Master's Thesis Research. By, Megha Suresh. Advisor: Mansoor M. Amiji, PhD

Transcription

1 In Vitro Model of Inflammatory Signaling and Cancer Stem Cell Signaling in Pancreatic Cancer using 3D Hetero-Cellular Spheroids and MicroRNA Delivery using Hyaluronic Acid-Based Nanoparticles Master's Thesis Research By, Megha Suresh Advisor: Mansoor M. Amiji, PhD Department of Pharmaceutical Sciences Northeastern University April,

2 ACKNOWLEDGEMENTS First and foremost, I would like to extend my deepest gratitude to my professor, Dr. Mansoor Amiji, for giving me this wonderful opportunity to learn from him and for being my biggest inspiration. He has been a constant motivator and guide, and has always been supportive of my professional goals both inside and outside of the lab. I am especially thankful to him for giving this project his time, attention, and for providing valuable comments. The last three years of working for him have been immensely gratifying and without his help, this thesis would not have been possible. Next, I would like to express my appreciation for the consistent support I have received from both my committee members, Dr. Samuel Gatley and Dr. Shanthi Ganesh. This project is an extension of Dr. Ganesh s work and I could not have asked for a better guide, inspiring me always. I want to specially thank Dr. Gatley for his unending belief in me, for his time and for always supporting me. I would like to thank my mentor, Dr. Arijit Chakravarty, whose timely guidance and perpetual support were crucial in expanding this project. I am thankful to him for being instrumental in strengthening my progress and driving me to push my boundaries. I am extremely grateful to members of Dr. Amiji s lab, especially to Dr. George Matthaiolampakis, who has helped me in every way without hesitation. I want to thank Amit, Srujan, Vanessa, Ed, Bill and Adwait for training me to get things started and always responding when I needed help. This note would be incomplete without mentioning Dr. Joelle Carlo, who has provided her expertise and guidance with numerous experiments. I also want to thank Huyen, Brijesh, Mei-Ju, Ruchi, Lara, Meghna, George Balaconis and Malav for their help. I want to thank the department members who have helped me as well. None of this would have been possible without the encouragement and support I received from Sarom, Nancy and Rosalee. Nearing the completion of my project, I was extremely lucky to have the most energetic, dedicated and hard-working masters students, Krina, Hima, Adhithi and Jing help me with my project and extending their timely assistance above and beyond the call of duty. It brings me great joy to see their enthusiasm and I wish them all the best with their projects. My friends who have provided their support through the last three years deserve a special mention. I would like to firstly thank my dearest friend, Abhijit Kulkarni, for always pushing me to excel and for being a relentless believer, and for his professional help with the NMRs. I would also like to thank my beloved friends, Bhagyashree Raut, Vaidehi Wani, Minah Iqbal, Sheena Singh and Mahek Shringhey for being there through it all. 2

3 None of this would have been possible without the limitless love of my family. My parents raised me to never be complacent and always strive for excellence in whatever I do and without them, I would not be here. My final appreciation is extended to Northeastern University, Bouvé college of Health Sciences, Department of Pharmaceutical Sciences and everyone who was a part of this project. Thank you for believing. 3

4 TABLE OF CONTENTS 1. SUMMARY OBJECTIVES AND SPECIFIC AIMS Statement of the Problem Objectives and Experimental Hypothesis Specific Aims BACKGROUND AND SIGNIFICANCE Pancreatic Cancer Incidence and Severity Current Therapy Inflammation and Pancreatic Cancer Role of Inflammatory Markers TGF-beta as an Anti-Inflammatory Cytokine TNF alpha as a Pro-Inflammatory Cytokine Interleukin 8 as a Pro-Inflammatory Cytokine Cancer Stem Cells Hypoxia and its Role in Maintenance of Cancer Stem Cells Inflammation as a Regulator of Cancer Stem Cells MicroRNAs Introduction to MicroRNAs Biogenesis and Function of MicroRNAs MicroRNAs in Cancer MicroRNA-34 and Cancer MiR34 and the Notch pathway 34 4

5 MiR34 and Wnt Signaling Pathway MicroRNA Therapy Three-dimensional Tumor Spheroids Hyaluronic acid-based Nanoparticle Formulations EXPERIMENTAL DESIGN AND METHODS Establishment of Homo-cellular and Hetero-cellular, 3-dimensional Tumor Spheroids a. Cell Culture of Panc1 (pancreatic ductal epithelioid carcinoma cell line), NIH/3T3 (murine fibroblast cell line) and J774 (murine macrophage cell line) b. Development of Homo-cellular and Hetero-cellular Spheroids by the Hanging Drop Method c. Confocal Size Measurements of Spheroids d. Distribution of Panc1, J774.A1 & NIH3T3 Cells within Hetero-cellular Spheroids Analysis of Inflammatory, Hypoxic and Cancer Stem Cell Phenotype of Spheroids through PCR a. Total RNA Extraction from Homo- and Hetero-cellular Spheroids b. cdna Synthesis from RNA Extracted from Day 5 Homo- and hetero-cellular spheroids c. RT-PCR of cdna from Day 5 Homo- and Hetero-cellular Spheroids d. Gel Electrophoresis to Observe Expression Levels of the Inflammatory Cytokines IL8, TNF-α, TGF-β in the Different Spheroid Models e. Analysis of LDH-A Expression in Day 5 Homo-cellular Panc1 Spheroids f. Selection of Housekeeping Gene g. qpcr to quantitatively assess levels of HIF-1α, HIF-2α, SCF and LDH-A levels in spheroids Analysis of CD24+ and Stem Cell Factor Expression in Day 5 Homo-cellular and Hetero-cellular Spheroids through Immunofluorescence

6 4.4. Synthesis and characterization of various polymer conjugates a. Synthesis of hyaluronic acid polyethyleneimine conjugate b. Synthesis of hyaluronic acid polyethylene glycol conjugate c. Synthesis of hyaluronic acid polyethyleneimine Cyanine 5 dye conjugate d. Nuclear magnetic resonance (NMR) to confirm conjugation of polymers e. Evaluation of the efficiency of cy5-ha-pei conjugation Preparation of HA-PEI: HA-PEG nanosystems a. Preparation of loaded HA-PEI: HA-PEG : microrna 34a mimic nanoparticles b. Preparation of blank HA-PEI: HA-PEG nanoparticles c. Preparation of loaded cy5-ha-pei: HA-PEG : microrna 34a mimic nanoparticles d. Preparation of loaded HA-PEI: microrna 34a mimic nanoparticles e. Preparation of blank HA-PEI nanoparticles Characterization of HA-PEI: HA-PEG Nanosystems a. Determination of size and charge of HA-PEI: HA-PEG nanosystems through Dynamic Light Scattering b. Morphological analysis of HA-PEI: HA-PEG nanosystems through transmission electron microscopy c. Determination of encapsulation of microrna within the HA-PEI: HA-PEG nanosystems d. Determination of release of microrna from the HA-PEI: HA-PEG nanosystems Uptake Study a. Uptake in cells grown in normoxic conditions b. Uptake in homo-cellular and hetero-cellular spheroids Transfection study

7 4.8.a. Transfection in cells grown in normoxic conditions b. Transfection in homo-cellular and hetero-cellular spheroids RESULTS Characterization of homo- and hetero-cellular spheroids a. Determination of spheroid diameter and thickness b. Arrangement of cell lines within Hetero-cellular Spheroids Analysis of Inflammatory, Hypoxic and Cancer Stem Cell Phenotype of Spheroids through PCR: a. Analysis of IL8, TNF-α and TGF-β gene levels across spheroid models using RT-PCR and gel electrophoresis b. Analysis of LDH-A expression in homo-cellular and hetero-cellular spheroids c. Comparison of housekeeping genes: Beta actin and 28srRNA d. qpcr to quantitatively assess HIF-1α, HIF-2α, SCF and LDH-A levels in spheroids Analysis of CD24+ and Stem Cell Factor Expression in Day 5 Homo-cellular and Heterocellular Spheroids through Immunofluorescence Evaluation of synthesized polymers a. Nuclear magnetic resonance (NMR) to evaluate conjugation of polymers b. Evaluation of the efficiency of cy5-ha-pei conjugation Evaluation of size and charge of HA-PEI: HA-PEG nanosystems through Dynamic Light Scattering a. Determination of size of loaded and blank HA-PEI: HA-PEG nanoparticles b. Determination of surface charge of loaded and blank HA-PEI: HA-PEG c. Analysis of nanoparticle morphology using TEM d. Analysis of encapsulation and release in the HA-based nanosystems d.a. Gel retardation assay to determine release of microrna from the HA-PEI: HA- PEG nanosystems

8 5.5.d.b. Evaluation of encapsulation of microrna from HA- based nanosystems Uptake study a. Uptake in normoxic cells b Uptake in spheroid models b.i. Uptake in Panc1 homo-cellular spheroids b.ii. Uptake in Panc1:J774.A1 hetero-cellular spheroids b.iii. Uptake in Panc1: NIH/3T3 hetero-cellular spheroids b.iv. Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids Transfection Study b.i. Transfection in 3 in 1 hetero-cellular spheroids b.ii. Transfection in normoxic Panc1 cells CONCLUSION REFERENCES

9 LIST OF FIGURES Figure 1: Schematic illustrating cancer stem cell initiation and their response to chemotherapeutic agents [1]..28 Figure 2: Signaling pathways involved with CSCs and EMT [2] 30 Figure 3: The biogenesis and function of mirna [3]..32 Figure 4: HA conjugated to 20 kda PEG [4]..38 Figure 5: HA conjugated to 10 kda PEI [4]...38 Figure 6: Hanging Drop Plates. 40 Figure 7: Hanging Drop Plate Assembly...40 Figure 8: Spheroid Plating and Formation..41 Figure 9: Plating of hetero cellular spheroids..42 Figure 10: Schematic illustrating the experimental design of the gel retardation assay with 2% and 4% PAA treatment for 10 and 20 minutes each with appropriate controls.57 Figure 11: Schematic of layout of uptake study in Panc1 cells grown in normoxic conditions 58 Figure 12: Schematic of layout of uptake study in spheroids 59 Figure 13: Schematic illustrating the layout of the transfection study.. 62 Figure 14: Day 5 Panc1 homo-cellular spheroid under 10X magnification - Average Diameter = microns and Average Thickness = 135 microns 63 Figure 15: Day 5 1:1 Panc1:J774.A1 hetero-cellular spheroid under 10X magnification - Average Diameter = microns and Thickness = 60 microns Figure 16: Day 5 1:1 Panc1: NIH/3T3 hetero-cellular spheroid under 10X magnification - Average Diameter = and Thickness = microns

10 Figure 17: Day 5 1:1:1 Panc1:J774.A1:NIH3T3 hetero-cellular spheroid under 10X magnification - Average Diameter microns and Average Thickness = 55 microns.. 64 Figure 18 and 19: Distribution of Panc1, J774.A1, NIH3T3 cells in the Day 5 1:1:1 co-culture spheroid.64 Figure 20: 2% Agarose E- gels with inflammatory marker expression analysis in Day 5 Panc1 homo-cellular spheroids..66 Figure 21: Relative expression of inflammatory markers in Day 5 Panc1 spheroids. All expressions were normalized to β-actin as housekeeping gene...66 Figure 22: 2% Agarose E- gels with inflammatory marker expression analysis in Day 5 1:1 Panc1: J774.A1 spheroids Figure 23: Relative expression of inflammatory markers in Day 5 1:1 Panc1: J774.A1 spheroids. All expressions were normalized to the respective β-actin as housekeeping gene..67 Figure 24: 2% Agarose E- gels with inflammatory marker expression analysis in Day 5 1:1 Panc1: NIH/3T3 spheroids 68 Figure 25: Relative expression of inflammatory markers in Day 5 1:1 Panc1: NIH/3T3 spheroids. All expressions were normalized to the respective β-actin as housekeeping gene 68 Figure 26: 2% Agarose E- gels with inflammatory marker expression analysis in Day 5 1:1:1 Panc1: J774.A1: NIH/3T3 spheroids...69 Figure 27: Relative expression of inflammatory markers in Day 5 1:1:1 Panc1: J774.A1: NIH/3T3 spheroids. All expressions were normalized to the respective β-actin as housekeeping gene.69 Figure 28: Statistical evaluation of expression of the IL-8 gene with 1-tailed test across all spheroid models. All expressions were normalized to the respective β-actin as housekeeping gene

11 Figure 29: Statistical evaluation of expression of the TNF-α gene with 1-tailed test across all spheroid models. All expressions were normalized to the respective β-actin as housekeeping gene...70 Figure 30: Statistical evaluation of expression of the TGF-β gene with 1-tailed test across all spheroid models. All expressions were normalized to the respective β-actin as housekeeping gene...71 Figure 31: 2% Agarose gels with LDH-A in Panc1 cells grown in a Normoxic monolayer, hypoxic monolayer and day5 homo- and hetero-cellular spheroids 71 Figure 32: Relative expression of LDH-A in Panc1 cells grown in a Normoxic monolayer, hypoxic monolayer on day 5 and day 5 homo- and hetero-cellular spheroids. All expressions were normalized to the respective β-actin as housekeeping gene. 72 Figure 33: Relative expression of Beta actin and 28srRNA in Panc1 cells grown in a Normoxic monolayer day 5 homo-cellular spheroids Figure 34: Relative expression of HIF1α across different models compared to Panc1 cells grown in normoxic conditions Figure 35: Relative expression of HIF2α across different models compared to Panc1 cells grown in normoxic conditions Figure 36: Relative expression of LDH-A across different models compared to Panc1 cells grown in normoxic conditions Figure 37: Relative expression of SCF across different models compared to Panc1 cells grown in normoxic conditions Figure 38a, 38b and 38c: CD24+ staining in Panc1 Day 5 homo-cellular Spheroids.78 Figure 39a, 39b and 39c: CD24+ staining in 1:1 Panc1:J774.A1 Day 5 hetero-cellular Spheroids78 11

