages (θ) of SKBR-3 cells by

free Herceptin (p) and Herceptin immunonanoparticles bearing 1.9 (�), 3.2 (r), 5.9 (�), and 9.4 (n)

antibodies exhibiting monovalent binding. The experimentally derived θ values closely match the

theoretically predicted θ values, with R2 = 0.99.

Looking forward, quantification of the binding isotherm also guides in vivo dosage requirements

by expressing the intratumoural particle concentration required to reach a desired fraction of

saturation binding. Approaching saturating particle levels maximizes receptor binding as a

gateway to receptor mediated cellular uptake, making this an important parameter for many

targeted drug delivery strategies [18].

87

5.6 Conclusions

The binding isotherms of Herceptin immunonanoparticles bound to HER2 overexpressing

SKBR-3 cells were measured at varying levels of antibody conjugation using a flow cytometric

immunoassay, thereby quantifying binding strength using a direct live cell assay. Based on these

measurements, a thermodynamic analysis of the binding valency was completed and the

resulting valency of antibody-receptor binding interactions of immunonanoparticles bearing

multiple targeting antibodies was investigated. Binding affinity increases with increasing

antibody conjugation density in a manner consistent with the theory for monovalent binding,

suggesting that multivalent interactions are not the primary cause of the amplified binding

strength. The Herceptin immunonanoparticle formulations tested can be selected for values of

Keq up to an order of magnitude greater than the value for free Herceptin. This method can also

be applied to other particle formulations having multiple targeting ligands to better understand

how the number of ligands affects the binding valency of a particular system, and how this

property can then be manipulated to control effective binding affinity. The resulting

understanding of the mechanism governing the increase in binding strength can be used in a

predictive manner to guide nanoparticle design.

5.7 Acknowledgements

We thank Simone Helke for her advice and technical support in developing our methods for

analysis of antibody binding using FACS. We are grateful to the Natural Sciences and

Engineering Research Council and the Canadian Institutes for Health Research for funding

through the Collaborative Health Research Program.

5.8 References

[1] Cairns R, Papandreou I, Denko N. Overcoming physiologic barriers to cancer treatment by molecularly targeting the tumor microenvironment. Mol Cancer Res 2006;4:61-70.

[2] Cho KJ, Wang X, Nie SM, Chen Z, Shin DM. Therapeutic nanoparticles for drug delivery in cancer. Clin Cancer Res 2008;14:1310-1316.

[3] Marcucci F, Lefoulon F. Active targeting with particulate drug carriers in tumor therapy: fundamentals and recent progress. Drug Discov Today 2004;9:219-228.

[4] Trail PA, King HD, Dubowchik GM. Monoclonal antibody drug immunoconjugates for targeted treatment of cancer. Cancer Immunol Immun 2003;52:328-337.

88

[5] Park JW, Kirpotin DB, Hong K, Shalaby R, Shao Y, Nielsen UB, et al. Tumor targeting using anti-her2 immunoliposomes. J Control Release 2001;74:95-113.

[6] Shi M, Wosnick JH, Ho K, Keating A, Shoichet MS. Immuno-polymeric nanoparticles by Diels-Alder chemistry. Angew Chem Int Edit 2007;46:6126-6131.

[7] Adams GP, Schier R, McCall AM, Simmons HH, Horak EM, Alpaugh RK, et al. High affinity restricts the localization and tumor penetration of single-chain Fv antibody molecules. Cancer Res 2001;61:4750-4755.

[8] Zhou Y, Daryl C, Zou H, Hayes ME, Adams GP, Kirpotin DB, et al. Impact of single-chain fv antibody fragment affinity on nanoparticle targeting of epidermal growth factor receptor-expressing tumor cells. J Mol Biol 2007;371:934-947.

[9] Bartlett DW, Davis ME. Physicochemical and biological characterization of targeted, nucleic acid-containing nanoparticles. Bioconjugate Chem 2007;18:456-468.

[10] Panyam J, Labhasetwar V. Biodegradable nanoparticles for drug and gene delivery to cells and tissue. Adv Drug Deliver Rev 2003;55:329-347.

[11] Nobs L, Buchegger F, Gurny R, Allemann E. Current methods for attaching targeting ligands to liposomes and nanoparticles. J Pharm Sci-Us 2004;93:1980-1992.

[12] Janeway CA, Travers P, Walport M, Shlomchik MJ. Immunobiology: the immune system in health and disease. 6 ed. New York: Garland Science; 2005.

[13] Kiessling LL, Gestwicki JE, Strong LE. Synthetic multivalent ligands in the exploration of cell-surface interactions. Current Opinion in Chemical Biology 2000;4:696-703.

[14] Adams GP, Schier R, Marshall K, Wolf EJ, McCall AM, Marks JD, et al. Increased affinity leads to improved selective tumor delivery of single-chain Fv antibodies. Cancer Res 1998;58:485-490.

[15] Kubetzko S, Balic E, Waibel R, Zangemeister-Wittke U, Pluckthun A. PEGylation and multimerization of the anti-p185(HER-2) single chain Fv fragment 4D5 - Effects on tumor targeting. J Biol Chem 2006;281:35186-35201.

[16] Beckman RA, Weiner LM, Davis HM. Antibody constructs in cancer therapy - Protein engineering strategies to improve exposure in solid tumors. Cancer 2007;109:170-179.

[17] Juweid M, Neumann R, Paik C, Perezbacete MJ, Sato J, Vanosdol W, et al. Micropharmacology of Monoclonal-Antibodies in Solid Tumors - Direct Experimental-Evidence for a Binding-Site Barrier. Cancer Res 1992;52:5144-5153.

[18] Paulos CM, Reddy JA, Leamon CP, Turk MJ, Low PS. Ligand binding and kinetics of folate receptor recycling in vivo: Impact on receptor-mediated drug delivery. Molecular Pharmacology 2004;66:1406-1414.

89

[19] Hong S, Leroueil PR, Majoros IJ, Orr BG, Baker JR, Holl MMB. The binding avidity of a nanoparticle-based multivalent targeted drug delivery platform. Chem Biol 2007;14:105-113.

[20] Zahnd C, Pecorari F, Straumann N, Wyler E, Pluckthun A. Selection and characterization of Her2 binding-designed ankyrin repeat proteins. J Biol Chem 2006;281:35167-35175.

[21] Soukka T, Harma H, Paukkunen J, Lovgren T. Utilization of kinetically enhanced monovalent binding affinity by immunoassays based on multivalent nanoparticle-antibody bioconjugates. Anal Chem 2001;73:2254-2260.

[22] Zhang HD, Williams PS, Zborowski M, Chalmers JJ. Binding affinities/avidities of antibody-antigen interactions: Quantification and scale-up implications. Biotechnology and Bioengineering 2006;95:812-829.

[23] Wong JY, Kuhl TL, Israelachvili JN, Mullah N, Zalipsky S. Direct measurement of a tethered ligand-receptor interaction potential. Science 1997;275:820-822.

[24] Arteaga CL. Trastuzumab, an appropriate first-line single-agent therapy for HER2-overexpressing metastatic breast cancer. Breast Cancer Res 2003;5:96-100.

[25] Baselga J, Albanell J. Mechanism of action of anti-HER2 monoclonal antibodies. Ann Oncol 2001;12:35-41.

