Connecting Polymeric Nanomedicine and Systems Biology: An

Clemson University Clemson University

TigerPrints TigerPrints

All Theses Theses

December 2019

Connecting Polymeric Nanomedicine and Systems Biology: An Connecting Polymeric Nanomedicine and Systems Biology: An

Innovative Approach to Glioblastoma Treatment Innovative Approach to Glioblastoma Treatment

Jesse James Westfall Clemson University

Follow this and additional works at: https://tigerprints.clemson.edu/all_theses

Recommended Citation Recommended Citation Westfall, Jesse James, "Connecting Polymeric Nanomedicine and Systems Biology: An Innovative Approach to Glioblastoma Treatment" (2019). All Theses. 3197. https://tigerprints.clemson.edu/all_theses/3197

This Thesis is brought to you for free and open access by the Theses at TigerPrints. It has been accepted for inclusion in All Theses by an authorized administrator of TigerPrints. For more information, please contact [email protected].

https://tigerprints.clemson.edu/

https://tigerprints.clemson.edu/all_theses

https://tigerprints.clemson.edu/theses

https://tigerprints.clemson.edu/all_theses?utm_source=tigerprints.clemson.edu%2Fall_theses%2F3197&utm_medium=PDF&utm_campaign=PDFCoverPages

https://tigerprints.clemson.edu/all_theses/3197?utm_source=tigerprints.clemson.edu%2Fall_theses%2F3197&utm_medium=PDF&utm_campaign=PDFCoverPages

mailto:[email protected]

CONNECTING POLYMERIC NANOMEDICINE AND SYSTEMS BIOLOGY: AN INNOVATIVE APPROACH TO GLIOBLASTOMA TREATMENT

A Thesis Presented to

the Graduate School of Clemson University

In Partial Fulfillment of the Requirements for the Degree

Master of Science Chemical and Biomolecular Engineering

by Jesse J. Westfall December 2019

Accepted by: Jessica Larsen, PhD, Committee Co-Chair

Marc Birtwistle, PhD, Committee Co-Chair Mark Blenner, PhD F. Alex Feltus, PhD

ii

ABSTRACT

Glioblastoma (GBM) remains a highly lethal brain tumor that continues have overall low survival

rate, with only 5% of patients to five or more years. This thesis proposes a drug- and polymersome-

loaded thermosensitive hydrogel as a therapeutic platform to target and eliminate post-surgical

GBM tumor cells. The experiments presented lay some foundational work in establishing the

feasibility of and optimizing such a platform. Polyethene glycol (PEG)-Polyester pH-responsive

polymersomes were synthesized, optimized and conjugated with peptide ligands to increase cellular

uptake in vitro. A Python workflow was designed, using RNAseq data from the Cancer Genome

Atlas (TCGA) and the Genotype-Tissue Expression project (GTEx) with online proteomic and

binding databases, to find GBM cellular surface targets and the ligands that bind them. A

biomimetic hydrogel, enhanced by chemotaxis signaling molecules, was synthesized as an

alternative tumor eliminating modality. The continuation of this work, merging modern

nanomedicine synthesis techniques and disease-specific data analysis promises to have positive

implications for GBM patient prognosis.

iii

ACKNOWLEDGMENTS

I would first like to thank my advisors, Dr. Jessica Larsen and Dr. Marc Birtwistle. We started our

adventure at Clemson University together and it has been a pleasure to have them both as role

models these last two years. They constantly pushed me to pursue the projects that I was passionate

about, mentoring me at every turn. They have helped me to become a better researcher,

communicator and person.

I would also like to thank my other committee members, Dr. Alex Feltus and Dr. Mark Blenner.

Dr. Feltus has been more than generous with his time and resources and helped me get my project

off the ground. Dr. Blenner has been a supportive role model to me since the first day we spoke.

Thank you to the following people in the Department of Chemical and Biomolecular Engineering

at Clemson University: Nick Gregorich, Dr. Eric Davis, Allison Yaguchi, Maxwell Hilbert, Bipin

Chakravarthy-Paruchuri, Deepraj Sarmah, Charles Wang, Dr. Cemal Erdem, Caitlin Anglin and

my Creative Inquiry team: Chad Eaton, Riley Rapert and Asia Paguntalan. Without their support,

both in and out of the laboratory, I would have never been able to write this text.

Thank you to the thousands of scientists and engineers that have built the foundation on which this

thesis rests, and that have dedicated their lives to the pursuit of helping other people. The results

shown here are in whole or part based upon data generated by the TCGA research network.

Finally, I would like to thank my parents, James and Renae Westfall. They have supported me at

every step of my career. They have given me unconditional love from the moment I drew breath

and have molded me into the man I am today. Better parents simply do not exist.

iv

TABLE OF CONTENTS Page

1 – INTRODUCTION .................................................................................................................. 1 Barriers to GBM Treatment ................................................................................................... 2 Local Delivery by Nanoparticle-Loaded Hydrogels ............................................................. 4 A Systems Biology Approach ............................................................................................... 6 Thesis and Overview ............................................................................................................. 7 2 – POLYMERSOMES FOR LOCAL DELIVERY IN GBM .................................................... 8 Introduction ........................................................................................................................... 8 Methods ................................................................................................................................. 9 Experimental Results and Discussion ................................................................................. 11 Conclusions and Future Outlook ......................................................................................... 17 3 – POLYMERSOME LIGAND ATTACHMENT FOR BIOLOGICAL FUNCTION ............ 20 Introduction ......................................................................................................................... 20 Methods ............................................................................................................................... 21 Experimental Results and Discussion ................................................................................. 23 Conclusions and Future Outlook ......................................................................................... 27 4 – TARGETING GBM USING mRNA Data ........................................................................... 29 Introduction ......................................................................................................................... 29 Methods ............................................................................................................................... 30 Experimental Results and Discussion ................................................................................. 32 Conclusions and Future Outlook ......................................................................................... 35 5 – THE CHEMOTAXIS HYDROGEL .................................................................................... 37 Introduction ......................................................................................................................... 37 Methods ............................................................................................................................... 38

v

TABLE OF CONTENTS (CONTINUED) Page

Experimental Results and Discussion ................................................................................. 40 Conclusions and Future Outlook ......................................................................................... 44 APPENDIX A – PYTHON CODES .......................................................................................... 46 REFERENCES ........................................................................................................................... 50

]

vi

LIST OF FIGURES

Page Figure

1.1 – Kaplan-Meier survival plots for glioblastoma multiforme cases according to year of diagnosis and age .......................................................................................................................... 1 1.2 – Nanoparticle loaded hydrogels ............................................................................................ 5 2.1 – Illustration of a cross section of a polymersome ................................................................. 8 2.2 – Size distribution (diameter) of rehydrated PEGPLA particles .......................................... 13 2.3 – Structure of PEG-based block copolymers ........................................................................ 15 2.4 – DLS measurements for PEGPCL polymersomes – solvent injection ............................... 15 2.5 – DLS measurements for PEGPCL polymersomes – emulsion evaporation method........... 16 3.1 – Illustrative examples of how polymersomes functionalized with targeting ligands bind to cellular surface proteins .............................................................................................................. 20 3.2 – Structure of maleimide-functionalized PEG-based block copolymers .............................. 25 3.3 – Size distribution (diameter) M-PEGPCL polymersomes .................................................. 25 3.4 – Fold change in mean fluorescence (FITC) per HEK293 cell treated with AF488- loaded CTAT-Functionalized and non-functionalized M-PEGPLGA polymersomes ........................... 27 4.1 – Illustrative view of workflow identifying candidate genes ............................................... 33 4.2 – Illustrative view of candidate gene/protein GEPIA verification and ligand selection ...... 34 5.1 – Illustration demonstrating that eukaryotic cells migrate along a gradient toward a higher concentration of chemoattractants .............................................................................................. 38 5.2 – Crosslinking of hydrogel by covalent bonding of maleimide functionalized 4-Arm PEG to sulfhydryl pendant group ............................................................................................................ 40 5.3 – Tissue stiffness for normal tissues ..................................................................................... 42 5.4 – Illustration of transwell invasion assay modified for chemotaxis hydrogel ...................... 43

vii

LIST OF TABLES Page

Table

1.1 – Summary of the Barriers to GBM Treatment ...................................................................... 3 2.1 – PEGPLA Polymersomes in 2%wt/v Mannitol – DLS Measurements............................... 12 2.2 – PEGPLA Polymersomes Before and After Lyophilization – DLS Measurements ........... 14 2.3 – Advantages and Disadvantages of Varying Copolymer, Lyoprotectant and Synthesis Method........................................................................................................................................ 18 3.1 – MPEGM - PEGPCL Polymersomes Before and After conjugation with CysTAT Peptide24 3.2 – M-PEGPLGA Polymersomes Before and After Conjugation with CysTAT Peptide .................................................................................................................................................... 26 5.1 – Chemotaxis Hydrogel Formulation and Young’s Modulus............................................... 41

1

1 – INTRODUCTION

After decades of research, glioblastoma (GBM) remains an aggressive, incurable and lethal brain

tumor with a median survival time of approximately 15 months1. Within two years, many patients

develop recurrent tumors in the periphery of the resection cavity, leading, in nearly every case, to

death2. GBM continues to have a low overall survival rate, with only 5% of patients living to five

or more years3,4. The current “gold standard” treatment for GBM is to surgically remove all

accessible tumor mass followed by a combination of radiotherapy and oral temozolomide (TMZ)

chemotherapy3,5–8. Treatment with TMZ only increases the median survival of patients by an

average of four months, as GBM cells rapidly become resistant to the cytotoxic drug9 (figure 1.1).

Figure 1.1 – Kaplan-Meier survival plots for glioblastoma multiforme cases according to year of diagnosis and age. (A) 20–44 years. (B) 45–64 years. (C) 65–79 years. (D) 80+ years. Note that

2005-2007 data is in the era of TMZ and reflects the slight increase in median survival time. Adapted from10.

2

Barriers to GBM Treatment

Treatment Gap

There is a time gap, generally 2-3 weeks, between surgical resection of the GBM tumor and when

oral TMZ and radiotherapy begins11–13. Despite recent advances in tumor surgery for GBM14,15,

even extensive tumor resections may leave behind invasive tumor cells, as they can evade detection

by radiological imaging2. A small population of tumor cells, free to migrate and proliferate for

weeks in a wound healing environment, could be responsible for seeding a recurrent tumor.

Blood Brain Barrier

The blood brain barrier, (BBB) selectively transports molecules and cells to and from the central

nervous system (CNS), and acts as a barrier to pathogens and harmful chemicals16,17. This barrier,

formed by endothelial cells that form tight junctions, prevents 98% of small molecule drugs from

crossing from plasma of blood into the interstitial or cerebrospinal fluid of the brain18,19. Generally,

only small lipophilic drugs can passively diffuse across the BBB20. This drastically lowers the

number of intravenous and oral therapeutic drug candidates available for treating GBM.

General Toxicity

Chemotherapeutics, when taken orally or intravenously, often have off-target toxicity in sensitive

systems, like the kidneys, lungs and heart21. This puts a limit on the drug dosages that doctors are

able to use to treat GBM tumors. TMZ has a particularly toxic effect on bone marrow and liver

during treatment of GBM22. In some cases, particularly among the elderly, TMZ actually shortens

the survival time of the patient due to adverse effects23. These drugs are not only acutely toxic to

sensitive systems, but can lead to long-term secondary malignancies and sterility24.

3

Temozolomide Resistance

Even, though TMZ is capable of crossing the BBB, at least half of all patients treated with TMZ

do not respond to the drug. Tumors that do respond quickly become resistant to the drug, in as few

as one dose25. The cytotoxic effect of alkylating drugs like TMZ is stopped by O6-methylguanine

DNA-methyltransferase (MGMT), a suicide repair protein25,26. Attempts to target MGMT with a

small molecule drug or inhibitor, to hinder TMZ resistance, have not been successful9.

GBM Tumor Cell Heterogeneity

GBM tumor cells are phenotypically heterogeneous. Within the same tumor, there can be several,

phenotypically diverse sub-populations of tumor cells27–29. Furthermore, there is evidence that

GBM tumor cells are plastic, and can shift phenotypes in response to environmental stimuli,

including surgery and chemotherapy30,31. The heterogeneous and dynamic nature of GBM tumor

cells makes them difficult to treat with a single drug and is the driving force behind combination

drug therapy research for GBM.