12 Figure 40a, 40b and 40c: CD24+ staining in 1:1:1 Panc1:J774.A1:NIH3T3 Day 5 hetero-cellular Spheroids Figure 41a, 41b and 41c: Stem Cell Factor staining in Panc1 Day 5 homo-cellular Spheroids 79 Figure 42a, 42b and 42c: Stem Cell Factor staining in 1:1 Panc1:J774.A1 Day 5 hetero-cellular Spheroids Figure 43a, 43b and 43c: Stem Cell Factor staining in 1:1:1 Panc1:J774.A1:NIH3T3 Day 5 heterocellular Spheroids Figure 44: NMR spectra of the HA polymer Figure 45: NMR spectra of the PEI polymer Figure 46: NMR spectra of HA polymer and PEI polymer Figure 47: NMR spectra of the HA-PEI conjugate Figure 48: NMR spectra of the 2kDa PEG polymer Figure 49: NMR spectra of HA polymer and 2kDa PEG polymer Figure 50: NMR spectra of the HA-2kDa PEG conjugate Figure 51: Standard curve plot of fluorescence values of various standards of cy5 in 1X PBS.. 85 Figure 52: Size distribution profile of blank HA-PEI: HA-PEG nanoparticles prepared in a 50:50 w/w HA-PEI: HA-PEG ratio Figure 53: Size distribution profile of HA-PEI: HA-PEG nanoparticles loaded with mir34a mimic prepared in a 50:50 w/w HA-PEI: HA-PEG ratio and a 27:27:1 w/w HA-PEI: HA-PEG: mir34a ratio 86 Figure 54: TEM images of loaded HA-PEI nanoparticles 88 Figure 55a: 4% Agarose EX gel showing released bands of microrna mimic from the HA-PEI: HA- PEG nanosystems with 2% and 4% PAA for 10 and 20 minutes each

13 Figure 55b: Bar plot showing changes in release of microrna mimic from nanosystems with different PAA treatments 90 Figure 56: Standard curve of fluorescence v/s concentration of all Pico green reagent standards...91 Figure 57a: Untreated Panc1 cells at 12 hours of incubation with 1X PBS..92 Figure 57b: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 50nM sirna concentration in Panc1 cells treated for 1 hour.92 Figure 57c: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 50nM sirna concentration in Panc1 cells treated for 3 hours..92 Figure 57d: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 50nM sirna concentration in Panc1 cells treated for 6 hours...93 Figure 57e: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 100nM sirna concentration in Panc1 cells treated for 6 hours...93 Figure 57f: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 100nM sirna concentration in Panc1 cells treated for 12 hours...93 Figure 58a: Untreated Panc1 homo-cellular spheroids incubated in 1X PBS for 24 hours 95 Figure 58b: Panc1 homo-cellular spheroids incubated in Cy5 dye in 1X PBS for 24 hours..95 Figure 58c: Uptake in Panc1 homo-cellular spheroids incubated in blank 50:50 w/w HA-PEI: HA- PEG nanoparticles in 1X PBS for 24 hours...96 Figure 58d: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA- PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 1 hour.96 Figure 58e: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA- 13

14 PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 3 hours.97 Figure 58f: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA- PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 6 hours.97 Figure 58g: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA- PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 12 hours.98 Figure 58h: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA- PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 24 hours.98 Figure 59a: Untreated Panc1:J774.A1 hetero-cellular spheroids incubated in 1X PBS for 24 hours 99 Figure 59b: Panc1:J774.A1 hetero-cellular spheroids incubated in Cy5 dye in 1X PBS for 24 hours...99 Figure 59c: Uptake in Panc1:J774.A1 hetero-cellular spheroids incubated in blank 50:50 w/w HA- PEI: HA-PEG nanoparticles in 1X PBS for 24 hours..100 Figure59d: Uptake in Panc1:J774.A1 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 1 hour Figure 59e: Uptake in Panc1:J774.A1 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 3 hours

15 Figure 59f: Uptake in Panc1:J774.A1 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 6 hours Figure 59g: Uptake in Panc1:J774.A1 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 24 hours Figure 60a: Untreated Panc1: NIH/3T3 hetero-cellular spheroids incubated in 1X PBS for 24 hours Figure 60b: Panc1: NIH/3T3 hetero-cellular spheroids incubated in Cy5 dye in 1X PBS for 24 hours Figure 60c: Uptake in Panc1: NIH/3T3 hetero-cellular spheroids incubated in blank 50:50 w/w HA-PEI: HA-PEG nanoparticles in 1X PBS for 24 hours Figure 60d: Uptake in Panc1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 1 hour Figure 60e: Uptake in Panc1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 3 hours. 105 Figure 60f: Uptake in Panc1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 6 hours Figure 60g: Uptake in Panc1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 12 hours

16 Figure 60h: Uptake in Panc1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 24 hours Figure 61a: Untreated 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids incubated in 1X PBS for 24 hours. 107 Figure 61b: Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids incubated in Cy5 dye in 1X PBS for 24 hours Figure 61c: Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids incubated in blank 50:50 w/w HA-PEI: HA-PEG nanoparticles in 1X PBS for 24 hours 108 Figure 61d: Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 1 hour Figure 61e: Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 3 hours 109 Figure 61f: Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 6 hours 109 Figure 61g: Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 12 hours 110 Figure 61h: Uptake in 1:1:1 Panc1: J774.A1: NIH/3T3 hetero-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 24 hours 110 Figure 62: mir34a transfection using HA-based nanosystems in 1:1:1 hetero-cellular spheroids. 112 Figure 63: mir34a transfection using HA-based nanosystems in normoxic Panc1 cells

17 LIST OF TABLES Table 1: Human and Mouse Primer sequences for IL8, TNF alpha, TGF beta and Beta actin. 43 Table 2: RT-PCR amplification conditions 44 Table 3: Human primer sequences for LDH-A and Beta actin45.45 Table 4: Human primer sequences for 28srRNA and Beta actin.46 Table 5: Human primer sequences for human HIF-1α, HIF-1β, SCF, LDH-A and β-actin 46 Table 6: Surface charge of blank and loaded HA-PEI: HA-PEG nanoparticles prepared in a 50:50 w/w HA-PEI: HA-PEG ratio 86 17

18 1. SUMMARY Pancreatic cancer is one of the most aggressive cancers worldwide with an extremely poor prognosis. There is a largely unmet clinical need for the development of effective therapies for its treatment. Most incidences of pancreatic cancer in patients are a direct result of chronic pancreatitis. This chronic inflammatory state is strongly implicated in bringing about a mutation of healthy cells into tumor cells. Also, the cross-talk between tumor cells and their in-vivo stromal microenvironment, containing macrophages, fibroblasts and other tumor-associated cells through various cytokines results in the development of a more aggressive and malignant phenotype. Additionally, the three dimensional environment of a tumor contains an outer proliferative zone, with a healthy supply of blood vessels and thereby, oxygen and other necessary nutrients that are required for cell growth. However, the inner core of a tumor is necrotic and hypoxic due to diminished blood supply. This oxygen gradient contributes towards the development of a stem cell like phenotype among the tumor cells, which is more metastatic and aggressive. These inflammatory and cancer stem cell signaling pathways are regulated by a number of small non-coding micro-rnas (mirnas) that are present endogenously. Exogenous delivery of mirna or mirna mimics to induce silencing of pro-oncogenic genes and to increase expression of tumor- suppressive genes is emerging as an efficacious platform to modulate the phenotype of tumor cells as these are highly selective to their targets compared to small molecule therapies. However, most anti-cancer therapies use cell monolayers as a model to test efficacy. These models do not recapitulate the complex signaling pathways that occur in vivo in tumors. Therefore, therapies that show promise in these monolayer platforms do not necessarily have translatable efficacies in vivo. Thus, the establishment of a hetero-cellular, three-dimensional invitro tumor model is essential to mimic in vivo hypoxic microenvironments and the dynamic interplay between tumor cells, macrophages and fibroblasts. 18

19 This MS dissertation project will focuses on the establishment of such an in vitro spheroid model, an analysis of the inflammatory and cancer stem cell-like phenotype and the evaluation of a nano-scale, hyaluronic acid (HA) based mirna delivery system to modulate the phenotype of the tumor cells. The delivery system will be evaluated for uptake, transfection within spheroid models. Finally downstream effects of the tumor suppressive microrna will be evaluated through cytotoxicity studies. 19

20 2. OBJECTIVES AND SPECIFIC AIMS 2.1. Statement of the problem Pancreatic cancer is the fourth deadliest cancer in terms of prognosis, with a 6% five year survival rate. With an incidence rate of 12.1 for every 100,000 men and women worldwide, it has nearly equal incidence and mortality rates saw an estimated 38,460 deaths related to pancreatic cancer and despite tremendous efforts to improve patient survival, surgery remains to be the most viable option to most patients, and this too provides only palliative relief for many. Non-surgical options like radiation therapy and chemotherapy have only been successful in prolonging survival from approximately two weeks to a few months. Current treatment options including combination drug therapies have also not been able to improve patient survival rates compared to single drug therapy.[5-9] The strong link between inflammation and pancreatic cancer has been discovered to be a driving force for tumor metastasis and the poor prognosis of the disease, and the microenvironment of chronic pancreatitis bears a strong resemblance to that of pancreatic cancer. Within a three-dimensional environment, tumor cells are also capable of switching between a non-stem cell-like and a stem cell-like phenotype. These cancer stem cells (CSCs) are more aggressive, more capable of metastasis and significantly contribute to drug resistance. In vitro cell monolayer platforms for testing chemotherapeutic drugs do not recapitulate the threedimensional environment of a tumor. Secondly, these models do not contain the inflammatory and fibrotic components of a tumor. A hetero-cellular, three-dimensional (3D) spheroid model consisting of tumor cells, tumor-associated macrophages (TAMs) and cancer-associated fibroblasts (CAFs) serves as a better platform for evaluation of inflammatory markers, the crosstalk between tumor cells and their microenvironment, and to evaluate a CSC-like phenotype. This model would also provide a better evaluation of the efficacy of anti-cancer therapeutic as well as selective targeting of the delivery system to tumor cells.[10, 11] 20

21 2.2. Objectives and experimental hypothesis The first aim of the project was to establish an in vitro model of a three dimensional spheroid composed of tumor cells, macrophages and fibroblasts, that can recapitulate certain components of the microenvironment of a pancreatic tumor more closely than two dimensional co-culture systems. Here, the hypothesis was that inflammatory and cancer stem cell signaling pathways are largely affected by the hypoxic conditions and by the arrangement and interaction of the different cell lines in the 3D structure of the spheroid. There are three theories that are being tested in the multi-cellular spheroid model: (1) the effect of the cross-talk between inflammatory cytokines and the resultant development or suppression of an aggressive, metastatic phenotype (2) the effect of hypoxia on the increase or decrease of cancer stem celllike phenotype and (4) the interplay of the inflammatory and cancer stem cell signaling pathways. Post confirmation of these two signaling pathways in the 3D spheroid model, a nano-scale hyaluronic acid-based, mirna delivery system was established as a therapeutic platform. This nanosystem was characterized and evaluated for its ability transfect tumor cells within the spheroids model through receptor-mediated endocytosis and bring about tumor cell-death due to downstream effects of the microrna. 21

22 2.3. Specific aims To develop an in vitro model of hetero-cellular spheroids composed of Panc-1 human pancreatic adenocarcinoma cells, J774.A1 murine macrophages and NIH/3T3 murine fibroblasts and to evaluate pro-and anti-inflammatory markers as well as CSC-specific markers. After establishing inflammatory and cancer stem cell signaling pathways in this model, a hyaluronic acid (HA)-based mirna delivery system will be developed and evaluated for its ability to modulate this signaling in order to down-regulate tumor-promoting pathways and up-regulate tumor-suppressive pathways. To validate the objectives, the specific aims of the project are: Aim 1: Establishment of Homo- and Hetero-Cellular 3-Dimensional Tumor Spheroids using Panc- 1, J774.A1, and NIH-3T3 Cells and Evaluation of Baseline Inflammatory Marker Expression Profiles a) Development of 3 dimensional homo-cellular spheroids of Panc-1 human pancreatic adenocarcinoma cells and hetero-cellular spheroids containing Panc-1, J774.A1 murine macrophages, and NIH-3T3 murine fibroblast cells using the hanging drop method. b) PCR analysis of inflammatory marker expression profiles in homo-cellular (Panc-1) and hetero-cellular (Panc-1, J774.A1 and NIH-3T3 cells) spheroids. Aim 2: Evaluation of Baseline Cancer Stem Cell Marker Expression Profile in Panc-1 Monolayer Cultures and Homo- and Hetero-Cellular Spheroids. a) PCR analysis of cancer stem cell marker LDH-A expression profiles in Panc-1 monolayer cell cultures in normoxic and hypoxic conditions and homo-cellular and hetero- cellular spheroid models. b) Qualitative analysis for evaluation of cancer stem cell markers CD24+ and Stem cell factor (SCF) through antibody staining and confocal microscopy. c) qpcr to quantitatively evaluate levels of hypoxia and cancer-stem cell markers, including HIF-1α, HIF-2α, LDH-A and SCF. 22

23 Aim 3: Formulation and characterization of Hyaluronic acid based nanosystems: a) Synthesis of various HA-based polymer conjugates and evaluation of conjugation using NMR b) Preparation and characterization of nanosystems using HA-based polymer conjugates through Dynamic Light Scattering and Transmission Electron Microscopy c) Evaluation of microrna deliverability in these formulations Aim: Uptake and Evaluation of Effects of mirna Therapy Administered in HA Nanoparticles in 3D Spheroid Model a) In vitro evaluation of delivery efficiency using uptake studies (using confocal microscopy) b) Evaluation of mirna transfection in the 3-in-1hetero-cellular spheroid model 23