[26] Carson WE, Parihar R, Lindemann MJ, Personeni N, Dierksheide J, Meropol NJ, et al. Interleukin-2 enhances the natural killer cell response to Herceptin-coated Her2/neu-positive breast cancer cells. Eur J Immunol 2001;31:3016-3025.

[27] Slamon DJ, Clark GM, Wong SG, Levin WJ, Ullrich A, Mcguire WL. Human-Breast Cancer - Correlation of Relapse and Survival with Amplification of the Her-2 Neu Oncogene. Science 1987;235:177-182.

[28] Nielsen UB, Kirpotin DB, Pickering EM, Drummond DC, D Marks J. A novel assay for monitoring internalization of nanocarrier coupled antibodies. Bmc Immunol 2006;7:24.

[29] Shi M, Shoichet, M.S. Furan-Functionalized Copolymers for Targeted Drug Delivery: Characterization, Self-Assembly and Drug Encapsulation. Journal of Biomaterials Science: Polymer Edition Accepted 2007.

[30] Nielsen UB, Adams GP, Weiner LM, Marks JD. Targeting of bivalent anti-ErbB2 diabody antibody fragments to tumor cells is independent of the intrinsic antibody affinity. Cancer Res 2000;60:6434-6440.

[31] Stupp Y, Yoshida T, Paul WE. Determination of Antibody-Hapten Equilibrium Constants by an Ammonium Sulfate Precipitation Technique. J Immunol 1969;103:625-627.

[32] Gabizon A, Horowitz AT, Goren D, Tzemach D, Mandelbaum-Shavit F, Qazen MM, et al. Targeting folate receptor with folate linked to extremities of poly(ethylene glycol)-grafted liposomes: In vitro studies. Bioconjugate Chem 1999;10:289-298.

90

[33] Mammen M, Choi SK, Whitesides GM. Polyvalent interactions in biological systems: Implications for design and use of multivalent ligands and inhibitors. Angew Chem Int Edit 1998;37:2755-2794.

[34] Davis HT. Statistical Mechanics of Phases, Interfaces and Thin Films. New York: VCH Publishers, Inc.; 1996.

[35] Mao SY. Conjugation of fluorochromes to antibodies. Methods Mol Biol 1999;115:35-38.

[36] Seddiki T, Ollivierbousquet M. Temperature-Dependence of Prolactin Endocytosis and Casein Exocytosis in Epithelial Mammary Cells. European Journal of Cell Biology 1991;55:60-70.

[37] Lee RJ, Low PS. Folate-Mediated Tumor-Cell Targeting of Liposome-Entrapped Doxorubicin in-Vitro. Biochimica Et Biophysica Acta-Biomembranes 1995;1233:134-144.

[38] Lee RJ, Low PS. Folate-targeted liposomes for drug delivery. Journal of Liposome Research 1997;7:455-466.

[39] Roberts CJ, Williams PM, Davies J, Dawkes AC, Sefton J, Edwards JC, et al. Real-Space Differentiation of Igg and Igm Antibodies Deposited on Microtiter Wells by Scanning Force Microscopy. Langmuir 1995;11:1822-1826.

[40] Muro S, Dziubla T, Qiu WN, Leferovich J, Cui X, Berk E, et al. Endothelial targeting of high-affinity multivalent polymer nanocarriers directed to intercellular adhesion molecule 1. Journal of Pharmacology and Experimental Therapeutics 2006;317:1161-1169.

[41] Willuda J, Kubetzko S, Waibel R, Schubiger PA, Zangemeister-Wittke U, Pluckthun A. Tumor targeting of mono-, di-, and tetravalent Anti-p185(HER-2) miniantibodies multimerized by self-associating peptides. J Biol Chem 2001;276:14385-14392.

[42] Weissleder R, Kelly K, Sun EY, Shtatland T, Josephson L. Cell-specific targeting of nanoparticles by multivalent attachment of small molecules. Nature Biotechnology 2005;23:1418-1423.

91

5.9 Supplementary information

Figure 5.5 In the case of monovalent binding, increasing the number of antibodies per particle, Ω, results

in an increase in the number of possible binding configurations where only one interaction occurs. (A)

The number of possible combinations of monovalent binding events increases with the number of

conjugated antibodies due to an amplified number of possible rotational binding orientations for N bound

nanoparticles (first term of Equation 5). (B) Also, given a lattice of M potential binding sites, the number

of distinct lattice configurations in which the particles can bind is given by a binomial coefficient (second

term of Equation 5).

For example, when Ω=2:

N=1 N=2 N=3 …

Possible combinations of bound

antibodies = ΩN

B

A

92

6 Discussion Cancer is a disease that is remarkably difficult to eradicate fully. Malignant cells are derived

from a patient’s own healthy cells and evolve slowly, developing a phenotype that can evade

immune recognition [1, 2], and invading or metastasizing into healthy tissue [3]. However, they

conserve many similarities to normal cells, and biochemical pathways are often interrelated; it is

difficult to find effective anti-cancer compounds that are specific to pathological biochemistry to

avoid killing healthy cells in the process. Unfortunately, when broad drug activity is coupled

with broad distribution, it leads to unacceptable systemic toxicity, limiting the dose and utility of

drug therapy. Localized administration to the primary tumour site is one approach to limiting

distribution [4], but by definition, distant metastases would be overlooked.

To combine the broad reach of systemic administration with the specific toxicity of localized

delivery, nanoparticle systems target tumour pathophysiology. Here, we examined a polymer

approach to drug targeting. We first selected an orthotopic tumour xenograft model of breast

cancer for its ability to replicate the pathophysiology required to observe EPR. Next, we

formulated docetaxel in poly(TMCC-co-LA)-g-PEG nanoparticles, and demonstrated that our

nanoparticles improved plasma circulation and tumour retention over the conventional

formulation. Finally, we demonstrated that after covalent modification with the Herceptin

antibody, we retained selective binding to live HER2 overexpressing cells, and that we could

predict and tune their binding strength using a mechanistic model. In the following discussion

we will explore the contributions of these findings on the field of anti-cancer drug targeting.

6.1 Validated pre-clinical tumour models

Pre-clinical evaluation of drug formulations relies on animal models to replicate relevant features

of clinical pathology. Nanoparticle accumulation in tumour tissue depends on hyperpermeable

tumour vasculature and reduced lymphatic drainage to lead to the EPR effect. Testing these

formulations in tumours that lack these features would lead to poor tumour accumulation

regardless of the true performance of the delivery system.

Several pharmacokinetic studies continue to be executed in healthy animals, yet there are global

changes to the biochemistry and protein expression profiles of tumour animals that extend

beyond the tumour site itself. Using tumour models enables tumour accumulation to be

93

measured in the same group of animals, saving time and resources. There is a resulting need for

simple but relevant tumour models for preliminary in vivo work.

Tumour xenograft models are attractive because they are simple to replicate and cohorts of mice

having well matched tumours are easily generated. In Chapter 3 we examined the EPR related

pathophysiology of two common preclinical models of breast cancer: orthotopic MFP and

ectopic SC human tumour xenografts in immune deficient mice. SC injection is simpler to

perform, while MFP and other orthotopic injections require surgical procedures and greater

technical skill. However, growing evidence suggests that the organ environment critically

influences cell engraftment and behaviour. Orthotopic models are increasingly being used to

study metastasis and treatment strategies for resulting secondary tumours because their strong

metastatic potential has been proven. If EPR has a similar dependence on the orthotopic cell

environment, then pharmacokinetic and biodistribution studies should also utilize orthotopic

tumour models.