Table 1.1 – Summary of Barriers to GBM Treatment

Treatment Gap

• Radiotherapy and TMZ treatment 2-3 weeks after tumor resection.

• Unresected GBM tumor cells can migrate and proliferate, leading to recurrent tumors.

General Toxicity • Oral or intravenous chemotherapeutics lead to toxicity

complications across vital systems in the body • Off-target toxicity limits doses of cytotoxic drugs

Blood Brain Barrier (BBB) • Prevents more than 98% of small molecule drugs from reaching the brain, limiting

TMZ Resistance • GBM tumor cells rapidly become resistant to cytotoxic agents, even after a single dose

Tumor Cell Heterogeneity

• GBM cells are phenotypically diverse, even in the same tumor.

• Tumors are dynamic and plastic, able to change phenotype in response to environment

4

Local Delivery by Nanoparticle-Loaded Hydrogels

Local drug delivery, in this case, delivering a drug directly into the GBM tumor resection site, has

the potential to overcome several treatment barriers. The drug can be implanted during resection

surgery, eliminating the treatment gap. This would also circumvent complications associated with

the BBB, as the drug would not have to cross from the blood plasma to the CNS. Moreover, as the

drug would not be circulating through the body in the bloodstream, concerns about toxicity in

sensitive systems would be eliminated. The question remains, as to how to deliver the drug to the

tumor cells with a sufficient dosage for long enough to eliminate the tumor mass.

Two drug delivery devices that have been researched for decades are drug-carrying hydrogels and

polymeric nanoparticles. A full review of these systems is beyond the scope of this text. However,

this section will focus on what makes these drug delivery tools attractive for local treatment of

GBM tumors.

Hydrogels

Hydrogels are networks of hydrophilic polymers that remain insoluble due to crosslinking between

the polymer chains. These networks are capable of holding aqueous solutions of drugs and other

therapeutic molecules32. These hydrogels can be synthesized from polymers that can undergo

conformational change when affected by an internal (pH, temperature, redox) or external (enzyme,

magnetic, light) stimulus33,34. Conformational changes can facilitate either the gelation or

degradation of the hydrogel, and when implanted, can be used as a mechanism for releasing a drug

solution into surrounding tissues35,36. A recently popular class of stimuli-responsive hydrogels for

the local treatment of GBM is thermosensitive hydrogels37–40. These solutions remain liquid at room

temperature, but gelate at physiological temperatures, holding a drug solution in place. This is

5

attractive for GBM treatment because a drug-loaded hydrogel can be injected directly into the

tumor resection site during surgery, where the drug can treat the remaining tumor cells.

Polymeric Nanoparticles

Polymeric nanoparticles are nanoscale, highly-dispersed, solid structures, made from synthetic or

organic polymers41. They have a variety of different functional geometries and are capable of

encapsulating hydrophilic and/or hydrophobic drugs41–45. Like hydrogels, when synthesized with

stimuli-responsive polymers, nanoparticles can undergo conformational changes or degradation, in

response to a stimulus, as a mechanism for releasing drugs46–50. Recently, several researchers have

begun to employ drug-carrying nanoparticles to target and treat GBM tumors51–53.

Nanoparticle-Loaded Hydrogels

In 2010, Arai et al. combined nanoparticles a with thermosensitive hydrogel, demonstrating that

the coupling of their properties resulted in prolonged release and therapeutic effects of doxorubicin

in vivo54. Since then, researchers have been combining nanoparticle and hydrogel systems to create

nanoparticle-loaded hydrogels and nanohydrogels (hydrogels synthesized by crosslinking

nanoparticles) 37,40,55,56 (Figure 1.2).

Figure 1.2 – Nanoparticle loaded hydrogels. (a) nanoparticles entrapped in polymer network (b) nanoparticles with physical interactions in polymer matrix (c) nanoparticles as the crosslinkers for

polymer network (d) nanogel made from crosslinked nanoparticles. Adapted from 40.

6

Recently, for treatment in GBM, Bastinach et al. demonstrated that a Lauroyl-gemcitabine

nanoparticle-based hydrogel (GemC12-LNC) lowered the rate of GBM recurrence when surgically

implanted in U87 mouse models. GemC12-LNC was also tolerated in these mouse models for up

to six months, suggesting a very low systemic toxicity38.

Nanoparticle-loaded hydrogels are an attractive platform for the treatment of GBM, as the

combination of stimuli-responsive modalities can facilitate combination therapy and more control

over drug release dynamics.

A Systems Biology Approach

TMZ resistance and GBM tumor cell heterogeneity are not barriers to GBM treatment that can be

overcome by formulating a new nanoparticle or hydrogel. These complications are a consequence

of the still ambiguous disease biology. To solve these problems, a greater understanding of GBMs

underlying biological mechanisms is required. Fortunately, systems biologists have been able to

employ genetic, transcriptomic, proteomic and metabolomic data to classify and quantify GBM

tumor cells in a number of useful ways.

For TMZ resistance, a number of different researchers are analyzing data and building models to

quantify and combat the mechanisms underlying MGMT repair and understand the prognostic

value of MGMT promoter methylation9,26,57,58. To better understand GBM heterogeneity,

researchers are combining new surgical sampling techniques with genomic analysis to track the

evolution dynamics of tumor cell populations31. Other biologists are using single-cell

transcriptomic data and gene-expression based analysis in an effort classify distinct GBM subtypes

for future researchers27,59. In addition, Databases like the Cancer Genome Atlas (TCGA), Cancer

Cell Line Encyclopedia (CCLE) and International Cancer Genome Consortium (ICGC) provide

7

enormous, curated, cancer-specific datasets that help inform the fields of drug discovery and

pharmacology60.

The continuing growth of publicly available biological data and improved analysis methodologies

is crucial to understanding GBM. This analysis can uncover methods to effectively target and

eliminate GBM tumor cells, and is fundamental in informing material, drug and delivery modalities

for future drug delivery systems.

Thesis and Overview

Using stimuli-responsive polymers to create a drug- and nanoparticle-loaded hydrogel, we can

overcome the barriers to improving GBM treatment and improve patient survival. The hydrogel is

to be injected intracranially, filling the cavity of tumor resection, where it releases both a cytotoxic

drug for a short period (<48 hrs), and targeting, drug-loaded polymersomes for a longer period (~40

days). This mode of localized drug delivery erases the complications of crossing the BBB and

systemic toxicity, while simultaneously eliminating the treatment time gap and post-surgical

proliferation of GBM tumor cells. Polymersomes with functionalized surfaces, encapsulating

targeted cancer drugs, facilitate drug delivery and elimination of the diverse array of GBM cell

phenotypes within a patient’s residual tumor.

This text will lay the foundation for the described drug delivery platform for treating GBM.

Chapters 1 focuses on synthesizing and optimizing nanoparticle properties for local treatment of

GBM. Chapter 2 extends the functionality of these particles through the conjugation of surface

ligands, in an effort to increase cellular particle uptake. Chapter 3 introduces a workflow to

elucidate GBM cellular surface targets and the ligands that bind them from transcriptomic datasets

and proteomic databases. Chapter 4 explores a novel tumor elimination modality using biomimetic

hydrogels loaded with chemotaxis signaling molecules.

8

2 – POLYMERSOMES FOR LOCAL DELIVERY IN GBM

Introduction

Amphiphilic block copolymers self-assemble, under aqueous conditions, into a variety of

nanoparticle geometries. However, for polyethylene glycol (PEG) based copolymers, when the

mass of the hydrophilic block is between ~35% ± 10% they assemble into a hollow vesicle known

as a polymersome61. The hydrophobic blocks of the copolymer forms a hydrophobic bilayer

surrounding an aqueous core. This allows the nanoparticle to carry hydrophobic drugs in the

membrane interior and hydrophilic drugs in the core62 (Figure 2.1). Being able to control the

chemical properties and molecular weight of the polymers, polymersomes typically have thicker,

more chemically stable membranes than earlier drug nanoparticles with similar geometries, like

liposomes63.

Figure 2.1 – Illustration of a cross section of a polymersome. The hydrophobic membrane (red) encapsulates hydrophobic molecules while the aqueous core encapsulates water-soluble

compounds. Adapted from 62.

There are a variety of polymers that can undergo conformational change when affected by a

stimulus. When block copolymers are synthesized using these stimuli-responsive polymers, they

9

can assemble into polymersomes that degrade or undergo conformational change in response to

internal (pH, temperature, redox) and external (enzyme, magnetic, light) stimuli by degradation or

conformational change. These mechanisms can be exploited to control drug release from either

aqueous or hydrophobic compartments of the nanoparticles46–49. There has been research

performed using these stimuli-responsive polymersomes for the treatment of GBM53. Using

internal stimuli, Jiang et al. engineered redox-responsive polymersomes to effectively deliver

saporin, a protein toxin therapeutic, in GBM mouse models64. In contrast, Luo et al. used high-

intensity focused ultrasound, an external stimuli, to trigger release and effective delivery of

doxorubicin and perfluorooctyl bromide in a U87 mouse model65.

PH-triggered drug release is attractive mechanism for cancer treatment, due to the acidity of both

the tumor microenvironment and tumor cell endosomes46. Polymersomes assembled from

copolymers with blocks degradable by acid hydrolysis, like polyesters, can be designed to degrade

at different rates in acidic conditions66. It has also been shown that nanoparticles with PEG on their

surface avoid uptake by phagocytic cells as part of the reticuloendothelial system in vivo67. This is

attractive for long-term local therapy of GBM, as microglia are phagocytes that digest particles

within the CNS68.

Polymersomes synthesized from PEG-polyester block copolymers can benefit from both controlled

pH-responsive controllable release and evasion of microglia in the brain. This chapter explores

material and synthesis design choices to for engineering polymersomes optimized for local delivery

to GBM tumor cells.

Methods

Solvent Injection Method for Synthesis of Polymersomes – 1 mg diblock copolymer methoxy

polyethylene glycol polylactide (1kDa-5kDa, Polysciences Inc #24381-1) polyethylene glycol

10

polycaprolactone (2kDA-5kDA) (Aldrich #900648) (PEGPCL) was dissolved in 100 μL miscible

organic solvent dimethyl sulfoxide (VWR BDH Chemicals #BDH1115-1LD) (DMSO), N-methyl-

2-pyrrolidone (VWR Life Science #89500-566) (NMP) or N,N-dimethylformamide (BDH

Chemicals #BDH83634.100) (DMF). In a 4 dram threaded vial, the solution was injected via 0.5

mL syringe (BD #305602) by nano-syringe pump (KD Scientific) into 10 mL continuously stirring

2wt%/v mannitol (BDH VWR Analytical #BDH9248) or inulin (Alfa Aesar #A18425) in water at

a constant rate of 5 μL/min.

Emulsion Evaporation Method for Synthesis of Polymersomes – 1 mg diblock copolymer

methoxy polyethylene glycol polylactide (1kDa-5kDa, Polysciences Inc #24381-1) (PEGPLA) or

methoxy polyethylene glycol polycaprolactone (2kDA-5kDA) (Aldrich #900648) (PEGPCL) was

dissolved in 1 mL immiscible solvent dichloromethane (VWR BDH Chemicals #BDH23373.100E)

(DCM). The solvent solution was added to 9 mL 0.5 wt%/v Pluronic F-68 (Gibco #24040-032) in

water in an 11 dram threaded vial. The mixture was emulsified using a homogenizer (Thermo

Fisher), three cycles of 5 sec on/5 sec off at 50% power. The vial containing the emulsion was

connected to a Rotavapor R-100 (Buchi) via 24/40 vial adapter (Chemglass #CG-1318-40) and

rotated under vacuum until bubbling stopped and solution was clear.

Lyophilization and Characterization of Polymersomes – Polymersome solution was frozen

slowly, first to -20°C for 6 hours, then to -80°C overnight, in a 50 mL conical tube. The solution

was then then lyophilized at 0.04 mbar and -105°C by FreeZone lyophilizer (LABCONCO).

Polymersome diameters and zeta potential were measured by dynamic light scattering (DLS) via

Zetasizer Nano ZS (Malvern).