24 3. BACKGROUND AND SIGNIFICANCE 3.1. Pancreatic Cancer Incidence and severity With an incidence rate nearly equaling the mortality rate, pancreatic cancer is one of the most aggressive forms of the gastro-intestinal tract cancers with a survival rate of below 5% due to early metastasis to neighboring sites such as the liver, improved angiogenesis, Multi-Drug Resistance (MDR) and activation of growth and invasion inducing factors [12]. Chemotherapy has not proved to be efficient in the treatment of pancreatic cancer due to late diagnosis and only 15% of the patients presenting symptoms are operable and the remaining 85% have a very poor chance of cure [13]. Therefore, the most preferred treatment method for tumors in the pancreatic duct remains as surgical resection [14, 15]. Higher incidences of pancreatic cancer are linked with heredity, chronic pancreatitis, alcohol consumption and smoking [12]. The ratio of pancreatic cancer diagnosis is 53:1 for patients with and without hereditary pancreatitis [16] Current therapy Due to the fact that a majority of pancreatic cancer patients are not operable at the time of first diagnosis, radiotherapy and chemotherapy serve as alternative treatment methods. However, even with highly toxic drugs such as gemcitabine and 5-fluorouracil, the survival rate of most patients is still less than a year as the densely fibrotic nature of the tumor acts as a mechanical barrier that limits the entry of chemotherapeutic agents. Gene based therapies and targeted systems like apoptosis modulators and angiogenesis inhibitors, which have been shown to be successful in vitro, fail to translate into effective clinical therapies [10]. The multifarious nature of pancreatic cancer and the dynamic interplay between the tumor cells and their microenvironment necessitates the development of an efficacious, targeted therapy. 24

25 3.2. Inflammation and pancreatic cancer Literature evidence suggests that inflammation has a critical role in driving the formation of solid tumors and in mediating the development and progression of pancreatic cancer due to the tumor conducive environment of inflammation [10]. Virchow discovered the presence of leukocytes in neoplasms and proposed that there exists a strong correlation between inflammation and the origin of cancer [16]. Consistent efforts have been made to identify and elucidate the multifaceted relationship between inflammatory mechanisms and tumorigenesis with varying successes. Fibroinflammatory responses occur subsequently upon cellular injury or disturbance in pancreatic cell homeostasis. Inflammation leads to rapid cell proliferation and induces genomic damage, which brings about a malignant transformation of pancreatic ductal cells. Various inflammatory cytokines, reactive oxygen species and other elements of the inflammatory pathway are involved in increasing cell cycling and metastasizing tumors [16]. Some of these inflammatory mediators including COX-2 and NF-kB are known to induce a decrease in the expression of tumor suppressor genes and an increase in oncogene levels. According to the 'Landscaper Theory', chronic inflammation of the pancreas damages stromal cells, which is followed by the healing process, during which these cells are exposed to multiple growth factors. The increased proliferation rates and cell damage leads to the formation of an abnormal microenvironment where stromal cells facilitate the growth of transformed cells [16]. In hereditary pancreatitis (less than 1% of all pancreatitis forms), the trypsinogen gene on chromosome 7 undergoes a mutation that slows breakdown of abnormally expressed trypsinogen or further enhances trypsinogen activation. This results in acinar cell damage and repetitive occurrences of acute inflammation followed by progressive damage, fibrosis and chronic pancreatitis. However, the leading cause of pancreatic cancer remains to be smoking of tobacco products. Cigarette smoking is known to be a cause of both chronic pancreatitis and pancreatic cancer and subsequently, most pancreatic cancer patients are cigarette smokers or heavy drinkers [17]. 25

26 Role of inflammatory markers The microenvironment of a pancreatic tumor is known to be associated with an upregulation of pro-inflammatory signaling [18]. However, diagnosing pancreatic cancer has proved to be a challenge due to the lack of a specific marker for the disease [18]. Previously, carbohydrate antigen 19-9 (CA19-9) and carcino-embryonic antigen (CEA) were used as markers, but were unreliable due to false positives. Hence, it is advantageous to use a combination of markers to specifically diagnose pancreatic tumors [19] TGF-beta signaling pathway TGF-beta is an anti-inflammatory cytokine that regulates cell growth, apoptosis and cell differentiation. In healthy cells and in early stages of tumor growth, TGF-beta has a tumor suppressive role, by preventing cell growth and bringing about apoptosis of damaged cells. It is expressed at high levels during latter stages of tumor growth although the cells do not respond to the growth inhibitory effects at this point, thereby implying that this cytokine switches its role to that of a tumor supporting one. In normal cells, TGF-beta binds to the TGF-beta Type I and Type II receptors, and this causes the hetero-dimerization of both these receptors. The Type II receptor phosphorylates the Type I receptor kinase domain, which leads to a cascade of reactions, the first step of which involves the phosphorylation of SMAD proteins by the Type I receptor. The phosphorylated SMAD proteins enter the nucleus, and the tumor suppressive effects are mediated by SMAD activation of the concerned target genes within the nucleus [10]. However, various pancreatic cancer cell lines express mutated TGF-beta receptors or deletion of the SMAD4 protein. Therefore, the impairment of the TGF-beta signaling pathway results in increased cell proliferation, improved cell survival of cancer cells, enhanced invasiveness and metastasis by upregulating matrix metallo-proteinase 2 (MMP2) and activating NF-κB. The TGFbeta pathway is therefore an attractive target for gene based and other drug therapies [10] TNF-alpha signaling pathway TNF- Alpha is a pro-inflammatory cytokine that is heavily involved in initiating the acute phase reaction in systemic inflammations. Chiefly produced by macrophages, it's primarily role is 26

27 to control the function of other immune cells. Due to its inherent pyrogenic nature, it induces fever, apoptosis, inflammation and inhibits tumorigenesis. Additionally, when cell lines were treated with TNF-alpha, an increased expression of TGF-alpha and its receptor EGFR (Epidermal Growth Factor Receptor) was seen. The signaling of the EGFR pathway pushes cells to transform into a more malignant phenotype [10] IL8 signaling pathway IL8 is another pro-inflammatory cytokine known to increase proliferation in pancreatic cancer cells. It is known to be up-regulated in pancreatic cancer cell lines and plays an important role in regulating angiogenesis and metastasis by inducing Vascular Endothelial Growth Factor (VEGF) and neuropilin-2. IL8 acts through the MAPK pathway and phosphorylates ERK2 [10] Cancer stem cells Stem cells are characterized by their ability to repair damaged or ageing cells in their body and are themselves capable of self-renewal. Their ability to self-renew is unaffected by their capacity to proliferate, with each subsequent asymmetric cell division forming one daughter stem cell and one daughter progenitor cell. Another property that distinguishes them from other cells is their increased resistance to chemotherapy. Because they are pluripotent, they differentiate into various cell types under the influence of different signaling molecules within their environment. Cancer stem cells have been isolated from various cancers including breast cancer, pancreatic cancer and chronic myeloid leukemia and it has been found that these cells remain viable even after treatment with chemotherapy, unlike somatic tumor cells [1]. Recently, a specific type of cancer cells, known as the Cancer Stem Cells have been discovered to be closely related in multiple aspects to stem cells. According to the tumor heterogeneity theory, a tumor is comprised of a heterogeneous population of tumor initiating and non-tumor initiating cells. These tumor initiating or cancer stem cells exhibit stem cell markers like CD24, CD44, CD133, Oct4, Nestin and ESA (Epithelial-specific antigen) and are capable of self-renewal and exhibiting drug resistance [1, 20, 21]. 27

28 Whether cancer stem cells are derived from normal stem cells or by differentiation of non-tumor initiating cells through epigenetic reprogramming is unclear. It is also hypothesized that both these processes may be involved. Cellular mutations and epigenetics bring about a multi-step transformation of normal cells to tumor cells and it could be presumed that cancer stem cells are derived from normal stem cells through the same process. However, once formed, cancer stem cells divide to form two distinct cell populations: one, capable of self-renewal and the other, non-tumor initiating [1]. Even after surgical resection of the primary tumor, these cancer stem cells have been proposed to be involved in tumor recurrence by getting activated under specific conditions or through reprogramming of damaged somatic cells into a progenitor state. Figure 1 illustrates the origin of cancer stem cells and the effect of chemotherapy on these cells [1]. Figure 1: Schematic illustrating cancer stem cell initiation and their response to chemotherapeutic agents [1] Both normal stem cells and cancer stem cells involve oncogenes and tumor suppressor genes and exploit the Notch, Bmi-1, Wnt and Sonic hedgehog pathways to modulate their selfrenewal capacities. Additionally, the Wnt beta-catenin pathway has also been found to be a contributor to this self-renewing characteristic. [22-26] 28

29 Hypoxia and its role in maintenance of cancer stem cells Reduced oxygen tension, also known as hypoxia, is a common phenomenon seen in tumors due to dysregulated tumor vasculature. In normal cells as well as tumor cells, hypoxia causes severe damage. However, cancer cells, unlike normal cells, have the ability to adapt to these hypoxic conditions and sustain their growth through alteration of their metabolic pathways. These changes may also be genetic, giving rise to the additional tumor cell characteristics of therapeutic resistance, enhanced angiogenesis and migration. Hypoxia marker identification in various tumor models has been correlated with poor prognosis. The switch in the metabolic pathway for glucose metabolism from aerobic tricarboxylic acid (TCA) cycle to anaerobic glycolysis helps in sustenance of cancer cell growth. According to the Warburg effect, it has been observed that cancer cells prefer anerobic pathways regardless of oxygen availability in their microenvironment [27-29]. Cancer stem cells are able to sustain their population by regulating their behavior in their unique microenvironments, known as niches. These niches have been found to be hypoxic, and hypoxia helps maintain the self-renewal property and the undifferentiated state of CSCs. Stem cells favor this hypoxic state in order to avoid DNA damage from Reactive Oxygen Species (ROS), which occurs at higher oxygen levels and under hypoxic conditions, the levels of Hypoxia Inducible Factors, HIF1α and HIF2α are elevated. However, CSCs have the ability to induce a hypoxia-like state regardless of the oxygen levels in their surroundings, through up-regulation of HIF2α. So, it can be said, that although cancer stem cells depend on their niche conditions, are also capable of modulating it to sustain their self-renewal properties and undifferentiated state [30-32] Inflammation as a regulator of cancer stem cells The inflammatory response associated with cancer is dynamic, with a two-way communication between cancer cells and stromal and immune cells through various cytokines and chemokines. Various transcription factors, including NF-κB, Snail, Slug, and a number of signaling pathways are known to be important regulators in driving epithelial-mesenchymal transition (EMT) as well as CSC formation. The NF-κB pathway provides a distinguished link 29

30 between EMT, inflammation and CSCs. Pro-inflammatory cytokines such as IL-6, TNF-α andil-1β stimulate the release of HIF1α in the cancer cells, leading to activation of the NF-κB pathway. The activation of the NF-κB pathway leads to the suppression of various apoptotic genes. Along with the NF-κB pathway, Snail and Slug (the two regulators of EMT) are also activated, leading to a decrease in adhesion protein, E-cadherin levels. Decrease in the levels of adhesion protein accounts for the ability of the cancer cells to break away from the tumor and metastasize in distal parts of the body [2, 33, 34]. Figure 2: Signaling pathways involved with CSCs and EMT [2] Inhibition of NF-κB results in the decrease of EMT and retardation of CSC proliferation. In pancreatic cells, TNF-α activates the Shh pathway, resulting in increased proliferation of tumor cells through up-regulation of Snail and other EMT regulators [35, 36]. Continuous activation of p53, a classical tumor suppressor gene, brings about release of high mobility group box 1 (HMGB1). HMGB1 stimulates the release of TNF-α from neighboring immune cells, leading to NFκB and Snail activation [37, 38]. Target NF-κB genes, such as CXCL1 attract myeloid cells towards 30

31 the tumor, which also promote EMT [39]. TGF-β, apart from being a regulator of EMT itself, is also involved in recruiting various immune cells to the tumor microenvironment, leading to release of pro-inflammatory cytokines like TNF-α [40, 41]. TNF-α up-regulates TGF-β levels, creating a positive feedback loop [38]. IL-6 mediated Stat3 activation is regulated by NF-κB and the IL6/Stat3 pathway is important for non-csc to CSC transition. Thus, the NF-κB and Stat3 pathway work synergistically to promote a CSC phenotype [42, 43]. In colorectal cancer, myofibroblasts cells from the stroma of the tumor secrete Hepatocyte Growth Factor (HGF), which mediates non-csc to CSC transformation [2, 44]. The CSCs have an efficient cytokine network, and by releasing IL8, MCP-1 and RANTES, they are capable of inducing proliferation of stromal cells as well as regulating chronic inflammation [45]. This suggests that the CSCs are important mediators of the inflammatory processes as well. Metastatic breast cancer cells in the lung have also been found to secrete TGFβ2/TGF-β3, which stimulate periostatin release from stromal fibroblasts. Periostatin upregulated Wnt ligands and the Wnt pathway are strong regulators of the CSC self-renewal property [2, 46] MicroRNA Introduction to microrna (mirna) Mature (mirna) are endogenously expressed single- stranded RNA molecules with the primary function of regulating post-transcriptional protein synthesis. They are typically nucleotides long, and do not code for proteins themselves. They are conserved evolutionally and are involved in controlling various biological processes such as apoptosis, differentiation, cell development and proliferation [3, 47, 48] Biogenesis and function of mirnas: As shown in the Figure 3, RNA Polymerase II transcribes primary mirnas (pri-mirnas), which are 1-3kb in length, from the intergenic or intragenic regions within the nucleus of the cell. These pri-mirnas are then acted upon by the Drosha and Pasha enzymes within the nucleus and cleaved to form pre-micrornas (pre-mirnas). Drosha is an RNase II enzyme, while Pasha is an 31