To validate these models in terms of EPR permissive pathophysiology, we used a high molecular

weight dextran (2 MDa) as a model nanocarrier. Studies of tumour vascular permeability often

use smaller model particles, such as labeled albumin (7 nm). However, the typical size range of

drug-loaded nanoparticles is 50-150 nm, which is large enough to evade rapid renal clearance

and small enough to cross gaps in tumour blood vessel walls. Consequently, we chose a larger

model material (80 nm) and demonstrated that EPR could still be observed. This assessment also

demonstrated that MFP tumours were more permissive to nanocarrier accumulation than size

matched SC tumours.

To form a more compelling case, we also investigated the underlying tumour pathophysiology

leading to enhanced tissue accumulation using immunostaining. Both the MFP and SC tumours

showed evidence of excess interstitial fluid, suggesting poor lymphatic drainage in both models.

We also observed detached pericytes in both models, suggesting vessel immaturity and lack of

control over vessel permeability. However, the MFP tumours showed measurable changes in

vascular structure that led to greater permeability: greater endothelium thickness, vascular

density, and thinner basement membranes were all observed in comparison to the SC model.

Several practical considerations also favoured MFP tumours for pre-clinical studies. The

orthotopic environment yielded faster tumour development time, and also avoided the skin

94

ulceration that was observed in SC tumour animals at later development times. These results

provide a rationale for selecting an orthotopic tumour model when studying nanoparticle

distribution.

6.2 Quantitative pharmacokinetics and biodistribution

Many pharmacokinetic and biodistribution studies use radiolabels or fluorescent tags to track

compounds after administration. While these are useful approaches that can provide a more

complete mass balance, these reporters can be cleaved and may not accurately reflect the final

location of the original intact molecule. Furthermore, large and hydrophobic fluorescent

molecules may alter the system behaviour. Instead of applying a tag directly to poly(TMCC-co-

LA)-g-PEG, we encapsulated docetaxel into the core of our nanoparticles, allowing us to track

the location of the intact drug compound tag-free using UPLC-MS.

These techniques introduced several tools for investigating the effectiveness of poly(TMCC-co-

LA)-g-PEG nanoparticles. This was the first time we successfully encapsulated an adequate

drug load in the core of our nanoparticles for a delivery study. We were able to match dosages

relevant to metronomic dosing (1.5 mg/kg), a low dose strategy that favours more frequent doses

over high doses with long latency times between treatments. In Chapter 2.5.2 we investigated

the behaviour of DOX-conjugated nanoparticles in vitro, where the drug compound was

covalently attached to the surface of our nanoparticles instead of being loaded in the core. This

produced several interesting and unexpected results in terms of toxicity and cellular trafficking.

However, the altered surface chemistry would likely negatively impact nanoparticle circulation,

as the high density of DOX molecules would counteract the PEG chains designed to extend

circulation time. This methodology also removed the obstacles associated with identifying

reactive groups on the drug compound to add maleimide functional groups for Diels-Alder

bonding to the nanoparticles, as well as regulatory concerns that arise from chemically altering

an approved drug molecule.

Furthermore, by monitoring the intact drug molecule directly using UPLC-MS, we could

distinguish docetaxel from metabolites. This is significant because drug fragments likely suffer

lost activity. Additionally, known metabolites that may cause non-specific toxicity can

simultaneously be investigated in the UPLC-MS spectra based on a distinct product peak.

Moreover, this technique is quantitative and sensitive down to nM levels, and only a small

95

sample volume (~10 uL) is required. This allowed us to monitor the plasma concentration at

many time points without sacrificing animals each time. By staggering the blood collection

schedule over several cohorts of animals we were able to collect a comprehensive set of samples

for pharmacokinetic analysis. These protocols can be broadly applied to a variety of drug

candidates given similar hydrophobicity (encapsulation) and size (UPLC-MS detection).

6.3 Long circulating polymer nanoparticles

Using the MFP tumour model and the UPLC-MS detection method, the pharmacokinetic

properties of docetaxel encapsulated in poly(TMCC-co-LA)-g-PEG nanoparticles were

quantified. As a benchmark, we compared this performance to docetaxel in the conventional

ethanolic polysorbate 80 formulation. This work validated poly(TMCC-co-LA)-g-PEG

nanoparticles as a long circulating drug delivery system, a property that we anticipated based on

high PEG density, low critical micelle concentration, and strong nanoparticle kinetic stability.

Nanoparticle encapsulation was also expected to inhibit docetaxel degradation by preventing

direct enzyme contact. Higher docetaxel concentration in plasma at longer times suggests that

this was achieved, because only the active compound was measured.

The pharmacokinetic parameters calculated showed a 1.6-fold increase in lambda half life (t1/2), a

3-fold increase in area under the curve (AUC∞), and an order of magnitude increase in the area

under the first moment curve (AUMC∞) for DTX formulated in nanoparticles. These parameters

all indicate greater drug exposure given an equal initial dose. Extended circulation encourages

greater tumour accumulation by allowing more passes through hyperpermeable tumour

vasculature.

The improved circulation properties of docetaxel reformulated in our nanoparticles implies

utility for delivery of other drug molecules as well. By formulating docetaxel in the nanoparticle

core, the surface properties of poly(TMCC-co-LA)-g-PEG nanoparticles was unaltered. The

size, geometry, and surface properties of the nanoparticles are the primary determinants of their

elimination rate. Other compounds having similar encapsulation stability should demonstrate

similar pharmacokinetic profiles.

96

6.4 Successful passive tumour targeting

In the same animals, biodistribution was investigated by collecting a panel of organs and

measuring their docetaxel content. Enhanced retention of docetaxel after administration in

unmodified poly(TMCC-co-LA)-g-PEG nanoparticles was exclusively observed in tumour

tissue, with a 5-fold decrease in the first-order elimination rate constant versus the free docetaxel

formulation. This has encouraging implications on the safety of docextaxel administered in our

nanoparticle formulation: in healthy tissues, increased accumulation was not observed over the

clinically prescribed formulation, except in transient cases. Potential for increased efficacy is

also implied because the tumour is exposed to a high docetaxel concentration for an extended

period of time. As a result, greater levels of toxicity are expected.

Of the healthy organ tissues observed, only the kidney saw greater accumulation of docetaxel

when formulated in nanoparticles, suggesting a partial shift to renal clearance. However, kidney

retention was not observed: the first order elimination rate constant was 1.3-fold higher for the

docetaxel formulated in nanoparticles, leading to an overall convergence of the two profiles.

Lack of accumulation in the RES organs (liver and spleen) suggest that the PEG coverage on

poly(TMCC-co-LA)-g-PEG nanoparticles successfully modulated the RES response.

However, introducing targeting ligands to the nanoparticle surface may impact circulation time

and biodistribution. Previous studies have shown that targeting ligands can decrease circulation

time, and these effects are likely more pronounced in immune competent animals [5, 6]. We

studied Herceptin-modified nanoparticles in immune compromised mice (Chapter 8), and

although we did not observe decreased circulation time, we did observe decreased tissue uptake

across our organ panel. Although the antibody modification did not result in increased particle

size, it likely reduced the ability of our nanoparticles to deform by steric hindrance, thereby

decreasing tissue accumulation [7, 8]. Using a smaller alternative targeting ligand, such as a Fab

antibody fragment, may better preserve the original distribution properties of unmodified

poly(TMCC-co-LA)-g-PEG nanoparticles.