Loading of Polymersomes – Lyophilized polymersomes were rehydrated in 1 mL aqueous

solution (0.16 mg/mL in water) of Alexa Fluor 488 carboxylic acid tris(triethylammonium) salt

11

(Invitrogen #A33077) (AF488). Unencapsulated AF488 was separated from loaded polymersomes

by centrifugal filter device (Amicon 0.5 mL 100K #UFC5100BK), spinning three times at 14000g

for 10 minutes in a 5424R centrifuge (Eppendorf). Encapsulation efficiency (EE) was determined

with the following equation:

𝐸𝐸𝐸𝐸(%) = 100%− (𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑚𝑚𝑇𝑇𝑚𝑚𝑚𝑚 (𝑚𝑚𝑚𝑚) 𝑇𝑇𝑜𝑜 𝐴𝐴𝐴𝐴488 𝑖𝑖𝑖𝑖 𝑜𝑜𝑖𝑖𝑇𝑇𝑇𝑇𝑓𝑓𝑇𝑇𝑇𝑇𝑓𝑓𝑚𝑚

0.16 𝑚𝑚𝑚𝑚 × 100%)

Mass of AF488 in filtrates was calculated using a linear calibration curve (relative fluorescence

units (RFU) vs. AF488 concentration). Fluorescence measurements were made with Synergy H1

microplate reader (BioTek).

Experimental Results and Discussion

The goal of these experiments is to optimize polymersome properties for local delivery in GBM.

The ideal polymersome will have a diameter between 10-200 nm, to prevent rapid clearance by the

renal system and opsonization by phagocytes in the reticuloendothelial system33, be monodisperse,

and be capable of carrying both hydrophobic and/or hydrophilic molecules for delivery. The

polymersomes should also have a significant negative zeta potential, a proxy for nanoparticle

surface charge, and indicator of colloidal stability in solution69. It is also important that the

polymersomes are capable of releasing these molecules in a pH-responsive manner for delivery in

either the acidic GBM tumor microenvironment or acidic intracellular vesicles. Another important

metric to be considered is whether the polymersomes can be lyophilized for long term storage and

rehydrated for future use. These experiments focus on choosing copolymers, synthesis methods

and lyoprotectants to synthesize polymersomes ideal for GBM delivery.

12

PEGPLA polymersomes in 2wt%/v mannitol in water – The solvent injection method employed

for synthesizing polymersomes is adapted from Kelly et al70. Table 2.1 shows the size,

polydispersity and zeta potential of the resulting particles.

Table 2.1 – PEGPLA Polymersomes in 2%wt/v Mannitol – DLS Measurements

Diameter (nm) PDI Zeta Potential (mV)

149.00 ± 13.63 0.088 ± 0.022 -30.56 ± 0.91

These polymersomes are in the desired size range, are monodisperse (polydispersity index (PDI) <

0.3)71 and have a large negative zeta potential. Also, the polylactic acid block comprising the

hydrophobic membrane is pH-responsive by acid hydrolysis. These characteristics make these

polymersomes a good candidate for GBM local delivery, assuming they can be loaded with

therapeutic molecules and lyophilized for long term storage.

Lyophilization and loading PEGPLA polymersomes – PEGPLA polymersomes were synthesized

via solvent injection using 2% mannitol and rehydrated by 0.16 mg/mL AF488 dye solution. AF488

is a hydrophilic molecule and is encapsulated in the aqueous core of the polymersomes. Dye

encapsulation efficiency of the polymersomes is shown to be as high as 38.6%, indicating that these

particles are promising carrier for soluble therapeutic molecules. However, DLS measurements

indicate that the rehydrated polymersomes are not monodisperse, likely destabilizing and forming

aggregates (Figure 2.2 (top)). Particles were synthesized in mannitol solutions up to 10%, with

similar results.

13

Figure 2.2 – Size distribution (Diameter) of rehydrated PEGPLA particles (top) in 2%wt/v mannitol in water(bottom) in 2%wt/v inulin in water.

PEGPLA polymersomes were re-synthesized via solvent injection, using 2wt%/v inulin in water as

the lyoprotectant. It has been shown that inulin, a fructose polymer, is effective at lyoprotection of

PEG-based polymersomes72,73. The particles are of similar size and dispersity as the mannitol-

protected particles. When these polymersomes are lyophilized and rehydrated using a 0.16 mg/mL

AF488 dye solution, they maintain their membrane stability and monodisperse nature (Figure 2.2

(bottom)), albeit with a slight increase in diameter. However, these polymersomes show a lower

capacity for soluble molecule encapsulation, having an AF488 encapsulation efficiency of 25.4%

± 2.2%. This is likely due to the larger size of inulin, making it less likely to leave the membrane

14

during rehydration.. Table 2.2 shows the results of DLS measurements on these PEGPLA

polymersomes.

Table 2.2 – PEGPLA Polymersomes Before and After Lyophilization – DLS Measurements

Before lyophilization Rehydrated after Lyophilization Lyoprotectant Diameter (nm) PDI Diameter (nm) PDI 2% mannitol 167.41 ± 25.29 0.108 ± 0.055 100.48 ± 42.01 0.611 ± 0.350 2% inulin 122.14 ± 7.74 0.119 ± 0.023 158.22 ± 6.13 0.175 ± 0.033

Another consideration when choosing a lyoprotectant is the characteristics of the solid product.

Polymersomes lyophilized in mannitol form a powdery substance, easy to store and transfer

between containers. Polymersomes lyophilized in inulin form a slow-flowing, translucent

amorphous, solid, which at room temperature can be difficult to handle and transfer.

While PEGPLA polymersomes are great candidates for pH-responsive delivery of drugs

intravenously, they may be less stable when subjected to extended time frames in the slightly acidic

pH conditions of the GBM tumor microenvironment. In this environment, the polymersome

membrane may prematurely destabilize and release a drug before it is able to reach the GBM tumor

cell. Thus, copolymer blocks made with polycaprolactone (PCL) are attractive as the polymer is

more hydrophobic than PLA. There is a longer carbon chain in each monomer, leading to stronger

hydrophobic interactions in an aqueous environment. There are also less ester groups, sites for acid

hydrolysis, per block (Figure 2.3). These characteristics should lead to stronger membrane

interactions, and longer stability in the acidic tumor microenvironment, while maintaining a pH-

responsive drug release mechanism.

15

Figure 2.3 – Structure of PEG-based block copolymers, (left) PEGPLA. (right) PEGPCL. Highlighted in red is the long carbon chain that leads to stronger hydrophobic interactions

between polymers in solution.

PEGPCL polymersome synthesis – PEGPCL polymersomes were synthesized via solvent injection

using two different organic solvents, NMP and DMF. This was necessary as PEGPCL is not soluble

in DMSO, like PEGPLA is Figure 2.4 shows the results of DLS measurement of the PEGPCL

polymersomes.

Although PEGPCL particles have smaller diameters when synthesized at the same concentration

as the PEGPLA polymersomes, it is shown that size can be modulated by controlling the

concentration of the copolymer in the organic phase during synthesis.

Figure 2.4 – DLS measurements for PEGPCL polymersomes – solvent injection.

PDI0.079

PDI0.086

PDI0.113

0

20

40

60

80

100

120

DMF 1mg/mL NMP 1mg/mL NMP 2mg/mL

Z-AV

G DI

AMET

ER (N

M)

16

While particles synthesized via solvent injection effectively encapsulate soluble compounds in their

aqueous core, there are GBM therapeutics that have very low solubility in water. There is another

mode of synthesis, single emulsion evaporation (EE) that simultaneously encapsulate hydrophobic

molecules with high efficiency74.

Emulsion evaporation synthesis of PEGPCL polymersomes – PEGPCL polymersomes were

prepared via EE method, adapted from Khalil et al75, with varying concentrations (0.7, 1, 1.3 mg

PEGPCL/1 mL DCM). Figure 2.5 shows the DLS measurements for the PEGPCL polymersomes.

Figure 2.5 – DLS measurements for PEGPCL polymersomes – emulsion evaporation method.

At 1 and 0.7 mg PEGPCL/1 mL DCM and lower concentrations, the polymersomes were desirable

size and monodisperse. At 1.3 mg PEGPCL/1 mL DCM, however, polymersome solutions became

polydisperse. More curious is that there is a negative correlation between polymer concentration

and average diameter. This is in contrast to the positive correlation between polymer concentration

and diameter for solvent injection method discussed earlier. This suggests that polymer

PDI0.299

PDI0.267 PDI

0.426

0

50

100

150

200

0.7 1 1.3

Z-AV

G DI

AMET

ER (N

M)

WT PEG-PCL/VOL DCM (MG/ML)

17

concentration may not be the driving force in the polymersome self-assembly process when using

the EE method. Although literature suggests that size modulation can be achieved through process

design (homogenization, surfactant content, solvents etc.)74,76,77.

Conclusions and Future Outlook

This chapter illustrates the types of material design choices a polymeric nanomedicine researcher

must make when designing drug delivery systems. Table 2.3 summarizes the material, lyoprotectant

and synthesis choices explored in this work.

18

Table 2.3 – Advantages and Disadvantages of Varying Copolymer, Lyoprotectant and Synthesis Method

Polymersome Synthesis Method Pros Cons Solvent

Injection

• Very monodisperse polymersome solutions

• Diameters easily modulated by polymer concentration

• Could be scaled reasonably

• Inefficient loading of hydrophilic molecules when synthesizing and loading simultainiously

Emulsion

Evaporation

• Synthesizing and loading of polymersomes with hydrophobic molecules simultaneous

• High entrapment of hydrophobic molecules

• Much higher PDI than solvent injection method

Lyoprotectant Pros Cons Mannitol • Cheap and simple molecule

• Through rehydration, high encapsulation efficiency of water-soluble molecules

• Lyophilized polymersomes are not monodisperse upon rehydration

• Lyophilized solutions will need to be filtered to remove aggregates

Inulin • Lyophilized polymersomes are monodisperse upon rehydration

• No need to filter rehydrated polymersome solutions

• Expensive, large polymer molecules

• Lower encapsulation of water-soluble molecules through rehydration

Amphiphilic Block Copolymer Pros Cons PEGPLA • pH-responsive acid hydrolysis

excellent mechanism for drug release

• Soluble in less toxic organic solvents.

• Polymersomes may prematurely destabilize in acidic tumor microenvironment.

PEGPCL • pH-responsive acid hydrolysis excellent mechanism for drug release

• Polymersomes will remain stable for longer in acidic tumor microenvironment

• Only soluble in more toxic solvents like NMP and DMF

19

Looking forward, for a local delivery system for the treatment of GBM, PEGPCL polymersomes

are the best nanoparticle on which to build a platform. It has been shown that they can be prepared

by different methods for the loading of both hydrophilic and hydrophobic therapeutic molecules.

These polymersomes can also be loaded and lyophilized, with inulin as a lyoprotectant, for long-

term storage. For local delivery in GBM, PEGPCL polymersome solutions can be used to solubilize

thermosensitive polymers to create polymersome-loaded, temperature-responsive hydrogels. These

hydrogels can be injected into GBM tumor resection sites for extended local treatment of remaining

tumor cells. Furthermore, these polymersomes can be functionalized with different ligands to

enhance biological function, including cell surface targeting and cellular uptake, which will be

explored in the next chapter.

20

3 – POLYMERSOME LIGAND ATTACHMENT FOR BIOLOGICAL FUNCTION

Introduction

Recent research have shown that using ligands to target membrane surface targets can lead to

increased biological function, like initiating receptor-mediated endocytosis and crossing the blood

brain barrier (BBB)78. For example, it was demonstrated that functionalizing nanoparticles with

ApoE, a lipoprotein, will increase the transport of those particles across the blood brain barrier via

interaction with low-density lipoprotein receptors79, which are overexpressed on BBB interfacial

cells17,18. The focus of this thesis, local delivery to GBM, is to bypass the BBB entirely, but we can

use the same concept of active targeting to increase effective drug delivery to GBM cells. In

particular, polymersomes are effective vessels for getting drugs through the membranes of cells

when their surfaces are functionalized with ligands, proteins, antibodies and peptides78 (Figure 3.1).

These functional molecules promote cell adhesion and receptor mediated endocytosis, increasing

the rate at which drugs are internalized by a cancer tumor cell44,51,78,80.

Figure 3.1 – Illustrative examples of how polymersomes functionalized with targeting ligands bind to cellular surface proteins. Adapted from 78.

21

Using active targeting ligands is an attractive modality for the local treatment of GBM. There are

cellular surface proteins, overexpressed in GBM, capable of being targeted (this be explored more

thoroughly in the next chapter). The binding affinities of the ligand to those proteins will facilitate

the preferential delivery of chemotherapeutics to GBM tumor cells.