32 RNA binding protein). These pre-mirnas are typically nucleotides long and contain a stem-loop in their structure [3]. Figure 3: The biogenesis and function of mirna [3] Exportin 5, a karyopherin protein molecule, then transports the pre-mirna from the nucleus to the cytoplasm. In the cytoplasm, the pre-mirna is acted upon by Dicer, an RNase II enzyme. Dicer converts the pre-mirnas into double-stranded molecules and this process mostly involves the loss of the stem-loop structure. At this stage, a mirna: mirna duplex is formed, which is unwound and separated into 2 strands, with one strand being incorporated into the RNA-Induced Silencing Complex (RISC), while the second "passenger" strand is degraded or shows a regulatory effect in controlling mirna homeostasis [3]. While in the RISC complex, the mirna is paired with its target mrna template either perfectly or imperfectly. This is the primary difference between silencing RNA or sirna and mirna, in that sirna requires perfect pairing in order to induce inhibition of mrna translation. This difference also gives a single mirna the ability to target multiple mrna sequences or conversely multiple mirnas may target the same mrna sequence [3, 49]. 32

33 Once the RISC complex matches the mirna to the mrna, at the 3' UTR region of the mrna, it inhibits translation of mrna or it sequesters or degrades the mrna. However, it has also been found that mirna are capable of acting as positive regulators of gene expression although the mechanism behind this phenomenon has not been elucidated yet [3] mirna in cancer Cancer is often associated with aberrant expression of mirnas, which contribute to the etiology, pathogenesis and prognosis of the disease. This altered mirna signature is notably different from mirna signatures in non-cancerous cells. In different cancers, different mirnas or combinations or mirnas are either up-regulated or down-regulated and thus mirnas can be used as promising biomarkers for cancer diagnostics. This dysregulation is a result of gene polymorphisms, genetic mutations, epigenetic changes, chromosomal irregularities and improper mirna biogenesis. Sometimes, mirna changes are common between different cancers, which indicates that these mirnas are involved in cancer-modified, rudimentary signaling pathways [3]. mirnas may either function as oncogenes or tumor suppressors and sometimes only a single mirna is required to cause tumorigenesis. Common examples of oncogenic mirna upregulation in cancers include mir-155 in B cell lymphomas and the mir cluster in lung cancers and lymphomas. Tumor suppressive mirnas are down-regulated in cancer, and these function as suppressors of oncogenes. These also inhibit genes that inhibit apoptosis or genes that support cell differentiation and development. Let-7 is an example of a tumor suppressive mirna which is down-regulated in various cancers including lung cancer [3] mir34 and cancer: The mir-34 family of mirnas consists of 3 members, including mir-34a, mir-34b and mir- 34c. All these members are directly controlled by tumor protein p53, and mir-34 plays the role of a tumor suppressor downstream of p53. In both primary and tumor cells, mir-34 induces cell cycle arrest by decreasing the activity of a family of genes that promote cell cycle progression. In cells with mutated p53, mir-34 expression is dramatically decreased leading to poor prognosis of 33

34 cancer. However, reinstituting the mir-34 expression inhibits cell division. Since it plays an important role in mediating apoptosis in these cancer cells, chemoresistance is correlated with a decrease in mir-34 levels [50]. Other targets of mir-34 include Notch, c-met and BCL-2, all of which are genes that control renewal and survival of cancer stem cells (CSCs). Pancreatic cancer cell lines including MiaPaCa2 and BxPC3 were found to express low levels of mir-34, but high levels of its targets BCL-2 and Notch-1. When these cell lines are transfected with mir-34, a decrease in the levels of the target proteins is seen without change in target mrna levels, which supports the concept that mirnas act post transcriptionally. Transfection with mir-34a mimics also leads to apoptosis and increases the sensitivity of these cells to chemotherapy and radiation [50]. CSCs express specific markers, including CD44 and CD133 on their surface, and this expression was found to decrease with mir-34 transfection. CSCs exhibit a more aggressive phenotype and this decrease in CD133 and CD44 was indicative of a decrease in the CSC population. MiR-34 transfection further led to a decrease in MiaPaCa2 tumor formation in nude mice [50] mir-34 and the Notch pathway The Notch signaling pathway is an important regulator of cell proliferation, differentiation and apoptosis. There are 4 Notch receptors, Notch 1-4 and five Notch ligands in mammals. It has been found that the Notch signaling pathway is strongly associated with pancreatic cancer stem cells (CSCs), which are known to be highly aggressive and drug resistant. Targeting specific signaling molecules of the Notch pathway could serve as a regulator of drug resistance and elimination of these CSCs [51]. The transmembrane Notch ligand from one cell binds to the transmembrane Notch receptor of a neighboring cell, bringing about activation of Notch signaling. Various proteolytic enzymes like metalloprotease, tumor necrosis factor-α-converting enzyme and ɣ-secretase cleave Notch, resulting in the formation of Notch extracellular truncation (NEXT). The ɣ-secretase complex then cleaves NEXT to release Notch intracellular domain (NICD), which is an active 34

35 fragment. NICD then translocates to the nucleus, where it binds to suppressor of hairless and lag- 1 (CSL), and recruits a co- activator complex p300, leading to activation of target Notch genes [51]. In epithelial pancreatic cancer cells, Notch pathway molecules are found to be upregulated and this pathway plays an oncogenic role by maintaining the progenitor state, and delaying cell differentiation. Studies using sirna and ɣ-secretase inhibitors (GSI) in order to down-regulate Notch signaling have shown a decrease in proliferation, decreased migration and invasion and increased susceptibility to apoptosis in pancreatic cancer cells [51-53] mir-34 and Wnt signaling pathway mir-34 is involved in regulating the Wnt signaling pathway which controls epithelial to mesenchymal transition (EMT) as well as metastasis. While WNT1, WNT3 and LRP-6 are directly inhibited by mir-34a and mir-34b/c, the mir-34 family are also capable of repressing LEF1 and beta-catenin downstream of Wnt [54, 55] Exogenous microrna therapy Major challenges associated with RNA interference therapies are: (i) selection and validation of RNA interference targets (ii) safety and (iii) delivery: accumulation at target site, intracellular uptake and cytoplasmic release of therapeutic [56]. Viral vectors have been used successfully for nucleic acid delivery and although their delivery efficiency is promising, their use is associated with high toxicity and immunogenicity. Polymeric or non-viral nano-sized carriers are not associated with these problems and have thus emerged as better and more viable options for RNA delivery [4]. Accumulation at target site: The efficiency of nanocarriers as RNA delivery systems is dependent on their size, charge, shape and material. Due to their small size (50-200nm) and the Enhanced Permeation and Retention (EPR) Effect seen in solid tumors, these nano-carriers have the capability to localize or accumulate in the target tumor tissue. Surface modification, such as that with PEG, may also add to this capability of localizing in the tumor tissue [56, 57]. 35

36 Intracellular uptake: After accumulation of RNA-loaded nano-carrier in the tumor, specific uptake by tumor cells is required. This is often facilitated by surface modification of the nanocarriers with ligands that target tumor cell specific receptors. This intracellular uptake may occur through receptor-mediated endocytosis, macropinocytosis or surface protein transporters [58]. Cytoplasmic payload release: After the nano-carrier has been taken up selectively by the tumor cell, it is trapped within the endosomal compartment. The endosome has an acidic environment and therefore the nano-carrier needs to escape degradation by the endosomal acids in order to release its payload in the cytoplasm [56]. Inhibition of translation: Once the mirnas are released into the cytoplasm, they elicit their action by binding to the 3'UTR of their target mrnas. MiRNA can either bind perfectly or imperfectly to their target mrna, which could either lead to translation repression or degradation of the target mrna. Due to this imperfect pairing, mirnas are able to exhibit their action on more than one target gene [59] Three-dimensional tumor spheroid models Tumors are composed of a proliferative outer zone and a necrotic central zone supported by the tumor stroma composed of extra-cellular matrix (ECMs), blood vessels, signaling molecules, Tumor Associated Macrophages (TAMs), Cancer Associated Fibroblasts (CAFs) (also called Activated fibroblasts or myofibroblasts) and other infiltrating immune cells [60, 61]. 2D cell monolayers are fairly simple to grow and provide for convenient systems for the screening of anti-neoplastic agents. However, they are morphologically and functionally different from tumors and are unable to reproduce the tumor microenvironment. Moreover, the gene expression profile of a cell is highly influenced by its shape and environment [62]. Therefore, therapies that show promise when tested on these 2 D models are often poorly translatable in vivo [63]. Multi-cellular heterospheroids that are formed by spontaneous aggregation of malignant tumor cells, fibroblasts and macrophages bear a sharper resemblance to the complex structures of avascularized tumors in their early stages of growth in terms of cell proliferation and growth, 36

37 cellular metabolism, oxygen and other nutrient gradients, cell-cell and cell-matrix interactions [64]. Therefore, 3 dimensional spheroid models are crucial in vitro resources that aid in gaining deeper insight into tumor biology. They serve as an efficacious platform for testing of gene therapy, small molecule therapy and immune-based therapy, thereby providing for higher reproducibility with in vivo testing [62] Hyaluronic acid-based nanoparticle formulations Due to numerous problems with naked RNA therapy, such as high molecular weight, negative charge, large size and instability due to endonuclease and exonuclease degradation, a delivery system that encapsulates and protects the RNA and targets tumor cells specifically is required. For this purpose, a non-toxic, non-immunogenic and nano-sized biodegradable system is required, that is capable of delivering the therapeutic RNA molecule to tumor cells and is further capable of inducing receptor-mediated intracellular uptake and bypassing endosomal/ lysosomal degradation to release the RNA molecule into the cell cytoplasm [4]. Low molecular weight hyaluronic acid polymer (HA) (20kDa) has emerged as a ideal polymer for tumor targeted RNA delivery. Hyaluronic acid is a naturally occurring, anionic mucopolysaccharide that is present in synovial fluid and extracellular matrix. It is made of alternating disaccharide units, D- glucoronic acid and N-acetyl D glucosamine (NAG) connected by Beta 1,4-glycosidic linkages. It complies with the requirements of being non-toxic, nonimmunogenic and biodegradable, and allows for chemical modification of its sugar residues through simple coupling chemistry. Functionalization of this molecule results in the formation of self-assembled nanocarriers that are capable of efficiently encapsulating RNA molecules. The polymer backbone contains various chemical moieties that have the ability to specifically target the CD44 ligand, which is a surface receptor expressed on the surface of various tumor cells [4]. Due to its anionic nature, however, encapsulation of negatively charged RNA molecules becomes a problem due to charge-charge repulsion. Therefore, modification of the polymer backbone by hydrophobic modification through the addition of fatty amine groups or cationic polyamine groups results in an overall decrease of the negative charge and improves RNA 37

38 encapsulation through self-assembly. This charge modification also improves endosomal escape and release into the cytoplasm after receptor-mediated endocytosis [4]. Hyaluronic acid Polyethylene glycol (PEG) modification of HA Figure 4: HA conjugated to 20 kda PEG [4] Hyaluronic acid HA conjugated to Polyethyleneimine (PEI) Figure 5: HA conjugated to 10 kda PEI [4] 38

39 4. EXPERIMENTAL DESIGN AND METHODS 4.1. Establishment of homo-cellular and hetero-cellular 3 dimensional tumor spheroids 4.1.a. Cell Culture of Panc1 (pancreatic ductal epitheloid carcinoma cell line), NIH/3T3 (murine fibroblast cell line) and J774 (murine macrophage cell line): In one vial each, of the Panc1, NIH/3T3 and J774.A1 cell lines was obtained from ATCC - American Type Culture Collection (Manassas, VA). Panc1 cells were grown in 1X DMEM (Dulbecco's Modification of Eagle's Medium) with 4.5g glucose and L- glutamine without sodium pyruvate obtained from Corning Cellgro (Manassas, VA). The growth medium was supplemented with 10% fetal bovine serum and 1% penicillin-streptomycin. Cells were passaged at around 80% confluency and detached using 0.25% trypsin 1X from Corning Cellgro (Manassas, VA) and replated. Cells beyond passage number 13 were not used for spheroid plating. NIH/3T3 cell line was grown in 1X DMEM (Dulbecco's Modification of Eagle's Medium) with 4.5g glucose and L- glutamine without sodium pyruvate obtained from Corning Cellgro (Manassas, VA). The growth medium was supplemented with 10% fetal calf serum and 1% penicillin-streptomycin. Cells were passaged at around 70% confluency and detached using 0.25% trypsin 1X from Corning Cellgro (Manassas, VA) and re-plated. J774.A1 cells were grown in 1X DMEM (Dulbecco's Modification of Eagle's Medium) with 4.5g glucose and L-glutamine without sodium pyruvate obtained from Corning Cellgro (Manassas, VA). The growth medium was supplemented with 10% fetal bovine serum and 1% penicillinstreptomycin. Cells were passaged at around 60% confluency and detached using a cell scraper and re-plated. Cells beyond passage number 4 were not used for spheroid plating. 39

40 4.1.b. Development of homo- and hetero-cellular spheroids: Plating of spheroids using a 96 well hanging drop plate Figure 6: Hanging Drop Plates (Image taken from 3dbiomatrix.com) 96-well Hanging Drop plates from 3DBiomatrix were used for producing homo-cellular and hetero-cellular spheroids. Figure 7: Hanging Drop Plate Assembly (Image taken from 3dbiomatrix.com) Sterilized water was added to the peripheral reservoirs on the bottom tray and the hanging drop plate. Using a multichannel pipette, 40µl of cell suspensions of appropriate concentrations were seeded into each well by inserting the pipette tip at least halfway into the well while tilting the pipette at a 45 angle and pipetting down into the well to form a hanging drop. The spheroid plate was observed under a microscope every day until the spheroids were harvested on day 5 of growth. 40