6.5 Quantitative live cell binding

In addition to tumour targeting, we were interested in evaluating poly(TMCC-co-LA)-g-PEG

nanoparticles after antibody modification to validate their selective binding to live cells. Using

97

flow cytometry, we were able to confirm live cell binding of Herceptin-nanoparticles to HER2

overexpressing cells. In addition, we were able to abrogate binding by blocking the HER2

receptor with pre-incubation with free Herceptin, or by replacing Herceptin with a non-specific

IgG1κ isotype control, confirming an antibody-antigen specific interaction. Also, in the absence

of HER2 expression, no binding was observed.

Notably, this means that the original binding activity of Herceptin was conserved. Poly(TMCC-

co-LA)-g-PEG was designed with this goal in mind. By designing the polymer composition to

include furan functional groups on the free ends of the PEG chains, site-specific modification of

the polymer with a maleimide-modified antibody was achieved using Diels-Alder chemistry.

This verified two important objectives that were set forth when designing the polymer: using the

mild reaction conditions for Diels-Alder chemistry preserved antibody binding activity during

the conjugation process; and antibody placement on the PEG chain termini positioned them

appropriately to bind their target antigen.

Taking these results and this technique one step further, we devised a method to quantify binding

strength. Investigating binding as a dose response, we were able to empirically fit the resulting

binding isotherms and calculate the equilibrium binding strength. This strategy allowed us to use

a semi-quantitative fluorescent method to extract quantitative information. Remarkably, we

were also able to apply this approach to live cells instead of antigens immobilized onto hard

synthetic substrates, taking full advantage of the natural membrane fluidity and receptor spacing

to produce the most accurate in vitro response.

6.6 Mechanistic predictions of binding behaviour

By developing a theoretical model of monovalent (one antibody-antigen interaction per particle)

and multivalent (multiple antibody-antigen interactions per particle) binding, we were able to

evaluate the empirical behaviour of Herceptin-nanoparticles. In our system, the equilibrium

binding constant (Keq) was directly proportional to the average number of antibodies per particle.

This is consistent with monovalent binding, where Keq increases due to increased available

configurations for binding. Multiple connections increase Keq exponentially through avidity.

Monovalent interaction suggests that although PEG has great flexibility, the overall nanoparticle

structure is rigid enough that antibody-modified chains do not migrate to the cell surface.

98

A mechanistic understanding of binding behaviour enables predictive nanoparticle design.

While strong binding demonstrates positive outcomes in vitro due to enhanced uptake, it may be

detrimental in vivo where tumour penetration becomes an important factor. High targeting

ligand density may also diminish the long-circulating properties of PEGylated nanocarriers:

while PEG extends circulation by providing a neutral charge and reducing opsonization,

targeting ligands must be presented beyond the PEG border to be available for binding, and this

alters the surface properties of the nanocarrier. It has previously been shown that molecular

targeting via natural ligands to a receptor [6] or antibodies [5] results in immune recognition and

decreased circulation times in immune competent animals. Being able to predict binding

strength gives the flexibility to tune particle binding as new information becomes available using

a parameter that is easily controlled on the bench via reaction time or feed ratios.

6.7 Conclusions

This work established for the first time the utility of poly(TMCC-co-LA)-g-PEG nanoparticles in

biological applications. We first established orthotopic MFP tumour xenografts as a valid pre-

clinical model of breast cancer for testing nanoparticle targeting based on the capacity to

demonstrate EPR. To better understand the vascular and lymphovascular properties the capture

this disease condition, we used immunostaining to confirm poor lymphatic drainage and blood

vessel immaturity, in addition to greater endothelium thickness, vascular density, and thinner

basement membranes than size-matched SC tumours. Using this model, we tested the

pharmacokinetics and biodistribution of docetaxel encapsulated in poly(TMCC-co-LA)-g-PEG

nanoparticles, demonstrating improved blood circulation over the conventional surfactant-based

formulation. We also examined a panel of tissues and found that in the tumour alone, there was

greater retention of docetaxel when reformulated in nanoparticles. Based on this observation,

enhanced efficacy is expected because extended exposure to high drug concentrations should

lead to greater toxicity. Next, we developed a live cell binding assay and verified that Herceptin-

nanoparticles selectively bind HER2 overexpressing cells based on a specific antibody-antigen

interaction. Based on binding isotherms, we quantified the binding strength of a series of

Herceptin-nanoparticles having varying antibody conjugation density. We found that Keq is

directly proportional to the average number of antibodies per particle, demonstrating empirically

that Herceptin-nanoparticles follow a theoretical model of monovalent binding. We also

developed a corresponding model for multivalent binding. Once the behaviour of a particular

99

system has been identified, these models can be applied widely to guide nanoparticle design.

These properties validate poly(TMCC-co-LA)-g-PEG as an alternative to surfactant-based

formulations with improved physical targeting to cancer on tissue and cellular levels.

6.8 Achievement of objectives

This research was motivated by the following hypothesis:

Poly(TMCC-co-LA)-g-PEG nanoparticle micelles will provide specific anti-cancer drug delivery

at both tissue and cellular levels

Herein, we described in vitro and in vivo tests that confirmed that poly(TMCC-co-LA)-g-PEG

nanoparticles provided tissue and cell specific targeting. Achievement of the objectives

originally laid out in Chapter 1 is summarized below.

(1) To confirm that the EPR effect can be observed in mice using in a tumour xenograft

model.

To observe nanoparticle targeting in a solid tumour model, we challenged MFP and SC

tumour xenografts using a model nanocarrier and monitored its accumulation in tumour

tissue sections. Based on these results, the MFP tumours demonstrated higher tumour

accumulation than size-matched SC tumours. This result was supported by the vascular

and lymphovascular properties revealed using immunostaining. Together these

observations validated MFP tumours as appropriate models for observing nanoparticle

targeting via EPR, and this model was carried forward to test objective 2.

(2) To demonstrate that poly(TMCC-co-LA)-g-PEG nanoparticles improve pharmacokinetics

and biodistribution over a surfactant-based drug formulation.

After loading docetaxel into poly(TMCC-co-LA)-g-PEG nanoparticles, we were able to

quantitatively monitor its distribution in MFP tumour-bearing mice using UPLC-MS.

Pharmacokinetic parameters established that the nanoparticle formulation improved blood

circulation over the conventional ethanolic polysorbate 80 formulation, improving the

odds of passive targeting by allowing multiple passes through the hyperpermeable tumour

vasculature. We also demonstrated enhanced retention of docetaxel reformulated in

100

nanoparticles only in tumour tissue. These results suggest potential for improved anti-

tumour efficacy based on extended exposure of cancer cells to a high drug concentration.

(3) To verify that antibody-modified poly(TMCC-co-LA)-g-PEG nanoparticles bind

selectively to cells overexpressing a target surface antigen.

Using live cells, we demonstrated selective binding of Herceptin-nanoparticles to HER2

overexpressing cells, and that this binding was antibody-antigen specific. Building on

this observation, we used binding isotherms to quantify the strength of these interactions.