This chapter will focus on demonstrating that, by functionalizing the surface of polymersomes

discussed in Chapter 2, cellular uptake of these polymersomes can be significantly increased in

vitro. This will provide a platform for synthesizing polymersomes functionalized with active

targeting ligands, and how to quantify their drug delivery efficiency in vitro.

Methods

Embedding Method for Synthesizing Maleimide-Functionalized Polymersomes – 1.5 mg

diblock copolymer methoxy polyethylene glycol polycaprolactone (2kDA-5kDA) (Aldrich

#900648) (PEGPCL) and varying masses (5,10,40 mg) of homobifunctional polyethylene

glycol(3500kDa)-(maleimide)2 (JENKEM Technology USA #A4010-1) (MPEGM) were dissolved

in 100 μL N,N-dimethylformamide (BDH Chemicals #BDH83634.100) (DMF). In a 2 dram

threaded vial, the solvent solution was injected via 0.5 mL syringe (BD #305602) by nano-syringe

pump (KD Scientific) into 1.4 mL continuously stirring Alexa Fluor 488 carboxylic acid

tris(triethylammonium) salt (Invitrogen #A33077) (AF488) (0.114 mg/mL) in water at a constant

rate of 5 μL/min. Unencapsulated AF488 and non-embedded MPEGM were separated from loaded

polymersomes by centrifugal filter device (Amicon 0.5 mL 100K #UFC5100BK), spinning three

times at 14000g for 10 minutes in a 5424R centrifuge (Eppendorf).

Synthesizing Maleimide-Functionalized Polymersomes with M-PEGPLGA/PCL – 1 mg

Poly(ε-caprolactone)-PEG-maleimide (2kDa-5kDa) (Nanosoft Polymers #2667) (M-PEGPCL), or

poly(lactide-co-glycolide)-PEG-maleimide (2kDa-5kDa) (M-PEGPLGA) (Nanosoft Polymers

22

#2794) was dissolved in N,N-dimethylformamide (BDH Chemicals #BDH83634.100) (DMF). In

a 2 dram threaded vial, the solvent solution was injected via 0.5 mL syringe (BD #305602) by nano-

syringe pump (KD Scientific) into 1.4 mL continuously stirring Alexa Fluor 488 carboxylic acid

tris(triethylammonium) salt (Invitrogen #A33077) (AF488) (0.114 mg/mL) in water at a constant

rate of 5 μL/min. Unencapsulated AF488 was separated from loaded polymersomes by centrifugal

filter device (Amicon 0.5 mL 100K #UFC5100BK), spinning three times at 14000g for 10 minutes

in a 5424R centrifuge (Eppendorf).

Encapsulation Efficiency and Polymersome Characterization – AF488 Encapsulation

efficiency (EE) was determined with the following equation:

𝐸𝐸𝐸𝐸(%) = 100%− (𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇𝑇 𝑚𝑚𝑇𝑇𝑚𝑚𝑚𝑚 (𝑚𝑚𝑚𝑚) 𝑇𝑇𝑜𝑜 𝐴𝐴𝐴𝐴488 𝑖𝑖𝑖𝑖 𝑜𝑜𝑖𝑖𝑇𝑇𝑇𝑇𝑓𝑓𝑇𝑇𝑇𝑇𝑓𝑓𝑚𝑚

0.16 𝑚𝑚𝑚𝑚 × 100%)

Mass of AF488 in filtrates was calculated using a linear calibration curve (relative fluorescence

units (RFU) vs. AF488 concentration). Fluorescence measurements were made with Synergy H1

microplate reader (BioTek). Polymersome diameters, polydispersity index (PDI) and zeta potential

were measured via Zetasizer Nano ZS (Malvern).

Conjugation of CysTAT ligand to Polymersomes – 0.25 mg CysTAT(47-57) (Genscript

#RP20343) (amino acid sequence: GRKKRRQRRRPQ) (CTAT) was added to aqueous solution of

0.5 mg loaded, maleimide-functionalized polymersomes in water. Mixture was stirred overnight

(~18 hours). Non-conjugated CTAT was separated from polymersomes by centrifugal filter device

(Amicon 0.5 mL 100K #UFC5100BK), spinning three times at 14000g for 10 minutes in a 5424R

centrifuge (Eppendorf).

HEK293 Cell Culture and Treatment – HEK293 cells (ATCC #CRL-1573, verified by short

tandem repeat profiling) were seeded in a 12-well plate, at 0.1 x 106 cells per well and incubated

23

for 24 hours at 37 °C and 5% CO2 in Delbecco’s modified eagle medium (Gibco #10313-021)

(+10% FBS). At ~50% confluence, the cells were treated with 0.5 mg loaded M-PEGPLGA

polymersomes in Delbecco’s phosphate buffered solution (Alfa Aesar #J61917) (PBS), 0.5 mg non-

loaded M-PEGPLGA polymersomes in PBS and 1 μL (2 mg/mL) AF488 in milli-Q water. Cells

were allowed to incubate for 3.5 hours at 37 °C and 5% CO2.

Flow Cytometry to Determine Polymersome Uptake – HEK293 cells removed from 12-well

plate surface by cell scraper and 1 mL of cell suspension per sample was collected. Flow cytometry

was performed to determine the mean cell FITC-A signal for each sample using a CytoFLEX LX

(Beckman Coulter).

Results and Discussion

This chapter’s series of experiments is designed to show that by conjugating ligands to our

polymersome drug carriers, we can introduce biological function, such as cell surface targeting and

enhanced cellular uptake, beneficial to drug delivery. It this case, transactivator of transcription

(TAT) of HIV will be the ligand and will act as a positively charged, cell-penetrating peptide (CPP)

to increase cellular uptake of polymersomes.

Embedding MPEGM in polymersome membrane – In an effort to locate maleimide functional

groups to the surface of PEGPCL polymersomes, Varying masses of homobifunctional PEG

polymers (MPEGM) were dissolved with 1.5 mg PEGPCL in the organic solvent. Polymersomes

are then synthesized using solvent injection and stirred overnight with 0.25 mg cysteine-terminal

TAT (CTAT). A thiol-maleimide conjugation reaction covalently bonds the CTAT to the PEG

polymer embedded in the polymersome membrane. Table 3.1 shows the polymersome

characteristics, measured by DLS both before and after conjugation with CysTAT peptide.

24

Table 3.1 – MPEGM - PEGPCL Polymersomes Before and After conjugation with CysTAT Peptide

mg MPEGM per 1.5 mg PEGPCL

CTAT Diameter (nm) PDI Zeta potential (mV)

5mg - 92.71 0.118 -31.4 + 88.87 0.081 -20.9

10mg - 92.33 0.103 - + 86.89 0.090 -20.3

40mg - 88.94 0.117 -17.8 + 105.9 0.136 -10.0

It would be expected, that if there was an appreciable amount of positively charged CysTAT peptide

on the surface of the polymersomes, the zeta potential, a proxy for surface charge, would be

positive, or at least have a large positive shift. However, after CysTAT conjugation, the

polymersomes still show a significant negative zeta potential. This underwhelming result could be

a consequence of either the lack of embedded homobifunctional PEG polymers. There may not be

sufficient driving force for much the hydrophilic MPEGM to become embedded in the

polymersome membrane. Instead the MPEGM would be free in solution and subsequently washed

away by the centrifugal device. Without enough embedded MPEGM, there are not enough

maleimides on the surface of the particles for the thiols of the CTAT to bind to.

To eliminate complications derived from incomplete or heterogeneous embedding of MPEGM in

the polymersome membrane, polymersomes were synthesized from PEG-based block copolymers

in which a maleimide functional group was already covalently bonded to the terminal end of the

PEG. For this purpose, maleimide functionalized PEG-polycaprolactone (M-PEGPCL) and PEG-

poly(lactide-co-glycolide) (M-PEGPLGA) were purchased (figure 3.2).

25

Figure 3.2 – Structure of Maleimide-Functionalized PEG-based Block Copolymers, (top) M-PEGPCL. (bottom) M-PEGPLGA.

Using solvent injection method, M-PEGPCL polymersomes are synthesized and characterized

using DLS. These polymersomes have surprisingly small diameters and are quite polydisperse

(figure 3.3 (top)). When a 50:50 blend of M-PEGPCL and PEGPCL is synthesized into

polymersomes by solvent injection method, the resulting nanoparticles are more monodisperse,

though still smaller than their non-functionalized counterparts from the previous chapter (figure 3.3

(bottom)).

Figure 3.3 – Size distribution (diameter) M-PEGPCL polymersomes (top) 100% M-PEGPCL (bottom) 50% M-PEGPCL, 50% PEGPCL.

26

When polymersomes were synthesized from M-PEGPLGA using solvent injection method, they

were in the target diameter, PDI and zeta potential ranges established in the previous chapter (See

Table 3.2 below). While PEGPCL is preferable because of its increased stability in the GBM tumor

microenvironment, the following in vitro experiments are performed at neutral pH. As the purpose

of this chapter is to demonstrate an increase in biological function by ligand attachment, the focus

will be on conjugating M-PEGPLGA polymersomes with CTAT in an effort to increase cellular

uptake.

M-PEGPLGA polymersomes were synthesized by solvent injection method then stirred overnight

with 0.25mg of CTAT peptide overnight. Table 3.2 shows diameter, PDI and zeta potential of these

polymersomes both before and after CTAT conjugation.

Table 3.2 – M-PEGPLGA Polymersomes Before and After Conjugation with CTAT Peptide

M-PEGPLGA polymersomes CTAT conjugation

Diameter (nm) PDI Zeta potential (mV)

CysTAT - 90.29 0.133 -33.6 CysTAT + 84.47 0.145 -0.723

A large positive shift, 32.88 mV, is observed after CTAT conjugation, indicating a large number

of covalently bonded CTAT peptide molecules on the surface of the polymersomes.

Observing enhanced cellular uptake by fluorescence flow cytometry – M-PEGPLGA particles are

synthesized by solvent injection with that the aqueous phase of the procedure containing 0.16 mg

AF488 dye. The resulting polymersomes, measured by DLS, had diameters of 87.87 ± 5.96 nm,

PDI of 0.145 ± 0.013, and zeta potentials of -39.2 ± 2.7 mV (n=3). Each batch of polymersomes

was split and half of each sample was stirred with 0.125 mg CysTAT overnight. HEK293 cells

were treated with the loaded polymersomes, both CysTAT- conjugated and unconjugated, for 3.5

27

hours. The cells were analyzed on a flow cytometer to determine the mean fluorescence (FITC) of

each cell population. Normalizing for cell autofluorescence, the mean fluorescence of cells treated

with CysTAT-conjugated increased nearly two-fold over those treated with nonconjugated

polymersomes (Figure 3.4).

Figure 3.4 – Fold change in mean fluorescence (FITC) per HEK293 cell treated with AF488- loaded CTAT-Functionalized and non-functionalized M-PEGPLGA polymersomes


This chapter illustrates that polymersomes are not just a platform for delivering drug combinations

and extending drug circulation times in patients, but rather they can be functionalized to perform

specific biological functions. In this case, CysTAT ligands increased the rate of cellular uptake of

polymersomes, albeit indiscriminate of cell type. The system is modular, and using similar

bioconjugation techniques, a variety of macromolecules, antibodies, peptides, proteins etc. can be

conjugated to a polymersome surface. In the context of local delivery, particularly in GBM, using

ligands that will increase cellular uptake or adhesion to tumor-specific cell types is an attractive

option.

0

0.5

1

1.5

2

Fold

Cha

nge

Without TAT With TAT

28

Moving forward, the importance of discovering biologically relevant ligands that are specific to the

treatment of specific diseases is paramount, as is optimizing bioconjugation techniques to attach

the ligands to therapeutic polymersomes. This flavor of biology-specific functionalization of

polymersomes could be used to increase drug delivery efficacy for a variety of genetic disorders

and cancers. For GBM, a methodology for finding tumor-specific targets, and the ligands that will

help target them, will be touched on in the next chapter.

29

4 – TARGETTING GBM USING mRNA DATA

Introduction

In previous chapters, it was demonstrated that PEG-polyester polymersomes were potentially

suitable candidates for delivering encapsulated drugs locally to GBM. Furthermore, those

polymersomes can be functionalized with ligands designed to target and improve drug delivery

efficiency to cells. The goal of this chapter is to elucidate GBM surface proteins that, when targeted,

will facilitate an increase in therapeutic efficacy of drug-loaded polymersomes.