41 Figure 8: Spheroid Plating and Formation (Image taken from 3dbiomatrix.com) Replacing media and maintaining humidity in the spheroid plate 20µl of fresh cell culture medium was added to each well of the spheroid plate on Day2 and Day4 of spheroid growth by inserting a multichannel pipette halfway into each well and dispensing the medium into the well at a 45 angle. Additionally, sterilized water was added only to the hanging drop plate, thereby maintaining humidity to prevent the drying up of the spheroid containing hanging drops. Harvesting spheroids 80µl of PBS (phosphate buffered saline) was aspirated into a multichannel pipette and used to harvest the spheroids by pipetting down into each well at a 90 angle. The spheroid droplets were collected in the bottom tray of the hanging drop apparatus. The harvested spheroids were then transferred into a 50ml centrifuge tube and spun down to form a pellet. The pellet was then stored at -80 C for total RNA extraction and further downstream analyses. Development of single cell (homo-cellular) spheroids For producing spheroids from single cell lines of Panc1, cells were detached, pelleted and re-suspended in culture medium to achieve a final concentration of 500,000 cells/ml or 20,000 cells in each well (40µl). This seeding number was optimized to obtain intact and well- packed spheroids on Day 5. 41

42 Development of hetero-cellular spheroids of Panc1 cells, NIH/3T3 and J774.A1 cells The cells from the 3 cell lines were detached upon reaching confluency and plated to produce spheroids in a 1:1:1 ratio with a total cell seeding number of 20,000 cells/well. Mixed cell suspensions were then plated in the same manner as the homo-cellular spheroids. Spheroid plates were observed for spheroid formation over a 5-day period, subsequently harvested, pelleted and stored in -80 C for RNA extraction and further downstream analyses. Figure 9: Plating of hetero cellular spheroids (Image taken from 3dbiomatrix.com) 4.1.c. Confocal Size Measurements of Day 5 homo-cellular and hetero-cellular spheroids Day 5 homo-cellular and hetero-cellular spheroids were harvested, washed with 1X PBS and mounted on depression microscopic slides using Shandon Immu-Mount (Thermo Scientific, PA), and covered with a cover-slip. They were observed under the Carl Zeiss LSM 700 confocal microscope under 10X magnification. The images were recorded and diameter measurements were taken using Image J software. 5µm optical slices were imaged through Z stacking and the total number of slices per spheroid was used to calculate the thickness of the spheroid. 4.1.d. Distribution of Panc1, J774.A1 and NIH3T3 cells within the hetero-cellular spheroids Cell Staining with membrane staining dyes and spheroid plating Panc1 cells in 1XDMEM were stained with Cellbrite Blue Cytoplasmic Membrane Labeling Kit (Biotium, CA) according to the protocol. J774.A1 cells in 1XDMEM and NIH3T3 were stained with Cellbrite Green Cytoplasmic Membrane Labeling Dye (Biotium, CA) and Cellbrite Red Cytoplasmic Membrane Labeling Dye (Biotium, CA) respectively. The cells were washed twice with 1X PBS, re-suspended in 1X DMEM and counted. These stained cells were plated in a 1:1:1 ratio to form hetero-cellular spheroids using the hanging drop method. 42

43 Analysis of arrangement of cells in hetero-cellular spheroids on Day 5 using confocal microscopy Upon harvesting the hetero-cellular spheroids in a 96 well plate, they were fixed with 4% Formalin and incubated for 10 minutes. The formalin was discarded and ice- cold PBS was used to wash the fixed spheroids. The washing step was repeated three times. The spheroids were then transferred to a microscope depression slide (Fisher Scientific, PA) using Shandon Immu- Mount (Thermo Scientific, PA), covered with a cover slip and images were taken using the Carl Zeiss LSM 700 confocal microscope. 43

44 4.2. Analysis of inflammatory, hypoxic and cancer stem cell phenotype of spheroids through PCR 4.2.a. Total RNA extraction from homo and hetero-cellular spheroids RNA extraction from Day 5 homo-cellular Panc1 spheroids and 1:1 Panc1: J774.A1 hetero-cellular spheroids Total RNA from day5 homo-cellular Panc1 spheroids and 1:1 Panc1: J774.A1 heterocellular spheroids was extracted using the Quick-RNA MiniPrep RNA extraction kit (Zymo Research, CA). Concentrations of the RNA were measured using the NanoDrop (Thermo Scientific, Willington, DE), recorded and the RNA was stored at -80 C. RNA extraction from Day 5 1:1 Panc1: NIH/3T3 hetero-cellular spheroids and 1:1:1 Panc1: J774: NIH/3T3 hetero-cellular spheroids Using the Direct-zol RNA MiniPrep RNA extraction kit (Zymo Research, CA), total RNA from day5 1:1 Panc1: NIH3T3 hetero-cellular spheroids and 1:1:1 Panc1: J774.A1: NIH3T3 heterocellular spheroids was extracted. RNA concentrations were measured and recorded using the NanoDrop (Thermo Scientific, Willington, DE) and the RNA was stored at -80 C. 4.2.b. cdna synthesis from RNA extracted from Day 5 homo and hetero-cellular spheroids RNA was thawed from -80 C on ice, and volumes containing 1µg of total RNA were used for cdna synthesis using the Verso cdna synthesis Kit (Thermo Scientific, PA) following the manufacturer's protocol. The cdna was then stored at -20 C until further use. 4.2.c. RT-PCR of cdna from Day 5 homo and hetero-cellular spheroids: Selection of primers For the evaluation of the expression of the inflammatory markers: IL8, TNF-α, TGF-β, the following primer sequences were selected for RT-PCR. 44

45 Gene Forward Primer Reverse Primer Annealing Temperatu re Human IL8 Human TNF alpha Human TGF beta Human Beta actin Mus muscul us IL8 Mus muscul us TNF alpha Mus muscul us TGF beta Mus muscul us Beta actin Produ ct Lengt h CCACCGGAAGGAACCATCTC TTCCTTGGGGTCCAGACAGA C 279 CCCAGGCAGTCAGATCATCTT C AGCTGCCCCTCAGCTTGA C 85 GGCAGTGGTTGAGCCGTGGA TGTTGGACAGCTGCTCCACCT C 531 CCTTTGCCGATCCGCCG CACCTCAAGAACATCCAGAGC T GGCAGGTCTACTTTGGAGTCA TTGC AACATGATCTGGGTCATCTTC TCGC CAAGCAGAACTGAACTACCAT CG ACATTCGAGGCTCCAGTGAAT TCGG 59.5 C C C 300 CGCCATCTATGAGAAAACC GTAACGCCAGGAATTGT 51.4 C 190 TGGAATCCTGTGGCATCCATG AAAC TAAAACGCAGCTCAGTAACAG TCCG 59.6 C 349 Table 1: Human and Mouse Primer sequences for IL8, TNF alpha, TGF beta and Beta actin RT-PCR of cdna from Day 5 homo-cellular Panc1 spheroids, 1:1 Panc1: J774.A1, 1:1 Panc1: NIH3T3 and 1:1:1 Panc1: J774.A1: NIH3T3 hetero-cellular spheroids RT-PCR was carried out using the Platinum PCR SuperMix (Life Technologies, NY) and using the human primers with β-actin as the housekeeping gene for the homo-cellular Day 5 Panc1 spheroids. RT-PCR for the hetero-cellular Day 5 spheroids was carried out using both human and mouse primers with β-actin as the housekeeping gene. 45

46 Thermo cycler (BioRad T100, Hercules, CA) was used to run RT-PCR with 94 C for 2 minutes to denature the template and activate the enzyme, followed by 35 cycles of PCR amplification as shown below: Cycle Time Denature 94 C for 30 seconds Anneal Gradient for 30 seconds Extend 72 C for 1 minute Table 2: RT- PCR amplification conditions The PCR products were then stored at -20 C. 4.2.d. Gel Electrophoresis to observe expression levels of the inflammatory cytokines IL8, TNFα, TGF-β in the different spheroid models A 2% Agarose E-Gel with SYBR Safe (Life Technologies, NY) was used to run 15μl of each of the PCR products. 20μl of the GeneRuler 100bp Ladder (Thermo Scientific, PA) was used to detect size of bands obtained. The bands thus obtained were observed using the ChemiDoc TM XRS imaging system (Bio-Rad, Hercules, CA). The intensity of the bands obtained for the human genes of IL8, TNF-α and TGF-β were expressed as a percentage of the human β-actin band intensity for the respective samples. The intensity of the bands obtained for the Mus musculus genes of IL8, TNF-α and TGF-β were expressed as a percentage of the murine β-actin band intensity for the respective samples. In order to understand the trends in expression of these markers between the four spheroid models, the 1-tailed student test was done for 3 batches of evaluation and results were recorded. 4.2.e. Analysis of LDH-A expression in homo- and hetero-cellular spheroids The expression of LDH-A, a marker for detecting presence of cancer stem cells (CSCs) was evaluated using RT-PCR to amplify the LDH-A gene and β- Actin in Panc1 cells grown in normoxic conditions, Panc1 cells in hypoxic conditions for 5 days and homo- and hetero-cellular Day 5 spheroids. Platinum PCR SuperMix (Life Technologies, NY) was used for RT-PCR carried out in Thermo cycler (BioRad T100, Hercules, CA) at 94 C for 2 minutes to denature the template and 46

47 activate the enzyme, followed by 35 cycles of PCR amplification (similar to RT-PCR for inflammatory markers). The following pair of primers was used to amplify the LDH-A gene: Gene Forward Primer Reverse Primer Annealing Temperature Product Length Human LDH- TGGAGTGGAATGAA ATAGCCCAGGATG 55.4 C 155 A TGTTGC TGTAGCC Human Beta actin CCTTTGCCGATCCGCC G AACATGATCTGGGTC ATCTTCTCGC 59.5 C 415 Table 3: Human primer sequences for LDH-A and Beta actin 15μl of the PCR products were run on 2% E-Gel with SYBR Safe (Life Technologies, NY) with 20μl of the GeneRuler 100bp Ladder (Thermo Scientific, PA) to detect size of bands obtained. The intensity of the LDH-A band of each condition was expressed as a percentage of the intensity of the β-actin band obtained under the same condition. 4.2.f. Selection of housekeeping gene In order to evaluate the most suitable housekeeping gene that has steady expression in both the monolayer cell culture and the spheroid model, a comparison between two commonly used housekeeping genes was performed using RT-PCR and gel electrophoresis. RNA obtained from Panc1 cells grown in normoxic conditions and that obtained from Panc1 homo-cellular spheroids was extracted. RT-PCR was carried out using primers for beta actin and 28srRNA. 15μl of each of the PCR products was run on a 2% Agarose E-Gel along with 20μl of the GeneRuler 100bp Ladder (Thermo Scientific, PA). The bands obtained were observed using the ChemiDocTM XRS imaging system. The ratio of the intensity of the beta actin band in spheroids and the intensity of the beta actin band in normoxia was compared to the ratio of the intensity of the 28srRNA band in spheroids to that in normoxia. The primers used for 28srRNA are shown below in Table 4 47

48 Gene Forward Primer Reverse Primer Annealing Temperature Product Length Human GGGTTTAGACCGTCGTGAG TCCTCAGCCAAGCACATACA C srRNA A Human Beta actin CCTTTGCCGATCCGCCG AACATGATCTGGGTCATCTTCT CGC 59.5 C 415 Table 4: Human primer sequences for 28srRNA and Beta actin 4.2.g. qpcr to quantitatively assess HIF-1α, HIF-2α, SCF and LDH-A levels in spheroids In order to quantitate the levels of HIF-1α, HIF-2α, SCF and LDH-A, qpcr analysis was carried out in normoxic cells, hypoxic cells, homo-cellular and hetero-cellular spheroids. Total RNA from all samples was extracted using the Quick-RNA MiniPrep RNA extraction kit. Concentrations of the RNA were measured using the NanoDrop (Thermo Scientific, Willington, DE), and cdna was synthesized using the Verso cdna synthesis kit. The LightCycler 480 SYBR Green I Master kit (Roche Diagnostics, Indianapolis, IN) was used to carry out qpcr along with the Roche Light Cycler 480 machine. Beta-actin was used as the loading control gene. The ΔΔCt Method was used for determining the expression of the aforementioned genes. The primer sequences used for HIF- 1α, HIF-2α, SCF and LDH-A are listed in Table 5. Gene Forward Primer Reverse Primer Annealing Temperature Product Length Human HIF- 1α GTGGTGGTTACTCAGCA CTT GGCTGTGTCGACTGAG GAAA 56.4 C 948 Human HIF- CCTCCGACTCCTTCC CGAATCTCCTCATG C 817 GACT GTCGCA 2α Human SCF GGATGACCTTGTGGA GCCCTTGTAAGACT C 417 GTGCG TGGCTG Human LDH- TGGAGTGGAATGAAT ATAGCCCAGGATGT 55.4 C 155 GTTGC GTAGCC A Human Beta actin CCTTTGCCGATCCGCCG AACATGATCTGGGTCA TCTTCTCGC 59.5 C 415 Table 5: Human primer sequences for human HIF-1α, HIF-1β, SCF, LDH-A and β-actin 48

49 4.3. Analysis of CD24+ and Stem Cell Factor expression in Day 5 homo-cellular and heterocellular spheroids through immunofluorescence: Antibody staining and Analysis of CD24+ and SCF expression Spheroids were harvested from the hanging drop plates into a 96 well Polystyrene assay plate (Corning Incorporated, NY) with one spheroid in each well. 0.2ml Cell Staining buffer (BioLegend, CA) was added to each well. The plate was left to stir for 10 minutes at 4 C (this step was done to prevent internalization of receptors). After the staining buffer was discarded carefully, 0.2ml of 4% formalin was added to each well and the plate was incubated for 1 hour at room temperature. 0.2ml of ice cold PBS was then added to each well to dilute the formalin and kept to stir for 10 minutes. The supernatant was then discarded and fresh 0.2ml of ice cold PBS was added to each well. The plate was kept overnight at 4 C. The ice cold PBS was discarded from each well and 0.1ml of 10% FBS in PBS solution (blocking buffer) was added followed by incubation at Room Temperature for 30 minutes. The blocking buffer was then discarded. 0.15ml of a 1:5 dilution of FITC Mouse Anti-Human CD24 (BD Pharmingen, CA) in 2% BSA (bovine serum albumin, Fisher, PA) in 1X PBS was added to a few wells. 0.15ml of a 1:25 dilution of the Anti- Human SCF Primary Antibody (Thermo Scientific, PA) diluted in 2% BSA (bovine serum albumin) in 1X PBS was added to the other wells. The plate was then incubated for 10 hours at 4 C with continuous stirring. The antibody solutions were then discarded and 0.2ml of ice-cold PBS was added to each well followed by 2 subsequent washing steps to remove any unbound antibody. The spheroids that were stained with the Anti-Human SCF primary antibody were then incubated with 0.2ml of the goat Anti-Rabbit IgG DyLite 594 conjugated Highly Cross Adsorbed antibody diluted in 2% BSA (bovine serum albumin, Fisher, PA) in 1X PBS for 2 hours. This was followed by two subsequent washings with ice-cold PBS. The spheroids were then stained with Hoechst Dye (Life Technologies, PA) to stain the nucleus, followed by two subsequent washes with ice-cold PBS. The spheroids were then mounted onto depression microscopic slides using Shandon Immu-Mount (Thermo Scientific, PA) and covered with a cover slip. The spheroids were observed 49