Based on this assessment, we concluded that empirical binding behaviour of Herceptin-

nanoparticles was consistent with a theoretical model of monovalent binding (one

antibody-antigen interaction per particle), demonstrating that binding strength can be

predicted and controlled.

6.9 References

[1] Strand S, Hofmann WJ, Hug H, Muller M, Otto G, Strand D, et al. Lymphocyte apoptosis induced by CD95 (APO-1/Fas) ligand-expressing tumor cells--a mechanism of immune evasion? Nat Med 1996;2:1361-1366.

[2] Huang B, Zhao J, Li HX, He KL, Chen YB, Mayer L, et al. Toll-like receptors on tumor cells facilitate evasion of immune surveillance. Cancer Res 2005;65:5009-5014.

[3] Seton-Rogers SE, Lu Y, Hines LM, Koundinya M, LaBaer J, Muthuswamy SK, et al. Cooperation of the ErbB2 receptor and transforming growth factor beta in induction of migration and invasion in mammary epithelial cells. P Natl Acad Sci USA 2004;101:1257-1262.

[4] Arias JL. Drug Targeting Strategies in Cancer Treatment: An Overview. Mini-Rev Med Chem 2011;11:1-17.

[5] Ferrari M. Nanogeometry: beyond drug delivery. Nat Nanotechnol 2008;3:131-132.

[6] McNeeley KM, Annapragada A, Bellamkonda RV. Decreased circulation time offsets increased efficacy of PEGylated nanocarriers targeting folate receptors of glioma. Nanotechnology 2007;18.

[7] Hwang HY, Kim IS, Kwon IC, Kim YH. Tumor targetability and antitumor effect of docetaxel-loaded hydrophobically modified glycol chitosan nanoparticles. J Control Release 2008;128:23-31.

[8] Moghimi SM, Hunter AC, Murray JC. Long-circulating and target-specific nanoparticles: Theory to practice. Pharmacol Rev 2001;53:283-318.

101

7 Limitations and recommendations for future work

7.1 Mixed cell populations in tumor xenograft models

In Chapter 3, we investigated human tumour xenografts in immune deficient mice and were able

to observe EPR and EPR permissive pathophysiology, especially in orthotopic MFP tumours.

However, there was a general lack of vascular capacity, which can be partially attributed to the

non-invasive nature of the cell line used [1]; a more metastatic variant may exert greater

basement membrane degradation which would enable more rapid vascular branching, leading to

improved blood perfusion [1, 2]. However, in our experience with a metastatic variant of MDA-

MB-231-H2N (LM2-4 [3]), we observed accelerated tumour development rates that made size

matching difficult and provided only short windows in which to perform studies because

unacceptable tumour burdens were quickly reached. This also raised debate over whether

adequate time for neovascularization and blood vessel remodeling had elapsed.

A possible solution is to replace a single clonal cell line with mixed populations of cells to

provide a blend of properties in the resulting tumour. A mixed model may also provide a more

clinically relevant response, as tumours are comprised of many cell types, and even within the

cancer cell population there are many subpopulations having varying phenotypes [4]. A mixed

cell population approach would also approach animal models utilizing primary tumour biopsy

cells [5] while remaining widely available to researchers without access to primary samples.

Moreover, many current studies will investigate cellular responses to nanocarriers with targeting

ligands by introducing two distinct tumours on either side of a single animal, one that is antigen

overexpressing and the other that is antigen negative [6]. However, this is less relevant than

targeting the antigen-overexpressing cells as a subpopulation within a tumour having mixed

cells, which is likely to be the case in a clinical setting. Selecting a target cell line with a

reporter, like GFP, would allow the separate populations to be distinguished so that the resulting

toxicities could be compared by microscopy or flow cytometry.

A mixed cell approach may also provide insight into targeting as it pertains to cancer stem cells.

There is increasing evidence that only a small fraction of cells in a tumour give rise to new cells

[7, 8]. In this case, if the cancer stem cells are left untreated, then remission will only be

102

transient. On the other hand, if only the cancer stem cells are targeted, then the initial response

may be negligible, as they represent such a small proportion of the overall tumour mass.

However, eventually the tumour will regress fully, having no source for new cancer cells to

propagate [9]. By creating a mixed cell population where the target cells represent only a small

fraction of the tumour as in the case of cancer stem cells, the delivery strategy would need to

overcome similar transport barriers as it would clinically.

7.2 Cellular uptake as a function of antibody density

In Chapter 5, we discussed how to control nanoparticle binding by adjusting the antibody

conjugation density. To demonstrate this, we looked at nanoparticle-cell interactions at

equilibrium while inhibiting endocytosis. With this information in hand, we can design a study

that instead looks at the kinetics of cellular uptake as a function of binding strength.

Adapting our previous secondary antibody detection methodology, it is possible to monitor

nanoparticles over time in live cells. To avoid kinetic effects of particle binding, cells can be

incubated with nanoparticles first at 4 °C to reach equilibrium binding, and then transferred to

37 °C to initiate internalization. Moreover, to ensure that receptor recycling does not contribute

to these measurements, geldanamycin can be used to inhibit HER2 recycling [10]. To capture

snapshots of particle localization, it is probable that cells will need to be fixed, but simply

transferring them to ice may stall further cellular trafficking events. There are two possible

methods to enable distinction between membrane-bound and internalized nanoparticles by

secondary antibody detection: (1) the signal from cells that have been permeabilized and that

have not been permeabilized represent the total signal and membrane-bound signal respectively

[11], and the internalized signal can be calculated by subtraction; (2) membrane-bound

nanoparticles can be removed via acid stripping, and the internalized and membrane-bound

signals measured separately [12].

It has previously been shown that endocytic rate constant (ke) is equal to the slope of the ratio of

internalized:surface-bound targeting ligand as a function of time [13]. Flow cytometry will

enable measurement of a large cell sample size, and has been successfully applied to simple

receptor-ligand internalization studies [12]. These measurements will allow us to calculate ke as

a function of binding strength at a given nanoparticle concentration, or as a function of

103

nanoparticle concentration at a set antibody conjugation density, which will provide estimates for

the antibody modification or dosage requirements to achieve a target uptake rate.

7.3 Tumour penetration as a function of antibody density

Heterogeneous drug distribution in a tumour mass leads to poor treatment uniformity and drug

resistance [14]. Poor vascular architecture already contributes to regions of hypoxia, which

correspond to the regions that are most distant from functioning blood flow [15]. Active

targeting strategies can further intensify the difficulty in distributing drugs to all areas of a

tumour, in part due to the slow diffusion of large nanocarriers, but also due to active binding and

depletion by the cells closest to an active blood vessel [16-18]. Even the free Herceptin antibody

can take up to 24 hours to distribute uniformly in a HER2 overexpressing tumour xenografts

[19]. Consequently, understanding how binding strength influences tumour penetration is an

important consideration in nanoparticle design.

While tumour penetration distance can be computationally difficult to measure in animal models,

a simplified approach would be to investigate concentration profiles in spherical cell aggregates

in vitro [15]. This also removes changing concentrations in blood circulation and transport

barriers across tumour blood vessels. These simplifications would enable direct measurement of

penetration as a function of binding strength by controlling antibody density.