The phenotypic heterogeneity of GBM tumor cells confounds the discovery of an effective cellular

surface target. Within a single GBM tumor, there are distinct populations of tumor cells with

varying levels of protein expression27–29,31,81. These populations of tumor cells change phenotypes

in response to stimuli, including immune responses to surgery and chemotherapy82–84. However,

systems biologists have been able to employ genetic, transcriptomic and proteomic data to classify

and quantify heterogeneity of GBM tumor cells in a number of useful ways. Researchers are

combining new surgical sampling techniques with genomic analysis to track the evolution

dynamics of tumor cell populations31. Other biologists are using single-cell transcriptomic data and

gene-expression based analysis in an effort classify distinct GBM subtypes for future

researchers27,59. Furthermore, databases like the Cancer Genome Atlas (TCGA), Cancer Cell Line

Encyclopedia (CCLE) and International Cancer Genome Consortium (ICGC) provide enormous,

curated, cancer-specific datasets that help inform the fields of drug discovery and pharmacology60.

This chapter will focus on using a combination of databases to identify a set of highly expressed

GBM tumor cell surface targets and the ligands that bind to them. These ligands can be conjugated

to drug-carrying polymersomes, creating a combination therapy platform designed to target

multiple GBM cell phenotypes in the same tumor.

30

Methods

Determining High Expression Transcripts in GBM – Feltus TCGA GEM analysis - A gene

expression matrix (GEM), describing fragments per kilobase million (FPKM) of 73599 mRNA

transcripts across 2016 cancerous tumor samples [bladder, thyroid, ovarian, low-grade glioma,

glioblastoma, data from The Cancer Genome Atlas (TCGA)] was built by William Poehlman

(Department of Genetics and Biochemistry, Clemson University), which had been previously

quantile normalized and values converted to log2 values (No outliers detected using Kolmogorov-

Smirnov test)85. From Dr. Alex Feltus (Department of Genetics and Biochemistry, Clemson

University), a key was obtained to convert the knowngene5 UC-Santa Cruz genome database gene

model identifiers to Ensembl gene identifiers, and the transcripts that had no Ensembl ID were

omitted from the GEM. A median FPKM value across the 174 GBM tumor samples was obtained

for each transcript. The transcripts with the 12000 highest median FPKM values were used to

determine genes of high expression in GBM. Westfall TCGA GEM Analysis – RNASeq expression

profiles for 174 GBM tumor samples were downloaded from the TCGA (10/30/2018). The profiles

were merged to form a single GEM. Using python, the data was quantile normalized, and FPKM

values were converted to log2. A median FPKM value across the tumor samples was obtained for

each transcript. The transcripts with the 12000 highest median FPKM values were used to

determine genes of high expression in GBM. (Python code: TCGA_analysis_FELTUS,

TCGA_Analysis_Westfall in Appendix A)

Comparing High Expression GBM Transcripts Against Normal/Low Expression in Normal

Brain Tissue – A GEM, describing FPKM of 56202 mRNA transcripts across 1671 non-diseased

brain tissue samples from the Genotype-Tissue Expression project (GTEx)86, was obtained from

Dr. Alex Feltus (Department of Genetics and Biochemistry, Clemson University), which had been

previously quantile normalized and converted to log2 values. A median FPKM value across the

31

tissue samples was obtained for each transcript. The transcripts with the 5000 highest median

FPKM values were used to determine genes of high expression in healthy brain tissue and were

omitted from the data. The remaining transcripts in the GEM were considered to be of normal or

low expression. The Ensembl gene IDs of the high expression transcripts in the Feltus/Westfall

TCGA GEMs were compared with the Ensembl gene IDs of the normal/low transcripts. Any

overlap in the set identifies a transcript that is highly expressed in GBM and not highly expressed

in non-diseased brain tissue. (Python Code: GTEX_analysis,

TCGA_GTEX_Comparison_FELTUS, TCGA_GTEX_Comparison_Westfall,

Westfall_Feltus_TCGA_COMPARE in Appendix A)

Gene Ontology Annotations to Find Candidate Surface Proteins and Corresponding Binding

Ligands – Ensemble gene IDs of transcripts categorized as highly expressed in GBM and

normal/low expressed in non-diseased brain tissue are uploaded to Uniprot87 where the gene

ontology (GO) annotations are listed for each gene. QuickGO88 is used to search for GO annotations

that identify surface proteins which can be potentially targeted with ligands. Gene IDs for the high

GBM/low non-diseased transcripts are filtered with the surface protein GO annotations to reveal

“candidate genes” as coding for proteins that are potential ligand binding sites for GBM drug

delivery. Candidate genes are cross-checked individually against the Gene Expression Profiling

Interactive Analysis (GEPIA) portal89 to verify that gene has high expression in GBM tumor

samples and normal/low expression in non-diseased brain tissue. To find corresponding binding

ligands, the candidate genes were searched for in the literature-informed online databases Binding

DB90 and RCSB Protein Data Bank91.

32


To find GBM surface proteins for active targeting and the ligands that can be used to target them,

it is necessary to leverage mRNA transcriptome data from the TCGA and GTEx. The workflow to

go from RNAseq data to GBM-specific ligand molecules incorporates a combination of Python

programming and online databases, including GEPIA, Uniprot, QuickGO, RCSB Protein Data

Bank, and Binding DB. The workflow shown in Figure 4.1 illustrates how Python isolates and

compares 73599 mRNA transcripts from TCGA and GTEx to identify genes that are highly

expressed in GBM tumor samples, but not in healthy, non-diseased brain tissues. Using this

method, depending on the source of the TCGA GEM (Feltus or Westfall), there are between 5400

and 7100 transcripts from candidate genes.

33

Figure 4.1 – Illustrative view of workflow identifying candidate genes

Genes identified in this way are uploaded to Uniprot, where groups of genes can be grouped

according to their Gene Ontology (GO) annotations. These annotations identify the many molecular

function, biological process, and cellular component attributes of the proteins these genes encode

for. GO annotations that can identify relevant cellular surface targets can be found using QuickGO.

Candidate genes found to code for the biologically relevant target proteins are then cross-checked

against the GEPIA online database, where expression of individual genes in tumor and healthy

tissue can be visualized based on RNAeq data. This is done as a second check to validate that the

initial python scripts correctly identified transcripts that were indeed abnormally expressed in GBM

tumor samples. These proteins for which candidate genes encode are then run on databases that

Top expressed transcripts from

TCGA GBM dataset

GO Annotations indicating a cellular

surface receptor capable of targetting

Lowest expressed transcripts from GTEX

dataset

34

have ligand binding information based on previous literature, most likely Binding DB or the RCSB

PDB. Figure 4.2 illustrates how gene expression is visualized in by boxplot in GEPIA (axis

representing transcripts per kilobase million (TPM+1)) and the identification of ligands with

binding affinity for the expressed protein.

Figure 4.2 – Illustrative view of candidate gene/protein GEPIA verification and ligand selection. (top) PLAUR (bottom) CXCR4. Boxplots are generated by GEPIA89. Axis is mRNA transcripts per kilobase million (log-2). GBM tumor samples are depicted in red, with healthy brain tissue

samples in gray.

Figure 4.2 also shows two candidate genes/proteins, Urokinase plasminogen activator surface

receptor (PLAUR) and C-X-C chemokine receptor type 4 (CXCR4). PLAUR is located in the

GEPIA RCSB PDB

GEPIA Binding

DB

35

plasma membrane and is shown to have high binding affinity to 2-Amino-5-phenyl-1-pyridin-2-

ylpyrrole-3-carbonitrile according to literature found using BindingDB92. CXCR4 is a

transmembrane protein that has a high binding affinity for (6,6-dimethyl-5,6-dihydroimidazo[2,1-

b][1,3]thiazol-3-yl)methyl N,N'-dicyclohexylimidothiocarbamate, per the literature found in RCSB

PDB93. According to data from GEPIA, transcripts for PLAUR and CXCR4 are found in

significantly higher amount in GBM tumor samples, when compared to healthy non-diseased brain

tissue93. These characteristics (location, binding affinity and mRNA expression) make PLAUR and

CXCR4 good candidates as part of a set of GBM surface targets for targeting with functionalized

polymersomes.


In this chapter, it has been shown that analysis of large RNAseq datasets can be combined with

extensive data from online transcriptomic and proteomic databases, revealing disease-specific

biologic targets for local drug delivery. This mRNA-to-ligand workflow is great for identifying

GBM targets but can easily be adopted to accommodate other cancers or diseases. It is not

unreasonable to image a workflow like this being adapted for patient- or tumor-derived RNAseq

data and used to develop personalized cancer-targeting drug carriers.

While promising, it is important understand the limitations of this analysis. While there is evidence

that intracellular mRNA levels are correlative with cellular protein levels94–97, there are also

researchers that caution that the correlation is poor and not properly understood98–100. They suggest

that mRNA to protein correlation may be affected by tissue specific factors and epigenetics.

However, as the methods for collecting transcriptomic and proteomic continue to improve, and

abundance of data available increases, it is likely that understanding of the mRNA and protein level

correlation will improve, giving strength to the type of analysis performed in this chapter.

36

Nevertheless, protein level measurements are crucial as additional confirmatory validation step in

future work.

Moving forward, this analysis must be expanded in two ways. First, it is important to find more

candidate genes for GBM. Finding these genes will likely be the result of identifying more GO

annotation combinations indicating a particular protein is a good biological candidate for targeting,

thus expanding the search scope through the highly expressed GBM genes. It is also important to

verify sets of candidate genes are not overexpressed exclusively in the same tumor samples. With

the current dataset of 174 samples, there is likely some overlap of overexpressed genes, but the

goal should be to identify a set of genes that minimizes correlative overexpression of the candidates.

Second, in vitro cellular uptake experiments can be performed on candidates as they are identified.

Cell lines can be transformed to overexpress the candidate protein on the cell surface. Drug or

fluorophore-loaded polymersomes will then be conjugated with the binding ligand associated with

that protein, as in Chapter 3. The expectation is that higher drug/fluorophore payloads will be

internalized by the transformed cells when the loaded polymersomes have been conjugated with

the targeting ligand, quantified by cell death or flow cytometry fluorescence.

37

5 – THE CHEMOTAXIS HYDROGEL

Introduction

In an earlier chapter, the concepts of drug-loaded and nanoparticle-loaded hydrogels were

introduced. Extensive research optimizing these systems, employing innovative materials and

informed by modern understanding of disease biology, has been performed over the last

decade37,38,40,54,56,101. However, the ultimate goal for these systems has remained the same: to

efficiently deliver therapeutics to a location/cell/tumor in the effort to treat or eliminate the disease.

However, modern technologies are allowing researchers to ponder quite the opposite: bringing the

diseased cells to the therapeutic location for elimination.

There are a few researchers currently creating and optimizing hydrogels that are great candidates

for cell migration: PEG based hydrogels, crosslinked by enzyme degradable peptides (EDP). These

PEG-EDP hydrogels mimic the extracellular matrix of the human body and are capable of

encapsulating cells and providing an environment for the cells to migrate. When the cell’s receptor

contacts the EDPs, mobility pathways in the cell are activated and the cell excretes the appropriate

degrading enzyme and migrates through the gel102–106.

This chapter focuses on taking the concepts of the PEG-EDP hydrogel and framing them as a tool

for local treatment of cancer. The hydrogels can act as a “black hole”, allowing invasive tumor

cells, particularly those left after tumor resection, to invade instead of migrating to the surrounding

tissue. The direction of the cell migration, that is toward the hydrogel, can be manipulated using a

chemoattractant gradient107,108 (Figure 5.1).

38

Figure 5.1 – Illustration demonstrating that eukaryotic cells migrate along a gradient toward a higher concentration of chemoattractants. Adapted from107.

Chemotaxis is the process by which that mammalian cells migrate toward the higher concentration

of chemical signals, which will in this case, be originating from the hydrogel.

Methods

Synthesis of Chemotaxis Hydrogel – 50 mg maleimide-functionalized 4-arm polyethylene glycol

(20KDa) (JENKEM Technology USA # A7029-1) (4armPEG) was dissolved in 0.5 mL milli-Q

water, varying concentrations of enzyme degradable peptide (Genscript #UI630DK260-1) (amino

acid sequence: CIPESLRAGC) (EDP) (20,40,100 mg/mL) was dissolved in 0.5 mL of milli-Q

water and added to 4armPEG solution and vortexed for 1 second. Mixture was quickly pipetted

into 12-well plate and allowed ~2 hours to gelate.