50 under confocal microscopy by excitation at 488nm for FITC, 593nm for DyLite and 405nm for Hoechst and images were captured and recorded. 50

51 4.4. Synthesis of hyaluronic acid based derivatives 4.4.a. Synthesis of hyaluronic acid polyethyleneimine conjugate 100mg of 20,000Da sodium hyaluronate obtained from Lifecore Biomedical Co. (Chaska, MN) was dissolved in 10 ml of 1M NaCl in a glass scintillation vial. 50mg of 1-Ethyl-3-(3- dimethylaminopropyl) carbodiimide (EDC) obtained from Thermo Scientific Corp.( Billerica, MA) and 50mg of sulfo-n-hydroxysuccinimide (NHS) obtained from Thermo Scientific Corp.( Billerica, MA) was added to this solution and allowed to reach for 30 minutes at RT. 15mg of 10,000Da polyethyleneimine (PEI) obtained from Polysciences Inc. (Warrington, PA) was added to this and allowed to react for 12 hours under constant stirring. This was followed by dialysis against 1M NaCl for 6 hours with a membrane of kda molecular weight cut-off (MWCO) (Spectra Por, Spectrum laboratories Inc., CA). Following this, dialysis against de-ionized water was carried out for 18 hours. The purified polymer conjugate was then frozen and subsequently freeze dried for 48 hours using the Freezone-6 lyophilizer (Labconco Inc., Kansas, MO) until a spongy white mass was obtained and stored at -20 C till use. 4.4.b. Synthesis of hyaluronic acid polyethylene glycol conjugate 10ml of 2-(N-morpholino)ethanesulfonic acid (MES) buffer was prepared and the ph was adjusted to 6 using 0.5M NaCl. 100mg of 20,000Da sodium hyaluronate obtained from Lifecore Biomedical Co. (Chaska, MN) was dissolved in 8ml of the MES buffer along with 50mg of 2,000 Da polyethylene glycol (PEG) (Laysan bio, AL). To the remaining 2ml of MES buffer, 50mg EDC and 50mg NHS were added. The two solutions were mixed and allowed to react for 12 hours under constant stirring. This was followed by dialysis against de-ionized water for 18 hours using the kda MWCO membrane. The purified polymer conjugate was then frozen and subsequently freeze-dried using the Freezone-6 lyophilizer (Labconco Inc., Kansas, MO) for 48 hours until a fluffy white mass was obtained and stored at -20 C till use. 4.4.c. Synthesis of hyaluronic acid polyethyleneimine Cyanine 5 dye conjugate 0.5 mg Sulfo-cyanine5 (cy5) NHS ester (Lumiprobe, FL) dissolved in 0.5ml 1X PBS and 10mg HA-PEI dissolved in 9.5ml 1X PBS were mixed and allowed to reach overnight at RT under constant 51

52 stirring. This was followed by dialysis against 1M NaCl followed by an overnight dialysis against de-ionized water using the kda MWCO membrane. The polymer conjugate was subsequently freeze dried using the Freezone-6 lyophilizer (Labconco Inc., Kansas, MO) for 48 hours and stored at -20 C till use. Throughout this process, care was taken to protect the sample from light. Samples from both dialysates were saved for later to measure cy5 grafting on HA-PEI. 4.4.d. Nuclear magnetic resonance (NMR) to confirm conjugation of polymers Preparation of HA-PEI samples for NMR In order to confirm conjugation of HA and PEI, the following 4 samples were prepared: (i) HA alone, (ii) PEI alone, (iii) HA and PEI added in a 100:15ratio and, (iv) HA-PEI conjugate itself. 7mg total weight of each of these polymers were added to 600µl of deuterated water and 1H- NMR spectroscopy (Varian Inc., CA) was carried out at 400 MHz. The peaks obtained were then compared to verify that new peaks indicating conjugation were seen in the sample. Preparation of HA-PEG samples for NMR Similarly, in order to confirm conjugation of HA and PEG, the following 4 samples were prepared: (i) HA alone, (ii) PEG alone, (iii) HA and PEG in a 2:1 ratio and, (iv) HA-PEG conjugate itself. 7mg total weight of each of these polymers were added to 600µl of deuterated water and 1H-NMR spectroscopy (Varian Inc., CA) was carried out at 400 MHz. Similar comparisons between controls an sample were made to verify conjugation between HA and PEG. 4.4.e. Evaluation of the efficiency of cy5-ha-pei conjugation In order to ensure conjugation between HA-PEI and the cy5 dye, 100µl of various concentrations of cy5 in de-ionized water, ranging from µg/ml (determined from expected maximum cy5 concentration in dialysates) were added to a 96-well assay plate. Along with this, 100µl of the stored dialysate solutions were also added to wells of the same 96-well assay plate. Fluorescence was measured at 633nm. Based on the fluorescence readout, a standard curve was plotted and used to determine the concentration of cy5 in the dialysates. Knowing the total volume of the dialysis solutions, the amount of cy5 in both dialysates was also 52

53 calculated and expressed as a % of cy5 that was not conjugated to the polymer. From this value, the % of cy5 conjugated to HA-PEI was determined Preparation of HA-PEI: HA-PEG nanosystems 53

54 4.5.a. Preparation of loaded HA-PEI: HA-PEG : microrna 34a mimic nanoparticles In order to prepare nanoparticles, the HA-PEI and HA-PEG conjugates were separately dissolved in de-ionized water at a concentration of 3 mg/ml. The HA-PEG solution was added to the 40µM mimic solution followed by briefly vortexing the mixture for 30 seconds. The HA-PEI solution was then added to this mixture and vortexed for 2-3 minutes. The sample was incubated at RT for 30 minutes. These volumes were selected so that the polymers were in a 50:50 w/w HA- PEI: HA-PEG ratio along with the appropriate volume of a 40µM microrna mimic mir34a (Sigma, MO) solution to give a 54:1 w/w total polymer to microrna ratio. The tube containing the sample was then sonicated for 15 seconds and passed through a 0.45 micron filter and subsequently processed for different applications. The volumes used in the formulation were scaled up if required based on the application. 4.5.b. Preparation of blank HA-PEI: HA-PEG nanoparticles In order to prepare blank nanoparticles, i.e., without microrna, the same w/w HA-PEI: HA-PEG ratio was used and the two 3 mg/ml polymer solutions were simply mixed in equal volumes followed by vortexed for 2-3 minutes, incubation at RT for 30 minutes, sonication for 15 seconds and filtration procedure. 4.5.c. Preparation of loaded cy5-ha-pei: HA-PEG : microrna 34a mimic nanoparticles These nanoparticles were prepared in the same manner as the loaded HA-PEI: HA-PEG: microrna 34a mimic nanoparticles, however the cy5 labeled HA-PEI was used instead of HA-PEI. 4.5.d. Preparation of loaded HA-PEI: microrna 34a mimic nanoparticles Loaded HA-PEI nanoparticles were prepared by vortexing a 3mg/ml HA-PEI solution with 40µM microrna mimic mir34a solution is a 27:1 w/w ratio for 2-3 minutes, followed by the same subsequent steps enlisted above. 4.5.e. Preparation of blank HA-PEI nanoparticles 54

55 Blank HA-PEI nanoparticles were prepared by vortexing a 3mg/ml HA-PEI solution for 2-3 minutes, followed by the same subsequent steps enlisted above Characterization of HA-PEI: HA-PEG nanosystems: 55

56 4.6.a. Determination of size and charge of HA-PEI: HA-PEG nanosystems through dynamic light scattering The average hydrodynamic diameter of the HA-PEI: HA-PEG nanosystems were characterized by preparing a 20X dilution of the nanoparticle preparations in 1X PBS. The size and polydispersity index of these samples were then analyzed using the dynamic light scattering (DLS) instrument (Malvern Zetasizer Nano-S, Malvern Inc., UK). In order to characterize the surface charge of the nanosystems, the same samples used for size characterization were transferred to an electrophoretic cell to measure electrophoretic mobility of particles in dispersion. The Malvern software converts this information to surface charge and provides zeta potential values. 4.6.b. Morphological analysis of HA-PEI: HA-PEG nanosystems through transmission electron microscopy In order to better understand the morphology of the HA-PEI nanosystems, 5µl of loaded nanoparticles were loaded onto 300µm mesh copper grids with a carbon backing film. The sample was then negatively stained with 5µl of 1.5mg/ml uranyl acetate stain for two minutes, dried and viewed under the transmission electron microscope (JEOL, JEM-1000, Tokyo, Japan) and the images were captured. 4.6.c. Determination of encapsulation of microrna within the HA-PEI: HA-PEG nanosystems For a quantitative estimation of encapsulation of microrna in the loaded HA-PEI only and HA-PEI: HA-PEG nanosystems, the Quant-iT PicoGreen dsdna Assay Kit (Life Technologies, Carlsbad, CA).was employed. The PicoGreen reagent is a fluorescent molecule that binds to microrna even at extremely low levels of microrna. In order to prepare a standard curve, varying concentrations of microrna in nucleasefree water ranging from 0.01 µg/ml to 1 µg/ml were used and incubated with the PicoGreen reagent for 5 minutes at RT protected from light. After preparation of the nanoparticles, they were centrifuged and dilutions of both, the supernatant and pellet were made. These samples 56

57 were also incubated with the PicoGreen reagent at the same time as the standards. The fluorescence was measured at 520nm minutes with the KC4 software and the Synergy HT microplate reader (Bio-Tek Inc., Winooski, VT). Based on the standard curve, encapsulation efficiencies were calculated and recorded. 4.6.d. Determination of release of microrna from the HA-PEI: HA-PEG nanosystems In order to analyze the release of microrna from the nanosystems, a gel retardation assay was carried out. PAA displaces the anionic polymer and releases microrna. The effect of varying concentrations and incubation periods with PAA on release was tested by treating the nanoparticles were first treated with both 2% and 4% polyacrylic acid (PAA) for 10 minutes and 20 minutes each. The nanoparticle samples were then spun at 13,000 r.p.m. and 5µl of supernatants were withdrawn carefully and diluted up to 20µl with nuclease-free water. The PAA treated samples were run along with the ultra-low range DNA ladder (10-100bp) (Thermo Scientific, PA), free microrna and untreated blank and loaded nanoparticles on 4% E-Gel EX agarose gels (Thermo Scientific, PA). The gel was then imaged and the bands were analyzed to evaluate release efficiency. The rest of the experiment was carried out in a similar fashion as mentioned above for treatment with 2% PAA for 10 minutes. Figure 10: Schematic illustrating the experimental design of the gel retardation assay with 2% and 4% PAA treatment for 10 and 20 minutes each with appropriate controls 57

58 4.7. Uptake study 4.7.a. Uptake in cells grown in normoxic conditions: A total of 20,000 Panc1 cells were seeded in each well of a Nunc Lab-Tek II Chamber Slide System (Thermo Scientific, PA) and allowed to double in number by the next day as shown in Figure 11. Nanoparticles were prepared in the same manner as described above using the cy5- labeled HA-PEI. They were diluted to obtain a mimic concentration of 50nM and 100nM. 500µl of these nanoparticle preparations were added to each well. After incubation with the formulation, the spheroids were washed twice with 1X PBS and fixed using 4% formalin. Counterstaining of nuclei was done using Vectashield anti-fade mounting medium with DAPI (Vector Labs, Burlingame, CA) at a concentration of 1.5µg/ml. Uptake was evaluated using the confocal microscope to capture images at the following time points: 1, 3, 6 and 12 hours for 50nM concentration and 12 and 6 hour time points for 100nM. The controls included for the uptake studies were as follows: (i) Untreated cells, (ii) Cells incubated with dye solution, (iii) Cells incubated with blank nanoparticles. The incubation period for all controls was 24 hours. Figure 11: Schematic of layout of uptake study in Panc1 cells grown in normoxic conditions 4.7.b. Uptake in homo-cellular and hetero-cellular spheroids: All spheroids were grown for 5 days as described above and harvested into 24-well plates such that each well contained 4 spheroids as shown in figure 12. They were washed twice with 1X PBS. Nanoparticles were also prepared in the same manner as described above using the cy5- labeled HA-PEI and diluted to obtain a mimic concentration of 100nM. 500µl of the nanoparticle preparation was added to each well. At the end of the incubation period with the formulation, 58

59 the spheroids were washed twice with 1X PBS and fixed using 4% formalin. Counter-staining of nuclei was similarly performed using Vectashield anti-fade mounting medium with a 1.5µg/ml of DAPI-Vectashield mounting medium. Uptake was evaluated at the following time points: 1, 3, 6, 12 and 24 hours. Similar controls, as enlisted above, were used, where spheroids were used instead of cell monolayers. The incubation period for all controls was 24 hours. Figure 12: Schematic of layout of uptake study in spheroids 4.8. Transfection assay: 59