To measure the nanoparticle concentration profile in cell aggregates, a flow cytometry approach

is still relevant. Submerging the aggregates in a non-specific stain, such as Hoechst 33342,

establishes a gradient of staining intensity that can be used to categorize cells according to depth

within the aggregate [20]. After dissociating the aggregates and using the secondary antibody

approach to co-stain for the nanoparticles, nanoparticle association with cells at different depths

can be determined by correlation with the Hoechst signal using flow cytometry [21].

This study would provide guidelines for the maximum antibody conjugation density that would

be permissive to uniform tumour distribution. Measurements of tumour vascular density

(Chapter 3) and in vivo tumour accumulation data (Chapter 4) also inform these estimates by

providing the approximate distance between blood vessels and the total nanoparticle load that

initially enters the tumour and is available for eventual distribution.

104

7.4 Evaluating alternative targeting ligands

Herceptin-nanoparticles demonstrated utility in vitro via receptor-specific binding and uptake,

but failed to accumulate in tumour tissue in vivo (Appendices). We were initially interested in

using a full antibody for three reasons: (1) by virtue of having two binding sites, antibodies can

enhance binding through avidity, and therefore generally have greater affinity for their targets

[22]; and (2) the full IgG structure of Herceptin and other therapeutically relevant antibodies is

required to induce antibody dependent cellular cytotoxicity (ADCC), an important mechanism

for their therapeutic activity [23, 24]; (3) site-specific modification of the saccharide chains on

the constant (non-binding) region of IgG provided a maleimide modification site that did not

interfere with binding activity [25]. However, using a fragment antigen binding (Fab fragment)

or single chain variable fragment (scFv) will retain the specificity of the parent IgG, and may

still allow inhibition of DNA repair, and therefore synergistic effects of combining HER2

targeting with chemotherapeutic delivery [26]. Moreover, if the amino acids in the binding

region are known, chemistry schemes can be tailored to avoid altering binding activity.

Replacing an antibody with an antibody fragment as the targeting ligand has the potential to

maintain receptor-specific targeting and uptake while reducing many of the negative changes we

observed in vivo. A full human IgG has a ~150 kDa molecular weight, whereas Fab fragments

are ~55 kDa, and scFv are ~30 kDa. These considerable reductions in molecular size may have a

significant impact on the ability of the modified nanoparticles to extravasate through

hyperpermeable tumour vasculature. These properties may also enhance tumour penetration by

accommodating greater nanoparticle flexibility [27]. A reduced influence on the nanoparticle

surface properties would also be expected. For these reasons, substituting for an alternative

targeting ligand may increase tumour uptake to levels similar to unmodified nanoparticles, while

still encouraging receptor-specific uptake and toxicity.

7.5 Safety and efficacy

In Chapter 4, we observed that poly(TMCC-co-LA)-g-PEG nanoparticles improved circulation

properties and tumour retention of docetaxel over the conventional surfactant-based formulation.

Simply by providing a suitable alternative to ethanolic polysorbate 80 while maintaining

adequate docetaxel concentrations for dosing, there is significant potential for reduced systemic

toxicity. The enhanced tumour retention has further potential influence on treatment efficacy, as

105

greater tumour toxicity is expected based on extended exposure to high docetaxel concentration.

Combining docetaxel encapsulation and a targeting ligand for selective uptake presents even

more exciting opportunities for selective toxicity, especially if greater tumour accumulation can

be achieved using an alternative ligand.

A metronomic dosing approach would be readily achieved using our existing protocols. While

1.5 mg/kg represents only a fraction of the 15-20 mg/kg mean tolerated dose [28], frequent (3

times per week) low doses of docetaxel have previously shown excellent tolerance and tumour

regression in models of ovarian cancer [29]. However, in our experience, higher doses (~6

mg/kg) could be achieved by particle concentration using a tangential flow filter.

A suggested dosing schedule is to administer dose equivalents of docetaxel at 1.5 mg/kg by bolus

IV injection three times a week for two weeks. To take advantage of EPR, treatment should

begin a minimum of 2-3 days after MFP tumours become palpable [30]. Animals should be

monitored daily for mortality, symptoms of humane endpoints, and weight loss. Plotting tumour

size as a function of time would be a useful measure of disease progression. Survival data

should also be recorded, although deaths are not anticipated under this treatment regimen. After

sacrifice, blood and tissue samples should be collected for analysis. A common toxicological

measure is toxicity in vital organs such as the liver, kidney, and spleen, which can be assessed

visually using hematoxylin and eosin staining [31].

This study would confirm whether drug targeting using poly(TMCC-co-LA)-g-PEG

nanoparticles translates into enhanced tumour regression and suppressed systemic toxicity, the

two primary end goals of tumour targeting. Together with the data presented here, these results

would provide a complete pre-clinical picture of our nanoparticle system.

7.6 References

[1] Abdelkarim M, Vintonenko N, Starzec A, Robles A, Aubert J, Martin M-L, et al. Invading Basement Membrane Matrix Is Sufficient for MDA-MB-231 Breast Cancer Cells to Develop a Stable <italic>In Vivo</italic> Metastatic Phenotype. PLoS ONE 2011;6:e23334.

[2] Lu P, Weaver VM, Werb Z. The extracellular matrix: a dynamic niche in cancer progression. J Cell Biol 2012;196:395-406.

106

[3] Munoz R, Man S, Shaked Y, Lee CR, Wong J, Francia G, et al. Highly efficacious nontoxic preclinical treatment for advanced metastatic breast cancer using combination oral UFT-cyclophosphamide metronomic chemotherapy. Cancer Res 2006;66:3386-3391.

[4] Evan GI, Vousden KH. Proliferation, cell cycle and apoptosis in cancer. Nature 2001;411:342-348.

[5] Marsden CG, Wright MJ, Carrier L, Moroz K, Pochampally R, Rowan BG. A novel in vivo model for the study of human breast cancer metastasis using primary breast tumor-initiating cells from patient biopsies. Bmc Cancer 2012;12.

[6] Gong HB, Kovar J, Little G, Chen HX, Olive DM. In Vivo Imaging of Xenograft Tumors Using an Epidermal Growth Factor Receptor-Specific Affibody Molecule Labeled with a Near-infrared Fluorophore. Neoplasia 2010;12:139-U159.

[7] Visvader JE, Lindeman GJ. Cancer stem cells in solid tumours: accumulating evidence and unresolved questions. Nat Rev Cancer 2008;8:755-768.

[8] Dick JE. Breast cancer stem cells revealed. P Natl Acad Sci USA 2003;100:3547-3549.

[9] Wicha MS, Liu SL, Dontu G. Cancer stem cells: An old idea - A paradigm shift. Cancer Research 2006;66:1883-1890.

[10] Austin CD, De Maziere AM, Pisacane PI, van Dijk SM, Eigenbrot C, Sliwkowski MX, et al. Endocytosis and sorting of ErbB2 and the site of action of cancer therapeutics trastuzumab and geldanamycin. Mol Biol Cell 2004;15:5268-5282.

[11] Hallden G, Andersson U, Hed J, Johansson SGO. A New Membrane Permeabilization Method for the Detection of Intracellular Antigens by Flow-Cytometry. J Immunol Methods 1989;124:103-109.

[12] Schmidt-Glenewinkel H, Reinz E, Eils R, Brady NR. Systems Biological Analysis of Epidermal Growth Factor Receptor Internalization Dynamics for Altered Receptor Levels. J Biol Chem 2009;284:17243-17252.