Determination of Hydrogel Stiffness – Stiffness of hydrogel was determined using Poroelastic

relaxation indentation (PRI). A custom-built apparatus for performing PRI was provided by Dr.

Eric Davis and operated by Nick Gregorich (Department of Chemical and Biomolecular

Engineering, Clemson University). A high-resolution linear actuator (M-230, physike intrumente)

was connected to a S-beam load cell (Futek LSB200, FUTEK Advanced Sensor Technology, Inc)

with a rigid glass indenter of radius of curvature R = 5.187mm. The actuator was mounted to a

39

high-performance linear stage with 46mm of travel range (M433, Newport). The indenter was

lowered by the actuator at a speed of 3 μm/s until contact with the hydrogel was reached. Upon

contact the indenter velocity changed to 10 μm/s until a load of 29.4mN was reached. The indenter

was held, fixed for 5 seconds, and released from the sample. A LabVIEW custom program, written

by Jaime Idarraga-Mora (Department of Chemical and Biomolecular Engineering, Clemson

University), was used to acquire load and indenter height data as a function of time from the S-

beam load cell during the experiments. Hooke’s law was used to calculate Young’s modulus from

collected data109.

Transwell Invasion Assay – Cell Culture – MDA-MB-231 cells and U-87 MG cells (ATCC

#HTB-26, ATCC #HTB-14, both verified by short tandem repeat profiling) were seeded in separate

T-75 flasks, at 2.0 x 106 cells per flask, and incubated for 36-48 hours at 37 °C and 5% CO2, in full

media Delbecco’s modified eagle medium (Gibco #10313-021) (DMEM +10% FBS +2mM L-

glutamine). Hydrogel inserts – 1 mL Chemotaxis hydrogel was synthesized and split into two

separate aliquots. 1 μg Doxorubicin-HCL (Fisher #BP2516-10) (DOX) was added to one of the

aliquots. For each cell line, 75 μL hydrogel was added to a 6.5 mm Transwell permeable support

with 80 μm pore (Costar #3422), and 75 μL hydrogel+DOX was added to another insert. Insert

hydrogels were allowed 2 hours to gelate. Each insert hydrogel was soaked overnight in starving

media (DMEM with no additives), or until hydrogel was visibly saturated. Cell migration – At

~70% confluence, MDAMB231 and U87 cells were lifted from T-75 flask with 0.25% Trypsin-

EDTA (Gibco #25200056), centrifuged, aspirated, then resuspended in starving media (DMEM

with no additives) to a concentration of 1.0 x 106 cells/mL. For each cell line, 100 μL (1.0 x 105

cells) were placed on top of a hydrogel inset, a hydrogel+DOX insert and a blank insert (directly

on membrane). Inserts were placed in 24-well plate containing 550 μL assay media (DMEM +0.5%

FBS +2mM L-glutamine +2ng/mL EGF). Plates, containing cells and inserts are incubated for 48

40

hours at 37 °C and 5% CO2. Plate wells were inspected under Revolve microscope (Echo

Laboratories) at 10x magnification to determine, qualitatively, whether cells had migrated through

the membrane and adhered to well surface.


The goal of the set of experiments in this chapter is to formulate a hydrogel from PEG and an

enzyme degradable peptide (EDP) that is degradable by one or more matrix metalloprotease. Also,

the hydrogel should have a stiffness similar to the tissue type it will be implanted into, which for

the purposes of treating GBM, is brain tissue. Furthermore, the hydrogel should facilitate cell

migration through degradation of the EDP, similar to the way a cell moves through the ECM.

Hydrogel formulation and gelation – This experiment was designed and performed by Chad Eaton

(Department of Bioengineering, Clemson University). A peptide sequence, CIPESLRAGC,

degradable by MMP-2105 (linker), was purchased and used to crosslink 4-Arm PEG functionalized

with terminal maleimides (4ARMPEG). Figure 5.2 illustrates the covalent bonding reaction

between the maleimide groups and the sulfhydryl pendant groups of the terminal cysteines of the

EDP.

Figure 5.2 – Crosslinking of hydrogel by covalent bonding of maleimide functionalized 4-Arm PEG to sulfhydryl pendant group

50 mg 4ARMPEG in 0.5 mL Milli-Q water is combined quickly with 20 mg linker in 0.5 mL Milli-

Q Water. The transparent hydrogel forms and gelates completely in ~2 hours.

41

Determining hydrogel stiffness – This experiment was designed and performed by Chad Eaton and

Riley Rapert (Department of Bioengineering, Clemson University) with assistance and equipment

provided by Nick Gregorich and Dr. Eric Davis (Department of Chemical and Biomolecular

Engineering, Clemson University). Now that a hydrogel can be formulated, it is important to

determine whether the hydrogel is a stiffness that is close to biological tissue, and whether that

stiffness can be modulated by formulation.

A set of hydrogels was synthesized by keeping the amount of 4ARMPEG constant and varying the

amount of linker. 50 mg 4ARMPEG in 50 mL Milli-Q water was combined with varying masses

of linker in 50 mL Milli-Q water and allowed to gelate for 2 hours (Table 5.1). Poroelastic

Relaxation Indentation (PRI) analysis was performed to determine Young’s modulus, a measure of

mechanical stiffness109,110. The following equation, a derivation of Hooke’s Law, was used:

𝐸𝐸 =(1 − 𝑣𝑣2)

𝜋𝜋∆(𝑃𝑃ℎ)∆𝑓𝑓

Where P is the load, h is the probe depth and r is the radius of contact of the spherical probe. An

assumption is made that the substance is isotropic, meaning that the stiffness will be the same in

the hydrogel regardless of direction. Thus, ν is assumed to be 0.5. Table 5.1 shows the Young’s

modulus of each formulation.

Table 5.1 – Chemotaxis Hydrogel Formulation and Young’s Modulus

Hydrogel 4ARMPEG (mg/mL H2O)

Linker (mg/mL H2O) Young’s modulus (kPa)

1 100 20 1.21 2 100 40 1.86 3 100 100 2.01

42

There is a positive correlation between the amount of linker used and the stiffness of the hydrogel.

Figure 5.3 shows the typical Young’s modulus for a variety of tissue types, according to their

Collagen-1 content111. GBM tumor varies from 0.5 - 2 kPa. This indicates that the hydrogels

formulated are similar in stiffness to GBM tumor. Also, it is shown that stiffness of the hydrogel

can be modulated by varying the amount of cross-linker available during gelation.

Figure 5.3 – Tissue stiffness for normal tissues. Highlighted in blue is a range of stiffnesses for GBM tumors. Adapted from111

Cell invasion assay – This experiment was designed and performed by Asia Paguntalan

(Department of Genetics and Biochemistry, Clemson University). The goal of this experiment is to

observe, and quantify, movement of cells into or through the hydrogel toward a chemotaxis signal.

43

A common way to quantify movement of cells toward chemoattractants is a transwell membrane

invasion assay. The assay was modified such that thin chemotaxis hydrogels would be formed

across the inserts above the membrane (Figure 5.4). 75 μL hydrogel (same formulation as hydrogel

#2 in table 5.1) was added and gelated on top of the membrane in the insert, then soaked in DMEM

with no additives. The receiver well was filled with DMEM +0.5% FBS +2mM L-glutamine

+2ng/mL Epidermal growth factor (EGF). The EGF is a chemoattractant for the invading

MDAMB231 breast cancer cells and the U87 glioblastoma cells. It is expected that the EGF would

diffuse through the hydrogel, creating a gradient, that would signal the cells to migrate toward the

receiver well through the hydrogel by degrading the EDP linker. Cells, in DMEM (with no

additives) were deposited on top of the hydrogel and incubated for 48 hours at 37 °C and 5% CO2.

Figure 5.4 – Illustration of transwell invasion assay modified for chemotaxis hydrogel

Unfortunately, in every trial, both U87 and MDAMB231 cells, microscope inspection, failed to

migrate through the hydrogel and membrane and into the receiver well. Upon further inspection,

44

the cells seemed to have remained above the hydrogel entirely. It may be unrealistic to assume

that in 48 hours, EGF would diffuse across the hydrogel and the cells would invade and

transverse through the hydrogel, which is approximately 2mm thick after soaking in media. The

assay can be modified such that the thickness of the gelated hydrogel in the insert is closer to that

of a thin film. Another modification to this experiment is to embed the cells in the hydrogel as it

gelates. Specifically, instead of dissolving 50mg of 4ARMPEG in 0.5 mL of Milli-Q water,

dissolve it in 0.5 mL of cell suspension in media. This approach would lower the time for

diffusing EGF to reach the cells, as they would already be dispersed in the hydrogel.


This chapter is an illustrative example of innovative materials design, generated by eschewing the

traditional “drug-loaded-hydrogel” delivery concept. These experiments are preliminary work

towards a goal of making a modular platform of disease-fighting chemotaxis hydrogels. The EDP

crosslinker chosen for this work can be interchanged with dozens of other EDPs, each specific to

mobility enzymes secreted by cells of varying tumor type. RNAseq analysis, described in the

previous chapter, can even reveal which chemotaxis signal receptors are expressed in a specific

tumor type, in turn, informing which chemoattractants would be most effective for the system.

Furthermore, this analysis shows that the chemotaxis hydrogel stiffness can be modulated through

formulation, allowing these hydrogels to be implanted in a variety of tissue types.

Moving forward, the most important focus will be observing and quantifying cell movement

through the hydrogel. Likely, the best way to accomplish this is thorough automated, 3D live-cell

imaging, where cells can be tracked in real-time. Hydrogels can be synthesized, using a cell

suspension in lieu of Milli-Q water, embedding the cells in the hydrogel, in multi-well plates. An

automated imaging system like the Cytation 5 and Biospa (BioTek) can be used to image the cells

45

in the hydrogels over several days, with the imaging software tracking selected cells. Furthermore,

chemotaxis signals can be injected at the edge of the well and allowed to diffuse, creating a gradient

across the hydrogel. This will show, qualitatively, whether migration direction of cells in the

hydrogel can be controlled by chemotaxis signaling.

46

APPENDIX A – PYTHON CODES

All Jupyter notebooks, containing these codes, are located on Birtwistle Lab Microsoft OneDrive (Folder: GBM Westfall) Dependent Files: All files are located on Birtwistle Lab Microsoft OneDrive (Folder: GBM Westfal) TCGAV1.txt TCGAIDtoESNID.txt GTEXV1.txt DATA (contains .txt files TCGA used for TCGA_analysis_Westfall) TCGA_analysis_FELTUS import pandas as pd df = pd.read_table("TCGAV1.txt", index_col=0) df.index.name = 'TCGA_ID' GBM={} for i in range(2016): if df.iloc[:,i].name.startswith('GBM'): GBM[df.iloc[:,i].name]=df.iloc[:,i] GBMData=pd.DataFrame(data=GBM) GBMData['Median'] = GBMData.median(axis=1) ensids = pd.read_table("TCGAIDtoENSID.txt", sep=' ', index_col=0) GBMData = GBMData.join(ensids, how='outer') GBMData.sort_values(by='Median', ascending=False, inplace=True) with_ensid = GBMData[GBMData['Ensembl_Gene_ID'].notnull()] largest_median_with_ensid = with_ensid[['Median','Ensembl_Gene_ID']][:12000] largest_median_with_ensid['Ensembl_Gene_ID'].to_csv(r'C:\Users\jjwes\Documents\Python Workspace\GBM\topTCGA.txt',index=0) largest_median_with_ensid.to_csv(r'C:\Users\jjwes\Documents\Python Workspace\GBM\TCGA_high.txt') TCGA_analysis_Westfall