60 4.8.a. Transfection in Panc1 cells grown in normoxia In order to assess transfection in normoxic Panc1 cells, 250,000 cells were seeded in each T -25 flask. The microrna dose used was 1.4µg for 250,000 cells. Serum-free DMEM was added to each flask. Nanoparticles were prepared as described above using non-labeled HA-PEI and HA-PEG and formulation was added to each of these wells into the serum- free medium. This was followed by incubation with the treatment for 18 hours, followed by washing and addition of complete media to each flask. The spheroids were then incubated up to 48 and 96 hours from the time of treatment initiation. The controls used in this experiment were as follows: (i) Untreated cells, (ii) Cells treated with mir34a mimic only, (iii) Cells treated with mir34a mimic + Lipofectamine, (iv Cells treated with nanoparticles loaded with negative control sirna sequence, (v) Cells treated with blank nanoparticles. The incubation period for all controls was 96 hours. The study layout is as illustrated in the schematic below in Figure 13. At the end of the incubation periods, cells were washed twice with 1X PBS, transferred to an Eppendorf tube and centrifuged at 2,500 r.p.m. for 5 minutes to obtain a cell pellet. Total RNA was extracted from these cell pellets using the Quick-RNA MiniPrep RNA extraction kit (Zymo Research, CA). Concentrations of the RNA were measured, recorded and the RNA, when not in use, was stored at -80 C. Reverse transcription was carried out using TaqMan microrna assay kit (Life Technologies, Foster City, CA). The cdna obtained was then used to carry out qpcr using the TaqMan kit probes and primers for the human sequences of mir34a and U6 (Thermo Fisher, PA). U6 was used as the housekeeping gene and the ΔΔCT method was employed for the analysis. 4.8.b. Transfection in 3 in 1 hetero-cellular spheroids 60

61 In order to assess transfection in the 3-in-1spheroid models, five day old spheroids were harvested and washed with 1X PBS twice. 8 spheroids were added to each well of a 24-well plate. The microrna dose used was 5.6µg for 8 spheroids. Serum-free DMEM was added to each well Nanoparticles were prepared as described above using non-labeled HA-PEI and HA-PEG and formulation was added to each of these wells into the serum- free medium. The spheroids were then incubated with the treatment for 18 hours after which they were washed and complete media was added to each well. The spheroids were then incubated up to 48 and 96 hours from the time of treatment initiation. The controls used in this experiment were as follows: (i) Untreated spheroids, (ii) Spheroids treated with mir34a mimic only, (iii) Spheroids treated with mir34a mimic + Lipofectamine, (iv) Spheroids treated with nanoparticles loaded with negative control sirna sequence, (v) Spheroids treated with blank nanoparticles. The incubation period for all controls was 96 hours. The study layout is as illustrated in the schematic below in Figure 13. At the end of the incubation periods, spheroids were washed twice with 1X PBS, transferred to an Eppendorf tube and centrifuged at 2,500 r.p.m. for 5 minutes to obtain a cell pellet. Total RNA was extracted from these cell pellets using the Quick-RNA MiniPrep RNA extraction kit (Zymo Research, CA). Concentrations of the RNA were measured, recorded and the RNA, when not in use, was stored at -80 C. Reverse transcription was carried out using TaqMan microrna assay kit (Life Technologies, Foster City, CA). The cdna obtained was then used to carry out qpcr using the TaqMan kit probes and primers for the human sequences of mir34a and U6 (Thermo Fisher, PA). U6 was used as the housekeeping gene and the ΔΔCT method was employed for the analysis. 61

62 Figure 13: Schematic illustrating the layout of the transfection study 62

63 5. RESULTS AND DISCUSSION 5.1. Characterization of homo- and hetero-cellular spheroids: 5.1.a. Determination of spheroid diameter and thickness: Figure 14: Day 5 Panc1 homo-cellular spheroid under 10X magnification - Average Diameter = microns and Average Thickness = 135 microns Figure 15: Day 5 1:1 Panc1:J774.A1 hetero-cellular spheroid under 10X magnification - Average Diameter = microns and Thickness = 60 microns Figure 16: Day 5 1:1 Panc1: NIH/3T3 hetero-cellular spheroid under 10X magnification - Average Diameter = and Thickness = microns 63

64 Figure 17: Day 5 1:1:1 Panc1: J774.A1: NIH3T3 hetero-cellular spheroid under 10X magnification - Average Diameter microns and Average Thickness = 55 microns The Panc1: NIH/3T3 hetero-cellular spheroids and the 1:1:1 Panc1: J774.A1: NIH3T3 hetero-cellular spheroids formed the smallest spheroids. These were however more intact than the Panc1: J774.A1 hetero-cellular spheroids and the Panc1 homo-cellular spheroids. The Panc1: J774.A1 hetero-cellular spheroids were the most loosely formed spheroids and were also observed to be non-uniform in shape. 5.1.b. Arrangement of Panc1, J774.A1 & NIH3T3 Cells within Hetero-cellular Spheroids: Figure 19a Figure 19b Figure 18 and 19: Distribution of Panc1, J774.A1, NIH3T3 cells in the Day 5 1:1:1 co-culture spheroid 64

65 The Panc1 cells were stained blue, the J774.A1 cells were stained green and the NIH/3T3 cells were stained red, as seen in Figure 19b. However, on superimposing the images, no specific arrangement of the three cell lines within the spheroid was observed. 65

66 5.2. Analysis of Inflammatory, Hypoxic and Cancer Stem Cell Phenotype of Spheroids through PCR: 5.2.a. Analysis of IL8, TNF-α and TGF-β gene levels across spheroid models using RT-PCR and gel electrophoresis: Figure 20: 2% Agarose E- gels with inflammatory marker expression analysis in Day 5 Panc1 homo-cellular spheroids % expression when normalized with human beta actin Inflammatory Marker Expression in Panc1 Day 5 spheroids 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 20.85% Human IL8 4.12% 42.06% Human TNF alpha Human TGF beta Inflammatory Markers Figure 21: Relative expression of inflammatory markers in Day 5 Panc1 spheroids. All expressions were normalized to β-actin as housekeeping gene. 66

67 High levels of about 42% of human TGF beta, intermediate levels of about 21% of IL8 were seen and very low levels ~4% of TNF alpha were observed. Figure 22: 2% Agarose E- gels with inflammatory marker expression analysis in Day 5 1:1 Panc1: J774.A1 spheroids % expression when normalized with beta actin Inflammatory Marker expression in 1:1 panc1: J774.A1 day5 spheroids % 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 21.43% 13.44% Human IL8 Human TNF alpha 26.12% Human TGF beta Inflammatory Markers Figure 23: Relative expression of inflammatory markers in Day 5 1:1 Panc1: J774.A1 spheroids. All expressions were normalized to the respective β-actin as housekeeping gene. 0% Mus musculus Il % Mus musculus TNF alpha 22.82% Mus musculus TGF beta 67

68 In the co-culture spheroid model of Panc1 cells and J774.A1 cells, a decrease in Human TGF beta levels, from 42% to 26% and an increase in TNF alpha levels, from 4% to 13% when compared to the homo-cellular spheroid was seen. Negligible fluctuation with respect to IL8 levels was observed, i.e % to 21.43%. Figure 24: 2% Agarose E- gels with inflammatory marker expression analysis in Day 5 1:1 Panc1: NIH/3T3 spheroids Inflammatory Marker expression in 1:1 panc1: NIH/3T3 day5 spheroids % expression when normalized with beta actin 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 18.71% 2.50% 56.45% Human IL8 Human TNFHuman TGF alpha beta 7.73% Mus musculus Il % Mus musculus TNF alpha 13.41% Mus musculus TGF beta Figure 25: Relative expression of inflammatory markers in Day 5 1:1 Panc1: NIH/3T3 spheroids. All expressions were normalized to the respective β-actin as housekeeping gene. 68

69 In the co-culture spheroid model of Panc1 cells and NIH/3T3 cells, a increase in human TGF beta levels, from 42% to 56% and a decrease in human TNF alpha levels, from 4% to 2.5% was seen, when compared to the homo-cellular spheroid. Negligible decrease with respect to human IL8 levels was observed, i.e % to 18.71%. Figure 26: 2% Agarose E-gels with inflammatory marker expression analysis in Day 5 1:1:1 Panc1: J774.A1: NIH/3T3 spheroids Inflammatory Marker expression in 1:1:1 panc1: J774.A1: NIH/3T3 day5 spheroids % expression when normalized with beta actin 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 27.01% Human IL8 5.06% Human TNF alpha 18.28% Human TGF beta 42.55% 56.48% Mus musculus Il8 Mus musculus TNF alpha 18.93% Mus musculus TGF beta Figure 27: Relative expression of inflammatory markers in Day 5 1:1:1 Panc1: J774.A1: NIH/3T3 spheroids. All expressions were normalized to the respective β-actin as housekeeping gene. 69

70 In the 1:1:1 Panc1: J774.A1: NIH/3T3 spheroid model, a decrease in human TGF beta levels, from 42% to 18% and an increase in human IL8 levels, from 21% to 27% was seen, when compared to the homo-cellular spheroid. Negligible increase was observed for human TNF alpha levels, i.e. from 4% to 5%. Figure 28: Statistical evaluation of expression of the IL-8 gene with 1-tailed test across all spheroid models. All expressions were normalized to the respective β-actin as housekeeping gene. Figure 29: Statistical evaluation of expression of the TNF-α gene with 1-tailed test across all spheroid models. All expressions were normalized to the respective β-actin as housekeeping gene. 70

71 Figure 30: Statistical evaluation of expression of the TGF-β gene with 1-tailed test across all spheroid models. All expressions were normalized to the respective β-actin as housekeeping gene. 5.2.b. Analysis of LDH-A expression in homo-cellular and hetero-cellular spheroids: Figure 31: 2% Agarose gels with LDH-A in Panc1 cells grown in a Normoxic monolayer, hypoxic monolayer and day5 homo- and hetero-cellular spheroids 71

72 % LDH-A expression when normalized with beta actin % 90.00% 80.00% 70.00% 60.00% 50.00% 40.00% 30.00% 20.00% 10.00% 0.00% 61.35% 67.34% 61.43% 57.31% 57.48% 63.20% Fluctuation in LDH-A gene levels when expressed as a % of beta actin Figure 32: Relative expression of LDH-A in Panc1 cells grown in a Normoxic monolayer, hypoxic monolayer on day 5 and day 5 homo- and hetero-cellular spheroids. All expressions were normalized to the respective β-actin as housekeeping gene. Among the spheroids, the highest levels of LDH-A ~ 63% of LDH-A were observed in the 1:1:1 Panc1: J774.A1: NIH/3T3 spheroid, followed closely by the levels observed in the Panc1 homo-cellular spheroid ~61%. LDH-A is a glycolytic markers and highest levels are seen under hypoxic conditions. A negligible increase in the levels of LDH-A were seen from Panc1:J774.A1 spheroids to Panc1: NIH/3T3 spheroids. It was hypothesized that the more intact a spheroid is, the higher the levels of hypoxia are. In this regard, the normoxic Panc1 monolayer should have expressed the lowest levels of LDH-A, which was not observed. 72

73 5.2.c. Comparison of housekeeping genes: Beta actin and 28srRNA: Figure 33: Relative expression of Beta actin and 28srRNA in Panc1 cells grown in a Normoxic monolayer day 5 homo-cellular spheroids. In order to determine the most suitable housekeeping gene to use as a control for gene analysis, beta actin and 28srRNA expression in normoxic Panc1 cells was compared to their expression in spheroids in order to ensure that expression does not vary vastly while moving from a 2D system to a 3D system. Therefore, a ratio of band intensities in both models were calculated for both genes, i.e. beta actin band intensity in spheroids/beta actin band intensity in normoxic cells. For beta actin, a higher ratio of 0.93 was observed, whereas a lower value of 0.86 was obtained for 28srRNA indicating that 28srRNA was a less suitable choice as a housekeeping gene due to decrease in observed expression while transitioning from a 2D to 3D sample. Beta actin was therefore used as a housekeeping gene for all PCR experiments. 73

74 5.2.d. qpcr to quantitatively assess HIF-1α, HIF-2α, SCF and LDH-A levels in spheroids Figure 34: Relative expression of HIF1α across different models. P values have been calculated to test differences compared to Panc1 cells grown in normoxic conditions at significance limit of α = 0.05 Although, highest levels of HIF-1α were expected to be seen in the hypoxia model and possibly in the Panc1: NIH/3T3 spheroid due to its highly intact form, significantly high levels were seen in the Panc1: J774.A1 and 3-in-1hetero-cellular spheroids. However, this increased could be explained by the effect of TNF-α levels on HIF-1α. TNF-α is known to up-regulate HIF-1α levels. [65] We have observed highest levels of TNF-α in the Panc1: J774.A1 spheroids. Second highest levels were seen in the 3-in-1hetero-cellular spheroids, which also contain J774.A1. 74

75 Figure 35: Relative expression of HIF-2α across different models. P values have been calculated to test differences compared to Panc1 cells grown in normoxic conditions at significance limit of α = 0.05 Higher significant differences were found in expression levels of HIF-2α compared to Panc1 cells grown in normoxic conditions. HIF-2α levels generally are elevated under hypoxic conditions. Therefore, it was expected that highest levels would be seen in cells cultured in hypoxic conditions. However, this was not observed while evaluating HIF-2α levels during qpcr. This could be a result of exposure of cells to normoxic conditions while collecting the cell pellet to extract RNA. Exposure to normoxic conditions even for brief periods have strong effects on HIF-2α levels. The Panc1: NIH/3T3 spheroids are the most intact spheroids and are strongly representative of the pancreatic tumor as the tumors are densely fibrotic (up to 60%). Here, we see extremely high levels of HIF-2α. But all other spheroid models also exhibit significantly high levels of HIF-2α. We don t expect loss of hypoxia-elevated genes grown in spheroids as they are 75