[13] Wiley HS, Cunningham DD. The Endocytotic Rate-Constant - a Cellular-Parameter for Quantitating Receptor-Mediated Endocytosis. J Biol Chem 1982;257:4222-4229.

[14] Tredan O, Galmarini CM, Patel K, Tannock IF. Drug resistance and the solid tumor microenvironment. J Natl Cancer I 2007;99:1441-1454.

[15] Minchinton AI, Tannock IF. Drug penetration in solid tumours. Nat Rev Cancer 2006;6:583-592.

[16] Baker JHE, Lindquist KE, Huxham L, Kyle AH, Sy JT, Minchinton AI. Direct visualization of heterogeneous extravascular distribution of trastuzumab in human epidermal growth factor receptor type 2 overexpressing xenografts. Clin Cancer Res 2008;14:2171-2179.

107

[17] Dreher MR, Liu WG, Michelich CR, Dewhirst MW, Yuan F, Chilkoti A. Tumor vascular permeability, accumulation, and penetration of macromolecular drug carriers. J Natl Cancer I 2006;98:335-344.

[18] Juweid M, Neumann R, Paik C, Perezbacete MJ, Sato J, Vanosdol W, et al. Micropharmacology of Monoclonal-Antibodies in Solid Tumors - Direct Experimental-Evidence for a Binding-Site Barrier. Cancer Res 1992;52:5144-5153.

[19] Lee CM, Tannock IF. The distribution of the therapeutic monoclonal antibodies cetuximab and trastuzumab within solid tumors. Bmc Cancer 2010;10.

[20] Durand RE. Use of Hoechst-33342 for Cell Selection from Multicell Systems. J Histochem Cytochem 1982;30:117-122.

[21] Olive PL, Chaplin DJ, Durand RE. Pharmacokinetics, Binding and Distribution of Hoechst 33342 in Spheroids and Murine Tumors. Brit J Cancer 1985;52:739-746.

[22] Hudson PJ. Recombinant antibody constructs in cancer therapy. Curr Opin Immunol 1999;11:548-557.

[23] Iannello A, Ahmad A. Role of antibody-dependent cell-mediated cytotoxicity in the efficacy of therapeutic anti-cancer monoclonal antibodies. Cancer Metast Rev 2005;24:487-499.

[24] Arnould L, Gelly M, Penault-Llorca F, Benoit L, Bonnetain F, Migeon C, et al. Trastuzumab-based treatment of HER2-positive breast cancer: an antibody-dependent cellular cytotoxicity mechanism? Br J Cancer 2006;94:259-267.

[25] Shi M, Wosnick JH, Ho K, Keating A, Shoichet MS. Immuno-polymeric nanoparticles by Diels-Alder chemistry. Angew Chem Int Edit 2007;46:6126-6131.

[26] Nahta R, Esteva FJ. Herceptin: mechanisms of action and resistance. Cancer Lett 2006;232:123-138.

[27] Hwang HY, Kim IS, Kwon IC, Kim YH. Tumor targetability and antitumor effect of docetaxel-loaded hydrophobically modified glycol chitosan nanoparticles. J Control Release 2008;128:23-31.

[28] Dykes DJ, Bissery MC, Harrison SD, Waud WR. Response of Human Tumor Xenografts in Athymic Nude-Mice to Docetaxel (Rp-56976, Taxotere(R)). Invest New Drug 1995;13:1-11.

[29] Kamat AA, Kim TJ, Landen CN, Lu CH, Han LY, Lin YG, et al. Metronomic chemotherapy enhances the efficacy of antivascular therapy in ovarian cancer. Cancer Res 2007;67:281-288.

[30] Schiffelers RM, Metselaar JM, Fens MHAM, Janssen APCA, Molema G, Storm G. Liposome-encapsulated prednisolone phosphate inhibits growth of established tumors in mice. Neoplasia 2005;7:118-127.

108

[31] Yang K, Wan JM, Zhang SA, Zhang YJ, Lee ST, Liu ZA. In Vivo Pharmacokinetics, Long-Term Biodistribution, and Toxicology of PEGylated Graphene in Mice. Acs Nano 2011;5:516-522.

109

Copyright Acknowledgements

Portions of Chapter 2 are reprinted with permission from Elsevier from the following publication:

Ho KS, Shoichet MS (2013). Design Considerations of Polymeric Nanoparticle Micelles for Targeted Chemotherapeutic Delivery. Current Opinion in Chemical Engineering, 2(1): 53-59.

Available online at: http://dx.doi.org/10.1016/j.coche.2013.01.003

Chapter 3 is reprinted with permission from Biomed Central from the following publication:

Ho KS, Poon PC, Owen SC, and Shoichet MS (2013) Blood vessel hyperpermeability and pathophysiology in human tumour xenograft models of breast cancer: a comparison of ectopic and orthotopic tumours. BMC Cancer, 12: 579.

Available online at: http://www.biomedcentral.com/1471-2407/12/579

Figures 2.5, 2.6, and 2.7 are reprinted with permission from John Wiley and Sons from the following publication:

Shi M, Wosnick JH, Ho K, Keating A, and Shoichet MS (2007). Immuno-polymeric nanoparticles by Diels-Alder chemistry. Angewandte Chemie International Edition, 46(32): 6126-6131.

Available online at: http://onlinelibrary.wiley.com/doi/10.1002/anie.200701032/abstract

Figures 2.8, 2.9, and 2.10, and Table 2.1 are reprinted with permission from John Wiley and Sons from the following publication:

Shi M, Ho K, Keating A, and Shoichet MS (2009). Doxorubicin-conjugated immuno-nanoparticles for intracellular anticancer drug delivery. Advanced Functional Materials, 19(11): 1689-1696.

Available online at: http://onlinelibrary.wiley.com/doi/10.1002/adfm.200801271/abstract

110

Chapter 4 is reprinted with permission from Elsevier from the following publication:

Ho KS, Aman AM, Al-awar RS, and Shoichet MS (2012) Amphiphilic micelles of poly(2-methyl-2-carboxytrimethyle carbonate-co-D,L-lactide)-graft-poly(ethylene glycol) deliver anti-cancer drugs to solid tumours. Biomaterials, 33 (7), 2223–2229.

Available online at: http://www.sciencedirect.com/science/article/pii/S0142961211014268

Chapter 5 is reprinted with permission from the Royal Society of Chemistry from the following publication:

Ho K, Lapitsky Y, Shi M, and Shoichet MS (2009). Tunable immunonanoparticle binding to cancer cells: thermodynamic analysis of targeted drug delivery vehicles. Soft Matter, 5(5): 1074-80.