47

import pandas as pd import numpy as np import glob files = glob.glob('**/*.FPKM.txt', recursive=True) base = {} base_frame = pd.DataFrame(base) for i in range(len(files)): base_frame[files[i]] = pd.read_table(files[i],index_col=0, names=['FPKM'])['FPKM'] FPKM=base_frame rank_mean = FPKM.stack().groupby(FPKM.rank(method='first').stack().astype(int)).mean() FPKM_Qnormal=FPKM.rank(method='min').stack().astype(int).map(rank_mean).unstack() FPKM_Qnormal_Nan = FPKM_Qnormal.replace(0, np.nan) Log2_Qnormal_Nan = np.log2(FPKM_Qnormal_Nan) Log2_Qnormal=Log2_Qnormal_Nan.replace(np.nan,0) Log2_Qnormal.index.name = 'Ensembl_Gene_ID' Log2_Qnormal['Median'] = Log2_Qnormal.median(axis=1) Log2_Qnormal.sort_values(by='Median', ascending=False, inplace=True) Log2_Qnormal.index=Log2_Qnormal.index.str[:15] largest_median_TCGA_WESTFALL = Log2_Qnormal[['Median']][:12000] largest_median_TCGA_WESTFALL.to_csv(r'C:\Users\jjwes\Documents\Python Workspace\GBM\TCGA_high_WESTFALL.txt') GTEX_analysis import pandas as pd GTEXData = pd.read_table("GTEXV1.txt") GTEXData['Median'] = GTEXData.median(axis=1) GTEXData.sort_values(by='Median', ascending=True, inplace=True)

48

GTEXData.index=GTEXData.index.str[:15] GTEX_smallest_median = GTEXData['Median'][:50000] GTEX_smallest_median.to_csv(r'C:\Users\jjwes\Documents\Python Workspace\GBM\GTEX_low.txt') TCGA_GTEX_Comparison_FELTUS import pandas as pd import numpy as np TCGA_high = pd.read_table('TCGA_high.txt', sep=',',index_col=0) GTEX_low = pd.read_table('GTEX_low.txt',sep=',',names=['Ensembl_Gene_ID','Median']) p=0 for i in range(12000): if GTEX_low['Ensembl_Gene_ID'].str.contains(TCGA_high['Ensembl_Gene_ID'].iloc[i]).any(): print (TCGA_high['Ensembl_Gene_ID'].iloc[i]) p=p+1 p TCGA_GTEX_Comparison_Westfall import pandas as pd import numpy as np TCGA_high_WESTFALL = pd.read_table('TCGA_high_WESTFALL.txt', sep=',',index_col=0) GTEX_low = pd.read_table('GTEX_low.txt',sep=',',names=['Ensembl_Gene_ID','Median']) p=0 for i in range(12000): if GTEX_low['Ensembl_Gene_ID'].str.contains(TCGA_high_WESTFALL.index[i]).any(): print (TCGA_high_WESTFALL.index[i]) p=p+1 p Westfall_Feltus_TCGA_COMPARE TCGA_high_F = pd.read_table('TCGA_high.txt', sep=',',index_col=0) TCGA_high_W = pd.read_table('TCGA_high_WESTFALL.txt', sep=',',index_col=0)

49

p=0 for i in range(12000): if TCGA_high_W.index.str.contains(TCGA_high_F['Ensembl_Gene_ID'].iloc[i]).any(): print (TCGA_high_F['Ensembl_Gene_ID'].iloc[i]) p=p+1 p

50

REFERENCES

1. Koshy, M. et al. Improved survival time trends for glioblastoma using the SEER 17

population-based registries. J. Neurooncol. 107, 207–12 (2012).

2. Berens, M. E. & Giese, A. “...those left behind.” Biology and Oncology of Invasive

Glioma Cells. Neoplasia 1, 208–219 (1999).

3. Yang, L.-J., Zhou, C.-F. & Lin, Z.-X. Temozolomide and Radiotherapy for Newly

Diagnosed Glioblastoma Multiforme: A Systematic Review. Cancer Invest. 32, 31–36

(2014).

4. Smoll, N. R., Schaller, K. & Gautschi, O. P. Long-term survival of patients with

glioblastoma multiforme (GBM). J. Clin. Neurosci. 20, 670–675 (2013).

5. Hottinger, A. F., Abdullah, K. G. & Stupp, R. Current Standards of Care in Glioblastoma

Therapy. in Glioblastoma 73–80 (Codon Publications, 2016). doi:10.1016/B978-0-323-

47660-7.00006-9

6. Darlix, A. et al. Prolonged administration of adjuvant temozolomide improves survival in

adult patients with glioblastoma. Anticancer Res. 33, 3467–74 (2013).

7. Gallego, O. Nonsurgical treatment of recurrent glioblastoma. Curr. Oncol. 22, e273-81

(2015).

8. Johnson, D. R. & O’Neill, B. P. Glioblastoma survival in the United States before and

during the temozolomide era. J. Neurooncol. 107, 359–364 (2012).

9. Tiek, D. M. et al. Alterations in Cell Motility, Proliferation, and Metabolism in Novel

Models of Acquired Temozolomide Resistant Glioblastoma. Sci. Rep. 8, (2018).

51

10. Darefsky, A. S., King, J. T. & Dubrow, R. Adult glioblastoma multiforme survival in the

temozolomide era: A population-based analysis of Surveillance, Epidemiology, and End

Results registries. Cancer 118, 2163–2172 (2012).

11. Wolbers, J. G. Novel strategies in glioblastoma surgery aim at safe, supra-maximum

resection in conjunction with local therapies. Chin. J. Cancer 33, 8–15 (2014).

12. Disease State. Available at: http://gliadel.com/hcp/disease-state.php. (Accessed: 16th

August 2019)

13. VandenDriessche, T. & Chuah, M. K. Glioblastoma: bridging the gap with gene therapy.

Lancet. Oncol. 14, 789–90 (2013).

14. Hu, S. et al. Real-Time Imaging of Brain Tumor for Image-Guided Surgery. Adv. Healthc.

Mater. 7, e1800066 (2018).

15. Lara-Velazquez, M. et al. Advances in Brain Tumor Surgery for Glioblastoma in Adults.

Brain Sci. 7, (2017).

16. von Roemeling, C., Jiang, W., Chan, C. K., Weissman, I. L. & Kim, B. Y. S. Breaking

Down the Barriers to Precision Cancer Nanomedicine. Trends in Biotechnology 35, 159–

171 (2017).

17. Dobbing, J. The Blood‐Brain Barrier. Developmental Medicine & Child Neurology 3,

311–314 (1961).

18. Abbott, N. J., Rönnbäck, L. & Hansson, E. Astrocyte-endothelial interactions at the blood-

brain barrier. Nature Reviews Neuroscience 7, 41–53 (2006).

19. Pardridge, W. M. Drug transport across the blood-brain barrier. J. Cereb. Blood Flow

52

Metab. 32, 1959–1972 (2012).

20. Wolak, D. J. & Thorne, R. G. Diffusion of macromolecules in the brain: Implications for

drug delivery. Mol. Pharm. 10, 1492–1504 (2013).

21. Peer, D. et al. Nanocarriers as an emerging platform for cancer therapy. Nat. Nanotechnol.

2, 751–760 (2007).

22. Niewald, M. et al. Toxicity after radiochemotherapy for glioblastoma using temozolomide

- a retrospective evaluation. Radiat. Oncol. 6, 141 (2011).

23. Saito, K. et al. Toxicity and outcome of radiotherapy with concomitant and adjuvant

temozolomide in elderly patients with glioblastoma: a retrospective study. Neurol. Med.

Chir. (Tokyo). 54, 272–9 (2014).

24. Hanna, N. et al. Long-Term Toxicity of Chemotherapy. (2003).

25. Zhang, J., F.G. Stevens, M. & D. Bradshaw, T. Temozolomide: Mechanisms of Action,

Repair and Resistance. Curr. Mol. Pharmacol. 5, 102–114 (2011).

26. D’Alessandris, Q. G., Montano, N., Larocca, L. M., Maira, G. & Pallini, R. Prognostic

impact of MGMT promoter methylation in glioblastoma - A systematic review. J. Cancer

Sci. Ther. 6, 136–141 (2014).

27. Patel, A. P. et al. Single-cell RNA-seq highlights intratumoral heterogeneity in primary

glioblastoma. Science (80-. ). 344, 1396–1401 (2014).

28. Qazi, M. A. et al. Intratumoral heterogeneity: pathways to treatment resistance and relapse

in human glioblastoma. Annals of oncology : official journal of the European Society for

Medical Oncology 28, 1448–1456 (2017).

53

29. Eskilsson, E. et al. EGFR heterogeneity and implications for therapeutic intervention in

glioblastoma. Neuro. Oncol. 20, 743–752 (2018).

30. Iwadate, Y. Epithelial-mesenchymal transition in glioblastoma progression. Oncology

Letters 11, 1615–1620 (2016).

31. Sottoriva, A. et al. Intratumor heterogeneity in human glioblastoma reflects cancer

evolutionary dynamics. Proc Natl Acad Sci U S A 110, 4009–4014 (2013).

32. Peppas, N. A., Bures, P., Leobandung, W. & Ichikawa, H. Hydrogels in pharmaceutical

formulations. Eur. J. Pharm. Biopharm. 50, 27–46 (2000).

33. Wagner, A. M., Spencer, D. S. & Peppas, N. A. Advanced architectures in the design of

responsive polymers for cancer nanomedicine. J. Appl. Polym. Sci. 135, 46154 (2018).

34. Cheng, W., Gu, L., Ren, W. & Liu, Y. Stimuli-responsive polymers for anti-cancer drug

delivery. Mater. Sci. Eng. C 45, 600–608 (2015).

35. Wolinsky, J. B., Colson, Y. L. & Grinstaff, M. W. Local drug delivery strategies for

cancer treatment: Gels, nanoparticles, polymeric films, rods, and wafers. Journal of

Controlled Release 159, 14–26 (2012).

36. Koetting, M. C., Peters, J. T., Steichen, S. D. & Peppas, N. A. Stimulus-responsive

hydrogels: Theory, modern advances, and applications. Mater. Sci. Eng. R Reports 93, 1–

49 (2015).

37. Bastiancich, C., Danhier, P., Préat, V. & Danhier, F. Anticancer drug-loaded hydrogels as

drug delivery systems for the local treatment of glioblastoma. Journal of Controlled

Release 243, 29–42 (2016).

54

38. Bastiancich, C. et al. Injectable nanomedicine hydrogel for local chemotherapy of

glioblastoma after surgical resection. J. Control. Release 264, 45–54 (2017).

39. Huang, H., Qi, X., Chen, Y. & Wu, Z. Thermo-sensitive hydrogels for delivering

biotherapeutic molecules: A review. Saudi Pharm. J. (2019).

doi:10.1016/j.jsps.2019.08.001

40. Basso, J. et al. Hydrogel-Based Drug Delivery Nanosystems for the Treatment of Brain

Tumors. Gels 4, 62 (2018).

41. Couvreur, P. Nanoparticles in drug delivery: Past, present and future. Advanced Drug

Delivery Reviews 65, 21–23 (2013).

42. Batrakova, E. V., Bronich, T. K., Vetro, J. A. & Kabanov, A. V. Polymer micelles as drug

carriers. Dep. Pharm. Sci. 57–93 (2006). doi:10.1142/9781860949074_0005

43. Cho, H. K., Cheong, I. W., Lee, J. M. & Kim, J. H. Polymeric nanoparticles, micelles and

polymersomes from amphiphilic block copolymer. Korean Journal of Chemical

Engineering 27, 731–740 (2010).

44. Yu, X. et al. Design of Nanoparticle-Based Carriers for Targeted Drug Delivery. Journal

of Nanomaterials 2016, (2016).

45. Letchford, K. & Burt, H. A review of the formation and classification of amphiphilic

block copolymer nanoparticulate structures: micelles, nanospheres, nanocapsules and

polymersomes. Eur. J. Pharm. Biopharm. 65, 259–269 (2007).

46. Meng, F., Zhong, Z. & Feijen, J. Stimuli-responsive polymersomes for programmed drug

delivery. Biomacromolecules 10, 197–209 (2009).

55

47. Kamaly, N., Yameen, B., Wu, J. & Farokhzad, O. C. Degradable controlled-release

polymers and polymeric nanoparticles: Mechanisms of controlling drug release. Chem.

Rev. 116, 2602–2663 (2016).

48. Shin, T. H. & Cheon, J. Synergism of nanomaterials with physical stimuli for biology and

medicine. Accounts of Chemical Research 50, 567–572 (2017).

49. Felber, A. E., Dufresne, M. H. & Leroux, J. C. PH-sensitive vesicles, polymeric micelles,

and nanospheres prepared with polycarboxylates. Advanced Drug Delivery Reviews 64,

979–992 (2012).

50. Jang, W. S. et al. Enzymatically triggered rupture of polymersomes. Soft Matter 12, 1014–

1020 (2016).