76 harvested with their form intact and the hypoxia is induced due to the 3 dimensional structure, which is thereby retained. Figure 36: Relative expression of LDH-A across different models. P values have been calculated to test differences compared to Panc1 cells grown in normoxic conditions at significance limit of α = 0.05 LDH-A levels in spheroid models were significantly down-regulated in the spheroid models, with the highest level among spheroids in the 3-in-1hetero-cellular model. Levels in the Panc1:J774.A1 and Panc1: NIH/3T3 and Panc1 homo-cellular spheroids were lower than in the 3- in-1model. The Panc1:J774.A1 spheroids are not intact, and therefore hypoxia levels are downregulated in this spheroid model. Low levels of LDH-A in this model may therefore be validated. However, higher LDH-a levels were expected in the Panc1: NIH/3T3 spheroids. 76

77 Figure 37: Relative expression of SCF across different models. P values have been calculated to test differences compared to Panc1 cells grown in normoxic conditions at significance limit of α = 0.05 SCF levels in all three co-culture spheroids were significantly high, indicating a strong influence of micro-environmental components in promoting the cancer-stem-cell like phenotype. Larger increases were seen in the spheroid models containing J774.A1 cells. Indicative of a higher influence of immune cells in propagating this phenotype. This was indicated by the highest levels in the Panc1:J774.A1 spheroids, followed by the second highest levels in the 3-in-1hetero-cellular spheroids. 77

78 5.3. Analysis of CD24+ and Stem Cell Factor Expression in Day 5 Homo-cellular and Heterocellular Spheroids through Immunofluorescence 38.a. Hoechst Staining 38.b. FITC Staining 38.c.Bright field image Figure 38a, 38b and 38c: CD24+ staining in Panc1 Day 5 homo-cellular Spheroids FITC staining of day 5 homo-cellular spheroids, followed by confocal microscopy, indicated the presence of CD24 on the surface of the tumor cells, as seen in Figure 38b. 39.a. Hoechst Staining 39.b. FITC Staining 39.c.Bright field image Figure 39a, 39b and 39c: CD24+ staining in 1:1 Panc1:J774.A1 Day 5 hetero-cellular Spheroids FITC staining of day 5 1:1 Panc1: J774.A1 spheroids, followed by confocal microscopy, indicated the presence of CD24 on the surface of the tumor cells, as seen in Figure 39b. 78

79 40.a. Hoechst Staining 40.b. FITC Staining 40.c.Bright field image Figure 40a, 40b and 40c: CD24+ staining in 1:1:1 Panc1:J774.A1:NIH3T3 Day 5 hetero-cellular Spheroids FITC staining of day 5 1:1:1 Panc1: J774.A1: NIH/3T3 spheroids, followed by confocal microscopy, indicated the presence of CD24 on the surface of the tumor cells, as seen in Figure 40b. 41.a.Hoechst Staining 41.b.DyLite Staining 41.c.Bright field image Figure 41a, 41b and 41c: Stem Cell Factor staining in Panc1 Day 5 homo-cellular Spheroids Dylite staining of day 5 homo-cellular spheroids, followed by confocal microscopy, indicated the presence of SCF on the surface of the tumor cells, as seen in Figure 41b. 79

80 42.a.Hoechst Staining 42.b.DyLite Staining 42.c.Bright field image Figure 42a, 42b and 42c: Stem Cell Factor staining in 1:1 Panc1:J774.A1 Day 5 hetero-cellular Spheroids Dylite staining of day 5 1:1 Panc1:J774.A1 spheroids, followed by confocal microscopy, indicated the presence of SCF on the surface of the tumor cells, as seen in Figure 42b. 43.a.Hoechst Staining 43.b.DyLite Staining 43.c.Bright field image Figure 43a, 43b and 43c: Stem Cell Factor staining in 1:1:1 Panc1:J774.A1:NIH3T3 Day 5 heterocellular Spheroids Dylite staining of day 5 1:1:1 Panc1:J774.A1: NIH/3T3 spheroids, followed by confocal microscopy, indicated the presence of SCF on the surface of the tumor cells, as seen in Figure 42b. 80

81 5.4. Evaluation of synthesized polymers: 5.4.a. Nuclear magnetic resonance (NMR) to evaluate conjugation of polymers: Figure 44: NMR spectra of the HA polymer Figure 45: NMR spectra of the PEI polymer 81

82 Figure 46: NMR spectra of the HA polymer and PEI polymer in deuterated water Figure 47: NMR spectra of the HA-PEI conjugate 82

83 NMR was carried out on HA, PEI, HA + PEI and HA-PEI conjugate in deuterated water. When compared with each other, the NMR spectra of the HA-PEI conjugate showed newly formed peaks that were absent in the aforementioned controls, indicating successful conjugation of PEI on the HA backbone. Figure 48: NMR spectra of the 2kDa PEG polymer Similarly, the NMR was carried out on HA, 2kDa PEG, HA +PEG and the HA-PEG conjugate in deuterated water, showed newly formed peaks in the NMR spectra of the HA-PEI conjugate that were absent in the aforementioned controls, indicating successful conjugation of PEG on the HA backbone. 83

84 Figure 49: NMR spectra of the HA polymer and 2kDa PEG polymer in deuterated water Figure 50: NMR spectra of the HA-2kDa PEG conjugate 84

85 Fluorescence 5.4.b. Evaluation of the efficiency of cy5-ha-pei conjugation 1400 Standard Curve y = 80583x R² = Concentration in µg/ml Figure 51: Standard curve plot of fluorescence values of various standards of cy5 in 1X PBS From the equation of the line, the following concentrations of cy5 in dialysates were obtained: Concentration of cy5 in 1M NaCl dialysate = µg/ml Concentration of cy5 in de-ionized water dialysate = µg/ml Total volume of 1M NaCl dialysate = 3000 ml. Therefore, total cy5 in 300ml of 1M NaCl dialysate = 33.66µg Total volume of de-ionized water dialysate = 3000 ml Therefore, total cy5 in 3000ml of de-ionized water dialysate = 7.68µg Total cy5 in both dialysates = 33.66µg µg = 41.34µg Total cy5 input into conjugation reaction = 0.5mg = 500µg % Cy5 not conjugated = µg/500µg X 100 = 8.26% % Cy5 conjugated = % = 91.74% Successful conjugation of cy5 to HA-PEI was observed through this experiment along with the extent of conjugation. 85

86 5.5. Evaluation of size and charge of HA-PEI: HA-PEG nanosystems through Dynamic Light Scattering: 5.5.a. Determination of size of loaded and blank HA-PEI: HA-PEG nanoparticles Figure 52: Size distribution profile of blank HA-PEI: HA-PEG nanoparticles prepared in a 50:50 w/w HA-PEI: HA-PEG ratio Figure 53: Size distribution profile of HA-PEI: HA-PEG nanoparticles loaded with mir34a mimic prepared in a 50:50 w/w HA-PEI: HA-PEG ratio and a 27:27:1 w/w HA-PEI: HA-PEG: mir34a ratio 86

87 Mean size of blank HA-PEI: HA-PEG nanoparticles was found to be 140.5nm with a PDI of while that of loaded nanoparticles was 116.6nm with a PDI of This decrease in size can be attributed to the electrostatic forces of attraction between the positively charged PEI moieties on the HA backbone and the negatively charged microrna molecules that are absent in the blank nanoparticles. It is therefore postulated that the loaded nanoparticles are more intact and smaller in size. Nanoparticle Type (n=3) Size in nm Poly Dispersity Index Blank HA-PEI: HA-PEG nanoparticles ± Loaded HA-PEI: HA-PEG: mir34a mimic nanoparticles ± Table 6a: Hydrodynamic diameter of blank and loaded HA-PEI: HA-PEG nanoparticles prepared in a 50:50 w/w HA-PEI: HA-PEG ratio 5.5.b. Determination of surface charge of loaded and blank HA-PEI: HA-PEG: Nanoparticle Type Surface Charge in mv Blank HA-PEI: HA-PEG nanoparticles ± 1.68 Loaded HA-PEI: HA-PEG: mir34a mimic nanoparticles ± 0.72 Table 6b: Surface charge of blank and loaded HA-PEI: HA-PEG nanoparticles prepared in a 50:50 w/w HA-PEI: HA-PEG ratio Upon comparing the surface charge of the loaded and blank nanoparticles, a large difference was seen as shown in Table 6.b.. The surface charge was much higher in loaded nanoparticles. Since surface charge of nanoparticles will affect cellular uptake due to the cell membranes being negatively charged, this increase in surface charge is preferable to ensure better uptake and lesser charge-charge repulsion at the cell membrane. 87

88 5.5.c. Analysis of nanoparticle morphology using TEM: Figure 54a: TEM images of loaded HA-PEI nanoparticles TEM analysis of loaded HA- PEI nanoparticles revealed spherical morphology of the HA- PEI-microRNA nanosystems with the mean diameter concurring with dynamic light scattering data. Negative staining of nanoparticles with uranyl acetate showed dark boundaries of the nanoparticles with brighter inner cores. 88

89 5.5.d. Analysis of encapsulation and release in the HA-based nanosystems: 5.5.d.a. Gel retardation assay to determine release of microrna from the HA-PEI: HA-PEG nanosystems: Figure 55a: 4% Agarose EX gel showing released bands of microrna mimic from the HA-PEI: HA-PEG nanosystems with 2% and 4% PAA for 10 and 20 minutes each Band intensity for mir34a mimic alone = Band intensity for mir34a mimic released from loaded nanoparticles lysed with 2% PAA for 10 minutes = % release of mir34a from HA-PEI: HA-PEG nanosystem lysed with 2% PAA for 10 minutes = 74% Band intensity for mir34a mimic released from loaded nanoparticles lysed with 2% PAA for 20 minutes = % release of mir34a from HA-PEI: HA-PEG nanosystem lysed with 2% PAA for 20 minutes = 102% Band intensity for mir34a mimic released from loaded nanoparticles lysed with 4% PAA for 10 minutes =

90 % release of mir34a from HA-PEI: HA-PEG nanosystem lysed with 4% PAA for 10 minutes = 79% Band intensity for mir34a mimic released from loaded nanoparticles lysed with 4% PAA for 20 minutes = % release of mir34a from HA-PEI: HA-PEG nanosystem lysed with 4% PAA for 20 minutes = 79% Figure 55b: Bar plot showing changes in release of microrna mimic from nanosystems with different PAA treatments As indicated in Figure 56, treatment with different PAA concentrations showed a modest effect in that a 5% increase is seen at the 10 minute treatment with 4% PAA compared to 2% PAA. However, it is seen that treatment with 2% PAA for 20 minutes seems to be the most efficient system to release microrna from the nano-system, indicating that lower concentration of PAA for higher incubation periods ensures higher release. 90

91 5.5.d.b. Evaluation of encapsulation of microrna within HA- based nanosystems: Evaluation of release from HA-PEI:Mir34a nanoparticles The plot of fluorescence v/s various pico green standards is shown below in Figure 56: Figure 56: Standard curve of fluorescence v/s concentration of all Pico green reagent standards From the equation of the line, and based on the fluorescence of nanoparticle samples, encapsulation efficiency of HA-PEI systems was calculated and found to be 117%. Since this experiment was carried out with very small volumes of nanoparticles, it is likely that there was a higher error rate. 91

92 5.6. Uptake study 5.6.a. Uptake in normoxic cells: DAPI Cy5 Merged Figure 57a: Untreated Panc1 cells at 12 hours of incubation with 1X PBS It can clearly been seen that no cy5 signal is observed in untreated Panc1 cells, as was expected. DAPI counterstaining in all images from here on also confirms the presence of cellular content in samples. DAPI Cy5 Merged Figure 57b: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 50nM sirna concentration in Panc1 cells treated for 1 hour After 1 hour of incubation with cy5 labeled nanosystems, very little uptake was seen in the Panc1 cells as seen in the second window of Figure 57b. DAPI Cy5 Merged 92

93 Figure 57c: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 50nM sirna concentration in Panc1 cells treated for 3 hours DAPI Cy5 Merged Figure 57d: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 50nM sirna concentration in Panc1 cells treated for 6 hours A slight increase in nanoparticle uptake was seen between 3 to 6 hours at 50nM concentration. DAPI Cy5 Merged Figure 57e: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 100nM sirna concentration in Panc1 cells treated for 6 hours Figure 57f: Uptake of HA-PEI: HA-PEG: negative control sirna nanoparticles at 100nM sirna concentration in Panc1 cells treated for 12 hours A large increase in uptake was seen from the 50nM and 100nM doses. An increase was also seen from the 6 to 12 hour time points. Since higher uptake was seen with the higher concentration of 100nM, this dose was selected as the optimal dose for carrying out uptake in spheroids. Since uptake was also elevated at longer periods of incubation, a 24-hour time point 93

94 was also included in the study in spheroids to investigate if uptake further increases beyond 12 hours. Since the spheroid s 3 dimensional structure offers a barrier to drug delivery, including the 24 hour time point was also deemed fit. 94

95 5.6.b Uptake in spheroid models: 5.6.b.i. Uptake in Panc1 homo-cellular spheroids: DAPI Cy5 Merged Figure 58a: Untreated Panc1 homo-cellular spheroids incubated in 1X PBS for 24 hours No cy5 signal was observed in untreated Panc1 spheroids, as was expected. DAPI Cy5 Merged Figure 58b: Panc1 homo-cellular spheroids incubated in Cy5 dye in 1X PBS for 24 hours A high cy5 signal was observed in Panc1 spheroids incubated with dye solution. Since cy5 is a small molecule, it tends to diffuse easily into the spheroid. 95

96 DAPI Cy5 Merged Figure 58c: Uptake in Panc1 homo-cellular spheroids incubated in blank 50:50 w/w HA-PEI: HA- PEG nanoparticles in 1X PBS for 24 hours DAPI Cy5 Merged Figure 58d: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 1 hour 96

97 DAPI Cy5 Merged Figure 58e: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 3 hours DAPI Cy5 Merged Figure 58f: Uptake in Panc1 homo-cellular spheroids treated with loaded 50:50 w/w HA-PEI: HA-PEG with a final mir34a mimic concentration of 100nm nanoparticles in 1X PBS for 6 hours 97