Available online at: http://pubs.rsc.org/en/Content/ArticleLanding/2009/SM/b814204a

111

8 Appendices

8.1 List of abbreviations α-SMA Alpha smooth muscle actin

ADCC Antibody-dependent cellular cytotoxicity

aHER2-DTX-NP Herceptin-modified nanoparticles containing encapsulated docetaxel

CMC Critical micelle concentration

DAB 3,3’-diaminobenzidine

DMF Dimethylformamide

DOX Doxorubicin

DTX Docetaxel

DTX-NP Nanoparticles containing encapsulated docetaxel

EPR Enhanced permeability and retention

FBS Fetal bovine serum

FITC Fluorescein isothiocyanate

HER2 Human epidermal growth factor receptor 2

HMEC-1 Healthy endothelial cell line

IgG Immunoglobulin G

IV Intravenous

LYVE-1 Lymphatic vessel endothelial hyaluronan receptor

MCF-7 HER2 normal breast cancer cell line

MDA-MB-231-H2N Tumourigenic and HER2 overexpressing cell line

MDA-MB-468 HER2 negative breast cancer cell line

MES Morpholinoethanesulfonic acid

MFP Mammary fat pad

MMP Matrix metalloproteinase

112

MTS assay Colourimetric metabolic assay using tetrazolium salt

NGS Normal goat serum

NP-aHER2 Herceptin-modified nanoparticles

NP-aHER2-DOX Herceptin- and doxorubicin-modified nanoparticles

NP-DOX Doxorubicin-modified nanoparticles

NSG mice NOD scid gamma mice

PBS Phosphate buffered saline, pH 7.4

PEG Poly(ethylene glycol)

PI Propidium iodide

PK Pharmacokinetics

PLA Poly(lactic acid)

Poly(TMCC-co-LA)-g-PEG Poly(2-methyl-2-carboxytrimethylene carbonate-co-D,L-lactide)-graft-poly(ethylene glycol)-furan

PS80 Polysorbate (Tween) 80

RES Reticuloendothelial system

SC Subcutaneous

SKBR-3 HER2 overexpressing breast cancer cell line

Trypsin-EDTA Trypsin-ethylenediamine tetraacetic acid

UPLC-MS Ultra performance liquid chromatography-coupled with mass spectrometry

8.2 List of parameters and mathematical notation

α Average number of antigen-antibody interactions

ε Molecular energy

Aµ Chemical potential of cell-bound nanoparticles

Sµ Chemical potential of nanoparticles in solution

113

Ω Average number of antibodies per particle

θ Fractional coverage, N/M

AUCall Area under the curve (to t = 8 h)

AUC∞ Area under the curve (to t = ∞)

AUMC∞ Area under the first moment curve (to t = ∞)

Co Initial plasma concentration

Cl Clearance

GΔ Molar Gibbs free energy of binding

MAXI Saturation fluorescence intensity

)( NPCI Concentration-dependent (measured) fluorescence intensity

kB Boltzmann constant

Kd Dissociation constant

eqK Equilibrium binding constant

Q(M,N,T) Canonical partition function for N nanoparticles binding to M binding sites at temperature T

M Available binding sites

N Number of bound nanoparticles

R Universal gas constant

T Temperature

t1/2, λ Lambda half life

Vd Volume of distribution

114

8.3 Additional data

Pharmacokinetics and biodistribution of docetaxel loaded in Herceptin-modified

nanoparticles

We investigated the PK and biodistribution of Herceptin-modified docetaxel-nanoparticles (NP-

aHER2-DTX-NP) using the same methods as in Chapter 4. There was no statistical difference

between the nanoparticle formulations in the plasma profile of docetaxel (Figure 8.1). Based on

this observation, we also anticipated similar tumour accumulation.

Figure 8.1 Plasma concentration profile for docetaxel in tumour-bearing mice. Orange symbols show

Herceptin-docetaxel-nanoparticles, black symbols show docetaxel nanoparticles, white symbols show

free docetaxel.

However, when the biodistribution panel was investigated, little accumulation was observed

across the panel, including in the tumour. The tissue profiles did not parallel those observed in

unmodified DTX-NP or free DTX (Figure 8.2).

115

Figure 8.2 Tissue concentration profile for docetaxel in tumour-bearing mice. Orange symbols show

Herceptin-docetaxel-nanoparticles, black symbols show docetaxel nanoparticles, white symbols show

free docetaxel.

This unexpected behaviour must result from physical changes to the nanoparticle system in the

presence of the Herceptin modification. It is unlikely that reduced kinetic stability led to DTX

release because the new behaviour did not resemble the profiles of free DTX. A more likely

explanation is that the altered nanoparticle surface properties presented new barriers to targeted

delivery. Particle composition and surface properties are key factors in determining

biodistribution and elimination rate [1]. As surface properties change (such as by addition of

IgG) the bound plasma protein profile changes [2, 3]; however, protein association is a balance

between what is most abundant and what has greatest affinity for the particle surface, meaning

that the profile will evolve over time [4]. The initial profile for both formulations is likely very

similar because proteins with general high concentration and high association rates will

dominate. Low concentration proteins will slowly displace some of these via higher affinity or

slower dissociation rates.

116

However, while extended circulation is a pre-requisite for EPR, passive targeting is not

guaranteed based on this property alone. Herceptin likely restricts the mobility of the

surrounding PEG chains through steric hindrance, limiting the flexibility of the construct. This

lack of deformation may have a direct impact on the ability of aHER2-DTX-NP to cross the gaps

presented by leaky tumour vasculature or other normal fenestrations in healthy organ tissue [5].

Additionally, particle geometry and flexibility impact the dynamics of motion in blood. As a

result the Herceptin modification may negatively impact the nanoparticles’ ability to enter and

drift to the edges of capillaries [6].

These results suggest that this formulation of poly(TMCC-co-LA)-g-PEG nanoparticles for

active targeting does not maximize the system’s potential for localized delivery. Although gains

in toxicity may still be observed via cellular uptake of the material that does enter the tumour [7],

a greater response would be expected with greater tumour accumulation. To return to particle

properties more similar to the unmodified nanoparticle formulation, it may be possible to use an

alternative targeting ligand to achieve active targeting. Using a smaller targeting ligand, such as

Fab antibody fragments, may provide improved tumour targeting by reducing steric resistance to

deformation or by making a lesser contribution to the surface properties of the modified

nanoparticles.

8.4 References

[1] Nel AE, Madler L, Velegol D, Xia T, Hoek EMV, Somasundaran P, et al. Understanding biophysicochemical interactions at the nano-bio interface. Nat Mater 2009;8:543-557.

[2] Schmidt S, Gonzalez D, Derendorf H. Significance of Protein Binding in Pharmacokinetics and Pharmacodynamics. J Pharm Sci-Us 2010;99:1107-1122.

[3] Aggarwal P, Hall JB, McLeland CB, Dobrovolskaia MA, McNeil SE. Nanoparticle interaction with plasma proteins as it relates to particle biodistribution, biocompatibility and therapeutic efficacy. Adv Drug Deliver Rev 2009;61:428-437.

[4] Rolan PE. Plasma protein binding displacement interactions--why are they still regarded as clinically important? Br J Clin Pharmacol 1994;37:125-128.

[5] Hwang HY, Kim IS, Kwon IC, Kim YH. Tumor targetability and antitumor effect of docetaxel-loaded hydrophobically modified glycol chitosan nanoparticles. J Control Release 2008;128:23-31.

[6] Caldorera-Moore M, Guimard N, Shi L, Roy K. Designer nanoparticles: incorporating size, shape and triggered release into nanoscale drug carriers. Expert Opin Drug Del 2010;7:479-495.

117

[7] Kirpotin DB, Drummond DC, Shao Y, Shalaby MR, Hong KL, Nielsen UB, et al. Antibody targeting of long-circulating lipidic nanoparticles does not increase tumor localization but does increase internalization in animal models. Cancer Res 2006;66:6732-6740.