51. Steichen, S. D., Caldorera-Moore, M. & Peppas, N. A. A review of current nanoparticle

and targeting moieties for the delivery of cancer therapeutics. European Journal of

Pharmaceutical Sciences 48, 416–427 (2013).

52. Jiang, Y., Zhang, J., Meng, F. & Zhong, Z. Apolipoprotein e Peptide-Directed Chimeric

Polymersomes Mediate an Ultrahigh-Efficiency Targeted Protein Therapy for

Glioblastoma. ACS Nano 12, 11070–11079 (2018).

53. Raucher, D., Dragojevic, S. & Ryu, J. Macromolecular drug carriers for targeted

glioblastoma therapy: Preclinical studies, challenges, and future perspectives. Frontiers in

Oncology 8, 624 (2018).

54. Arai, T. et al. Novel local drug delivery system using thermoreversible gel in combination

with polymeric microspheres or liposomes. Anticancer Res. 30, 1057–1064 (2010).

55. Jin, S. et al. Redox/pH stimuli-responsive biodegradable PEGylated P(MAA/BACy)

56

nanohydrogels for controlled releasing of anticancer drugs. Colloids Surfaces A

Physicochem. Eng. Asp. 484, 47–55 (2015).

56. Pitorre, M. et al. Recent advances in nanocarrier-loaded gels: Which drug delivery

technologies against which diseases? Journal of Controlled Release 266, 140–155 (2017).

57. Binabaj, M. M. et al. The prognostic value of MGMT promoter methylation in

glioblastoma: A meta-analysis of clinical trials. J. Cell. Physiol. 233, 378–386 (2018).

58. Zhang, K., Wang, X. Q., Zhou, B. & Zhang, L. The prognostic value of MGMT promoter

methylation in Glioblastoma multiforme: A meta-analysis. Familial Cancer 12, 449–458

(2013).

59. Verhaak, R. G. W. et al. Integrated Genomic Analysis Identifies Clinically Relevant

Subtypes of Glioblastoma Characterized by Abnormalities in PDGFRA, IDH1, EGFR,

and NF1. Cancer Cell 17, 98–110 (2010).

60. Yang, Y. et al. Databases and web tools for cancer genomics study. Genomics, Proteomics

and Bioinformatics 13, 46–50 (2015).

61. Discher, D. E. & Ahmed, F. Polymersomes. Annu. Rev. Biomed. Eng. 8, 323–41 (2006).

62. Lee, J. S. & Feijen, J. Polymersomes for drug delivery: Design, formation and

characterization. Journal of Controlled Release 161, 473–483 (2012).

63. Rideau, E., Dimova, R., Schwille, P., Wurm, F. R. & Landfester, K. Liposomes and

polymersomes: a comparative review towards cell mimicking. Chem. Soc. Rev. 47, 8572–

8610 (2018).

64. Jiang, Y., Yang, W., Zhang, J., Meng, F. & Zhong, Z. Protein Toxin Chaperoned by LRP-

57

1-Targeted Virus-Mimicking Vesicles Induces High-Efficiency Glioblastoma Therapy In

Vivo. Adv. Mater. 30, 1800316 (2018).

65. Luo, Z. et al. On-Demand Drug Release from Dual-Targeting Small Nanoparticles

Triggered by High-Intensity Focused Ultrasound Enhanced Glioblastoma-Targeting

Therapy. ACS Appl. Mater. Interfaces 9, 31612–31625 (2017).

66. Ahmed, F. & Discher, D. E. Self-porating polymersomes of PEG-PLA and PEG-PCL:

Hydrolysis-triggered controlled release vesicles. J. Control. Release 96, 37–53 (2004).

67. Sanchez, L., Yi, Y. & Yu, Y. Effect of partial PEGylation on particle uptake by

macrophages. Nanoscale 9, 288–297 (2017).

68. Galloway, D. A., Phillips, A. E. M., Owen, D. R. J. & Moore, C. S. Phagocytosis in the

Brain: Homeostasis and Disease. Front. Immunol. 10, 790 (2019).

69. Tutorials, E. The Zeta Potential. Building 1–4 (1999).

70. Kelly, J. M., Gross, A. L., Martin, D. R. & Byrne, M. E. Polyethylene glycol-b-poly(lactic

acid) polymersomes as vehicles for enzyme replacement therapy. Nanomedicine 12,

2591–2606 (2017).

71. Bhattacharjee, S. DLS and zeta potential - What they are and what they are not? Journal

of Controlled Release 235, 337–351 (2016).

72. Kelly, J. M., Pearce, E. E., Martin, D. R. & Byrne, M. E. Lyoprotectants modify and

stabilize self-assembly of polymersomes. Polymer (Guildf). 87, 316–322 (2016).

73. Ayen, W. Y. & Kumar, N. A systematic study on lyophilization process of polymersomes

for long-term storage using doxorubicin-loaded (PEG) 3 -PLA nanopolymersomes. Eur. J.

58

Pharm. Sci. 46, 405–414 (2012).

74. Li, M., Rouaud, O. & Poncelet, D. Microencapsulation by solvent evaporation: State of

the art for process engineering approaches. International Journal of Pharmaceutics 363,

26–39 (2008).

75. Khalil, N. M. et al. Pharmacokinetics of curcumin-loaded PLGA and PLGA-PEG blend

nanoparticles after oral administration in rats. Colloids Surfaces B Biointerfaces 101, 353–

360 (2013).

76. Lee, Y., Chang, J. B., Kim, H. K. & Park, T. G. Stability studies of biodegradable

polymersomes prepared by emulsion solvent evaporation method. Macromol. Res. 14,

359–364 (2006).

77. Iqbal, M., Zafar, N., Fessi, H. & Elaissari, A. Double emulsion solvent evaporation

techniques used for drug encapsulation. International Journal of Pharmaceutics 496, 173–

190 (2015).

78. Bazak, R., Houri, M., El Achy, S., Kamel, S. & Refaat, T. Cancer active targeting by

nanoparticles: a comprehensive review of literature. J. Cancer Res. Clin. Oncol. 141, 769–

784 (2015).

79. Neves, A. R., Queiroz, J. F., Lima, S. A. C. & Reis, S. Apo E-Functionalization of Solid

Lipid Nanoparticles Enhances Brain Drug Delivery: Uptake Mechanism and Transport

Pathways. Bioconjug. Chem. 28, 995–1004 (2017).

80. Liu, Y., Xu, C. F., Iqbal, S., Yang, X. Z. & Wang, J. Responsive Nanocarriers as an

Emerging Platform for Cascaded Delivery of Nucleic Acids to Cancer. Adv. Drug Deliv.

Rev. 115, 98–114 (2017).

59

81. De Vleeschouwer, S. Glioblastoma. (Codon Publications, 2017).

doi:http://dx.doi.org/10.15586/codon.glioblastoma.2017 Edited

82. Iwadate, Y. Epithelial-mesenchymal transition in glioblastoma progression. Oncol. Lett.

11, 1615–1620 (2016).

83. Wang, Q. et al. Tumor Evolution of Glioma-Intrinsic Gene Expression Subtypes

Associates with Immunological Changes in the Microenvironment. Cancer Cell 32, 42-

56.e6 (2017).

84. Ceccarelli, M. et al. Molecular Profiling Reveals Biologically Discrete Subsets and

Pathways of Progression in Diffuse Glioma. Cell 164, 550–563 (2016).

85. Roche, K. E., Weinstein, M., Dunwoodie, L. J., Poehlman, W. L. & Feltus, F. A. Sorting

Five Human Tumor Types Reveals Specific Biomarkers and Background Classification

Genes. Sci. Rep. 8, 8180 (2018).

86. Carithers, L. J. et al. A Novel Approach to High-Quality Postmortem Tissue Procurement:

The GTEx Project. Biopreserv. Biobank. 13, 311–319 (2015).

87. Bateman, A. UniProt: A worldwide hub of protein knowledge. Nucleic Acids Res. 47,

D506–D515 (2019).

88. Binns, D. et al. QuickGO: A web-based tool for Gene Ontology searching. Bioinformatics

25, 3045–3046 (2009).

89. Tang, Z. et al. GEPIA: A web server for cancer and normal gene expression profiling and

interactive analyses. Nucleic Acids Res. 45, W98–W102 (2017).

90. Gilson, M. K. et al. BindingDB in 2015: A public database for medicinal chemistry,

60

computational chemistry and systems pharmacology. Nucleic Acids Res. 44, D1045–

D1053 (2016).

91. Berman, H. M. et al. The protein data bank. Acta Crystallogr. Sect. D Biol. Crystallogr.

58, 899–907 (2002).

92. Omar, M. A. et al. Synthesis and docking studies of novel antitumor benzimidazoles.

Bioorganic Med. Chem. 20, 6989–7001 (2012).

93. Wu, B. et al. Structures of the CXCR4 chemokine GPCR with small-molecule and cyclic

peptide antagonists. Science 330, 1066–1071 (2010).

94. Liu, Y., Beyer, A. & Aebersold, R. On the Dependency of Cellular Protein Levels on

mRNA Abundance. Cell 165, 535–550 (2016).

95. Greenbaum, D., Colangelo, C., Williams, K. & Gerstein, M. Comparing protein

abundance and mRNA expression levels on a genomic scale. Genome Biol. 4, 117 (2003).

96. Edfors, F. et al. Gene‐specific correlation of RNA and protein levels in human cells and

tissues . Mol. Syst. Biol. 12, 883 (2016).

97. Vogel, C. & Marcotte, E. M. Insights into the regulation of protein abundance from

proteomic and transcriptomic analyses. Nat. Rev. Genet. 13, 227–232 (2012).

98. Nancy Kendrick. A gene’s mRNA level does not usually predict its protein leve. Kendrick

labs.Inc (2014).

99. Perl, K. et al. Reduced changes in protein compared to mRNA levels across non-

proliferating tissues. BMC Genomics 18, 305 (2017).

100. Pascal, L. E. et al. Correlation of mRNA and protein levels: Cell type-specific gene

61

expression of cluster designation antigens in the prostate. BMC Genomics 9, 246 (2008).

101. Shen, M., Xu, Y. Y., Sun, Y., Han, B. S. & Duan, Y. R. Preparation of a Thermosensitive

Gel Composed of a mPEG-PLGA-PLL-cRGD Nanodrug Delivery System for Pancreatic

Tumor Therapy. ACS Appl. Mater. Interfaces 7, 20530–20537 (2015).

102. Kyburz, K. A. & Anseth, K. S. Synthetic Mimics of the Extracellular Matrix: How Simple

is Complex Enough? Ann. Biomed. Eng. 43, 489–500 (2015).

103. Tibbitt, M. W. & Anseth, K. S. Dynamic microenvironments: The fourth dimension.

Science Translational Medicine 4, (2012).

104. Singh, S. P. et al. A synthetic modular approach for modeling the role of the 3D

microenvironment in tumor progression. Sci. Rep. 5, (2015).

105. Jansen, L. E., Negrón-Piñeiro, L. J., Galarza, S. & Peyton, S. R. Control of thiol-

maleimide reaction kinetics in PEG hydrogel networks. Acta Biomater. 70, 120–128

(2018).

106. Deforest, C. A., Polizzotti, B. D. & Anseth, K. S. Sequential click reactions for

synthesizing and patterning three-dimensional cell microenvironments. Nat. Mater. 8,

659–664 (2009).

107. Wang, Y., Chen, C. L. & Iijima, M. Signaling mechanisms for chemotaxis. Development

Growth and Differentiation 53, 495–502 (2011).

108. Jin, T. Gradient sensing during chemotaxis. Curr. Opin. Cell Biol. 25, 532–537 (2013).

109. Sakai, M. & Shimizu, S. Indentation rheometry for glass-forming materials. J. Non. Cryst.

Solids 282, 236–247 (2001).

62

110. Nadermann, N. K., Davis, E. M., Page, K. A., Stafford, C. M. & Chan, E. P. Using

Indentation to Quantify Transport Properties of Nanophase-Segregated Polymer Thin

Films. Adv. Mater. 27, 4924–4930 (2015).

111. Pfeifer, C. R., Alvey, C. M., Irianto, J. & Discher, D. E. Genome variation across cancers

scales with tissue stiffness - An invasion-mutation mechanism and implications for

immune cell infiltration. Current Opinion in Systems Biology 2, 103–114 (2017).